business speech science research

Chapter 6: Researching Your Speech

Learning Objectives

• Learn that research is not only useful, but fun.
• Describe how to establish research needs before beginning research.
• Identify appropriate scholarly and popular sources.
• Differentiate between primary and secondary research.
• Understand how to incorporate sources within a speech and how to use sources ethically.
• Differentiate between direct quotations and paraphrases of information within a speech.
• Explain twelve strategies for avoiding plagiarism.
• CRAAP Method
• Direct Quotation
• “Drive-by” Quoting
• Encyclopedias
• General-Interest Periodicals
• Interlibrary Loan
• Peer-reviewed Sources
• Popular Sources
• Primary Research
• Representative Sample
• Research Log
• Scholarly Sources
• Secondary Research
• Special-Interest Periodicals
• Subheadings
• Topic Sentence
• World Wide Web

style scholars in the various social science fields (e.g., psychology, human communication, business) are more likely to use

CRAAP stands for “currency,” “relevance,” “authority,” “accuracy,” and “purpose,” or the five ways that you should evaluate each source to determine if it represents the best information available at the time

an online searchable collection of information

when you cite the actual words from a source with no changes

a practice that disorients your audience by not giving them everything they need to understand how the source is relevant to your own claims

information sources that provide short, very general information about a topic and are available in both print and electronic formats

magazines and newsletters published on a fairly systematic basis

a title at the head of a page or section of a book

a process where librarians are able to search other libraries to locate the book a researcher is trying to find

a scholarly publication containing articles written by researchers, professors and other experts

a word or concept of great significance

the style scholars in the various humanities fields (e.g., English, philosophy, rhetoric) are more likely to use

to take a source’s basic idea and condense it using your own words

an article that has been reviewed by a group of experts in the field, sometimes called a board of editors

(also called non-scholarly) sources inform and entertain the public or allow practitioners to share industry, practice, and production information

carried out to discover or revise facts, theories, and applications and is reported by the person conducting the research

a group or set chosen from a larger statistical population or group of factors or instances that adequately replicates the larger group according to whatever characteristic or quality is under study

scholarly investigation into a topic in order to discover, revise, or report facts, theories, and applications

step-by-step account of the process of identifying, obtaining, and evaluating sources for a specific project, similar to a lab note-book in an experimental setting

are written by experts in their field, usually professors in a specific discipline

research carried out to discover or revise facts, theories, and applications—similar to primary research—but it is reported by someone not involved in conducting the actual research

magazines and newsletters that are published for a narrower audience

a heading given to a subsection of a piece of writing

clear sentence that restates the preview statement in past tense, outlining the main points that were addressed in the speech

The first sentence of each paragraph is the topic sentence, which is basically a paragraph’s thesis statement: well-written topic sentences tell the reader what the entire paragraph is about.

an interconnected system of public webpages accessible through the Internet

It’s About Them: Public Speaking in the 21st Century Copyright © 2022 by LOUIS: The Louisiana Library Network is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License , except where otherwise noted.

Share This Book

Research & Innovation Office

• About the Research & Innovation Office
• Vice President for Research & Innovation
• RIO Leadership Team
• E4 by Design
• UMN Research Statistics
• Administrative Support Units
• Academic Centers & Institutes
• Council of Research Associate Deans (CRAD)
• Community Oversight Board
• Research Integrity & Safety Collaborative (RISC)
• Research Facilities & Expertise
• Research Tools & Applications
• The Office of Research Information Systems
• ORIS Staff List
• Research Animal Resources
• About BSL-3
• Frequently Asked Questions
• BSL-3/ABSL-3 Research Laboratory Suites
• Related Resources
• Contact BSL-3
• International Research Support & Guidance
• Disclosing Foreign Support to Federal Science Agencies
• Foreign Talent Recruitment Programs
• Hosting International Visitors, Scholars & Researchers
• Traveling Abroad
• International Institutional Agreements
• Export Controls & Sanctions Compliance
• Helpful Resources & Contacts
• About Artist-in-Residence Awards
• Award Recipients
• About Biotechnology & Biomanufacturing Seed Grants
• BBIC Awardees
• About Grant-in-Aid
• Exemplary Proposals
• Awarded Equipment
• NC A&T & UMN Research Partnership
• About the Program
• Past Awards: 2023
• Past Awards: 2022
• Past Awards: 2021
• Past Awards: 2020
• Past Awards: 2019
• Past Awards: 2018
• Past Awards: 2017
• Past Awards: 2016
• Past Awards: 2015
• Past Awards: 2014
• Past Awards: 2013
• Past Awards: 2011
• About Social Justice Impact Grants
• SJIG Awardees
• About the Innovation Impact Case Award
• 2022 Awardees
• 2023 Awardees
• Research Technical Staff Award
• Matching Funds
• Research Computing Funding
• COVID-19 Rapid Response Grants
• About the International Institute for Biosensing
• About Minnesota Futures
• Minnesota Futures: Past Awards
• Review Committee
• MnDRIVE Funding
• U of M Internal Funding
• Limited Extramural Funding
• Proposals & Awards (SPA)
• F&A Costs, Effort Reporting
• Research Integrity & Compliance (RIC)
• Human Research Protection
• Institutional Review Board (IRB)
• Animal Care & Use (IACUC)
• Biotechnology Activities Oversight (OBAO)
• Export Controls
• Conflict of Interest
• Reporting Research Misconduct
• Research Compliance Training Guide
• Research Ethics Week
• Partnerships for Corporations & Industry
• Building Successful Industry Partnerships
• Research Agreements
• Confidentiality Agreements
• Starting a Company (Venture Center)
• Technology Commercialization
• About MnDrive
• Louis Stokes North Star STEM Alliance/MnDRIVE Partnership
• Transdisciplinary Research Program
• Sponsor Research (MN-IP)
• News & Announcements
• All Past News & Announcements
• UMN Research Newsletter
• UMN Washington Update

Speaking Science: Inspiring Interest in Academic Research

Photo: Julie Hanus/UMN Institute on the Environment

University researchers are constantly pushing the limits of knowledge in their fields, making discoveries that stand to change how people think and the way they live. For research to inform the decisions people make, however, they must first be able to understand it.

Earlier this month, the University of Minnesota hosted the latest conference in a series designed to help scientists more effectively share their knowledge and research with audiences outside of academia. “Speaking Science: Communicating with Media, Funders, Policymakers, and the Public” brought together faculty, postdoctoral researchers, and graduate students interested in honing their communications and storytelling skills.

The one-day event, hosted by a collection of colleges and units from across the University, featured interactive training exercises, breakout sessions, and a keynote presentation by award-winning scientist Hope Jahren, PhD, who detailed her paleobiology career to date in the national bestselling memoir Lab Girl .

Making Them Care

While it’s tempting to think that data alone will change the way people think and behave, often it takes a little bit more—the audience needs to care.

In “The Science of What Makes People Care,” University of Florida Center for Public Interest Communications experts Ellen Nodine and Matt Sheehan distilled principles from behavioral, cognitive, and social science into six guidelines that scientists can use to help get people to care about their work and what it means.

Among these principles was shaping the communications approach to fit the audience scientists aim to reach. By understanding the values and identity of these individuals or groups of people, Sheehan said, scientists can approach them from a familiar standpoint and better connect and communicate with them. This step can help avoid “solution aversion,” where people dismiss an idea because it goes against their worldview or makes them uncomfortable.

“Effective communication is no longer about just pushing your message,” Sheehan said. “We need to find where our values intersect.”

Another important lesson was to “talk in pictures.” This includes employing actual photographs and visuals, but also written and spoken language that helps the audience picture an environment or an action in their mind. Nodine and Sheehan suggest painting a more vibrant picture using imagery like “sit down together at a table” and “join hands” to deliver a more memorable point to the audience than by speaking in abstraction.

“We remember material far better when it’s expressed in concrete language that allows people to make more visual images,” Nodine said. “What kinds of images and metaphors can you use to help people understand and connect with your ideas?”

In addition to using imagery, a message will make more of an impact when it draws on emotion. Nodine and Sheehan suggest considering how intentional use of emotions like anger, humor, and sadness can amplify the message. They also suggest considering how less commonly used emotions could factor in. What role can hope, pride, or awe play?

When it comes to the message itself, cut back on field-specific jargon and instead define just the key terms and concepts the audience needs to know. Think also about what you want the audience to take away from your message. Is there a call to action? If so, give them a specific action relevant to their lives that they can take, and then show how their participation will support a larger goal.

Finally, remember to tell stories. Think about how exposition, characters, conflict, and resolution can help connect people to the meaning behind the facts and data.

“Our societies are built on the stories that we tell ourselves,” Sheehan said. “We are uniquely wired to respond to stories, not to facts.”

The Art of a Likable Science Talk

When it comes time to present your research, will people stay tuned in? Will they like it? As Hope Jahren told attendees during her keynote presentation, these questions matter.

“A good talk can make or break a career,” said Jahren, who got her bachelor’s degree at UMN Twin Cities and is currently a professor at Norway’s University of Oslo. “This is our oral tradition. It is the talks that we give, not the papers we write, that really let us share who we are and what we do.”

To discover how to give a scientific talk that people will like, Jahren and her team studied the composition of dozens of scientific talks, recording data about the length of the presentation, what worked well, and what didn’t. While the concept of “liking” something is inherently subjective, several researchers evaluated each talk to ensure no individual bias swayed the numbers. The team also tallied the types of slides used to gain a clearer picture as to exactly how presenters were conveying their information: text, photos, maps, graphs, tables, equations, or videos.

What Jahren and her team found was that there’s no clear “recipe” for how many of each type of slide presenters should use to achieve a likable talk. Some excellent talks were heavy on photos, while others used lots of graphs. There was, however, a consistent length for good presentations—one that has become a “contract” of sorts between the speaker and the audience—which averages to three quarters of an hour long, with about 33 slides total and just over one minute spent on each slide.

Jahren’s study also found that people generally want to like a scientific talk when they hear one, giving the presenter the benefit of the doubt, though she noted that this could vary based on the culture, the specific academic discipline, or the particular individuals in the audience.

At the end of a scientific talk, Jahren said it’s important to remember to thank the audience for choosing to spend their time listening to and discussing scientific ideas.

“We should all feel good about what we just accomplished,” she said. “We learned and we shared and we came together as people who are all interested in the same thing. And that in itself is a wonderful, wonderful way to spend the day.”

More Tips for Speaking Science

Want more pointers on speaking about science with a general audience? Check out these five tips from the conference’s inaugural year, and eight more from its second year.

• Admissions Overview
• Undergraduate Admissions
• Graduate Degree Programs
• International Student Admissions
• Academics Overview
• Undergraduate Majors & Minors
• Graduate School
• Purdue Online Learning
• Tour Purdue’s Campus
• Research and Innovation Overview
• Research & Partnerships
• Corporate & Global Partnerships
• Purdue Research Foundation
• About Purdue
• Office of the President
• Commitment to Free Speech
• Student Life at Purdue
• Purdue Activity & Wellness
• Campus Inclusion
• Prospective Students
• Current Students
• Faculty and Staff
• Purdue Northwest
• Purdue Fort Wayne
• Purdue Global
• Purdue Online

Department of Speech, Language, and Hearing Sciences Research

In Purdue’s Department of Speech, Language, and Hearing Sciences (SLHS), faculty and students collaborate to make discoveries that define cutting-edge clinical practice and advance basic scientific knowledge related to human communication.

Research Areas

Hearing science; hearing disorders.

Faculty who study hearing science and disorders in the Department of Speech, Language, and Hearing Sciences focus on a variety of topics, including computational modeling, hidden hearing loss, neurophysiology and electrophysiology, psychophysics and speech perception.

Language Science; Language Disorders and Disabilities

Researchers in the Department of Speech, Language, and Hearing Sciences study language science, disorders and disabilities to advance discovery across topics such as aphasia, autism spectrum disorder (ASD), language development, language disorders, linguistics and sign language.

Speech, Swallowing and Voice Science; Speech, Swallowing, Voice Disorders

To improve quality of life and help people communicate effectively, faculty in the Department of Speech, Language, and Hearing Sciences investigate dysphagia, fluency disorders, neurogenic problems, speech sound disorders, and voice disorders.

Center and Institutes/Laboratories

Acquired Brain Injury Communication and Cognition (ABC) Lab

The Acquired Brain Injury Communication and Cognition (ABC) Lab studies the behavioral and neural factors that support the recovery of language and cognition in people who have aphasia and/or a traumatic brain injury. The lab conducts both basic and applied research studies with the goal of developing interventions for language comprehension and cognition.

Arianna LaCroix

Lab Website

Attention and Neurodevelopmental Disorders (AtteND) Lab

The Attention and Neurodevelopmental Disorders Lab investigates attentional strengths and weaknesses in individuals who are at risk for or diagnosed with autism spectrum disorder (ASD) to identify new targets for intervention as well as improve current intervention strategies. This research provides insight into how attention affects the development of social and communicative abilities in typically developing children and children with ASD.

Brandon Keehn

Aphasia Research Laboratory

The Aphasia Research Laboratory focuses on how language processing is affected by aging and acquired neurological conditions (stroke, Parkinson’s disease) and identifies the factors that facilitate language recovery in individuals with aphasia. The findings provide insight into how language is stored and processed in the brain and aid in the development of intervention approaches for people with aphasia.

Auditory Cognitive Neuroscience Laboratory

When listening to speech in a noisy public place, most adults find it easier to understand the speaker if they can see the person’s face because facial movements can provide a great deal of information about speech content. However, this ability to use visual speech cues when the sound quality is poor is not present at birth and develops gradually in children. To better understand this, the Auditory Cognitive Neuroscience Laboratory studies how and when children learn to use visual speech information and how their language development may be negatively affected if they fail to acquire this skill. We work with typically developing children and also with children who had delayed language acquisition.

Natalya Kaganovich

Auditory Electrophysiology Laboratory

The Auditory Electrophysiology Laboratory uses electrophysiological measures to understand the neural representation of complex sounds in normal and impaired ears at the brainstem and cortical levels and how these representations are shaped by experience. The long-term objective is to enhance the encoding of behaviorally relevant dimensions of sounds and determine their relative roles in the temporal structure of sound. We are also interested in evaluating the nature of the interplay between early sensory level processes and later cognitive levels of processing.

Ravi Krishnan

Auditory Neurophysiology and Modeling Lab

Research in the Auditory Neurophysiology and Modeling Laboratory involves the coordinated use of neurophysiology, psychoacoustics and computational modeling. This multidisciplinary approach provides a powerful framework to enhance our understanding of the effects of different types of sensorineural hearing loss on neural and perceptual responses to sound. This knowledge will be extremely valuable for developing diagnostic tests, evaluating the limitations of current hearing aids and suggesting novel strategies for hearing aids and cochlear implants.

