long essay question length

Find what you need to study

AP World Long Essay Question (LEQ) Overview

15 min read • may 10, 2022

Zaina Siddiqi

Exam simulation mode

Prep for the AP exam with questions that mimic the test!

Overview of the Long Essay Question (LEQ)

Section II of the AP Exam includes three Long Essay Question (LEQ) prompts. You will choose to write about just one of these. 

The formatting of prompts varies somewhat between the AP Histories, though the rubric does not. In AP World History, the prompt includes a sentence that orients the writer to the time, place, and theme of the prompt topic, while prompts in AP US History and AP European History typically do not. However, the rubrics and scoring guidelines are the same for all Histories.

Your answer should include the following:

A valid thesis

A discussion of relevant historical context

Use of evidence supports your thesis

Use of a reasoning skill to organize and structure the argument

Complex understanding of the topic of the prompt

We will break down each of these aspects in the next section. For now, the gist is that you need to write an essay that answers the prompt, using evidence. You will need to structure and develop your essay using one of the course reasoning skills.

Many of the skills you need to write a successful LEQ essay are the same skills you will use on the DBQ. In fact, some of the rubric points are identical, so you can use a lot of the same strategies on both writing tasks!

You will have three choices of prompts for your LEQ. All three prompts will focus on the same reasoning skills, but the time periods will differ in each prompt. Prompt topics may span across time periods specified in the course outline, and the time period breakdowns for each prompt are as follows:

Writing time on the AP Exam includes both the Document Based Question (DBQ) and the (LEQ), but it is suggested that you spend 40 minutes completing the LEQ. You will need to plan and write your essay in that time.

A good breakdown would be 5 min. (planning) + 35 min. (writing) = 40 min.

The LEQ is scored on a rubric out of six points, and is weighted at 15% of your overall exam score. We’ll break down the rubric next.

How to Rock the LEQ: The Rubric

The LEQ is scored on a six point rubric, and each point can be earned independently. That means you can miss a point on something and still earn other points with the great parts of your essay.

Note: all of the examples in this section will be for this prompt from AP World History: Modern. You could use similar language, structure, or skills to write samples for prompts in AP US History or AP European History.

Let’s break down each rubric component...

What is it?

The thesis is a brief statement that introduces your argument or claim, and can be supported with evidence and analysis. This is where you answer the prompt.

Where do I write it?

This is the only element in the essay that has a required location. The thesis needs to be in your introduction or conclusion of your essay. It can be more than one sentence, but all of the sentences that make up your thesis must be consecutive in order to count.

How do I know if mine is good?

The most important part of your thesis is the claim , which is your answer to the prompt. The description the College-Board gives is that it should be “historically defensible,” which really means that your evidence must be plausible. On the LEQ, your thesis needs to be related topic of the prompt.

Your thesis should also establish your line of reasoning. Translation: address why or how something happened - think of this as the “because” to the implied “how/why” of the prompt. This sets up the framework for the body of your essay, since you can use the reasoning from your thesis to structure your body paragraph topics later.

The claim and reasoning are the required elements of the thesis. And if that’s all you can do, you’re in good shape to earn the point. 

Going above-and-beyond to create a more complex thesis can help you in the long run, so it’s worth your time to try. One way to build in complexity to your thesis is to think about a counter-claim or alternate viewpoint that is relevant to your response. If you are using one of the course reasoning process to structure your essay (and you should!) think about using that framework for your thesis too.

In a causation essay, a complex argument addresses causes and effects.

In a comparison essay, a complex argument addresses similarities and differences.

In a continuity and change over time essay, a complex argument addresses change and continuity .

This counter-claim or alternate viewpoint can look like an “although” or “however” phrase in your thesis.

Powers in both land-based and maritime empires had to adapt their rule to accommodate diverse populations. However, in this era land-based empires were more focused on direct political control, while the maritime empires were more based on trade and economic development.

This thesis works because it clearly addresses the prompt (comparing land and maritime empires). It starts by addressing a similarity, and then specifies a clear difference with a line of reasoning to clarify the actions of the land vs. maritime empires.

Contextualization

Contextualization is a brief statement that lays out the broader historical background relevant to the prompt.

There are a lot of good metaphors out there for contextualization, including the “previously on…” at the beginning of some TV shows, or the famous text crawl at the beginning of the Star Wars movies.

Both of these examples serve the same function: they give important information about what has happened off-screen that the audience needs to know to understand what is about to happen on-screen.

In your essay, contextualization is the same. You give your reader information about what else has happened, or is happening, in history that will help them understand the specific topic and argument you are about to make.

There is no specific requirement for where contextualization must appear in your essay. The easiest place to include it, however, is in your introduction . Use context to get your reader acquainted with the time, place, and theme of your essay, then transition into your thesis.

Good contextualization doesn’t have to be long, and it doesn’t have to go into a ton of detail, but it does need to do a few very specific things.

Your contextualization needs to refer to events, developments and/or processes outside the time and place of the prompt. It could address something that occurred in an earlier era in the same region as the topic of the prompt, or it could address something happening at the same time as the prompt, but in a different place. Briefly describe this outside information.

Then, connect it to your thesis/argument. The language from the College Board is that contextualization must be “relevant to the prompt,” and in practical terms this means you have to show the connection. A transition sentence or phrase is useful here (plus, this is why contextualization makes the most sense in the introduction!)

Also, contextualization needs to be multiple consecutive sentences, so it’s all one argument (not sprinkled around in a paragraph). The introduction is the best place for contextualization, but not the only place. 

Basically, choose a connected topic that “sets the stage” for your thesis, and briefly describe it in a couple sentences. Then, make a clear connection to the argument of your thesis from that outside information.

In the period 1450-1750, both European and Asian powers expanded their reach and created large empires across the world. In Asia, the trend was toward large, land-based empires which were controlled from a central capital city. Europeans built empires that stretched across oceans included territories in the Americas, Africa, and Asia.

This contextualization works by addressing the time period of the prompt and establishing basic definitions for empire-building and the types of empires (land and maritimes.) These definitions will be valuable context for seeing the comparisons developed in the thesis and body paragraphs of this essay.

Evidence: Provide Specific Examples

For this point, the focus is simply about having evidence. Evidence is the historical detail you include in your writing. The specific facts and examples that prove your argument. In the LEQ, your evidence comes your knowledge of history. 

Evidence goes in your body paragraphs. In fact, the bulk of your body paragraphs will be made up of evidence and supporting analysis or commentary that connects that evidence to other evidence and/or to the argument you are making.

Good evidence is specific, accurate, and relevant to the prompt. For this point, simply including multiple pieces of quality evidence is enough. If you’re a numbers person, a good starting point is to aim for two pieces of quality evidence in each body paragraph and go up from there.

In order for your evidence to count for this point, it needs to be really specific. Using course-specific vocabulary is a great strategy here to know that you are writing specific evidence. If you can’t remember a specific vocabulary term, describe what you mean in plain language with as much detail as possible.

Though the Ottoman Sultans were Muslims, they ruled over a population that included fellow Muslims, Christians, and Jews.

This evidence works because it includes specific and relevant details, namely the religions of both the Ottoman rulers and the diverse population they ruled over.

Evidence: Supports an Argument with Evidence

In addition to having evidence, this point is about using that evidence to support an argument in response to the prompt. Basically, connect your evidence back to your topic sentence and/or thesis.

Supporting statements go with your evidence in your body paragraph. Ideally, a connecting statement comes right before or after a piece of evidence. 

This point is harder to earn than the previous evidence point, because it’s a little more difficult to explain fully. 

One way to know if you are doing this at all is to look at the topic sentences of your body paragraphs. First of all, do you have one? You should. The first sentence of your body paragraph should make it clear what you are talking about in that paragraph. It should relate to some aspect of your thesis, and it should be connected to the reasoning skill you have chosen to organize your argument.

One characteristic shared by both kinds of empires was the need to adapt to diverse populations. As the Ottoman empire expanded its influence, it took over territory previously controlled by the Byzantines. Though the Ottoman Sultans were Muslims, they now ruled over a population that included fellow Muslims, Christians, and Jews. In order to keep peace within their empire, the Ottomans allowed people to continue practicing their traditional faiths. Ottoman cities such as Istanbul had areas of the cities set aside where different groups could live and worship without interference from others .

This section works because it defines the adaptation made by Ottoman rulers to effectively rule a diverse population, and elaborates on both how and why that adaptation was made.

Following your topic sentence, your body paragraph should elaborate on the idea in that topic sentence, using the evidence to prove your point. At first, you may rely on phrases like “this shows…” or “this means…,” which can get repetitive, but may also help you know when you are making the connections between evidence and argument explicit.

Analysis and Reasoning: Historical Reasoning

A good argument needs structure, and yours needs to use one of the course reasoning skills to create that structure. You can choose whichever skill works best for a particular prompt: causation , comparison , or continuity and change over time .

Strong reasoning goes throughout an essay, so this will be the overarching structure of your writing from the thesis through your body paragraphs.

The reasoning doesn’t necessarily have to be completely balanced or even in order to count, which gives you room to write about what you know best. For example, in an essay structured around continuity and change, you might spend most of your time addressing changes and relatively little time addressing continuity. And that’s ok.

The best essays do address both “sides” of the historical reasoning, and yours should too. If you created a complex thesis in your introduction, you can extend those ideas into your body paragraphs. Even if you don’t have equal sentences or paragraphs for each topic, as long as you address the reasoning process in your essay, you’re on the right track.

Analysis and Reasoning: Complexity

The second part of the Analysis and Reasoning scoring category is complexity. This is by far the most challenging part of the LEQ, and the point earned by the fewest students. It isn’t impossible, just difficult. Part of the difficulty is that it is the least concrete skill to teach and practice.

If you’re already feeling overwhelmed by the time limits of the LEQ, don’t stress about complexity. Focus on writing the best essay you can that answers the prompt. Plenty of students earn 5’s without the complexity point.

If you are ready to tackle this challenge, keep reading!

The College Board awards this point for essays that “demonstrate a complex understanding” of the topic of the prompt.

Complexity cannot be earned with a single sentence or phrase. It must show up throughout the essay. 

A complex argument starts with a complex thesis. A complex thesis must address the topic of the prompt in more than one way. Including a counter-claim or alternate viewpoint in the thesis is a good way to set up a complex argument, because it builds in room within the structure of your essay to address more than one idea (provided your body paragraphs follow the structure of your thesis!)

A complex argument may include corroboration - evidence that supports or confirms the premise of the argument. Clear explanation that connects each piece of evidence to the thesis will help do this. In the LEQ, your evidence is all from your knowledge of history, so it’s up to you to fully explain how that evidence backs up your thesis. Consistent, thoughtful explanation can go a long way toward the complexity point.

A complex argument may also include qualification - evidence that limits or counters an initial claim. This isn’t the same as undoing or undermining your claim. Qualifying a claim shows that it isn’t universal. An example of this might be including continuity in an essay that is primarily about change.

A final way to introduce complexity to your argument is through modification - using evidence to change your claim or argument as it develops. Modification isn’t quite as extreme as qualification, but it shows that the initial claim may be too simple to encompass the reality of history.

Since no single sentence can demonstrate complexity on its own, it’s difficult to show examples of complex arguments. Fully discussing your claim and its line of reasoning, and fairly addressing your counter-claim or alternate view is the strongest structure to aim for a complexity point. Explain everything as you go and aim for success!

How to Rock the LEQ: The Process

Before you start writing....

It is tempting to just start writing at the beginning of your LEQ time, especially if you took extra time to write your DBQ and you’re feeling some pressure. It’s actually better to take a few minutes to analyze the prompt and plan your essay before you start writing to give yourself the best shot at success. You might surprise yourself with how quickly an essay comes together after you create a solid plan.

The very first thing you should do with any prompt is to be sure you understand the question . Misunderstanding the time period, topic, or geographic region of a prompt can kill a thoughtful and well-argued essay. When you’re practicing early in the year, go ahead and re-write the prompt as a question. Later on you can re-phrase it mentally without all the work.

As you think about the question, start thinking about which reasoning skill might apply best for this prompt: causation, comparison, or continuity and change over time. 

Original prompt - Develop an argument that compares the process of empire building in land-based and maritime empires in the period 1450-1750 CE. 

Revised - What were the key similarities and differences in the ways that land-based (Asian) and maritime (European) empires built their governments and power between 1450-1750?

Now that you know what you’re writing about, take a few minutes to brainstorm what you know about that topic. You can make a simple graphic organizer to help you see relationships between information (i.e. a Venn diagram, T-chart, timeline, etc.), or just jot down ideas as they come to mind.

Go back over you list and mark which ideas work best as context (generally broader and less related to the prompt) and which ideas work best as evidence (more specific.)

If you have time, brainstorm a sample thesis and/or outline for how you want to structure your ideas. This may seem like an extravagance with limited time, but it can be a great cheat sheet for you if you lose your way in the middle of a body paragraph.

When you have a plan you like, start writing!

Writing the essay

Your introduction should include your contextualization and thesis. Start with a statement that establishes your time and place in history, and follow that with a brief description of the historical situation. Connect that broader context to the theme and topic of the prompt. Then, make a claim that answers the prompt, with an overview of your reasoning and any counter-claim you plan to address.

Body paragraphs will vary in length, depending on how many documents or other pieces of evidence you include, but should follow a consistent structure. Start with a topic sentence that introduces the specific aspect of the prompt that paragraph will address. There aren’t specific points for topic sentences, but they will help you stay focused.

Follow your topic sentence with a piece of evidence and connect it back to your topic sentence and/or thesis. Continue with 1-2 pieces of evidence and more explanation until you have completed the argument of your topic sentence. Then start a new paragraph with a new topic sentence.

Each body paragraph will follow this general format, and there is no set number of paragraphs for the LEQ (minimum or maximum.) Write as many paragraphs as you need to fully answer the prompt by developing the argument (and counter-argument if applicable) from your thesis.

If you have time, you may choose to write a conclusion . It isn’t necessary, so you can drop it if you’re rushed. BUT, the conclusion is the only place where you can earn the thesis point outside the introduction, so it’s not a bad idea. You could re-state your thesis in new wording, or give any final thoughts in terms of analysis about your topic. You might solidify your complexity point in the conclusion if written well.

Since most people write the DBQ first, when you finish the LEQ you’re done with your AP Exam. Congratulations!

Sample Prompts

AP World History: Modern

In the period 1450-1750 CE, empires achieved increased scope and influence around the world, shaping and being shaped by the diverse populations they incorporated.

Develop an argument that compares the process of empire building in land-based and maritime empires in the period 1450-1750 CE.

AP US History

Evaluate the extent to which Massachusetts and Virginia differed in the ways that economic development affected their politics between 1607 and 1750.

AP European History

Evaluate the effectiveness of challenges to royal authority in Eastern Europe in the seventeenth and eighteenth centuries.

The LEQ Rubric (Quick Reference)

Stay Connected

© 2024 Fiveable Inc. All rights reserved.

AP® and SAT® are trademarks registered by the College Board, which is not affiliated with, and does not endorse this website.

Choose Your Test

Sat / act prep online guides and tips, the best college essay length: how long should it be.

College Essays

Figuring out your college essay can be one of the most difficult parts of applying to college. Even once you've read the prompt and picked a topic, you might wonder: if you write too much or too little, will you blow your chance of admission? How long should a college essay be?

Whether you're a terse writer or a loquacious one, we can advise you on college essay length. In this guide, we'll cover what the standard college essay length is, how much word limits matter, and what to do if you aren't sure how long a specific essay should be.

How Long Is a College Essay? First, Check the Word Limit

You might be used to turning in your writing assignments on a page-limit basis (for example, a 10-page paper). While some colleges provide page limits for their college essays, most use a word limit instead. This makes sure there's a standard length for all the essays that a college receives, regardless of formatting or font.

