initial coding in qualitative research example

Qualitative Data Coding 101

How to code qualitative data, the smart way (with examples).

By: Jenna Crosley (PhD) | Reviewed by:Dr Eunice Rautenbach | December 2020

As we’ve discussed previously , qualitative research makes use of non-numerical data – for example, words, phrases or even images and video. To analyse this kind of data, the first dragon you’ll need to slay is qualitative data coding (or just “coding” if you want to sound cool). But what exactly is coding and how do you do it?

Overview: Qualitative Data Coding

In this post, we’ll explain qualitative data coding in simple terms. Specifically, we’ll dig into:

What exactly qualitative data coding is
What different types of coding exist
How to code qualitative data (the process)
Moving from coding to qualitative analysis
Tips and tricks for quality data coding

What is qualitative data coding?

Let’s start by understanding what a code is. At the simplest level, a code is a label that describes the content of a piece of text. For example, in the sentence:

“Pigeons attacked me and stole my sandwich.”

You could use “pigeons” as a code. This code simply describes that the sentence involves pigeons.

So, building onto this, qualitative data coding is the process of creating and assigning codes to categorise data extracts. You’ll then use these codes later down the road to derive themes and patterns for your qualitative analysis (for example, thematic analysis ). Coding and analysis can take place simultaneously, but it’s important to note that coding does not necessarily involve identifying themes (depending on which textbook you’re reading, of course). Instead, it generally refers to the process of labelling and grouping similar types of data to make generating themes and analysing the data more manageable.

Makes sense? Great. But why should you bother with coding at all? Why not just look for themes from the outset? Well, coding is a way of making sure your data is valid . In other words, it helps ensure that your analysis is undertaken systematically and that other researchers can review it (in the world of research, we call this transparency). In other words, good coding is the foundation of high-quality analysis.

What are the different types of coding?

Now that we’ve got a plain-language definition of coding on the table, the next step is to understand what overarching types of coding exist – in other words, coding approaches . Let’s start with the two main approaches, inductive and deductive .

With deductive coding, you, as the researcher, begin with a set of pre-established codes and apply them to your data set (for example, a set of interview transcripts). Inductive coding on the other hand, works in reverse, as you create the set of codes based on the data itself – in other words, the codes emerge from the data. Let’s take a closer look at both.

Deductive coding 101

With deductive coding, we make use of pre-established codes, which are developed before you interact with the present data. This usually involves drawing up a set of codes based on a research question or previous research . You could also use a code set from the codebook of a previous study.

For example, if you were studying the eating habits of college students, you might have a research question along the lines of

“What foods do college students eat the most?”

As a result of this research question, you might develop a code set that includes codes such as “sushi”, “pizza”, and “burgers”.

Deductive coding allows you to approach your analysis with a very tightly focused lens and quickly identify relevant data . Of course, the downside is that you could miss out on some very valuable insights as a result of this tight, predetermined focus.

Inductive coding 101

But what about inductive coding? As we touched on earlier, this type of coding involves jumping right into the data and then developing the codes based on what you find within the data.

For example, if you were to analyse a set of open-ended interviews , you wouldn’t necessarily know which direction the conversation would flow. If a conversation begins with a discussion of cats, it may go on to include other animals too, and so you’d add these codes as you progress with your analysis. Simply put, with inductive coding, you “go with the flow” of the data.

Inductive coding is great when you’re researching something that isn’t yet well understood because the coding derived from the data helps you explore the subject. Therefore, this type of coding is usually used when researchers want to investigate new ideas or concepts , or when they want to create new theories.

A little bit of both… hybrid coding approaches

If you’ve got a set of codes you’ve derived from a research topic, literature review or a previous study (i.e. a deductive approach), but you still don’t have a rich enough set to capture the depth of your qualitative data, you can combine deductive and inductive methods – this is called a hybrid coding approach.

To adopt a hybrid approach, you’ll begin your analysis with a set of a priori codes (deductive) and then add new codes (inductive) as you work your way through the data. Essentially, the hybrid coding approach provides the best of both worlds, which is why it’s pretty common to see this in research.

Need a helping hand?

initial coding in qualitative research example

How to code qualitative data

Now that we’ve looked at the main approaches to coding, the next question you’re probably asking is “how do I actually do it?”. Let’s take a look at the coding process , step by step.

Both inductive and deductive methods of coding typically occur in two stages: initial coding and line by line coding .

In the initial coding stage, the objective is to get a general overview of the data by reading through and understanding it. If you’re using an inductive approach, this is also where you’ll develop an initial set of codes. Then, in the second stage (line by line coding), you’ll delve deeper into the data and (re)organise it according to (potentially new) codes.

Step 1 – Initial coding

The first step of the coding process is to identify the essence of the text and code it accordingly. While there are various qualitative analysis software packages available, you can just as easily code textual data using Microsoft Word’s “comments” feature.

Let’s take a look at a practical example of coding. Assume you had the following interview data from two interviewees:

What pets do you have?

I have an alpaca and three dogs.

Only one alpaca? They can die of loneliness if they don’t have a friend.

I didn’t know that! I’ll just have to get five more.

I have twenty-three bunnies. I initially only had two, I’m not sure what happened.

In the initial stage of coding, you could assign the code of “pets” or “animals”. These are just initial, fairly broad codes that you can (and will) develop and refine later. In the initial stage, broad, rough codes are fine – they’re just a starting point which you will build onto in the second stage.

While there are various analysis software packages, you can just as easily code text data using Word's "comments" feature.

How to decide which codes to use

But how exactly do you decide what codes to use when there are many ways to read and interpret any given sentence? Well, there are a few different approaches you can adopt. The main approaches to initial coding include:

In vivo coding

Process coding

Open coding

Descriptive coding

Structural coding.

Value coding

Let’s take a look at each of these:

In vivo coding

When you use in vivo coding , you make use of a participants’ own words , rather than your interpretation of the data. In other words, you use direct quotes from participants as your codes. By doing this, you’ll avoid trying to infer meaning, rather staying as close to the original phrases and words as possible.

In vivo coding is particularly useful when your data are derived from participants who speak different languages or come from different cultures. In these cases, it’s often difficult to accurately infer meaning due to linguistic or cultural differences.

For example, English speakers typically view the future as in front of them and the past as behind them. However, this isn’t the same in all cultures. Speakers of Aymara view the past as in front of them and the future as behind them. Why? Because the future is unknown, so it must be out of sight (or behind us). They know what happened in the past, so their perspective is that it’s positioned in front of them, where they can “see” it.

In a scenario like this one, it’s not possible to derive the reason for viewing the past as in front and the future as behind without knowing the Aymara culture’s perception of time. Therefore, in vivo coding is particularly useful, as it avoids interpretation errors.

Next up, there’s process coding , which makes use of action-based codes . Action-based codes are codes that indicate a movement or procedure. These actions are often indicated by gerunds (words ending in “-ing”) – for example, running, jumping or singing.

Process coding is useful as it allows you to code parts of data that aren’t necessarily spoken, but that are still imperative to understanding the meaning of the texts.

An example here would be if a participant were to say something like, “I have no idea where she is”. A sentence like this can be interpreted in many different ways depending on the context and movements of the participant. The participant could shrug their shoulders, which would indicate that they genuinely don’t know where the girl is; however, they could also wink, showing that they do actually know where the girl is.

Simply put, process coding is useful as it allows you to, in a concise manner, identify the main occurrences in a set of data and provide a dynamic account of events. For example, you may have action codes such as, “describing a panda”, “singing a song about bananas”, or “arguing with a relative”.

Descriptive coding aims to summarise extracts by using a single word or noun that encapsulates the general idea of the data. These words will typically describe the data in a highly condensed manner, which allows the researcher to quickly refer to the content.

Descriptive coding is very useful when dealing with data that appear in forms other than traditional text – i.e. video clips, sound recordings or images. For example, a descriptive code could be “food” when coding a video clip that involves a group of people discussing what they ate throughout the day, or “cooking” when coding an image showing the steps of a recipe.

Structural coding involves labelling and describing specific structural attributes of the data. Generally, it includes coding according to answers to the questions of “ who ”, “ what ”, “ where ”, and “ how ”, rather than the actual topics expressed in the data. This type of coding is useful when you want to access segments of data quickly, and it can help tremendously when you’re dealing with large data sets.

For example, if you were coding a collection of theses or dissertations (which would be quite a large data set), structural coding could be useful as you could code according to different sections within each of these documents – i.e. according to the standard dissertation structure . What-centric labels such as “hypothesis”, “literature review”, and “methodology” would help you to efficiently refer to sections and navigate without having to work through sections of data all over again.

Structural coding is also useful for data from open-ended surveys. This data may initially be difficult to code as they lack the set structure of other forms of data (such as an interview with a strict set of questions to be answered). In this case, it would useful to code sections of data that answer certain questions such as “who?”, “what?”, “where?” and “how?”.

Let’s take a look at a practical example. If we were to send out a survey asking people about their dogs, we may end up with a (highly condensed) response such as the following:

Bella is my best friend. When I’m at home I like to sit on the floor with her and roll her ball across the carpet for her to fetch and bring back to me. I love my dog.

In this set, we could code Bella as “who”, dog as “what”, home and floor as “where”, and roll her ball as “how”.

Values coding

Finally, values coding involves coding that relates to the participant’s worldviews . Typically, this type of coding focuses on excerpts that reflect the values, attitudes, and beliefs of the participants. Values coding is therefore very useful for research exploring cultural values and intrapersonal and experiences and actions.

To recap, the aim of initial coding is to understand and familiarise yourself with your data , to develop an initial code set (if you’re taking an inductive approach) and to take the first shot at coding your data . The coding approaches above allow you to arrange your data so that it’s easier to navigate during the next stage, line by line coding (we’ll get to this soon).

While these approaches can all be used individually, it’s important to remember that it’s possible, and potentially beneficial, to combine them . For example, when conducting initial coding with interviews, you could begin by using structural coding to indicate who speaks when. Then, as a next step, you could apply descriptive coding so that you can navigate to, and between, conversation topics easily.

Step 2 – Line by line coding

Once you’ve got an overall idea of our data, are comfortable navigating it and have applied some initial codes, you can move on to line by line coding. Line by line coding is pretty much exactly what it sounds like – reviewing your data, line by line, digging deeper and assigning additional codes to each line.

With line-by-line coding, the objective is to pay close attention to your data to add detail to your codes. For example, if you have a discussion of beverages and you previously just coded this as “beverages”, you could now go deeper and code more specifically, such as “coffee”, “tea”, and “orange juice”. The aim here is to scratch below the surface. This is the time to get detailed and specific so as to capture as much richness from the data as possible.

In the line-by-line coding process, it’s useful to code everything in your data, even if you don’t think you’re going to use it (you may just end up needing it!). As you go through this process, your coding will become more thorough and detailed, and you’ll have a much better understanding of your data as a result of this, which will be incredibly valuable in the analysis phase.

Moving from coding to analysis

Once you’ve completed your initial coding and line by line coding, the next step is to start your analysis . Of course, the coding process itself will get you in “analysis mode” and you’ll probably already have some insights and ideas as a result of it, so you should always keep notes of your thoughts as you work through the coding.

When it comes to qualitative data analysis, there are many different types of analyses (we discuss some of the most popular ones here ) and the type of analysis you adopt will depend heavily on your research aims, objectives and questions . Therefore, we’re not going to go down that rabbit hole here, but we’ll cover the important first steps that build the bridge from qualitative data coding to qualitative analysis.

When starting to think about your analysis, it’s useful to ask yourself the following questions to get the wheels turning:

What actions are shown in the data?
What are the aims of these interactions and excerpts? What are the participants potentially trying to achieve?
How do participants interpret what is happening, and how do they speak about it? What does their language reveal?
What are the assumptions made by the participants?
What are the participants doing? What is going on?
Why do I want to learn about this? What am I trying to find out?
Why did I include this particular excerpt? What does it represent and how?

The type of qualitative analysis you adopt will depend heavily on your research aims, objectives and research questions.

Code categorisation

Categorisation is simply the process of reviewing everything you’ve coded and then creating code categories that can be used to guide your future analysis. In other words, it’s about creating categories for your code set. Let’s take a look at a practical example.

If you were discussing different types of animals, your initial codes may be “dogs”, “llamas”, and “lions”. In the process of categorisation, you could label (categorise) these three animals as “mammals”, whereas you could categorise “flies”, “crickets”, and “beetles” as “insects”. By creating these code categories, you will be making your data more organised, as well as enriching it so that you can see new connections between different groups of codes.

Theme identification

From the coding and categorisation processes, you’ll naturally start noticing themes. Therefore, the logical next step is to identify and clearly articulate the themes in your data set. When you determine themes, you’ll take what you’ve learned from the coding and categorisation and group it all together to develop themes. This is the part of the coding process where you’ll try to draw meaning from your data, and start to produce a narrative . The nature of this narrative depends on your research aims and objectives, as well as your research questions (sounds familiar?) and the qualitative data analysis method you’ve chosen, so keep these factors front of mind as you scan for themes.

Themes help you develop a narrative in your qualitative analysis

Tips & tricks for quality coding

Before we wrap up, let’s quickly look at some general advice, tips and suggestions to ensure your qualitative data coding is top-notch.

Before you begin coding, plan out the steps you will take and the coding approach and technique(s) you will follow to avoid inconsistencies.
When adopting deductive coding, it’s useful to use a codebook from the start of the coding process. This will keep your work organised and will ensure that you don’t forget any of your codes.
Whether you’re adopting an inductive or deductive approach, keep track of the meanings of your codes and remember to revisit these as you go along.
Avoid using synonyms for codes that are similar, if not the same. This will allow you to have a more uniform and accurate coded dataset and will also help you to not get overwhelmed by your data.
While coding, make sure that you remind yourself of your aims and coding method. This will help you to avoid directional drift , which happens when coding is not kept consistent.
If you are working in a team, make sure that everyone has been trained and understands how codes need to be assigned.

Psst... there’s more!

This post was based on one of our popular Research Bootcamps . If you're working on a research project, you'll definitely want to check this out ...

You Might Also Like:

31 Comments

I appreciated the valuable information provided to accomplish the various stages of the inductive and inductive coding process. However, I would have been extremely satisfied to be appraised of the SPECIFIC STEPS to follow for: 1. Deductive coding related to the phenomenon and its features to generate the codes, categories, and themes. 2. Inductive coding related to using (a) Initial (b) Axial, and (c) Thematic procedures using transcribe data from the research questions

Thank you so much for this. Very clear and simplified discussion about qualitative data coding.

This is what I want and the way I wanted it. Thank you very much.

All of the information’s are valuable and helpful. Thank for you giving helpful information’s. Can do some article about alternative methods for continue researches during the pandemics. It is more beneficial for those struggling to continue their researchers.

Thank you for your information on coding qualitative data, this is a very important point to be known, really thank you very much.

Very useful article. Clear, articulate and easy to understand. Thanks

This is very useful. You have simplified it the way I wanted it to be! Thanks

Thank you so very much for explaining, this is quite helpful!

hello, great article! well written and easy to understand. Can you provide some of the sources in this article used for further reading purposes?

You guys are doing a great job out there . I will not realize how many students you help through your articles and post on a daily basis. I have benefited a lot from your work. this is remarkable.

Wonderful one thank you so much.

Hello, I am doing qualitative research, please assist with example of coding format.

This is an invaluable website! Thank you so very much!

Well explained and easy to follow the presentation. A big thumbs up to you. Greatly appreciate the effort 👏👏👏👏

Thank you for this clear article with examples

Thank you for the detailed explanation. I appreciate your great effort. Congrats!

Ahhhhhhhhhh! You just killed me with your explanation. Crystal clear. Two Cheers!

D0 you have primary references that was used when creating this? If so, can you share them?

Being a complete novice to the field of qualitative data analysis, your indepth analysis of the process of thematic analysis has given me better insight. Thank you so much.

Excellent summary

Thank you so much for your precise and very helpful information about coding in qualitative data.

Thanks a lot to this helpful information. You cleared the fog in my brain.

Glad to hear that!

This has been very helpful. I am excited and grateful.

I still don’t understand the coding and categorizing of qualitative research, please give an example on my research base on the state of government education infrastructure environment in PNG

Wahho, this is amazing and very educational to have come across this site.. from a little search to a wide discovery of knowledge.

Thanks I really appreciate this.

Thank you so much! Very grateful.

This was truly helpful. I have been so lost, and this simplified the process for me.

Just at the right time when I needed to distinguish between inductive and

deductive data analysis of my Focus group discussion results very helpful

Very useful across disciplines and at all levels. Thanks…

Submit a Comment Cancel reply

Your email address will not be published. Required fields are marked *

Save my name, email, and website in this browser for the next time I comment.

Print Friendly

Coding Qualitative Data: A Beginner’s How-To + Examples

When gathering feedback, whether it’s from surveys , online reviews, or social mentions , the most valuable insights usually come from free-form or open-ended responses.

Though these types of responses allow for more detailed feedback, they are also difficult to measure and analyse on a large scale. Coding qualitative data allows you to transform these unique responses into quantitative metrics that can be compared to the rest of your data set.

Read on to learn about this process.

What is Qualitative Data Coding?

Qualitative data coding is the process of assigning quantitative tags to the pieces of data. This is necessary for any type of large-scale analysis because you 1) need to have a consistent way to compare and contrast each piece of qualitative data, and 2) will be able to use tools like Excel and Google Sheets to manipulate quantitative data.

For example, if a customer writes a Yelp review stating “The atmosphere was great for a Friday night, but the food was a bit overpriced,” you can assign quantitative tags based on a scale or sentiment. We’ll get into how exactly to assign these tags in the next section.

Inductive Coding vs Deductive Coding

When deciding how you will scale and code your data, you’ll first have to choose between the inductive or deductive methods. We cover the pros and cons of each method below.

Inductive Coding

Inductive coding is when you don’t already have a set scale or measurement with which to tag the data. If you’re analysing a large amount of qualitative data for the first time, such as the first round of a customer feedback survey, then you will likely need to start with inductive coding since you don’t know exactly what you will be measuring yet.

Inductive coding can be a lengthy process, as you’ll need to comb through your data manually. Luckily, things get easier the second time around when you’re able to use deductive coding.

Deductive Coding

Deductive coding is when you already have a predetermined scale or set of tags that you want to use on your data. This is usually if you’ve already analysed a set of qualitative data with inductive reasoning and want to use the same metrics.

To continue from the example above, say you noticed in the first round that a lot of Yelp reviews mentioned the price of food, and, using inductive coding, you were able to create a scale of 1-5 to measure appetisers, entrees, and desserts.

When analysing new Yelp reviews six months later, you’ll be able to keep the same scale and tag the new responses based on deductive coding, and therefore compare the data to the first round of analysis.

3 Steps for Coding Qualitative Data From the Top-Down

For this section, we will assume that we’re using inductive coding.

1. Start with Broad Categories

The first thing you will want to do is sort your data into broad categories. Think of each of these categories as specific aspects you want to know more about.

To continue with the restaurant example, your categories could include food quality, food price, atmosphere, location, service, etc.

Or for a business in the B2B space, your categories could look something like product quality, product price, customer service, chatbot quality, etc.

2. Assign Emotions or Sentiments

The next step is to then go through each category and assign a sentiment or emotion to each piece of data. In the broadest terms, you can start with just positive emotion and negative emotion.

Remember that when using inductive coding, you’re figuring out your scale and measurements as you go, so you can always start with broad analysis and drill down deeper as you become more familiar with your data.

3. Combine Categories and Sentiments to Draw Conclusions

Once you’ve sorted your data into categories and assigned sentiments, you can start comparing the numbers and drawing conclusions.

For example, perhaps you see that out of the 500 Yelp reviews you’ve analysed, 300 fall into the food price/negative sentiment section of your data. That’s a pretty clear indication that customers think your food is too expensive, and you may see an improvement in customer retention by dropping prices.

The three steps outlined above cover just the very basics of coding qualitative data, so you can understand the theory behind the analysis.

In order to gain more detailed conclusions, you’ll likely need to dig deeper into the data by assigning more complex sentiment tags and breaking down the categories further. We cover some useful tips and a coding qualitative data example below.

4 Tips to Keep in Mind for Accurate Qualitative Data Coding

4-tips-to-keep-in-mind-for-accurate-coding

Here are some helpful reminders to keep on hand when going through the three steps outlined above.

1. Start with a Small Sample of the Data

You’ll want to start with a small sample of your data to make sure the tags you’re using will be applicable to the rest of the set. You don’t want to waste time by going through and manually tagging each piece of data, only to realise at the end that the tags you’ve been using actually aren’t accurate.

Once you’ve broken up your qualitative data into the different categories, choose 10-20% of responses in each category to tag using inductive coding.

Then, continue onto the analysis phase using just that 10-20%.

If you’re able to find takeaways and easily compare the data with that small sample size , then you can continue coding the rest of the data in that same way, adding additional tags where needed.

2. Use Numerical Scales for Deeper Analysis

Instead of just assigning positive and negative sentiments to your data points, you can break this down even further by utilising numerical scales.

Exactly how negative or how positive was the piece of feedback? In the Yelp review example from the beginning of this article, the reviewer stated that the food was “a bit overpriced.” If you’re using a scale of 1-5 to tag the category “food price,” you could tag this as a ⅗ rating.

You’ll likely need to adjust your scales as you work through your initial sample and get a clearer picture of the review landscape.

Having access to more nuanced data like this is important for making accurate decisions about your business.

If you decided to stick with just positive and negative tags, your “food price” category might end up being 50% negative, indicating that a massive change to your pricing structure is needed immediately.

But if it turns out that most of those negative reviews are actually ⅗’s and not ⅕’s, then the situation isn’t as dire as it might have appeared at first glance.

3. Remember That Each Data Point Can Contain Multiple Pieces of Information

Remember that qualitative data can have multiple sentiments and multiple categories (such as the Yelp review example mentioning both atmosphere and price), so you may need to double or even triple-sort some pieces of data.

That’s the beauty of and the struggle with handling open-ended or free-form responses.

However, these responses allow for more accurate insights into your business vs narrow multiple-choice questions.

4. Be Mindful of Having Too Many Tags

Remember, you’re able to draw conclusions from your qualitative data by combining category tags and sentiment tags.

An easy mistake for data analysis newcomers to make is to end up with so many tags that comparing them becomes impossible. This usually stems from an overabundance of caution that you’re tagging responses accurately.

For example, say you’re tagging a review that’s discussing a restaurant host’s behavior. You put it in the category “host/hostess behavior” and tag it as a ⅗ for the sentiment.

Then, you come across another review discussing a server’s behaviour that’s slightly more positive, so you tag this as “server behaviour” for the category and 3.75/5 for the sentiment.

By getting this granular, you’re going to end up with very few data points in the same category and sentiment, which defeats the purpose of coding qualitative data.

In this example, unless you’re very specifically looking at the behaviour of individual restaurant positions, you’re better off tagging both responses as “customer service” for the category and ⅗ for the sentiment for consistency’s sake.

Coding Qualitative Data Example

Below we’ll walk through an example of coding qualitative data, utilising the steps and tips detailed above.

Step 1: Read through your data and define your categories. For this example, we’ll use “customer service,” “product quality,” and “price.”

Step 2: Sort a sample of the data into the above categories. Remember that each data point can be included in multiple categories.

“This software is amazing, does exactly what I need it to [Product Quality]. However, I do wish they’d stop raising prices every year as it’s starting to get a little out of my budget [Price].”
“Love the product [Product Quality], but honestly I can’t deal with the terrible customer service anymore [Customer Service]. I’ll be shopping around for a new solution.”
“Meh, this software is okay [Product Quality] but cheaper competitors [Price] are just as good with much better customer service [Customer Service].”

Step 3: Assign sentiments to the sample. For more in-depth analysis, use a numerical scale. We’ll use 1-5 in this example, with 1 being the lowest satisfaction and 5 being the highest.

Product Quality:
“This software is amazing, does exactly what I need it to do” [5/5]
“Love the product” [5/5]
“Meh, this software is okay [⅖]
Customer Service:
“Honestly I can’t deal with the terrible customer service anymore [⅕]
“...Much better customer service,” [⅖]
“However, I do wish they’d stop raising prices every year as it’s starting to get a little out of my budget.” [⅗]
“Cheaper competitors are just as good.” [⅖]

Step 4: After confirming that the established category and sentiment tags are accurate, continue steps 1-3 for the rest of your data, adding tags where necessary.

Step 5: Identify recurring patterns using data analysis. You can combine your insights with other types of data , like demographic and psychographic customer profiles.

Step 6: Take action based on what you find! For example, you may discover that customers aged 20-30 were the most likely to provide negative feedback on your customer service team, equating to ⅖ or ⅕ on your coding scale. You may be able to conclude that younger customers need a more streamlined way to communicate with your company, perhaps through an automated chatbot service.

Step 7: Repeat this process with more specific research goals in mind to continue digging deeper into what your customers are thinking and feeling . For example, if you uncover the above insight through coding qualitative data from online reviews, you could send out a customer feedback survey specifically asking free-form questions about how your customers would feel interacting with a chatbot instead.

How AI tools help with Coding Qualitative Data

Now that you understand the work that goes into coding qualitative data, you’re probably wondering if there’s an easier solution than manually sorting through every response.

The good news is that, yes, there is. Advanced AI-backed tools are available to help companies quickly and accurately analyse qualitative data at scale, such as customer surveys and online reviews.

These tools can not only code data based on a set of rules you determine, but they can even do their own inductive coding to determine themes and create the most accurate tags as they go.

These capabilities allow business owners to make accurate decisions about their business based on actual data and free up the necessary time and employee bandwidth to act on these insights.

The infographic below gives a visual summary of how to code qualitative data and why it’s essential for businesses to learn how:

Try Chattermill today today to learn how our AI-powered software can help you make smarter business decisions.

5 learnings of (nearly) a decade working with marketplaces

Zepz, the fintech behind global remittance brands WorldRemit and Sendwave, chooses Chattermill to align their team around customer insights

What to do if you are number 2 in the market

We look at what to do if you are the number 2 brand in the market...and provide a word of caution to market leaders.

See Chattermill in action

Understand the voice of your customers in realtime with Customer Feedback Analytics from Chattermill.

Chapter 18. Data Analysis and Coding

Introduction.

Piled before you lie hundreds of pages of fieldnotes you have taken, observations you’ve made while volunteering at city hall. You also have transcripts of interviews you have conducted with the mayor and city council members. What do you do with all this data? How can you use it to answer your original research question (e.g., “How do political polarization and party membership affect local politics?”)? Before you can make sense of your data, you will have to organize and simplify it in a way that allows you to access it more deeply and thoroughly. We call this process coding . [1] Coding is the iterative process of assigning meaning to the data you have collected in order to both simplify and identify patterns. This chapter introduces you to the process of qualitative data analysis and the basic concept of coding, while the following chapter (chapter 19) will take you further into the various kinds of codes and how to use them effectively.

To those who have not yet conducted a qualitative study, the sheer amount of collected data will be a surprise. Qualitative data can be absolutely overwhelming—it may mean hundreds if not thousands of pages of interview transcripts, or fieldnotes, or retrieved documents. How do you make sense of it? Students often want very clear guidelines here, and although I try to accommodate them as much as possible, in the end, analyzing qualitative data is a bit more of an art than a science: “The process of bringing order, structure, and interpretation to a mass of collected data is messy, ambiguous, time-consuming, creative, and fascinating. It does not proceed in a linear fashion: it is not neat. At times, the researcher may feel like an eccentric and tormented artist; not to worry, this is normal” ( Marshall and Rossman 2016:214 ).

To complicate matters further, each approach (e.g., Grounded Theory, deep ethnography, phenomenology) has its own language and bag of tricks (techniques) when it comes to analysis. Grounded Theory, for example, uses in vivo coding to generate new theoretical insights that emerge from a rigorous but open approach to data analysis. Ethnographers, in contrast, are more focused on creating a rich description of the practices, behaviors, and beliefs that operate in a particular field. They are less interested in generating theory and more interested in getting the picture right, valuing verisimilitude in the presentation. And then there are some researchers who seek to account for the qualitative data using almost quantitative methods of analysis, perhaps counting and comparing the uses of certain narrative frames in media accounts of a phenomenon. Qualitative content analysis (QCA) often includes elements of counting (see chapter 17). For these researchers, having very clear hypotheses and clearly defined “variables” before beginning analysis is standard practice, whereas the same would be expressly forbidden by those researchers, like grounded theorists, taking a more emergent approach.

All that said, there are some helpful techniques to get you started, and these will be presented in this and the following chapter. As you become more of an expert yourself, you may want to read more deeply about the tradition that speaks to your research. But know that there are many excellent qualitative researchers that use what works for any given study, who take what they can from each tradition. Most of us find this permissible (but watch out for the methodological purists that exist among us).

Qualitative Data Analysis as a Long Process!

Although most of this and the following chapter will focus on coding, it is important to understand that coding is just one (very important) aspect of the long data-analysis process. We can consider seven phases of data analysis, each of which is important for moving your voluminous data into “findings” that can be reported to others. The first phase involves data organization. This might mean creating a special password-protected Dropbox folder for storing your digital files. It might mean acquiring computer-assisted qualitative data-analysis software ( CAQDAS ) and uploading all transcripts, fieldnotes, and digital files to its storage repository for eventual coding and analysis. Finding a helpful way to store your material can take a lot of time, and you need to be smart about this from the very beginning. Losing data because of poor filing systems or mislabeling is something you want to avoid. You will also want to ensure that you have procedures in place to protect the confidentiality of your interviewees and informants. Filing signed consent forms (with names) separately from transcripts and linking them through an ID number or other code that only you have access to (and store safely) are important.

Once you have all of your material safely and conveniently stored, you will need to immerse yourself in the data. The second phase consists of reading and rereading or viewing and reviewing all of your data. As you do this, you can begin to identify themes or patterns in the data, perhaps writing short memos to yourself about what you are seeing. You are not committing to anything in this third phase but rather keeping your eyes and mind open to what you see. In an actual study, you may very well still be “in the field” or collecting interviews as you do this, and what you see might push you toward either concluding your data collection or expanding so that you can follow a particular group or factor that is emerging as important. For example, you may have interviewed twelve international college students about how they are adjusting to life in the US but realized as you read your transcripts that important gender differences may exist and you have only interviewed two women (and ten men). So you go back out and make sure you have enough female respondents to check your impression that gender matters here. The seven phases do not proceed entirely linearly! It is best to think of them as recursive; conceptually, there is a path to follow, but it meanders and flows.

Coding is the activity of the fourth phase . The second part of this chapter and all of chapter 19 will focus on coding in greater detail. For now, know that coding is the primary tool for analyzing qualitative data and that its purpose is to both simplify and highlight the important elements buried in mounds of data. Coding is a rigorous and systematic process of identifying meaning, patterns, and relationships. It is a more formal extension of what you, as a conscious human being, are trained to do every day when confronting new material and experiences. The “trick” or skill is to learn how to take what you do naturally and semiconsciously in your mind and put it down on paper so it can be documented and verified and tested and refined.