Brain Research in Auditory Neuroscience (BRAiN) Lab

The BRAiN Lab investigates auditory perception and speech processing in adults with cochlear implants. The goal of this research is to better understand the individual differences in auditory perception that impact speech-recognition outcomes in cochlear implant recipients.

Maureen Shader

Child Language Research Laboratory

In the Child Language Research Laboratory, we study how children learn to produce and understand words and sentences. We are especially interested in discovering the reasons behind why children with language impairments experience difficulties — and finding ways to help these children overcome their language learning problems. Our studies also include children who are developing language without difficulty, so we can have a clear idea of the learning patterns associated with typical development.

Laurence B. Leonard

Child Phonology Laboratory

Research in the Child Phonology Laboratory investigates how monolingual and bilingual children learn to produce speech sounds with the goal of developing best practices for assessing and treating children with phonological disorders. In particular, we are interested in better understanding how the ability to perceive speech sounds affects the accurate production of speech.

Françoise Brosseau-Lapré

Cranial Sensorimotor Control and Neurodegeneration Lab – Schaser Research Group

The Schaser Research Group uses advanced imaging tools and an alpha-synuclein fibril seeding approach to study protein aggregation in a mouse model of Parkinson’s disease and Dementia with Lewy Bodies. We specialize in merging clinical issues, animal behavior, and exploration of pathology in the cranial sensorimotor system to characterize and treat voice, communication, and swallowing deficits related to neurodegeneration.

Allison Schaser

In the Purdue I-EaT Lab, we investigate the underlying mechanisms that control swallowing function and eating in adults and children with and without swallowing disorders (dysphagia). The primary studies involve patients with cerebral palsy, stroke and Parkinson’s disease. We aim to use this knowledge to develop and evaluate novel rehabilitative treatments that will be effective, accessible and adhered to by patients to improve their health and quality of life.

Georgia Malandraki

Language Learning and Meaning Acquisition (LLAMA) Lab

Because children have an exceptional ability to learn language, the LLAMA Lab explores the cognitive mechanisms that support this ability. Ongoing research topics include how children understand connections between word meanings, how general learning mechanisms and experience support speech comprehension and early markers of risk for poor language and reading outcomes.

Arielle Borovsky

Motor Speech Lab

The Motor Speech Lab covers a wide range of topics related to quality of life for older adults and individuals with Parkinson’s disease in an effort to treat the changes to speech and cognition that occur as a part of typical aging or as a result of aging-related diseases, such as Parkinson’s disease.

Psychoacoustics Lab

Research in the Psychoacoustics Lab focuses on behavioral measures of peripheral auditory processes in listeners with normal hearing and listeners with cochlear hearing impairment. We are particularly interested in studying dynamic adjustments in response to background noise and use models of auditory signal processing to connect behavior and physiology.

Elizabeth Strickland

Purdue Experimental Amplification Research (EAR) Lab

Research in the Experimental Amplification Research (EAR) Laboratory focuses on auditory processes that contribute to speech perception deficiencies in hearing-impaired listeners and hearing aid processing. Ongoing projects include work on frequency-lowering techniques, wide dynamic range compression and speech enhancement techniques.

Josh Alexander

Purdue Infant Speech Lab

The Purdue Infant Speech Lab explores how language comes to the child by focusing on whether measures of early speech perception, production and the input relate to later language in both typical development and in children at-risk for autism spectrum disorders.

Speech Perception and Cognitive Effort (SPACE) Lab

The Speech Perception and Cognitive Effort Lab focuses on the contribution of cognitive mechanisms — such as working memory and selective attention — to understanding speech in difficult circumstances. For example, this could include a talker with an unfamiliar accent or the presence of competing sounds. We use behavioral and psychophysiological measures to assess speech understanding, cognitive effort and stress in younger and older adults (with and without hearing impairment) under a range of listening conditions. Results of this research provide insight into the cognitive foundations of spoken language understanding and contribute to improving methods for the assessment and treatment of hearing impairment in older listeners.

Alex Francis

Sign Language Research Lab

Research in the Sign Language Research Lab uses theoretical and experimental methods to investigate aspects of sign languages and their similarities and differences to spoken languages. Results from this research are applied to improving deaf education and the quality of life of members of the deaf community. Projects include experimental studies of sign language structure, perception and production. Our work incorporates online questionnaires, psycholinguistic methods, motion capture analysis and neurolinguistics, as well as collaboration with engineers toward automatic recognition of sign language.

Ronnie Wilbur

Voice Lab — Sivasankar Research Group

The goal of the Sivasankar Research Group is to understand why some speakers experience voice disruptions related to prolonged speaking, aging, environmental exposures and disease. We utilize a multidisciplinary approach to understand the causes of voice problems in order to better prevent and treat this common communication disorder.

Preeti M. Sivasankar

Join Our Research Registry

SLHS researchers study how people of all ages hear, speak and understand language to advance our knowledge of how we communicate while also helping those who struggle in these areas.  

WHO CAN SIGN UP? Our studies include healthy adults and children of all ages as well as individuals with concerns or disorders affecting their ability to speak, hear, swallow or communicate. 

HOW DO I SIGN UP? Complete this linked questionnaire , where you will be asked to provide some basic information so we can contact you about eligible studies. By signing up for the registry, you are expressing your interest in learning more about research opportunities. When contacted, you can make a decision whether or not to participate. You can opt out at any time by emailing  [email protected]  or calling 765-494-4229.

Faculty by Research Area

Interdisciplinary training program in auditory neuroscience.

The Interdisciplinary Training Program in Auditory Neuroscience provides graduate student training and research experience to prepare students for independent research careers that can advance understanding of auditory system function using innovative tools and technologies. Graduates of this training program will develop creative solutions, devices and strategies to assist with and prevent hearing loss.

Communicative Disorders Training Program

The Communicative Disorders Training Program is designed to develop the research skills of clinicians and engage basic scientists in the study of communication disorders. Through this program, we can increase the proportion of PhD researchers who make substantive scientific contributions to knowledge of the causes, diagnosis and treatment of communication disorders.

• Apply to UBC
• Canadian high school students
• International high school students
• International Baccalaureate students
• Advanced Placement students
• University or college transfer students
• Mature students
• Academic English development programs
• UBC Conditional Admission Program
• UBC-Langara Indigenous Transfer Partnership
• Plan for UBC
• Choose what to study
• First-year credit
• Write your personal profile
• Complete the application
• Disability-related considerations
• Submit your documents
• First-year study options
• Accept your offer
• Apply for a study permit
• Plan your finances
• Find a place to live
• Take care of your health and wellness
• Explore your academic options
• Plan your degree
• Register for your courses
• Get ready for UBC life
• What we look for
• Dates and deadlines
• How UBC determines your tuition fees
• What will your first year cost?
• Beyond Tomorrow Scholars Program
• Indigenous student awards
• Loran Awards
• Presidential Scholars Awards
• Schulich Leader Scholarships
• UBC Centennial Scholars Entrance Awards
• International Scholars
• Loans for Canadian and US students
• Awards for students with disabilities
• Funding for former youth in care
• Funding for veterans
• Financial schedule
• Living in residence
• Living off campus
• Interdisciplinary and innovative programs
• Sports and recreation
• Work programs
• Leadership and volunteering
• Clubs and societies
• Orientations
• Students with disabilities
• Arts and culture
• Indigenous students
• Former youth in care
• UBC stories
• Admissions Blog
• Campus tours
• Information sessions
• Student experience events
• School visits
• Student meetings
• Self-guided virtual tours
• Create an account
• Get your Guide
• Ask UBC a question

Speech Sciences

Speech Sciences is an interdisciplinary program with courses from Linguistics, Psychology, and the School of Audiology and Speech Sciences. You'll study research methods, language structure, child development and language acquisition, anatomy and physiology, experimental psychology, and instrumental phonetics.

Program information

• Campus: Vancouver
• Length 4 yrs
• Co-op Yes You can combine your studies with full-time, paid work at top local and international organizations.
• Honours No You can study intense specialization in a single field.

The Speech Sciences program is designed to prepare you for graduate work in speech-language pathology or audiology. The program has an interdisciplinary structure administered by the Linguistics Department, with courses from Linguistics, Psychology, and the School of Audiology and Speech Sciences.

You will study research methods, language structure, child development and language acquisition, anatomy and physiology, experimental psychology, and instrumental phonetics. Coursework focusses on normal language. Work on language disorders is not generally undertaken until the graduate program.

Campus features

As an undergraduate student in the Speech Sciences program, you’ll have an opportunity to explore and gain experience in various research labs housed at the Faculty of Medicine’s School of Audiology and Speech Sciences, including the Adult Language Processing and Disorders Lab and the Child Language Lab.

Your future

Career opportunities vary widely across a range of fields including audiology, speech-language pathology, occupational therapy, physiotherapy, public health, computational linguistics, communications, education, and others.

There are many career paths that can combine your academics, skills, and experience with your different interests, including:

• Art or Music therapist
• Artificial intelligence designer
• Audiologist
• Audiometric technician
• Communication disorders assistant
• Communications manager
• Education consultant
• English-as-a-second-language teacher
• Human resources specialist
• Speech pathologist

Careers with Speech Sciences

Related programs

Program requirements.

• Canadian high schools
• International high schools
• International baccalaureate

English-language requirements

English is the language of instruction at UBC. All prospective students must demonstrate English-language competency prior to admission. There are numerous ways to meet the  English Language Admission Standard .

General admission requirements

IB Diploma Programme

• Completed IB Diploma, including at least three Higher Level courses.

IB Certificate Courses

• IB Certificate courses (Standard and Higher Level) may be used in an admissions average if you are graduating from a recognized high school curriculum that can be used as your basis of admission.
• IB Math Applications and Interpretations SL, or IB Math Studies, do not satisfy the math requirement for admission to UBC’s science-based programs, the Faculty of Management, the UBC Sauder School of Business, or the Vancouver School of Economics.

Degree-specific requirements: Arts

• No specific courses required beyond those needed for general admission

Related courses

The following subject categories are particularly relevant for this degree. Consider taking courses in these areas in your junior year and senior year.

• Language Arts
• Mathematics and Computation
• Second Languages
• Social Studies
• Visual and Performing Arts

You’ll find these pages helpful

Tour ubc’s okanagan and vancouver campuses from anywhere, meet professor schreyer, a linguist and anthropologist who helped develop superman's mother tongue, ready to choose your degree.

• MSc Program
• PhD Program
• Why Study with Us?
• Externship Schedules
• MSc Course Registration
• MSc Graduation Requirements
• PhD Requirements
• News & Events
• Academic Faculty
• Clinical Faculty
• Doctoral Students
• Emeriti and Previous Faculty
• In Memoriam
• Colloquium Series
• 50th Anniversary
• Clinical Educators
• Call for Research Participants
• Adult Language Processing & Disorders Lab
• Amplification Lab
• Auditory Perception & Speech Lab
• Bilingualism Research Group
• Child Language & Cognition Lab
• Child Phonology, Phonetics & Language Acquisition
• Language and Development Lab
• Living with Aphasia Lab
• Middle Ear Lab
• Pediatric Audiology Lab (coming soon)
• Swallowing Innovations Lab
• Professional Organizations
• Resources for Practitioners
• Resources for the Public
• Resources for Industry
• Aphasia Mentors Program
• SLP Referral Form
• SLP Clinic Fees
• COVID Safety Information – SLP Clinic
• Contact Us – SLP Clinic
• Request an Appointment Form
• COVID Safety Information – Audiology Clinic
• Contact Us – Audiology Clinic

The School of Audiology and Speech Sciences at UBC owes its reputation for excellence in research to the outstanding research programs of faculty members . SASS faculty are engaged in research in all areas of human communication and its disorders.

Student Involvement in Research

At SASS, student involvement in research is a crucial aspect of education. Graduate students and postdoctoral fellows are valued as key members of SASS research teams. The School’s faculty believe that educating independent, successful researchers requires providing not only strong guidance, but also enough freedom for student researchers to explore uncharted territory, and to discover new knowledge.

Research Settings and Partnerships

Research is conducted both in well-equipped university laboratories, and in clinical settings in the community. The School’s relationships with numerous clinical faculty members — located in schools, health units, private practices and hospitals across Canada — allow faculty members and students to work at the forefront of the field, implementing innovative, research-based solutions in clinical settings. Given the interdisciplinary and inter-professional nature of some of the School’s work, faculty members and students also collaborate with academics from other disciplines (e.g. linguistics, nursing, otolaryngology, psychology, and social work).

The School’s research activities are well supported by a variety of competitive private and public sources, including Tri-Council funding. Faculty members’ research findings are disseminated in a wide variety of forms to many different audiences, including academic, professional, and public audiences.

Speech, Language, and Hearing Science

Research in speech language and hearing sciences.

Aphasia Research and Treatment Lab (ARTlab)

Principal Investigator:  Maya Henry , PhD., CCC-SLP

Research in the lab is directed at improving our understanding of how the brain supports speech and language processes and how targeted treatment programs may improve communication impairments caused by stroke or neurodegenerative disease. The lab utilizes current approaches in cognitive neuroscience, neuroimaging, and cognitive rehabilitation to address these issues.

Arthur M. Blank Center for Stuttering Education and Research

Principal Investigator:  Courtney Byrd , Ph.D., CCC-SLP

The Arthur M. Blank Center for Education and Research generates new knowledge, trains the future, and provides innovative treatment that enables participants of all ages worldwide to stutter openly, speak confidently, communicate effectively, and advocate meaningfully.

Augmentative and Alternative Communication Laboratory

Principal Investigator:  Rajinder Koul , Ph.D.

The Augmentative and Alternative Communication Lab’s aim is to improve the efficacy of augmentative and alternative communication intervention for persons with severe speech and language impairment as a result of developmental and acquired conditions. We study the variables that influence the outcomes of such interventions, including symbol identification, the perception of synthetic speech, dynamic display configurations, and alternative access methods such as eye-tracking and brain-computer interface systems.

Central Sensory Processes Lab

Principal Investigator:  Julia Campbell , Ph.D.

Our research is focused on cortical plasticity, or the ability of the brain to adapt to changes in the environment.  Specifically, we are interested in how a typical brain with no sensory disorders might process various sensory input such as audition and vision over the lifespan, and how these cortical functions interact.

Children's Language, Literacy & Learning Lab (CL3 Lab)

Principal Investigator:  Mary Beth Schmitt , Ph.D., CCC-SLP

Dr. Schmitt’s research focuses on identifying active ingredients of language therapy that lead to improved language abilities for preschool and school-age children with language disorders. Much of her work happens in the field (i.e., in the schools). Currently, Dr. Schmitt and her team are investigating the role of children’s behavior regulation, peer effects, and service delivery models as they relate to children’s language development funded through NIH and local funding agencies.