In the simplest terms, your college essay should be pretty close to, but not exceeding, the word limit in length. Think within 50 words as the lower bound, with the word limit as the upper bound. So for a 500-word limit essay, try to get somewhere between 450-500 words. If they give you a range, stay within that range.

College essay prompts usually provide the word limit right in the prompt or in the instructions.

For example, the University of Illinois says :

"You'll answer two to three prompts as part of your application. The questions you'll answer will depend on whether you're applying to a major or to our undeclared program , and if you've selected a second choice . Each response should be approximately 150 words."

As exemplified by the University of Illinois, the shortest word limits for college essays are usually around 150 words (less than half a single-spaced page). Rarely will you see a word limit higher than around 650 words (over one single-spaced page). College essays are usually pretty short: between 150 and 650 words. Admissions officers have to read a lot of them, after all!

Weigh your words carefully, because they are limited!

How Flexible Is the Word Limit?

But how flexible is the word limit? What if your poignant anecdote is just 10 words too long—or 100 too short?

Can I Go Over the Word Limit?

If you are attaching a document and you need one or two extra words, you can probably get away with exceeding the word limit by such a small amount. Some colleges will actually tell you that exceeding the word limit by 1-2 words is fine. However, I advise against exceeding the word limit unless it's explicitly allowed for a few reasons:

First, you might not be able to. If you have to copy-paste it into a text box, your essay might get cut off and you'll have to trim it down anyway.

If you exceed the word limit in a noticeable way, the admissions counselor may just stop reading your essay past that point. This is not good for you.

Following directions is actually a very important part of the college application process. You need to follow directions to get your letters of recommendation, upload your essays, send supplemental materials, get your test scores sent, and so on and so forth. So it's just a good general rule to follow whatever instructions you've been given by the institution. Better safe than sorry!

Can I Go Under the Word Limit?

If you can truly get your point across well beneath the word limit, it's probably fine. Brevity is not necessarily a bad thing in writing just so long as you are clear, cogent, and communicate what you want to.

However, most college essays have pretty tight word limits anyways. So if you're writing 300 words for an essay with a 500-word limit, ask yourself: is there anything more you could say to elaborate on or support your points? Consult with a parent, friend, or teacher on where you could elaborate with more detail or expand your points.

Also, if the college gives you a word range, you absolutely need to at least hit the bottom end of the range. So if you get a range from the institution, like 400-500 words, you need to write at least 400 words. If you write less, it will come across like you have nothing to say, which is not an impression you want to give.

What If There Is No Word Limit?

Some colleges don't give you a word limit for one or more of your essay prompts. This can be a little stressful, but the prompts generally fall into a few categories:

Writing Sample

Some colleges don't provide a hard-and-fast word limit because they want a writing sample from one of your classes. In this case, a word limit would be very limiting to you in terms of which assignments you could select from.

For an example of this kind of prompt, check out essay Option B at Amherst :

"Submit a graded paper from your junior or senior year that best represents your writing skills and analytical abilities. We are particularly interested in your ability to construct a tightly reasoned, persuasive argument that calls upon literary, sociological or historical evidence. You should NOT submit a laboratory report, journal entry, creative writing sample or in-class essay."

While there is usually no word limit per se, colleges sometimes provide a general page guideline for writing samples. In the FAQ for Option B , Amherst clarifies, "There is no hard-and-fast rule for official page limit. Typically, we anticipate a paper of 4-5 pages will provide adequate length to demonstrate your analytical abilities. Somewhat longer papers can also be submitted, but in most cases should not exceed 8-10 pages."

So even though there's no word limit, they'd like somewhere in the 4-10 pages range. High school students are not usually writing papers that are longer than 10 pages anyways, so that isn't very limiting.

Implicit Length Guideline

Sometimes, while there's no word (or even page) limit, there's still an implicit length guideline. What do I mean by this?

See, for example, this Western Washington University prompt :

“Describe one or more activities you have been involved in that have been particularly meaningful. What does your involvement say about the communities, identities or causes that are important to you?”

While there’s no page or word limit listed here, further down on page the ‘essay tips’ section explains that “ most essay responses are about 500 words, ” though “this is only a recommendation, not a firm limit.” This gives you an idea of what’s reasonable. A little longer or shorter than 500 words would be appropriate here. That’s what I mean by an “implicit” word limit—there is a reasonable length you could go to within the boundaries of the prompt.

But what's the proper coffee-to-paragraph ratio?

Treasure Hunt

There is also the classic "treasure hunt" prompt. No, it's not a prompt about a treasure hunt. It's a prompt where there are no length guidelines given, but if you hunt around on the rest of the website you can find length guidelines.

For example, the University of Chicago provides seven "Extended Essay" prompts . You must write an essay in response to one prompt of your choosing, but nowhere on the page is there any guidance about word count or page limit.

However, many colleges provide additional details about their expectations for application materials, including essays, on FAQ pages, which is true of the University of Chicago. On the school’s admissions Frequently Asked Questions page , they provide the following length guidelines for the supplemental essays: 

“We suggest that you note any word limits for Coalition or Common Application essays; however, there are no strict word limits on the UChicago Supplement essays. For the extended essay (where you choose one of several prompts), we suggest that you aim for around 650 words. While we won't, as a rule, stop reading after 650 words, we're only human and cannot promise that an overly wordy essay will hold our attention indefinitely. For the “Why UChicago?” essay, we suggest about 250-500 words. The ideas in your writing matter more than the exact number of words you use!”

So there you go! You want to be (loosely) in the realm of 650 for the extended essay, and 250-500 words for the “Why UChicago?” essay.

Help! There Really Is No Guidance on Length

If you really can't find any length guidelines anywhere on the admissions website and you're at a loss, I advise calling the admissions office. They may not be able to give you an exact number (in fact, they probably won't), but they will probably at least be able to tell you how long most of the essays they see are. (And keep you from writing a panicked, 20-page dissertation about your relationship with your dog).

In general, 500 words or so is pretty safe for a college essay. It's a fairly standard word limit length, in fact. (And if you're wondering, that's about a page and a half double-spaced.) 500 words is long enough to develop a basic idea while still getting a point across quickly—important when admissions counselors have thousands of essays to read!

"See? It says 500 words right there in tiny font!"

The Final Word: How Long Should a College Essay Be?

The best college essay length is usually pretty straightforward: you want to be right under or at the provided word limit. If you go substantially past the word limit, you risk having your essay cut off by an online application form or having the admissions officer just not finish it. And if you're too far under the word limit, you may not be elaborating enough.

What if there is no word limit? Then how long should a college essay be? In general, around 500 words is a pretty safe approximate word amount for a college essay—it's one of the most common word limits, after all!

Here's guidance for special cases and hunting down word limits:

If it's a writing sample of your graded academic work, the length either doesn't matter or there should be some loose page guidelines.

There also may be implicit length guidelines. For example, if a prompt says to write three paragraphs, you'll know that writing six sentences is definitely too short, and two single-spaced pages is definitely too long.

You might not be able to find length guidelines in the prompt, but you could still hunt them up elsewhere on the website. Try checking FAQs or googling your chosen school name with "admissions essay word limit."

If there really is no word limit, you can call the school to try to get some guidance.

With this advice, you can be sure you've got the right college essay length on lockdown!

Hey, writing about yourself can even be fun!

What's Next?

Need to ask a teacher or friend for help with your essay? See our do's and dont's to getting college essay advice .

If you're lacking in essay inspiration, see our guide to brainstorming college essay ideas . And here's our guide to starting out your essay perfectly!

Looking for college essay examples? See 11 places to find college essay examples and 145 essay examples with analysis !

Ellen has extensive education mentorship experience and is deeply committed to helping students succeed in all areas of life. She received a BA from Harvard in Folklore and Mythology and is currently pursuing graduate studies at Columbia University.

Student and Parent Forum

Our new student and parent forum, at ExpertHub.PrepScholar.com , allow you to interact with your peers and the PrepScholar staff. See how other students and parents are navigating high school, college, and the college admissions process. Ask questions; get answers.

Ask a Question Below

Have any questions about this article or other topics? Ask below and we'll reply!

Improve With Our Famous Guides

• For All Students

The 5 Strategies You Must Be Using to Improve 160+ SAT Points

How to Get a Perfect 1600, by a Perfect Scorer

Series: How to Get 800 on Each SAT Section:

Score 800 on SAT Math

Score 800 on SAT Reading

Score 800 on SAT Writing

Series: How to Get to 600 on Each SAT Section:

Score 600 on SAT Math

Score 600 on SAT Reading

Score 600 on SAT Writing

Free Complete Official SAT Practice Tests

What SAT Target Score Should You Be Aiming For?

15 Strategies to Improve Your SAT Essay

The 5 Strategies You Must Be Using to Improve 4+ ACT Points

How to Get a Perfect 36 ACT, by a Perfect Scorer

Series: How to Get 36 on Each ACT Section:

36 on ACT English

36 on ACT Math

36 on ACT Reading

36 on ACT Science

Series: How to Get to 24 on Each ACT Section:

24 on ACT English

24 on ACT Math

24 on ACT Reading

24 on ACT Science

What ACT target score should you be aiming for?

ACT Vocabulary You Must Know

ACT Writing: 15 Tips to Raise Your Essay Score

How to Get Into Harvard and the Ivy League

How to Get a Perfect 4.0 GPA

How to Write an Amazing College Essay

What Exactly Are Colleges Looking For?

Is the ACT easier than the SAT? A Comprehensive Guide

Should you retake your SAT or ACT?

When should you take the SAT or ACT?

Stay Informed

Get the latest articles and test prep tips!

Looking for Graduate School Test Prep?

Check out our top-rated graduate blogs here:

GRE Online Prep Blog

GMAT Online Prep Blog

TOEFL Online Prep Blog

Holly R. "I am absolutely overjoyed and cannot thank you enough for helping me!”

What are your chances of acceptance?

Calculate for all schools, your chance of acceptance.

Your chancing factors

Extracurriculars.

How Long Should Your College Essay Be? What Is the Ideal Length?

What’s covered: , personal statement length vs. supplemental essay length, are college essay word limits hard, what if a college essay word count isn’t given, what if you need to submit a graded paper, where to get your essays edited.

Students often spend hours agonizing over the best topics for their college essays. While it’s natural to wonder whether your personal statement is original or compelling enough, there’s one aspect of the process that shouldn’t cause you undue stress—how many words should a college essay be? Fortunately, with a little research, you can uncover the ideal college essay length for all your applications.

Unlike high school assignments, which typically have a strict page requirement, most colleges provide a word limit or word range for their application essays. This practice helps ensure that essays are the same length regardless of font or formatting. A good guideline is that students should strive to get as close as possible to the upper limit of the word range without exceeding it. Keep reading to learn more about best practices for college essay length.

How many words should a college essay be? Personal statements are generally 500-650 words. For example, the Common Application , which can be used to apply to more than 800 colleges, requires an essay ranging from 250-650 words . Similarly, the Coalition Application , which has 150 member schools, features an essay with a recommended length of 500-650 words.

650 words is the most common limit for your personal statement, but some schools may ask students to write more or less. For example, ApplyTexas , a platform used to apply to Texas public universities and other select colleges, requests essays with requirements that vary by school. For example, students applying to UT Austin will need to submit an essay of 500-700 words, along with three short-answer questions of 250-300 words each.

On the other hand, the University of California (UC) application includes a Personal Insight section with eight prompts . Students are asked to respond to any four of these prompts, with each response topping out at 350 words.

Additionally, some schools request a few supplemental essays, which are typically shorter than a personal statement. These questions are designed to gain more information about a student’s interests and abilities, and may include topics like your reasons for wanting to attend their school, your desired major, or your favorite activity.

Most schools require 1-3 supplemental essays, though some may require more or none at all (see our list of top colleges without supplemental essays ). These essays tend to be around 250 words, but some may be just as long as your main essay. For example, Cornell requires applicants to write a second supplemental essay (of 650 words max) that is specific to the program they’re applying to. The exception to this is the Cornell College of Engineering, for which applicants are required to compose two supplemental essays of 250 words max each.

For best results, keep your essays within the word range provided. While you don’t have to hit the count exactly, you should aim to stay within a 10% difference of the upper limit—without including fluff or filler. For example, if the school requests 500 words, try to ensure that your essay is between 450 and 500 words.

For the Common App, try to stay within 550-650 words, even though the given range is 250-650. Any submission shorter than 500 words will make it look as though you simply didn’t care enough to give your best effort. An essay shorter than 500 words won’t be long enough to truly share who you are and what matters to you.

Exceeding the word count isn’t an option—the application portal cuts off anything over the maximum number of allowed words. This is something you want to be particularly careful of if you’re drafting your essay in a Word or Google document and pasting it into the application.

Although most schools provide applicants with a specific word count, some offer more general guidelines. For example, a college may ask for a particular number of pages or paragraphs.

If you aren’t given a word count, try to adhere to the best practices and conventions of writing. Avoid writing especially short or overly long paragraphs—250 words per paragraph is generally a safe upper limit. If you’re asked to write a certain number of pages, single- or double-spaced, stick to a standard font and font size (like 12-point Times New Roman).

In the event that the college doesn’t offer any guidelines at all, aim for an essay length of around 500 words.

While essays are the most commonly requested writing sample, some colleges ask for additional pieces of content. For example, Princeton University requires students to submit a previously graded paper for evaluation .

Princeton offers guidelines that cover length, but if another school requests an old paper and doesn’t offer length requirements, a paper ranging from 3-5 pages should yield the best results. The goal is to select a paper long enough to showcase your writing skills and unique voice, but short enough that the admissions officer doesn’t get bored reading it.

Is your essay effective while staying within the required word count? It’s hard to evaluate your own writing, especially after rereading it numerous times. CollegeVine’s free Peer Essay Review provides an opportunity to have your essay reviewed by a fellow student, for free. Similarly, you can help other students by reviewing their essays—this is a great way to refine your own writing skills.

Expert advice is also available. CollegeVine’s advisors are prepared to help you perfect your personal statement and submit a successful application to your top schools. Find the right advisor for you to improve your chances of getting into your dream school!

Related CollegeVine Blog Posts

How Long Should My Paper Be?

• Writing Research Papers
• Writing Essays
• English Grammar
• M.Ed., Education Administration, University of Georgia
• B.A., History, Armstrong State University

It's really annoying when a teacher or professor gives a writing assignment and doesn't offer specific instruction about how long the response should be. There is a reason for this, of course. Teachers like for students to focus on the meaning of the work and not just fill a given amount of space.

But students like guidance! Sometimes, if we don't have parameters to follow, we are lost when it comes to getting started. For this reason, I'll share these general guidelines pertaining to test answers and paper length. I've asked several professors to explain what they really mean when they say the following:

"Short answer essay" - We often see short answer essays on exams. Focus on the "essay" more than the "short" on this one. Write an essay that contains at least five sentences. Cover about a third of a page to be safe.

"Short answer" - You should respond to a "short answer" question on an exam with two or three sentences. Be sure to explain what , when , and why .

"Essay question" - An essay question on an exam should be at least a full page in length, but longer is probably better. If you're using a blue book, the essay should be at least two pages long.

"Write a short paper" - A short paper is normally three to five pages long.

"Write a paper" - How unspecific can a teacher be? But when they give such generic instruction, it means they really want to see some meaningful writing. Two pages of great content will work better than six or ten pages of fluff.