At the conclusion of the coding phase, your material will be searchable, intelligible, and ready for deeper analysis. You can begin to offer interpretations based on all the work you have done so far. This fifth phase might require you to write analytic memos, beginning with short (perhaps a paragraph or two) interpretations of various aspects of the data. You might then attempt stitching together both reflective and analytical memos into longer (up to five pages) general interpretations or theories about the relationships, activities, patterns you have noted as salient.

As you do this, you may be rereading the data, or parts of the data, and reviewing your codes. It’s possible you get to this phase and decide you need to go back to the beginning. Maybe your entire research question or focus has shifted based on what you are now thinking is important. Again, the process is recursive , not linear. The sixth phase requires you to check the interpretations you have generated. Are you really seeing this relationship, or are you ignoring something important you forgot to code? As we don’t have statistical tests to check the validity of our findings as quantitative researchers do, we need to incorporate self-checks on our interpretations. Ask yourself what evidence would exist to counter your interpretation and then actively look for that evidence. Later on, if someone asks you how you know you are correct in believing your interpretation, you will be able to explain what you did to verify this. Guard yourself against accusations of “ cherry-picking ,” selecting only the data that supports your preexisting notion or expectation about what you will find. [2]

The seventh and final phase involves writing up the results of the study. Qualitative results can be written in a variety of ways for various audiences (see chapter 20). Due to the particularities of qualitative research, findings do not exist independently of their being written down. This is different for quantitative research or experimental research, where completed analyses can somewhat speak for themselves. A box of collected qualitative data remains a box of collected qualitative data without its written interpretation. Qualitative research is often evaluated on the strength of its presentation. Some traditions of qualitative inquiry, such as deep ethnography, depend on written thick descriptions, without which the research is wholly incomplete, even nonexistent. All of that practice journaling and writing memos (reflective and analytical) help develop writing skills integral to the presentation of the findings.

Remember that these are seven conceptual phases that operate in roughly this order but with a lot of meandering and recursivity throughout the process. This is very different from quantitative data analysis, which is conducted fairly linearly and processually (first you state a falsifiable research question with hypotheses, then you collect your data or acquire your data set, then you analyze the data, etc.). Things are a bit messier when conducting qualitative research. Embrace the chaos and confusion, and sort your way through the maze. Budget a lot of time for this process. Your research question might change in the middle of data collection. Don’t worry about that. The key to being nimble and flexible in qualitative research is to start thinking and continue thinking about your data, even as it is being collected. All seven phases can be started before all the data has been gathered. Data collection does not always precede data analysis. In some ways, “qualitative data collection is qualitative data analysis.… By integrating data collection and data analysis, instead of breaking them up into two distinct steps, we both enrich our insights and stave off anxiety. We all know the anxiety that builds when we put something off—the longer we put it off, the more anxious we get. If we treat data collection as this mass of work we must do before we can get started on the even bigger mass of work that is analysis, we set ourselves up for massive anxiety” ( Rubin 2021:182–183 ; emphasis added).

The Coding Stage

A code is “a word or short phrase that symbolically assigns a summative, salient, essence-capturing, and/or evocative attribute for a portion of language-based or visual data” ( Saldaña 2014:5 ). Codes can be applied to particular sections of or entire transcripts, documents, or even videos. For example, one might code a video taken of a preschooler trying to solve a puzzle as “puzzle,” or one could take the transcript of that video and highlight particular sections or portions as “arranging puzzle pieces” (a descriptive code) or “frustration” (a summative emotion-based code). If the preschooler happily shouts out, “I see it!” you can denote the code “I see it!” (this is an example of an in vivo, participant-created code). As one can see from even this short example, there are many different kinds of codes and many different strategies and techniques for coding, more of which will be discussed in detail in chapter 19. The point to remember is that coding is a rigorous systematic process—to some extent, you are always coding whenever you look at a person or try to make sense of a situation or event, but you rarely do this consciously. Coding is the process of naming what you are seeing and how you are simplifying the data so that you can make sense of it in a way that is consistent with your study and in a way that others can understand and follow and replicate. Another way of saying this is that a code is “a researcher-generated interpretation that symbolizes or translates data” ( Vogt et al. 2014:13 ).

As with qualitative data analysis generally, coding is often done recursively, meaning that you do not merely take one pass through the data to create your codes. Saldaña ( 2014 ) differentiates first-cycle coding from second-cycle coding. The goal of first-cycle coding is to “tag” or identify what emerges as important codes. Note that I said emerges—you don’t always know from the beginning what will be an important aspect of the study or not, so the coding process is really the place for you to begin making the kinds of notes necessary for future analyses. In second-cycle coding, you will want to be much more focused—no longer gathering wholly new codes but synthesizing what you have into metacodes.

You might also conceive of the coding process in four parts (figure 18.1). First, identify a representative or diverse sample set of interview transcripts (or fieldnotes or other documents). This is the group you are going to use to get a sense of what might be emerging. In my own study of career obstacles to success among first-generation and working-class persons in sociology, I might select one interview from each career stage: a graduate student, a junior faculty member, a senior faculty member.

Second, code everything (“ open coding ”). See what emerges, and don’t limit yourself in any way. You will end up with a ton of codes, many more than you will end up with, but this is an excellent way to not foreclose an interesting finding too early in the analysis. Note the importance of starting with a sample of your collected data, because otherwise, open coding all your data is, frankly, impossible and counterproductive. You will just get stuck in the weeds.

Third, pare down your coding list. Where you may have begun with fifty (or more!) codes, you probably want no more than twenty remaining. Go back through the weeds and pull out everything that does not have the potential to bloom into a nicely shaped garden. Note that you should do this before tackling all of your data . Sometimes, however, you might need to rethink the sample you chose. Let’s say that the graduate student interview brought up some interesting gender issues that were pertinent to female-identifying sociologists, but both the junior and the senior faculty members identified as male. In that case, I might read through and open code at least one other interview transcript, perhaps a female-identifying senior faculty member, before paring down my list of codes.

This is also the time to create a codebook if you are using one, a master guide to the codes you are using, including examples (see Sample Codebooks 1 and 2 ). A codebook is simply a document that lists and describes the codes you are using. It is easy to forget what you meant the first time you penciled a coded notation next to a passage, so the codebook allows you to be clear and consistent with the use of your codes. There is not one correct way to create a codebook, but generally speaking, the codebook should include (1) the code (either name or identification number or both), (2) a description of what the code signifies and when and where it should be applied, and (3) an example of the code to help clarify (2). Listing all the codes down somewhere also allows you to organize and reorganize them, which can be part of the analytical process. It is possible that your twenty remaining codes can be neatly organized into five to seven master “themes.” Codebooks can and should develop as you recursively read through and code your collected material. [3]

Fourth, using the pared-down list of codes (or codebook), read through and code all the data. I know many qualitative researchers who work without a codebook, but it is still a good practice, especially for beginners. At the very least, read through your list of codes before you begin this “ closed coding ” step so that you can minimize the chance of missing a passage or section that needs to be coded. The final step is…to do it all again. Or, at least, do closed coding (step four) again. All of this takes a great deal of time, and you should plan accordingly.

Researcher Note

People often say that qualitative research takes a lot of time. Some say this because qualitative researchers often collect their own data. This part can be time consuming, but to me, it’s the analytical process that takes the most time. I usually read every transcript twice before starting to code, then it usually takes me six rounds of coding until I’m satisfied I’ve thoroughly coded everything. Even after the coding, it usually takes me a year to figure out how to put the analysis together into a coherent argument and to figure out what language to use. Just deciding what name to use for a particular group or idea can take months. Understanding this going in can be helpful so that you know to be patient with yourself.

—Jessi Streib, author of The Power of the Past and Privilege Lost

Note that there is no magic in any of this, nor is there any single “right” way to code or any “correct” codes. What you see in the data will be prompted by your position as a researcher and your scholarly interests. Where the above codes on a preschooler solving a puzzle emerged from my own interest in puzzle solving, another researcher might focus on something wholly different. A scholar of linguistics, for example, may focus instead on the verbalizations made by the child during the discovery process, perhaps even noting particular vocalizations (incidence of grrrs and gritting of the teeth, for example). Your recording of the codes you used is the important part, as it allows other researchers to assess the reliability and validity of your analyses based on those codes. Chapter 19 will provide more details about the kinds of codes you might develop.

Saldaña ( 2014 ) lists seven “necessary personal attributes” for successful coding. To paraphrase, they are the following:

Having (or practicing) good organizational skills
Perseverance
The ability and willingness to deal with ambiguity
Flexibility
Creativity, broadly understood, which includes “the ability to think visually, to think symbolically, to think in metaphors, and to think of as many ways as possible to approach a problem” (20)
Commitment to being rigorously ethical
Having an extensive vocabulary [4]

Writing Analytic Memos during/after Coding

Coding the data you have collected is only one aspect of analyzing it. Too many beginners have coded their data and then wondered what to do next. Coding is meant to help organize your data so that you can see it more clearly, but it is not itself an analysis. Thinking about the data, reviewing the coded data, and bringing in the previous literature (here is where you use your literature review and theory) to help make sense of what you have collected are all important aspects of data analysis. Analytic memos are notes you write to yourself about the data. They can be short (a single page or even a paragraph) or long (several pages). These memos can themselves be the subject of subsequent analytic memoing as part of the recursive process that is qualitative data analysis.

Short analytic memos are written about impressions you have about the data, what is emerging, and what might be of interest later on. You can write a short memo about a particular code, for example, and why this code seems important and where it might connect to previous literature. For example, I might write a paragraph about a “cultural capital” code that I use whenever a working-class sociologist says anything about “not fitting in” with their peers (e.g., not having the right accent or hairstyle or private school background). I could then write a little bit about Bourdieu, who originated the notion of cultural capital, and try to make some connections between his definition and how I am applying it here. I can also use the memo to raise questions or doubts I have about what I am seeing (e.g., Maybe the type of school belongs somewhere else? Is this really the right code?). Later on, I can incorporate some of this writing into the theory section of my final paper or article. Here are some types of things that might form the basis of a short memo: something you want to remember, something you noticed that was new or different, a reaction you had, a suspicion or hunch that you are developing, a pattern you are noticing, any inferences you are starting to draw. Rubin ( 2021 ) advises, “Always include some quotation or excerpt from your dataset…that set you off on this idea. It’s happened to me so many times—I’ll have a really strong reaction to a piece of data, write down some insight without the original quotation or context, and then [later] have no idea what I was talking about and have no way of recreating my insight because I can’t remember what piece of data made me think this way” ( 203 ).

All CAQDAS programs include spaces for writing, generating, and storing memos. You can link a memo to a particular transcript, for example. But you can just as easily keep a notebook at hand in which you write notes to yourself, if you prefer the more tactile approach. Drawing pictures that illustrate themes and patterns you are beginning to see also works. The point is to write early and write often, as these memos are the building blocks of your eventual final product (chapter 20).

In the next chapter (chapter 19), we will go a little deeper into codes and how to use them to identify patterns and themes in your data. This chapter has given you an idea of the process of data analysis, but there is much yet to learn about the elements of that process!

Qualitative Data-Analysis Samples

The following three passages are examples of how qualitative researchers describe their data-analysis practices. The first, by Harvey, is a useful example of how data analysis can shift the original research questions. The second example, by Thai, shows multiple stages of coding and how these stages build upward to conceptual themes and theorization. The third example, by Lamont, shows a masterful use of a variety of techniques to generate theory.

Example 1: “Look Someone in the Eye” by Peter Francis Harvey ( 2022 )

I entered the field intending to study gender socialization. However, through the iterative process of writing fieldnotes, rereading them, conducting further research, and writing extensive analytic memos, my focus shifted. Abductive analysis encourages the search for unexpected findings in light of existing literature. In my early data collection, fieldnotes, and memoing, classed comportment was unmistakably prominent in both schools. I was surprised by how pervasive this bodily socialization proved to be and further surprised by the discrepancies between the two schools.…I returned to the literature to compare my empirical findings.…To further clarify patterns within my data and to aid the search for disconfirming evidence, I constructed data matrices (Miles, Huberman, and Saldaña 2013). While rereading my fieldnotes, I used ATLAS.ti to code and recode key sections (Miles et al. 2013), punctuating this process with additional analytic memos. ( 2022:1420 )

Example 2:” Policing and Symbolic Control” by Mai Thai ( 2022 )

Conventional to qualitative research, my analyses iterated between theory development and testing. Analytical memos were written throughout the data collection, and my analyses using MAXQDA software helped me develop, confirm, and challenge specific themes.…My early coding scheme which included descriptive codes (e.g., uniform inspection, college trips) and verbatim codes of the common terms used by field site participants (e.g., “never quit,” “ghetto”) led me to conceptualize valorization. Later analyses developed into thematic codes (e.g., good citizens, criminality) and process codes (e.g., valorization, criminalization), which helped refine my arguments. ( 2022:1191–1192 )

Example 3: The Dignity of Working Men by Michèle Lamont ( 2000 )

To analyze the interviews, I summarized them in a 13-page document including socio-demographic information as well as information on the boundary work of the interviewees. To facilitate comparisons, I noted some of the respondents’ answers on grids and summarized these on matrix displays using techniques suggested by Miles and Huberman for standardizing and processing qualitative data. Interviews were also analyzed one by one, with a focus on the criteria that each respondent mobilized for the evaluation of status. Moreover, I located each interviewee on several five-point scales pertaining to the most significant dimensions they used to evaluate status. I also compared individual interviewees with respondents who were similar to and different from them, both within and across samples. Finally, I classified all the transcripts thematically to perform a systematic analysis of all the important themes that appear in the interviews, approaching the latter as data against which theoretical questions can be explored. ( 2000:256–257 )

Sample Codebook 1

This is an abridged version of the codebook used to analyze qualitative responses to a question about how class affects careers in sociology. Note the use of numbers to organize the flow, supplemented by highlighting techniques (e.g., bolding) and subcoding numbers.

01. CAPS: Any reference to “capitals” in the response, even if the specific words are not used

01.1: cultural capital 01.2: social capital 01.3: economic capital

(can be mixed: “0.12”= both cultural and asocial capital; “0.23”= both social and economic)

01. CAPS: a reference to “capitals” in which the specific words are used [ bold : thus, 01.23 means that both social capital and economic capital were mentioned specifically

02. DEBT: discussion of debt

02.1: mentions personal issues around debt 02.2: discusses debt but in the abstract only (e.g., “people with debt have to worry”)

03. FirstP: how the response is positioned

03.1: neutral or abstract response 03.2: discusses self (“I”) 03.3: discusses others (“they”)

Sample Coded Passage:

* Question: What other codes jump out to you here? Shouldn’t there be a code for feelings of loneliness or alienation? What about an emotions code ?

Sample Codebook 2

This is an example that uses "word" categories only, with descriptions and examples for each code

10.6 Qualitative Coding, Analysis, and Write-up: The How to Guide

This section provides an abbreviated set of steps and directions for coding, analyzing, and writing up qualitative data, taking an inductive approach. The following material is adapted from Research Rundowns, retrieved from https://researchrundowns.com/qual/qualitative-coding-analysis/ .

Step1: Open coding

At this first level of coding, the researcher is looking for distinct concepts and categories in the data, which will form the basic units of the analysis. In other words, the researcher is breaking down the data into first level concepts, or master headings, and second-level categories, or subheadings.

Researchers often use highlighters to distinguish concepts and categories. For example, if interviewees consistently talk about teaching methods, each time an interviewee mentions teaching methods, or something related to a teaching method, the researcher uses the same colour highlight. Teaching methods would become a concept, and other things related (types, etc.) would become categories – all highlighted in the same colour. It is valuable to use different coloured highlights to distinguish each broad concept and category. At the end of this stage, the transcripts contain many different colours of highlighted text. The next step is to transfer these into a brief outline, with main headings for concepts and subheadings for categories.

Step 2: Axial (focused) coding

In open coding, the researcher is focused primarily on the text from the interviews to define concepts and categories. In axial coding, the researcher is using the concepts and categories developed in the open coding process, while re-reading the text from the interviews. This step is undertaken to confirm that the concepts and categories accurately represent interview responses.

In axial coding, the researcher explores how the concepts and categories are related. To examine the latter, you might ask: What conditions caused or influenced concepts and categories? What is/was the social/political context? What are the associated effects or consequences? For example, let us suppose that one of the concepts is Adaptive Teaching , and two of the categories are tutoring and group projects . The researcher would then ask: What conditions caused or influenced tutoring and group projects to occur? From the interview transcripts, it is apparent that participants linked this condition (being able to offer tutoring and group projects) with being enabled by a supportive principle. Consequently, an axial code might be a phrase like our principal encourages different teaching methods . This discusses the context of the concept and/or categories and suggests that the researcher may need a new category labeled “supportive environment.” Axial coding is merely a more directed approach to looking at the data, to help make sure that the researcher has identified all important aspects.

Step 3: Build a data table

Table 10.4 illustrates how to transfer the final concepts and categories into a data table. This is a very effective way to organize results and/or discussion in a research paper. While this appears to be a quick process, it requires a lot of time to do it well.

Table 10.4 Major categories and associated concept

Step 4: Analysis & write-up

Not only is Table 10.4 an effective way to organize the analysis, it is also a good approach for assisting with the data analysis write-up. The first step in the analysis process is to discuss the various categories and describe the associated concepts. As part of this process, the researcher will describe the themes created in the axial coding process (the second step).

There are a variety of ways to present the data in the write-up, including: 1) telling a story; 2) using a metaphor; 3) comparing and contrasting; 4) examining relations among concepts/variables; and 5) counting. Please note that counting should not be a stand-alone qualitative data analysis process to use when writing up the results, because it cannot convey the richness of the data that has been collected. One can certainly use counting for stating the number of participants, or how many participants spoke about a specific theme or category; however, the researcher must present a much deeper level of analysis by drawing out the words of the participants, including the use of direct quotes from the participants´ interviews to demonstrate the validity of the various themes.

Here are some links to demonstrations on other methods for coding qualitative data:

https://www.youtube.com/watch?reload=9&v=phXssQBCDls
https://www.youtube.com/watch?v=lYzhgMZii3o
http://qualisresearch.com/DownLoads/qda.pdf

When writing up the analysis, it is best to “identify” participants through a number, alphabetical letter, or pseudonym in the write-up (e.g. Participant #3 stated …). This demonstrates that you drawing data from all of the participants. Think of it this way, if you were doing quantitative analysis on data from 400 participants, you would present the data for all 400 participants, assuming they all answered a specific question. You will often see in a table of quantitative results (n=400), indicating that 400 people answered the question. This is the researcher’s way of confirming, to the reader, how many participants answered a particular research question. Assigning participant numbers, letters, or pseudonyms serves the same purpose in qualitative analysis.

Research Methods for the Social Sciences: An Introduction Copyright © 2020 by Valerie Sheppard is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License , except where otherwise noted.

Share This Book

University Library, University of Illinois at Urbana-Champaign

Qualitative Data Analysis: Coding

Atlas.ti web
R for text analysis
Microsoft Excel & spreadsheets
Other options
Planning Qual Data Analysis
Free Tools for QDA
QDA with NVivo
QDA with Atlas.ti
QDA with MAXQDA
PKM for QDA
QDA with Quirkos
Working Collaboratively
Qualitative Methods Texts
Transcription
Data organization
Example Publications

Coding Qualitative Data

Planning your coding strategy.

Coding is a qualitative data analysis strategy in which some aspect of the data is assigned a descriptive label that allows the researcher to identify related content across the data. How you decide to code - or whether to code- your data should be driven by your methodology. But there are rarely step-by-step descriptions, and you'll have to make many decisions about how to code for your own project.

Some questions to consider as you decide how to code your data:

What will you code?

What aspects of your data will you code? If you are not coding all of your available data, how will you decide which elements need to be coded? If you have recordings interviews or focus groups, or other types of multimedia data, will you create transcripts to analyze and code? Or will you code the media itself (see Farley, Duppong & Aitken, 2020 on direct coding of audio recordings rather than transcripts).

Where will your codes come from?

Depending on your methodology, your coding scheme may come from previous research and be applied to your data (deductive). Or you my try to develop codes entirely from the data, ignoring as much as possible, previous knowledge of the topic under study, to develop a scheme grounded in your data (inductive). In practice, however, many practices will fall between these two approaches.

How will you apply your codes to your data?

You may decide to use software to code your qualitative data, to re-purpose other software tools (e.g. Word or spreadsheet software) or work primarily with physical versions of your data. Qualitative software is not strictly necessary, though it does offer some advantages, like:

Codes can be easily re-labeled, merged, or split. You can also choose to apply multiple coding schemes to the same data, which means you can explore multiple ways of understanding the same data. Your analysis, then, is not limited by how often you are able to work with physical data, such as paper transcripts.
Most software programs for QDA include the ability to export and import coding schemes. This means you can create a re-use a coding scheme from a previous study, or that was developed in outside of the software, without having to manually create each code.
Some software for QDA includes the ability to directly code image, video, and audio files. This may mean saving time over creating transcripts. Or, your coding may be enhanced by access to the richness of mediated content, compared to transcripts.
Using QDA software may also allow you the ability to use auto-coding functions. You may be able to automatically code all of the statements by speaker in a focus group transcript, for example, or identify and code all of the paragraphs that include a specific phrase.

What will be coded?

Will you deploy a line-by-line coding approach, with smaller codes eventually condensed into larger categories or concepts? Or will you start with codes applied to larger segments of the text, perhaps later reviewing the examples to explore and re-code for differences between the segments?

How will you explain the coding process?

Regardless of how you approach coding, the process should be clearly communicated when you report your research, though this is not always the case (Deterding & Waters, 2021).
Carefully consider the use of phrases like "themes emerged." This phrasing implies that the themes lay passively in the data, waiting for the researcher to pluck them out. This description leaves little room for describing how the researcher "saw" the themes and decided which were relevant to the study. Ryan and Bernard (2003) offer a terrific guide to ways that you might identify themes in the data, using both your own observations as well as manipulations of the data.

How will you report the results of your coding process?

How you report your coding process should align with the methodology you've chosen. Your methodology may call for careful and consistent application of a coding scheme, with reports of inter-rater reliability and counts of how often a code appears within the data. Or you may use the codes to help develop a rich description of an experience, without needing to indicate precisely how often the code was applied.

How will you code collaboratively?

If you are working with another researcher or a team, your coding process requires careful planning and implementation. You will likely need to have regular conversations about your process, particularly if your goal is to develop and consistently apply a coding scheme across your data.

Coding Features in QDA Software Programs

Atlas.ti (Mac)
Atlas.ti (Windows)
NVivo (Windows)
NVivo (Mac)
Coding data See how to create and manage codes and apply codes to segments of the data (known as quotations in Atlas.ti).

Search and Code Using the search and code feature lets you locate and automatically code data through text search, regular expressions, Named Entity Recognition, and Sentiment Analysis.
Focus Group Coding Properly prepared focus group documents can be automatically coded by speaker.
Inter-Coder Agreement Coded text, audio, and video documents can be tested for inter-coder agreement. ICA is not available for images or PDF documents.
Quotation Reader Once you've coded data, you can view just the data that has been assigned that code.

Find Redundant Codings (Mac) This tool identifies "overlapping or embedded" quotations that have the same code, that are the result of manual coding or errors when merging project files.
Coding Data in Atlas.ti (Windows) Demonstrates how to create new codes, manage codes and applying codes to segments of the data (known as quotations in Atlas.ti)
Search and Code in Atlas.ti (Windows) You can use a text search, regular expressions, Named Entity Recognition, and Sentiment Analysis to identify and automatically code data in Atlas.ti.
Focus Group Coding in Atlas.ti (Windows) Properly prepared focus group transcripts can be automatically coded by speaker.
Inter-coder Agreement in Atlas.ti (Windows) Coded text, audio, and video documents can be tested for inter-coder agreement. ICA is not available for images or PDF documents.
Quotation Reader in Atlas.ti (Windows) Once you've coded data, you can view and export the quotations that have been assigned that code.
Find Redundant Codings in Atlas.ti (Windows) This tool identifies "overlapping or embedded" quotations that have the same code, that are the result of manual coding or errors when merging project files.
Coding in NVivo (Windows) This page includes an overview of the coding features in NVivo.
Automatic Coding in Documents in NVivo (Windows) You can use paragraph formatting styles or speaker names to automatically format documents.
Coding Comparison Query in NVivo (Windows) You can use the coding comparison feature to compare how different users have coded data in NVivo.
Review the References in a Node in NVivo (Windows) References are the term that NVivo uses for coded segments of the data. This shows you how to view references related to a code (or any node)
Text Search Queries in NVivo (Windows) Text queries let you search for specific text in your data. The results of your query can be saved as a node (a form of auto coding).
Coding Query in NVivo (Windows) Use a coding query to display references from your data for a single code or multiples of codes.
Code Files and Manage Codes in NVivo (Mac) This page offers an overview of coding features in NVivo. Note that NVivo uses the concept of a node to refer to any structure around which you organize your data. Codes are a type of node, but you may see these terms used interchangeably.
Automatic Coding in Datasets in NVivo (Mac) A dataset in NVivo is data that is in rows and columns, as in a spreadsheet. If a column is set to be codable, you can also automatically code the data. This approach could be used for coding open-ended survey data.
Text Search Query in NVivo (Mac) Use the text search query to identify relevant text in your data and automatically code references by saving as a node.
Review the References in a Node in NVivo (Mac) NVivo uses the term references to refer to data that has been assigned to a code or any node. You can use the reference view to see the data linked to a specific node or combination of nodes.
Coding Comparison Query in NVivo (Mac) Use the coding comparison query to calculate a measure of inter-rater reliability when you've worked with multiple coders.

The MAXQDA interface is the same across Mac and Windows devices.

The "Code System" in MAXQDA This section of the manual shows how to create and manage codes in MAXQDA's code system.
How to Code with MAXQDA

Display Coded Segments in the Document Browser Once you've coded a document within MAXQDA, you can choose which of those codings will appear on the document, as well as choose whether or not the text is highlighted in the color linked to the code.
Creative Coding in MAXQDA Use the creative coding feature to explore the relationships between codes in your system. If you develop a new structure to you codes that you like, you can apply the changes to your overall code scheme.
Text Search in MAXQDA Use a Text Search to identify data that matches your search terms and automatically code the results. You can choose whether to code only the matching results, the sentence the results are in, or the paragraph the results appear in.
Segment Retrieval in MAXQDA Data that has been coded is considered a segment. Segment retrieval is how you display the segments that match a code or combination of codes. You can use the activation feature to show only the segments from a document group, or that match a document variable.
Intercorder Agreement in MAXQDA MAXQDA includes the ability to compare coding between two coders on a single project.
Create Tags in Taguette Taguette uses the term tag to refer to codes. You can create single tags as well as a tag hierarchy using punctuation marks.
Highlighting in Taguette Select text with a document (a highlight) and apply tags to code data in Taguette.

Useful Resources on Coding

Deterding, N. M., & Waters, M. C. (2021). Flexible coding of in-depth interviews: A twenty-first-century approach. Sociological Methods & Research , 50 (2), 708–739. https://doi.org/10.1177/0049124118799377

Farley, J., Duppong Hurley, K., & Aitken, A. A. (2020). Monitoring implementation in program evaluation with direct audio coding. Evaluation and Program Planning , 83 , 101854. https://doi.org/10.1016/j.evalprogplan.2020.101854

Ryan, G. W., & Bernard, H. R. (2003). Techniques to identify themes. Field Methods , 15 (1), 85–109. https://doi.org/10.1177/1525822X02239569.

<< Previous: Data organization
Next: Citations >>
Last Updated: May 20, 2024 4:12 PM
URL: https://guides.library.illinois.edu/qualitative

A guide to coding qualitative research data

Last updated

12 February 2023

Reviewed by

Each time you ask open-ended and free-text questions, you'll end up with numerous free-text responses. When your qualitative data piles up, how do you sift through it to determine what customers value? And how do you turn all the gathered texts into quantifiable and actionable information related to your user's expectations and needs?

Qualitative data can offer significant insights into respondents’ attitudes and behavior. But to distill large volumes of text / conversational data into clear and insightful results can be daunting. One way to resolve this is through qualitative research coding.

Streamline data coding

Use global data tagging systems in Dovetail so everyone analyzing research is speaking the same language

What is coding in qualitative research?

This is the system of classifying and arranging qualitative data . Coding in qualitative research involves separating a phrase or word and tagging it with a code. The code describes a data group and separates the information into defined categories or themes. Using this system, researchers can find and sort related content.

They can also combine categorized data with other coded data sets for analysis, or analyze it separately. The primary goal of coding qualitative data is to change data into a consistent format in support of research and reporting.

A code can be a phrase or a word that depicts an idea or recurring theme in the data. The code’s label must be intuitive and encapsulate the essence of the researcher's observations or participants' responses. You can generate these codes using two approaches to coding qualitative data: manual coding and automated coding.

Why is it important to code qualitative data?

By coding qualitative data, it's easier to identify consistency and scale within a set of individual responses. Assigning codes to phrases and words within feedback helps capture what the feedback entails. That way, you can better analyze and understand the outcome of the entire survey.

Researchers use coding and other qualitative data analysis procedures to make data-driven decisions according to customer responses. Coding in customer feedback will help you assess natural themes in the customers’ language. With this, it's easy to interpret and analyze customer satisfaction .

How do inductive and deductive approaches to qualitative coding work?

Before you start qualitative research coding, you must decide whether you're starting with some predefined code frames, within which the data will be sorted (deductive approach). Or, you may plan to develop and evolve the codes while reviewing the qualitative data generated by the research (inductive approach). A combination of both approaches is also possible.

In most instances, a combined approach will be best. For example, researchers will have some predefined codes/themes they expect to find in the data, but will allow for a degree of discovery in the data where new themes and codes come to light.

Inductive coding

This is an exploratory method in which new data codes and themes are generated by the review of qualitative data. It initiates and generates code according to the source of the data itself. It's ideal for investigative research, in which you devise a new idea, theory, or concept.

Inductive coding is otherwise called open coding. There's no predefined code-frame within inductive coding, as all codes are generated by reviewing the raw qualitative data.

If you're adding a new code, changing a code descriptor, or dividing an existing code in half, ensure you review the wider code frame to determine whether this alteration will impact other feedback codes. Failure to do this may lead to similar responses at various points in the qualitative data, generating different codes while containing similar themes or insights.

Inductive coding is more thorough and takes longer than deductive coding, but offers a more unbiased and comprehensive overview of the themes within your data.

Deductive coding

This is a hierarchical approach to coding. In this method, you develop a codebook using your initial code frames. These frames may depend on an ongoing research theory or questions. Go over the data once again and filter data to different codes.

After generating your qualitative data, your codes must be a match for the code frame you began with. Program evaluation research could use this coding approach.