Fostering Inclusive Education and Linguistic Diversity Lab (FIELD Lab)

Principal Investigator:  Chelsea Privette , Ph.D., CCC-SLP

The FIELD Lab focuses on developing culturally responsive assessment strategies and inclusive classroom practices for children with diverse language experiences. We are particularly interested in language development in environments where multiple languages and dialects are spoken. We use mixed methods to understand the experiences of children, families, and teachers from marginalized communities.

Hamilton Lab

Principal Investigator:  Liberty Hamilton , Ph.D.

The Hamilton lab investigates how the human brain processes speech sounds using intracranial electrocorticography (ECoG) recordings from patients with intractable epilepsy who are undergoing surgery to treat their epilepsy.  We use a combination of electrophysiology, behavior, neuroimaging, and computational modeling to ask how different features of sounds are combined to form the words that we speak and hear, and how this changes during development.

Multilingual Aphasia and Dementia Research Lab (MADRlab)

Principal Investigator:  Stephanie Grasso , PhD., CCC-SLP

Research in the lab focuses on developing treatment approaches for bilingual adults with aphasia and neurodegenerative disorders. In addition, we are interested in better characterizing the manifestation of aphasia and neurodegenerative disorders in individuals from culturally and linguistically diverse backgrounds. Research in the lab also addresses bilingualism as a contributor to cognitive and neural reserve and utilizes neuroimaging to investigate variability in treatment responsiveness. 

Principal Investigator:  Jesse Franco , Ph.D., CCC-SLP

Skills and Knowledge of Intervention for Language Learning Success (SKILLS) Lab explores evidence based interventions, such as parent-directed treatment, for children with Autism Spectrum Disorders and other Developmental Delays.

Speech Disorders and Technology Lab

Principal Investigator:  Jun Wang , Ph.D.

The lab is dedicated to develop assistive speech technologies including silent speech interfaces and to conduct basic speech science and disorders research on neurogenic motor speech disorders and neural processing for speech communication. 

Speech Psychophysics Laboratory

Principal Investigator:  Chang Liu , Ph.D.

The lab focuses on auditory processing of speech and non-speech sounds in a broad range of listeners including:

• Native and non-native English listeners,
• Listeners with hearing impairment, and
• Children with typical development and speech disorders.

We are also interested in:

• Speech acoustics for native and non-native speakers,
• Speech technology for speech enhancement, and
• Computational models for speech recognition.

Stuttering and Linguistic Processing Lab

Principal Investigator:  Zoi Gkalitsiou , Ph.D., CCC-SLP

The Stuttering and Linguistic Processing Lab’s research includes the investigation of linguistic and cognitive factors that contribute to the development and maintenance of stuttering using eye-tracking methodology, the manifestation of stuttering in bilingual speakers, and the improvement of evidence-based practices in stuttering.

Swallow Modulation Lab

Principle Investigator: Corrine Jones , Ph.D.

The Swallow Modulation Lab uses motor learning principles and sensitive tools to study swallowing physiology in health and neurodegeneration in order to identify opportunities for novel dysphagia rehabilitation models.

Texas Auditory Neuroscience Laboratory

Principal Investigator:  Spencer Smith , Ph.D., Au.D.

Research in the Texas Auditory Neuroscience Lab focuses on neural processes, from inner ear to cortex, involved in speech perception in noise.  We employ objective (otoacoustic emission and electroencephalography) and behavioral techniques to study relationships between neurophysiological function and perception.  Ongoing work in our lab is currently funded by the National Institutes of Health.

UT Care Lab

Principal Investigator:   Dr. Srikanta Mishra, Ph. D.

The UT CARE Lab develops scientifically rigorous tools and methods to improve speech understanding in noisy backgrounds for children with clinically normal audiograms. By integrating expertise from the fields of audiology, engineering, and neuroscience, we advance our understanding of speech perception mechanisms and enhance communication management. Utilizing advanced signal processing and acoustics techniques, we create intuitive tools to improve speech intelligibility and mitigate background noise effects. We also use demographic variables and psychoacoustic assessments to predict speech perception in noise among pediatrics. Our goal, through rigorous experimentation and data analysis, is to empower children with otherwise clinically normal hearing by providing evidence-based tools that optimize their ability to understand speech in challenging listening conditions.

UT Voice Lab

Principal Investigator:  Rosemary Lester-Smith , Ph.D.

We study voice production in individuals with neurological disorders, healthy speakers, and singers to identify factors that improve or impair vocal control. Our interdisciplinary research aims to advance assessment and treatment of neurological voice disorders.

Lab Archive

Principal Investigator:  Craig Champlin , Ph.D.

The aim of the Hearing Function Lab is to understand the workings of typical and atypical hearing systems. By understanding the underlying mechanisms of such systems, we are able to devise more sensitive tests used to characterize and track changes in hearing function over time.

Principal Investigator: Bharath Chandrasekaran, Ph.D.

In the SoundBrain lab, researchers study the sensory and cognitive processes that underlie speech and music perception. Using functional magnetic resonance imaging (fMRI), event-related potentials (ERPs), brainstem electrophysiology and behavioral methods, they study the representation of speech and music in the human brain, and how these representations are modified by listening experiences. The Sound Brain Laboratory is no longer active on UT campus.

Main Office CMA 4.114 Phone: (512) 471-4119 Email: [email protected]

• Engineering
• Write For Us
• Privacy Policy

Business Speech: Types with Examples, Informative, Special, Persuasive

Good presentation and speaking habits may be considered soft skills in the workplace or in any type of organization. Today in this article, we have shared what is business speech and how many types of business speeches are there.

Anybody can relate to all these types of business speech because these all are equally important in social life as well. So let’s start our topic with the basics of business speech.

► What is Business Speech?

Speech refers to that action when a person stands among a great number of people and starts delivering any kind of information or statement. It may be or may not be useful for the whole audience but most of the time it is valuable for them.

A speech that is delivered in the workplace or in any business organization for some specific purpose is known as Business Speech.

This is one of the forms of Business Communication and the audience has to sit quietly while the speech is being delivered. Most of the time audience knows very well that the speech must contain anything that will be beneficial for them.

► Types of Business Speech:

There are mainly three types of speech that are as follow;

• Informative Speech
• Persuasive Speech
• Special Occasion Speech

◉ Informative Speech

Informative business speech can be defined as speech that comprises the purpose to deliver useful information to the audience.

For Example  – In any organization, an Executive Coach or Trainer speaking about the new trends in the market to his trainees. It can be hard to understand for few trainees, but the fact is that he is delivering something informative that is beneficial for them.

Informative Speech is further divided into four types;

• Speeches about Objects
• Speeches about Events
• Speeches about Processes
• Speeches about Concepts

The following are known kinds of informative speech.

✔ Speeches about Objects :

It can be about any object related to that particular organization where the speech is being delivered.

For Example  – how various wildlife animals look, what is the smell of medicine, information about any product.

✔ Speeches about Events :

Those speeches that inform the audience about any events like historical incidents or about any situations are called speeches about the event.

For Example  – New President’s speech about future goals after the oath-taking ceremony.

✔ Speeches about Processes :

The main purpose of this type of informative speech is to inform the audience about anything which is currently happening or about how to do any particular task or work.

For Example  – a Yoga teacher explaining how to perform specific yoga poses.

✔ Speeches about Concepts :

Speeches about concepts are those speeches that inform the audience about any concept such as the peace of the world, freedom of rights, or love, fundamentals of any study topic.

For Example – a Science teacher explains Einstein’s theory of general relativity to his students in the class.

Must Read : Skills of HR Manager

◉ Persuasive Speech

Persuasive Speech refers to those speeches where the intention of the speech is to convince the audience to accept the particular opinion or fact and create influence on the audience to do anyhow.

In short, the speech which influences the listeners or audience to follow a certain idea is called a persuasive speech.

Persuasive speech is also an informative speech. because here speaker gives information in a lucrative manner to influence others.

For Example  –  in any debate, every person is try to persuade others to follow their given point of view. It is a form of persuasive speech.

In another example, During the advertising and promotional functions of any business, the sales manager or speaker uses his persuasion skills to influence the audience. Here the main purpose of speech is to change the thinking, beliefs, or behaviors of the audience towards his product.

Persuasive speech can be divided into three types that are as follows:

• Factual Persuasive Speech
• Value Persuasive Speech
• Policy Persuasive Speech

✔ Factual Persuasive Speech:

The Factual Persuasive Speech is such a speech that contains facts and it is based on a concrete proof about the certainty of anything that had happened.

The main purpose of this factual persuasive speech is to persuade the listeners whether the certain thing happened or not, exists or doesn’t exist.

For Example – If a student is giving a speech about the first man, who landed on the surface of the Moon. Nobody in the class knows whether it did happen or not, yet it possesses concrete proof.

✔ Value Persuasive Speech:

A Value Persuasive Speech is such a speech that tells the listeners about anything, whether it is wrong or right. The purpose of this speech is to challenge the ethical or moral aspects of a certain issue.

For Example –  If someone is giving a speech about capital punishment, whether it is moral or immoral, right or wrong, done or prevented. this type of speech is a value persuasive speech.

✔ Policy Persuasive Speech:

The policy persuasive speech refers to that speech where the speaker is trying to persuade the audience to either following a policy or rejecting it. It is not limited to just a policy, but it can be about accepting or rejecting a rule or a candidate is also a policy persuasive speech.

For Example – Suppose If the President of a country is not satisfied with the present foreign policy and wants to change it. The president gives a speech to higher authorities for convincing them to change the current foreign policy and support the new policy then it is known as policy persuasive speech.

Must Read : Types of Communication

◉ Special Occasion Speech:

Special Occasion speech refers to that speech which is given on the special occasion like;  A speech of farewell allows someone to say good-bye to one part of his or her life as he or she is moving on to the next part of life. Maybe you’ve accepted a new job and are leaving your current job.

Special occasion business speech is something which anyone can face at some point in their lives.

For example –  If your company won an award of the year for excellence. And you are receiving that award on the behalf of your company. The speech given by you after getting the award can be considered as a special occasion business speech.

In another example, If you are getting retirement from your job and want to thank your subordinates, superiors, and top management at the farewell party.

University of South Florida

Department of Communication Sciences & Disorders

College of Behavioral and Community Sciences

Main Navigation

Research facilities - auditory & speech sciences laboratory.

The Auditory & Speech Sciences Laboratory emphasizes basic and applied research to solve communication problems through the combination of behavioral sciences, electrophysiology, and assistive technology. Current projects are funding by NIH NIA, NIH NIDCD, NSF, Industry, and USF.

Visit the ASSL website...

Locations: The facility consists of two spaces. One ~1500 square feet facility is in Suite 210 Business Partnership Building (in the USF Research Park). The second facility is in room 3008A on the third floor within the Department of Communication Sciences & Disorders. 

Auditory & Speech Perception Research: The laboratory includes seven listening stations in five sound booths and a developmental station. Each research station is equipped with sound generation equipment (eight Tucker-Davis Technology (TDT) hardware/software systems; a variety of headphones (Etymotic ER-1, ER-2, and ER-3A insert phones; TDH 50P supra-aural earphones; Sennheiser HDA-200; HD-265; HD-250 circumaural headphones; KEF Q100, Behringer Truth B2010A, and an assortment of other loudspeakers). Each system has multiple human subject interfaces (touchscreen monitors, button response boxes, PC keyboards and PC mice). For added flexibility, the lab has 8 and 24-channel MOTU (Moon Over The Unicorn) external sound cards.

Spatial Hearing Systems: The laboratory features a 24-loudspeaker array, head-tracking system, and variable-degree reverberation sound treated booth for advanced spatial hearing research. The laboratory has extensive custom and commercial software for experimental design and execution including TDT SykofizX and a series of custom interfaces designed in MATLAB. We also use virtual spatial environments based on head-related transfer functions (HRTFs) and headphone sound presentation.

Auditory Electrophysiology Research: Both facilities house equipment for multi-channel auditory evoked/event-related potential recordings and corresponding brain mapping software including 4 and 16 channel systems from Tucker-Davis Technologies (TDT) and a 64-channel system by Advanced Neuro Technology (ANT Neuro). These systems include custom MATLAB-based software for stimulus generation, presentation, hardware triggering, and simultaneous behavioral response collection integrated with ASA-LabTM via the ASA ExMan feature. We also have an extensive set of image processing software for data analysis and interpretation including ASA-LabTM, ASA-ProTM, and ExMan from ANT-Neuro and MATLAB-based EEGLab and ERPLab.

Diagnostic & Research Audiology: Both facilities house complete audiometric equipment, including audiometers, middle ear analyzers, an otoacoustic emission systems, evoked potential systems, cerumen management equipment, and an array of speech audiometry materials. The laboratory also has a series of special audiometric tests including the HINT, the LISN-S, tests of auditory processing, and the NIH Toolbox Auditory Assessment (developed in house).

Hearing Enhancement Device Research & Development: The facilities house a complete hearing aid dispensary for research including earmold impression and modification equipment, hearing aid fitting software, wireless instrument interfaces (e.g., NoahLinkTM, HiPro, and proprietary interfaces) hearing aid testing equipment (Audioscan VF-1 and VF-2), and an assortment of experimental and commercial hearing instruments as well as prototype hearing enhancement devices.

Acoustic Analysis and Measurement: The facilities include a variety of equipment for acoustic signal measurement and analyses from leading manufacturers (B&K, Larson-Davis, G.R.A.S., Quest, Fluke, Tektronix, B&K Instruments) including pressure and free-field microphones, calibrators, sound level meters, couplers, ear-simulators, KEMAR acoustic manikin, probe-microphone systems (Etymotic ER-7Cs), and signal conditioners. The laboratory also has acoustic analysis software including EASERATM, Adobe Audition, B&K PulseTM, and custom MATLAB analysis software.

Database Utilities: The laboratory features three SQL databases to support ongoing and future research. The first is a large database of over 1400 individuals previously characterized in terms of audiometric status, medical and hearing history, and perceptual abilities that is available for retrospective research. The second is a large and growing database with similar information from subjects who are current or potential participants. The third is a robust resource sharing database for recruiting subjects and scheduling subject visits with specific personnel and specific laboratory resources (e.g., listening stations) for a variety of research projects. Recruitment may also draw from the USF Audiology clinic and the 2.4 million people of the greater Tampa area.

An official website of the United States government

The .gov means it’s official. Federal government websites often end in .gov or .mil. Before sharing sensitive information, make sure you’re on a federal government site.

The site is secure. The https:// ensures that you are connecting to the official website and that any information you provide is encrypted and transmitted securely.