• How Long Should Your Web Page Be?
• How Long Should Your Common Application Short Answer Essay Be?
• Ideal College Application Essay Length
• Preparing for Final Exams
• Top 10 Tips for Passing the AP US History Exam
• 3 Steps to Ace Your Next Test
• Creating and Scoring Essay Tests
• How Long Should You Study a Subject?
• How to Write and Format an MBA Essay
• How to Develop a Research Paper Timeline
• How to Write a Research Paper That Earns an A
• Sample College Application Short Answer Essay
• Short Answer Mistakes
• Sample Short Answer Essay on Running
• Private School Application Essay Tips
• 10 Tips for Art History Students

How Long Is an Essay? The Ultimate Essay Length Guide

It’s safe to say that most students struggle with the word limit within an essay. Sometimes, it’s hard to find ideas for a text and meet the word requirement for every part of the paper. With so many factors influencing essay length, it’s easy to get confused.

Our specialists will write a custom essay specially for you!

Luckily, our custom-writing team has your back. In this article, our custom-writing experts will answer all your questions regarding essay length. We will also help you write papers with an ideal number of words!

📜 Is Essay Length Important?

📏 essay parts: recommended length.

• 🤔 How to Make Essays Shorter or Longer
• 📑 Essay Length & Formatting
• ❓ Different Academic Levels FAQ
• 📚 Essay Length: Different Types
• ⭐ Other Aspects
• 📝 Essay Examples

🔍 References

Often, the phrase “word limit” causes panic among students. After all, if an essay is too long or too short, your grade will be lowered. However, in reality, there’s nothing to worry about. When it comes to words, limitations are beneficial for both the students and the professors.

Let’s see what exactly it means.

Many people believe that the longer an essay is, the better. However, according to Frontiers, research shows that it’s a bias that couldn’t be further from the truth. A perfect-length paper is one that allows students to express their ideas and showcase their knowledge fully while keeping it clean and simple.

What Influences Essay Length

Various factors determine the length of an essay. Here are the most important ones:

Just in 1 hour! We will write you a plagiarism-free paper in hardly more than 1 hour

Let’s start with the essentials. Usually, assignment length is given as a number of words rather than pages. Unless your supervisor or instructor mentions any specific limitations, it’s acceptable to be 10% below or above the word limit.

It’s also worth knowing the 80/20 rule . According to it, the body should constitute 80% of the text, while the intro and the conclusion take up the remaining 20%.

Keep reading to learn more about the recommended length of each essay part. The main numbers are shown in the table below:

How Long Should an Introduction Be?

An introduction is the first section and the face of your essay. For that reason, it needs to be compelling and well-thought-out. Usually, it consists of 3 to 5 sentences or 50 to 80 words .

An introduction must have a hook, some background information, and a thesis statement. While the attention grabber and the thesis are usually brief, you may need 2 to 3 sentences for the background. To avoid going overboard, try to stay on topic and don’t add any filler.

Receive a plagiarism-free paper tailored to your instructions. Cut 15% off your first order!

How Long Is a Body Paragraph in an Essay?

The length of a body paragraph may vary. Sometimes, it can be limited to a single sentence. In other cases, it may take up a whole page. Usually, it’s recommended to have between 80 and 200 words (5-8 sentences) per body paragraph.

Since the paper’s body contains the most information, it’s necessary to explain and support your ideas properly. That’s why it’s no big deal if your body paragraphs go slightly over the word limit.

How Many Body Paragraphs Should Be in an Essay?

Like the word count, the number of paragraphs is determined by the type of paper and its topic. The minimum is 1. Generally, however, the body consists of 3-5 paragraphs , 1 for each argument.

To improve your paper’s structure, ensure that there are as many paragraphs as there are points in your thesis statement. Each one should have a purpose and support your arguments. If there’s any fluff, it’s better to get rid of it.

How Long Should a Conclusion Be?

Like the introduction, the conclusion consists of 50-80 words . It’s essential to keep it simple and only mention the central ideas. A weak concluding sentence may affect the reader’s understanding of the topic and spoil the overall impression of your paper.

Get an originally-written paper according to your instructions!

🤔 How to Make Essays Shorter or Longer: Best Tips

Undoubtedly the essay’s content is more important than the number of words you use. But there are times when students go more than 10-15% below or over the limit. Is there a solution to this problem?

Yes, there is! In this section, we will share the most useful tips to help you stay on point with your paper’s word count.

How to Make Essays Longer

Since having enough words is essential for a good grade, we’ve collected the best tips that can help you lengthen your essay without teachers noticing:

• Use relevant quotations.  You don’t need to litter your essay with citations, but using them whenever appropriate is a great idea. For instance, if you’re working on a book analysis, referencing a couple of direct quotes from the source text will make your essay more credible and increase the word count.
• Give examples.  Go through the claims in your paper and provide additional evidence where possible. It will make your essay longer and more informative.
• Use transitional expressions.  Adding transition words and phrases is a natural way of increasing the number of words. It will also improve your essay’s readability. 
• Add more references.  Providing references is always a good idea when writing a formal essay. That way, you will increase the number of words and make your paper more credible.
• Work on your descriptions.  If you struggle to develop new ideas, go over what you’ve already written and consider adding some descriptive words. It’s a great idea for creative essays to include more imagery. 

How to Shorten an Essay

Another struggle of academic writing is cutting down the number of words in your essay to meet a set limit. We are here to tell you that it’s not that hard. Writing straightforwardly and keeping your sentences short is a key to concise content. Here are several strategies you may use to tighten a lengthy essay:

• Choose the active voice.  It takes up less space than passive voice. Using it also makes your writing more professional and compelling.
• Remove needless transitions.  Transitions can indeed maintain the flow of the paper. But some transitional phrases can be easily removed.
• Get rid of unnecessary adverbs and adjectives.  Some students tend to overuse adjectives and adverbs. It adds wordiness to their writing.
• Avoid running starts.  Some students like to start their sentences with long phrases like: “there are,” “it is believed,” or “the fact that.” Getting rid of them makes texts much more concise.
• Delete “that.”  In most cases, the word “that” can often be easily removed from texts.

Another cool trick is to use our summarizing tool as essay shortener. Try it out!

📑 How Long Is an Essay Depending on Formatting?

As we mentioned earlier, the essay’s length is usually limited by the number of words. But sometimes, a teacher may ask you to write a specific number of pages. This is trickier because the amount of text you can place on the page depends on the formatting. By using the font size and spacing properly, it’s possible to make the paper visually longer or shorter. Let’s discuss it in more detail.

Essay Spacing: How Does It Affect the Length?

• Adjusting the spacing between lines.  Try to make the changes as slight as possible. For instance, if you were asked to double-space the paper, use 2.1 or 2.2 spacing instead. Another option is to slightly extend spaces between paragraphs.
• Extending the margin size.  You can increase the right and bottom margins by a quarter to make very subtle changes in length. For example, if the margins are 1 inch , you can set them at 1.25 inches instead. 
• Increasing the spacing between characters.  It is less noticeable than the line spacing. Still, try not to overdo it and keep the numbers between 1.2 and 1.5 . 
• Adjusting the footer.  Add a footer with page numbers to stretch the bottom margin even further.
• Lengthening the header.  You can extend your header by adding your name, e-mail address, or other relevant information. Another option is double-spacing it.

Length of an Essay: Font and Size

• Using the right type of font.  If your instructor didn’t specify which font you should use, go for the bigger ones. We suggest Arial, Bangla Sangam MN, Cambria, or Quicksand. They will make your text look longer without being too on the nose.  
• Using a bigger font size.  This is another technique that can come in handy. However, be careful and don’t increase your font by more than 0.1-0.5 pt.  
• Increasing the size of periods and commas.   This is one of the less noticeable tricks you can use. For instance, if your paper’s font is 12 pt. , increase it to 14 pt. only for punctuation marks. Italicizing periods and commas will also add several lines of length to your essay. 

What to Do if There Are No Length Guidelines

Sometimes a teacher sets no word limit for a written work. What to do in that case? Well, first, you can ask your professor to confirm if they have simply forgotten to mention it. But if that’s not the case, here are a couple of helpful solutions:

• Think of the paragraph number.  Sometimes, you may be given the number of paragraphs instead of words. In that case, you can decide on the number of words depending on how many paragraphs you have. 
• Think about the topic’s complexity.  The length of your paper is also directly dependent on the theme. If the topic is simple, 4-5 paragraphs will be enough. A more complex issue may require an in-depth explanation, so your essay can be 6-8 paragraphs long.

❓ Essay Length for Different Academic Levels FAQ

The length of the elementary school essay is usually short. Usually, a paper needs to have around 3-5 paragraphs, with 4-5 sentences per paragraph. Primary school essays can be 1-2 paragraphs long.

The word limit for a middle school essay is usually between 300 to 1000 words. The most common essay length is 500 words, which is about 5 paragraphs. However, it may differ from school to school.

The length of the high school essay may differ depending on the school and the complexity of the task itself. Usually, however, a paper can be between 300 to 1000 words long.

The length of the undergraduate college essay often falls within the range of 1500 to 2100 words. It translates into roughly 5-7 pages. 5 pages is the most common essay length at this level.

When it comes to the graduate school admission essay, the word limit is usually between 500 and 1000 words. It’s possible to go slightly over or below the set limit; however, it’s best to stick to the requirements as close as possible.

📚 How Long Should an Essay Be: Different Types

Now, let’s talk about different types of essays. How long should they be? Keep reading to learn about the length of college essays, short and extended ones, scholarship essays, and research papers.

How Long Is a College Essay?

When it comes to a college essay, it’s more important to stick to the word limit than with any other paper. Some teachers may refuse to read it unless it meets all the requirements.

The shortest limit for a college essay is about 250 words which is the shortest length of a Common App personal statement. It’s also rare to see a good college essay with over 650 words . So, an average piece usually has between 150 and 650 words ; you can go over or below the limit by 50.

How Long Is a Paragraph in College Essays?

A college essay usually consists of 4-5 paragraphs . One paragraph takes about 1/3 of the page, which is roughly 5 sentences . Each sentence corresponds with one of the following components:

• Topic sentence.
• Explanation.
• Transitions.

College Essay Length Requirements: Top 5 Schools

To understand the requirements for a college application essay even better, take a look at the table below. It showcases the top 5 schools and their length criteria for personal statements. Keep it in mind when writing your college essay:

How Long Is a Short Essay?

A short essay is usually 500 words long. Using 12pt Times New Roman font with standard margins and double spacing should result in about 2 pages of text.

Extended Essay Length

An extended essay is different from a short or a standard one. It requires extensive research and thorough explanation. That’s why the upper limit for this kind of essay is 4000 words . In this case, a typical essay length is 3500 words or 18 paragraphs .

Scholarship Essay Length

Generally, scholarship papers have a limit of 500 words , which is 1 page in length. Most scholarship programs provide additional requirements that indicate the minimum number of words or pages. If there are no set limitations, you can stick to the limit.

How Long Is a Research Paper?

Typically, a research paper is between 4000 and 6000 words long. Sometimes, there are shorter papers, which have around 2000 words, or in-depth ones with over 10000 words.

⭐ Other Aspects of Essay Length

When it comes to essay length, many different aspects come into play. Here, we’ve gathered all the essential information regarding an essay’s number of pages, paragraphs, words, and references.

How Many Paragraphs Are in an Essay?

Sometimes, it is more convenient to count paragraphs rather than words. Let’s now figure out how many paragraphs are in essays of different lengths. You may also check out the examples to see what such an essay looks like:

How to Count Paragraphs in an Essay Based on Word Count

You can also count the number of body paragraphs for your essay using the formula below:

Number of body paragraphs (average) = (TWC – TWC*0.16)/100

• TWC – total word count
• 0.16 – an average percentage of total word count for introduction and conclusion
• 100 – an average number of words per paragraph

How Many Pages Are in an Essay?

The number of pages in your essay may vary from subject to subject. But it’s still possible to determine the number of pages based on word count. Check out the numbers below to see the conversions with bonus examples:

You can also use a specialized calculator such as Word Counter to determine a number of pages in your essay.

What Does an Essay Look Like when Typed?

You might be wondering: what do essays of different lengths look like when typed? Well, here’s the table where you can find out the metrics for single- and double-spaced papers.

How Many Pages Are in a Handwritten Essay?

In case you need to turn in a handwritten paper, you should check out the table below.

Counting Words in a Handwritten Essay

If you don’t have enough time to count the words in your handwritten essay one by one, here’s what you can do:

• Count how many words there are in one line. Take the first and last lines and a line in the middle of a page. Let’s say there are 15, 14, and 15 words in them. Then, the average number of words per line is 15.
• Next, count how many lines there are on one page. Let’s say there are 17 lines on a page.
• Take the number of words per line and multiply it by the number of lines per page. In our case, we multiply 15 by 17. So, there are 255 words per page on average.
• Finally, multiply the number of words per page by the number of pages. If your essay has 3 pages, it is approximately 765 words long.

How Long Does it Take to Write an Essay?

It is crucial to know how long writing will take you, especially if you are working on an exam essay or just short on time. Note that you need to consider the time for typing and researching necessary to complete a piece. Research time may vary. Usually, it’s 1-2 hours for 200-250 words .

Below, we’ve gathered the average writing time for average and slower writing speed:

And here are the results in pages:

How Many References Does an Essay Need?

Another essential part of any composition is the reference list. Different academic levels require different references. You’ll find out how many of them should be in your paper in the table below!

📝 Essay Examples: Different Length

Finally, we’ve gathered some excellent sample essays of different lengths. Make sure to check them out!

We also recommend you check out our free essay samples sorted by pages:

• 1-Page Essay Examples
• 2-Page Essay Examples
• 3-Page Essay Examples
• 4-Page Essay Examples
• 5-Page Essay Examples
• 10-Page Essay Examples
• 20-Page Essay Examples
• 30-Page Essay Examples
• 40-Page Essay Examples
• 50-Page Essay Examples

Now you know all about essay length, word limits, and ways to lengthen or shorten your text. If you know other interesting tricks, make sure to share them in a comment! Good luck with your writing assignments!

You may also like:

• How to Write a Process Analysis Essay: Examples & Outline
• How to Write a Precis: Definition, Guide, & Examples 
• How to Write a Critical Analysis Essay: Examples & Guide
• How to Write a Narrative Essay Outline: Template & Examples
• How to Write a Formal Essay: Format, Rules, & Example
• Word Limits and Assignment Length: Massey University
• The Paragraph in the College Essay: California State University, Long Beach
• Introductions & Conclusions: The University of Arizona Global Campus
• How Long Should a Paragraph Be?: Daily Writing Tips
• Paragraphing (Length Consistency): Purdue University
• Hitting the Target Word Count in Your College Admission Essay: Dummies.com
• How Long Should Your College Essay Be? What is the Ideal Length?: College Vine
• Writing Personal Statements Online: Issues of Length and Form: Penn State University
• Pen Admissions: Essays: University of Pennsylvania
• Essay Questions: University of Michigan
• Essay Structure: Harvard University
• Components of a Good Essay: University of Evansville
• Write Your Essay: UNSW Sydney
• College Writing: University of North Carolina at Chapel Hill
• 21 Helpful and Easy Tips to Make an Essay Longer: Seventeen
• How to Make a College Paper Longer: ThoughtCo
• Share to Facebook
• Share to Twitter
• Share to LinkedIn
• Share to email

Do you dream of being successfully admitted to college? Well, you probably know that admission is a long process that includes many challenging steps. One of them is a recommendation letter: a document written on your behalf that describes your characteristics as a student. If done well, such a letter...

Almost everyone feels insecure when they start working on their first resume. Suddenly there are lots of questions that need to be answered. What should I write? How should I write it? What should I include in my resume to make it more appealing? If you’re looking for answers to...

What’s the first thing we do when facing the unknown? We Google it, of course! Google is fundamental to our experience of the Internet. According to the statistics, more than 100 000 people press “search” on Google every second! At first glance, the process is straightforward. You type in what...