Inductive and deductive approaches

Research studies usually blend both inductive and deductive coding approaches. For instance, you may use a deductive approach for your initial set of code sets, and later use an inductive approach to generate fresh codes and recalibrate them while you review and analyze your data.

What are the practical steps for coding qualitative data?

You can code qualitative data in the following ways:

1. Conduct your first-round pass at coding qualitative data

You need to review your data and assign codes to different pieces in this step. You don't have to generate the right codes since you will iterate and evolve them ahead of the second-round coding review.

Let's look at examples of the coding methods you may use in this step.

Open coding : This involves the distilling down of qualitative data into separate, distinct coded elements.

Descriptive coding : In this method, you create a description that encapsulates the data section’s content. Your code name must be a noun or a term that describes what the qualitative data relates to.

Values coding : This technique categorizes qualitative data that relates to the participant's attitudes, beliefs, and values.

Simultaneous coding : You can apply several codes to a single piece of qualitative data using this approach.

Structural coding : In this method, you can classify different parts of your qualitative data based on a predetermined design to perform additional analysis within the design.

In Vivo coding : Use this as the initial code to represent specific phrases or single words generated via a qualitative interview (i.e., specifically what the respondent said).

Process coding : A process of coding which captures action within data. Usually, this will be in the form of gerunds ending in “ing” (e.g., running, searching, reviewing).

2. Arrange your qualitative codes into groups and subcodes

You can start organizing codes into groups once you've completed your initial round of qualitative data coding. There are several ways to arrange these groups.

You can put together codes related to one another or address the same subjects or broad concepts, under each category. Continue working with these groups and rearranging the codes until you develop a framework that aligns with your analysis.

3. Conduct more rounds of qualitative coding

Conduct more iterations of qualitative data coding to review the codes and groups you've already established. You can change the names and codes, combine codes, and re-group the work you've already done during this phase.

In contrast, the initial attempt at data coding may have been hasty and haphazard. But these coding rounds focus on re-analyzing, identifying patterns, and drawing closer to creating concepts and ideas.

Below are a few techniques for qualitative data coding that are often applied in second-round coding.

Pattern coding : To describe a pattern, you join snippets of data, similarly classified under a single umbrella code.

Thematic analysis coding : When examining qualitative data, this method helps to identify patterns or themes.

Selective coding/focused coding : You can generate finished code sets and groups using your first pass of coding.

Theoretical coding : By classifying and arranging codes, theoretical coding allows you to create a theoretical framework's hypothesis. You develop a theory using the codes and groups that have been generated from the qualitative data.

Content analysis coding : This starts with an existing theory or framework and uses qualitative data to either support or expand upon it.

Axial coding : Axial coding allows you to link different codes or groups together. You're looking for connections and linkages between the information you discovered in earlier coding iterations.

Longitudinal coding : In this method, by organizing and systematizing your existing qualitative codes and categories, it is possible to monitor and measure them over time.

Elaborative coding : This involves applying a hypothesis from past research and examining how your present codes and groups relate to it.

4. Integrate codes and groups into your concluding narrative

When you finish going through several rounds of qualitative data coding and applying different forms of coding, use the generated codes and groups to build your final conclusions. The final result of your study could be a collection of findings, theory, or a description, depending on the goal of your study.

Start outlining your hypothesis , observations , and story while citing the codes and groups that served as its foundation. Create your final study results by structuring this data.

What are the two methods of coding qualitative data?

You can carry out data coding in two ways: automatic and manual. Manual coding involves reading over each comment and manually assigning labels. You'll need to decide if you're using inductive or deductive coding.

Automatic qualitative data analysis uses a branch of computer science known as Natural Language Processing to transform text-based data into a format that computers can comprehend and assess. It's a cutting-edge area of artificial intelligence and machine learning that has the potential to alter how research and insight is designed and delivered.

Although automatic coding is faster than human coding, manual coding still has an edge due to human oversight and limitations in terms of computer power and analysis.

What are the advantages of qualitative research coding?

Here are the benefits of qualitative research coding:

Boosts validity : gives your data structure and organization to be more certain the conclusions you are drawing from it are valid

Reduces bias : minimizes interpretation biases by forcing the researcher to undertake a systematic review and analysis of the data

Represents participants well : ensures your analysis reflects the views and beliefs of your participant pool and prevents you from overrepresenting the views of any individual or group

Fosters transparency : allows for a logical and systematic assessment of your study by other academics

What are the challenges of qualitative research coding?

It would be best to consider theoretical and practical limitations while analyzing and interpreting data. Here are the challenges of qualitative research coding:

Labor-intensive: While you can use software for large-scale text management and recording, data analysis is often verified or completed manually.

Lack of reliability: Qualitative research is often criticized due to a lack of transparency and standardization in the coding and analysis process, being subject to a collection of researcher bias.

Limited generalizability : Detailed information on specific contexts is often gathered using small samples. Drawing generalizable findings is challenging even with well-constructed analysis processes as data may need to be more widely gathered to be genuinely representative of attitudes and beliefs within larger populations.

Subjectivity : It is challenging to reproduce qualitative research due to researcher bias in data analysis and interpretation. When analyzing data, the researchers make personal value judgments about what is relevant and what is not. Thus, different people may interpret the same data differently.

What are the tips for coding qualitative data?

Here are some suggestions for optimizing the value of your qualitative research now that you are familiar with the fundamentals of coding qualitative data.

Keep track of your codes using a codebook or code frame

It can be challenging to recall all your codes offhand as you code more and more data. Keeping track of your codes in a codebook or code frame will keep you organized as you analyze the data. An Excel spreadsheet or word processing document might be your codebook's basic format.

Ensure you track:

The label applied to each code and the time it was first coded or modified

An explanation of the idea or subject matter that the code relates to

Who the original coder is

Any notes on the relationship between the code and other codes in your analysis

Add new codes to your codebook as you code new data, and rearrange categories and themes as necessary.

How do you create high-quality codes?

Here are four useful tips to help you create high-quality codes.

1. Cover as many survey responses as possible

The code should be generic enough to aid your analysis while remaining general enough to apply to various comments. For instance, "product" is a general code that can apply to many replies but is also ambiguous.

Also, the specific statement, "product stops working after using it for 3 hours" is unlikely to apply to many answers. A good compromise might be "poor product quality" or "short product lifespan."

2. Avoid similarities

Having similar codes is acceptable only if they serve different objectives. While "product" and "customer service" differ from each other, "customer support" and "customer service" can be unified into a single code.

3. Take note of the positive and the negative

Establish contrasting codes to track an issue's negative and positive aspects separately. For instance, two codes to identify distinct themes would be "excellent customer service" and "poor customer service."

4. Minimize data—to a point

Try to balance having too many and too few codes in your analysis to make it as useful as possible.

What is the best way to code qualitative data?

Depending on the goal of your research, the procedure of coding qualitative data can vary. But generally, it entails:

Reading through your data

Assigning codes to selected passages

Carrying out several rounds of coding

Grouping codes into themes

Developing interpretations that result in your final research conclusions

You can begin by first coding snippets of text or data to summarize or characterize them and then add your interpretative perspective in the second round of coding.

A few techniques are more or less acceptable depending on your study’s goal; there is no right or incorrect way to code a data set.

What is an example of a code in qualitative research?

A code is, at its most basic level, a label specifying how you should read a text. The phrase, "Pigeons assaulted me and took my meal," is an illustration. You can use pigeons as a code word.

Is there coding in qualitative research?

An essential component of qualitative data analysis is coding. Coding aims to give structure to free-form data so one can systematically study it.

Should you be using a customer insights hub?

Do you want to discover previous research faster?

Do you share your research findings with others?

Do you analyze research data?

Start for free today, add your research, and get to key insights faster

Editor’s picks

Last updated: 11 January 2024

Last updated: 15 January 2024

Last updated: 17 January 2024

Last updated: 25 November 2023

Last updated: 12 May 2023

Last updated: 30 April 2024

Last updated: 13 May 2024

Latest articles

Related topics, .css-je19u9{-webkit-align-items:flex-end;-webkit-box-align:flex-end;-ms-flex-align:flex-end;align-items:flex-end;display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-flex-direction:row;-ms-flex-direction:row;flex-direction:row;-webkit-box-flex-wrap:wrap;-webkit-flex-wrap:wrap;-ms-flex-wrap:wrap;flex-wrap:wrap;-webkit-box-pack:center;-ms-flex-pack:center;-webkit-justify-content:center;justify-content:center;row-gap:0;text-align:center;max-width:671px;}@media (max-width: 1079px){.css-je19u9{max-width:400px;}.css-je19u9>span{white-space:pre;}}@media (max-width: 799px){.css-je19u9{max-width:400px;}.css-je19u9>span{white-space:pre;}} decide what to .css-1kiodld{max-height:56px;display:-webkit-box;display:-webkit-flex;display:-ms-flexbox;display:flex;-webkit-align-items:center;-webkit-box-align:center;-ms-flex-align:center;align-items:center;}@media (max-width: 1079px){.css-1kiodld{display:none;}} build next, decide what to build next.

Users report unexpectedly high data usage, especially during streaming sessions.

Users find it hard to navigate from the home page to relevant playlists in the app.

It would be great to have a sleep timer feature, especially for bedtime listening.

I need better filters to find the songs or artists I’m looking for.

Log in or sign up

Get started for free

404 Not found

Qualitative Data Coding

Saul Mcleod, PhD

Editor-in-Chief for Simply Psychology

BSc (Hons) Psychology, MRes, PhD, University of Manchester

Saul Mcleod, PhD., is a qualified psychology teacher with over 18 years of experience in further and higher education. He has been published in peer-reviewed journals, including the Journal of Clinical Psychology.

Learn about our Editorial Process

Olivia Guy-Evans, MSc

Associate Editor for Simply Psychology

BSc (Hons) Psychology, MSc Psychology of Education

Olivia Guy-Evans is a writer and associate editor for Simply Psychology. She has previously worked in healthcare and educational sectors.

Coding is the process of analyzing qualitative data (usually text) by assigning labels (codes) to chunks of data that capture their essence or meaning. It allows you to condense, organize and interpret your data.

A code is a word or brief phrase that captures the essence of why you think a particular bit of data may be useful. A good analogy is that a code describes data like a hashtag describes a tweet.

Coding is an iterative process, with researchers refining and revising their codes as their understanding of the data evolves.

The ultimate goal is to develop a coherent and meaningful coding scheme that captures the richness and complexity of the participants’ experiences and helps answer the research questions.

Step 1: Familiarize yourself with the data

Read through your data (interview transcripts, field notes, documents, etc.) several times. This process is called immersion.
Think and reflect on what may be important in the data before making any firm decisions about ideas, or potential patterns.

Step 2: Decide on your coding approach

Will you use predefined deductive codes (based on theory or prior research), or let codes emerge from the data (inductive coding)?
Will a piece of data have one code or multiple?
Will you code everything or selectively? Broader research questions may warrant coding more comprehensively.

If you decide not to code everything, it’s crucial to:

Have clear criteria for what you will and won’t code
Be transparent about your selection process in research reports
Remain open to revisiting uncoded data later in analysis

Step 3: Do a first round of coding

Go through the data and assign initial codes to chunks that stand out
Create a code name (a word or short phrase) that captures the essence of each chunk
Keep a codebook – a list of your codes with descriptions or definitions
Be open to adding, revising or combining codes as you go

Descriptive codes

In vivo coding / Semantic coding : This method uses words or short phrases directly from the participant’s own language as codes. It deals with the surface-level content, labeling what participants directly say or describe. It identifies keywords, phrases, or sentences that capture the literal content. Participant : “I was just so overwhelmed with everything.” Code : “overwhelmed”
Process coding : Uses gerunds (“-ing” words) to connote observable or conceptual action in the data. Participant : “I started by brainstorming ideas, then I narrowed them down.” Codes : “brainstorming ideas,” “narrowing down”
Open coding : A form of initial coding where the researcher remains open to any possible theoretical directions indicated by the data. Participant : “I found the class really challenging, but I learned a lot.” Codes : “challenging class,” “learning experience”
Descriptive coding : Summarizes the primary topic of a passage in a word or short phrase. Participant : “I usually study in the library because it’s quiet.” Code : “study environment”

Step 4: Review and refine codes

Look over your initial codes and see if any can be combined, split up, or revised
Ensure your code names clearly convey the meaning of the data
Check if your codes are applied consistently across the dataset
Get a second opinion from a peer or advisor if possible

Interpretive codes

Interpretive codes go beyond simple description and reflect the researcher’s understanding of the underlying meanings, experiences, or processes captured in the data.

These codes require the researcher to interpret the participants’ words and actions in light of the research questions and theoretical framework.

For example, latent coding is a type of interpretive coding which goes beyond surface meaning in data. It digs for underlying emotions, motivations, or unspoken ideas the participant might not explicitly state

Latent coding looks for subtext, interprets the “why” behind what’s said, and considers the context (e.g. cultural influences, or unconscious biases).

Example: A participant might say, “Whenever I see a spider, I feel like I’m going to pass out. It takes me back to a bad experience as a kid.” A latent code here could be “Feelings of Panic Triggered by Spiders” because it goes beyond the surface fear and explores the emotional response and potential cause.

It’s useful to ask yourself the following questions:

What are the assumptions made by the participants?
What emotions or feelings are expressed or implied in the data?
How do participants relate to or interact with others in the data?
How do the participants’ experiences or perspectives change over time?
What is surprising, unexpected, or contradictory in the data?
What is not being said or shown in the data? What are the silences or absences?

Theoretical codes

Theoretical codes are the most abstract and conceptual type of codes. They are used to link the data to existing theories or to develop new theoretical insights.

Theoretical codes often emerge later in the analysis process, as researchers begin to identify patterns and connections across the descriptive and interpretive codes.

Structural coding : Applies a content-based phrase to a segment of data that relates to a specific research question. Research question : What motivates students to succeed? Participant : “I want to make my parents proud and be the first in my family to graduate college.” Interpretive Code : “family motivation” Theoretical code : “Social identity theory”
Value coding : This method codes data according to the participants’ values, attitudes, and beliefs, representing their perspectives or worldviews. Participant : “I believe everyone deserves access to quality healthcare.” Interpretive Code : “healthcare access” (value) Theoretical code : “Distributive justice”

Pattern codes

Pattern coding is often used in the later stages of data analysis, after the researcher has thoroughly familiarized themselves with the data and identified initial descriptive and interpretive codes.

By identifying patterns and relationships across the data, pattern codes help to develop a more coherent and meaningful understanding of the phenomenon and can contribute to theory development or refinement.

For Example

Let’s say a researcher is studying the experiences of new mothers returning to work after maternity leave. They conduct interviews with several participants and initially use descriptive and interpretive codes to analyze the data. Some of these codes might include:

“Guilt about leaving baby”
“Struggle to balance work and family”
“Support from colleagues”
“Flexible work arrangements”
“Breastfeeding challenges”

As the researcher reviews the coded data, they may notice that several of these codes relate to the broader theme of “work-family conflict.”

They might create a pattern code called “Navigating work-family conflict” that pulls together the various experiences and challenges described by the participants.

Research Methodology

What Is a Focus Group?

Cross-Cultural Research Methodology In Psychology

What Is Internal Validity In Research?

Research Methodology , Statistics

What Is Face Validity In Research? Importance & How To Measure

Criterion Validity: Definition & Examples

Convergent Validity: Definition and Examples

AI & NLP
Churn & Loyalty
Customer Experience
Customer Journeys
Customer Metrics
Feedback Analysis
Product Experience
Product Updates
Sentiment Analysis
Surveys & Feedback Collection
Try Thematic

Welcome to the community

Coding Qualitative Data: How to Code Qualitative Research

How many hours have you spent sitting in front of Excel spreadsheets trying to find new insights from customer feedback?

You know that asking open-ended survey questions gives you more actionable insights than asking your customers for just a numerical Net Promoter Score (NPS) . But when you ask open-ended, free-text questions, you end up with hundreds (or even thousands) of free-text responses.

How can you turn all of that text into quantifiable, applicable information about your customers’ needs and expectations? By coding qualitative data.

Keep reading to learn:

What coding qualitative data means (and why it’s important)
Different methods of coding qualitative data
How to manually code qualitative data to find significant themes in your data

What is coding in qualitative research?

Coding is the process of labeling and organizing your qualitative data to identify different themes and the relationships between them.

When coding customer feedback , you assign labels to words or phrases that represent important (and recurring) themes in each response. These labels can be words, phrases, or numbers; we recommend using words or short phrases, since they’re easier to remember, skim, and organize.

Coding qualitative research to find common themes and concepts is part of thematic analysis . Thematic analysis extracts themes from text by analyzing the word and sentence structure.

Within the context of customer feedback, it's important to understand the many different types of qualitative feedback a business can collect, such as open-ended surveys, social media comments, reviews & more.

What is qualitative data analysis?

Qualitative data analysis is the process of examining and interpreting qualitative data to understand what it represents.

Qualitative data is defined as any non-numerical and unstructured data; when looking at customer feedback, qualitative data usually refers to any verbatim or text-based feedback such as reviews, open-ended responses in surveys , complaints, chat messages, customer interviews, case notes or social media posts

For example, NPS metric can be strictly quantitative, but when you ask customers why they gave you a rating a score, you will need qualitative data analysis methods in place to understand the comments that customers leave alongside numerical responses.

Methods of qualitative data analysis

Thematic analysis.

This refers to the uncovering of themes, by analyzing the patterns and relationships in a set of qualitative data. A theme emerges or is built when related findings appear to be meaningful and there are multiple occurences. Thematic analysis can be used by anyone to transform and organize open-ended responses, online reviews and other qualitative data into significant themes.

Content analysis:

This refers to the categorization, tagging and thematic analysis of qualitative data. Essentially content analysis is a quantification of themes, by counting the occurrence of concepts, topics or themes. Content analysis can involve combining the categories in qualitative data with quantitative data, such as behavioral data or demographic data, for deeper insights.

Narrative analysis:

Some qualitative data, such as interviews or field notes may contain a story on how someone experienced something. For example, the process of choosing a product, using it, evaluating its quality and decision to buy or not buy this product next time. The goal of narrative analysis is to turn the individual narratives into data that can be coded. This is then analyzed to understand how events or experiences had an impact on the people involved.

Discourse analysis:

This refers to analysis of what people say in social and cultural context. The goal of discourse analysis is to understand user or customer behavior by uncovering their beliefs, interests and agendas. These are reflected in the way they express their opinions, preferences and experiences. It’s particularly useful when your focus is on building or strengthening a brand , by examining how they use metaphors and rhetorical devices.

Framework analysis:

When performing qualitative data analysis, it is useful to have a framework to organize the buckets of meaning. A taxonomy or code frame (a hierarchical set of themes used in coding qualitative data) is an example of the result. Don't fall into the trap of starting with a framework to make it faster to organize your data. You should look at how themes relate to each other by analyzing the data and consistently check that you can validate that themes are related to each other .

Grounded theory:

This method of analysis starts by formulating a theory around a single data case. Therefore the theory is “grounded’ in actual data. Then additional cases can be examined to see if they are relevant and can add to the original theory.

Why is it important to code qualitative data?

Coding qualitative data makes it easier to interpret customer feedback. Assigning codes to words and phrases in each response helps capture what the response is about which, in turn, helps you better analyze and summarize the results of the entire survey.

Researchers use coding and other qualitative data analysis processes to help them make data-driven decisions based on customer feedback. When you use coding to analyze your customer feedback, you can quantify the common themes in customer language. This makes it easier to accurately interpret and analyze customer satisfaction.

What is thematic coding?

Thematic coding, also called thematic analysis, is a type of qualitative data analysis that finds themes in text by analyzing the meaning of words and sentence structure.

When you use thematic coding to analyze customer feedback for example, you can learn which themes are most frequent in feedback. This helps you understand what drives customer satisfaction in an accurate, actionable way.

To learn more about how Thematic analysis software helps you automate the data coding process, check out this article .

Automated vs. Manual coding of qualitative data

Methods of coding qualitative data fall into three categories: automated coding and manual coding, and a blend of the two.

You can automate the coding of your qualitative data with thematic analysis software . Thematic analysis and qualitative data analysis software use machine learning, artificial intelligence (AI) , and natural language processing (NLP) to code your qualitative data and break text up into themes.

Thematic analysis software is autonomous , which means…

You don’t need to set up themes or categories in advance.
You don’t need to train the algorithm — it learns on its own.
You can easily capture the “unknown unknowns” to identify themes you may not have spotted on your own.

…all of which will save you time (and lots of unnecessary headaches) when analyzing your customer feedback.

Businesses are also seeing the benefit of using thematic analysis software. The capacity to aggregate data sources into a single source of analysis helps to break down data silos, unifying the analysis and insights across departments . This is now being referred to as Omni channel analysis or Unified Data Analytics .

Use Thematic Analysis Software

Try Thematic today to discover why leading companies rely on the platform to automate the coding of qualitative customer feedback at scale. Whether you have tons of customer reviews, support chat or open-ended survey responses, Thematic brings every valuable insight to the surface, while saving you thousands of hours.

Advances in natural language processing & machine learning have made it possible to automate the analysis of qualitative data, in particular content and framework analysis. The most commonly used software for automated coding of qualitative data is text analytics software such as Thematic .

While manual human analysis is still popular due to its perceived high accuracy, automating most of the analysis is quickly becoming the preferred choice. Unlike manual analysis, which is prone to bias and doesn’t scale to the amount of qualitative data that is generated today, automating analysis is not only more consistent and therefore can be more accurate, but can also save a ton of time, and therefore money.

Our Theme Editor tool ensures you take a reflexive approach, an important step in thematic analysis. The drag-and-drop tool makes it easy to refine, validate, and rename themes as you get more data. By guiding the AI, you can ensure your results are always precise, easy to understand and perfectly aligned with your objectives.

Thematic is the best software to automate code qualitative feedback at scale.

Don't just take it from us. Here's what some of our customers have to say:

I'm a fan of Thematic's ability to save time and create heroes. It does an excellent job using a single view to break down the verbatims into themes displayed by volume, sentiment and impact on our beacon metric, often but not exclusively NPS.

It does a superlative job using GenAI in summarizing a theme or sub-theme down to a single paragraph making it clear what folks are trying to say. Peter K, Snr Research Manager.

Thematic is a very intuitive tool to use. It boasts a robust level of granularity, allowing the user to see the general breadth of verbatim themes, dig into the sub-themes, and further into the sentiment of the open text itself. Artem C, Sr Manager of Research. LinkedIn.

AI-powered software to transform qualitative data at scale through a thematic and content analysis.

How to manually code qualitative data

For the rest of this post, we’ll focus on manual coding. Different researchers have different processes, but manual coding usually looks something like this:

Choose whether you’ll use deductive or inductive coding.
Read through your data to get a sense of what it looks like. Assign your first set of codes.
Go through your data line-by-line to code as much as possible. Your codes should become more detailed at this step.
Categorize your codes and figure out how they fit into your coding frame.
Identify which themes come up the most — and act on them.

Let’s break it down a little further…

Deductive coding vs. inductive coding

Before you start qualitative data coding, you need to decide which codes you’ll use.

What is Deductive Coding?

Deductive coding means you start with a predefined set of codes, then assign those codes to the new qualitative data. These codes might come from previous research, or you might already know what themes you’re interested in analyzing. Deductive coding is also called concept-driven coding.

For example, let’s say you’re conducting a survey on customer experience . You want to understand the problems that arise from long call wait times, so you choose to make “wait time” one of your codes before you start looking at the data.

The deductive approach can save time and help guarantee that your areas of interest are coded. But you also need to be careful of bias; when you start with predefined codes, you have a bias as to what the answers will be. Make sure you don’t miss other important themes by focusing too hard on proving your own hypothesis.

What is Inductive Coding?

Inductive coding , also called open coding, starts from scratch and creates codes based on the qualitative data itself. You don’t have a set codebook; all codes arise directly from the survey responses.

Here’s how inductive coding works:

Break your qualitative dataset into smaller samples.
Read a sample of the data.
Create codes that will cover the sample.
Reread the sample and apply the codes.
Read a new sample of data, applying the codes you created for the first sample.
Note where codes don’t match or where you need additional codes.
Create new codes based on the second sample.
Go back and recode all responses again.
Repeat from step 5 until you’ve coded all of your data.

If you add a new code, split an existing code into two, or change the description of a code, make sure to review how this change will affect the coding of all responses. Otherwise, the same responses at different points in the survey could end up with different codes.

Sounds like a lot of work, right? Inductive coding is an iterative process, which means it takes longer and is more thorough than deductive coding. A major advantage is that it gives you a more complete, unbiased look at the themes throughout your data.

Combining inductive and deductive coding

In practice, most researchers use a blend of inductive and deductive approaches to coding.

For example, with Thematic, the AI inductively comes up with themes, while also framing the analysis so that it reflects how business decisions are made . At the end of the analysis, researchers use the Theme Editor to iterate or refine themes. Then, in the next wave of analysis, as new data comes in, the AI starts deductively with the theme taxonomy.

Categorize your codes with coding frames

Once you create your codes, you need to put them into a coding frame. A coding frame represents the organizational structure of the themes in your research. There are two types of coding frames: flat and hierarchical.

Flat Coding Frame

A flat coding frame assigns the same level of specificity and importance to each code. While this might feel like an easier and faster method for manual coding, it can be difficult to organize and navigate the themes and concepts as you create more and more codes. It also makes it hard to figure out which themes are most important, which can slow down decision making.

Hierarchical Coding Frame

Hierarchical frames help you organize codes based on how they relate to one another. For example, you can organize the codes based on your customers’ feelings on a certain topic:

In this example:

The top-level code describes the topic (customer service)
The mid-level code specifies whether the sentiment is positive or negative
The third level details the attribute or specific theme associated with the topic

Hierarchical framing supports a larger code frame and lets you organize codes based on organizational structure. It also allows for different levels of granularity in your coding.

Whether your code frames are hierarchical or flat, your code frames should be flexible. Manually analyzing survey data takes a lot of time and effort; make sure you can use your results in different contexts.

For example, if your survey asks customers about customer service, you might only use codes that capture answers about customer service. Then you realize that the same survey responses have a lot of comments about your company’s products. To learn more about what people say about your products, you may have to code all of the responses from scratch! A flexible coding frame covers different topics and insights, which lets you reuse the results later on.

Tips for manually coding qualitative data

Now that you know the basics of coding your qualitative data, here are some tips on making the most of your qualitative research.

Use a codebook to keep track of your codes

As you code more and more data, it can be hard to remember all of your codes off the top of your head. Tracking your codes in a codebook helps keep you organized throughout the data analysis process. Your codebook can be as simple as an Excel spreadsheet or word processor document. As you code new data, add new codes to your codebook and reorganize categories and themes as needed.

Make sure to track:

The label used for each code
A description of the concept or theme the code refers to
Who originally coded it
The date that it was originally coded or updated
Any notes on how the code relates to other codes in your analysis

How to create high-quality codes - 4 tips

1. cover as many survey responses as possible..

The code should be generic enough to apply to multiple comments, but specific enough to be useful in your analysis. For example, “Product” is a broad code that will cover a variety of responses — but it’s also pretty vague. What about the product? On the other hand, “Product stops working after using it for 3 hours” is very specific and probably won’t apply to many responses. “Poor product quality” or “short product lifespan” might be a happy medium.

2. Avoid commonalities.

Having similar codes is okay as long as they serve different purposes. “Customer service” and “Product” are different enough from one another, while “Customer service” and “Customer support” may have subtle differences but should likely be combined into one code.

3. Capture the positive and the negative.

Try to create codes that contrast with each other to track both the positive and negative elements of a topic separately. For example, “Useful product features” and “Unnecessary product features” would be two different codes to capture two different themes.

4. Reduce data — to a point.

Let’s look at the two extremes: There are as many codes as there are responses, or each code applies to every single response. In both cases, the coding exercise is pointless; you don’t learn anything new about your data or your customers. To make your analysis as useful as possible, try to find a balance between having too many and too few codes.

Group responses based on themes, not words

Make sure to group responses with the same themes under the same code, even if they don’t use the same exact wording. For example, a code such as “cleanliness” could cover responses including words and phrases like:

Looked like a dump
Could eat off the floor

Having only a few codes and hierarchical framing makes it easier to group different words and phrases under one code. If you have too many codes, especially in a flat frame, your results can become ambiguous and themes can overlap. Manual coding also requires the coder to remember or be able to find all of the relevant codes; the more codes you have, the harder it is to find the ones you need, no matter how organized your codebook is.

Make accuracy a priority

Manually coding qualitative data means that the coder’s cognitive biases can influence the coding process. For each study, make sure you have coding guidelines and training in place to keep coding reliable, consistent, and accurate .

One thing to watch out for is definitional drift, which occurs when the data at the beginning of the data set is coded differently than the material coded later. Check for definitional drift across the entire dataset and keep notes with descriptions of how the codes vary across the results.

If you have multiple coders working on one team, have them check one another’s coding to help eliminate cognitive biases.

Conclusion: 6 main takeaways for coding qualitative data

Here are 6 final takeaways for manually coding your qualitative data:

Coding is the process of labeling and organizing your qualitative data to identify themes. After you code your qualitative data, you can analyze it just like numerical data.
Inductive coding (without a predefined code frame) is more difficult, but less prone to bias, than deductive coding.
Code frames can be flat (easier and faster to use) or hierarchical (more powerful and organized).
Your code frames need to be flexible enough that you can make the most of your results and use them in different contexts.
When creating codes, make sure they cover several responses, contrast one another, and strike a balance between too much and too little information.
Consistent coding = accuracy. Establish coding procedures and guidelines and keep an eye out for definitional drift in your qualitative data analysis.

Some more detail in our downloadable guide

If you’ve made it this far, you’ll likely be interested in our free guide: Best practices for analyzing open-ended questions.

The guide includes some of the topics covered in this article, and goes into some more niche details.

If your company is looking to automate your qualitative coding process, try Thematic !

If you're looking to trial multiple solutions, check out our free buyer's guide . It covers what to look for when trialing different feedback analytics solutions to ensure you get the depth of insights you need.

Happy coding!

Authored by Alyona Medelyan, PhD – Natural Language Processing & Machine Learning

CEO and Co-Founder

Alyona has a PhD in NLP and Machine Learning. Her peer-reviewed articles have been cited by over 2600 academics. Her love of writing comes from years of PhD research.

We make it easy to discover the customer and product issues that matter.

Unlock the value of feedback at scale, in one platform. Try it for free now!

Questions to ask your Feedback Analytics vendor
How to end customer churn for good
Scalable analysis of NPS verbatims
5 Text analytics approaches
How to calculate the ROI of CX

Our experts will show you how Thematic works, how to discover pain points and track the ROI of decisions. To access your free trial, book a personal demo today.

Uncomplicated Reviews of Educational Research Methods

Qualitative Coding & Analysis

.pdf version of this page

This review is in the form of an abbreviated set of directions for initial coding and analysis. There are many ways to accomplish both actions. This approach assumes you are using interview data. For a more detailed treatment of these and related analysis concepts, click here .