• Publications
• Account settings

Preview improvements coming to the PMC website in October 2024. Learn More or Try it out now .

• Advanced Search
• Journal List
• J Speech Lang Hear Res

Essential Statistical Concepts for Research in Speech, Language, and Hearing Sciences

Jacob j. oleson.

a Department of Biostatistics, University of Iowa, Iowa City

Grant D. Brown

Ryan mccreery.

b Boys Town National Research Hospital, Omaha, NE

Associated Data

Clinicians depend on the accuracy of research in the speech, language, and hearing sciences to improve assessment and treatment of patients with communication disorders. Although this work has contributed to great advances in clinical care, common statistical misconceptions remain, which deserve closer inspection in the field. Challenges in applying and interpreting traditional statistical methods with behavioral data from humans have led to difficulties with replication and reproducibility in other allied scientific fields, including psychology and medicine. The importance of research in our fields of study for advancing science and clinical care for our patients means that the choices of statistical methods can have far-reaching, real-world implications.

The goal of this article is to provide an overview of fundamental statistical concepts and methods that are used in the speech, language, and hearing sciences.

We reintroduce basic statistical terms such as the p value and effect size, as well as recommended procedures for model selection and multiple comparisons.

Conclusions

Research in the speech, language, and hearing sciences can have a profound positive impact on the lives of individuals with communication disorders, but the validity of scientific findings in our fields is enhanced when data are analyzed using sound statistical methods. Misunderstanding or misinterpretation of basic statistical principles may erode public trust in research findings. Recommendations for practices that can help minimize the likelihood of errors in statistical inference are provided.

Supplemental Material

https://doi.org/10.23641/asha.7849223

Statistical techniques are not magical or mysterious; rather, they are tools designed to quantify scientific evidence in various ways. Each method is built upon a mathematical foundation and has well-defined appropriate uses and requirements. Understanding more about these foundations, as well as the assumptions made by statistical procedures, can help investigators to adopt the most appropriate statistical method for the problem at hand, leading to more reliable and replicable results. Most traditional statistical methods follow the frequentist philosophy, in which models are fit via maximum likelihood or ordinary least squares. There are, of course, alternate perspectives on statistical inference, including Bayesian statistics and algorithmic modeling/machine learning. In addition, there are many techniques in nonparametric inference and variations on likelihoods (e.g., partial likelihood, semipartial likelihood, quasilikelihood, pseudolikelihood). Our focus in this article, however, is primarily on frequentist hypothesis testing using maximum likelihood, which is at the core of most applied science, as well as introductory statistics curricula. We expect that readers of this article have had at least one basic course in statistics. Researchers in speech, language, and hearing sciences are used to designing studies to learn about specific topics of interest, and many do perform their own statistical analyses to answer their hypothesis questions. Therefore, our goal is not to review all of the basics of statistics; rather, our goal is to highlight common errors and misconceptions in statistical approaches that can help scientists to avoid common statistical pitfalls in their research. In a companion article, some advanced methods for analyzing data in speech, language, and hearing sciences will be highlighted ( Oleson, Brown, & McCreery, 2019 ).

The use of statistical methods with underlying assumptions that do not match the data can have potentially serious consequences for the accuracy and reproducibility of scientific results. A recent analysis in major behavioral psychology journals indicated that approximately half of articles published between 1985 and 2013 contained at least one statistical error, and around 12% of published articles contained a statistical error that would have altered key findings of the study ( Nuijten, Hartgerink, van Assen, Epskamp, & Wicherts, 2016 ). Widespread findings of errors in statistical reporting and interpretation are believed to have contributed to an inability to replicate key scientific findings from the literature in psychology ( Pashler & Wagenmakers, 2012 ) and medicine ( J. P. Loannidis, 2005 ). Problems with replication of statistical results erode public trust in science and can reduce the impact of scientific findings. Fortunately, many current problems related to the lack of transparency and reproducibility in scientific research can be resolved through increasing statistical proficiency of scientists and promoting open and transparent practices in the sharing of the code used for statistical analyses and data ( Peng, 2015 ). To be clear, clinical and scientific experts do not need to become statistical experts, but they should recognize the statistical principles involved in their study design and their statistical analysis plan. Moreover, they should seek out collaborations with statistical experts to be involved with the study design and the statistical analyses. Study teams should include statistical expertise early in the development process to help design the study, to set up a clear and appropriate analysis plan, and to ensure appropriately analyzed and presented results.

Consider a research study in which we want to compare differences in the mean speech perception for children with hearing loss to children without hearing loss. There are multiple statistical approaches that could be applied to analyze differences between groups. Although scientists in nearly every discipline are familiar with the use of a two-sample Student's t test in this situation, confusion may arise when the assumptions of independence, normality, and equal variances are violated. An investigator must decide whether to immediately jump to a nonparametric analysis, implement a Welch adjustment for unequal variances, employ a repeated-measures analysis of variance (ANOVA), or transform the outcome variable, among many other options. All of these possible alternatives may produce the same conclusion, meaning the statistical approach we choose is inconsequential in many cases. On the other hand, we might obtain wildly different results depending on our choice of statistical method; understanding why the violation of assumptions can alter the behavior of statistical procedures is critically important to good statistical practice and making inferences based on the results of statistical analyses.

In most cases, there is not a single statistical approach that must be applied to solve a given research question, even when there is a clear tradition or common practice. Importantly, every statistical method comes with advantages and limitations. Practitioners should be familiar with the benefits and drawbacks of the procedures they employ and use that knowledge to choose methods that credibly answer the research question of interest. Our goal is not to provide detailed mathematical explanations of these tradeoffs but rather to focus on intuitive and practical recommendations when performing analyses, reporting results, and interpreting findings. We begin by discussing significance reporting and introducing the basic ideas behind hypothesis testing. We next discuss the ubiquitous use of the p value and what every user of a p value should know. The discussion of statistical significance via the p value is followed by measures of clinical significance through the appropriate use of effect sizes. We conclude by offering thoughts on regression analysis, model selection, and multiple comparisons.

Statistical and Clinical Significance

Significance reporting.

Research findings are often reported in terms of statistical and clinical significance, but it is very easy for both the producers and consumers of research to misuse or misinterpret these tools. Statistical and clinical (or practical) significance provide different pieces of important information. Ideally, both statistical and clinical significance should be described in reports and analytical results. Recall that statistical tests are generally devised as a choice between the null hypothesis (e.g., the average speech perception score for children with hearing loss is the same as the average speech perception score for children with normal hearing) and an alternative hypothesis (e.g., the average speech perception scores differ between children with and without hearing loss). To test a hypothesis is to ask whether we have enough scientific evidence to “reject the null hypothesis” and conclude that there is enough scientific evidence to conclude that the alternative hypothesis must be true (e.g., the speech perception means for children with hearing impairment and children with normal hearing are different). When we “do not reject the null hypothesis,” we write that there is not enough evidence to conclude that the group means differ; we do not make positive conclusions (e.g., “the two group means are actually equal”). The phrase “statistical significance” simply indicates that we have satisfied a prespecified rule (how far apart the means are relative to the standard error), which allows us to reject the null hypothesis in favor of the alternative hypothesis. This is most commonly assessed using p values, which we discuss in more detail in the p Values section. The result of our reject or do-not-reject decision can be either correct or incorrect. If we reject the null hypothesis in favor of the alternative hypothesis, but the null hypothesis is actually true, then we have made a Type I error. For example, a Type I error occurs when we claim two groups have different means when the means are in fact equal. If we do not reject the null hypothesis, but the alternative hypothesis is the true one, then we have made a Type II error. This transpires when we do not claim that two group means are different when they really are different.

Clinicians and researchers should recognize that statistically significant findings are not always practically or clinically significant. Statistical and clinical significance are often decoupled for large data sets because p values are strongly affected by the sample size. For example, we may find that the mean ages between two participant groups are significantly different statistically but that the observed mean difference is only 0.08 years (1 month). Unless this difference occurs at a point in development where the outcome of interest is likely to change rapidly over a short period, an age difference of 1 month between groups is unlikely to be clinically important. Reporting just the p value without highlighting the practical implications of the 1-month difference could be misleading.

Confidence intervals can help to contextualize significance tests and effects. Confidence intervals give a range of plausible values for a particular effect of interest. As such, they can be used to test hypotheses by checking whether the lower and upper bounds contain the hypothesized value of the quantity of interest. In addition, confidence intervals can also give an impression of clinical significance, which is discussed in more detail below in the Effect Sizes section in the context of “effect size.” Briefly, an effect size can be as simple as reporting parameter estimates such as the sample mean, the mean difference between two groups, or the slope estimate in a regression analysis. Confidence intervals are subject to the same Type I and Type II error rates as significance testing more broadly, and the specific intervals given are often misinterpreted by researchers. Nevertheless, confidence intervals provide additional clinical effect information beyond the simple statistical significance of a finding and are discussed in more detail in the Effect Sizes section.

Unlike p values, the practical size of an effect does not change based on a sample size and provides a crucial piece of information that should be reported. In general, more time and space in scientific reports should be devoted to the discussion of the practical significance and scientific impact of detected effects. To be most impactful, scientific publications should convey both the evidence of an effect (statistical significance) and the practical impact of the detected effect (clinical significance). In addition, because different scientific audiences may have different tolerances for Type I errors, we recommend that researchers report actual p values, rather than just whether or not an effect is significant. Whereas preference or convention leads many researchers to be satisfied with a 5% Type I error rate ( p < .05), others may prefer 1% or 0.1%, especially if the consequences of such an error would have a major impact on the field or alter an established scientific premise or clinical practice. In the next two subsections, we discuss p values and effect sizes in greater detail.

Although every introductory statistics class requires students to memorize their definition, p values have long been misunderstood and are often misused, misinterpreted, and increasingly criticized. At least one journal (Basic and Applied Social Psychology) has even taken the step of banning p values ( Trafimow & Marks, 2015 ), and another has suggested using a more conservative p -value threshold to determine statistical significance to .005 ( J. P. A. Loannidis, 2018 ). These controversies have led to ongoing debate and discussion across a wide range of scientific fields about the appropriateness of p values as the main criterion for judgments about statistical significance. To shed more light on the proper use and interpretation of p values, the American Statistical Association released a statement in 2016 ( Wasserstein & Lazar, 2016 ) with six principles to consider for the proper use and interpretation of the p value. They are the following:

• p values can indicate how incompatible the data are with a specified statistical model.
• p values do not measure the probability that the studied hypothesis is true or the probability that the data were produced by random chance alone.
• Scientific conclusions and business or policy decisions should not be based only on whether a p value passes a specific threshold.
• Proper inference requires full reporting and transparency.
• A p value, or statistical significance, does not measure the size of an effect or the importance of a result.
• By itself, a p value does not provide a good measure of evidence regarding a model or hypothesis.

As we tackle the question of the usefulness of the p value, we need to understand what it is and what it is not. By definition, the p value is the probability of observing a test statistic, which is as extreme as or more extreme than that which was observed, assuming that the null hypothesis is true. If this probability is sufficiently low (e.g., below an acceptable Type I error rate) for the purposes of a given study, then we consider this sufficient evidence that the null hypothesis is likely not true and that we may favor the alternative hypothesis. The p value is not a tool for deciding which of two competing hypotheses is more likely, given the observed data.

In the frequentist statistical paradigm (the most common approach to statistics and data analysis), the hypothesis is not random, so we do not assign a measure of probability directly to it. From this perspective, the actual hypothesis is either true or false, and this fact does not change simply due to our inability to know the truth with certainty (e.g., the two means are equal, or they are not equal). Instead, we measure the likelihood that the data could have arisen if the null hypothesis were true. In order to accomplish this, we assume that the null hypothesis is true (e.g., the mean speech perception rates are equal for children with hearing impairment and children with normal hearing) and then define a rule to reject the said hypothesis if the observed data would be sufficiently improbable if that hypothesis were indeed true (e.g., these means are more than 2 SD s apart). If the observed data would have been unlikely to have arisen under the conditions specified in the null hypothesis (e.g., the sample means for our two groups are too far apart for them to have come from two distributions with the same mean), we consider that evidence against the null hypothesis. This process does not allow for a probability comparing the two competing hypotheses.

In some settings, researchers do desire a method to compare the likelihood of competing hypotheses. In such settings, Bayesian statistical approaches may be more useful. In the Bayesian paradigm, one can evaluate the probability of the null hypothesis being true versus the probability of the alternative hypothesis being true, because the Bayesian interpretation of probability is fundamentally different from the frequentist interpretation. Broadly speaking, Bayesian probability is a tool to quantify knowledge and uncertainty. Bayesian statistics uses probability to describe what is known before data are collected and to update that knowledge based on how much is learned after collecting data. More detail on Bayesian statistics is given in this issue in Oleson et al. (2019) and McMillan and Cannon (2019) .

p values suffer from other drawbacks besides the confusing definition. Consider a two-sample problem to compare two group means such as testing whether the mean speech perception score of the population that is hard of hearing is significantly different from the mean speech perception score of the population with normal hearing. Although the null assumption that population means are equal is statistically useful, in real-world settings where we cannot randomly assign group membership, we know a priori that the two population group means are not the exact same value. In our example of speech perception scores, it seems improbable that a language outcome score for the group with hearing impairment has the same population mean as the group with normal hearing. Our primary interest is to determine whether a statistically detectable and clinically relevant difference exists. Even in experimental settings, if there is a difference worth testing, there is often a good reason to suspect that at least an infinitesimal difference between group means exists. Nevertheless, we assume that the population means are equal in a null hypothesis. In order to demonstrate evidence of a difference between them, we must first naively assume that there is no difference. Whenever the population means are even slightly different, we know that we can make a p value small enough to declare statistical significance simply by choosing a large enough sample size (e.g., a high enough power). Indeed, researchers who fail to reject null hypotheses are often advised to increase their sample size. This practice is not statistically appropriate in part because it does inflate the Type I error. In this sense, p values by themselves often tell us more about sample size than anything practically meaningful about the size of the effect. Moreover, p values are constructed under the statistical model that we assumed: Any departures from the assumptions of that model impact the reliability of the p value.

If we use common, incorrect statistical practices in speech, language, and hearing science research, including sampling from finite populations without correction, obtaining correlated samples that are not accounted for, or sampling from populations with different distributions from those assumed by particular statistical procedures, the applicability and reliability of p values are greatly diminished. Therefore, the statistical analysis that is used should be carefully determined to appropriately reflect the problem at hand and that the relevant assumptions should always be considered and checked prior to making inferences based on statistical results.

Effect Sizes

An “effect” or “effect size” is a measurement of a phenomenon of interest, whether an expected change in an outcome in response to a treatment, a difference between group means, or a ratio of the odds of an outcome in two groups. Importantly, the specific effect size that is chosen should provide insight concerning a meaningful study question.

