research based personality test

PUBLISHED Jan. 16, 2019, at 9:00 AM

Most Personality Quizzes Are Junk Science. Take One That Isn’t.

Compare your results to those of your friends and family..

By Maggie Koerth and Julia Wolfe

Your groups

You’re part of the ${group} group. its members are taking a personality quiz to see how they compare to the group’s average. once you’ve taken the quiz, your score will be added to the group average..

What’s your personality, and what can it tell you about your true self? Those questions have launched a thousand online personality quizzes. But you can do better than those specious — yet irresistible — quizzes. You can take a personality quiz backed by science.

Meet the Big Five, the way most psychologists measure and test personality. It’s a system built on decades of research about how people describe one another and themselves. (You can read more about it in this article we published last year.) There are a couple of things that make it — and this quiz — different.

First, the Big Five doesn’t put people into neat personality “types,” because that’s not how personalities really work. Instead, the quiz gives you a score on five different traits: extraversion, agreeableness, conscientiousness, negative emotionality and openness to experience. For each of those traits, you’re graded on a scale from 0 to 100, depending on how strongly you associate with that trait. So, for example, this quiz won’t tell you whether you’re an extravert or an introvert — instead, it tells you your propensity toward extraversion. Every trait is graded on a spectrum, with a few people far out on the extremes and a lot of people in the middle.

The other thing that makes the Big Five different is it lets you easily compare your score to others’. We’re going to show you how your personality compares with that of the average American. And once you get your results, you can invite friends and relatives to compare your personality to theirs. (In the meantime, you’ll be stuck comparing yourself to the average FiveThirtyEight staffer.)

So what are you waiting for? Find out who you really are — take the quiz for yourself!

Share with family and friends

No matter how extraverted you are, this quiz is more fun when you can compare yourself with people you know to start sharing your results, create a group below., name your group.

Additional design and development by Rachael Dottle

Illustration by Sonnie Kozlover

Big Five Inventory-2 items copyright 2015 by Oliver John and Christopher J. Soto. Reprinted with permission.

Related Stories

Comments .

• About Nielsen Measurement
• Terms of Use
• Privacy Policy
• Your California Privacy Rights
• Children's Online Privacy Policy
• Interest-Based Ads

© 2018 ABC News Internet Ventures. All rights reserved.

October 10, 2018

How Accurate Are Personality Tests?

Precious few personality assessments are known to be reliable, and researchers say their use outside academia is debatable

By Angus Chen

Tim Robberts Getty Images

If you’re looking for insight into the true you, there’s a buffet of personality questionnaires available. Some are silly—like the internet quiz that tells everyone who takes it that they are procrastinators at the core. Other questionnaires, developed and sold as tools to help people hire the right candidate or find love, take themselves more seriously.

The trouble is, if you ask the experts, most of these might not be worth the money. “You should be skeptical,” says Simine Vazire, a personality researcher at the University of California, Davis. “Until we test them scientifically we can’t tell the difference between that and pseudoscience like astrology.”

One famous example of a popular but dubious commercial personality test is the Myers–Briggs Type Indicator. This questionnaire divides people into 16 different “types” and, often, the assessment will suggest certain career or romantic pairings. It costs $15 to $40 for an individual, but psychologists say the questionnaire is one of the worst personality tests in existence for a wide range of reasons. It is unreliable because a person’s type may change from day to day. It gives false information (“bogus stuff,” one researcher puts it). The questions are confusing and poorly worded. Vazire sums it up as “shockingly bad.”

On supporting science journalism

If you're enjoying this article, consider supporting our award-winning journalism by subscribing . By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.

Personality questionnaires began evolving about a century ago, says Jim Butcher, an emeritus psychologist at the University of Minnesota. “They started asking questions about an individual’s thinking and behavior during World War I,” he says. “These were to study personality problems and mental health problems.” And importantly, he adds, the U.S. military wanted the questionnaires to help weed out soldiers who weren’t fit to fly military aircraft.

According to Butcher, during the first half of the 20th century many academics started creating different personality scales. “Not just on mental health diagnoses, but what personality is like,” he says. The problem with practically all of the assessments at the time was they were a built on the creators’ subjective feelings about personality, he notes. “Then people started to raise questions about do they really measure what they think they’re measuring? How reliable are those conclusions, and are they valid?”

Butcher describes what followed as a mass culling of personality systems and questionnaires by the scientific method. There is one personality model that did survive the 20th century, though. It is popular among academics today, and is what Vazire uses in her research. It’s called the Big 5 Personality Traits (aka 5-Factor Model), and it was developed over three decades beginning in 1961 at Brooks Air Force Base. From then to the 1990s, several psychologists including Lewis Goldberg, Warren Norman, Paul Costa and Robert McCrae helped develop the model into its modern form.