Interview Data (Identifying & Coding Themes )

Open coding

At this first level of coding, you are looking for distinct concepts and categories in the data, which will form the basic units of your analysis. In other words, you are breaking down the data into first level concepts, or master headings, and second-level categories, or subheadings.

Researchers often use highlights to distinguish concepts and categories. For example, if interviewees consistently talk about teaching methods, each time an interviewee mentions teaching methods, or something related to a teaching method, you would use the same color highlight. Teaching methods would become a concept, and other things related (types, etc.) would become categories – all highlighted the same color. Use different colored highlights to distinguish each broad concept and category. What you should have at the end of this stage are transcripts with 3-5 different colors in lots of highlighted text. Transfer these into a brief outline, with concepts being main headings and categories being subheadings.

Axial coding

In open coding, you were focused primarily on the text to define concepts and categories. In axial coding, you are using your concepts and categories while re-reading the text to 1. Confirm that your concepts and categories accurately represent interview responses and, 2. Explore how your concepts and categories are related. To examine the latter, you might ask, What conditions caused or influenced concepts and categories? What is/was the social/political context? or What are the associated effects or consequences?

For example, if one of your concepts is Adaptive Teaching , and two of your categories are tutoring and group projects , an axial code might be a phrase like “our principal encourages different teaching methods.” This discusses the context of the concept and/or categories, and suggests that you may need a new category labeled “supportive environment.” Axial coding is merely a more directed approach at looking at the data, to help make sure that you have identified all important aspects. Have your highlights ready for revision/addition.

Create a table

Transfer final concepts and categories into a data table, such as this one (Aulls, 2004). Note how the researcher listed the major categories, then explained them after the table. This is a very effective way to organize results and/or discussion in a research paper. Here is an excellent comprehensive guide (think desk reference) to creating data displays for qualitative research.

Note: Be patient. This appears to be a quick process, but it should not be. After you are satisfied with your coding procedures, I suggest submitting your table to an expert for review, or perhaps even one of the participants (if interviewing) to promote validity.

Cookies on GOV.UK

We use some essential cookies to make this website work.

We’d like to set additional cookies to understand how you use GOV.UK, remember your settings and improve government services.

We also use cookies set by other sites to help us deliver content from their services.

You have accepted additional cookies. You can change your cookie settings at any time.

You have rejected additional cookies. You can change your cookie settings at any time.

Education, training and skills
Further and higher education, skills and vocational training
Sector-based Work Academy Programme: qualitative case study research
Department for Work & Pensions

SWAP Qualitative Case Study Research: Annexes

Updated 16 May 2024

This publication is licensed under the terms of the Open Government Licence v3.0 except where otherwise stated. To view this licence, visit nationalarchives.gov.uk/doc/open-government-licence/version/3 or write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email: [email protected] .

Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned.

This publication is available at https://www.gov.uk/government/publications/sector-based-work-academy-programme-qualitative-case-study-research/swap-qualitative-case-study-research-annexes

Annex 1: SWAP Theory of Change Logic Model

The flow diagram presents the following information:.

Inputs mainly under the heading ‘Government (Continued benefits, training costs, barriers to participation, analytical resources)’:

‘Employer -Time for set up/management/ feedback’

‘Training Provider - Time for set up/management’

‘Employer Advisor’, ‘Work Coach’

These inputs flow into the following activities:

‘Local Labour market analysis’
‘Arrange SWAP ’
‘Disseminate available SWAP with JCP ’
‘Awareness of available SWAP ’
‘Sell SWAP to claimant’
‘Refer claimant’

These activities flow to the output: ‘Claimant agrees to participate in SWAP ’. The claimant agreement flows on to the claimant input: ‘Investment of time and initial travel costs (reimbursable)’ and the following work coach activities:

‘Barrier Assessment’
‘Record referral’

Then the following employer activities:

‘Background checks (if applicable)’
‘Workplace adjustment (if applicable)’

These all flow into the output ‘Claimant starts SWAP ’.

The main SWAP portion of the theory of change starts here with the activities:

‘Pre-employment Training’ (with side flow to ‘Claimant gains certification (if applicable)’)
‘Work placement’
‘Guaranteed Interview’

The main SWAP activities flow to the following three possible outputs:

‘Claimant completes SWAP : Interview successful’
‘Claimant completes SWAP : Interview unsuccessful’
‘Claimant does not complete SWAP ’)

Regardless of which output, the diagram shows a flow to the activity ‘WC completes SWAP Tracker’.

If Claimant completes SWAP with an unsuccessful interview, or the Claimant does not complete the SWAP , these outputs flow into the short term outcome of ‘additional needs identified’. The unsuccessful interview also flows into the short term outcome ‘interview feedback from employer’.

Both of these short term outcomes flow into the activity ‘Reflection with WC’ and this can flow into the activity ‘Claimant applies for other jobs in sector’. Claimant applies for other jobs in sector can flow into a successful short term outcome ‘claimant gains employment in new sector’ or an unsuccessful short term outcome, in which case ‘claimant reengages with WC’.

Claimants unsuccessful at SWAP or subsequent interviews (through engagement with WC activities) flow into either:

the short-term outcome ‘Increased employability’
the medium term outcome ‘improve claimants employability’.

Collectively the short and medium term outcomes flow into the impact ‘Value for unsuccessful claimants’.

If the claimant is successful at either the guaranteed or subsequent interview, these flow into the output ‘ Claimant enters work in new sector’ which flows onto the following four short-term outcomes:

‘Change in attitude towards working in new sector’
‘Claimant has skills to succeed at new job’
‘Increased earning’
‘Reduced UC /benefit’

This may flow into the following medium term outcomes:

‘ UC ends (or is maintained at reduced level)’
‘Sustained employment (18 months)’
‘Career progression’

These medium term outcomes flow into the impacts:

‘Increased employment’
‘Reduced UC costs’

If the claimant is successful at the guaranteed interview, and subject to the assumption ‘Employer outputs and outcomes are dependent on the SWAP meeting employer expectations’, the following employer outputs are recorded:

‘Reduced vacancies’
‘Employer social responsibility goals met’
‘Development of local workforce’

‘Reduced vacancies’ and ‘Employer social responsibility goals met’ flow to the short-term outcome ‘Employer satisfied with SWAP experience’. This then flows to the activity ‘ DWP gathers feedback from claimants and employers’ and the short-term outcome ‘Collated employer success stories’.

These short-term outcomes flow to the medium term outcomes:

‘Businesses return for additional SWAP ’
‘Increased employer uptake in SWAP or other provisions’
‘Improved Attitudes towards hiring DWP claimants’
‘Employers approaching DWP with vacancies more readily’

‘Development of local workforce’ output flows to the following short-term outcomes:

‘ SWAP aligns with local market need including sector shortages’
‘Sector pathways identified’
‘Change in attitude towards working in new sector’, which flow to the medium term outcome ‘Improved fit between employers and claimants’.

Medium term outcomes in this employer focussed part of the theory of change flow to the impacts ‘Improved DWP relationship with business sector’ and ‘increased employment’.

Annex 2: Participant characteristics

Table 3: employer, training provider and claimant participants by swap sector, table 4: claimant participant characteristics, annex 3: additional methodology details.

This annex includes additional information about how the case study research was conducted.

Contacting claimants

A random sample of 150 claimants was drawn for each case study area (600 claimants in total across the four areas) in order to achieve 10 claimant interviews in each district. This sample size was in line with previous, similar research (in terms of mode, length and recruitment approach), which achieved a response rate of approximately 1 in 15 claimants. The sample was sourced from the SWAP manual trackers completed by each district which detail which claimants are referred each week to the programme. Claimant identification numbers were then linked to centrally held contact information (for example, postal address and telephone number).

The stratification of the sample was limited by the quality of data DWP holds on certain claimant characteristics (for example, ethnicity and disability information was not available) as well as claimants’ SWAP journey (only claimant start dates on the pre-employment training ( PET ) were consistently recorded by all areas). As a result, it was impossible to identify in advance claimants who had dropped out of a SWAP part-way through, or claimants who were successful at the guaranteed interview stage, which limited the study’s ability to explore these aspects in detail. The sample drawn was, therefore, broadly reflective (rather than representative) of the claimant population who started on a SWAP in terms of gender and sector of SWAP , and consisted of individuals who had started the SWAP PET within the previous 12 weeks of the sample being drawn. This time period was agreed in order to ensure the feasibility of obtaining a sample of 150 claimants from each area, while minimising as much as possible the risk of recall bias within claimant accounts of their experience.

It is important to note the claimant sample was delivered in two separate stages, to reflect the gap in fieldwork between Area 2 and Area 3. The samples for Areas 3 and 4, was additionally stratified by age (18 to 24 years vs. 25+ years) to account for the small number of potential participants aged 18 to 24 years provided in the sample for Areas 1 and 2.

All claimants in the sample were sent an advance letter to the address held on DWP ’s central records. This letter provided further information about the research, what their participation would involve, data processing information and an email address to which they could write if they wanted to opt-out. Claimants were called using the telephone numbers provided in each sample, and while formal quotas for recruiting participants were not used, calls were targeted to achieve a spread in terms of claimant gender, age and sector of SWAP (the latter was obtained from the SWAP manual trackers and was therefore dependent on DWP staff interpretations of this at the local level). Claimants were called up to three times without a response before they were not contacted any further. During the calls, researchers emphasised their independence from benefits processing and that decisions regarding participation would not affect claimants’ benefits in any way. Each interview lasted approximately 30 to 45 minutes and claimants received a £20 voucher for their time.

In Area 4, fieldwork was terminated early due to an underlying issue with this sample in which few claimants could be contacted (many claimants did not pick up the phone) and of those who did, few recalled the programme or had actually started the SWAP to which they had been referred. Only two interviews were completed from 207 recruitment calls, compared to 10 interviews completed from 88 recruitment calls in Area 3. The study team attempted to unpick the reasoning for the issues with the underlying sample in subsequent meetings and interviews with the local operational contacts, however, it was difficult to pinpoint this exactly. The information gathered suggested that the issue was likely a result of error(s) completing the local manual SWAP trackers. As a result, fieldwork was terminated early so that the findings could be reported to the timetable agreed.

Contacting employers and training providers

As described in the main report, the study was reliant on the case study areas to supply the contact details of employers and training providers who had taken part in a SWAP in their districts, as there was no alternative way of identifying these organisations. Within each area, the study aimed to interview a total of 7 employers, and 3 training providers, and so local contacts were asked to provide approximately 15 to 20 employer contacts and 5 to 10 training provider contacts to account for uncertainty in likely response rates. Obtaining contacts was more difficult in some of the case study areas and was affected by factors such as local record keeping of this information (for example, some training providers were listed as employers, and other contact information was out of date), and busyness of the staff involved. In all areas, subsequent samples of employers were requested due to poor response rates for this participant group.

To counter the risk of staff supplying only contacts for similar organisations, and therefore similar experiences of the programme, contact information for a range of organisations in terms of key characteristics (size, sector of SWAP , length of SWAP , number of SWAPs involved in, and how the SWAP was initiated) was requested. Organisations were then approached by researchers to ensure a spread across these characteristics, although achieving this was limited by response rates, particularly among employers.

Organisations were initially emailed using a template which explained the purpose of the research and asked if they were able to participate. Where organisations agreed to take part, they were then sent an additional information sheet and booked in for an interview at a convenient date and time. Where no response was received, a follow-up email was sent a few days later prompting them about the study. Finally, where the target number of interviews had not yet been reached, organisations were contacted by telephone for up to a maximum of two attempts. Where this was the case, the researcher verbally communicated the key information about the study contained in the initial emails.

For most areas, the first time employers and training providers heard about the research was when they were contacted via email about the study. In Area 4, however, DWP staff approached employers in advance before handing contact details over to the study team. This approach was taken as DWP staff in this area felt it would be beneficial in securing employer participation and minimised any risk to their relationships with these contacts if the study team were to contact them without warning. It should be noted that this may have increased the risk that some employers may have felt obligated to take part in the research and/or restricted their feedback due to a perceived lack of separation between DWP researchers and operational staff leading on SWAPs . As with all areas, researchers in Area 4 emphasised their independence from jobcentres ( JCPs ) and SWAP policy decision-making during each contact with participants, and the questions asked during data collection were framed in a way to encourage and enable participants to be honest about their experience. Despite this, it’s likely that a certain level of bias related to this aspect remains in the dataset obtained.

Each interview lasted approximately 30 minutes to an hour, depending on how much each organisation wanted to share. To ensure the most appropriate person was spoken to, the information shared during the recruitment stages requested that the participant was an individual who had knowledge of, or was responsible for, the SWAP that their organisation had been involved in.

Contacting staff

Once JCP Service Leaders had agreed for fieldwork to take place in their district areas, the study team were signposted to operational staff who would be able to facilitate the research. These individuals became key contacts for the study team during fieldwork. In initial meetings, these local contacts provided a broad overview of the SWAP set-up in their district, and the types of staff involved in delivery, from which a list of different staff roles to speak to as part of the fieldwork was agreed. Due to the varying nature of the local staffing models, it was easier to understand how SWAP delivery was organised in some areas more than others.

The project manager and case study leads maintained regular contact with these local contacts while fieldwork took place in each area. In Areas 1 and 2 this mostly consisted of contact via email, whereas for Areas 3 and 4 this took the form of a weekly scheduled meeting. In Areas 2, 3 and 4, a follow-up meeting took place with the local contacts to check the study team’s understanding of local SWAP delivery obtained through data collection, and to clarify any aspects of delivery that remained unclear.

The local contacts provided a list of suggested staff who could be approached for the fieldwork based on their role and involvement in local SWAP delivery. The study team then arranged the interviews and focus groups for these staff around their availability. In setting up the interviews and focus groups, an information sheet was provided about the research, and it was emphasised that their participation was voluntary. Despite this, some staff may have only participated in the study because they felt obligated to. As with other participant groups, the independence of the study team was emphasised, and participants were offered the opportunity to withdraw from the study if they wanted to.

Piloting [footnote 1] interviews were conducted with two members of DWP staff, one employer and one training provider who agreed to this. These interviews were conducted to test the length and appropriateness of the topic guides for these participant groups, and the quality of the data obtained. These individuals were recruited from a separate JCP district to the case study areas and so the information collected was not used in the analysis and reporting of this study. The topic guides were amended following these pilots.

For claimants, the first week of fieldwork was considered a pilot. Minimal changes were made to the topic guide following these interviews, and so, unlike the other fieldwork strands, the data collected during these interviews was analysed and reported on. It should, however, be noted that the topic guides and fieldwork processes were continually reviewed and modified throughout the data collection periods to ensure they were as efficient and effective as possible. The study team met multiple times a week to reflect on interviews, and formal debrief sessions were held within the study team, and separately with wider supporting researchers, following the end of data collection in each study area. This process allowed learning from each area to be implemented in subsequent fieldwork.

Analysis and Reporting

Once all interviews had been conducted, the interview notes formed the final dataset. The dataset was explored using a thematic analysis approach. As there were multiple researchers involved in the coding of the data, a coding framework (Annex 4) was developed to ensure consistency in coding across the study team. The research questions were used as a guide to ensure the framework aligned with the objectives of the research, and the framework was tested with an initial sample of interviews before a final version was agreed for coding the rest of the data (although this was still subject to ongoing tweaks as coding progressed).

Members of the study team were paired up to code a specific strand of data (for example, employers) and each pair coded the same initial set of notes to check alignment in their coding approach, before separately coding the remaining data. A separate member of the study team then examined a selection of coded notes from each pair to quality assure the completed coding. Feedback on the coding approach, particularly inconsistencies within each pair, was provided to the coders so that this could be incorporated into the analysis of future notes.

The project team met multiple times to discuss and agree the themes identified within the coded data. The themes identified via this process of analysis structured the findings within this report. When analysing the data, findings were explored by participant group (for example, claimants vs. employers) as well as by case study area (for example, Area 1 vs. Area 2), and these were included in the reporting where relevant.

A Quality Assurance ( QA ) panel was established to review the work of the research team during the analysis stage. The panel included researchers external to the project, senior researchers, and a fieldworker external to the study team, who was involved in conducting the research. This panel was engaged to review the initial coding framework that had been developed, and again to review how the codes had been applied to a sample of the data collected. This ensured that the approach taken to analysis had been peer reviewed, and that the data analysis conducted was of good quality. The final report was separately quality assured by an academic on secondment to the In-House Research Unit ( IHRU ), as well as senior researchers in the unit.

Annex 4: Initial coding framework

A pilot is a small-scale, preliminary study that is used as a test run for a particular research instrument to ensure its efficacy. ↩

Is this page useful?

Yes this page is useful
No this page is not useful

Help us improve GOV.UK

Don’t include personal or financial information like your National Insurance number or credit card details.

To help us improve GOV.UK, we’d like to know more about your visit today. Please fill in this survey (opens in a new tab) .

Qualitative Insights into Organizational Value Creation: Decoding Characteristics of Metaverse Platforms

Open access
Published: 15 May 2024

Cite this article

You have full access to this open access article

Fabian Tingelhoff ORCID: orcid.org/0000-0002-7235-6001 1 ,
Raphael Schultheiss 1 ,
Sofia Marlena Schöbel 2 &
Jan Marco Leimeister 1

141 Accesses

Explore all metrics

The significance of metaverse platforms is growing in both research and practical applications. To utilize the chances and opportunities metaverse platforms offer, research and practice must understand how these platforms create value, which has not been adequately explored. Our research explores the characteristics of metaverse platforms that facilitate value creation for organizations in both B2B and B2C sectors. Employing a qualitative inductive approach, we conducted 15 interviews with decision-makers from international corporations active in the metaverse. We identified 26 metaverse platform characteristics, which we categorized into six dimensions based on the DeLone and McLean Information Systems success model. Subsequently, we provide examples to illustrate the application of these identified characteristics within metaverse platforms. This study contributes to the academic discourse by uncovering the characteristics that shape the competitive landscape of emerging metaverse platforms. Leveraging these characteristics may offer metaverse providers a competitive edge in attracting complementary organizations to their platforms.

Three layers of abstraction: a conceptual framework for theorizing digital multi-sided platforms

Creating the Foundation for a Functioning Internal Platform

Tracing the Roots of the Organizational Benefits of IT Services

Avoid common mistakes on your manuscript.

1 Introduction

Recently, companies have begun to invest in metaverse platforms. McKinsey and J.P. Morgan project the metaverse as a trillion-dollar opportunity (McKinsey, 2022 ; Moy & Gadgil, 2022 ), while Gartner predicts 30% of companies will serve customers via metaverse platforms by 2027 (Gartner, 2022 ). This culminates in JP Morgan’s assessment, emphasizing that “metaverses will likely infiltrate every sector in some way in the coming years” (Moy & Gadgil, 2022 ). Indeed, with annual spending over $54 billion, expenditure on metaverses nearly doubled that of music purchases in 2021 (Moy & Gadgil, 2022 ). Researchers attribute this growth to metaverses revolutionizing virtual societal and economic interactions (Di Pietro & Cresci, 2021 ). Specifically, the metaverse benefit from network effects, becoming more appealing as user and organizational participation increases (Gawer & Cusumano, 2014 ; Tingelhoff et al., 2024 ). Accordingly, metaverse users benefit from the presence of more participating organizations, and reciprocally, having many users renders the metaverse more attractive for organizations (Kim et al., 2016 ).

While research deems metaverses only as viable once many users and complementors participate, about 500 companies already joined a metaverse platform, which only amounts to about 0.00015% of all companies (Brimco, 2024 ; Newzoo, 2022 ). A major contributing circumstance to the discrepancy of predicted and actual predicted and actual corporate engagement in metaverses is the inconsistencies of how different metaverse platforms influence organizational value creation. While past studies have examined how organizations mitigate challenges when creating value on metaverse platforms (e.g., Schöbel & Tingelhoff, 2023 ), organizations find it challenging to join metaverses due to the diverse ways in which various metaverses support value creation. Ultimately, it has not been fully explored how organizations can utilize characteristics of metaverse platforms (e.g., personalization and immersion) in their offerings to maximize their value proposition. When selecting a metaverse platform, organizations must prioritize its characteristics carefully to guarantee it fits their strategy. This study aims to uncover which metaverse platform features aid or hinder organizational value creation, leading to the following research questions:

RQ1 : What platform characteristics influence an organization’s ability to create value on a metaverse platform?

RQ2 : How have these characteristics already been configured in existing metaverse platforms?

This study seeks to contribute towards the research questions using qualitative data. We analyzed interviews with 15 decision-makers, all tasked with strategizing for corporations' engagements on metaverse platforms. We built on the DeLone and McLean Information Systems success model (D&M IS success model) (DeLone & McLean, 1992 , 2002 , 2003 ) to structure our data collection and analysis. From this, we identified 26 platform characteristics across six dimensions. Furthermore, we discovered these characteristics, such as accessibility and privacy, play different roles in metaverse platforms compared to traditional ones, a topic we expand on in our discussion. To substantiate our findings, we illustrate how these characteristics are divergently implemented in two existing metaverse platforms: Roblox and Decentraland.

Our study deepens the understanding of how organizations adopt metaverse platforms, laying groundwork for future research into corporate activities within these platforms. The study also broadens our knowledge of how emerging technologies' features support value-creation, beyond known platform mechanisms. Our results can help metaverse providers design more effective and appealing platforms. Moreover, by clarifying which platform features affect value creation, our study aids organizations in deciding how to offer their products or services on metaverse platforms. In summary, this study enables informed decisions about engaging with metaverse platforms for organizations.

This paper is structured as follows. The next section introduces foundational knowledge on metaverse platform ecosystems and outlines the structuring of their characteristics. Following that, we detail the methodology, then report and analyze the study's results. Subsequently, we exemplify our findings with case studies of two prominent metaverse platforms. In the sixth section, we explore the study's limitations and suggest directions for future research. The paper concludes with a summary of key insights.

2 Theoretical Background

2.1 the metaverse platform ecosystem and how it can create organizational value.

The metaverse is a multi-user virtual platform built on Web3 technology to revolutionize how people interact in any context (Di Pietro & Cresci, 2021 ). The malleable features of this emerging platform generate an immersive user experience, mimicking the real world sans physical constraints (Jaynes et al., 2003 ), thus enabling new and richer kinds of social and business interactions (e.g., through immersive 3D communication) (Bourlakis et al., 2009 ). By combining technologies like self-sovereign identity, virtual reality, and blockchain, these platforms evolve into parallel societies and economies. Though social media platforms (e.g., Facebook) or e-commerce platforms (e.g., Amazon) have already impacted human interactions and purchasing behaviors, this unique technology blend is unprecedented in virtual environments (Wang et al., 2021 ), enabling users to construct a parallel life with all its facets (i.e., work and leisure) in one platform for the first time.

The organizational structures around metaverse platforms classify them as an emergent platform ecosystem type (Schöbel & Leimeister, 2023 ). An ecosystem can be described as the underlying structure among partners, designed to enhance interactions and deliver a core value proposition (Adner, 2017 ). Generally, Typically, a platform is any product, service, or technology used by ecosystem actors to innovate and create complementary offerings (Gawer & Cusumano, 2014 ). It is typically integrated into ecosystems to help actors fulfill their value propositions (Gawer & Cusumano, 2014 ). For instance, the Apple App Store is integrated into the Apple ecosystem, including hardware and operating systems, enabling partners to create valuable apps and services.

Value is defined by an individual’s overall utility assessment based on their perception of input and output (Zeithaml, 1988 ). Two processual value components exist: use value and exchange value (Bowman & Ambrosini, 2000 ). Use value describes the recipient’s perception of a product or service based on their needs and values. Thus, use value stems from an organization's products and services, shaped by the recipient's personal and situational evaluation. Conversely, exchange value is what the recipient is willing to give in return for the use value. It can take other forms than monetary compensation, such as brand awareness or customer relationship, and is realized through the reciprocal transfer via the platform (Sanders & Simons, 2009 ).

Platform ecosystems comprise three major roles: platform leaders, complementors, and users. Platform leaders (or orchestrators) are the providers of the platform ecosystem (Gawer & Cusumano, 2002 ; Oliveira et al., 2019 ). They orchestrate the platform and are responsible for its regulation, maintenance, and adoption of standards within the ecosystem. Complementors provide products and services built for the platform, enhancing the ecosystem's core value. Lastly, users procure the complementors’ products and services via the platform. Usually, they access the platform and its complementary offering directly through a user interface. Consequently, in a metaverse platform ecosystem, value creation occurs through the reciprocal exchange of use and exchange values between complementors and users. Users receive products or services in the metaverse (use value) and are willing to compensate complementors with money and attention (exchange value). Accordingly, complementors are the source, users are the target, and the metaverse platform is the locus of value creation.

Some researchers contend that metaverse platform ecosystems challenge traditional role allocations (Schöbel & Leimeister, 2023 ). Notably, metaverse platforms represent early examples of decentralized governance models (Goldberg & Schär, 2023 ). Decentralized platforms are collectively owned by stakeholders rather than a single orchestrator. This collective ownership model is known as a Decentralized Autonomous Organization (DAO). On such a platform, users and complementors can impact decisions about platform governance, such as technical standards or the maximum amount of land created for users. This shifts the power balance among ecosystem participants, unlike in traditional platforms. In decentralized platforms, power is distributed; complementors and customers can propose and implement changes, not just the orchestrator (Goldberg & Schär, 2023 ). reducing the orchestrator's role also affects the direct relationships between customers and complementors (Yoo et al., 2023 ). While traditionally, both parties only interacted through the platform and, hence, the orchestrator, metaverses enable customers and complementors to interact directly. This is further emphasized by the fact that a metaverse does not require transaction intermediaries (e.g., banks). Specifically, the metaverse enables the direct exchanges of value items (Tapscott & Tapscott, 2017 ). This uniqueness to metaverse platforms potentially leads to simpler, more flexible, and efficient business processes by reducing communication complexities.

While many researchers contributed conceptualizing the metaverse, Schöbel et al. ( 2023 ) made the first effort to distinguish different types of metaverse platforms. Their taxonomy evaluates the technologies in a metaverse platform's infrastructure to highlight differences in value propositions. For example, they argue that a metaverse platform ecosystem like Decentraland focuses more on the creator economy and, hence, the creation and distribution of value items. In contrast, game-based platforms like Roblox focus on entertaining experiences that deepen user-brand relationships, alongside transactional value. This visualizes the paucity of platform types and business foci under the metaverse umbrella.

2.2 The DeLone and McLean IS Success Model

Identifying key success factors for organizational value creation is essential for insights that stay relevant in the rapidly changing metaverse platform ecosystems (Schöbel & Leimeister, 2023 ). In this context, past research employed value creation or value co-creation theory, predominantly focusing on aspects of value creation that are within an organization’s control. For instance, value creation theory typically centers on the consumer's perceived assessment, which organizations can influence by modifying their offerings or engaging in value co-creation with consumers. This aspect of how organizations can steer this process has been explored in existing literature, including in the context of the metaverse (e.g., Schöbel & Tingelhoff, 2023 ; Tingelhoff et al., 2024 ).

Conversely, our study examines platform characteristics that impact an organization's ability to create value. These external factors, often outside organizations' control, significantly influence value creation dynamics in technology-driven settings like metaverse platforms. In this context, the DeLone and McLean IS success model (D&M IS success model) (DeLone & McLean, 1992 , 2002 , 2003 ) has emerged as a cornerstone in the domain of Information Systems (Wang, 2008 ), enabling analysis of platform characteristics and their effects on organizational value creation. It provides a multifaceted perspective, deconstructing information systems into six key dimensions: information quality, system quality, service quality, usage intentions, user satisfaction, and net benefits.

Firstly, metaverse platforms are inherently complex digital ecosystems that thrive on exchanging information (Schöbel & Leimeister, 2023 ). The dimension of information quality is directly tied to the nature of data presented in metaverses, its accuracy, timeliness, and relevance. High-quality information will invariably influence user and organizational decision-making within a metaverse platform (Balica et al., 2022 ). System quality mirrors the technical prowess of a metaverse platform. As immersive environments, metaverse platforms demand high system performance, ease of navigation, and reliability. The better the system quality, the more seamless the user experience, thus attracting more complementors (Schöbel & Tingelhoff, 2023 ). Service quality in the metaverse context indicates the support structures in place. This could involve technical support, user guidelines, and assistance in content creation (Park & Kim, 2022 ). A high-quality service structure can significantly enhance the desirability of a metaverse platform for users and, in turn, organizations (Jo & Park, 2022 ).

The next dimensions, usage intentions and user satisfaction , arise from the interaction of the prior dimensions. Usage intentions resonate with how frequently users and organizations engage with the metaverse platform. A platform with a higher intent of usage becomes a hotspot for value creation and exchange, which is paramount in a networked environment like the metaverse (Ataman et al., 2023 ). Conversely, increased user satisfaction levels in a metaverse indicate that the platform effectively meets or surpasses the multifaceted expectations of users and organizations. When users are satisfied with their experiences, they are more likely to invest time, resources, and encourage peer participation, thus amplifying the platform's network effects and cementing its value-creation potential for organizations (Golf-Papez et al., 2022 ).

Lastly, the dimension of net benefits encapsulates the tangible and intangible outcomes that organizations derive from their engagement with an information system. When organizations can measure the positive impact of their involvement in a metaverse platform—whether in terms of revenue, brand recognition, skills acquisition, or social connections—it reinforces their commitment to the platform and ensures sustained engagement (Polyviou & Pappas, 2022 ). This sustained engagement, powered by recognized net benefits, can indicate the platform's long-term value-creation potential for organizations (Periyasami & Periyasamy, 2022 ).

While the D&M IS success model is tried and tested across various IS contexts, its inherent adaptability makes it particularly apt for metaverse platforms. The model’s validity at both individual and organizational levels (Petter et al., 2008 ) makes it a versatile tool for understanding an emergent and complex environment like the metaverse. Furthermore, previous validations of this model in diverse IS environments, from enterprise systems to e-commerce platforms (Wang, 2008 ), have shown its robustness and adaptability (Ahlan, 2014 ; Al-Kofahi et al., 2020 ). Applying it to the metaverse, an amalgamation of various IS types, seems like a logical progression.

The essence of the D&M IS success model is its capability to elucidate key dimensions that contribute to its capability to support complementors to create and deliver value through an information system (Wang, 2008 ). Yet, the interplay of these dimensions over time and the progressive stages of adoption need to be considered for a more comprehensive understanding. This aspect is vital as it bridges the D&M IS success model with real-world technological adoption behaviors and patterns, making it more contextually relevant, especially for evolving digital realms like the metaverse. By integrating process steps, one can understand not just what supports value creation (as indicated by the D&M IS success model) but also how and when these characteristics manifest and influence organizational value creation over the adoption lifecycle. To expand theories for a temporal interplay of characteristics, Ahlan ( 2014 ) proposed three process steps: system creation, system use, and system impact. This hierarchical order of influence characteristics can be applied to the D&M IS success model, where information, system, and service quality correspond to the system creation, usage intention and user satisfaction to system usage, and net benefits to system impact. The combined model is depicted in Fig. 1 .