Ever heard of a smart egg tray? Welcome to 2024! Today, there are gadgets for everything, and the variety of intelligent devices available increases every day. Some of them can improve your life. Others seem a bit useless. Will a smart egg tray actually add value to your life, or...

Dorms become like a second home to many students during their university experience. And while dorm life can mean making new friends and having exciting adventures, adjusting to new routines is not always easy. Since most students who come to the dorms have no prior experience living outside their parents’...

Imagine you’ve just written a paper. You’re going to proofread it, right? But is proofreading enough to ensure the absence of any factual errors or unfounded conclusions? Can you say with 100% certainty that you cited every reference correctly? If you want to learn how to fact-check like a pro,...

Every year, millions of U.S. students decide which university they want to attend. Around 20% of students rely on college rankings to make their final decision. Over the last two decades, such rankings have significantly grown in popularity. But are these lists reliable? Which criteria can be trusted? And which...

Have you ever seen someone else’s success and thought to yourself: “Why can’t I do the same?” What was your thought process like? Perhaps you believed you lacked some innate quality. Or did you think you could have achieved the same thing if you had put in enough effort? The difference in these attitudes is defined by two distinct mindsets—growth or fixed. You will read all about them in this article by Custom-Writing.org. You’ll also...

Are you interested in having a side hustle? If you have a hobby you are passionate about, look no further. A hobby-based side hustle is a great way to spend time doing something you love, a great source of income, and a way to boost your resume! In this article by Custom-Writing.org, you will find: a list of 7 inexpensive hobbies available to anyone; ways to monetize them; their benefits for...

In 2024, online education remains one of the most searched terms. Not only can you undertake short courses, but it is now commonplace to get Bachelor’s and Master’s degrees online. Is this the right fit for you? Keep reading to find out! In this article, our custom writing team will:...

What is the number one reason for becoming a volunteer? You may already know that charitable organizations help people in need and allow everyone to make a difference. However, there are some material benefits as well. Volunteer experience can boost your career prospects: it increases your chances of getting into...

Are you excited about grammar? Have you ever read the dictionary for fun? Do you become enthralled by your textbooks? Whether you are a language nerd or not, crossword puzzles are here to help. In this article, you will find dozens of exciting puzzle resources to support your learning. You...

Confirm Password *

By registering, you agree to the Terms of Service and Privacy Policy . *

Username or email *

Forgot Password

Lost your password? Please enter your email address. You will receive a link and will create a new password via email.

Sorry, you do not have permission to ask a question, You must login to ask a question.

Please briefly explain why you feel this question should be reported.

Please briefly explain why you feel this answer should be reported.

Please briefly explain why you feel this user should be reported.

English Notes

English notes latest questions, how long should an “essay question” be.

as long as it needs to be. There is no definitive length for an essay question. You need to come up with a topic that you feel is important enough to your audience and relevant enough to your course to make the effort worth it. Essay questions can be very short or very long. They can be one sentence, one paragraph, one page, or even several pages.

There is no right or wrong length. However, if you want to make it clear that you are interested in what they think about the topic, then you need to give them something more than just a statement of fact. An essay question can be a prompt that asks students to write an answer to a specific prompt, or it can be a more open-ended prompt that asks students to write about a topic. An essay question can be as short as a sentence or as long as several paragraphs.

• Share on WhatsApp
• Share on Facebook
• Share on Twitter
• Share on LinkedIn

You must login to add an answer.

What are the six different essay lengths?

This is the second of three chapters about Essays . To complete this reader, read each chapter carefully and then unlock and complete our materials to check your understanding.   

– Discuss why essays might vary in length

– Outline the six major lengths of academic essay

– Provide defining features for each essay length

Chapter 1: What is an academic essay?

Chapter 2: What are the six different essay lengths?

Chapter 3: What are the seven different types of academic essay?

Before you begin reading...

• video and audio texts
• knowledge checks and quizzes
• skills practices, tasks and assignments

The length of essay you’re assigned will likely depend on where you are exactly in your academic course. Generally, assignments at the start of a bachelor’s degree will be shorter than those required in a student’s final years, just like master’s  and doctoral-level essays will continue to increase in both length and difficulty.

1. The One-Paragraph Essay

Generally about 150 to 250 words in length, the one-paragraph essay may be assigned by academic tutors simply in order to practise the basic concepts of paragraph writing, or it may also be used for specific purposes such as to practise summarising an article that’s been read in class or to write an extended definition of a concept. Additionally, one-paragraph essays may also be used as a diagnostic to quickly determine the level of a student’s writing. Unlike other essay lengths, for the one-paragraph essay, you’ll need to include at least some introductory, body and conclusive elements within the same paragraph.    

2. The Three-Paragraph Essay

Usually around 500 words in length, the three-paragraph essay is generally used to introduce students to the concept that all essays should maintain an introduction , body section and conclusion if the writer wishes to produce cohesive and a logical writing. The introduction and conclusion will be the first and last paragraphs and tend to be a little shorter in length, while the central body paragraph will contain the essay’s content or argument. A simple table explaining the balance of content in a three-paragraph essay has been provided below:

3. The Five-Paragraph Essay

Around 1,000 words in length, the five-paragraph essay is generally set by tutors who are content that their students understand the introduction-body-conclusion essay  structure and wish to allow more freedom to expand the ideas and arguments presented by the writer in the body section of the essay. This length of essay still only dedicates one paragraph to the introduction and conclusion , but allows three paragraphs to be dedicated to the exploration of the theme in the essay’s body. At this length, certain essay types such as cause and effect essays or compare and contrast essays may now be utilised. The following is a simple diagram of the balance of paragraph lengths in a five-paragraph essay.

4. The Extended Essay

The extended essay is the most common type of essay that’s assigned during a bachelor’s or master’s degree , and it may be of any length – although it’s unusual for such essays to be above 5,000 words. The most common lengths for an extended essay are 1,500, 3,000 and 5,000 words, with a word count allowance of plus or minus 10%. Such essay types will most certainly require research and referencing skills , and may also begin to follow more complex structures such as are found in dissertations and theses rather than simply following the introduction-body-conclusion structure of shorter essays.

5. The Dissertation

Generally assigned as the final project for both bachelor’s   and master’s degree , the typical length of an academic dissertation is 10,000 or 15,000 words. Unlike shorter essay types , dissertations have more complex structures and are almost always based around primary research (original research that the writer has conducted themselves). The following table demonstrates some of the key parts of a dissertation as well as the rough word count  percentages for each section:

6. The Thesis

Finally, the thesis is the longest academic essay type and the one that’s reserved for doctorate students studying PhDs. Generally between 40,000 and 60,000 words in length, the doctorate thesis may contain all the elements of a dissertation but in much more detail and with more careful investigation. Such essays  are almost certainly original and are based on primary research , with a larger focus on the accuracy of the literature review , data collection and data analysis . Many students will never encounter this essay type. 

Once you can recognise which essay length you’ve been assigned, the next question covered in Chapter 3 is about determining the type of essay you have to write. This is because each essay type will require particular styles, structures, foci and language.

To reference this reader:

Academic Marker (2022) Essays . Available at: https://academicmarker.com/academic-guidance/assignments/essays/ (Accessed: Date Month Year).

• Harvard Writing Center
• Leeds University Library
• Purdue Online Writing Lab

Downloadables

Once you’ve completed all three chapters about essays , you might also wish to download our beginner, intermediate and advanced worksheets to test your progress or print for your students. These professional PDF worksheets can be easily accessed for only a few Academic Marks .

Our essays academic reader (including all four chapters about this topic) can be accessed here at the click of a button.

Gain unlimited access to our essays  beginner worksheet, with activities and answer keys designed to check a basic understanding of this reader’s chapters.

To check a confident understanding of this reader’s chapters, click on the button below to download our  essays   intermediate worksheet with activities and answer keys.

Our essays  advanced worksheet with activities and answer keys has been created to check a sophisticated understanding of this reader’s chapters. 

To save yourself 5 Marks , click on the button below to gain unlimited access to all of our essays chapters and worksheets. The All-in-1 Pack includes every chapter in this reader, as well as our beginner, intermediate and advanced worksheets in one handy PDF.

Click on the button below to gain unlimited access to our essays  teacher’s PowerPoint, which should include everything you’d need to successfully introduce this topic.

Collect Academic Marks

• 100 Marks for joining
• 25 Marks for daily e-learning
• 100-200 for feedback/testimonials
• 100-500 for referring your colleages/friends

ORIGINAL RESEARCH article

Is a long essay always a good essay the effect of text length on writing assessment.

• 1 Department of Educational Research and Educational Psychology, Leibniz Institute for Science and Mathematics Education, Kiel, Germany
• 2 Institute for Psychology of Learning and Instruction, Kiel University, Kiel, Germany
• 3 School of Education, Institute of Secondary Education, University of Applied Sciences and Arts Northwestern Switzerland, Brugg, Switzerland

The assessment of text quality is a transdisciplinary issue concerning the research areas of educational assessment, language technology, and classroom instruction. Text length has been found to strongly influence human judgment of text quality. The question of whether text length is a construct-relevant aspect of writing competence or a source of judgment bias has been discussed controversially. This paper used both a correlational and an experimental approach to investigate this question. Secondary analyses were performed on a large-scale dataset with highly trained raters, showing an effect of text length beyond language proficiency. Furthermore, an experimental study found that pre-service teachers tended to undervalue text length when compared to professional ratings. The findings are discussed with respect to the role of training and context in writing assessment.

Introduction

Judgments of students’ writing are influenced by a variety of text characteristics, including text length. The relationship between such (superficial) aspects of written responses and the assessment of text quality has been a controversial issue in different areas of educational research. Both in the area of educational measurement and of language technology, text length has been shown to strongly influence text ratings by trained human raters as well as computer algorithms used to score texts automatically ( Chodorow and Burstein, 2004 ; Powers, 2005 ; Kobrin et al., 2011 ; Guo et al., 2013 ). In the context of classroom language learning and instruction, studies have found effects of text length on teachers’ diagnostic judgments (e.g., grades; Marshall, 1967 ; Osnes, 1995 ; Birkel and Birkel, 2002 ; Pohlmann-Rother et al., 2016 ). In all these contexts, the underlying question is a similar one: Should text length be considered when judging students’ writing – or is it a source of judgment bias? The objective of this paper is to investigate to what degree text length is a construct-relevant aspect of writing competence, or to what extent it erroneously influences judgments.

Powers (2005) recommends both correlational and experimental approaches for establishing the relevance of response length in the evaluation of written responses: “the former for ruling out response length (and various other factors) as causes of response quality (by virtue of their lack of relationship) and the latter for establishing more definitive causal links” (p. 7). This paper draws on data from both recommended approaches: A correlational analysis of a large-scale dataset [MEWS; funded by the German Research Foundation (Grant Nr. CO 1513/12-1) and the Swiss National Science Foundation (Grant Nr. 100019L_162675)] based on expert text quality ratings on the one hand, and an experimental study with untrained pre-service teachers on the other. It thereby incorporates the measurement perspective with the classroom perspective. In the past, (language) assessment research has been conducted within different disciplines that rarely acknowledged each other. While some assessment issues are relevant for standardized testing in large-scale contexts only, others pertain to research on teaching and classroom instruction as well. Even though their assessments may serve different functions (e.g., formative vs. summative or low vs. high stakes), teachers need to be able to assess students’ performance accurately, just as well as professional raters in standardized texts. Thus, combining these different disciplinary angles and looking at the issue of text length from a transdisciplinary perspective can be an advantage for all the disciplines involved. Overall, this paper aims to present a comprehensive picture of the role of essay length in human and automated essay scoring, which ultimately amounts to a discussion of the elusive “gold standard” in writing assessment.

Theoretical Background

Writing assessment is about identifying and evaluating features of a written response that indicate writing quality. Overall, previous research has demonstrated clear and consistent associations between linguistic features on the one hand, and writing quality and development on the other. In a recent literature review, Crossley (2020) showed that higher rated essays typically include more sophisticated lexical items, more complex syntactic features, and greater cohesion. Developing writers also show movements toward using more sophisticated words and more complex syntactic structures. The studies presented by Crossley (2020) provide strong indications that linguistic features in texts can afford important insights into writing quality and development. Whereas linguistic features are generally considered to be construct-relevant when it comes to assessing writing quality, there are other textual features whose relevance to the construct is debatable. The validity of the assessment of students’ competences is negatively affected by construct-irrelevant factors that influence judgments ( Rezaei and Lovorn, 2010 ). This holds true for professional raters in the context of large-scale standardized writing assessment as well as for teacher judgments in classroom writing assessment (both formative or summative). Assigning scores to students’ written responses is a challenging task as different text-inherent factors influence the accuracy of the raters’ or teachers’ judgments (e.g., handwriting, spelling: Graham et al., 2011 ; length, lexical diversity: Wolfe et al., 2016 ). Depending on the construct to be assessed, the influence of these aspects can be considered judgment bias. One of the most relevant and well-researched text-inherent factors influencing human judgments is text length. Crossley (2020) points out that his review does “not consider text length as a linguistic feature while acknowledging that text length is likely the strongest predictor of writing development and quality.” Multiple studies have found a positive relationship between text length and human ratings of text quality, even when controlling for language proficiency ( Chenoweth and Hayes, 2001 ; McCutchen et al., 2008 ; McNamara et al., 2015 ). It is still unclear, however, whether the relation between text length and human scores reflects a true relation between text length and text quality (appropriate heuristic assumption) or whether it stems from a bias in human judgments (judgment bias assumption). The former suggests that text length is a construct-relevant factor and that a certain length is needed to effectively develop a point of view on the issue presented in the essay prompt, and this is one of the aspects taken into account in the scoring ( Kobrin et al., 2007 ; Quinlan et al., 2009 ). The latter claims that text length is either completely or partly irrelevant to the construct of writing proficiency and that the strong effect it has on human judgment can be considered a bias ( Powers, 2005 ). In the context of large-scale writing assessment, prompt-based essay tasks are often used to measure students’ writing competence ( Guo et al., 2013 ). These essays are typically scored by professionally trained raters. These human ratings have been shown to be strongly correlated with essay length, even if this criterion is not represented in the assessment rubric ( Chodorow and Burstein, 2004 ; Kobrin et al., 2011 ). In a review of selected studies addressing the relation between length and quality of constructed responses, Powers (2005) showed that most studies found correlations within the range of r = 0.50 to r = 0.70. For example, he criticized the SAT essay for encouraging wordiness as longer essays tend to score higher. Kobrin et al. (2007) found the number of words to explain 39% of the variance in the SAT essay score. The authors argue that essay length is one of the aspects taken into account in the scoring as it takes a certain length to develop an argument. Similarly, Deane (2013) argues in favor of regarding writing fluency a construct-relevant factor (also see Shermis, 2014 ; McNamara et al., 2015 ). In an analytical rating of text quality, Hachmeister (2019) could showed that longer texts typically contain more cohesive devices, which has a positive impact on ratings of text quality. In the context of writing assessment in primary school, Pohlmann-Rother et al. (2016) found strong correlations between text length and holistic ratings of text quality ( r = 0.62) as well as the semantic-pragmatic analytical dimension ( r = 0.62). However, they found no meaningful relationship between text length and language mechanics (i.e., grammatical and orthographical correctness; r = 0.09).

Text length may be considered especially construct-relevant when it comes to writing in a foreign language. Because of the constraints of limited language knowledge, writing in a foreign language may be hampered because of the need to focus on language rather than content ( Weigle, 2003 ). Silva (1993) , in a review of differences between writing in a first and second language, found that writing in a second language tends to be “more constrained, more difficult, and less effective” (p. 668) than writing in a first language. The necessity of devoting cognitive resources to issues of language may mean that not as much attention can be given to higher order issues such as content or organization (for details of this debate, see Weigle, 2003 , p. 36 f.). In that context, the ability of writing longer texts may be legitimately considered as indicative of higher competence in a foreign language, making text length a viable factor of assessment. For example, Ruegg and Sugiyama (2010) showed that the main predictors of the content score in English foreign language essays were first, organization and second, essay length.