Effect sizes can be subdivided into two groups: “standardized” or “relative” effect sizes and “unstandardized” or “absolute” effect sizes. Unfortunately, which kind of effect is being presented in a given analysis is not always clear. Cohen's d is a commonly used relative effect size, whereas our previous example of 1-month difference in age is an example of an absolute effect size. Many formulas for the power of statistical procedures and sample size of study designs are given in terms of Cohen's d , because the power to reject a null hypothesis depends on the relative extremeness of the alternative hypothesis; Cohen's d gives a standardized measure of how far an estimated effect is from a hypothesized value, relative to its standard error. This gives a convenient scale on which to compare and interpret mean differences but suffers from the same problem as p values: Cohen's d does not measure the practical significance of a difference but is instead more closely related to statistical significance. Using our previous example, if we detect a statistically significant difference in mean age between two groups, we can equivalently expect that the difference in means is large relative to its standard error. Lenth (2012) provides an excellent summary of the shortcomings of comparisons of relative effect sizes. A clinically meaningful effect is typically better evaluated by examining the magnitude of the absolute mean difference.

Absolute effect sizes for hypothesis tests comparing means are simply the difference between the means of interest. Statistical software for most statistical modeling frameworks will output absolute effect sizes based simply on the parameters reported by statistical software. Obtaining relative effect sizes in the style of Cohen's d in modern statistical models can be somewhat more complex given the potential difficulty of estimating the standard error of the difference between groups, which is the denominator term in the calculation of Cohen's d . Although methods exist to compute relative effect sizes for more complicated models ( Brysbaert & Stevens, 2018 ; Selya, Rose, Dierker, Hedeker, & Mermelstein, 2012 ; Westfall, Kenny, & Judd, 2014 ), we generally prefer to rely on more interpretable absolute effect sizes.

Confidence intervals provide another perspective on absolute effect sizes. Although they suffer from some of the same interpretability challenges as the p value, they do give us more direct information on the range of plausible values for the parameter of interest. Consider the difference between speech perception scores between two groups again. For two independent groups, a two-sample independent groups confidence interval is typically computed using the difference in the means and a weighted average of the variances of the two groups (a pooled variance). Note that we say typically because we are assuming equal variances between the two groups and that the difference in means approximately follows a normal distribution, and any deviations from those assumptions would require alternate methods. The confidence interval is then centered at the mean difference, and we add and subtract the standard error (square root of the pooled variance) times our critical value. The result is an upper limit and a lower limit of confidence of where we believe the true population mean difference might be. The interpretation challenges come from the fact that the performance guarantees (confidence levels) concern the repeated process of how confidence intervals are constructed in general, rather than the specific numbers produced in one single statistical analysis. For example, in a study where the mean difference in language standard scores between two groups of children is 4.51, with a valid 95% confidence interval of [3.30, 5.72], we have confidence that intervals constructed in this way will contain the true language standard score 95% of the time. No such statement can be made about the specific interval of 3.30–5.72. Even so, this range does give plausible values that are consistent with the observed data—a valuable addition to the practical understanding of the absolute size of an effect.

When reporting study results, it is important to report measures representing both the statistical significance and clinical significance of the study. Absolute effect sizes generally carry the most information regarding clinical importance for the study and should be reported. Generally, this should be in terms of the parameter estimate and the confidence interval for the estimate. Relative effect sizes can also be informative in some situations but carry additional complexity and can conflate statistical and practical significance for practitioners and readers.

Regression Topics

Regression and anova.

A common statistical technique employed in the field is regression analysis. In its most basic form, a regression analysis includes one dependent variable that is related to the outcome via a line, although regression models can be made much more complex. The goal is generally to test whether all or some of the independent/explanatory variables in the model are related to the dependent/outcome variable. Those independent variables may be continuous or categorical. In the event that there is a single categorical independent variable, then the resulting regression model is commonly referred to as ANOVA . Although classical regression and ANOVA were developed separately, ANOVA is simply a special case of regression, where regression can accommodate more complicated relationships between the dependent variable and the independent variables.

In a regression-type model, the slope parameter estimates themselves provide a measure of effect size. In cases where multiple groups are present (such as in ANOVA or regression with categorical variables), the effect of interest may be given by the estimate of a contrast or a single parameter estimate comparing one group to a reference group. See Bring (1994) for a full discussion of regression effect sizes. In Walker et al. (2014) , regression models were used to investigate relationships among predictor variables and service delivery for a group of children who were hard of hearing. The independent variables were gender, test site, maternal education level, immediate family history of hearing loss, and degree of hearing loss measured by better-ear pure-tone average. They found that, after controlling for all of the variables, only degree of hearing loss was significantly related to the dependent variables age at first diagnostic evaluation (β = −0.36, p = .003), age at hearing loss confirmation (β = −0.42, p = .001), and age at hearing aid fitting (β = −0.37, p = .011). By reporting the β (slope) values, we can immediately determine how better-ear pure-tone average is related to each of these age-related outcome variables. They also found that only gender was significantly related to length of the delay between hearing loss confirmation and enrolling in early intervention (β = −3.34, p = .024). Through reporting of the β value as an effect size, we immediately know that girls had an estimated delay that was 3.34 months shorter than boys, and we can readily assess the clinical impact those months will have on the children.

In ANOVA, we are often confronted with the need to adjust our outcome variable of interest by a covariate that is continuous. A classical approach to performing this adjustment is analysis of covariance (ANCOVA). Recall that ANOVA is designed to compare group means by measuring the between-group variance relative to the within-group variance. ANCOVA is a method that mathematically adjusts the group means to take into account the values of the covariate (e.g., age) and then performs ANOVA on the covariate-adjusted means. However, just as ANOVA is equivalent to a regression analysis that includes only a categorical variable, ANCOVA simply adds a continuous variable to be adjusted for to that regression model. For example, an ANCOVA that examines differences in speech recognition for children who are hard of hearing and children with normal hearing adjusting for age is the same as a linear regression analysis that includes age and hearing status (normal hearing or hard of hearing) as predictor variables. General linear hypothesis tests can still be performed within the regression model framework. General linear hypothesis tests include what many in the field call post hoc tests. Technically, post hoc tests are those tests, usually pairwise comparisons, that are only considered after the results of the global tests are found. If these pairwise comparisons are planned from the beginning of the study, then they are not technically post hoc tests.

Moreover, regression models and their generalizations offer a great deal of flexibility to include additional covariates, add random effects, modify the residual error structure, do nonlinear transformations, and more. The addition of random effects to a regression model to create subject-specific curves and account for within-subject correlation is referred to as linear mixed-effects regression models or linear mixed models. More details on mixed models can be read in our companion article ( Oleson et al., 2019 ). Although ANOVA and ANCOVA are familiar statistical procedures with a lot of history, it is preferable to use the regression framework for the task of comparing group means.

Model Selection

Often, we encounter the situation of deciding what independent variables to include in our regression model. This process of deciding which variables to include, and which to not include, is known as model selection . For example, Walker et al. (2014) were interested in predictors of the dependent variable age at hearing aid fitting. The independent variables were gender, site at testing, maternal education level, immediate family history of hearing loss, and degree of hearing loss. The variables included in the final model were included because of specific hypotheses about the factors that influence hearing aid use from theory and the previous literature on hearing aid use in children. Model selection refers to deciding what subset of the full list of independent variables should be included in the final model.

Model selection is broader than just regression and arises in all forms of statistical inference. Investigators must identify a model that can address the research questions of interest using the variables that were collected as data. Ideally, scientific theory should be the foundation for the process of model selection, comparing among a small set of scientifically plausible models. Some approaches to model selection are based on the principle of parsimony: The best statistical model is that which includes all the essential variables and nothing more. In practice, however, the most parsimonious model for a specific research question can be difficult to identify. In many exploratory studies, we gather data without a full understanding of what the important relationships are, which covariates need to be adjusted for, and what patterns of correlation in longitudinal studies are likely to be appropriate. Although model selection is a large topic area and an active field of statistical research, common techniques in a frequentist setting can be broadly divided into two categories: algorithmic approaches and researcher-driven exploration. The process of model selection is crucially important and can impact the outcome and reliability of hypothesis testing procedures, as well as the inferences and conclusions of a research study.

At the core of any model selection procedure is the ability to compare selected models of interest to determine which better fits the data. This comparison can be accomplished for nested models in a regression setting using F tests, which test the null hypothesis that the collection of variables to be added to the model does not explain a significant amount of variability in the outcome above and beyond the currently included variables. More generally, we can compare frequentist models using information criteria such as the Akaike information criterion (AIC). Models with smaller AIC values are considered to have better fit to the data. Algorithmic model selection approaches build on these concepts by sequentially considering modifications to a model. For example, forward selection starts with a base model and proceeds to consider adding terms to a model. Terms are added at each step if they meet an inclusion threshold, which can be based, for example, on p values or information criteria. Similarly, backward selection starts with a large model and considers terms for removal. Researcher-driven exploration may use some of the same tools to compare models, but decisions on what models to explore and select are made by the researcher. This process of model selection may be informed by scientific knowledge, formal model comparisons, or conscious/subconscious preference for certain model features or results.

Although it is clear why informal exploratory analysis can lead to problems with inflated Type I error rates and reproducibility, many practitioners are surprised to discover that formal techniques such as stepwise model selection are similarly problematic. In each case, the output generated by statistical software after a model selection routine is performed is indistinguishable from output obtained from a prespecified analysis. For most statistical procedures, the reported results are based on the assumption that data were collected according to a simple random sample from a large population, and then a single model was fit. These assumptions are not met when data-driven model decisions are made, and these assumption violations can endanger the validity of statistical inference.

To illustrate this point, we highlight several simulation results from a regression example. Replicate data sets (50,000) are simulated with 10 hypothetical covariates and a normally distributed outcome measure and are generated under the global null hypotheses that all of the regression coefficients are zero. Each data set has n = 100 observations. Two primary hypothesis tests are of interest. First, we consider the setting where one covariate addresses a primary study question, and the other nine covariates are included to adjust for potential confounding. In the second, we consider the overall F test for the (true) null hypothesis that all of the covariate beta coefficients equal zero. For prespecified analyses conducted without model selection, we have theoretical guarantees that the Type I error rate will not exceed whatever nominal threshold we specify.

For a chosen Type I error rate of 0.05, we observe that choosing large or small models presents no problems in the absence of model selection. In line with our theoretical expectations, the observed Type I error rate for the analysis of just the single variable of interest produced a Type I error rate of 0.0502, whereas an analysis of all 10 covariates with pairwise interactions produced a comparable Type I error rate for this comparison of 0.0496. The overall F -test results were similar.

In the presence of forward stepwise model selection based on AIC, however, the hypothesis test concerning our primary variable of interest had a Type I error rate of 0.0645, and the overall F test had a Type I error rate of 0.4314. Therefore, just by applying basic model selection algorithms, we are increasing our probability of incorrectly rejecting the null hypothesis. Although these results are quite troubling, especially for overall F tests, the situation in real analyses may be even worse, as actual data tend to exhibit correlation among the independent variables or various types of confounding, driving up variability of parameter estimates and the likelihood of incorrect conclusions. With this in mind, it is incumbent upon practitioners to clearly delineate between exploratory analyses and confirmatory analyses in scientific reports and to clearly describe any model selection procedures that were employed in a study. By default, p values only attain their advertised performance for prespecified models. Code and complete results of the simulation are provided in Supplemental Material S1 .

Multiple Comparisons

Statistical models may require comparisons of several effects within an overall model. Traditionally, scientists have been trained to adjust their statistical assumptions to be more conservative to account for multiple comparisons to avoid the increasing risk of statistical errors as the number of comparisons increases (e.g., Aickin & Gensler, 1996 ). However, there are also costs to many approaches to accounting for multiple comparisons, including reducing statistical power. There are different schools of thought when it comes to multiple comparisons and many relevant summary articles. See overviews by Saville (1990) , Bender and Lange (2001) , or Cao and Zhang (2014) for more in-depth discussions of this issue. Rather than provide a review of the myriad methods of accounting for multiple comparisons in statistical models, we lay out some general points to consider regarding the topic.

The primary reason that researchers are advised to do a multiple-comparisons adjustment is to strictly control the overall (familywise) Type I error rate. In summarizing the alternative approaches for adjusting for multiple comparisons, we will consider three different general approaches. We could (a) perform no adjustment and accept individual Type I error rates, (b) adjust our alpha level to preserve a predetermined familywise error rate, or (c) adjust the alpha level to allow for a specified acceptable false-positive error rate.

The first approach to multiple testing is applicable when an alpha level adjustment may not be required. Bender and Lange (2001) argue that multiple comparisons should be used in confirmatory studies for the primary outcome of interest and that they are not necessarily required for exploratory studies. The researcher should clearly define the primary outcome and identify which comparisons correspond to that outcome. This may be a small subset of analyses that are performed, where the rest of the analyses are secondary and perhaps dependent upon the results of the primary question. If we have a truly controlled and confirmatory analysis, then we do want to reasonably control this alpha level. There is also the question of making Type II errors where we do not detect an important difference. In speech, language, and hearing studies, where sample sizes tend to be small, the alpha level adjustments reduce the significance level, which also increases the Type II error. It is critical to report absolute effect sizes in these situations.

The most common approach implemented for multiple comparisons is to adjust the alpha level to preserve a predetermined familywise Type I error rate. The Bonferroni adjustment is the most common technique used. For all pairwise comparisons, some will choose a Tukey–Kramer adjustment, but there are many more ways of adjusting alpha to control familywise error rates. Although it is clear that these adjustments are conservative, it may not be as clear the assumptions that are being made by such an adjustment. In an adjustment to control the overall Type I error, a critical assumption is that all null hypotheses are correct and that we want to jointly make only a 5% chance of falsely rejecting at least one of those. However, we happen to know that it is highly unlikely that all null hypotheses are true, as we outlined in the Significance Reporting section. These adjustment techniques make the most sense in a highly controlled confirmatory study with a single outcome of interest. In many studies, especially those that include various covariates (e.g., age, hearing loss severity, gender), we do expect many of the null hypotheses to be false and would be surprised not to find significant evidence against them. For example, a study in which age is adjusted for as a known confounder is expected to find a significant effect due to age. As an alternative example, a study with a placebo arm, a known effective treatment arm, and a novel treatment would be expected to reject the null hypothesis that the existing treatment and placebo produce the same mean. Assuming the global null hypothesis as a basis for Type 1 error rate correction is often unrealistic and unnecessarily conservative.