Research Model, based on DeLone and McLean ( 1992 , 2002 , 2003 ) and Ahlan ( 2014 )

The initial phase of any technological endeavor involves its conceptualization, development, and implementation (Weber et al., 2023 ). This phase is particularly crucial within the metaverse context as it lays the foundation for the user experience. Information, system, and service quality become essential metrics at this juncture. The quality and relevance of information guide the design and functionalities of the metaverse platform (Bayraktar et al., 2023 ). System quality ensures the platform’s technical soundness and scalability, which is vital to handling the dynamic nature of the metaverse’s ever-evolving virtual landscapes and foundational technologies (Peukert et al., 2022 ). Service quality, conversely, pertains to the support mechanisms, ensuring that complementors have a smooth onboarding process and immediate resolution to any technical hitches (Xi et al., 2023 ). A robust system creation phase, bolstered by these quality metrics, sets the tone for subsequent adoption and usage.

Once a system has been created and launched, its success is predominantly gauged by its acceptance and the extent of its use. Here, usage intention is a precursor, indicating initial interest and potential adoption rates (Jeong & Kim, 2023 ). However, for the system to embed itself into the daily routines of users and organizations, satisfaction becomes pivotal (Xi et al., 2023 ). As users engage with the metaverse platform, their experiences, which are shaped by immersive interactions, realistic representations, and the fulfillment of intended purposes, dictate their satisfaction levels. Satisfied users not only continue their engagement but also promote organic growth through positive word-of-mouth and peer recommendations (Mladenović et al., 2023 ).

For metaverse platforms, the system impact, as denoted by net benefits, encapsulates how effectively platform characteristics support organizational value creation (Polyviou & Pappas, 2022 ). This could manifest in diverse ways, from driving innovation in product or service offerings, facilitating unique customer engagement models, to fostering new revenue streams or enhancing brand visibility within the virtual realms (Hadi et al., 2023 ). Moreover, it might also encompass intangible benefits such as enhanced collaborative potentials, access to new market segments, or the ability to test and iterate offerings in risk-mitigated virtual scenarios (Yoo et al., 2023 ). Thus, system impact, in this context, underscores whether the metaverse platform meets the technical and experiential needs of its complementors and provides a conducive environment for organizations to harness its potential and realize tangible value.

In summary, the adapted D&M IS success model provides an exhaustive framework to unpack, understand, and measure how metaverse platform characteristics influence organizational value creation. By mapping its six dimensions to the specific characteristics of metaverse platforms along Ahlan ( 2014 )’s process steps, this study aims to clarify the characteristics that influence organizational value creation in this burgeoning digital frontier.

3 Methodology

3.1 ensuring the quality of qualitative data.

This study investigates how metaverse platform characteristics impact organizational value creation. The phenomenon of the metaverse is relatively new, and less is known about how organizations create value on metaverse platforms. To gain exploratory insights into organizational value creation, we adopted a qualitative research method (Draper, 2004 ).

To ensure our qualitative data's validity and reliability, we adhered to Lincoln and Guba ( 1985 )’s criteria: credibility, transferability, dependability, and confirmability. Credibility refers to the data's internal validity and its alignment with reality (Merriam & Grenier, 2019 ). We selected our interviewees based on three criteria. First, we considered whether interviewees had a holistic overview of organizational value creation. This includes individuals who are not only involved in strategic decision-making but also with a broad understanding of how various organizational units contribute to value creation. Second, as we explored metaverse platforms, our participants needed substantial experience in the metaverse sector. This ensures that our data comes from individuals who are not only familiar with the concept but are actively engaged in its application and development within their organizations. Third, we focused on participants whose organizations actively offer products or services on metaverse platforms (e.g., Roblox or Decentraland). This practical involvement ensures our insights are grounded in real-world experiences and challenges. To guarantee a similar level of abstraction and comparability of our findings, we selected interview partners based on similar levels of responsibility regarding the metaverse. Furthermore, to increase credibility and mitigate possible biases in the data collection and analysis (Valenzuela & Shrivastava, 2002 ), the first three authors of the paper coded the data independently. Specifically, we employed the qualitative inductive approach described by Gioia et al. ( 2013 ), which is designed to enhance qualitative rigor in conducting and presenting inductive research. This approach is particularly suitable for inductively developing grounded theory, offering rich and detailed theoretical descriptions, which, in our case, pertains to the value creation on metaverse platforms. In line with the method proposed by Gioia et al. ( 2013 ), we constructed first-order concepts, second-order themes, and aggregate dimensions, with the latter reflecting the theoretical constructs.

To ensure intercoder reliability, the coders critically discussed their initial coding results until the first-order codes were sufficiently reviewed, and a second iteration could start that led to a detailed description of our second-order constructs. During this process, the coders aimed for the codes to be on a comparable level of abstraction and still reflect the individual experiences of each informant. Our analysis was rooted in a hermeneutic approach, emphasizing the shared understanding of the nature of metaverse platforms. Following the insights of Paterson and Higgs ( 2005 ), we embraced hermeneutics for its dialogical nature, which in our research manifested through the interviews conducted between experts and the interviewer.

To improve the transferability—whether the results of one study can be transferred to different settings with different participants—of our results, we have provided sufficient contextual information about each informant and their interview in Appendix 1. Dependability describes the extent to which the findings of a study can be replicated (Merriam & Grenier, 2019 ). We noticed a high degree of concept/coding saturation (95% of codes surfaced) after the 11th interview, which aligns with previous studies on the method (Guest et al., 2006 ) and the topic at hand (Schöbel & Tingelhoff, 2023 ). Finally, confirmability mainly concerns objectivity and addressing potential research biases that can result, for instance, from inherent values and beliefs but also the mere presence or timing of follow-up questions (Valenzuela & Shrivastava, 2002 ). We addressed confirmability by independently executing tasks simultaneously and comparing results (e.g., for data coding and interpretation). Further, all the researchers were at least aware of their potential biases and influences, which we critically, openly, and proactively addressed during the planning and execution of this research.

3.2 Data Collection

In our qualitative study, we interviewed 15 metaverse decision-makers from multilateral organizations in both B2B and B2C markets. These participants represent the mission and vision of an organization; they have the highest responsibility for their organization’s metaverse strategy, were familiar with key metaverse platforms, and understood how different organizational units contribute to value creation. In other words, we interviewed CEOs and senior leaders who are deeply familiar with their organizations and can insightfully discuss how metaverse platforms align with their business models to create value. Interviewees explicitly consented to their interviews being recorded and their personal data being published, both before and after the interviews. This procedure adhered to ethical research standards, ensuring transparency and reliability in data collection.

We refer to individual participants using the abbreviation IP (interview partner) and their number (1–15). We used a structured interview guide to ensure interviewees fully understood the platform characteristics being studied. The guide featured open-ended, concise questions, supplemented with follow-up probes to resolve ambiguities. Additionally, we provided contextual information and examples where necessary to aid comprehension. We began by asking participants about their demographics. This was followed by general questions about the metaverse (how they defined it, for instance). We then focused on which platform characteristics they deemed crucial for strategic decisions regarding their metaverse offerings. We explicitly referenced the three process steps of system creation, system use, and system impact. In addition to discussing the impact on their organizations, we invited opinions on future metaverse platform developments and their relevance to the interviewee’s organization.

Interviews were held online (Lo Iacono et al., 2016 ) and conducted in the interviewees' native languages when possible (Harzing & Maznevski, 2002 ). To address the potential for translation-induced alterations, we employed a rigorous process. First, we conducted 66% of the interviews in the native languages of both interviewees and interviewers, reducing the necessity for translation. The remaining interviews were conducted in English. For interviews needing translation, a bilingual, topic-knowledgeable co-author handled the task. A second bilingual co-author reviewed and verified the transcripts against the audio recordings for translation accuracy. The interviews lasted between 26 and 56 min (mean: 34). In one interview (IP 5), technical difficulties obstructed the recording. The interviewer—in this case, the second author— immediately after the interview reconstructed the interview details from notes and memory.

We based our coding, for which we used the Atlas.ti22 software, on the transcripts (Gioia et al., 2013 ). The first three authors coded simultaneously and independently. According to Gioia et al. ( 2013 ), the first-order coding was an open coding mechanism that remained faithful to the words of each interviewee. Ending up with a high number of codes prompted us to summarize those codes where the informants addressed the same topic but with different terminology. Since the initial open coding was executed independently, ongoing researcher discussions shaped the final constructs. After identifying platform characteristics, we assessed their relative importance. To this end, we presented the ordered variables to our interview partners for an ordinal rating of relative strategic importance. To substantiate our findings, we showcased the highest-rated characteristics using two practical examples. Roblox and Decentraland serve as leading examples of early metaverse platforms (Schöbel et al., 2023 ). As current recipients of immense funding and publicity from practitioners (Kamin, 2021 ; Wang & Ho, 2022 ), they are ideal subjects for further investigation.

In line with our research goals, we identified 26 platform characteristics and categorized them according to the six dimensions of the D&M IS success model. Additionally, our metaverse experts ranked these characteristics based on their importance and relevance to organizational value creation in metaverse environments. Figure 2 reports the characteristics per dimension ranked by their average importance (with 1 indicating the highest importance). However, the rankings reflect the experts' average views. Though indicative of relative importance, they are not meant as precise statistical measures. Given this, we advise that small differences in rankings, particularly those within 0.1 or 0.2, may be negligible.

In subsequent sections, we detail our empirical findings and analyze each of the six success dimensions.

Adapted Research Model (including the average ranks by IPs)

4.1 System Creation: System Quality, Information Quality, and Service Quality

The system quality dimension encompassed the most platform characteristics. In this dimension, interview partners (IPs 1, 2, 4, 8, 9, 12, 13, 14, and 15) ranked accessibility —ease of access for complementors and users—as the most crucial. The interviewees specifically mentioned that platforms must be accessible through various devices, including tablets, smartphones, and desktop PCs (IPs 1, 2, and 9). IP 9 elaborates:

It is all about being available across the devices. The more devices you have available for the platform, the easier it is to enter. I mean, as we all know, traffic is coming from mobile and desktop. […] So, you know, if your platform is only compatible with PCs, you’re missing out on a big portion of the market. Footnote 1

However, accessibility encompasses not only device compatibility but also the ease of navigating the platform's functions. “One should be able to move around and execute actions without having to read a one-hour manual,” argued IP8 when asked about the necessary functionalities regarding the system quality of a metaverse platform. Similarly, IP 14 emphasized that operability is currently more important than functionality.

Interviewees (IPs 4, 12, 14, and 15) also deemed an integrated economic system essential. The interviewees highlighted the necessity of integrated payment systems (IP 12) and their ability to create new user experiences (IP 15). This includes a currency that offers voting rights on decentralized platforms. Further, economic systema enable new functionalities, such as play-to-earn, where users are rewarded with items or tokens of monetary value. In this way, a platform can ensure its most active users receive a higher voting share. Additionally, new business models—for instance, transferring goods from the virtual to the physical world and vice versa—require digital ownership structures that metaverse platforms must provide (IP 14).

Across all dimensions, platform stability emerged as the most frequently mentioned characteristic (IPs 2, 3, 4, 6, 7, 9, 11, 12, 14, and 15) . Interviewees expressed concerns that platform bugs, like a faulty login page, could harm their brand image (IPs 4, 6, and 9). While some worried that these “bugs lead to users exiting the metaverse altogether” (IP11), others cautioned against entry barriers: IP12 warned that “[t]he metaverse-guest’s fear about crossing the threshold into the virtual immersion must be addressed for them to use it at all.”

Information quality represents the second dimension in the D&M IS success model. Content quality emerged as the primary concern in the interviews (IPs 3, 4, 6, 7, 8, 10, 12, 13, 14, and 15). Organizations worry about losing potential customers if the platform's overall user experience and other complementors' offerings are unsatisfactory. “So, you know, if the content is not attractive,” IP 14 explained, “people will just hop onto it, take a look, and probably never come back. If you don’t engage your participants, that becomes an issue.” Specifically, IP14 worries that a bad experience on one platform could deter customers from using any metaverse platform. Compared to other platforms, this concern is particularly pertinent for metaverse platforms given their current stage of development. Additionally, companies consider the connotation of the content produced on a platform. For instance, companies may hesitate to join game-based metaverses due to concerns that users could associate the game's characteristics with the company. IP12 specified this concern: “We try hard to avoid becoming a Disneyland. This is not just a game and fun. It is totally okay if it is around fun; that is not wrong. […] Even though a platform might have entertainment aspects, we also want to be able to educate our customers in a serious manner.”

Personalization (IPs 1, 8, 11, and 14) is another integral metaverse platform characteristic. This encompasses two facets of customization. First, organizations aim to align their virtual presence with their brand identity: “Nobody wants to access a generic platform with this great template-world. No, everybody wants their own logos, their own trees, their own information channels” (IP 11). In addition to platform design flexibility, organizations also aim to offer personalized products to their customers. This introduces several technological challenges, as discussed by Duan et al. ( 2021 ). To provide personalized features effectively, orchestrators must balance individual customization with maintaining a cohesive design across the platform. This creates tension among stakeholders, as personalization challenges the standardization and regulations needed for a unified look. IP11 cited Meta’s AI builder as a notable example of addressing these challenges:

The amount that individualization and customization are wanted by the customers is extremely effortful. […] There is now this AI builder by Meta. It’s a prototype based on voice recognition. So, Mark Zuckerberg is standing there [on the metaverse platform] as an avatar and says: ‘Hey, I need an island,’ and then an island appears. And then he says, ‘I want trees,’ and a palm tree appears. […] And this is how you could build customized yet standardized objects for your customers in a few minutes.

Service quality marks the last dimension within the system creation phase. Here, aspects of privacy and security (IPs 1, 3, 6, 11, 12, 13, and 14) emerged as the focal platform characteristic. This focus stems from the unique types of data collected by metaverse platforms. With advancements enabling the collection of micro-movements and face-tracking data, organizations can now access unprecedented types of personal data. Aware of these risks, interviewees unanimously agreed that “if it [a metaverse platform] is not built safely, it won’t be successful” (IP 6). In other words, our interviewees require security and privacy concepts to develop a metaverse platform “into a safe space” (IP 13). Nevertheless, organizations often require sensitive data to effectively serve their customers. For instance, a hotel company testing its designs in the metaverse necessitates personal data on user movements within the platform. Consequently, organizations are concerned that customer reluctance to share data on metaverse platforms, coupled with their cautious approach to data collection, may restrict product sales (IPs 1 and 14). This creates tension between the need for privacy and safety versus the drive for product offerings and sales, often prioritizing the latter at the expense of the former. Therefore, striking the correct balance becomes crucial for complementors, given its significant influence on a platform's operational capabilities.

4.2 System Use: Usage Intention and User Satisfaction

Usage intention determines the target users of a platform and their manner of engagement. Naturally, organizations aim to reach their target audience upon entering a metaverse platform (IPs 1, 6, 10, 11, 12, and 14). IP 10 explains that large corporations might easily financially engage with any major platform, whereas small and medium enterprises (SMEs) must carefully choose platforms where their selected target audience is most active. For instance, while Roblox’s age demographic is age 13 to 25, the users of Meta’s Horizon are potentially older (IP 6). Additionally, complementors seek platforms with an active user base (IPs 3, 4, 6, 7, 8, 9, and 15). Platforms are appealing if “used by active users on a daily basis” (IP 4), as complementors aim to “engage and interact with users to enter a common dialogue and exchange” (IP 7). In other words, complementors view the metaverse as a venue for initiating interactions with current and potential customers. In this context, the metaverse acts as an additional channel for organizations to actively engage with their target audience (Hadi et al., 2023 ). Some companies aim to educate customers and create leads in the metaverse (IP 12), others focus on product-centric community creation to foster customer loyalty (IPs 3 and 6). This is consistent with the metaverse's focus on social interaction and immersion (Schöbel et al., 2023 ).

The user satisfaction dimension relates to the benefits users perceive from a platform. This perception shapes how users view the complementors on the platform. As a result, organizations favor platforms that cater to user needs. In this regard, the most critical platform characteristic concerns usability and user experience (IPs 1, 3, 4, 7, 8, 9, 10, 12, 13, and 14). This aspect is closely linked with system availability, a key part of the system quality dimension. It emphasizes a user-friendly experience throughout the application phase and the entire user journey (IP 7). IP 8 desires platforms that include “elements to positively surprise the user,” a point further elaborated by IP 4:

“Then another point would be the platform usage, this is all about how to access this platform and send the user out to discover the world and the many, many leads of the platform. How could we even make the consumer journey way easier to access the metaverse? […] How could we make this […] extend reality? How could we remove all the initial steps to make it like a natural, accessible environment?”

IP 4's concerns focus on how a metaverse platform's unique features are practically implemented. More specifically, metaverses are meant to extend reality by adding features for business and leisure (Bourlakis et al., 2009 ). For instance, making crypto wallet sign-up optional could increase platform accessibility. However, this could restrict the functionality of the platform's economic system, as seen in Decentraland's guest login feature. Therefore, platforms and complementors must find compromises to balance this tension.

Most interviewees agreed that a metaverse platform must offer added value to its users, essentially making it useful (IPs 1, 2, 3, 4, 6, 7, 8, 12, and 14). A metaverse platform “needs to provide an inherent added value to its users to convince [complementors] to have potential” (IP8). This implies that a metaverse platform must offer value to users even without the contributions of complementors. Here, value means any desirable outcome users perceive from using the platform. This challenges Bowman and Ambrosini ( 2000 )’ value theory that positions complementors as the exclusive source of use value in traditional ecosystems. Nonetheless, complementors prefer metaverse platforms that are already part of a functional, value-adding ecosystem (IPs 1, 4, 7, and 8).

The rationale behind these characteristics lies in the competitive dynamics of early-stage platform markets. IP 3 noted that platforms offering the highest perceived value will outlast others, becoming attractive investments for complementors. Although complementors are crucial to a digital platform's value proposition, they approach metaverse platforms with skepticism, demanding evidence of past performance. Complementors aim to minimize investment and brand reputation risks by insisting on metaverse platforms demonstrating success (i.e., adding user value) before participating. Such requirements pose significant challenges for emerging platforms. Larger orchestrators like Meta (formerly Facebook) and Microsoft benefit from established credibility, whereas smaller, entrepreneurial platforms face stricter scrutiny, complicating the development of strong complementor ecosystems.

4.3 System Impact: Net Benefits

Net benefits describe the positive outcomes expected by complementors from their participation in a platform. Unlike previous dimensions, which focus on the metaverse's structure and customer relations, this dimension concentrates on the direct interaction between complementors and the platform. Foremost, complementors value community building (IPs 3, 4, 6, and 14). Our experts view the metaverse as enhancing bidirectional relationships between brands and communities, beyond what is typical on social media. IP3 highlighted this advantage, noting it “allowing our fans to give feedback to us.” More specifically, companies can nurture their communities while profiting from their responses. Product development is a field where this potential is already visible (see customer engagement ). As an example, Starwood Hotels first constructed their properties in the metaverse world of “Second Life” (Gates ( n.d. )). Subsequently, they invited customers to experience the space in 3D and offer feedback. Only after several iterations did Starwood begin real-world construction. Moreover, complementors can benefit from community building in various other use cases. For instance, the energy drink brand Prime engages its audience through the metaverse and NFTs (see customer engagement ), fostering community feelings and boosting beverage consumption (see increased sales ). Thus, community building serves as a steppingstone, not just an end goal. This aligns with previous studies on community building and social media (e.g., Guo et al., 2016 ; Sledgianowski & Kulviwat, 2008 ).

Brand building (IPs 3, 6, 7, 10, and 14) involves marketing activities with enduring effects. IP 14, among others, highlighted marketing as the most viable current business model outcome in the metaverse. Interviewees see metaverses as crucial for educating customers about a brand (IPs 3, 6, 7, 14). As with every online marketing, the metaverse can be used to engage customers with the core principles of one’s brand (IP 14), to provide information and experiences about products (IPs 7 and 10), and to create strong emotional bonds, such as customer loyalty (IP 6). Yet, metaverses go beyond traditional functions by offering real-life brand experiences. For instance, in the metaverse, customers can test-drive a BMW (BMW, 2023 ), feel the difference of Nike shoes (Nike, 2023 ), or experience Givenchy's brand values at a virtual pool party (Smith, 2022 ). Immersing users to the point where fiction and reality blur (Lee et al., 2022 ) allows for deeper and more impactful brand experiences (Hadi et al., 2023 ). Therefore, interviewees stressed the importance of a metaverse platform's brand-building tools in choosing the right metaverse.

Although many interviewees (IPs 1, 2, 3, 6, 7, 8, 10, and 15) mentioned increased sales as a significant platform benefit, it received the lowest ranking across all dimensions, falling into the 80th percentile. This finding was surprising, given the portrayal of the metaverse as a prime venue for trading, evidenced by trends in NFTs and virtual real estate. However, our interviewees either did not see its revenue potential (yet) or appreciated its other unique opportunities, disregarding financial benefit. For instance, IP3 asserted that “there is no revenue potential out of this [metaverse],” contrasting with IP2's statement that “we just don’t want to burn money.” Ultimately, all interviewees agreed that sales increase was among the least important net benefits. However, metaverse usage becomes much more apparent when comparing the results of B2B companies to B2C companies. In our sample, all B2C interview partners used metaverse platforms for community or brand-building, indicating that consumer brands predominantly utilize these platforms to foster their existing assets (mainly customer base and brand equity). This emphasizes that, for these companies, the metaverse represents not a quick profit avenue but a long-term, sustainable investment. Conversely, B2B organizations show no clear tendencies. While five interviewees (IPs 6, 8, 11, 13, and 14) evaluated community and brand-building as the most important, two indicated reaching new customers and customer engagement as their highest priorities (IPs 3 and 2, respectively). However, none ranked increased sales first. This is in line with previous research on organizational value creation in metaverses. For instance, according to Schöbel and Tingelhoff ( 2023 ), organizations lack the foundational knowledge to understand a metaverse platform’s opportunities and challenges, thus hindering efficient decision-making in the B2B sector. Alternatively, the diversity in organizational needs within the B2B sector might explain the varied preferences. Nonetheless, the unanimous lack of priority for increased sales in both B2B and B2C groups highlights the metaverse's broader benefits, offering extensive marketing and customer interaction possibilities beyond mere e-commerce.

5 Discussion of Results: Illustrating How to Transfer Platform Characteristics

Building on earlier discussions about crucial metaverse platform characteristics for complementors, certain orchestrators are now incorporating features that align with these metaverse attributes. Decentraland and Roblox are two leading examples of early metaverse platforms. With 42 million daily active users globally, Roblox stands as one of the most popular metaverse platforms (Gollmer, 2022 ). Conversely, Decentraland has approximately 8,000 daily users, according to internal Decentraland Foundation data (Decentraland, 2022 ). Despite smaller user numbers, Decentraland remains a leading name among decentralized metaverse platforms (Brooke, 2022 ). Furthermore, Decentraland (IPs 2, 3, 4, 6, 7, 8, 9, and 15) and Roblox (IPs 3, 4, 6, and 13) were the most frequently mentioned metaverse platforms by our interviewees, referencing these platforms continuously as metaverse examples.

Studies indicate that both platforms possess the technological capabilities essential for metaverse classification, such as immersion or creator economy aspects (Schöbel et al., 2023 ). For instance, Decentraland encourages creativity by allowing users to own and develop land parcels. Each platform offers a comprehensive visualization of its virtual world. Collaborating with these platforms allows organizations to bypass the need for hiring game designers or creating immersive worlds from scratch. Thus, both platforms serve as accessible entry points into the metaverse. The key distinction lies in Roblox's centralized structure versus Decentraland's decentralized governance. The Roblox Corporation, a US-based public software company, controls significant updates to Roblox. In contrast, Decentraland operates as a DAO, where ownership and decision-making powers are distributed among users and complementors via the virtual currency, MANA. This model enables stakeholders to vote on governance issues and how the treasury is allocated (Brooke, 2022 ). We will use these two platforms as practical examples to illustrate our interview findings.

5.1 Illustration of System Creation

Both platforms’ user experiences are quite diverging. Roblox stands out for its accessibility , supporting web browsers, mobile devices, and Xbox consoles, aligning with our interviewees' preferences. Additionally, Roblox offers both 2D and immersive 3D experiences via virtual reality (VR) headsets. Despite its maturity, Roblox has yet to incorporate augmented reality (AR) technology (Shin, 2022 ; Yang et al., 2022a , b ). Metaverses aim to blend virtual and physical realities, enabling seamless information exchange between the two (Marabelli & Newell, 2022 ). AR offers significant advantages by overlaying virtual content onto the real world, intertwining both realities. Body sensors can further enhance user immersion by integrating physical movements into the virtual experience (Park & Kim, 2022 ). Despite our interviewees' emphasis on accessibility, metaverse platforms infrequently implement these immersive features (Park & Kim, 2022 ).

Although Roblox boasts an open and accessible design, it restricts users with a compulsory sign-in process that demands personal information, including date of birth. Additionally, signing in necessitates users' acceptance of Roblox’s terms of use and privacy policy. This requirement raises privacy and security concerns due to the submission of private data for platform access (IP 1, 3, 6, 11, 12, 13, and 14). Past research highlights the critical role of data security measures like privacy guidelines, stressing the need for user empowerment in decisions regarding data collection and storage (Guidi & Michienzi, 2022 ; Ning et al., 2021 ). Figure 3 shows Roblox’s access page.

Roblox’s Web Access

Decentraland features a distinct economic system, a priority reflected in its user account setup. Unlike other platforms, Decentraland registration requires a third-party crypto wallet, such as MetaMask, instead of a platform-specific account. This shows that Decentraland, unlike Roblox, is not interested in being the proprietor of its users’ data. A crypto wallet is essential for owning and trading virtual goods and currencies, a fundamental aspect of metaverse platforms (Di Pietro & Cresci, 2021 ; Oliver et al., 2010 ; Tayal et al., 2022 ; Vidal-Tomás, 2022 ). Using a crypto wallets as accounts, Decentraland’s users can execute peer-to-peer economic transactions without intermediaries. This aligns with our interview results, where interviewees highlighted the need for integrated payment systems (IP 12) as enabler for unique user experiences (IP 15).

User-generated content (UGC) is a defining characteristic of metaverse platforms. UGC empowers users to enrich the platform with their creations, ranging from services and products to various content. While some researchers argue that all content must be user-generated in a metaverse (Ayiter, 2012 ), others believe metaverses should primarily enable user creativity (Dionisio et al., 2013 ) to facilitate UGC (Oliver et al., 2010 ). For instance, Roblox incorporates gaming elements (Getchell et al., 2010 ) and actively supports users in crafting their games from the ground up (Metcalf, 2022 ). The Roblox engine allows users to seamlessly transition between diverse game genres, like puzzles and sports —this is, in part, related to the metaverse vision of persistency and interoperability between different platforms. Additionally, Roblox offers resources for users to learn programming, further empowering them to design their games. Roblox's open creation platform has led to the development of over 32 million unique virtual experiences, indicating its focus on freedom of creativity . Greater creative freedom for content creators fosters more innovation in the design of experiences (Orgaz et al., 2012 ).

Similarly, creativity extends to the creation and sale of digital goods (Boughzala et al., 2012 ; Kim, 2021 ). Roblox, along with its users and complementors, can create and sell these digital items. Roblox thus empowers its users to transact and monetize their content, a vital characteristic of metaverse platforms (Popescu et al., 2022 ; Tayal et al., 2022 ). Yet, Roblox conducts all transactions with its proprietary currency, Robux, which can be purchased via the platform using traditional payment methods, such as credit cards. Studies indicate that transactions in FIAT currency benefit users by eliminating currency conversion and lock-in mechanisms (Gadalla et al., 2013 ; Hwang & Lee, 2022 ; Papagiannidis et al., 2008 ; Vidal-Tomás, 2022 ; Yang et al., 2022a , b ). Despite the emphasized importance of interoperability by research (Di Pietro & Cresci, 2021 ; Kim, 2021 ) and our interviewees (IPs 1, 7, and 14), Roblox lacks this feature. Although purchases on Roblox are persistent (Braud et al., 2021 ; Falchuk et al., 2018 ), they are not transferable to other platforms (Wang et al., 2021 ). Figure 4 shows the Roblox Avatar shop.

Roblox’s Avatar Shop

Decentraland also prioritizes personalization . For instance, the platform offers extensive character customization options available even to guest users. Beyond many free features, Decentraland also offers paid personalization options, such as skins or collectibles. This is an essential feature of metaverse platforms, as scholars agree on the importance of designing avatars resembling the actual appearance of the user (e.g., Dionisio et al., 2013 ; Gadalla et al., 2013 ; Hwang & Lee, 2022 ; Shin, 2022 ). However, some researchers describe real-time 3D scans as the pinnacle of avatar customization (Schöbel et al., 2023 ). Photorealistic depictions of users offer significant advantages. This approach mitigates safety concerns related to anonymity and potential irresponsible behavior in virtual spaces (Falchuk et al., 2018 ; Guidi & Michienzi, 2022 ; Sykownik et al., 2022 ). Yet, achieving photorealism introduces several technological challenges. Implementing face scans requires infrastructure like depth sensors, while photorealistic rendering demands high computing power and significant data storage, both of which are already current bottlenecks in expanding virtual worlds. Decentraland's approach to character customization, depicted in Fig. 5 , strikes a balance between photorealism and anonymized character appearances.

Decentraland’s Character Customization Page

Transitioning between locations on both metaverse platforms resulted in significant loading times. During our Decentraland test, specifically when creating the character, the platform crashed, erasing all progress (see Fig. 6 ). While platform stability was mentioned by almost every interviewee (IPs 2, 3, 4, 6, 7, 9, 11, 12, 14, and 15), it cannot always be guaranteed. These issues stem largely from the high demands for computational power and data transfer (Choi & Kim, 2017 ). The integration of technologies like VR, decentralized ledgers, photorealism, and platform interconnectivity means metaverses will eventually generate more data than current storage capacities can handle (Schöbel et al., 2023 ). Therefore, “hosting and handling a metaverse platform will require significantly more effort than organizing traditional two-sided market platforms” (Schöbel et al., 2023 , p. 8). These challenges can already be observed in the platforms’ governing decisions. To manage computational and data-sharing loads, both platforms cap the number of users per server, as computational power, rendering, and data traffic requirements increase exponentially with each additional user. Imposing these limitations, however, contradicts fundamental metaverse principles like unrestricted user movement (Jaynes et al., 2003 ; Owens et al., 2011 ) and independence (Davis et al., 2009 ; Khansulivong et al., 2022 ). Further, social interactions are a cornerstone of the metaverse concept (Davis et al., 2009 ; Wang et al., 2021 ). The present state of technology and governance in metaverse platforms indicates that a fully realized metaverse remains a distant goal (Peukert et al., 2022 ; Schöbel et al., 2023 ).