The relevance of this issue has further increased as systems of automated essay scoring (AES) have become more widely used in writing assessment. These systems offer a promising way to complement human ratings in judging text quality ( Deane, 2013 ). However, as the automated scoring algorithms are typically modeled after human ratings, they are also affected by human judgment bias. Moreover, it has been criticized that, at this point, automated scoring systems mainly count words when computing writing scores ( Perelman, 2014 ). Chodorow and Burstein (2004) , for example, showed that 53% of the variance in human ratings can be explained by automated scoring models that use only the number of words and the number of words squared as predictors. Ben-Simon and Bennett (2007) provided evidence from National Assessment of Educational Progress (NAEP) writing test data that standard, statistically created e-rater models weighed essay length even more strongly than human raters (also see Perelman, 2014 ).

Bejar (2011) suggests that a possible tendency to reward longer texts could be minimized through the training of raters with responses at each score level that vary in length. However, Barkaoui (2010) and Attali (2016) both compared the holistic scoring of experienced vs. novice raters and – contrary to expectations – found that the correlation between essay length and scores was slightly stronger for the experienced group. Thus, the question of whether professional experience and training counteract or even reinforce the tendency to overvalue text length in scoring remains open.

Compared to the amount of research on the role of essay length in human and automated scoring in large-scale high-stakes contexts, little attention has been paid to the relation of text length and quality in formative or summative assessment by teachers. This is surprising considering the relevance of the issue for teachers’ professional competence: In order to assess the quality of students’ writing, teachers must either configure various aspects of text quality in a holistic assessment or hold them apart in an analytic assessment. Thus, they need to have a concept of writing quality appropriate for the task and they need to be aware of the construct-relevant and -irrelevant criteria (cf. the lens model; Brunswik, 1955 ). To our knowledge, only two studies have investigated the effect of text length on holistic teacher judgments, both of which found that longer texts receive higher grades. Birkel and Birkel (2002) found significant main effects of text length (long, medium, short) and spelling errors (many, few) on holistic teacher judgments. Osnes (1995) reported effects of handwriting quality and text length on grades.

Whereas research on the text length effect on classroom writing assessment is scarce, a considerable body of research has investigated how other text characteristics influence teachers’ assessment of student texts. It is well-demonstrated, for example, that pre-service and experienced teachers assign lower grades to essays containing mechanical errors ( Scannell and Marshall, 1966 ; Marshall, 1967 ; Cumming et al., 2002 ; Rezaei and Lovorn, 2010 ). Scannell and Marshall (1966) found that pre-service teachers’ judgments were affected by errors in punctuation, grammar and spelling, even though they were explicitly instructed to grade on content alone. More recently, Rezaei and Lovorn (2010) showed that high quality essays containing more structural, mechanical, spelling, and grammatical errors were assigned lower scores than texts without errors even in criteria relating solely to content. Teachers failed to distinguish between formal errors and the independent quality of content in a student essay. Similarly, Vögelin et al. (2018 , 2019) found that lexical features and spelling influenced not only holistic teacher judgments of students’ writing in English as a second or foreign language, but also their assessment of other analytical criteria (e.g., grammar). Even though these studies do not consider text length as a potential source of bias, they do show that construct-irrelevant aspects influence judgments of teachers.

This Research

Against this research background, it remains essential to investigate whether the relation between essay length and text quality represents a true relationship or a bias on the part of the rater or teacher ( Wolfe et al., 2016 ). First, findings of correlational studies can give us an indication of the effect of text length on human ratings above and beyond language proficiency variables. Second, going beyond correlational findings, there is a need for experimental research that examines essay responses on the same topic differing only in length in order to establish causal relationships ( Kobrin et al., 2007 ). The present research brings together both of these approaches.

This paper comprises two studies investigating the role of essay length in foreign language assessment using an interdisciplinary perspective including the fields of foreign language education, computer linguistics, educational research, and psychometrics. Study 1 presents a secondary analysis of a large-scale dataset with N = 2,722 upper secondary school students in Germany and Switzerland who wrote essays in response to “independent writing” prompts of the internet-based Test of English as a Foreign Language (TOEFL iBT). It investigates the question of how several indicators of students’ English proficiency (English grade, reading and listening comprehension, self-concept) are related to the length of their essays (word count). It further investigates whether or not essay length accounts for variance in text quality scores (expert ratings) even when controlling for English language proficiency and other variables (e.g., country, gender, cognitive ability). A weak relationship of proficiency and length as well as a large proportion of variance in text quality explained by length beyond proficiency would be in favor of the judgment bias assumption.

Study 2 focused on possible essay length bias in an experimental setting, investigating the effect of essay length on text quality ratings when there was (per design) no relation between essay length and text quality score. Essays from Study 1 were rated by N = 84 untrained pre-service teachers, using the same TOEFL iBT rubric as the expert raters. As text quality scores were held constant within all essay length conditions, any significant effect of essay length would indicate a judgment bias. Both studies are described in more detail in the following sections.

This study investigates the question of judgment bias assumption vs. appropriate heuristic assumption in a large-scale context with professional human raters. A weak relationship between text length and language proficiency would be indicative of the former assumption, whereas a strong relationship would support the latter. Moreover, if the impact of text length on human ratings was significant and substantial beyond language proficiency, this might indicate a bias on the part of the rater rather than an appropriate heuristic. Thus, Study 1 aims to answer the following research questions:

(1) How is essay length related to language proficiency?

(2) Does text length still account for variance in text quality when English language proficiency is statistically controlled for?

Materials and Methods

Sample and procedure.

The sample consisted of N = 2,722 upper secondary students (11th grade; 58.1% female) in Germany ( n = 894) and Switzerland ( n = 1828) from the interdisciplinary and international research project Measuring English Writing at Secondary Level (MEWS; for an overview see Keller et al., 2020 ). The target population were students attending the academic track of general education grammar schools (ISCED level 3a) in the German federal state Schleswig-Holstein as well as in seven Swiss cantons (Aargau, Basel Stadt, Basel Land, Luzern, St. Gallen, Schwyz, Zurich). In a repeated-measures design, students were assessed at the beginning (T1: August/September 2016; M age = 17.34; SD age = 0.87) and at the end of the school year (T2: May/June 2017; M age = 18.04; SD age = 0.87). The students completed computer-based tests on writing, reading and listening skills, as well as general cognitive ability. Furthermore, they completed a questionnaire measuring background variables and individual characteristics.

Writing prompt

All students answered two independent and two integrated essay writing prompts of the internet-based Test of English as a Foreign Language (TOEFL iBT ® ) that is administered by the Educational Testing Service (ETS) in Princeton. The task instruction was as follows: “In the writing task below you will find a question on a controversial topic. Answer the question in an essay in English. List arguments and counter-arguments, explain them and finally make it clear what your own opinion on the topic is. Your text will be judged on different qualities. These include the presentation of your ideas, the organization of the essay and the linguistic quality and accuracy. You have 30 min to do this. Try to use all of this time as much as possible.” This task instruction was followed by the essay prompt. The maximum writing time was 30 min according to the official TOEFL iBT ® assessment procedure. The essays were scored by trained human raters on the TOEFL 6-point rating scale at ETS. In addition to two human ratings per essay, ETS also provided scores from their automated essay scoring system (e-rater ® ; Burstein et al., 2013 ). For a more detailed description of the scoring procedure and the writing prompts see Rupp et al. (2019) and Keller et al. (2020) . For the purpose of this study, we selected the student responses to the TOEFL iBT independent writing prompt “Teachers,” which showed good measurement qualities (see Rupp et al., 2019 ). Taken together, data collections at T1 and T2 yielded N = 2,389 valid written responses to the following prompt: “A teacher’s ability to relate well with students is more important than excellent knowledge of the subject being taught.”

Text quality and length

The rating of text quality via human and machine scoring was done by ETS. All essays were scored by highly experienced human raters on the operational holistic TOEFL iBT rubric from 0 to 5 ( Chodorow and Burstein, 2004 ). Essays were scored high if they were well-organized and individual ideas were well-developed, if they used specific examples and support to express learners’ opinion on the subject, and if the English language was used accurately to express learners’ ideas. Essays were assigned a score of 0 if they were written in another language, were generally incomprehensible, or if no text was entered.

Each essay received independent ratings by two trained human raters. If the two ratings showed a deviation of 1, the mean of the two scores was used; if they showed a deviation of 2 or more, a third rater (adjudicator) was consulted. Inter-rater agreement, as measured by quadratic weighted kappa (QWK), was satisfying for the prompt “Teachers” at both time points (QWK = 0.67; Hayes and Hatch, 1999 ; see Rupp et al., 2019 for further details). The mean text quality score was M = 3.35 ( SD = 0.72).

Word count was used to measure the length of the essays. The number of words was calculated by the e-Rater scoring engine. The mean word count was M = 311.19 ( SD = 81.91) and the number of words ranged from 41 to 727. We used the number of words rather than other measures of text length (e.g., number of letters) as it is the measure which is most frequently used in the literature: 9 out of 10 studies in the research review by Powers (2005) used word count as the criterion (also see Kobrin et al., 2007 , 2011 ; Crossley and McNamara, 2009 ; Barkaoui, 2010 ; Attali, 2016 ; Wolfe et al., 2016 ; Wind et al., 2017 ). This approach ensures that our analyses can be compared with previous research.

English language proficiency and control variables

Proficiency was operationalized by a combination of different variables: English grade, English writing self-concept, reading and listening comprehension in English. The listening and reading skills were measured with a subset of items from the German National Assessment ( Köller et al., 2010 ). The tasks require a detailed understanding of long, complex reading and listening texts including idiomatic expressions and different linguistic registers. The tests consisted of a total of 133 items for reading, and 118 items for listening that were administered in a multi-matrix-design. Each student was assessed with two rotated 15-min blocks per domain. Item parameters were estimated using longitudinal multidimensional two-parameter item response models in M plus version 8 ( Muthén and Muthén, 1998–2012 ). Student abilities were estimated using 15 plausible values (PVs) per person. The PV reliabilities were 0.92 (T1) and 0.76 (T2) for reading comprehension, and 0.85 (T1) and 0.72 (T2) for listening comprehension. For a more detailed description of the scaling procedure see Köller et al. (2019) .

General cognitive ability was assessed at T1 using the subtests on figural reasoning (N2; 25 items) and on verbal reasoning (V3; 20 items) of the Cognitive Ability Test (KFT 4–12 + R; Heller and Perleth, 2000 ). For each scale 15 PVs were drawn in a two-dimensional item response model. For the purpose of this study, the two PVs were combined to 15 overall PV scores with a reliability of 0.86.

The English writing self-concept was measured with a scale consisting of five items (e.g., “I have always been good at writing in English”; Eccles and Wigfield, 2002 ; Trautwein et al., 2012 ; α = 0.90). Furthermore, country (Germany = 0/Switzerland = 1), gender (male = 0/female = 1) and time of measurement (T1 = 0; T2 = 1) were used as control variables.

Statistical Analyses

All analyses were conducted in M plus version 8 ( Muthén and Muthén, 1998–2012 ) based on the 15PV data sets using robust maximum likelihood estimation to account for a hierarchical data structure (i.e., students clustered in classes; type = complex). Full-information maximum likelihood was used to estimate missing values in background variables. Due to the use of 15PVs, all analyses were run 15 times and then averaged (see Rubin, 1987 ).

Confirmatory factor analysis was used to specify a latent proficiency factor. All four proficiency variables showed substantial loadings in a single-factor measurement model (English grade: 0.67; writing self-concept: 0.73; reading comprehension: 0.42; listening comprehension: 0.51). As reading and listening comprehension were measured within the same assessment framework and could thus be expected to share mutual variance beyond the latent factor, their residuals were allowed to correlate. The analyses yielded an acceptable model fit: χ 2 (1) = 3.65, p = 0.06; CFI = 0.998, RMSEA = 0.031, SRMR = 0.006.

The relationship between text length and other independent variables was explored with correlational analysis. Multiple regression analysis with latent and manifest predictors was used to investigate the relations between text length, proficiency, and text quality.

The correlation of the latent proficiency factor and text length (word count) was moderately positive: r = 0.36, p < 0.01. This indicates that more proficient students tended to write longer texts. Significant correlations with other variables showed that students tended to write longer texts at T1 ( r = -0.08, p < 0.01), girls wrote longer texts than boys ( r = 0.11, p < 0.01), and higher cognitive ability was associated with longer texts ( r = 0.07, p < 0.01). However, all of these correlations were very weak as a general rule. The association of country and text length was not statistically significant ( r = -0.06, p = 0.10).

Table 1 presents the results of the multiple linear regression of text quality on text length, proficiency and control variables. The analysis showed that proficiency and the covariates alone explained 38 percent of the variance in text quality ratings, with the latent proficiency factor being by far the strongest predictor (Model 1). The effect of text length on the text quality score was equally strong when including the control variables but not proficiency in the model (Model 2). When both the latent proficiency factor and text length were entered into the regression model (Model 3), the coefficient of text length was reduced but remained significant and substantial, explaining an additional 24% of the variance (ΔR 2 = 0.24 from Model 1 to Model 3). Thus, text length had an incremental effect on text quality beyond a latent English language proficiency factor.

Table 1. Linear regression of text quality on text length, English language proficiency, and control variables: standardized regression coefficients (β) and standard errors (SE).

Study 1 approached the issue of text length by operationalizing the construct of English language proficiency and investigating how it affects the relationship of text length and text quality. This can give us an idea of how text length may influence human judgments even though it is not considered relevant to the construct of writing competence. These secondary analyses of an existing large-scale dataset yielded two central findings: First, text length was only moderately associated with language proficiency. Second, text length strongly influenced writing performance beyond proficiency. Thus, it had an impact on the assigned score that was not captured by the construct of proficiency. These findings could be interpreted in favor of the judgment bias assumption as text length may include both construct-irrelevant and construct-relevant information.

The strengths of this study were the large sample of essays on the same topic and the vast amount of background information that was collected on the student writers (proficiency and control variables). However, there were three major limitations: First, the proficiency construct captured different aspects of English language competence (reading and listening comprehension, writing self-concept, grade), but that operationalization was not comprehensive. Thus, the additional variance explained by text length may still have been due to other aspects that could not be included in the analyses as they were not in the data. Further research with a similar design (primary or secondary analyses) should use additional variables such as grammar/vocabulary knowledge or writing performance in the first language.

The second limitation was the correlational design, which does not allow a causal investigation of the effect of text length on text quality ratings. Drawing inferences which are causal in nature would require an experimental environment in which, for example, text quality is kept constant for texts of different lengths. For that reason, Study 2 was conducted exactly in such a research design.

Last but not least, the question of transferability of these findings remains open. Going beyond standardized large-scale assessment, interdisciplinary research requires us to look at the issue from different perspectives. Findings pertaining to professional raters may not be transferable to teachers, who are required to assess students’ writing in a classroom context. Thus, Study 2 drew on a sample of preservice English teachers and took a closer look at how their ratings were impacted by text length.

Research Questions

In Study 2, we investigated the judgment bias assumption vs. the appropriate heuristic assumption of preservice teachers. As recommended by Powers (2005) , we conducted an experimental study in addition to the correlational design used in Study 1. As text quality scores were held constant within all essay length conditions, any significant effect of essay length would be in favor of the judgment bias assumption. The objective of this study was to answer the following research questions:

(1) How do ratings of pre-service teachers correspond to expert ratings?