The third approach is to adjust the alpha level to allow for an acceptable error rate. This differs from approaches that attempt to control the overall Type I error rate because the goal is to control the proportion of our “significant” results that are incorrect, rather than the probability of making any such mistakes. The false discovery rate (FDR; Benjamini & Hochberg, 1995 ) and the procedures for controlling it were developed for this purpose. This method makes more intuitive sense for how we think about testing, but it does not guarantee an overall prespecified Type I error rate. FDR-adjusted p values control the number of false discoveries (Type I errors). For example, an FDR of 0.05 implies that, on average, 5% of significant tests will result in false positives under the null hypothesis. The FDR-adjusted p value can be far less conservative than a Bonferroni adjustment but still addresses the goals of controlling Type I errors when multiple comparisons are made.

The goals of this article were to (a) present statistical concepts and methods that we regularly see implemented and misunderstood in speech, language, and hearing sciences research and (b) offer additional insights or alternatives to those concepts and methods. Like other disciplines, the field of statistics has evolved over time to accommodate the realities of modern research involving human subjects. Although the traditional statistical methods that we have discussed in this article still have relevance and specific uses, often other tools can better and more accurately answer the research question of interest and offer greater flexibility for more complex research designs. It is up to scientists in the field to continue educating themselves in modern statistical practice to find statistical approaches that fit their specific goals and to work with statistical experts throughout the research process.

This education begins with a better understanding of the p value and its worth. When the p value is reported in conjunction with an appropriate measure of the effect size, then it gives the researcher important information about the study findings, both statistically and clinically. It is imperative to consider the information actually conveyed by a particular measure of effect size and how it informs the statistical results and practical implications of a given study. Although many of the concepts described in this review are basic statistical principles, the misunderstanding or misinterpretation of these concepts is a substantial threat to the validity of our research findings. To help minimize the potential for statistical errors, researchers in the speech, language, and hearing sciences can do the following:

• Understand how p values are calculated and what p values represent. The 2016 recommendations of the American Statistical Association regarding p values ( Wasserstein & Lazar, 2016 ) can help scientists to avoid common misconceptions about p values and understand the difference between statistical and clinical significance.
• Report measures of statistical and clinical or practical significance as part of articles or reports summarizing research findings. Presented together, metrics of statistical and clinical significance enhance the interpretation of research and make the effects meaningful for clinicians and patients who are much more likely to be interested in the magnitude of the effect in terms of a specific outcome rather than the statistical significance alone.
• Include confidence intervals for effect sizes. The appropriate use of confidence intervals conveys the potential range of plausible values around an effect and can allow consumers of research to understand the influence of variability on the precision of reported effect sizes.
• Choose statistical approaches that allow for modeling complexity over methods with more rigid assumptions such as ANOVA or ANCOVA. Using flexible methods such as regression or mixed models can expand the breadth of research questions that can be evaluated beyond examining differences between groups or across conditions.
• Develop model selection procedures based on scientific knowledge and theory that are parsimonious solutions to complex phenomena. Different iterations of the same statistical model should be compared using established information-based methods, such as comparisons using AIC or Bayesian information criterion.
• Use methods for controlling for multiple comparisons within statistical models that are specific to the goals of the research rather than always using overly conservative approaches that control for familywise error rate, such as Bonferroni adjustment. Decisions about the appropriate method for controlling for multiple comparisons should occur prior to the statistical analysis based on the goals of the comparisons and design of the study.

The goal of reporting results from statistical analyses in articles should be to present new clinically relevant findings or suggest future research opportunities. Readers of the work should be able to replicate the experiment and the analysis. The statistical methods should be written with enough detail that a data analyst could read it and replicate the analysis. Furthermore, code and data should be provided when feasible to promote transparency and reproducibility. Statistical methods are designed to provide good results under uncertainty but always include the possibility of error. With this in mind, replication is an essential part of scientific progress, and wherever possible, researchers should facilitate these efforts. There are many opportunities to apply these principles in the speech, language, and hearing sciences, and it is our hope that this article will help orient investigators to the problem of selecting appropriate and robust statistical models.

Supplementary Material

Supplemental material 1., acknowledgments.

This research was supported by grants from the National Institute on Deafness and Other Communication Disorders Grant R01 DC013591, awarded to Ryan McCreery. Additionally, we thank the editor and two anonymous referees for many helpful comments and suggestions on drafts of this article.

Funding Statement

This research was supported by grants from the National Institute on Deafness and Other Communication Disorders Grant R01 DC013591, awarded to Ryan McCreery.

• Aickin M., & Gensler H. (1996). Adjusting for multiple testing when reporting research results: The Bonferroni vs Holm methods . American Journal of Public Health , 86 ( 5 ), 726–728. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Bender R., & Lange S. (2001). Adjusting for multiple testing—When and how . Journal of Clinical Epidemiology , 54 ( 4 ), 343–349. [ PubMed ] [ Google Scholar ]
• Benjamini Y., & Hochberg Y. (1995). Controlling the false discovery rate: A practical and powerful approach to multiple testing . Journal of the Royal Statistical Society: Series B, Statistical Methodology , 57 ( 1 ), 289–300. [ Google Scholar ]
• Bring J. (1994). How to standardize regression coefficients . The American Statistician , 48 ( 3 ), 209–213. [ Google Scholar ]
• Brysbaert M., & Stevens M. (2018). Power analysis and effect size in mixed effects models: A tutorial . Journal of Cognition , 1 ( 1 ), 9 https://doi.org/10.5334/joc.10 [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Cao J., & Zhang S. (2014). Multiple comparison procedures . Journal of the American Medical Association , 312 ( 5 ), 543–544. [ PubMed ] [ Google Scholar ]
• Lenth R. V. (2012). Some practical guidelines for effective sample size determination . The American Statistician , 55 ( 3 ), 187–193. [ Google Scholar ]
• Loannidis J. P. (2005). Contradicted and initially stronger effects in highly cited clinical research . Journal of the American Medical Association , 294 ( 2 ), 218–228. [ PubMed ] [ Google Scholar ]
• Loannidis J. P. A. (2018). The proposal to lower p value thresholds to .005 . Journal of the American Medical Association , 319 ( 14 ), 1429–1430. [ PubMed ] [ Google Scholar ]
• McMillan G. P., & Cannon J. B. (2019). Bayesian applications in auditory research . Journal of Speech, Language, and Hearing Research , 62 , 577–586. https://doi.org/10.1044/2018_JSLHR-H-ASTM-18-0228 [ PubMed ] [ Google Scholar ]
• Nuijten M. B., Hartgerink C. H., van Assen M. A., Epskamp S., & Wicherts J. M. (2016). The prevalence of statistical reporting errors in psychology (1985–2013) . Behavior Research Methods , 48 ( 4 ), 1205–1226. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Oleson J. J., Brown G. D., & McCreery R. (2019). The evolution of statistical methods in speech, language, and hearing sciences . Journal of Speech, Language, and Hearing Research , 62 , 498–506. https://doi.org/10.1044/2018_JSLHR-H-ASTM-18-0378 [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Pashler H., & Wagenmakers E.-J. (2012). Editors' introduction to the special section on replicability in psychological science: A crisis of confidence . Perspectives on Psychological Science , 7 ( 6 ), 528–530. [ PubMed ] [ Google Scholar ]
• Peng R. (2015). The reproducibility crisis in science: A statistical counterattack . Significance , 12 ( 3 ), 30–32. [ Google Scholar ]
• Saville D. J. (1990). Multiple comparison procedures: The practical solution . The American Statistician , 44 ( 2 ), 174–180. [ Google Scholar ]
• Selya A. S., Rose J. S., Dierker L. C., Hedeker D., & Mermelstein R. J. (2012). A practical guide to calculating Cohen's f 2 , a measure of local effect size, from PROC MIXED . Frontiers in Psychology , 3 , 111. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Trafimow D., & Marks M. (2015). Editorial . Basic and Applied Social Psychology , 37 ( 1 ), 1–2. [ Google Scholar ]
• Walker E. A., Holte L., Spratford M., Oleson J., Welhaven A., & Harrison M. (2014). Timeliness of service delivery for children with later-identified mild-to-severe hearing loss . American Journal of Audiology , 23 ( 1 ), 116–128. [ PMC free article ] [ PubMed ] [ Google Scholar ]
• Wasserstein R. L., & Lazar N. A. (2016). The ASA's statement on p -values: Context, process, and purpose . The American Statistician , 70 ( 2 ), 129–133. [ Google Scholar ]
• Westfall J., Kenny D. A., & Judd C. M. (2014). Statistical power and optimal design in experiments in which samples of participants respond to samples of stimuli . Journal of Experimental Psychology: General , 143 ( 5 ), 2020–2045. [ PubMed ] [ Google Scholar ]

• SUGGESTED TOPICS
• The Magazine
• Newsletters
• Managing Yourself
• Managing Teams
• Work-life Balance
• The Big Idea
• Data & Visuals
• Reading Lists
• Case Selections
• HBR Learning
• Topic Feeds
• Account Settings
• Email Preferences

Embracing Business Practices That Actually Improve the World

• Caitlin McElroy,
• Roberta Roesler,

What organizations need to know about adopting “regenerative” practices that actively help ecosystems and communities — not just minimize harm to them.

The science is clear that the track we’re on is not good enough to prevent further catastrophic effects from climate change. We’re beyond a point where we can merely aim to do less bad; we need to actively regenerate the areas that have experienced significant degradation. Regenerative businesses aim to improve ecosystems and communities, rather than simply minimize harm to them. But in this rapidly expanding, philosophically attractive, and still unsettled space of regenerative business, those who want to take action on regeneration are working from many definitions and approaches. The authors unpack some of the competing definitions of regeneration and show how certifications can help organizations ensure their regeneration strategies and practices support a truly regenerative future.

At the COP28 conference late last year, regeneration emerged as a focus for business leaders. Regenerative businesses aim to improve ecosystems and communities, rather than simply minimize harm to them. It’s no wonder it’s a hot topic — the science is screaming at us that the track we’re on is not good enough to prevent further catastrophic effects from climate change. According to the Stockholm Resilience Centre, we’ve already crossed six of the nine planetary boundaries , “ processes that are critical for maintaining the stability and resilience of [the] Earth system as a whole.”

• Char Love is global director of advocacy at Natura &Co and executive in residence at Saïd Business School, University of Oxford.
• CM Caitlin McElroy , PhD, is a departmental research lecturer at the Smith School of Enterprise and the Environment, University of Oxford.
• RR Roberta Roesler , PhD, is head of make-up development, process, and formula sustainability at Avon International and was previously global R&D director for regeneration and circularity at Natura &Co.
• EF Eve Fraser is a climate policy analyst at the NewClimate Institute and was previously a research assistant at the Smith School of Enterprise and the Environment, University of Oxford.

Partner Center

The Best Business Speech Topics for Students

As a student, you can create your ideal future by honing your communication abilities and embracing the world of public speaking.

One perfect approach to do this is to master the art of delivering high-impact business presentations as a keynote speaker. You can start by learning to brainstorm themes and ideas when sifting through different business speech topics for your school or workplace project.

A business speech is a purposefully delivered verbal address within a workplace or business organization. It is an essential form of business communication where the audience attentively listens to the oration as it is being given.

In this article, we share a comprehensive list containing the best possible business speech topics by focusing on themes and scholarly areas of specialization. We recommend that you pick out subjects that appeal to current industry trends and blend this with your unique points of view that will capture everyone's attention.

We hope our guidelines help you pick an attractive but practical business topic.

Understanding Business Speech Writing

Before we delve deeper into the topic, we must explain the essence of business speech writing.

Speech writers in the business world are responsible for crafting carefully structured scripts that guide speakers' addresses. Although speech writing is commonly associated with political contexts, the demand for this skill extends beyond politics. Various sectors, ranging from business executives to motivational speakers, seek assistance in developing impactful speeches tailored to engage and resonate with their specific audience.

This is one of the reasons why the topic of speech writing is being studied across different institutions of higher learning.

Key Considerations for Business Speech Topic Selection

Now let's take a look at what to consider when selecting a hot business topic for your speech:

●       The intended audience: First, you must consider your target demographic as well as their interests and knowledge level. This ensures that the chosen topic is both meaningful and engaging for them

●       The purpose of the business speech: Ensure you understand what you want from the address. Clarity of purpose is always paramount

●       Relevance: As clichéd as this may sound, ensure that you choose a topic that associates with relevant issues in the business landscape or industry

●       Length and Scope: Finally, evaluate the time allotted for your speech and work within this time frame. This ensures that you choose a topic with enough content and that is relevant to make it worthwhile

Different Topics You Can Select For Your Address

The following sections offer practical business speech-writing topics to get you started. We initially focus on creative, informative, and persuasive business speech sample topic ideas, then on specific scholarly areas of specialization.

Creative Business Speech Topics

1. Transforming traditional business operations through a retail renaissance

2. Leading through the storm: Navigating management challenges in crisis

3. Staffing innovations: Embracing creative solutions for talent acquisition

4. Unconventional team-building methods that foster collaboration

5. Unlocking the hidden potential of knowledge management

6. Franchise fusion: Crafting agreements for unprecedented win-win ventures

7. Strategic e-marketing that makes waves across industries

8. E-marketing quandaries: Addressing strategic challenges with finesse

9. Small wonders: Unleashing entrepreneurial potential with creative ideas

10. The power of place: Crafting your ideal working environment

11. Beyond boundaries: Unveiling the ever-evolving global business environment

12. Persuasive powerplay: Mastering the art of influential business speeches

13. Untold stories of triumph from inspiring business owners

14. The presentation puzzle: Unconventional approaches to captivate audiences

15. Change catalysts: Managing organizational transformation with finesse

16. How can you craft compelling research papers in business?

17. Ideas in action: Empowering employees to implement creative solutions

18. Consumer chronicles: Unraveling the mysteries of consumer behavior

19. Embracing change in business practices through managerial metamorphosis

20. Communication compass: Navigating the path to effective business dialogue

21. Plan for prosperity: Unleashing the power of strategic business plans

22. Beyond paychecks: Rethinking performance-related incentives for motivation

Informative Business Speech Topics

1.      Successful retail business operations: Key strategies and best practices

2.      Leadership and management challenges in today's business environment

3.      Recruitment and staffing decisions: How to find the right talent

4.      Creative team-building methods for enhanced collaboration and performance

5.      Unveiling desirable business topics: Engaging ideas for presentations

6.      Knowledge management: Unlocking the power of information in organizations

7.      Exploring good franchising business agreements for sustainable growth

8.      Strategic e-marketing: Navigating digital landscapes for business success

9.      Informative business speech topics: Inspiring ideas for engaging audiences

10.  Strategic e-marketing issues and solutions: Insights for effective campaigns

11.  Generating small business ideas: From vision to profitable ventures

12.  Designing your working environment: Enhancing productivity and satisfaction

13.  How to handle dissatisfied customers

14.  Understanding the business environment: Trends, challenges, and impacts

15.  What are the challenges, strategies, and success stories of small businesses?

16.  How to captivate audiences and deliver impact in meetings

17.  Techniques for effective communication

18.  How to manage organizational change to have smooth transitions

19.  Critical components for successful business academic writing

Persuasive Speech Topics

1.      Why you need advanced strategies for retail business operations

2.      Importance of unleashing potential and driving success through leadership

3.      Proven recruitment and staffing methods that are effective

4.      Creative and proven methods for collaboration and productivity

5.      Why you must maximize the value of knowledge management

6.      Why you should make symbiotic agreements for win-win partnerships

7.      How to shape the business world like a champion

8.      Why the branding concept of a company matters

9.      Using strategic e-marketing to level up your business game

10.  Crack the code of e-marketing challenges like a professional

11.  Top ideas to boost your entrepreneurial journey

12.  Unleash the secret powers of a productive work environment

13.  Be a business communication superstar: Secrets to clear and impactful messaging

14.  Navigating trends and transformations to shape business environments

15.  Empowering small business owners: Strategies for success and sustainability

16.  Developing a comprehensive business plan

17.  Fair performance-related pay: Motivating employees through rewarding results

18.  Achieving work-life balance: Promoting employee well-being and productivity

19.  Emergency or controversial issues: Strategies for effective crisis communication and management

20.  Exploring globalization trade opportunities: Benefits and challenges for businesses

21.  Leveraging strategic e-marketing for competitive advantage: Case studies and insights

22.  Promoting corporate responsibility: Ethical practices for sustainable success

23.  Innovating to reduce production costs: Strategies for operational efficiency

24.  Five essential employer responsibilities: Creating a supportive work environment

Breaking Down Business Speech Topics Based on Your Area of Specialization

We also recommend that students select topics that align with their field of study. For instance, a business administration student can focus on customer acquisition, finance, or marketing issues.