Error Message while using Decentraland

5.2 Illustration of System Usage

Past research has identified trust as a critical component for the functionality of an ecosystem (Lang et al., 2019 ; Tawaststjerna & Olander, 2021 ). Stakeholders rely on trustful interactions among ecosystem actors, particularly in e-commerce and social media platforms (Bonina & Eaton, 2020 ). E-commerce platforms serve as centers for financial transactions. Customers and complementors must trust these platforms to accurately and error-free conduct their transactions. Given that many transactions on these platforms are one-time events, complementors face challenges in building trust through rapport with customers. Hence, e-commerce platforms (such as Amazon) act as trusted intermediaries (Friedrich et al., 2019 ; Molla & Licker, 2001 ; Torkzadeh & Dhillon, 2002 ). Users prepay for products, trusting in timely delivery, whereas complementors trust the platform to compensate them for sales. Thus, trust among its actors is essential for an e-commerce platform's proper function (Friedrich et al., 2019 ). Similarly, social media platforms manage valuable data. Unlike e-commerce platforms, social media platforms facilitate the sharing of personal information, like preferences and opinions, rather than financial transactions. On social media, users express their identities by customizing profiles, engaging with content, and sharing their posts (Krasnova et al., 2017 ; Lin & Lu, 2011 ). The importance of trust in social media was underscored when Twitter sold verification checkmarks without vetting accounts. As users trusted the check would verify an individual’s or company’s authenticity, many took announcements from fake accounts as serious news. This led to global stock market turmoil, with some companies losing billions in market value due to misinformation (Mac et al., 2022 ). Experts concluded that Twitter, as an information mediator, “undermine[s] the original purpose […] – to help users know they can trust information being shared” (Duffy, 2022 ). This instance highlights the crucial need for trust within social media ecosystems (Sledgianowski & Kulviwat, 2008 ).

The metaverse stands out as a platform that integrates functionalities from e-commerce, social media, collaboration, and education into a single environment (Tingelhoff et al., 2024 ). Consequently, it is designed to support both financial transactions and the exchange of private information. Therefore, trust may play an even more critical role for ecosystem actors in the metaverse than in other digital environments (Badruddoja et al., 2022 ; Wang et al., 2022 ). Existing research consistently demonstrates how user experience influences trust in digital ecosystems (Seckler et al., 2015 ) and automated systems (Yang et al., 2017 ).

At its heart, the metaverse is focused on delivering a distinctive user experience (Tingelhoff et al., 2024 ). Immersion and automation technologies make user experience even more crucial in the metaverse compared to other digital platforms. Our interview partners mirrored this. For instance, IP4 emphasized the importance of removing obstacles to enhance the accessibility and enjoyment of the virtual world. Moreover, IP9 identified user experience as essential for the success of both the platform and its complementors. IP8 summarized: “The deciding factor is clearly customer experience, meaning that a user can develop a positive feeling on the metaverse platform and then leave it with a smile on their face.” This can be further supported by anecdotal evidence from our research team. When the previously described error message occurred while testing Decentraland (see Fig. 6 ), our first author exclaimed: “If the platform isn't stable enough to create an avatar, how can I trust it with my financial transactions?”.

5.3 Illustration of System Impact

Given the vast range of integration options Roblox and Decentraland provide for complementors, their business models and resulting net benefits can significantly vary. Even within a single platform, complementors often pursue varied objectives through their participation. For instance, Adidas launched an NFT collection to potentially attract new customers and boost sales (Bain, 2023 ),, whereas Nike concentrated on brand building . On Roblox, Nike created Nikeland, a virtual world where users engage in games themed around Nike shoes and interact with one another. Players navigate the map using features like sprinting or hoverboards, underscoring sportiness and innovation—qualities Nike aims to embody (Marr, 2022 ).

Nike additionally focuses on community building . Specifically, it nurtures a sense of community by blending collaborative and competitive dynamics. Competitive features are key in metaverses, engaging users, encouraging social interaction, and presenting challenges to tackle (Martins et al., 2022 ; Quintín et al., 2016 ). Nikeland features a leaderboard, allowing users to continuously compare achievements with peers. Users can also team up for challenges, promoting user collaboration (Martins et al., 2022 ). These game elements are proven to influence brand image within the metaverse (Oliver et al., 2010 ).

Beyond subconsciously building brand associations (Kim et al., 2022 ; Wagner et al., 2009 ), Nike actively educates customers on its brand values during gameplay. According to our interviewees (e.g., IP 12) and existing research (e.g., Tingelhoff et al., 2024 ), metaverse platforms deeply engage and educate customers through immersive content (Lee et al., 2022 ). Specifically, in Nikeland, the content is centered on Nike-related content. For instance, players interact with a guide named “Nike Coach.” Wearing a Nike shirt, the coach is presented as an authority on health and sports. The coach assigns quests and offers advice on increasing physical activity. Through this, Nike seeks to solidify its reputation as a sports authority. Additionally, the coach explicitly conveys Nike's values and goals, educating customers. Experts (Hazan et al., 2022 ) and researchers (Dwivedi et al., 2022 ) have underscored the metaverse's unique ability to blend gaming with e-commerce in that manner. Figure 7 depicts a screenshot of Nikeland with the previously discussed elements highlighted.

Playing games in Nikeland, Nike’s virtual world on the Roblox platform

We caution against assuming that the net benefits for complementors can be definitively assessed from an external perspective. Nevertheless, Decentraland and Roblox demonstrate consideration for several identified characteristics conducive to complementors' value creation. Both platforms also exhibit limitations within certain dimensions, which, from our observation, detract from their overall performance. This may stem from the platforms' evolving nature, underscoring the imperative for ongoing enhancement. Moreover, the presence of the variables from our analysis in both platforms is irrespective of their governance structure . This reinforces the view that the organizational structure of a metaverse platform is not the primary distinguishing factor for complementors. Our interview findings, illustrated in Fig. 2 , support this observation.

6 Contributions and Implications

Our study offers several contributions to researchers and practitioners. Theoretically, our study enriches understanding of ecosystem dynamics in metaverse platforms. Specifically, we explored the role and relationships of complementors within metaverse platforms. We applied the D&M IS success model, a well-established framework, to examine the platform-business model fit across different contexts. From its six dimensions, we pinpointed 26 characteristics specific to metaverse platforms. These characteristics significantly impact complementors' value creation in metaverse environments. They progress the conceptualization of the of metaverse’s capabilities and the tensions with which complementors must deal within the metaverse ecosystem. Consequently, our study encourages researchers to explore the metaverse's nature, meaning, and ecosystem players further.

From a practitioner’s perspective, our study guides both orchestrators and complementors. For metaverse decision-makers, our findings highlight essential platform characteristics vital for effective value creation. This understanding can lead to complementors aligning with platforms that resonate with their business models and, conversely, enable orchestrators to refine their platforms to support organizational value creation better. The implications are dual: improving the alignment between business models and platforms, and fostering platform evolution to better facilitate value creation. Ultimately, these insights could promote stronger business cases in metaverse platforms, accelerating adoption among businesses and consumers.

7 Limitations and Future Research

Our study’s limitations provide grounds for future research. First, a consensus on defining the metaverse remains elusive. While conceptual papers discuss the metaverse's nature (e.g., Hadi et al., 2023 ), others like Peukert et al. ( 2022 ) highlight its ongoing evolution, complicating its definition. While this study aims to contribute to the technological design choices during its development, we also want to highlight the need to replicate our findings in the future to determine their validity over time. Second, our sample spanned experts from the B2B and B2C contexts and several industries, resulting in a reasonably general framework. Future research could replicate this study with industry-specific experts to explore how metaverse applications can be customized to meet distinct industry needs. Third, most of our interviewees were from Europe, with only one from Hong Kong and another from the USA. Therefore, cultural differences might yield varied findings. Given the concentration of metaverse developments in North America and Asia, conducting studies in these regions could uncover additional insights. Finally, our interviewees worked at companies already involved in metaverses or actively considering them, which could have biased our results. Future research should engage experts holding critical views on the metaverse to contrast their perspectives with our findings.

8 Conclusion

This study aimed to highlight key characteristics of metaverse platforms crucial for complementors seeking to maximize their value creation. Our research draws on platform ecosystem and value creation theories to provide a foundational understanding of how complementors generate value on metaverse platforms. We interviewed 15 metaverse decision-makers across various organizations, identifying 26 characteristics of metaverse platforms that impact complementors' ability to create value. We structured these characteristics according to the six dimensions of the DeLone and McLean IS success model and exemplified them through Decentraland and Roblox.

The journey to fully comprehend metaverse platforms is ongoing. This study clarifies the design characteristics of metaverse platforms and introduces a framework to aid complementors in choosing suitable platforms. Additionally, it empowers researchers and practitioners to design new metaverse platforms purposefully. Ultimately, our research seeks to lay the groundwork for future inquiries in this domain.

Data Availability

All data supporting the findings of this study are available within the paper and its Appendix. Any further data are available from the corresponding author upon reasonable request.

All quotes were translated to English for this paper.

Adner, R. (2017). Ecosystem as structure: An actionable construct for strategy. Journal of Management, 43 (1), 39–58.

Article Google Scholar

Ahlan, A. R. (2014). Implementation of input-process-output model for measuring information system project success. TELKOMNIKA Indonesian Journal of Electrical Engineering, 12 (7), 5603–5612.

Google Scholar

Al-Kofahi, M. K., Hassan, H., Mohamad, R., Intan, T., & Com, M. (2020). Information systems success model: A review of literature. International Journal of Innovation, Creativity and Change, 12 (8)

Ataman, K., Myhr, N., & Nistor, C. (2023). Disruptive Innovation as a Network Dilemma: A Conceptual Model. Journal of Behavioral and Applied Management, 23 (2), 104–113.

Ayiter, E. (2012). Further dimensions: Text, typography and play in the Metaverse. International Conference on Cyberworlds

Badruddoja, S., Dantu, R., He, Y., Thompson, M., Salau, A., & Upadhyay, K. (2022). Trusted AI with Blockchain to Empower Metaverse. Fourth International Conference on Blockchain Computing and Applications (BCCA) .

Bain, M. (2023). How Web3 Is Shaking Up Digital Marketing. Business of Fashion. https://www.businessoffashion.com/articles/technology/the-state-of-fashion-2023-report-web3-digital-marketing-nfts/

Balica, R. -Ş, Majerová, J., & Cuţitoi, A.-C. (2022). Metaverse applications, technologies, and infrastructure: Predictive algorithms, real-time customer data analytics, and virtual navigation tools. Linguistic and Philosophical Investigations, 21 , 219–235.

Bayraktar, E., Tatoglu, E., Aydiner, A. S., & Delen, D. (2023). Business analytics adoption and technological intensity: An efficiency analysis. Information Systems Frontiers, 1–18

BMW. (2023). The iX1 Metaverse Experience. Retrieved February 2023 from https://www.bmw.com/en/events/dreams.html

Bonina, C., & Eaton, B. (2020). Cultivating open government data platform ecosystems through governance: Lessons from Buenos Aires, Mexico City and Montevideo. Government Information Quarterly, 37 (3), 101479.

Boughzala, I., de Vreede, G.-J., & Limayem, M. (2012). Team collaboration in virtual worlds: Editorial to the special issue. Journal of the Association for Information Systems, 13 (10), 715–734.

Bourlakis, M., Papagiannidis, S., & Li, F. (2009). Retail spatial evolution: Paving the way from traditional to metaverse retailing. Electronic Commerce Research, 9 (1), 135–148.

Bowman, C., & Ambrosini, V. (2000). Value creation versus value capture: Towards a coherent definition of value in strategy. British Journal of Management, 11 (1), 1–15.

Braud, A., Fromentoux, G., Radier, B., & Le Grand, O. (2021). The Road to European Digital Sovereignty with Gaia-X and IDSA. IEEE Network, 35 (2), 4–5.

Brimco. (2024). Business Statistics You Should Know in 2024. Retrieved March 2024 from https://www.brimco.io/business-statistics-to-know/#:~:text=There%20are%20an%20estimated%20334%20million%20companies%20worldwide&text=This%20is%20a%20marginal%20growth%20from%202021%20when%20there%20were%20314.21%20million .

Brooke, C. (2022). Best Metaverse Platforms to Play in 2022 . Business2Community. https://www.business2community.com/cryptocurrency/best-metaverse-platforms

Choi, H.-S., & Kim, S.-H. (2017). A content service deployment plan for metaverse museum exhibitions—Centering on the combination of beacons and HMDs. International Journal of Information Management, 37 (1), 1519–1527.

Davis, A., Murphy, J. D., Owens, D., Khazanchi, D., & Zigurs, I. (2009). Avatars, people, and virtual worlds: Foundations for research in metaverses. Journal of the Association for Information Systems, 10 (2), 90–117.

Decentraland. (2022). How Many DAU Does Decentraland Have? https://decentraland.org/blog/announcements/how-many-dau-does-decentraland-have/

DeLone, W. H., & McLean, E. R. (2002). Information systems success revisited. Proceedings of the 35th Annual Hawaii International Conference on System Sciences

DeLone, W. H., & McLean, E. R. (1992). Information systems success: The quest for the dependent variable. Information Systems Research, 3 (1), 60–95.

DeLone, W. H., & McLean, E. R. (2003). The DeLone and McLean model of information systems success: A ten-year update. Journal of Management Information Systems, 19 (4), 9–30.

Dionisio, J. D. N., III. (2013). 3D virtual worlds and the metaverse: Current status and future possibilities. ACM Computing Surveys (CSUR), 45 (3), 1–38.

Di Pietro, R., & Cresci, S. (2021). Metaverse: Security and Privacy Issues. IEEE International Conference on Trust, Privacy and Security in Intelligent Systems and Applications (TPS-ISA), 3 , 281–288

Draper, J. (2004). The relationship between research question and research design. In Research into Practice: Essential Skills for Reading and Applying Research in Nursing and Health Care (Vol. 2, pp. 69–84). Bailliere Tindall

Duan, H., Li, J., Fan, S., Lin, Z., Wu, X., & Cai, W. (2021, October). Metaverse for social good: A university campus prototype. In: Proceedings of the 29th ACM international conference on multimedia, pp 153–161. https://doi.org/10.1145/3474085.3479238

Duffy, C. (2022). Twitter relaunches option to pay for blue check marks. CNN Business . https://edition.cnn.com/2022/12/12/tech/twitter-verification-relaunch/index.html

Dwivedi, Y. K., Hughes, L., Wang, Y., Alalwan, A. A., Ahn, S. J., Balakrishnan, J., Barta, S., Belk, R., Buhalis, D., & Dutot, V. (2022). Metaverse marketing: How the metaverse will shape the future of consumer research and practice. Psychology & Marketing, 40 (4), 750–76.

Falchuk, B., Loeb, S., & Neff, R. (2018). The social metaverse: Battle for privacy. IEEE Technology and Society Magazine, 37 (2), 52–61.

Friedrich, T., Schlauderer, S., & Overhage, S. (2019). The impact of social commerce feature richness on website stickiness through cognitive and affective factors: An experimental study. Electronic Commerce Research and Applications, 36 , 100861.

Gadalla, E., Keeling, K., & Abosag, I. (2013). Metaverse-retail service quality: A future framework for retail service quality in the 3D internet. Journal of Marketing Management, 29 (13–14), 1493–1517.

Gartner. (2022). Estimated metaverse use case among consumers and businesses worldwide in 2026. Statista. Retrieved January 2022 from https://www.statista.com/statistics/1290160/projected-metaverse-use-reach-global-consumers-businesses/

Gawer, A., & Cusumano, M. A. (2002). Platform leadership: How Intel, Microsoft, and Cisco drive industry innovation (Vol. 5). Harvard Business School Press Boston, MA

Gawer, A., & Cusumano, M. A. (2014). Industry platforms and ecosystem innovation. Journal of Product Innovation Management, 31 (3), 417–433.

Getchell, K., Oliver, I., Miller, A., & Allison, C. (2010). Metaverses as a platform for game based learning. IEEE International Conference on Advanced Information Networking and Applications, 24 , 1195–1202.

Gioia, D. A., Corley, K. G., & Hamilton, A. L. (2013). Seeking qualitative rigor in inductive research: Notes on the Gioia methodology. Organizational Research Methods, 16 (1), 15–31.

Goldberg, M., & Schär, F. (2023). Metaverse governance: An empirical analysis of voting within Decentralized Autonomous Organizations. Journal of Business Research, 160 , 113764.

Golf-Papez, M., Heller, J., Hilken, T., Chylinski, M., de Ruyter, K., Keeling, D. I., & Mahr, D. (2022). Embracing falsity through the metaverse: The case of synthetic customer experiences. Business Horizons, 65 (6), 739–749.

Gollmer, P. (2022). Web3: Die Internet-Revolution, die aus der Abhängigkeit von Big Tech führen soll. Neue Züricher Zeitung. https://www.nzz.ch/technologie/ist-web3-die-zukunft-des-internets-oder-nur-ein-weiterer-hype-ld.1664275?reduced=true

Guest, G., Bunce, A., & Johnson, L. (2006). How many interviews are enough? An experiment with data saturation and variability. Field Methods, 18 (1), 59–82.

Guidi, B., & Michienzi, A. (2022). SocialFi: towards the new shape of social media. ACM SIGWEB Newsletter (Summer), 1–8

Guo, J., Liu, Z., & Liu, Y. (2016). Key success factors for the launch of government social media platform: Identifying the formation mechanism of continuance intention. Computers in Human Behavior, 55 , 750–763.

Hadi, R., Melumad, S., & Park, E. S. (2023). The Metaverse: A New Digital Frontier for Consumer Behavior. Journal of Consumer Psychology, AHEAD-OF-PRINT.

Harzing, A.-W., & Maznevski, M. (2002). The interaction between language and culture: A test of the cultural accommodation hypothesis in seven countries. Language and Intercultural Communication, 2 (2), 120–139.

Hazan, E., Kelly, G., Khan, H., Spillecke, D., & Yee, L. (2022). Marketing in the metaverse: An opportunity for innovation and experimentation.

Hwang, R., & Lee, M. (2022). The Influence of Music Content Marketing on User Satisfaction and Intention to Use in the Metaverse: A Focus on the SPICE Model. Businesses, 2 (2), 141–155.

Jaynes, C., Seales, W. B., Calvert, K., Fei, Z., & Griffioen, J. (2003). The Metaverse: a networked collection of inexpensive, self-configuring, immersive environments. Proceedings of the Workshop on Virtual Environments.

Jeong, S.-H., & Kim, H.-K. (2023). Effect of Trust in Metaverse on Usage Intention through Technology Readiness and Technology Acceptance Model. Tehnički Vjesnik, 30 (3), 837–845.

Jo, H., & Park, S. (2022). Roles of Information Security and Benefits in Forming User Satisfaction of Metaverse. International Conference on Computational Science and Computational Intelligence (CSCI)

Kamin, D. (2021). Investors Snap Up Metaverse Real Estate in a Virtual Land Boom. New York Times. https://www.nytimes.com/2021/11/30/business/metaverse-real-estate.html

Khansulivong, C., Wicha, S., & Temdee, P. (2022). Adaptive of New Technology for Agriculture Online Learning by Metaverse: A Case Study in Faculty of Agriculture, National University of Laos. Joint International Conference on Digital Arts, Media and Technology with ECTI Northern Section Conference on Electrical, Electronics, Computer and Telecommunications Engineering (ECTI DAMT & NCON) , 428–432.

Kim, J. (2021). Advertising in the Metaverse: Research Agenda. Journal of Interactive Advertising, 21 (3), 141–144.

Kim, H. J., Kim, I., & Lee, H. (2016). Third-party mobile app developers’ continued participation in platform-centric ecosystems: An empirical investigation of two different mechanisms. International Journal of Information Management, 36 (1), 44–59.

Kim, D. J., Salvacion, M., Salehan, M., & Kim, D. W. (2022). An empirical study of community cohesiveness, community attachment, and their roles in virtual community participation. European Journal of Information Systems, 1–28.

Krasnova, H., Veltri, N. F., Eling, N., & Buxmann, P. (2017). Why men and women continue to use social networking sites: The role of gender differences. The Journal of Strategic Information Systems, 26 (4), 261–284.

Lang, N., von Szczepanski, K., & Wurzer, C. (2019). The emerging art of ecosystem management. Boston Consulting Group, 1–20

Lee, H.-K., Park, S., & Lee, Y. (2022). A proposal of virtual museum metaverse content for the MZ generation. Digital Creativity, 33 (2), 1–17.

Lin, K.-Y., & Lu, H.-P. (2011). Why people use social networking sites: An empirical study integrating network externalities and motivation theory. Computers in Human Behavior, 27 (3), 1152–1161.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry . Sage.

Book Google Scholar

Lo Iacono, V., Symonds, P., & Brown, D. H. (2016). Skype as a tool for qualitative research interviews. Sociological Research Online, 21 (2), 103–117.

Mac, R., Mullin, B., Conger, K., & Isaac, M. (2022). A Verifiable Mess: Twitter Users Create Havoc by Impersonating Brands. New York Times. https://www.nytimes.com/2022/11/11/technology/twitter-blue-fake-accounts.html

Marabelli, M., & Newell, S. (2022). Everything you Always Wanted to Know About the Metaverse* (*But Were Afraid to Ask). Academy of Management Annual Meeting, 1–40.

Marr, B. (2022). The Amazing Ways Nike Is Using The Metaverse, Web3 And NFTs. Forbes. https://www.forbes.com/sites/bernardmarr/2022/06/01/the-amazing-ways-nike-is-using-the-metaverse-web3-and-nfts/?sh=62e049e256e9

Martins, D., Oliveira, L., & Amaro, A. C. (2022). Life in Second Life through the eyes of the residents. Iberian Conference on Information Systems and Technologies (CISTI), 17 , 1–6.

McKinsey. (2022). Value creation in the metaverse: The real business of the virtual world. McKinsey Research. https://www.mckinsey.com/~/media/mckinsey/business%20functions/marketing%20and%20sales/our%20insights/value%20creation%20in%20the%20metaverse/Value-creation-in-the-metaverse.pdf

Merriam, S. B., & Grenier, R. S. (2019). Qualitative research in practice: Examples for Discussion and Analysis. John Wiley & Sons

Metcalf, M. (2022). How Creativity & Deep Audience Knowledge Made Roblox A Sensation. Latana. https://latana.com/post/roblox-deep-dive/

Mladenović, D., Ismagilova, E., Filieri, R., & Dwivedi, Y. K. (2023). MetaWOM–toward a sensory word-of-mouth (WOM) in the metaverse. International Journal of Contemporary Hospitality Management

Molla, A., & Licker, P. S. (2001). E-commerce systems success: An attempt to extend and respecify the Delone and MacLean model of IS success. Journal of Electronic Commerce Research, 2 (4), 131–141.

Moy, C., & Gadgil, A. (2022). Opportunities in the metaverse: How businesses can explore the metaverse and navigate the hype vs. reality . Onyx by J.P. Morgan. Retrieved March 2022 from https://www.jpmorgan.com/content/dam/jpm/treasury-services/documents/opportunities-in-the-metaverse.pdf

Newzoo. (2022). The Metaverse, Blockchain Gaming, and NFTs: Navigating the Internet’s Uncharted Waters . Retrieved March 2024 from https://resources.newzoo.com/hubfs/Reports/Metaverse/2022_Newzoo_Metaverse_Blockchain_NFT_Report.pdf?utm_campaign=Metaverse%20Report%202022&utm_medium=email&_hsmi=217738092&_hsenc=p2ANqtz--lsLjwfkn5vmoHZ3vlIlMt5jH5co1McK_yqbxz5Sjt76ruFm6Hc1lnkiXT1Iu_HKHixVUYNGAFyG0THr5G4wRIMUc0SZ4GcdoW--edcMxde4v1eq8&utm_content=217738092&utm_source=hs_automation

Nike. (2023). Nikeland on Roblox. https://www.nike.com/kids/nikeland-roblox

Ning, H., Wang, H., Lin, Y., Wang, W., Dhelim, S., Farha, F., Ding, J., & Daneshmand, M. (2021). A Survey on Metaverse: the State-of-the-art, Technologies, Applications, and Challenges. arXiv preprint , arXiv:2111.09673 , 1–34.

Oliver, I. A., Miller, A. H., & Allison, C. (2010). Virtual worlds, real traffic: interaction and adaptation. Proceedings of the Annual ACM SIGMM Conference on Multimedia Systems

Oliveira, S., Barros, M. I., Lima, G. D., Farias, L. B. (2019). Investigations into Data Ecosystems: A systematic mapping study. Knowledge and Information Systems, 61 (2), 589–630.

Orgaz, G. B., R-moreno, M. D., Camacho, D., & Barrero, D. F. (2012). Clustering avatars behaviours from virtual worlds interactions. Proceedings of the 4th International Workshop on Web Intelligence & Communities

Owens, D., Mitchell, A., Khazanchi, D., & Zigurs, I. (2011). An empirical investigation of virtual world projects and metaverse technology capabilities. ACM SIGMIS Database: The DATABASE for Advances in Information Systems, 42 (1), 74–101.

Papagiannidis, S., Bourlakis, M., & Li, F. (2008). Making real money in virtual worlds: MMORPGs and emerging business opportunities, challenges and ethical implications in metaverses. Technological Forecasting and Social Change, 75 (5), 610–622.

Park, S.-M., & Kim, Y.-G. (2022). A Metaverse: Taxonomy, components, applications, and open challenges. IEEE Access, 10 , 4209–4251.

Paterson, M., & Higgs, J. (2005). Using hermeneutics as a qualitative research approach in professional practice. Qualitative Report, 10 (2), 339–357.

Periyasami, S., & Periyasamy, A. P. (2022). Metaverse as future promising platform business model: Case study on fashion value chain. Businesses, 2 (4), 527–545.

Petter, S., DeLone, W., & McLean, E. (2008). Measuring information systems success: Models, dimensions, measures, and interrelationships. European Journal of Information Systems, 17 , 236–263.

Peukert, C., Weinhardt, C., Hinz, O., & van der Aalst, W. M. (2022). Metaverse: How to Approach Its Challenges from a BISE Perspective. Business & Information Systems Engineering, 64 , 401–406.

Polyviou, A., & Pappas, I. O. (2022). Chasing metaverses: Reflecting on existing literature to understand the business value of metaverses. Information Systems Frontiers, 25 , 2417–28.

Popescu, G. H., Valaskova, K., & Horak, J. (2022). Augmented Reality Shopping Experiences, Retail Business Analytics, and Machine Vision Algorithms in the Virtual Economy of the Metaverse. Journal of Self-Governance & Management Economics, 10 (2), 67–81.

Quintín, E., Sanz, C., & Zangara, A. (2016). The impact of role-playing games through Second Life on the oral practice of linguistic and discursive sub-competences in English. International Conference on Collaboration Technologies and Systems (CTS), 148–155

Sanders, L., & Simons, G. (2009). A social vision for value co-creation in design. Open Source Business Resource(December 2009).

Schöbel, S. M., & Leimeister, J. M. (2023). Metaverse platform ecosystems. Electronic Markets, 33 (1), 12.

Schöbel, S. M., & Tingelhoff, F. (2023). Overcoming Challenges to Enable the Potential of Metaverse Platforms — A Qualitative Approach to Understand Value Creation. AIS Transactions of Human-Computer Interaction, 15 (1), 1–23.

Schöbel, S. M., Karatas, J., Tingelhoff, F., & Leimeister, J. M. (2023). Not everything is a Metaverse?! A Practitioners Perspective on Characterizing Metaverse Platforms. Hawaii International Conference on System Science, 56 , 544–553.

Seckler, M., Heinz, S., Forde, S., Tuch, A. N., & Opwis, K. (2015). Trust and distrust on the web: User experiences and website characteristics. Computers in Human Behavior, 45 , 39–50.

Shin, D. (2022). The actualization of meta affordances: Conceptualizing affordance actualization in the metaverse games. Computers in Human Behavior, 133 , 1–13.

Sledgianowski, D., & Kulviwat, S. (2008). Social network sites: Antecedents of user adoption and usage

Smith, S. (2022). Givenchy Parfums enters Metaverse with Roblox. TheIndustry.Beauty. https://theindustry.beauty/givenchy-parfums-enters-metaverse-with-roblox/

Stephen Gates. (n.d.). Aloft Hotels in Second Life . Retrieved December 2022 from https://stephengates.com/portfolio/aloft-hotels-second-life/

Sykownik, P., Maloney, D., Freeman, G., & Masuch, M. (2022). Something Personal from the Metaverse: Goals, Topics, and Contextual Factors of Self-Disclosure in Commercial Social VR. CHI Conference on Human Factors in Computing Systems , 1–17.

Tapscott, D., & Tapscott, A. (2017). How blockchain will change organizations. MIT Sloan Management Review, 58 (2), 10–13.

Tawaststjerna, T., & Olander, H. (2021). Managing digital transformation in digital business ecosystems. International Journal of Innovation Management, 25 (10), 2140003.

Tayal, S., Rajagopal, K., & Mahajan, V. (2022). Virtual Reality based Metaverse of Gamification. International Conference on Computing Methodologies and Communication (ICCMC) , 1597–1604

Tingelhoff, F., Klug, S., & Elshan, E. (2024). The Diffusion of the Metaverse: How YouTube Influencers Shape Metaverse Mass Adoption and What This Means for Organizations. California Management Review , forthcoming.

Torkzadeh, G., & Dhillon, G. (2002). Measuring factors that influence the success of Internet commerce. Information Systems Research, 13 (2), 187–204.

Valenzuela, D., & Shrivastava, P. (2002). Interview as a method for qualitative research (Presentation). Southern Cross University and the Southern Cross Institute of Action Research (SCIAR) . https://d1wqtxts1xzle7.cloudfront.net/31820583/Interview_as_a_Method_for_QR-libre.pdf?1390950384=&response-content-disposition=inline%3B+filename%3DInterview_as_a_Method_for_QR.pdf&Expires=1698950691&Signature=HRcpfWOlmyQlAQWJesqeOUaT9WdxxfRUOJ-o~ADETWEgE-Gd7QQ2ILhPYEyXo8DGGTEBwfy4g4Fa4nyQoxiOTSybbmNAZ6Dvyfd19KsekeygVuJtJCJgQA2jtMD21L1d4m3Fqts8Vvee5uTwFQ5sUlMzQsEnhOzBW-0Kr7N-zFDO9nl4MA3treOlaaTiXScP84p~EN~lnnhFPUsPM1sNPkWAvFT17Nj8m0PGQE8Yx07XsstcnWDAHLbSdJIBChNoDpNsHoEHLgVEW4mEE-wn2YouqJMyf~gdzGreBlza-XFcD3AGgJzJwDTxPTF3p78MQGqXunb7UeBDP4fKvIhCNw__&Key-Pair-Id=APKAJLOHF5GGSLRBV4ZA

Vidal-Tomás, D. (2022). The new crypto niche: NFTs, play-to-earn, and metaverse tokens. Finance Research Letters, 47 , 1–13.