(2) Is there an effect of text length on the text quality ratings of preservice English teachers, when there is (per design) no relation between text length and text quality (main effect)?

(3) Does the effect differ for different levels of writing performance (interaction effect)?

Participants and Procedure

The experiment was conducted with N = 84 pre-service teachers ( M Age = 23 years; 80% female), currently enrolled in a higher education teacher training program at a university in Northern Germany. They had no prior rating experience of this type of learner texts. The experiment was administered with the Student Inventory ASSET ( Jansen et al., 2019 ), an online tool to assess students’ texts within an experimental environment. Participants were asked to rate essays from the MEWS project (see Study 1) on the holistic rubric used by the human raters at ETS (0–5; https://www.ets.org/s/toefl/pdf/toefl_writing_rubrics.pdf ). Every participant had to rate 9 out of 45 essays in randomized order, representing all possible combinations of text quality and text length. Before the rating process began, participants were given information about essay writing in the context of the MEWS study (school type; school year; students’ average age; instructional text) and they were presented the TOEFL writing rubric as the basis for their judgments. They had 15 min to get an overview of all nine texts before they were asked to rate each text on the rubric. Throughout the rating process, they were allowed to highlight parts of the texts.

The operationalization of text quality and text length as categorical variables as well as the procedure of selecting an appropriate essay sample for the study is explained in the following.

Text Length and Text Quality

The essays used in the experiment were selected on the basis of the following procedure, which took both text quality and text length as independent variables into account. The first independent variable of the essay (overall text quality) was operationalized via scores assigned by two trained human raters from ETS on a holistic six-point scale (0–5; see Study 1 and Appendix A). In order to measure the variable as precisely as possible, we only included essays for which both human raters had assigned the same score, resulting in a sample of N = 1,333 essays. As a result, three gradations of text quality were considered in the current study: lower quality (score 2), medium quality (score 3) and higher quality (score 4). The corpus included only few texts (10.4%) with the extreme scores of 0, 1, and 5; these were therefore excluded from the essay pool. We thus realized a 3 × 3 factorial within-subjects design. The second independent variable text length was measured via the word count of the essays, calculated by the e-rater (c) scoring engine. As with text quality, this variable was subdivided in three levels: rather short texts (s), medium-length texts (m), and long texts (l). All available texts were analyzed regarding their word count distribution. Severe outliers were excluded. The remaining N = 1308 essays were split in three even groups: the lower (=261 words), middle (262–318 words) and upper third (=319 words). Table 2 shows the distribution of essays for the resulting combinations of text length and text score.

Table 2. Distribution of essays in the sample contingent on text quality and text length groupings.

Selection of Essays

For each text length group (s, m, and l), the mean word count across all three score groups was calculated. Then, the score group (2, 3, or 4) with the smallest number of essays in a text length group was taken as reference (e.g., n = 22 short texts of high quality or n = 15 long texts of low quality). Within each text length group, the five essays being – word count-wise – closest to the mean of the reference were chosen for the study. This was possible with mostly no or only minor deviations. In case of multiple possible matches, the essay was selected at random. This selection procedure resulted in a total sample of 45 essays, with five essays for each combination of score group (2, 3, 4) and length group (s, m, l).

A repeated-measures ANOVA with two independent variables (text quality and text length) was conducted to test the two main effects and their interaction on participants’ ratings (see Table 3 ). Essay ratings were treated as a within-subject factor, accounting for dependencies of the ratings nested within raters. The main effect of text quality scores on participants’ ratings showed significant differences between the three text quality conditions ( low , medium , high ) that corresponded to expert ratings; F (2, 82) = 209.04, p < 0.001, d = 4.52. There was also a significant main effect for the three essay length conditions ( short , medium , long ); F (2, 82) = 9.14, p < 0.001, d = 0.94. Contrary to expectations, essay length was negatively related to participants’ ratings, meaning that shorter texts received higher scores than longer texts. The interaction of text quality and text length also had a significant effect; F (4, 80) = 3.93, p < 0.01, d = 0.89. Post-hoc tests revealed that texts of low quality were especially impacted by essay length in a negative way (see Figure 1 ).

Table 3. Participants’ ratings of text quality: means (M) and standard deviations (SD).

Figure 1. Visualization of the interaction between text length and text quality.

The experiment conducted in Study 2 found a very strong significant main effect for text quality, indicating a high correspondence of pre-service teachers’ ratings with the expert ratings of text quality. The main effect of text length was also significant, but was qualified by a significant interaction effect text quality x text length, indicating that low quality texts were rated even more negative the longer they were. This negative effect of text length was contrary to expectations: The pre-service teachers generally tended to assign higher scores to shorter texts. Thus, they seemed to value shorter texts over longer texts. However, this was mainly true for texts of low quality.

These findings were surprising against the research background that would suggest that longer texts are typically associated with higher scores of text quality, particularly in the context of second language writing. Therefore, it is even more important to discuss the limitations of the design before interpreting the results: First, the sample included relatively inexperienced pre-service teachers. Further research is needed to show whether these findings are transferable to in-service teachers with reasonable experience in judging students’ writing. Moreover, further studies could use assessment rubrics that teachers are more familiar with, such as the CEFR ( Council of Europe, 2001 ; also see Fleckenstein et al., 2020 ). Second, the selection process of essays may have reduced the ecological validity of the experiment. As there were only few long texts of low quality and few short texts of high quality in the actual sample (see Table 2 ), the selection of texts in the experimental design was – to some degree – artificial. This could also have influenced the frame of reference for the pre-service teachers as the distribution of the nine texts was different from what one would find naturally in an EFL classroom. Third, the most important limitation of this study is the question of the reference norm, a point which applies to studies of writing assessment in general. In our study, writing quality was operationalized using expert ratings, which have been shown to be influenced by text length in many investigations as well as in Study 1. If the expert ratings are biased themselves, the findings of this study may also be interpreted as pre-service teachers (unlike expert raters) not showing a text length bias at all: shorter texts should receive higher scores than longer ones if the quality assigned by the expert raters is held constant. We discuss these issues concerning the reference norm in more detail in the next section.

All three limitations may have affected ratings in a way that could have reinforced a negative effect of text length on text quality ratings. However, as research on the effect of text length on teachers’ judgments is scarce, we should consider the possibility that the effect is actually different from the (positive) one typically found for professional human raters. There are a number of reasons to assume differences in the rating processes that are discussed in more detail in the following section. Furthermore, we will discuss what this means in terms of the validity of the gold standard in writing assessment.

General Discussion

Combining the results of both studies, we have reason to assume that (a) text length induces judgment bias and (b) the effect of text length largely depends on the rater and/or the rating context. More specifically, the findings of the two studies can be summarized as follows: Professional human raters tend to reward longer texts beyond the relationship of text length and proficiency. Compared to this standard, inexperienced EFL teachers tend to undervalue text length, meaning that they sanction longer texts especially when text quality is low. This in turn may be based on an implicit expectation deeply ingrained in the minds of many EFL teachers: that writing in a foreign language is primarily about avoiding mistakes, and that longer texts typically contain more of them than shorter ones ( Keller, 2016 ). Preservice teachers might be particularly afflicted with this view of writing as they would have experienced it as learners up-close and personal, not too long ago. Both findings point toward the judgment bias assumption, but with opposite directions. These seemingly contradictory findings lead to interesting and novel research questions – both in the field of standardized writing assessment and in the field of teachers’ diagnostic competence.

Only if we take professional human ratings as reliable benchmark scores can we infer that teachers’ ratings are biased (in a negative way). If we consider professional human ratings to be biased themselves (in a positive way), then the preservice teachers’ judgments might appear to be unbiased. However, it would be implausible to assume that inexperienced teachers’ judgments are less biased than those of highly trained expert raters. Even if professional human ratings are flawed themselves, they are the best possible measure of writing quality, serving as a reference even for NLP tools ( Crossley, 2020 ). It thus makes much more sense to consider the positive impact of text length on professional human ratings – at least to a degree – an appropriate heuristic. This means that teachers’ judgments would generally benefit from applying the same heuristic when assessing students’ writing, as long as it does not become a bias.

In his literature review, Crossley (2020) sees the nature of the writing task to be among the central limitations when it comes to generalizing findings in the context of writing assessment. Written responses to standardized tests (such as the TOEFL) may produce linguistic features that differ from writing samples produced in the classroom or in other, more authentic writing environments. Moreover, linguistic differences may also occur depending on a writing sample being timed or untimed. Timed samples provide fewer opportunities for planning, revising, and development of ideas as compared to untimed samples, where students are more likely to plan, reflect, and revise their writing. These differences may surface in timed writing in such a way that it would be less cohesive and less complex both lexically and syntactically.

In the present research, such differences may account for the finding that pre-service teachers undervalue text length compared to professional raters. Even though the participants in Study 2 were informed about the context in which the writing samples were collected, they may have underestimated the challenges of a timed writing task in an unfamiliar format. In the context of their own classrooms, students rarely have strict time limitations when working on complex writing tasks. If they do, in an exam consisting of an argumentative essay, for example, it is usually closer to 90 min than to 30 min (at least in the case of the German pre-service teachers who participated in this study). Thus, text length may not be a good indicator of writing quality in the classroom. On the contrary, professional raters may value length as a construct-relevant feature of writing quality in a timed task, for example as an indicator of writing fluency (see Peng et al., 2020 ).

Furthermore, text length as a criterion of quality cannot be generalized over different text types at random. The genres which are taught in EFL courses, or assessed in EFL exams, differ considerably with respect to expected length. In five paragraph essays, for example, developing an argument requires a certain scope and attention to detail, so that text length is a highly salient feature for overall text quality. The same might not be true for e-mail writing, a genre frequently taught in EFL classrooms ( Fleckenstein et al., in press ). E-mails are usually expected to be concise and to the point, so that longer texts might seem prolix, or rambling. Such task-specific demands need to be taken into account when it comes to interpreting our findings. The professional raters employed in our study were schooled extensively for rating five-paragraph essays, which included a keen appreciation of text length as a salient criterion of text quality. The same might not be said of classroom teachers, who encounter a much wider range of genres in their everyday teaching and might therefore be less inclined to consider text length as a relevant feature. Further research should consider different writing tasks in order to investigate whether text length is particularly important to the genre of the argumentative essay.

Our results underscore the importance of considering whether or not text length should be taken into account for different contexts of writing assessment. This holds true for classroom assessment, where teachers should make their expectations regarding text length explicit, as well as future studies with professional raters. Crossley (2020) draws attention to the transdisciplinary perspective of the field as a source for complications: “The complications arise from the interdisciplinary nature of this type of research which often combines writing, linguistics, statistics, and computer science fields. With so many fields involved, it is often easy to overlook confounding factors” (p. 428). The present research shows how the answer to one and the same research question – How does text length influence human judgment? – can be very different from different perspectives and within different areas of educational research. Depending on the population (professional raters vs. pre-service teachers) and the methodology (correlational analysis vs. experimental design), our findings illustrate a broad range of possible investigations and outcomes. Thus, it is a paramount example of why interdisciplinary research in education is not only desirable but imperative. Without an interdisciplinary approach, our view of the text length effect would be uni-dimensional and fragmentary. Only the combination of different perspectives and methods can live up to the demands of a complex issue such as writing assessment, identify research gaps, and challenge research traditions. Further research is needed to investigate the determinants of the strength and the direction of the bias. It is necessary to take a closer look at the rating processes of (untrained) teachers and (trained) raters, respectively, in order to investigate similarities and differences. Research pertaining to judgment heuristics/biases can be relevant for both teacher and rater training. However, the individual concerns and characteristics of the two groups need to be taken into account. This could be done, for example, by directly comparing the two groups in an experimental study. Both in teacher education and in text assessment studies, we should have a vigorous discussion about how appropriate heuristics of expert raters can find their way into the training of novice teachers and inexperienced raters in an effort to reduce judgement bias.

Data Availability Statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

Ethics Statement

The studies involving human participants were reviewed and approved by the Ministry of Education, Science and Cultural Affairs of the German federal state Schleswig-Holstein. Written informed consent to participate in this study was provided by the participants’ legal guardian/next of kin.

Author Contributions

JF analyzed the data and wrote the manuscript. TJ and JM collected the experimental data for Study 2 and supported the data analysis. SK and OK provided the dataset for Study 1. TJ, JM, SK, and OK provided feedback on the manuscript. All authors contributed to the article and approved the submitted version.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Attali, Y. (2016). A comparison of newly-trained and experienced raters on a standardized writing assessment. Lang. Test. 33, 99–115. doi: 10.1177/0265532215582283

CrossRef Full Text | Google Scholar

Barkaoui, K. (2010). Explaining ESL essay holistic scores: a multilevel modeling approach. Lang. Test. 27, 515–535. doi: 10.1177/0265532210368717

Bejar, I. I. (2011). A validity-based approach to quality control and assurance of automated scoring. Assess. Educ. 18, 319–341. doi: 10.1080/0969594x.2011.555329

Ben-Simon, A., and Bennett, R. E. (2007). Toward more substantively meaningful automated essay scoring. J. Technol. Learn. Asses. 6, [Epub ahead of print].

Google Scholar

Birkel, P., and Birkel, C. (2002). Wie einig sind sich Lehrer bei der Aufsatzbeurteilung? Eine Replikationsstudie zur Untersuchung von Rudolf Weiss. Psychol. Erzieh. Unterr. 49, 219–224.

Brunswik, E. (1955). Representative design and probabilistic theory in a functional psychology. Psychol. Rev. 62, 193–217. doi: 10.1037/h0047470

PubMed Abstract | CrossRef Full Text | Google Scholar

Burstein, J., Tetreault, J., and Madnani, N. (2013). “The E-rater ® automated essay scoring system,” in Handbook of Automated Essay Evaluation , eds M. D. Shermis and J. Burstein (Abingdon: Routledge), 77–89.

Chenoweth, N. A., and Hayes, J. R. (2001). Fluency in writing: generating text in L1 and L2. Written Commun. 18, 80–98. doi: 10.1177/0741088301018001004

Chodorow, M., and Burstein, J. (2004). Beyond essay length: evaluating e-rater ® ’s performance on toefl ® essays. ETS Res. Rep. 2004, i–38. doi: 10.1002/j.2333-8504.2004.tb01931.x

Council of Europe (2001). Common European Framework of Reference for Languages: Learning, Teaching and Assessment. Cambridge, MA: Cambridge University Press.

Crossley, S. (2020). Linguistic features in writing quality and development: an overview. J. Writ. Res. 11, 415–443. doi: 10.17239/jowr-2020.11.03.01

Crossley, S. A., and McNamara, D. S. (2009). Computational assessment of lexical differences in L1 and L2 writing. J. Second. Lang. Writ. 18, 119–135. doi: 10.1016/j.jslw.2009.02.002

Cumming, A., Kantor, R., and Powers, D. E. (2002). Decision making while rating ESL/EFL writing tasks: a descriptive framework. Modern Lang. J. 86, 67–96. doi: 10.1111/1540-4781.00137

Deane, P. (2013). On the relation between automated essay scoring and modern views of the writing construct. Assess. Writ. 18, 7–24. doi: 10.1016/j.asw.2012.10.002

Eccles, J. S., and Wigfield, A. (2002). Motivational beliefs, values, and goals. Annu. Rev.Psychol. 53, 109–132. doi: 10.1146/annurev.psych.53.100901.135153

Fleckenstein, J., Keller, S., Krüger, M., Tannenbaum, R. J., and Köller, O. (2020). Linking TOEFL iBT ® writing scores and validity evidence from a standard setting study. Assess. Writ. 43:100420. doi: 10.1016/j.asw.2019.100420

Fleckenstein, J., Meyer, J., Jansen, T., Reble, R., Krüger, M., Raubach, E., et al. (in press). “Was macht Feedback effektiv? Computerbasierte Leistungsrückmeldung anhand eines Rubrics beim Schreiben in der Fremdsprache Englisch,” in Tagungsband Bildung, Schule und Digitalisierung , eds K. Kaspar, M. Becker-Mrotzek, S. Hofhues, J. König, and D. Schmeinck (Münster: Waxmann

Graham, S., Harris, K. R., and Hebert, M. (2011). It is more than just the message: presentation effects in scoring writing. Focus Except. Child. 44, 1–12.