Here are more examples.

Business Administration Speech Ideas

1. How to power businesses through organizational change management

2. Unleashing the power of knowledge management in operations

3. Crafting effective franchising business agreements for success and growth

4. A discussion on strategic e-marketing in business administration

5. Supercharge your leadership skills to conquer management challenges

Marketing and Advertising Speech Ideas

1. Strategic e-marketing: What are the Opportunities and challenges in advertising?

2. Harnessing the power of persuasive speeches in marketing

3. Crafting an effective business plan: Essential steps for marketing success

4. What are the effects of consumer behavior on advertising strategies?

5. Trade opportunities and marketing strategies for business growth

6. Developing persuasive speeches for workplace meetings

Entrepreneurship and Small Business Speech Ideas

1.      How to come up with and execute small business ideas

2.      What makes a business plan successful?

3.      Exporting issues related to government regulations

4.      How do small companies overcome challenges and implement workers' ideas?

5.      Unveiling the power of creative team building

6.      Addressing the opportunities and challenges in today's market for small businesses

7.      Unique manufacturing methods that reduce production costs

8.      Enhancing logistics and transportation is critical for economic growth

Leadership and Management Speech Topic Ideas for 2023

1.     Overcoming leadership and management challenges in today's business environment

2.     Effective recruitment and staffing decisions: Building high-performing teams

3.     Inspiring leadership: Strategies for motivating small business owners

4.     How to avoid wastage due to churning frozen food products

5.     Navigating organizational change: Effective management for smooth transitions

6.     Harnessing emotional intelligence: The key to effective leadership and management

Finance and Accounting Speech Topic Ideas for 2023

1.     Exploring financial strategies for successful retail business practice

2.     Financial leadership: Overcoming challenges in a dynamic business environment

3.     Effective decision-making in recruitment and staffing: An economic perspective

4.     Maximizing profitability: Innovative approaches to performance-related pay

5.     Navigating government regulations: Ensuring compliance and financial success

Human Resources and Organizational Behavior Speech Topic Ideas for 2023

1.     Effective recruitment strategies: Making successful staffing decisions for business growth

2.     Nurturing a positive work environment: Enhancing employee motivation and satisfaction

3.     Managing organizational change: Strategies for smooth transitions and adaptation

4.     Building high-performing teams: Unlocking the power of creative team building methods

5.     Ethical leadership: Balancing corporate responsibility and business success

6.     Should you have a ceiling on weekly working hours?

Enjoy Custom Speech Topic Assistance

In the business world, great speeches can make a lasting impression during any meeting presentation.

Delivering an impactful and persuasive presentation starts with selecting the right topic. To do this effectively, you must consider your audience, define your purpose, pick timely and relevant subjects, and align the speech's length and scope to maximize effectiveness.

Whether it's creative or informative content that you're after, ensuring every word is chosen for maximum impact will help you resonate with listeners. This is where My Custom Essays come in. Our experienced professional writers are here to help create compelling business speeches that captivate audiences every time.

So don't miss out on this opportunity – contact us today for assistance crafting powerful presentations that get results and any help writing theses.

1.      Can I use these business speech topics for my college assignments?

Yes, absolutely! We have carefully curated them to suit different interests and scholarly areas of specialization, making them perfect for your assignments or presentations.

2.      How can I make my business speech more engaging?

Consider incorporating storytelling techniques, using visual aids, and involving the audience.

3.      Are any additional resources available to help me improve my public speaking skills?

Yes. You can explore online courses, join public speaking clubs or organizations, and practice regularly.

Recent Posts

28 Oct, 2023

19 Oct, 2023

05 Oct, 2023

20 Sep, 2023

18 Sep, 2023

13 Sep, 2023

12 Sep, 2023

11 Sep, 2023

21 Aug, 2023

29 Jul, 2023

28 Jul, 2023

25 Jul, 2023

21 Jun, 2023

15 Jun, 2023

02 May, 2023

24 Apr, 2023

18 Apr, 2023

12 Apr, 2023

26 Mar, 2023

24 Mar, 2023

22 Feb, 2023

17 Feb, 2023

16 Feb, 2023

24 Oct, 2022

26 Sep, 2022

25 Sep, 2022

• #1 Nursing dissertation helpers
• Marketing Dissertation Helpers
• Buy Expository Essay
• Buy Book Reports
• Best MBA Essay Editing Service Providers
• Essay Outline Writing Service
• Chemical Engineering Assignment Help
• Mechanical Engineering Assignment Helpers
• Extended essay writing service
• Buy case study
• Buy personal statement online
• Financial Report Writing Services
• Buy Annotated Bibliography
• Buy dissertation proposal
• Professional Biology Essay Writers
• Buy Response Essay
• Write My Philosophy Paper for Me
• Psychology Essay Writing Service
• Buy Lab Report Online
• Professional Article Review Writing Service
• Need Graduate School Homework Help?
• Buy High-Quality Reflective Essays
• Buy Dissertation Chapters Online
• Buy Movie Reviews
• History Essay Writer
• Buy a Persuasive Essay
• Can I pay someone to do my chemistry homework?
• Buy Critical Essay
• Buy Coursework Online
• Buy Narrative Essays Online
• Buy Cheap Argumentative Essay
• Do my final exam for me
• Pay someone to take my ProctorU Exam
• Hire someone to take GRE for me
• Pay Someone to Take my Physics Exam
• MBA Essay Writing Service

Communication Sciences and Disorders

Us news & world report ranks iowa audiology, speech-language pathology among the top in the nation.

Two UIowa programs in the Department of Communication Sciences and Disorders (CSD) are once again recognized among the best in their field according to the U.S. News & World Report Best Graduate Schools rankings for 2024 .  

Iowa’s audiology program is again ranked second best in the nation and remains the top public institution training audiologists. The speech-language pathology program climbed in the rankings to fifth and is now the third-ranked public institution. CSD is housed in the College of Liberal Arts and Sciences.   

New rankings of nation-wide  Audiology and  Speech-Language Pathology graduate programs is available on the USNWR website.

“We are focused on training outstanding audiologists and speech-language pathologists,” said Eric Hunter , DEO and Harriet B. and Harold S. Brady Chair in Liberal Arts and Sciences. “These rankings recognize the daily efforts of our remarkable faculty and staff, who consistently strive to provide exceptional education to our students.”  

“Together, we are shaping the future of audiology and speech-language pathology, and we will continue to lead the way in delivering excellence,” Hunter added.  

Hunter, who has been at Iowa since August 2023 and is a nationally recognized expert in the field, was the first hire under the University of Iowa’s Transformational Faculty Hiring Program, which is aimed at attracting world-class faculty to strategic programs and areas of excellence. He received a PhD from Iowa’s Department of Communication Sciences and Disorders program in 2001.  

Iowa has long been a leader in communication sciences and disorders dating back to 1897, when the university, led by Carl Seashore’s pioneering work, developed speech pathology as a discipline of study.  

“CLAS is proud of our faculty and staff who continue to guide this storied program into the future,” said Dean Sara Sanders. “Because of their tremendous talent and dedication to their research and teaching, CSD continues to be at the forefront of audiology and speech-language pathology.”   

The university, college, and department continue to lead, ensuring students access to unparalleled opportunities. Construction has started on a new $249 million building that will provide a state-of-the-art learning space for Iowa students studying in the Department of Communication Sciences and Disorders, Department of Health and Human Physiology, and the Carver College of Medicine’s Department of Physical Therapy and Rehabilitation Science.  

Substantial completion of the building is anticipated in 2025.  

NOTICE: The University of Iowa Center for Advancement is an operational name for the State University of Iowa Foundation, an independent, Iowa nonprofit corporation organized as a 501(c)(3) tax-exempt, publicly supported charitable entity working to advance the University of Iowa. Please review its full disclosure statement.

Duke CNAP Welcomes First Speech-Language Pathologist

Duke’s Cognitive Neuroscience Admitting Program (CNAP) is set to welcome its first speech-language pathologist as a doctoral student. Shanika Phillips Fullwood , a clinically certified SLP with over a decade of experience, will begin her studies in the fall of 2024.

Her enrollment reflects a promising shift toward innovative, interdisciplinary approaches in brain and language research.

Phillips Fullwood works at Moses Cone Memorial Hospital in Greensboro, treating adult patients who have neurogenic disorders, associated with stroke, traumatic brain injury, and spinal cord injury. She serves as the acute rehab neuroscience diagnostic specialist, a role that involves coordinating with physical therapists, occupational therapists, and SLPs to ensure the use of evidence-based practices.

In addition to her clinical work, she is an adjunct lecturer at her alma mater, Brooklyn College, CUNY, and she contributes to research on neurogenic communication disorders in the lab of Jamila Minga, PhD, CCC-SLP . Phillips Fullwood’s work in the lab includes training undergraduates in transcription checks for the Right-Hemisphere Brain Damage (RHD) Bank and participating in brain mapping studies to identify regions associated with specific deficits.

She’s also co-authoring manuscripts focused on the impact of culture on communication and the development of a diagnostic code for apragmatism, work that aligns with her research interests.

As a doctoral student, Phillips Fullwood plans to focus on the cognitive aspects of communication, particularly how right-hemisphere strokes affect language and communication abilities. Her interest in this area developed during her time as an SLP in Indiana, where she encountered a number of patients with right-hemisphere strokes who had trouble with communication.

These patients struggled to stay on topic and understand questions, even though their language center – which is typically associated with the left hemisphere – was not affected.

“I would bring up these concerns during our interdisciplinary team meetings to the physical medicine doctor, and the response often was, ‘Well, they didn't have a left-hemisphere stroke, they had a right-hemisphere stroke, so they shouldn't have communication difficulties,’” she said.

This led her to delve deeper into the research to understand whether these kinds of communication deficits have been documented in the literature.

“There was some evidence supporting that, but still, there's a bias toward the left hemisphere being the only hemisphere that's responsible for communication,” she said. “I thought, ‘Well, it's time for me to go back and do my PhD,’ which was always my goal.”

Known for its interdisciplinary approach, CNAP provides students with the opportunity to engage with faculty across multiple departments and fields at Duke, such as neurobiology, psychology, psychiatry, biomedical engineering, and philosophy .

Students undergo 18 months of coursework and laboratory rotations, gaining foundational knowledge in cognitive neuroscience. After that, they select a department and two advisors to guide their thesis research, furthering their specific interests through an affiliate graduate program or department.

“The fact that CNAP is so interdisciplinary was really attractive to me,” Phillips Fullwood said. “I felt that this program would serve me a lot better than a PhD in communication sciences because I'll be able to tap into some of the techniques they're using in neuroscience to get a better understanding of the communication piece.”

Her enrollment in CNAP demonstrates the interdisciplinary approach necessary for advancing the field of cognitive neuroscience. It also highlights the opportunity to integrate speech-language pathology into Duke's broader research and academic offerings. Currently, the school does not offer a training program for communication sciences and disorders at the master’s or doctoral level.

Through her work, Phillips Fullwood hopes to help bridge the gap between clinical speech-language pathology and cognitive neuroscience in treating patients with right-hemisphere strokes who have communication deficits.

“ I believe that we must have a good understanding of why this patient is doing what they're doing to inform how we evaluate and treat them,” she said. “ If we improve this piece, then we'll be able to help translate that into better diagnostic tools, better assessment methods, and better, more formal ways for treating this than we currently have.”    

College of Engineering

Georgia tech ai makerspace.

Using an approach unlike any other in higher education, Georgia Tech’s College of Engineering has created a digital sandbox for students to understand and use artificial intelligence in the classroom.

The AI Makerspace is a supercomputer hub that gives students access to computing resources typically available only to researchers or tech companies. It means hands-on experience for our students, deepening their skills and preparing them to be the new generation of AI professionals.

With the resources in the AI Makerspace, the College can redesign courses to incorporate practical AI tools and develop new ones that impart the essential principles of AI to all students.

The initiative is in collaboration with NVIDIA , one of the country’s largest suppliers of AI hardware and software — and a substantial investment. Students and faculty receive support through NVIDIA Deep Learning Institute resources, including faculty-run NVIDIA workshops, certifications, a university ambassador program, curriculum-aided teaching kits, and a developer community network.

The AI Makerspace also enables Georgia Tech to enhance or redesign courses to incorporate practical AI tools, along with develop new courses — both foundational and advanced — that impart the essential principles of AI to all students. The partnership between Georgia Tech and NVIDIA signifies a substantial investment. The allocated funds will be utilized for technology, including NVIDIA graphics processing units (GPUs), and infrastructure. S tudents and faculty will receive support through NVIDIA Deep Learning Institute resources, including faculty-run NVIDIA workshops, certifications, a university ambassador program, curriculum-aided teaching kits, and a developer community network.

The collaboration is part of the College’s commitment to nurturing a vibrant AI-powered university that will shape the future generation of AI professionals.