Wagner, C., Liu, L., Schneider, C., Prasarnphanich, P., & Chen, H. (2009). Creating a successful professional virtual community: A sustainable digital ecosystem for idea sharing. 3rd IEEE International Conference on Digital Ecosystems and Technologies.

Wang, Y. S. (2008). Assessing e-commerce systems success: A respecification and validation of the DeLone and McLean model of IS success. Information Systems Journal, 18 (5), 529–557.

Wang, M., & Ho, C. (2022). Metaverse Platforms — Roblox Overtakes Decentraland (MANA) and Sandbox (SAND). Bybit . https://blog.bybit.com/en-US/post/metaverse-platforms-roblox-overtakes-decentraland-mana-and-sandbox-sand--bltc3d581e1ae0fa13e/

Wang, D., Yan, X., & Zhou, Y. (2021). Research on Metaverse: Concept, development and standard system. International Conference on Electronics, Communications and Information Technology (CECIT) , 2 , 983–991

Wang, Y., Su, Z., Zhang, N., Xing, R., Liu, D., Luan, T. H., & Shen, X. (2022). A survey on metaverse: Fundamentals, security, and privacy. IEEE Communications Surveys & Tutorials 25

Weber, M., Engert, M., Schaffer, N., Weking, J., & Krcmar, H. (2023). Organizational Capabilities for AI Implementation—Coping with Inscrutability and Data Dependency in AI. Information Systems Frontiers, 25 (4), 1549–1569.

Xi, N., Chen, J., Gama, F., Riar, M., & Hamari, J. (2023). The challenges of entering the metaverse: An experiment on the effect of extended reality on workload. Information Systems Frontiers, 25 (2), 659–680.

Yang, D., Zhou, J., Chen, R., Song, Y., Song, Z., Zhang, X., Wang, Q., Wang, K., Zhou, C., & Sun, J. (2022a). Expert consensus on the metaverse in medicine. Clinical eHealth, 5 , 1–9.

Yang, Q., Zhao, Y., Huang, H., Xiong, Z., Kang, J., & Zheng, Z. (2022b). Fusing blockchain and AI with metaverse: A survey. IEEE Open Journal of the Computer Society, 3 , 122–136.

Yang, X. J., Unhelkar, V. V., Li, K., & Shah, J. A. (2017). Evaluating effects of user experience and system transparency on trust in automation. Proceedings of the 2017 ACM/IEEE international conference on human-robot interaction.

Yoo, K., Welden, R., Hewett, K., & Haenlein, M. (2023). The merchants of meta: A research agenda to understand the future of retailing in the metaverse. Journal of Retailing, 99 (2), 173–192.

Zeithaml, V. A. (1988). Consumer perceptions of price, quality, and value: A means-end model and synthesis of evidence. Journal of Marketing, 52 (3), 2–22.

Download references

Acknowledgements

For this research, Fabian Tingelhoff received funding from the Konrad-Adenauer-Foundation (KAS).

Open access funding provided by University of St.Gallen

Author information

Authors and affiliations.

Institute of Information Systems and Digital Business, University of St.Gallen, St. Gallen, Switzerland

Fabian Tingelhoff, Raphael Schultheiss & Jan Marco Leimeister

Information Systems, University of Osnabrück, Osnabrück, Germany

Sofia Marlena Schöbel

You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Fabian Tingelhoff .

Ethics declarations

Conflict of interest.

The authors declare that they had no conflict of interest when constructing this manuscript.

Additional information

Publisher's note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

See Table 1

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/ .

Reprints and permissions

About this article

Tingelhoff, F., Schultheiss, R., Schöbel, S.M. et al. Qualitative Insights into Organizational Value Creation: Decoding Characteristics of Metaverse Platforms. Inf Syst Front (2024). https://doi.org/10.1007/s10796-024-10494-x

Download citation

Accepted : 02 May 2024

Published : 15 May 2024

DOI : https://doi.org/10.1007/s10796-024-10494-x

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Metaverse Platform Ecosystem
Organizational Value Creation
Platform Characteristics
Complementors
Find a journal
Publish with us
Track your research

Open access
Published: 18 May 2024

Determinants of appropriate antibiotic and NSAID prescribing in unscheduled outpatient settings in the veterans health administration

Michael J. Ward 1 , 2 , 3 , 4 ,
Michael E. Matheny 1 , 4 , 5 , 6 ,
Melissa D. Rubenstein 3 ,
Kemberlee Bonnet 7 ,
Chloe Dagostino 7 ,
David G. Schlundt 7 ,
Shilo Anders 4 , 8 ,
Thomas Reese 4 &
Amanda S. Mixon 1 , 9

BMC Health Services Research volume 24 , Article number: 640 ( 2024 ) Cite this article

45 Accesses

1 Altmetric

Metrics details

Despite efforts to enhance the quality of medication prescribing in outpatient settings, potentially inappropriate prescribing remains common, particularly in unscheduled settings where patients can present with infectious and pain-related complaints. Two of the most commonly prescribed medication classes in outpatient settings with frequent rates of potentially inappropriate prescribing include antibiotics and nonsteroidal anti-inflammatory drugs (NSAIDs). In the setting of persistent inappropriate prescribing, we sought to understand a diverse set of perspectives on the determinants of inappropriate prescribing of antibiotics and NSAIDs in the Veterans Health Administration.

We conducted a qualitative study guided by the Consolidated Framework for Implementation Research and Theory of Planned Behavior. Semi-structured interviews were conducted with clinicians, stakeholders, and Veterans from March 1, 2021 through December 31, 2021 within the Veteran Affairs Health System in unscheduled outpatient settings at the Tennessee Valley Healthcare System. Stakeholders included clinical operations leadership and methodological experts. Audio-recorded interviews were transcribed and de-identified. Data coding and analysis were conducted by experienced qualitative methodologists adhering to the Consolidated Criteria for Reporting Qualitative Studies guidelines. Analysis was conducted using an iterative inductive/deductive process.

We conducted semi-structured interviews with 66 participants: clinicians ( N = 25), stakeholders ( N = 24), and Veterans ( N = 17). We identified six themes contributing to potentially inappropriate prescribing of antibiotics and NSAIDs: 1) Perceived versus actual Veterans expectations about prescribing; 2) the influence of a time-pressured clinical environment on prescribing stewardship; 3) Limited clinician knowledge, awareness, and willingness to use evidence-based care; 4) Prescriber uncertainties about the Veteran condition at the time of the clinical encounter; 5) Limited communication; and 6) Technology barriers of the electronic health record and patient portal.

Conclusions

The diverse perspectives on prescribing underscore the need for interventions that recognize the detrimental impact of high workload on prescribing stewardship and the need to design interventions with the end-user in mind. This study revealed actionable themes that could be addressed to improve guideline concordant prescribing to enhance the quality of prescribing and to reduce patient harm.

Peer Review reports

Adverse drug events (ADEs) are the most common iatrogenic injury. [ 1 ] Efforts to reduce these events have primarily focused on the inpatient setting. However, the emergency department (ED), urgent care, and urgent primary care clinics are desirable targets for interventions to reduce ADEs because approximately 70% of all outpatient encounters occur in one of these settings. [ 2 ] Two of the most commonly prescribed drug classes during acute outpatient care visits that have frequent rates of potentially inappropriate prescribing include antibiotics and non-steroidal anti-inflammatory drugs (NSAIDs). [ 3 , 4 ]

An estimated 30% of all outpatient oral antibiotic prescriptions may be unnecessary. [ 5 , 6 ] The World Health Organization identified overuse of antibiotics and its resulting antimicrobial resistance as a global threat. [ 7 ] The Centers for Disease Control and Prevention (CDC) conservatively estimates that in the US there are nearly 3 million antibiotic-resistant infections that cause 48,000 deaths annually. [ 8 ] Antibiotics were the second most common source of adverse events with nearly one ADE resulting in an ED visit for every 100 prescriptions. [ 9 ] Inappropriate antibiotic prescriptions (e.g., antibiotic prescription for a viral infection) also contribute to resistance and iatrogenic infections such as C. difficile (antibiotic associated diarrhea) and Methicillin-resistant Staphylococcus aureus (MRSA) . [ 8 ] NSAID prescriptions, on the other hand, result in an ADE at more than twice the rate of antibiotics (2.2%), [ 10 ] are prescribed to patients at an already increased risk of potential ADEs, [ 4 , 11 ] and frequently interact with other medications. [ 12 ] Inappropriate NSAID prescriptions contribute to serious gastrointestinal, [ 13 ] renal, [ 14 ] and cardiovascular [ 15 , 16 ] ADEs such as gastrointestinal bleeding, acute kidney injury, and myocardial infarction or heart failure, respectively. Yet, the use of NSAIDs is ubiquitous; according to the CDC, between 2011 and 2014, 5% of the US population were prescribed an NSAID whereas an additional 2% take NSAIDs over the counter. [ 11 ]

Interventions to reduce inappropriate antibiotic prescribing commonly take the form of antimicrobial stewardship programs. However, no such national programs exist for NSAIDs, particularly in acute outpatient care settings. There is a substantial body of evidence supporting the evidence of such stewardship programs. [ 17 ] The CDC recognizes that such outpatient programs should consist of four core elements of antimicrobial stewardship, [ 18 ] including commitment, action for policy and practice, tracking and reporting, and education and expertise. However, the opportunities to extend antimicrobial stewardship in EDs are vast. Despite the effectiveness, there is a recognized need to understand which implementation strategies and how to implement multifaceted interventions. [ 19 ] Given the unique time-pressured environment of acute outpatient care settings, not all antimicrobial stewardship strategies work in these settings necessitating the development of approaches tailored to these environments. [ 19 , 20 ]

One particularly vulnerable population is within the Veterans Health Administration. With more than 9 million enrollees in the Veterans Health Administration, Veterans who receive care in Veteran Affairs (VA) hospitals and outpatient clinics may be particularly vulnerable to ADEs. Older Veterans have greater medical needs than younger patients, given their concomitant medical and mental health conditions as well as cognitive and social issues. Among Veterans seen in VA EDs and Urgent Care Clinics (UCCs), 50% are age 65 and older, [ 21 ] nearly three times the rate of non-VA emergency care settings (18%). [ 22 ] Inappropriate prescribing in ED and UCC settings is problematic with inappropriate antibiotic prescribing estimated to be higher than 40%. [ 23 ] In a sample of older Veterans discharged from VA ED and UCC settings, NSAIDs were found to be implicated in 77% of drug interactions. [ 24 ]

Learning from antimicrobial stewardship programs and applying to a broader base of prescribing in acute outpatient care settings, it is necessary to understand not only why potentially inappropriate prescribing remains a problem for antibiotics, but for medications (e.g., NSAIDs) which have received little stewardship focus previously. This understanding is essential to develop and implement interventions to reduce iatrogenic harm for vulnerable patients seen in unscheduled settings. In the setting of the Veterans Health Administration, we sought to use these two drug classes (antibiotics and NSAIDs) that have frequent rates of inappropriate prescribing in unscheduled outpatient care settings, to understand a diverse set of perspectives on why potentially inappropriate prescribing continues to occur.

Selection of participants

Participants were recruited from three groups in outpatient settings representing emergency care, urgent care, and urgent primary care in the VA: 1) Clinicians-VA clinicians such as physicians, advanced practice providers, and pharmacists 2) Stakeholders-VA and non-VA clinical operational and clinical content experts such as local and regional medical directors, national clinical, research, and administrative leadership in emergency care, primary care, and pharmacy including geriatrics; and 3) Veterans seeking unscheduled care for infectious or pain symptoms.

Clinicians and stakeholders were recruited using email, informational flyers, faculty/staff meetings, national conferences, and snowball sampling, when existing participants identify additional potential research subjects for recruitment. [ 25 ] Snowball sampling is useful for identifying and recruiting participants who may not be readily apparent to investigators and/or hard to reach. Clinician inclusion criteria consisted of: 1) at least 1 year of VA experience; and 2) ≥ 1 clinical shift in the last 30 days at any VA ED, urgent care, or primary care setting in which unscheduled visits occur. Veterans were recruited in-person at the VA by key study personnel. Inclusion criteria consisted of: 1) clinically stable as determined by the treating clinician; 2) 18 years or older; and 3) seeking care for infectious or pain symptoms in the local VA Tennessee Valley Healthcare System (TVHS). TVHS includes an ED at the Nashville campus with over 30,000 annual visits, urgent care clinic in Murfreesboro, TN with approximately 15,000 annual visits, and multiple primary care locations throughout the middle Tennessee region. This study was approved by the VA TVHS Institutional Review Board as minimal risk.

Data collection

Semi-structured interview guides (Supplemental Table 1) were developed using the Consolidated Framework for Implementation Research (CFIR) [ 26 ] and the Theory of Planned Behavior [ 27 , 28 ] to understand attitudes and beliefs as they relate to behaviors, and potential determinants of a future intervention. Interview guides were modified and finalized by conducting pilot interviews with three members of each participant group. Interview guides were tailored to each group of respondents and consisted of questions relating to: 1) determinants of potentially inappropriate prescribing; and 2) integration into practice (Table. 1 ). Clinicians were also asked about knowledge and awareness of evidence-based prescribing practices for antibiotics and NSAIDs. The interviewer asked follow-up questions to elicit clarity of responses and detail.

Each interview was conducted by a trained interviewer (MDR). Veteran interviews were conducted in-person while Veterans waited for clinical care so as not to disrupt clinical operations. Interviews with clinicians and stakeholders were scheduled virtually. All interviews (including in-person) were recorded and transcribed in a manner compliant with VA information security policies using Microsoft Teams (Redmond, WA). The audio-recorded interviews were transcribed and de-identified by a transcriptionist and stored securely behind the VA firewall using Microsoft Teams. Study personnel maintained a recording log on a password-protected server and each participant was assigned a unique participant ID number. Once 15 interviews were conducted per group, we planned to review interviews with the study team to discuss content, findings, and to decide collectively when thematic saturation was achieved, the point at which no new information was obtained. [ 29 ] If not achieved, we planned to conduct at least 2 additional interviews prior to group review for saturation. We estimated that approximately 20–25 interviews per group were needed to achieve thematic saturation.

Qualitative data coding and analysis was managed by the Vanderbilt University Qualitative Research Core. A hierarchical coding system (Supplemental Table 2) was developed and refined using an iterative inductive/deductive approach [ 30 , 31 , 32 ] guided by a combination of: 1) Consolidated Framework for Implementation Research (CFIR) [ 26 ]; 2) the Theory of Planned Behavior [ 27 , 28 ]; 3) interview guide questions; and 4) a preliminary review of the transcripts. Eighteen major categories (Supplemental Table 3) were identified and were further divided into subcategories, with some subcategories having additional levels of hierarchical division. Definitions and rules were written for the use of each of the coding categories. The process was iterative in that the coding system was both theoretically informed and derived from the qualitative data. The coding system was finalized after it was piloted by the coders. Data coding and analysis met the Consolidated Criteria for Reporting Qualitative Research (COREQ) guidelines. [ 33 ]

Four experienced qualitative coders were trained by independently coding two transcripts from each of the three participant categories. Coding was then compared, and any discrepancies resolved by reconciliation. After establishing reliability in using the coding system, the coders divided and independently coded the remaining transcripts in sequential order. Each statement was treated as a separate quote and could be assigned up to 21 different codes. Coded transcripts were combined and sorted by code.

Following thematic saturation, the frequency of each code was calculated to understand the distribution of quotes. Quotes were then cross-referenced with coding as a barrier to understand potential determinants of inappropriate prescribing. A thematic analysis of the barriers was conducted and presented in an iterative process with the research team of qualitative methodologists and clinicians to understand the nuances and refine the themes and subthemes from the coded transcripts. Transcripts, quotations, and codes were managed using Microsoft Excel and SPSS version 28.0.

We approached 132 individuals and 66 (50%) agreed to be interviewed. Participants included 25 clinicians, 24 stakeholders, and 17 Veterans whose demographic characteristics are presented in Table 2 . The clinicians were from 14 VA facilities throughout the US and 20 physicians, and five advanced practice providers. Of the clinicians, 21 (84%) worked in either an ED or urgent care while the remainder practiced in primary care. The 24 stakeholders included 13 (54%) clinical service chiefs or deputy chief (including medical directors), five (21%) national directors, and six (25%) experts in clinical content and methodology. The 17 Veterans interviewed included 15 (88%) who were seen for pain complaints.

Results are organized by the six thematic categories with several subthemes in each category. Themes and subthemes are presented in Table 3 and are visually represented in Fig. 1 . The six themes were: 1) perceived versus actual Veterans expectations about prescribing, 2) the influence of a time-pressured clinical environment on prescribing stewardship, 3) limited clinician knowledge, awareness, and willingness to use evidence-based care, 4) uncertainties about the Veteran condition at the time of the clinical encounter, 5) limited communication, and 6) technology barriers.

Visual representation of themes and subthemes from 66 clinician, stakeholder, and Veteran interviews

Theme 1: Perception that Veterans routinely expect a medication from their visit, despite clinical inappropriateness

According to clinicians, Veterans frequently expect to receive a prescription even when this decision conflicts with good clinical practice.

Certainly lots of people would say you know if you feel like you’re up against some strong expectations from the patients or caregivers or families around the utility of an antibiotic when it’s probably not indicated…In the emergency department the bias is to act and assume the worst and assume like the worst for the clinical trajectory for the patient rather than the reverse. [Clinician 49, Physician, ED]

In addition, stakeholders further stated that patient prescription expectations are quite influential and are likely shaped by Veterans’ prior experiences.

I think the patients, particularly for antibiotics, have strong feelings about whether they should or shouldn’t get something prescribed. [Stakeholder 34] You know I think the biggest challenge, I think, is adjusting patients’ expectations because you know they got better the last time they were doing an antibiotic. [Stakeholder 64]

Patient satisfaction and clinician workload may also influence the clinician’s prescription decision.

We have a lot of patients that come in with back pain or knee pain or something. We’ll get an x-ray and see there’s nothing actually wrong physically that can be identified on x-ray at least and you have to do something. Otherwise, patient satisfaction will dip, and patients leave angry. [Clinician 28, Physician, urgent care clinic] For some clinicians it’s just easier to prescribe an antibiotic when they know that’s the patient’s expectation and it shortens their in-room discussion and evaluation. [Clinician 55, Physician, ED]

Despite clinician perception, Veterans communicated that they did not necessarily expect a prescription and were instead focused on the clinical interaction and the clinician’s decision.

I’m not sure if they’ll give me [unintelligible] a prescription or what they’ll do. I don’t care as long as they stop the pain. [Patient 40, urgent care clinic] I don’t expect to [receive a prescription], but I mean whatever the doctor finds is wrong with me I will follow what he says. [Patient 31, ED]

Theme 2: Hectic clinical environments and unique practice conditions in unscheduled settings provide little time to focus on prescribing practices

Clinicians and stakeholders reported that the time-constrained clinical environment and need to move onto the next patient were major challenges to prescribing stewardship.

The number one reason is to get a patient out of your office or exam bay and move on to the next one. [Stakeholder 28] It takes a lot of time and you have to be very patient and understanding. So, you end up having to put a fair bit of emotional investment and intelligence into an encounter to not prescribe. [Stakeholder 1]

Stakeholders also noted that unique shift conditions and clinician perceptions that their patients were “different” might influence prescribing practices.

A common pushback was ‘well my patients are different.’ [Stakeholder 4] Providers who worked different types of shifts, so if you happened to work on a Monday when the clinics were open and had more adults from the clinics you were more likely to prescribe antibiotics than if you worked over night and had fewer patients. Providers who worked primarily holidays or your Friday prescribing pattern may be very different if you could get them into a primary care provider the next day. [Stakeholder 22]

Clinicians also reported that historical practices in the clinical environment practices may also contribute to inappropriate prescribing.

I came from working in the [outpatient] Clinic as a new grad and they’re very strict about prescribing only according to evidence-based practice. And then when I came here things are with other colleagues are a little more loose with that type of thing. It can be difficult because you start to adopt that practice to. [Clinician 61, Nurse Practitioner, ED]

Theme 3: Clinician knowledge, awareness, and willingness to use evidence-based care

Stakeholders felt that clinicians had a lack of knowledge about prescribing of NSAIDs and antibiotics.

Sometimes errors are a lack of knowledge or awareness of the need to maybe specifically dose for let’s say impaired kidney function or awareness of current up to date current antibiotic resistance patterns in the location that might inform a more tailored antibiotic choice for a given condition. [Stakeholder 37] NSAIDs are very commonly used in the emergency department for patients of all ages…the ED clinician is simply not being aware that for specific populations this is not recommended and again just doing routine practice for patients of all ages and not realizing that for older patients you actually probably should not be using NSAIDs. [Stakeholder 40]

Some clinicians may be unwilling to change their prescribing practices due to outright resistance, entrenched habits, or lack of interest in doing so.

It sounds silly but there’s always some opposition to people being mandated to do something. But there are some people who would look and go ‘okay we already have a handle on that so why do we need something else? I know who prescribes inappropriately and who doesn’t. Is this a requirement, am I evaluated on it? That would come from supervisors. Is this one more thing on my annual review?’ [Stakeholder 28] If people have entrenched habits that are difficult to change and are physicians are very individualistic people who think that they are right more often than the non-physician because of their expensive training and perception of professionalism. [Stakeholder 4]

Theme 4: Uncertainty about whether an adverse event will occur

Clinicians cited the challenge of understanding the entirety of a Veteran’s condition, potential drug-drug interactions, and existing comorbidities in knowing whether an NSAID prescription may result in an adverse event.

It’s oftentimes a judgement call if someone has renal function that’s right at the precipice of being too poor to merit getting NSAIDs that may potentially cause issues. [Clinician 43, Physician, inpatient and urgent care] It depends on what the harm is. So, for instance, you can’t always predict allergic reactions. Harm from the non-steroidals would be more if you didn’t pre-identify risk factors for harm. So, they have ulcer disease, they have kidney problems where a non-steroidal would not be appropriate for that patient. Or potential for a drug-drug interaction between that non-steroid and another medication in particular. [Clinician 16, Physician, ED]

Rather than be concerned about the adverse events resulting from the medication itself, stakeholders identified the uncertainty that clinicians experience about whether a Veteran may experience an adverse event from an infection if nothing is done. This uncertainty contributes to the prescription of an antibiotic.

My experience in working with providers at the VA over the years is that they worry more about the consequences of not treating an infection than about the consequences of the antibiotic itself. [Stakeholder 19] Sometimes folks like to practice conservatively and they’ll say even though I didn’t really see any hard evidence of a bacterial infection, the patient’s older and sicker and they didn’t want to risk it. [Stakeholder 16]

Theme 5: Limited communication during and after the clinical encounter

The role and type of communication about prescribing depended upon the respondent. Clinicians identified inadequate communication and coordination with the Veteran’s primary care physician during the clinical encounter.

I would like to have a little more communication with the primary doctors. They don’t seem to be super interested in talking to anyone in the emergency room about their patients… A lot of times you don’t get an answer from the primary doctor or you get I’m busy in clinic. You can just pick something or just do what you think is right. [Clinician 25, Physician, ED]

Alternatively, stakeholders identified post-encounter patient outcome and clinical performance feedback as potential barriers.

Physicians tend to think that they are doing their best for every individual patient and without getting patient by patient feedback there is a strong cognitive bias to think well there must have been some exception and reason that I did it in this setting. [Stakeholder 34] It’s really more their own awareness of like their clinical performance and how they’re doing. [Stakeholder 40]

Veterans, however, prioritized communication during the clinical encounter. They expressed the need for clear and informative communication with the clinician, and the need for the clinician to provide a rationale for the choice and medication-specific details along with a need to ask any questions.

I expect him to tell me why I’m taking it, what it should do, and probably the side effects. [Patient 25, ED] I’d like to have a better description of how to take it because I won’t remember all the time and sometimes what they put on the bottle is not quite as clear. [Patient 22, ED]

Veterans reported their desire for a simple way to learn about medication information. They provided feedback on the current approaches to educational materials about prescriptions.

Probably most pamphlets that people get they’re not going to pay attention to them. Websites can be overwhelming. [Patient 3, ED] Posters can be offsetting. If you’re sick, you’re not going to read them…if you’re sick you may glance at that poster and disregard it. So, you’re not really going to see it but if you give them something in the hand people will tend to look at it because it’s in their hand. [Patient 19, ED] It would be nice if labels or something just told me what I needed to know. You know take this exactly when and reminds me here’s why you’re taking it for and just real clear and not small letters. [Patient 7, ED]

Theme 6: Technology barriers limited the usefulness of clinical decision support for order checking and patient communication tools

Following the decision to prescribe a medication, clinicians complained that electronic health record pop-ups with clinical decision support warnings for potential safety concerns (e.g., drug-drug interactions) were both excessive and not useful in a busy clinical environment.

The more the pop ups, the more they get ignored. So, it’s finding that sweet spot right where you’re not constantly having to click out of something because you’re so busy. Particularly in our clinical setting where we have very limited amount of time to read the little monograph. Most of the time you click ‘no’ and off you go. (Clinician 16, Physician, ED) Some of these mechanisms like the EMR [electronic medical record] or pop-up decision-making windows really limit your time. If you know the guidelines appropriately and doing the right thing, even if you’re doing the right thing it takes you a long time to get through something. (Clinician 19, Physician, Primary care clinic)

For post-encounter communication that builds on Theme 5 about patient communication, patients reported finding using the VA patient portal (MyHealtheVet) challenging for post-event communication with their primary care physician and to review the medications they were prescribed.

I’ve got to get help to get onto MyHealtheVet but I would probably like to try and use that, but I haven’t been on it in quite some time. [Patient 22, ED] I tried it [MyHealtheVet] once and it’s just too complicated so I’m not going to deal with it. [Patient 37, Urgent care]

This work examined attitudes and perceptions of barriers to appropriate prescribing of antibiotics and NSAIDs in unscheduled outpatient care settings in the Veterans Health Administration. Expanding on prior qualitative work on antimicrobial stewardship programs, we also included an examination of NSAID prescribing, a medication class which has received little attention focused on prescribing stewardship. This work seeks to advance the understanding of fundamental problems underlying prescribing stewardship to facilitate interventions designed to improve not only the decision to prescribe antibiotics and NSAIDs, but enhances the safety checks once a decision to prescribe is made. Specifically, we identified six themes during these interviews: perceived versus actual Veteran expectations about prescribing, the influence of a time-pressured clinical environment on prescribing stewardship, limited clinician knowledge, awareness, and willingness to use evidence-based care, uncertainties about the Veteran condition at the time of the clinical encounter, limited communication, and technology barriers.

Sensitive to patient expectations, clinicians believed that Veterans would be dissatisfied if they did not receive an antibiotic prescription, [ 34 ] even though most patients presenting to the ED for upper respiratory tract infections do not expect antibiotics. [ 35 ] However, recent work by Staub et al. found that among patients with respiratory tract infections, receipt of an antibiotic was not independently associated with improved satisfaction. [ 36 ] Instead, they found that receipt of antibiotics had to match the patient’s expectations to affect patient satisfaction and recommended that clinicians communicate with their patients about prescribing expectations. This finding complements our results in the present study and the importance of communication about expectations is similarly important for NSAID prescribing as well.

A commitment to stewardship and modification of clinician behavior may be compromised by the time-pressured clinical environment, numerous potential drug interactions, comorbidities of a vulnerable Veteran population, and normative practices. The decision to prescribe medications such as antibiotics is a complex clinical decision and may be influenced by both clinical and non-clinical factors. [ 34 , 37 , 38 ] ED crowding, which occurs when the demand for services exceeds a system’s ability to provide care, [ 39 ] is a well-recognized manifestation of a chaotic clinical environment and is associated with detrimental effects on the hospital system and patient outcomes. [ 40 , 41 ] The likelihood that congestion and wait times will improve is unlikely as the COVID-19 pandemic has exacerbated the already existing crowding and boarding crisis in EDs. [ 42 , 43 ]

Another theme was the uncertainty in the anticipation of adverse events that was exacerbated by the lack of a feedback loop. Feedback on clinical care processes and patient outcomes is uncommonly provided in emergency care settings, [ 44 ] yet may provide an opportunity to change clinician behavior, particularly for antimicrobial stewardship. [ 45 ] However, the frequent use of ineffective feedback strategies [ 46 ] compromises the ability to implement effective feedback interventions; feedback must be specific [ 47 ] and address the Intention-to-Action gap [ 48 ] by including co-interventions to address recipient characteristics (i.e., beliefs and capabilities) and context to maximize impact. Without these, feedback may be ineffective.

An additional barrier identified from this work is the limited communication with primary care following discharge. A 2017 National Quality Forum report on ED care transitions [ 49 ] recommended that EDs and their supporting hospital systems should expand infrastructure and enhance health information technology to support care transitions as Veterans may not understand discharge instructions, may not receive post-ED or urgent care, [ 50 , 51 , 52 ] or may not receive a newly prescribed medication. [ 24 ] While there are existing mechanisms to communicate between the ED and primary care teams such as notifications when a Veteran presents to the ED and when an emergency clinician copies a primary care physician on a note, these mechanisms are insufficient to address care transition gaps and are variable in best practice use. To address this variability, the VA ED PACT Tool was developed using best practices (standardized processes, "closed-loop" communication, embedding into workflow) to facilitate and standardize communication between VA EDs and follow-up care clinicians. [ 53 ] While the ED PACT Tool is implemented at the Greater Los Angeles VA and can create a care coordination order upon ED discharge, its use is not yet widely adopted throughout the VA.

In the final theme about technology barriers, once the decision has been made to prescribe a medication, existing electronic tools that are key components of existing stewardship interventions designed to curtail potentially inappropriate prescriptions may be compromised by their lack of usability. For example, clinician and stakeholder interview respondents described how usability concerns were exacerbated in a time-pressured clinical environment (e.g., electronic health record clinical decision support tools). Clinical decision support is an effective tool to improve healthcare process measures in a diverse group of clinical environments; [ 54 ] however, usability remains a barrier when alerts must be frequently overridden. [ 55 , 56 ] Alert fatigue, as expressed in our interviews for order checking and recognized within the VA’s EHR, [ 57 , 58 ] may contribute to excessive overrides reducing the benefit of clinical decision support, [ 56 , 59 ] there was a notable lack of discussion about the decision to initiate appropriate prescriptions, which is a key action of the CDC’s outpatient antibiotic stewardship campaign. [ 18 ] Thus, a potentially more effective, albeit challenging approach, is to “nudge” clinicians towards appropriate prescribing and away from the initial decision to prescribe (e.g., inappropriate antibiotic prescribing for viral upper respiratory tract infections) with either default order sets for symptom management or to enhance prescription decisions through reminders about potential contraindications to specific indications (e.g., high risk comorbidities). Beyond EHR-based solutions that might change clinician behavior, the CDC’s outpatient antibiotic stewardship program provides a framework to change the normative practices around inappropriate prescribing and includes a commitment to appropriate prescribing, action for policy and change, tracking and reporting, and education and expertise. [ 18 ]

Another technical barrier faces patients through patient-facing electronic tools such as the VA’s MyHealtheVet portal, which was developed to enhance patient communication following care transitions and to allow Veterans to review their medications and to communicate with their primary care clinical team. Patient portals can be an effective tool for medication adherence [ 60 ] and offer promise to provide patient education [ 61 ] following a clinical encounter. However, they are similarly limited by usability concerns, representing an adoption barrier to broader Veteran use after unscheduled outpatient care visits [ 62 ], particularly in an older patient population.