Guo, L., Crossley, S. A., and McNamara, D. S. (2013). Predicting human judgments of essay quality in both integrated and independent second language writing samples: a comparison study. Assess. Writ. 18, 218–238. doi: 10.1016/j.asw.2013.05.002

Hachmeister, S. (2019). “Messung von Textqualität in Ereignisberichten,” in Schreibkompetenzen Messen, Beurteilen und Fördern (6. Aufl) , eds I. Kaplan and I. Petersen (Münster: Waxmann Verlag), 79–99.

Hayes, J. R., and Hatch, J. A. (1999). Issues in measuring reliability: Correlation versus percentage of agreement. Writt. Commun. 16, 354–367. doi: 10.1177/0741088399016003004

Heller, K. A., and Perleth, C. (2000). KFT 4-12+ R Kognitiver Fähigkeitstest für 4. Bis 12. Klassen, Revision. Göttingen: Beltz Test.

Jansen, T., Vögelin, C., Machts, N., Keller, S. D., and Möller, J. (2019). Das Schülerinventar ASSET zur Beurteilung von Schülerarbeiten im Fach Englisch: Drei experimentelle Studien zu Effekten der Textqualität und der Schülernamen. Psychologie in Erziehung und Unterricht 66, 303–315. doi: 10.2378/peu2019.art21d

Keller, S. (2016). Measuring Writing at Secondary Level (MEWS). Eine binationale Studie. Babylonia 3, 46–48.

Keller, S. D., Fleckenstein, J., Krüger, M., Köller, O., and Rupp, A. A. (2020). English writing skills of students in upper secondary education: results from an empirical study in Switzerland and Germany. J. Second Lang. Writ. 48:100700. doi: 10.1016/j.jslw.2019.100700

Kobrin, J. L., Deng, H., and Shaw, E. J. (2007). Does quantity equal quality? the relationship between length of response and scores on the SAT essay. J. Appl. Test. Technol. 8, 1–15. doi: 10.1097/nne.0b013e318276dee0

Kobrin, J. L., Deng, H., and Shaw, E. J. (2011). The association between SAT prompt characteristics, response features, and essay scores. Assess. Writ. 16, 154–169. doi: 10.1016/j.asw.2011.01.001

Köller, O., Fleckenstein, J., Meyer, J., Paeske, A. L., Krüger, M., Rupp, A. A., et al. (2019). Schreibkompetenzen im Fach Englisch in der gymnasialen Oberstufe. Z. Erziehungswiss. 22, 1281–1312. doi: 10.1007/s11618-019-00910-3

Köller, O. Knigge, M. and Tesch B. (eds.) (2010). Sprachliche Kompetenzen im Ländervergleich. Germany: Waxmann.

Marshall, J. C. (1967). Composition errors and essay examination grades re-examined. Am. Educ. Res. J. 4, 375–385. doi: 10.3102/00028312004004375

McCutchen, D., Teske, P., and Bankston, C. (2008). “Writing and cognition: implications of the cognitive architecture for learning to write and writing to learn,” in Handbook of research on Writing: History, Society, School, Individual, Text , ed. C. Bazerman (Milton Park: Taylor & Francis Group), 451–470.

McNamara, D. S., Crossley, S. A., Roscoe, R. D., Allen, L. K., and Dai, J. (2015). A hierarchical classification approach to automated essay scoring. Assess. Writ. 23, 35–59. doi: 10.1016/j.asw.2014.09.002

Muthén, L. K., and Muthén, B. O. (1998–2012). Mplus user’s Guide. Los Angeles: Muthén & Muthén.

Osnes, J. (1995). “Der Einflus von Handschrift und Fehlern auf die Aufsatzbeurteilung,” in Die Fragwürdigkeit der Zensurengebung (9. Aufl., S) , ed. K. Ingenkamp (Göttingen: Beltz), 131–147.

Peng, J., Wang, C., and Lu, X. (2020). Effect of the linguistic complexity of the input text on alignment, writing fluency, and writing accuracy in the continuation task. Langu. Teach. Res. 24, 364–381. doi: 10.1177/1362168818783341

Perelman, L. (2014). When “the state of the art” is counting words. Assess. Writ. 21, 104–111. doi: 10.1016/j.asw.2014.05.001

Pohlmann-Rother, S., Schoreit, E., and Kürzinger, A. (2016). Schreibkompetenzen von Erstklässlern quantitativ-empirisch erfassen-Herausforderungen und Zugewinn eines analytisch-kriterialen Vorgehens gegenüber einer holistischen Bewertung. J. Educ. Res. Online 8, 107–135.

Powers, D. E. (2005). Wordiness”: a selective review of its influence, and suggestions for investigating its relevance in tests requiring extended written responses. ETS Res. Rep. i–14.

Quinlan, T., Higgins, D., and Wolff, S. (2009). Evaluating the construct-coverage of the e-rater ® scoring engine. ETS Res. Rep. 2009, i–35. doi: 10.1002/j.2333-8504.2009.tb02158.x

Rezaei, A. R., and Lovorn, M. (2010). Reliability and validity of rubrics for assessment through writing. Assess. Writ. 15, 18–39. doi: 10.1016/j.asw.2010.01.003

Rubin, D. B. (1987). The calculation of posterior distributions by data augmentation: comment: a noniterative sampling/importance resampling alternative to the data augmentation algorithm for creating a few imputations when fractions of missing information are modest: the SIR algorithm. J. Am. Stat. Assoc. 82, 543–546. doi: 10.2307/2289460

Ruegg, R., and Sugiyama, Y. (2010). Do analytic measures of content predict scores assigned for content in timed writing? Melbourne Papers in Language Testing 15, 70–91.

Rupp, A. A., Casabianca, J. M., Krüger, M., Keller, S., and Köller, O. (2019). Automated essay scoring at scale: a case study in Switzerland and Germany. ETS Res. Rep. Ser. 2019, 1–23. doi: 10.1002/ets2.12249

Scannell, D. P., and Marshall, J. C. (1966). The effect of selected composition errors on grades assigned to essay examinations. Am. Educ. Res. J. 3, 125–130. doi: 10.3102/00028312003002125

Shermis, M. D. (2014). The challenges of emulating human behavior in writing assessment. Assess. Writ. 22, 91–99. doi: 10.1016/j.asw.2014.07.002

Silva, T. (1993). Toward an understanding of the distinct nature of L2 writing: the ESL research and its implications. TESOL Q. 27, 657–77. doi: 10.2307/3587400

Trautwein, U., Marsh, H. W., Nagengast, B., Lüdtke, O., Nagy, G., and Jonkmann, K. (2012). Probing for the multiplicative term in modern expectancy–value theory: a latent interaction modeling study. J. Educ. Psychol. 104, 763–777. doi: 10.1037/a0027470

Vögelin, C., Jansen, T., Keller, S. D., Machts, N., and Möller, J. (2019). The influence of lexical features on teacher judgements of ESL argumentative essays. Assess. Writ. 39, 50–63. doi: 10.1016/j.asw.2018.12.003

Vögelin, C., Jansen, T., Keller, S. D., and Möller, J. (2018). The impact of vocabulary and spelling on judgments of ESL essays: an analysis of teacher comments. Lang. Learn. J. 1–17. doi: 10.1080/09571736.2018.1522662

Weigle, S. C. (2003). Assessing Writing. Cambridge: Cambridge University Press.

Wind, S. A., Stager, C., and Patil, Y. J. (2017). Exploring the relationship between textual characteristics and rating quality in rater-mediated writing assessments: an illustration with L1 and L2 writing assessments. Assess. Writ. 34, 1–15. doi: 10.1016/j.asw.2017.08.003

Wolfe, E. W., Song, T., and Jiao, H. (2016). Features of difficult-to-score essays. Assess. Writ. 27, 1–10. doi: 10.1016/j.asw.2015.06.002

Keywords : text length, writing assessment, text quality, judgment bias, English as a foreign language, human raters, pre-service teachers

Citation: Fleckenstein J, Meyer J, Jansen T, Keller S and KÖller O (2020) Is a Long Essay Always a Good Essay? The Effect of Text Length on Writing Assessment. Front. Psychol. 11:562462. doi: 10.3389/fpsyg.2020.562462

Received: 15 May 2020; Accepted: 31 August 2020; Published: 25 September 2020.

Reviewed by:

Copyright © 2020 Fleckenstein, Meyer, Jansen, Keller and Köller. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Johanna Fleckenstein, [email protected]

Disclaimer: All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article or claim that may be made by its manufacturer is not guaranteed or endorsed by the publisher.

Essay / Long Answer Question Type

• Check it out in an example survey !
• Add a survey with this question to your account!

An Essay question is an open-text field used to gather longer open-text responses.

By default, the Essay question does not have a word/character limit. You can determine the response length by setting the maximum number of words or characters that the respondent can input. To learn more check out our Min/Max Word and Character Count Tutorial .

You can also set the width and the height of the Essay input box. 

When To Use It

Use Essay Questions to collect a paragraph or two of text. For example, ask respondents to provide additional feedback in their own words. Still not sure which text field to select? Check out our tutorial on Textboxes vs. Essay .

Min/Max Character & Word Counts

Use the Validation tab to specify:

• Min & Max Character Count*
• Min & Max Word Count

*Spaces are counted towards the character count.

Adjust Essay Box Size

The Essay question textbox has a default size of 40 x 5 (width x height in pixels). Use the Layout tab to adjust the sizing if needed. Input different pixel values or use the up/down arrows to increment the sizing up/down by one.

Survey Taking

On a desktop and most laptops, the Essay question type looks like so.

When optimized for mobile devices the Essay question type looks very similar.

By default, survey questions show one at a time on mobile devices to prevent the need for scrolling on smaller screens. You can turn off this one-at-a-time interaction if you wish.

In a Standard Report, Essay questions report as a Word Cloud, by default, with a list of the responses below. In the list of responses the initial display of each response is limited to 1500 characters.  There will be a Read More link at the end of the 1500 characters available to display the full text. 

Learn about your customization options for the Word Cloud element . 

  See additional compatible chart types

Within the Standard Report there are various chart types available for visualizing your data. The below grid shows which of the chart types Essay questions are compatible with. 

  See what other report types are compatible

The below grid shows which of the report types Essay questions are compatible with. If you plan to do some specific analysis within SurveyGizmo this report compatibility chart should help you choose the right question types.

In the CSV/Excel export, the data is displayed in a single column within the spreadsheet.

• Leading Practices: Textbox Vs. Essay Which Question Type Should I Use?
• Add Placeholder Text to Textbox and Essay Questions
• Textbox Question Type
• Custom Group Question Type
• Custom Table Question Type
• Author: Bri Hillmer

• Last updated: 04/04/2018 3:46 pm MDT

— Mike.cordeiro on 03/25/2020

@Paul: Thanks so much for your question! While the Essay Question does not provide this functionality, we do have a File Upload Question Type allowing respondents to upload up to 10 documents into a survey response. Follow the link below for more information on this question type! https://help.surveygizmo.com/help/file-upload#feature-table I hope this helps! Michael Technical Writer SurveyGizmo Learning and Development

— Lindsey on 03/25/2020

Is it possible for respondents to upload documents (Word files, for example) either here or in some other question type?

— Mike.cordeiro on 01/19/2020

@Megan: Thank you for your question! I'm afraid that there is not a way to format an essay textbox in the way you described using out-of-the-box features to have the respondents be able to answer in bullet point form with the Essay Question Type. This type of formatting is possible but would require custom scripting via the survey Style tab. If you don't have access to web development resources, our Professional Services team is available for these types of custom projects: https://www.surveygizmo.com/professional-services/ I hope this helps Clarify! Michael Cordeiro SurveyGizmo Support

— Megan on 01/16/2020

Is it possible to allow respondents to use numbered or bulleted lists in their responses?

— Dave Domagalski on 10/03/2019

@Jim: Thank you for your question! I'm sorry for the trouble! To include open-text answers in a PDF or Excel download of a report, you must check the box to 'Include lists of open text answers' after selecting the download method: https://help.surveygizmo.com/help/report-downloads My apologies if you are already doing this! If you continue to see issues with this, please don't hesitate to reach out to our Support team: https://help.surveygizmo.com/help/surveygizmo-support-hours I hope this helps! David Technical Writer SurveyGizmo Learning & Development

— Jim on 10/03/2019

how do i get the essay responses to show in the .pdf report? They show up online when i click on show responses, but i can't get them to show on the .pdf or excel downloads.

— Dave Domagalski on 03/28/2019

@Product: Thank you for your question! By default, the Essay question will report as a Word Cloud element. This Word Cloud can be customized to include a list of responses, as noted in the introduction here: https://help.surveygizmo.com/help/standard-report-word-cloud If you continue to see the 'no data to display' message, it might be something that our Support team needs to take a closer look at. Please don't hesitate to reach out if this is the case: https://help.surveygizmo.com/help/surveygizmo-support-hours I hope this helps! David Technical Writer SurveyGizmo Customer Experience

— Product on 03/28/2019

How can I see all the individual essay responses as part of a report? Now, it says: "No data to display." But if I go to the "Results" tab and click on "Individual Responses," I see them all. I just can't get them to show up in a report.

— Dave Domagalski on 09/05/2018

@Sean: Thank you for your question! I'm afraid that there is not a way to format an essay textbox in the way you described using out-of-the-box features. This type of formatting is possible but would require extensive custom CSS modification via the survey Style tab. If you don't have access to web development resources, our Programming Services team is available for these types of custom projects: https://www.surveygizmo.com/programming-services/ I hope this helps to clarify! David Documentation Specialist SurveyGizmo Customer Experience

— Sean.mcnally on 09/04/2018

Is there a way to format the actual text box? I'd like to show the response box in a speech bubble, or as part of an image. For example, in an image that shows a person making a comment and a blank speech bubble for a person replying to the comment. The approach is used qualitative research , but I'd like to extend it to quantitative studies.

— Dave Domagalski on 04/25/2018

@Joe: Thank you for your question! I'm afraid that a setting to restrict certain characters/emojis is not currently available as part of the Essay / Long Answer question type. This is something that would currently require a custom solution via JavaScript. If you don't have development resources available, this is something that our Programming Services team could build, for a fee: https://www.surveygizmo.com/programming-services/ I'm sorry for the trouble! David Documentation Specialist SurveyGizmo Customer Experience

— Joe on 04/25/2018

Is there a way to restrict people from submitting emoji's on there replies? thank you

— Dave Domagalski on 03/21/2018

@Yehonatan: Thank you for your note! I'm sorry for the trouble. In my testing, Android devices were not covering up the 'Next' button on the mobile version (one-at-a-time), whereas iOS devices were covering the buttons, but did present a 'done' button to hide the keyboard. In any case, I'm afraid that there is not a setting or built-in feature that would change this behavior. If you could share an example of a device type that you have seen this behavior on with our Support team, we can explore our options. Alternatively, one option is to provide instructions to the survey respondents on best practices for closing the virtual keyboard. I hope this helps! David Documentation Specialist SurveyGizmo Customer Experience

— Yehonatan on 03/21/2018

I'm having a problem with this question type in mobile survies. when the user types his response, the virtual keyboard covers the 'next' button of the survey. the user then has to somehow 'pull down' his keyboard (which sometimes involves pressing the back button of his phone) and if he does incorrectly it sometimes counts as having his phone retrive previous page / app. and we don't want that to happen. is there a a good way to work around the problem?