Georgia Tech Unveils New AI Makerspace

By giving students access to powerful supercomputers, Georgia Tech will teach AI to undergraduates in a way unlike any other university in the nation.

What Sets the Georgia Tech AI Makerspace Apart?

Educational Empowerment

In an era where AI is increasingly ingrained in our daily lives, the AI Makerspace democratizes access to heavyweight computing resources.

Training the AI Workforce

The AI Makerspace takes a dedicated approach to workforce development through curriculum-based study as well as independent exploration. 

National Security

Harnessing the power of AI is a strategic imperative for national security. As nations strive to secure their positions as global leaders in the field, investing in AI education is critical for U.S. competitiveness.

Interdisciplinary Focus

The AI Makerspace offers a unique opportunity for students to harness the power of AI technologies in ways that extend beyond traditional computing applications.

The Georgia Tech AI Makerspace is a dedicated computing cluster paired with NVIDIA AI Enterprise software. The software technology resides on an advanced AI infrastructure that is designed, built, and deployed by  Penguin Solutions , providing a virtual gateway to a high-performance computing environment. 

The first phase of the endeavor is powered by 20 NVIDIA HGX H100 systems, housing 160 NVIDIA H100 Tensor Core GPUs, one of the most powerful computational accelerators capable of enabling and supporting advanced AI and machine learning efforts. The system is interconnected with an NVIDIA Quantum-2 InfiniBand networking platform, featuring in-network computing. 

Infrastructure support is led by Georgia Tech’s Partnership for an Advanced Computing Environment (PACE) .

It would take a single NVIDIA H100 GPU one second to come up with a multiplication operation that would take Georgia Tech’s 50,000 students 22 years to achieve.

20 NVIDIA H100-HGX servers, each with:

• 8 x NVIDIA H100 GPUs (SXM5 form-factor)
• 2 x 32-Core Intel Sapphire Rapids CPUs (2.8 GHz)
• 2TB 4800 MHz DDR5 DRAM
• 3 x 3.84 TB NVMe storage
• 1 x ConnectX-7 IB NIC (400 Gbps)

Total System:

• 160 NVIDIA H100 GPUs
• 1,280 Intel Sapphire Rapids CPU cores
• 40TB 4800 MHz DDR5 DRAM
• 230.4 TB NVMe storage

Frequently Asked Questions

What are gpus and cpus.

GPUs (graphics processing units) are specialized processors designed to handle certain complex computations efficiently, commonly used in tasks such as rendering high-resolution graphics and performing parallel computations in fields like machine learning and artificial intelligence. CPUs (central processing units) are the central component of a computer responsible for executing instructions, managing tasks, and coordinating the operation of various hardware components, serving as the brain of the computer.

GPUs have become prominent due to their exceptional parallel processing capabilities, which make them highly efficient for high-performance computing (HPC) tasks. Additionally, advancements in GPU technology have led to significant improvements in graphics rendering, gaming experiences, and visual computing applications, further driving their prominence in various industries and fields.

How many GPUs are in the Georgia Tech AI Makerspace and what makes them important?

Phase I of the Georgia Tech AI Makerspace comprises a total of 160 NVIDIA H100 Tensor Core GPUs. 20 NVIDIA H100-HGX servers contain 8 GPUs each. The benefit of GPUs is that they provide extremely performant accelerators designed specifically for AI, with a very large unified memory space that can accommodate very big models.

It’s also noteworthy that an important capability of AI is low-precision performance. These nodes provide roughly 640 petaflops (PF) of theoretical 8-bit floating-point for 8-bit integer (FP8/INT8) capability, combined with the 640 gigabytes of GPU memory per server.

Why are there both GPUs and CPUs in the Georgia Tech AI Makerspace? 

CPUs and GPUs are optimized for different kinds of calculations, so it’s useful to have both available. Optimized software will perform certain steps of code on the CPU and others on the GPU to maximize performance.

CPUs are “standard” general-purpose chips that work well for many calculations. GPUs are specialized. A server cannot run without a CPU. The CPU handles all the tasks required for all software on the server to run correctly. 

GPUs are accelerators with more focused computational hardware that rely on a separate host system to operate.

Who will manage the infrastructure of the AI Makerspace? 

The AI Makerspace infrastructure will be supported by Georgia Tech’s Partnership for an Advanced Computing Environment (PACE). PACE provides sustainable leading-edge Research Computing and Data (RCD) cyberinfrastructure, software, and support for research and education requiring high performance computing and other advanced research computing infrastructure. 

PACE is a collaboration between Georgia Tech faculty and the Office of Information Technology (OIT) focused on HPC.

Is the AI Makerspace scalable?

Yes. Each GPU can be physically partitioned into 7 GPUs (with 1/8 the capability of the whole). With 160 total GPUs, the AI Makerspace can provide 1,120 concurrent GPUs to allow large numbers of students access simultaneously. 

How much power does the AI Makerspace require?

The new servers will draw about 140kW of power, compared to the 800kW PACE’s five existing clusters draw.

The theoretical 64-bit performance of the new hardware is 5.5 PF (petaflops, a measurement of computer speed of performing calculations). The existing PACE clusters altogether have about 4-4.5 PF of performance. This means that the new servers are significantly more energy efficient for the same computational capability than older systems.

Related Content

Minor Degree in AI and Machine Learning Available Summer 2024

The new minor degree program is a partnership between the College of Engineering and the Ivan Allen College of Liberal Arts, teaching AI technical skills alongside ethics and policy considerations.

College Adds, Reimagines AI Courses for Undergraduates

In response to demand from its students, initiatives within faculty research, and increasing needs from industry, the College has created and reimagined more than a dozen courses to strengthen its AI and machine learning education.

Purdue’s I-EaT Research Lab examining swallowing and speech in children with cerebral palsy

Georgia Malandraki

Written by: Tim Brouk, [email protected]

Understanding the role head and neck muscles play in eating and talking for children with cerebral palsy (CP) is a research effort led by Georgia Malandraki , professor in the Purdue University Department of Speech, Language, and Hearing Sciences , and her I-EaT Lab . The work was funded by the National Institutes of Health, the American Academy for Cerebral Palsy and Developmental Medicine, and the Purdue College of Health and Human Sciences .

“More than 50% of children and adults with cerebral palsy may have speech, feeding and swallowing difficulties,” Malandraki said. “Although there has been a lot of research on the gross motor and hand functions of this population and how to best rehabilitate them, there has been much less work on the underlying mechanisms that cause feeding, swallowing and speech difficulties. This has limited our treatment options for these children and was a gap that we wanted to address.”

Speech difficulties can result in children and adults with CP not being understood; feeling isolated; and having reduced academic, social and employment fulfillment. Swallowing difficulties can also lead to serious consequences, such as being undernourished, dehydrated or in severe cases, developing respiratory infections (pneumonia) from foods and liquids not being safely swallowed. These types of difficulties — swallowing and speech — may also co-occur.

Most of the available treatments for swallowing and speech difficulties in CP are compensatory in nature and mostly address the symptoms, according to Malandraki. For example, simple changes in the child’s head and neck position during meals can aid in swallowing, and specialized, padded eating utensils could be an easy investment for families if a child with CP is having feeding issues. Malandraki added that making meals with softer, easier-to-swallow food, such as a fruit smoothie instead of fruit that needs to be chewed, is another simple strategy when children have difficulty chewing. Speaking slower or using exaggerated mouth movements are some common speech strategies as well.

“To be able to effectively improve the speech and swallowing functions of children and adults with CP, it is first imperative to understand how their muscles and brains work so we can use that information as we develop new treatments that will not just focus on the symptoms,” Malandraki stated.

Understanding underlying mechanisms

A child wears small muscle activity readers on their face to study how muscles activate during eating and speech tasks.

Through electromyography (small muscle activity readers), Malandraki and her team measured the head and neck muscle activity in 16 children with CP ages 7-12 and 16 children without the condition but the same ages.

The children with CP were tested at two sites, Malandraki’s lab and Nationwide Children’s Hospital in Columbus, Ohio. All had unilateral CP, a form of the condition where one side of the body is weaker than the other due to brain lesions affecting primarily one side of the brain. These children are usually more physically functional than children with more severe types of CP, such as quadriplegic CP, where both sides of the body are affected.

The children were able to feed themselves, and they were given a variety of food and liquids to swallow and were also asked to complete simple speech tasks. Through sensors attached around the mouth and under the chin, the researchers first found children with CP use these muscles much more than children without CP for both swallowing and speaking. This over-activation means these children were spending more energy to eat and speak than the typically developing control group.

“They show increased amplitude of muscle activity, indicating a lot of muscle effort,” Malandraki said. “So, they’re expending a lot of energy on both sides — the affected and unaffected — of the head and neck to be able to do simple tasks, such as eat, swallow and speak.”

For swallowing testing, the children were given sips of water to drink and pudding and pretzels to eat. Repetition of simple sounds and words was also requested. While used differently, the same muscles — perilabial (around the lips) and submental (under the chin) — are in play for both tasks.

Along with how much the muscles are fired, the researchers looked at timing — how quickly do the muscles react, and how long are they being contracted during a task? Swallowing and speech both require very delicate muscular coordination and timing. This biomechanic dance is easy for most but for children with CP, the steps can be arduous.

 “What we saw was that in addition to using more muscle activity, they also used the muscles with less coordination between left and right sides compared to the typically developing group,” Malandraki said. “Our findings show that there may be a muscle coordination issue in some of these children. They can contract the muscles but have a hard time isolating the specific muscles that they need for a specific task.”

Recently published in the Journal of Neurophysiology , the work can inform evaluation processes and improve treatments for the children as they grow and develop.

“We want to better understand what the underlying mechanisms of the speech, feeding and swallowing difficulties that we see in these children are and also what are the interactions between speech and swallowing,” Malandraki said. “We are interested in finding more about a potential cross-system interaction between these two vital functions that both share many of the same mechanisms and muscles in the head and neck but use these muscles a little bit differently — or they’re supposed to use them a little bit differently for each function.”

Brain analysis results next

The study also collected magnetic resonance imaging (MRI) scans of the brains of children with CP and their typically developing peers. Since CP is typically caused by lesions on the brain formed during infancy or even while in the fetal state, this arm of the work is being analyzed now and will reveal neurological understanding of how children with varying levels of CP swallow and speak.

Tackling potentially life-threatening issues for children with CP is a difficult task because no child with the condition is like another. In her work, Malandraki said the severity of CP and of the speech and feeding profiles within the 16 children tested for the recent work ranged widely. Studying data gathered from muscle and brain MRI scans will help understand how some of these children can develop better swallowing and speech skills than others, and give us insights into how children and their families can adapt as the children grow into teenagers and young adults.

“Hopefully we can use these findings to develop treatments that will target these mechanisms,” Malandraki explained. “Most current treatments are not improving their speech and swallowing to the level that hopefully they could if they were based on their underlying mechanisms. Our work is trying to change that.”

Malandraki is currently expanding this work to other neurological populations to guide physiology-based treatment development for debilitating swallowing and speech disorders across the life span.

U.S. News & World Report ranks UNI among 2024 Best Graduate Schools

CEDAR FALLS, Iowa –  U.S. News & World Report has once again ranked the University of Northern Iowa among its list of 2024 Best Graduate Schools. Designed for prospective students looking to further their education beyond an undergraduate degree, the  Best Graduate Schools rankings evaluate programs in a variety of disciplines, including business, education, engineering, law, medicine and nursing. 

UNI’s honored programs for graduate education are a combination of face-to-face, entirely online and hybrid degrees offered in partnership with all UNI colleges,  Online and Distance Education , and the  College of Graduate, Research and Online Education where more than 80 degree, certificate and endorsement programs are housed. 

“It’s incredibly gratifying to have UNI recognized as a national leader in graduate education,” said Steph Huffman, associate vice president for strategic initiatives and dean of the College of Graduate, Research and Online Education. “UNI has a strong, long-standing history of developing educators, innovators, strategic leaders and other professionals who serve our schools, businesses and organizations, and communities. This ranking reflects our commitment to providing high-quality, impactful graduate educational opportunities.”

By collecting graduate school data annually, U.S. News is able to present the most current figures on enrollment, job placement, faculty and other critical quality indicators that help prospective students make informed decisions.

UNI’s  College of Education offers 10 master's programs and a doctoral program which are represented within the graduate rankings of Best Education Schools. These range from advanced degrees like elementary and early childhood education to programs that prepare school administrators, school psychologists and teacher librarians.

"The ranking continues to represent our outstanding faculty, innovative curriculum, and our ability to meet students where they are to deliver an experience that is individualized, relevant and meaningful,” said Amy Nielsen, the College of Education’s interim associate dean of graduate studies.

UNI also ranks among the top in the country for its  part-time MBA in the Wilson College of Business. 

"I am proud to see our MBA program recognized again as one of the top programs in the nation," said Leslie Wilson, dean of the Wilson College. "I know our faculty are working diligently to ensure our curriculum meets the needs of students preparing to be global business leaders. From the use of AI, addressing supply chain disruptions, and managing a remote workforce, we seek to provide our students with the necessary skills and knowledge to address the changing landscape and excel in their careers. This recognition is a testament to the hard work and dedication of our faculty, staff and students, who have helped make our program successful."

UNI’s  Speech-Language Pathology M.A. is among the top 100 speech pathology programs in the rankings. Students benefit from valuable clinical experience at the Roy Eblen Speech and Hearing Clinic located conveniently on campus.

“We take pride in offering students a comprehensive educational experience that combines hands-on clinical exposure, rigorous coursework, and valuable research opportunities with faculty members,” said Jennifer Garrett, head of the UNI Department of Communication Sciences and Disorders. “This recognition reaffirms our department's commitment to excellence in training future speech-language pathologists. With the experiences they gain here, our students graduate as highly competent professionals, ready to make meaningful contributions to improving the lives of individuals with communication disorders.”

UNI’s  Master of Public Policy Program is recognized as one of the Best Public Affairs Programs by U.S. News & World Report. With its fully online format, the program is ideal for those seeking career advancement within public service and nonprofit organizations in the field of public policy. 

“As the only nationally accredited program in the state, earning this recognition reflects our commitment to excellence and innovation in preparing future leaders for impactful public policy careers,” said Chris Larimer, MPP program coordinator and professor in the Department of Political Science. 

The University of Northern Iowa ranks highly in a number of U.S. News categories, including ranking  second best among regional public universities in the Midwest . It also ranks among the best value schools, best colleges for veterans and most innovative schools.

Media Contact: Adam Amdor

[email protected], uni grad tees up tournament operations for pga’s john deere classic.

TheatreUNI presents “Suzette Who Set to Sea”

UNI physics faculty awarded National Science Foundation and Dept. of Energy research grants