These interviews further underscored that lack of usability of clinical decision support for order checking that arises from ineffective design and is a key barrier preventing health information technology from reaching its promise of improving patient safety. [ 63 ] A common and recognized reason for these design challenges include the failure to place the user (i.e., acute care clinician) at the center of the design process resulting in underutilization, workarounds, [ 64 ] and unintended consequences, [ 65 ] all of which diminish patient safety practices and fail to change clinician behavior (i.e., prescribing). Complex adaptive systems work best when the relative strengths of humans (e.g., context sensitivity, situation specificity) are properly integrated with the information processing power of computerized systems. [ 66 ] One potential approach to address usability concerns is through the integration of user-centered design into technology design represents an opportunity to design more clinician- and patient-centric systems of care to advance prescribing stewardship interventions that may have lacked broader adoption previously. As antimicrobial stewardship and additional prescribing stewardship efforts focus on time-pressured environments where usability is essential to adoption, taking a user-centered design approach to not only the development of electronic tools but also in addressing the identified barriers in prescribing represents a promising approach to enhance the quality of prescribing.

Limitations

The study findings should be considered in light of its limitations. First, the setting for this work was the Veterans Health Administration, the largest integrated health system in the US. Also, while we focused on the stewardship of two drug classes, there are numerous additional drug classes that are prescribed in these settings. Studies in other settings or on other drug classes may not generalize to other settings and drug classes. Second, while clinicians and stakeholder perspectives included diverse, national representation, the Veterans interviewed were local to the Tennessee Valley Healthcare System. Given the concurrent COVID-19 pandemic at the time of enrollment, most of the Veterans were seen for pain-related complaints, and only two infectious-related complaints were included. However, we also asked them about antibiotic prescribing. Clinician and stakeholder narratives may not completely reflect their practice patterns as their responses could be influenced by social desirability bias. Third, responses may be subject to recall bias and may influence the data collected. Finally, the themes and subthemes identified may overlap and have potential interactions. While we used an iterative process to identify discrete themes and subthemes, prescription decisions represent a complex decision process that are influenced by numerous patient and contextual factors and may not be completely independent.

Despite numerous interventions to improve the quality of prescribing, the appropriate prescription of antibiotics and NSAIDs in unscheduled outpatient care settings remains a challenge. Using the Veterans Health Administration, this study found that challenges to high quality prescribing include perceived Veteran expectations about receipt of medications, a hectic clinical environment deprioritizing stewardship, limited clinician knowledge, awareness, and willingness to use evidence-based care, uncertainty about the potential for adverse events, limited communication, and technology barriers. Findings from these interviews suggest that interventions should consider the detrimental impact of high workload on prescribing stewardship, clinician workflow, the initial decision to prescribe medications, and incorporate end-users into the intervention design process. Doing so is a promising approach to enhance adoption of high quality prescribing practices in order to improve the quality and patient outcomes from NSAID and antibiotic prescribing.

Availability of data and materials

De-identified datasets used and/or analysed during the current study will be made available from the corresponding author on reasonable request.

Leape LL, Brennan TA, Laird N, et al. The nature of adverse events in hospitalized patients. Results of the Harvard Medical Practice Study II. N Engl J Med. 1991;324(6):377–384.

Article CAS PubMed Google Scholar

Pitts SR, Carrier ER, Rich EC, Kellermann AL. Where Americans get acute care: increasingly, it’s not at their doctor’s office. Health Aff (Millwood). 2010;29(9):1620–9.

Article PubMed Google Scholar

Palms DL, Hicks LA, Bartoces M, et al. Comparison of antibiotic prescribing in retail clinics, urgent care centers, emergency departments, and traditional ambulatory care settings in the United States. Jama Intern Med. 2018;178(9):1267–9.

Article PubMed PubMed Central Google Scholar

Davis JS, Lee HY, Kim J, et al. Use of non-steroidal anti-inflammatory drugs in US adults: changes over time and by demographic. Open Heart. 2017;4(1):e000550.

Fleming-Dutra KE, Hersh AL, Shapiro DJ, et al. Prevalence of inappropriate antibiotic prescriptions among US ambulatory care visits, 2010–2011. JAMA. 2016;315(17):1864–73.

Shively NR, Buehrle DJ, Clancy CJ, Decker BK. Prevalence of Inappropriate Antibiotic Prescribing in Primary Care Clinics within a Veterans Affairs Health Care System. Antimicrob Agents Chemother. 2018;62(8):e00337–18. https://doi.org/10.1128/AAC.00337-18 . https://pubmed.ncbi.nlm.nih.gov/29967028/ .

World Health Organization. Global antimicrobial resistance and use surveillance system (GLASS) report: 2022. 2022.

Centers for Disease Control and Prevention. COVID-19: U.S. Impact on Antimicrobial Resistance, Special Report 2022. Atlanta: U.S. Department of Health and Human Services, CDC; 2022.

Google Scholar

Shehab N, Lovegrove MC, Geller AI, Rose KO, Weidle NJ, Budnitz DS. US emergency department visits for outpatient adverse drug events, 2013–2014. JAMA. 2016;316(20):2115–25.

Fassio V, Aspinall SL, Zhao X, et al. Trends in opioid and nonsteroidal anti-inflammatory use and adverse events. Am J Manag Care. 2018;24(3):e61–72.

PubMed Google Scholar

Centers for Disease Control and Prevention. Chronic Kidney Disease Surveillance System—United States. http://www.cdc.gov/ckd . Accessed 21 March 2023.

Cahir C, Fahey T, Teeling M, Teljeur C, Feely J, Bennett K. Potentially inappropriate prescribing and cost outcomes for older people: a national population study. Br J Clin Pharmacol. 2010;69(5):543–52.

Gabriel SE, Jaakkimainen L, Bombardier C. Risk for Serious Gastrointestinal Complications Related to Use of Nonsteroidal Antiinflammatory Drugs - a Metaanalysis. Ann Intern Med. 1991;115(10):787–96.

Zhang X, Donnan PT, Bell S, Guthrie B. Non-steroidal anti-inflammatory drug induced acute kidney injury in the community dwelling general population and people with chronic kidney disease: systematic review and meta-analysis. BMC Nephrol. 2017;18(1):256.

McGettigan P, Henry D. Cardiovascular risk with non-steroidal anti-inflammatory drugs: systematic review of population-based controlled observational studies. PLoS Med. 2011;8(9): e1001098.

Article CAS PubMed PubMed Central Google Scholar

Holt A, Strange JE, Nouhravesh N, et al. Heart Failure Following Anti-Inflammatory Medications in Patients With Type 2 Diabetes Mellitus. J Am Coll Cardiol. 2023;81(15):1459–70.

Davey P, Marwick CA, Scott CL, et al. Interventions to improve antibiotic prescribing practices for hospital inpatients. Cochrane Database Syst Rev. 2017;2(2):CD003543.

Sanchez GV, Fleming-Dutra KE, Roberts RM, Hicks LA. Core Elements of Outpatient Antibiotic Stewardship. MMWR Recomm Rep. 2016;65(6):1–12.

May L, Martin Quiros A, Ten Oever J, Hoogerwerf J, Schoffelen T, Schouten J. Antimicrobial stewardship in the emergency department: characteristics and evidence for effectiveness of interventions. Clin Microbiol Infect. 2021;27(2):204–9.

May L, Cosgrove S, L'Archeveque M, et al. A call to action for antimicrobial stewardship in the emergency department: approaches and strategies. Ann Emerg Med. 2013;62(1):69–77 e62.

Veterans Health Administration Emergency Medicine Management Tool. EDIS GeriatricsAgeReport v3.

Cairns C KK, Santo L. National Hospital Ambulatory Medical Care Survey: 2020 emergency department summary tables. NHAMCS Factsheets - EDs Web site. https://www.cdc.gov/nchs/data/nhamcs/web_tables/2020-nhamcs-ed-web-tables-508.pdf . Accessed 20 Dec 2022.

Lowery JL, Alexander B, Nair R, Heintz BH, Livorsi DJ. Evaluation of antibiotic prescribing in emergency departments and urgent care centers across the Veterans’ Health Administration. Infect Control Hosp Epidemiol. 2021;42(6):694–701.

Hastings SN, Sloane RJ, Goldberg KC, Oddone EZ, Schmader KE. The quality of pharmacotherapy in older veterans discharged from the emergency department or urgent care clinic. J Am Geriatr Soc. 2007;55(9):1339–48.

Goodman LA. Snowball sampling. The annals of mathematical statistics. 1961. pp. 148–170.

Damschroder LJ, Aron DC, Keith RE, Kirsh SR, Alexander JA, Lowery JC. Fostering implementation of health services research findings into practice: a consolidated framework for advancing implementation science. Implement Sci. 2009;4:50.

Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50(2):179–211.

Article Google Scholar

Ajzen I. The theory of planned behaviour: reactions and reflections. Psychol Health. 2011;26(9):1113–27. https://doi.org/10.1080/08870446.2011.613995 . https://www.tandfonline.com/doi/full/10.1080/08870446.2011.613995 .

Morse JM. The significance of saturation. Qual Health Res. 1995;5(2):147–9.

Azungah T. Qualitative research: deductive and inductive approaches to data analysis. Qual Res J. 2018;18(4):383–400.

Tjora A. Qualitative research as stepwise-deductive induction. Routledge; 2018. https://www.routledge.com/Qualitative-Research-as-Stepwise-Deductive-Induction/Tjora/p/book/9781138304499 .

Fereday J, Muir-Cochrane E. Demonstrating rigor using thematic analysis: A hybrid approach of inductive and deductive coding and theme development. Int J Qual Methods. 2006;5(1):80–92.

Tong A, Sainsbury P, Craig J. Consolidated criteria for reporting qualitative research (COREQ): a 32-item checklist for interviews and focus groups. Int J Qual Health Care. 2007;19(6):349–57.

Patel A, Pfoh ER, Misra Hebert AD, et al. Attitudes of High Versus Low Antibiotic Prescribers in the Management of Upper Respiratory Tract Infections: a Mixed Methods Study. J Gen Intern Med. 2020;35(4):1182–8.

May L, Gudger G, Armstrong P, et al. Multisite exploration of clinical decision making for antibiotic use by emergency medicine providers using quantitative and qualitative methods. Infect Control Hosp Epidemiol. 2014;35(9):1114–25.

Staub MB, Pellegrino R, Gettler E, et al. Association of antibiotics with veteran visit satisfaction and antibiotic expectations for upper respiratory tract infections. Antimicrob Steward Healthc Epidemiol. 2022;2(1): e100.

Schroeck JL, Ruh CA, Sellick JA Jr, Ott MC, Mattappallil A, Mergenhagen KA. Factors associated with antibiotic misuse in outpatient treatment for upper respiratory tract infections. Antimicrob Agents Chemother. 2015;59(7):3848–52.

Hruza HR, Velasquez T, Madaras-Kelly KJ, Fleming-Dutra KE, Samore MH, Butler JM. Evaluation of clinicians’ knowledge, attitudes, and planned behaviors related to an intervention to improve acute respiratory infection management. Infect Control Hosp Epidemiol. 2020;41(6):672–9.

American College of Emergency Physicians Policy Statement. Crowding. https://www.acep.org/globalassets/new-pdfs/policy-statements/crowding.pdf . Published 2019. Accessed 11 Oct 2023.

Bernstein SL, Aronsky D, Duseja R, et al. The effect of emergency department crowding on clinically oriented outcomes. Acad Emerg Med. 2009;16(1):1–10.

Rasouli HR, Esfahani AA, Nobakht M, et al. Outcomes of crowding in emergency departments; a systematic review. Arch Acad Emerg Med. 2019;7(1):e52.

PubMed PubMed Central Google Scholar

Janke AT, Melnick ER, Venkatesh AK. Monthly Rates of Patients Who Left Before Accessing Care in US Emergency Departments, 2017–2021. JAMA Netw Open. 2022;5(9): e2233708.

Janke AT, Melnick ER, Venkatesh AK. Hospital Occupancy and Emergency Department Boarding During the COVID-19 Pandemic. JAMA Netw Open. 2022;5(9): e2233964.

Lavoie CF, Plint AC, Clifford TJ, Gaboury I. “I never hear what happens, even if they die”: a survey of emergency physicians about outcome feedback. CJEM. 2009;11(6):523–8.

Ivers N, Jamtvedt G, Flottorp S, et al. Audit and feedback: effects on professional practice and healthcare outcomes. Cochrane Database Syst Rev. 2012;(6):CD000259. https://doi.org/10.1002/14651858.CD000259.pub3 .

Hysong SJ, SoRelle R, Hughes AM. Prevalence of Effective Audit-and-Feedback Practices in Primary Care Settings: A Qualitative Examination Within Veterans Health Administration. Hum Factors. 2022;64(1):99–108.

Presseau J, McCleary N, Lorencatto F, Patey AM, Grimshaw JM, Francis JJ. Action, actor, context, target, time (AACTT): a framework for specifying behaviour. Implement Sci. 2019;14(1):102.

Desveaux L, Ivers NM, Devotta K, Ramji N, Weyman K, Kiran T. Unpacking the intention to action gap: a qualitative study understanding how physicians engage with audit and feedback. Implement Sci. 2021;16(1):19.

National Quality Forum. Emergency Department Transitions of Care: A Quality Measurement Framework—Final Report: DHHS contract HHSM‐500–2012–000091, Task Order HHSM‐500‐T0025. Washington, DC: National Quality Forum; 2017.

Kyriacou DN, Handel D, Stein AC, Nelson RR. Brief report: factors affecting outpatient follow-up compliance of emergency department patients. J Gen Intern Med. 2005;20(10):938–42.

Vukmir RB, Kremen R, Ellis GL, DeHart DA, Plewa MC, Menegazzi J. Compliance with emergency department referral: the effect of computerized discharge instructions. Ann Emerg Med. 1993;22(5):819–23.

Engel KG, Heisler M, Smith DM, Robinson CH, Forman JH, Ubel PA. Patient comprehension of emergency department care and instructions: are patients aware of when they do not understand? Ann Emerg Med. 2009;53(4):454–461 e415.

Cordasco KM, Saifu HN, Song HS, et al. The ED-PACT Tool Initiative: Communicating Veterans’ Care Needs After Emergency Department Visits. J Healthc Qual. 2020;42(3):157–65.

Bright TJ, Wong A, Dhurjati R, et al. Effect of clinical decision-support systems: a systematic review. Ann Intern Med. 2012;157(1):29–43.

Weingart SN, Toth M, Sands DZ, Aronson MD, Davis RB, Phillips RS. Physicians’ decisions to override computerized drug alerts in primary care. Arch Intern Med. 2003;163(21):2625–31.

van der Sijs H, Aarts J, Vulto A, Berg M. Overriding of drug safety alerts in computerized physician order entry. J Am Med Inform Assoc. 2006;13(2):138–47.

Shah T, Patel-Teague S, Kroupa L, Meyer AND, Singh H. Impact of a national QI programme on reducing electronic health record notifications to clinicians. BMJ Qual Saf. 2019;28(1):10–4.

Lin CP, Payne TH, Nichol WP, Hoey PJ, Anderson CL, Gennari JH. Evaluating clinical decision support systems: monitoring CPOE order check override rates in the Department of Veterans Affairs’ Computerized Patient Record System. J Am Med Inform Assoc. 2008;15(5):620–6.

Middleton B, Bloomrosen M, Dente MA, et al. Enhancing patient safety and quality of care by improving the usability of electronic health record systems: recommendations from AMIA. J Am Med Inform Assoc. 2013;20(e1):e2-8.

Han HR, Gleason KT, Sun CA, et al. Using Patient Portals to Improve Patient Outcomes: Systematic Review. JMIR Hum Factors. 2019;6(4): e15038.

Johnson AM, Brimhall AS, Johnson ET, et al. A systematic review of the effectiveness of patient education through patient portals. JAMIA Open. 2023;6(1):ooac085.

Lazard AJ, Watkins I, Mackert MS, Xie B, Stephens KK, Shalev H. Design simplicity influences patient portal use: the role of aesthetic evaluations for technology acceptance. J Am Med Inform Assoc. 2016;23(e1):e157-161.

IOM. Health IT and Patient Safety: Building Safer Systems for Better Care. Washington, DC: NAP;2012.

Koppel R, Wetterneck T, Telles JL, Karsh BT. Workarounds to barcode medication administration systems: their occurrences, causes, and threats to patient safety. J Am Med Inform Assoc. 2008;15(4):408–23.

Ash JS, Sittig DF, Poon EG, Guappone K, Campbell E, Dykstra RH. The extent and importance of unintended consequences related to computerized provider order entry. J Am Med Inform Assoc. 2007;14(4):415–23.

Hollnagel E, Woods D. Joint Cognitive Systems: Foundations of Cognitive Systems Engineering. Boca Raton: CRC Press; 2006.

Download references

Acknowledgements

This material is based upon work supported by the Department of Veterans Affairs, Veterans Health Administration, Office of Research and Development, Health Services Research and Development (I01HX003057). The content is solely the responsibility of the authors and does not necessarily represent the official views of the VA.

Author information

Authors and affiliations.

Education, and Clinical Center (GRECC), VA , Geriatric Research, Tennessee Valley Healthcare System, 2525 West End Avenue, Ste. 1430, Nashville, TN, 37203, USA

Michael J. Ward, Michael E. Matheny & Amanda S. Mixon

Medicine Service, Tennessee Valley Healthcare System, Nashville, TN, USA

Michael J. Ward

Department of Emergency Medicine, Vanderbilt University Medical Center, Nashville, TN, USA

Michael J. Ward & Melissa D. Rubenstein

Department of Biomedical Informatics, Vanderbilt University Medical Center, Nashville, TN, USA

Michael J. Ward, Michael E. Matheny, Shilo Anders & Thomas Reese

Department of Biostatistics, Vanderbilt University Medical Center, Nashville, TN, USA

Michael E. Matheny

Division of General Internal Medicine & Public Health, Vanderbilt University Medical Center, Nashville, TN, USA

Department of Psychology, Vanderbilt University, Nashville, TN, USA

Kemberlee Bonnet, Chloe Dagostino & David G. Schlundt

Center for Research and Innovation in Systems Safety, Vanderbilt University Medical Center, Nashville, TN, USA

Shilo Anders

Section of Hospital Medicine, Vanderbilt University Medical Center, Nashville, TN, USA

Amanda S. Mixon

You can also search for this author in PubMed Google Scholar

Contributions

Conceptualization: MJW, ASM, MEM, DS, SA. Methodology: MJW, ASM, MEM, DS, KB, SA, TR. Formal analysis: KB, DS, CD, MJW. Investigation: MJW, MDR, DS. Resources: MJW, MEM. Writing—Original Draft. Preparation: MJW, ASM, KB, MDR. Writing—Review & Editing: All investigators. Supervision: MJW, ASM, MEM. Funding acquisition: MJW, MEM.

Corresponding author

Correspondence to Michael J. Ward .

Ethics declarations

Ethics approval and consent to participate.

This study was approved by the VA Tennessee Valley Healthcare System Institutional Review Board as minimal risk (#1573619). A waiver of informed consent was approved and each subject was verbally consented prior to interviews. The IRB determined that all requirements set forth in 38CFR16.111 in accordance for human subjects research have been satisfied. All the methods were carried out according the relevant guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note.

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Supplementary material 1., rights and permissions.

Reprints and permissions

About this article

Cite this article.

Ward, M.J., Matheny, M.E., Rubenstein, M.D. et al. Determinants of appropriate antibiotic and NSAID prescribing in unscheduled outpatient settings in the veterans health administration. BMC Health Serv Res 24 , 640 (2024). https://doi.org/10.1186/s12913-024-11082-0

Download citation

Received : 11 October 2023

Accepted : 07 May 2024

Published : 18 May 2024

DOI : https://doi.org/10.1186/s12913-024-11082-0

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

Non-Steroidal Anti-Inflammatory Drugs
Antibiotics
Qualitative Methods
Emergency Department
Urgent Care
Primary Care
Prescribing Stewardship

BMC Health Services Research

ISSN: 1472-6963

General enquiries: [email protected]

COMMENTS

Qualitative Data Coding 101 (With Examples)
Step 1 - Initial coding. The first step of the coding process is to identify the essence of the text and code it accordingly. While there are various qualitative analysis software packages available, you can just as easily code textual data using Microsoft Word's "comments" feature.
Coding Qualitative Data: A Beginner's How-To + Examples
Below we'll walk through an example of coding qualitative data, utilising the steps and tips detailed above. Step 1: Read through your data and define your categories. For this example, we'll use "customer service," "product quality," and "price.". Step 2: Sort a sample of the data into the above categories.
How To Do Initial Coding
Initial coding, also known as "open coding," is the first step of the coding process, particularly in qualitative methods such as grounded theory. Read our guide for how to do initial coding. ... and you can code specific excerpts of data in various ways. Here are some examples: Line by line coding: Code every line of your data. Or try ...
Chapter 18. Data Analysis and Coding
The following three passages are examples of how qualitative researchers describe their data-analysis practices. The first, by Harvey, is a useful example of how data analysis can shift the original research questions. The second example, by Thai, shows multiple stages of coding and how these stages build upward to conceptual themes and ...
PDF 1 An Introduction to Codes and Coding
Chapter Summary. This chapter first presents the purposes and goals of The Coding Manual for Qualitative Researchers. It then provides definitions and examples of codes and categories and their roles in qualitative data analysis. The procedures and mechanics of coding follow, along with discussions of analytic software and team collaboration.
The Living Codebook: Documenting the Process of Qualitative Data
For example, Weston et al. (2001) discuss their process in developing a coding system and codebook in their qualitative interview research team. They outline how Weston developed the initial coding scheme, the graduate student team members coded the materials, and that the coding was an iterative process, where Weston would draft and revise the ...
10.6 Qualitative Coding, Analysis, and Write-up: The How to Guide
Step 2: Axial (focused) coding. In open coding, the researcher is focused primarily on the text from the interviews to define concepts and categories. In axial coding, the researcher is using the concepts and categories developed in the open coding process, while re-reading the text from the interviews.
PDF Coding in Qualitative Research
Coding by Qualitative Design. Several types of qualitative approaches/designs. Ethnography, Phenomenology, Grounded Theory, Narrative Research, Case Study, Qualitative Descriptive. We will go through an example "chunk" of data, and code it together as if it were a Grounded Theory study.
Coding
Coding is a qualitative data analysis strategy in which some aspect of the data is assigned a descriptive label that allows the researcher to identify related content across the data. How you decide to code - or whether to code- your data should be driven by your methodology. But there are rarely step-by-step descriptions, and you'll have to ...
PDF Tips & Tools #18: Coding Qualitative Data
Coding Qualitative Data . ... Figure 1 is an example of the initial coding and marginal remarks on a hard copy of interview transcripts. Typically, in an interview transcript like the one shown (Figure 1), you will read the ... Introduction to Qualitative Research Methods. Wiley and Sons: Hoboken, NJ. Author: tdsatter Created Date: 8/5/2014 1 ...
Essential Guide to Coding Qualitative Data
See how to do in vivo coding and an in vivo coding qualitative data example. Process Coding. Capturing an action. With Process coding, you use codes to communicate an action in the qualitative data. Codes will typically be gerunds that end with "ing". See how to do process coding. Open Coding. An initial round of loose and tentative coding
A Guide to Coding Qualitative Research Data
The primary goal of coding qualitative data is to change data into a consistent format in support of research and reporting. A code can be a phrase or a word that depicts an idea or recurring theme in the data. The code's label must be intuitive and encapsulate the essence of the researcher's observations or participants' responses.
PDF Analyzing Qualitative Data 4 Thematic coding and categorizing
• see the central role of coding in qualitative analysis; • see from the close examination of an example the importance of creating codes that are analytic and theoretical and not merely descriptive; and • know two techniques that can be used to promote the move from description to analysis: constant comparison and line-by-line coding ...
PDF Coding Qualitative Data
12.2 When and How Coding is Used. Simply put, coding is qualitative analysis. Coding is the analytical phase where researchers become immersed in their data, take the time to fully get to know it (Basit, 2003; Elliott, 2018), and allow its sense to be discerned. A code is "...a word or short phrase that symbolically assigns a summative ...
Qualitative Data Coding 101 (With Examples)
Deductive coding 101. With deductive engraving, we make use of pre-established codes, which are made for your interact with the present data. These usually involves drawing up a fixed of codes based for a research question or previous research.You could additionally use a encipher setting from the codebook of a previous study.
Qualitative Coding Tutorial: How To Code Qualitative Data For ...
Learn how to code qualitative data the right way. We explain the qualitative coding process in simple, easy to understand terms. Learn about the different co...
Qualitative Data Coding
Qualitative Data Coding. Coding is the process of analyzing qualitative data (usually text) by assigning labels (codes) to chunks of data that capture their essence or meaning. It allows you to condense, organize and interpret your data. A code is a word or brief phrase that captures the essence of why you think a particular bit of data may be ...
(PDF) Qualitative Data Coding
2. WORKSHOP. Qualitative Data Coding. ABSTRACT. In the quest to address a research problem, meeting the purpose of the study, and answer ing. qualitative research question (s), we actively look ...
Coding qualitative data: a synthesis guiding the novice
that can help pave the way to the researcher's interpretive judgements and improve the ir quality. By using this paper, novice researchers will be able to reflect more carefully on the ...
Coding Qualitative Data: How To Guide
Create new codes based on the second sample. Go back and recode all responses again. Repeat from step 5 until you've coded all of your data. If you add a new code, split an existing code into two, or change the description of a code, make sure to review how this change will affect the coding of all responses.
Thematic Analysis & Coding: An Overview of the Qualitative Paradigm
Hence effective coding requires at least an initial coding and . ... Coding in qualitative research is a systematic process ... For example, with the research question the overall goal is ...
Qualitative Coding & Analysis
Open coding. At this first level of coding, you are looking for distinct concepts and categories in the data, which will form the basic units of your analysis. In other words, you are breaking down the data into first level concepts, or master headings, and second-level categories, or subheadings. Researchers often use highlights to distinguish ...
How to Do Thematic Analysis
When to use thematic analysis. Thematic analysis is a good approach to research where you're trying to find out something about people's views, opinions, knowledge, experiences or values from a set of qualitative data - for example, interview transcripts, social media profiles, or survey responses. Some types of research questions you might use thematic analysis to answer:
How to do pattern coding in qualitative analysis
Qualitative coding tools like Delve can simplify the pattern coding process, by allowing you to seamlessly group codes together: Pattern coding is an appropriate method to facilitate this process and pull together the story of your data with a narrative that centers around an overarching category.
SWAP Qualitative Case Study Research: Annexes
The samples for Areas 3 and 4, was additionally stratified by age (18 to 24 years vs. 25+ years) to account for the small number of potential participants aged 18 to 24 years provided in the ...
Qualitative Insights into Organizational Value Creation: Decoding
Since the initial open coding was executed independently, ongoing researcher discussions shaped the final constructs. After identifying platform characteristics, we assessed their relative importance. ... Merriam, S. B., & Grenier, R. S. (2019). Qualitative research in practice: Examples for Discussion and Analysis. John Wiley & Sons. Metcalf ...
Determinants of appropriate antibiotic and NSAID prescribing in
In a sample of older Veterans discharged from VA ED and UCC settings, NSAIDs were found to be implicated in 77% of drug interactions. ... Qualitative data coding and analysis was managed by the Vanderbilt University Qualitative Research Core. ... Azungah T. Qualitative research: deductive and inductive approaches to data analysis. Qual Res J ...

Qualitative Data Coding 101

Overview: Qualitative Data Coding

What is qualitative data coding?

What are the different types of coding?

Deductive coding 101

Inductive coding 101

A little bit of both… hybrid coding approaches

Need a helping hand?

How to code qualitative data

Step 1 – Initial coding

How to decide which codes to use

Process coding

Descriptive coding

In vivo coding

Values coding

Step 2 – Line by line coding

Moving from coding to analysis

Code categorisation

Theme identification

Tips & tricks for quality coding

Psst... there’s more!

You Might Also Like:

31 Comments

Submit a Comment Cancel reply

Coding Qualitative Data: A Beginner’s How-To + Examples

What is Qualitative Data Coding?

Inductive Coding vs Deductive Coding

Inductive Coding

Deductive Coding

3 Steps for Coding Qualitative Data From the Top-Down

1. Start with Broad Categories

2. Assign Emotions or Sentiments

3. Combine Categories and Sentiments to Draw Conclusions

4 Tips to Keep in Mind for Accurate Qualitative Data Coding

1. Start with a Small Sample of the Data

2. Use Numerical Scales for Deeper Analysis

3. Remember That Each Data Point Can Contain Multiple Pieces of Information

4. Be Mindful of Having Too Many Tags

Coding Qualitative Data Example

How AI tools help with Coding Qualitative Data

Try Chattermill today today to learn how our AI-powered software can help you make smarter business decisions.

5 learnings of (nearly) a decade working with marketplaces

Zepz, the fintech behind global remittance brands WorldRemit and Sendwave, chooses Chattermill to align their team around customer insights

What to do if you are number 2 in the market

See Chattermill in action

Chapter 18. Data Analysis and Coding

Qualitative Data Analysis as a Long Process!

The Coding Stage

Writing Analytic Memos during/after Coding

Qualitative Data-Analysis Samples

Example 1: “Look Someone in the Eye” by Peter Francis Harvey ( 2022 )

Example 2:” Policing and Symbolic Control” by Mai Thai ( 2022 )

Example 3: The Dignity of Working Men by Michèle Lamont ( 2000 )

Sample Codebook 1

Sample Codebook 2

Further Readings

10.6 Qualitative Coding, Analysis, and Write-up: The How to Guide

Step1: Open coding

Step 2: Axial (focused) coding

Step 3: Build a data table

Step 4: Analysis & write-up

Share This Book

University Library, University of Illinois at Urbana-Champaign

Qualitative Data Analysis: Coding

Coding Qualitative Data

What will you code?

Where will your codes come from?

How will you apply your codes to your data?

What will be coded?

How will you explain the coding process?

How will you report the results of your coding process?

How will you code collaboratively?

Coding Features in QDA Software Programs

Useful Resources on Coding

A guide to coding qualitative research data

Streamline data coding

Inductive coding

Deductive coding

Inductive and deductive approaches

1. Conduct your first-round pass at coding qualitative data