— Bri Hillmer on 11/03/2016

@Kimberley: At this time there is not a way to work around this other than the quotes that you suggested. You're not the first customer to run into this. We've been tracking the number of customers affected by this issue. I will add you to the list! Bri Documentation Coordinator/Survey Sorceress SurveyGizmo Customer Experience Team

— Kimberley on 11/03/2016

Hello, I have a response to an essay question that is 1757 characters long with spaces. In the standard report this response is cut off at 997 characters. (Word Cloud is turned off in the report's global options and the responses are shown in a summary table). Is there a way to stop this response from being cut off? I could add this response as a quote but then I have some of the text twice.

— Dave Domagalski on 10/13/2016

@Brad: Thank you for your feedback! The following article should help provide specifics regarding the Word Cloud element: https://help.surveygizmo.com/help/word-cloud I have now linked to the Word Cloud content from this article under the Reporting section. Thank you again for your feedback and helping us improve SurveyGizmo Documentation. Best regards, David Documentation Specialist/Survey Explorer SurveyGizmo Customer Experience

— Brad on 10/13/2016

Although you referenced that Essay question responses are generated as a "Word Cloud" but you do not provide any guidance on how to stop having a Word Cloud displayed - Frankly, altho it might be "flashy"; it is worthless when all you want is to show what the individual answers were per the essay type question.

• Learn how to use our help!
• SurveyGizmo Glossary
• Need to contact support?
• Send Your Survey Via Email
• Set Up Question Logic
• Create a Survey
• Merge Codes: Dynamically Populate Data
• Using URL Variables to Pass Data
• Enterprise Experience Theme
• Individual Responses: Details Tab
• Return to Work Assessment Template (Covid-19)
• Retiring TLS 1.0 and TLS 1.1 SSL protocols
• Gabor-Granger Pricing Method
• See more...
• Search Help
• Community Questions
• Documentation
• Developer Resources
• Terms of Use
• Privacy Policy

Our approach

• Responsibility
• Infrastructure
• Try Meta AI

RECOMMENDED READS

• 5 Steps to Getting Started with Llama 2
• The Llama Ecosystem: Past, Present, and Future
• Introducing Code Llama, a state-of-the-art large language model for coding
• Meta and Microsoft Introduce the Next Generation of Llama
• Today, we’re introducing Meta Llama 3, the next generation of our state-of-the-art open source large language model.
• Llama 3 models will soon be available on AWS, Databricks, Google Cloud, Hugging Face, Kaggle, IBM WatsonX, Microsoft Azure, NVIDIA NIM, and Snowflake, and with support from hardware platforms offered by AMD, AWS, Dell, Intel, NVIDIA, and Qualcomm.
• We’re dedicated to developing Llama 3 in a responsible way, and we’re offering various resources to help others use it responsibly as well. This includes introducing new trust and safety tools with Llama Guard 2, Code Shield, and CyberSec Eval 2.
• In the coming months, we expect to introduce new capabilities, longer context windows, additional model sizes, and enhanced performance, and we’ll share the Llama 3 research paper.
• Meta AI, built with Llama 3 technology, is now one of the world’s leading AI assistants that can boost your intelligence and lighten your load—helping you learn, get things done, create content, and connect to make the most out of every moment. You can try Meta AI here .

Today, we’re excited to share the first two models of the next generation of Llama, Meta Llama 3, available for broad use. This release features pretrained and instruction-fine-tuned language models with 8B and 70B parameters that can support a broad range of use cases. This next generation of Llama demonstrates state-of-the-art performance on a wide range of industry benchmarks and offers new capabilities, including improved reasoning. We believe these are the best open source models of their class, period. In support of our longstanding open approach, we’re putting Llama 3 in the hands of the community. We want to kickstart the next wave of innovation in AI across the stack—from applications to developer tools to evals to inference optimizations and more. We can’t wait to see what you build and look forward to your feedback.

Our goals for Llama 3

With Llama 3, we set out to build the best open models that are on par with the best proprietary models available today. We wanted to address developer feedback to increase the overall helpfulness of Llama 3 and are doing so while continuing to play a leading role on responsible use and deployment of LLMs. We are embracing the open source ethos of releasing early and often to enable the community to get access to these models while they are still in development. The text-based models we are releasing today are the first in the Llama 3 collection of models. Our goal in the near future is to make Llama 3 multilingual and multimodal, have longer context, and continue to improve overall performance across core LLM capabilities such as reasoning and coding.

State-of-the-art performance

Our new 8B and 70B parameter Llama 3 models are a major leap over Llama 2 and establish a new state-of-the-art for LLM models at those scales. Thanks to improvements in pretraining and post-training, our pretrained and instruction-fine-tuned models are the best models existing today at the 8B and 70B parameter scale. Improvements in our post-training procedures substantially reduced false refusal rates, improved alignment, and increased diversity in model responses. We also saw greatly improved capabilities like reasoning, code generation, and instruction following making Llama 3 more steerable.

*Please see evaluation details for setting and parameters with which these evaluations are calculated.

In the development of Llama 3, we looked at model performance on standard benchmarks and also sought to optimize for performance for real-world scenarios. To this end, we developed a new high-quality human evaluation set. This evaluation set contains 1,800 prompts that cover 12 key use cases: asking for advice, brainstorming, classification, closed question answering, coding, creative writing, extraction, inhabiting a character/persona, open question answering, reasoning, rewriting, and summarization. To prevent accidental overfitting of our models on this evaluation set, even our own modeling teams do not have access to it. The chart below shows aggregated results of our human evaluations across of these categories and prompts against Claude Sonnet, Mistral Medium, and GPT-3.5.

Preference rankings by human annotators based on this evaluation set highlight the strong performance of our 70B instruction-following model compared to competing models of comparable size in real-world scenarios.

Our pretrained model also establishes a new state-of-the-art for LLM models at those scales.

To develop a great language model, we believe it’s important to innovate, scale, and optimize for simplicity. We adopted this design philosophy throughout the Llama 3 project with a focus on four key ingredients: the model architecture, the pretraining data, scaling up pretraining, and instruction fine-tuning.

Model architecture

In line with our design philosophy, we opted for a relatively standard decoder-only transformer architecture in Llama 3. Compared to Llama 2, we made several key improvements. Llama 3 uses a tokenizer with a vocabulary of 128K tokens that encodes language much more efficiently, which leads to substantially improved model performance. To improve the inference efficiency of Llama 3 models, we’ve adopted grouped query attention (GQA) across both the 8B and 70B sizes. We trained the models on sequences of 8,192 tokens, using a mask to ensure self-attention does not cross document boundaries.

Training data

To train the best language model, the curation of a large, high-quality training dataset is paramount. In line with our design principles, we invested heavily in pretraining data. Llama 3 is pretrained on over 15T tokens that were all collected from publicly available sources. Our training dataset is seven times larger than that used for Llama 2, and it includes four times more code. To prepare for upcoming multilingual use cases, over 5% of the Llama 3 pretraining dataset consists of high-quality non-English data that covers over 30 languages. However, we do not expect the same level of performance in these languages as in English.

To ensure Llama 3 is trained on data of the highest quality, we developed a series of data-filtering pipelines. These pipelines include using heuristic filters, NSFW filters, semantic deduplication approaches, and text classifiers to predict data quality. We found that previous generations of Llama are surprisingly good at identifying high-quality data, hence we used Llama 2 to generate the training data for the text-quality classifiers that are powering Llama 3.

We also performed extensive experiments to evaluate the best ways of mixing data from different sources in our final pretraining dataset. These experiments enabled us to select a data mix that ensures that Llama 3 performs well across use cases including trivia questions, STEM, coding, historical knowledge, etc.

Scaling up pretraining

To effectively leverage our pretraining data in Llama 3 models, we put substantial effort into scaling up pretraining. Specifically, we have developed a series of detailed scaling laws for downstream benchmark evaluations. These scaling laws enable us to select an optimal data mix and to make informed decisions on how to best use our training compute. Importantly, scaling laws allow us to predict the performance of our largest models on key tasks (for example, code generation as evaluated on the HumanEval benchmark—see above) before we actually train the models. This helps us ensure strong performance of our final models across a variety of use cases and capabilities.

We made several new observations on scaling behavior during the development of Llama 3. For example, while the Chinchilla-optimal amount of training compute for an 8B parameter model corresponds to ~200B tokens, we found that model performance continues to improve even after the model is trained on two orders of magnitude more data. Both our 8B and 70B parameter models continued to improve log-linearly after we trained them on up to 15T tokens. Larger models can match the performance of these smaller models with less training compute, but smaller models are generally preferred because they are much more efficient during inference.

To train our largest Llama 3 models, we combined three types of parallelization: data parallelization, model parallelization, and pipeline parallelization. Our most efficient implementation achieves a compute utilization of over 400 TFLOPS per GPU when trained on 16K GPUs simultaneously. We performed training runs on two custom-built 24K GPU clusters . To maximize GPU uptime, we developed an advanced new training stack that automates error detection, handling, and maintenance. We also greatly improved our hardware reliability and detection mechanisms for silent data corruption, and we developed new scalable storage systems that reduce overheads of checkpointing and rollback. Those improvements resulted in an overall effective training time of more than 95%. Combined, these improvements increased the efficiency of Llama 3 training by ~three times compared to Llama 2.

Instruction fine-tuning

To fully unlock the potential of our pretrained models in chat use cases, we innovated on our approach to instruction-tuning as well. Our approach to post-training is a combination of supervised fine-tuning (SFT), rejection sampling, proximal policy optimization (PPO), and direct preference optimization (DPO). The quality of the prompts that are used in SFT and the preference rankings that are used in PPO and DPO has an outsized influence on the performance of aligned models. Some of our biggest improvements in model quality came from carefully curating this data and performing multiple rounds of quality assurance on annotations provided by human annotators.

Learning from preference rankings via PPO and DPO also greatly improved the performance of Llama 3 on reasoning and coding tasks. We found that if you ask a model a reasoning question that it struggles to answer, the model will sometimes produce the right reasoning trace: The model knows how to produce the right answer, but it does not know how to select it. Training on preference rankings enables the model to learn how to select it.

Building with Llama 3

Our vision is to enable developers to customize Llama 3 to support relevant use cases and to make it easier to adopt best practices and improve the open ecosystem. With this release, we’re providing new trust and safety tools including updated components with both Llama Guard 2 and Cybersec Eval 2, and the introduction of Code Shield—an inference time guardrail for filtering insecure code produced by LLMs.

We’ve also co-developed Llama 3 with torchtune , the new PyTorch-native library for easily authoring, fine-tuning, and experimenting with LLMs. torchtune provides memory efficient and hackable training recipes written entirely in PyTorch. The library is integrated with popular platforms such as Hugging Face, Weights & Biases, and EleutherAI and even supports Executorch for enabling efficient inference to be run on a wide variety of mobile and edge devices. For everything from prompt engineering to using Llama 3 with LangChain we have a comprehensive getting started guide and takes you from downloading Llama 3 all the way to deployment at scale within your generative AI application.

A system-level approach to responsibility

We have designed Llama 3 models to be maximally helpful while ensuring an industry leading approach to responsibly deploying them. To achieve this, we have adopted a new, system-level approach to the responsible development and deployment of Llama. We envision Llama models as part of a broader system that puts the developer in the driver’s seat. Llama models will serve as a foundational piece of a system that developers design with their unique end goals in mind.

Instruction fine-tuning also plays a major role in ensuring the safety of our models. Our instruction-fine-tuned models have been red-teamed (tested) for safety through internal and external efforts. ​​Our red teaming approach leverages human experts and automation methods to generate adversarial prompts that try to elicit problematic responses. For instance, we apply comprehensive testing to assess risks of misuse related to Chemical, Biological, Cyber Security, and other risk areas. All of these efforts are iterative and used to inform safety fine-tuning of the models being released. You can read more about our efforts in the model card .

Llama Guard models are meant to be a foundation for prompt and response safety and can easily be fine-tuned to create a new taxonomy depending on application needs. As a starting point, the new Llama Guard 2 uses the recently announced MLCommons taxonomy, in an effort to support the emergence of industry standards in this important area. Additionally, CyberSecEval 2 expands on its predecessor by adding measures of an LLM’s propensity to allow for abuse of its code interpreter, offensive cybersecurity capabilities, and susceptibility to prompt injection attacks (learn more in our technical paper ). Finally, we’re introducing Code Shield which adds support for inference-time filtering of insecure code produced by LLMs. This offers mitigation of risks around insecure code suggestions, code interpreter abuse prevention, and secure command execution.

With the speed at which the generative AI space is moving, we believe an open approach is an important way to bring the ecosystem together and mitigate these potential harms. As part of that, we’re updating our Responsible Use Guide (RUG) that provides a comprehensive guide to responsible development with LLMs. As we outlined in the RUG, we recommend that all inputs and outputs be checked and filtered in accordance with content guidelines appropriate to the application. Additionally, many cloud service providers offer content moderation APIs and other tools for responsible deployment, and we encourage developers to also consider using these options.

Deploying Llama 3 at scale

Llama 3 will soon be available on all major platforms including cloud providers, model API providers, and much more. Llama 3 will be everywhere .

Our benchmarks show the tokenizer offers improved token efficiency, yielding up to 15% fewer tokens compared to Llama 2. Also, Group Query Attention (GQA) now has been added to Llama 3 8B as well. As a result, we observed that despite the model having 1B more parameters compared to Llama 2 7B, the improved tokenizer efficiency and GQA contribute to maintaining the inference efficiency on par with Llama 2 7B.

For examples of how to leverage all of these capabilities, check out Llama Recipes which contains all of our open source code that can be leveraged for everything from fine-tuning to deployment to model evaluation.

What’s next for Llama 3?

The Llama 3 8B and 70B models mark the beginning of what we plan to release for Llama 3. And there’s a lot more to come.

Our largest models are over 400B parameters and, while these models are still training, our team is excited about how they’re trending. Over the coming months, we’ll release multiple models with new capabilities including multimodality, the ability to converse in multiple languages, a much longer context window, and stronger overall capabilities. We will also publish a detailed research paper once we are done training Llama 3.

To give you a sneak preview for where these models are today as they continue training, we thought we could share some snapshots of how our largest LLM model is trending. Please note that this data is based on an early checkpoint of Llama 3 that is still training and these capabilities are not supported as part of the models released today.

We’re committed to the continued growth and development of an open AI ecosystem for releasing our models responsibly. We have long believed that openness leads to better, safer products, faster innovation, and a healthier overall market. This is good for Meta, and it is good for society. We’re taking a community-first approach with Llama 3, and starting today, these models are available on the leading cloud, hosting, and hardware platforms with many more to come.

Try Meta Llama 3 today

We’ve integrated our latest models into Meta AI, which we believe is the world’s leading AI assistant. It’s now built with Llama 3 technology and it’s available in more countries across our apps.

You can use Meta AI on Facebook, Instagram, WhatsApp, Messenger, and the web to get things done, learn, create, and connect with the things that matter to you. You can read more about the Meta AI experience here .

Visit the Llama 3 website to download the models and reference the Getting Started Guide for the latest list of all available platforms.

You’ll also soon be able to test multimodal Meta AI on our Ray-Ban Meta smart glasses.

As always, we look forward to seeing all the amazing products and experiences you will build with Meta Llama 3.

Our latest updates delivered to your inbox

Subscribe to our newsletter to keep up with Meta AI news, events, research breakthroughs, and more.

Join us in the pursuit of what’s possible with AI.

Product experiences

Foundational models

Latest news

Meta © 2024