Vazire says in developing the Big 5 Personality model, psychologists tried to avoid pitfalls that plagued early personality researchers—like selecting criteria based primarily on intuition. Instead, the Big 5 model took a holistic tack by compiling every word that could be considered a personality trait and creating simple, straightforward questions about them. For example, on a scale of 1 to 5, are you outgoing, sociable? Have a forgiving nature? Based on how people answered initial surveys, researchers used statistical methods to group traits that seemed to cluster together (like  “talkative” and “sociable”) into five basic categories: extraversion, conscientiousness, agreeableness, neuroticism and openness to experience. The other model, the HEXACO model of personality structure created in 2000 by psychologists Kibeom Lee at the University of Calgary and Michael Ashton at Brock University in Ontario , is similar but adds an extra category: honesty–humility.

The key to the Big 5 model is its simplicity. It doesn’t sort anybody into a “type,” just informs them where they fall on a continuum of personality traits. There are no tricks and no surprises to be revealed, Vazire says. “In a way, it’s disappointing. It just means that personality tests can only tell you what you tell it.” You won’t learn anything that you didn’t already know about yourself, she adds, and its accuracy comes entirely from how honest and self-reflective you were with your answers.

At best, Vazire says you could use it as a comparative tool that can tell you how you rank on extroversion compared with others who have taken the same test. There have been studies that show certain Big 5 factor scores correlate with certain outcomes—conscientiousness correlates with longer life, for instance, and extroversion correlates with higher sales for sales reps. “But that doesn’t mean someone with high extroversion will be a better salesperson,” Vazire says. Correlations are just that; they could be incidental. But commercial personality assessments seem to depend heavily on such correlations. For example, one assessment from The Predictive Index, a company that measures behavioral characteristics and matches personality profiles to jobs, views such correlations in their own studies as a measure of success. “[We showed] in one client, a retail jeweler, that increases in dominance or aggression was responsible for $125,000 in revenue,” says Thad Peterson, one of the company’s executives. The idea behind The Index, Peterson says, is to use those measures to help “marry people to [job] positions.”

Such personality assessments—particularly those targeted toward hiring recruiters and managers—aim to uncover a kind of “hidden truth about the person,” says Randy Stein, a psychologist at California Polytechnic State University, Pomona. “They assume that there is an essence of you and an essence of the job, and you should be matching up those two things in hiring,” he says. “But I don’t think there is a hidden truth—and even if there is, a personality test doesn’t do it.”

Like the Big 5 model, any personality or behavior assessment can’t know things you haven’t explicitly answered in the questionnaire, Stein says. Sometimes commercial personality tests ask odd questions—like, Do you identify with snakes? or How do you react to a certain color?—and try to draw inferences from your answers. Those kinds of conclusions venture into the pseudoscientific, Stein says.

There are other reasons why Stein thinks some personality assessments may be pseudoscientific. “What those tests will tell people is true or false is determined by what people are willing to pay for,” he says. “Their process as a company is to tell people whatever will sell the product.” By contrast, the Big 5 and HEXACO models were shaped by an empirical process and independent peer review that showed people’s scores tended to be consistent, and predictions made using the models are reproducible. Without that, Stein says personality tests should be treated with extreme suspicion.

Some companies like The Predictive Index say their product meets such standards. The company invested in an audit, paying over $20,000 dollars to Norwegian classification firm DNV GL to review their product and certify that it complies with a standard set by the European Federation of Psychologists’ Associations. Two Index representatives, Greg Barnett and Austin Fossey, also say predictions based on their methods are accurate.

Perhaps. U.C. Davis’s Vazire says it is fairly easy to reach some level of validity. “If I just asked you to make a questionnaire on extroversion, you would probably do a pretty good job,” she says. It is because we are all judges of character, and we often do well at intuiting whom to date or hire and who we are, Vazire says. If the process seems confusing or if questions veer off into the abstract, that’s a red flag. Personality, she says, is just not that mysterious.

The SAPA Project

Take the test., explore your personality., advance the study of individual differences..

• Start the test

FAQ about the test

We won't sort you into a house or match you to a harmonious date. But we will give you feedback based on modern psychological theory. Learn more...

The research behind SAPA

Each customized report is generated on the basis of participant's responses, but each participant gets a slightly different subset of all 7,000 items. Learn more...

Individual Differences

Why do individuals differ in the ways they think, feel, and act? How do people differ in response to the same situations? Learn why individual differences matter ...

   Recent Blog Posts

• Parsimony in Personality
• Myers-Briggs Type Indicator® (MBTI®)
• Presidential Candidacy & Personality
• see more...

Another personality test?

   The SAPA Project is a collaborative research tool for studying patterns of human behavior. Our goal is to find patterns among the vast number of ways that people differ from one another in terms of their thoughts, feelings , interests , abilities , desires, values, and preferences. Personality psychologists have been theorizing about these domains of differences for decades but the data you provide will help us to empirically integrate these domains into a single comprehensive framework.

   The test on this site will give you feedback about your personality on two levels. First, you will receive scores on the 27 personality dimensions that were identified through statistical analyses of 300,000 SAPA participants. Then we will also give you scores on the so-called Big Five factors of personality plus an overall cognition score. Most people spend 15 to 25 minutes on the full test (you can quit in the middle if you want but the feedback won't be as accurate). The test is entirely free (absolutely no strings attached) and completely anonymous (no cookies, no tracking, nothing).

   If you enjoy this test, please share it with others. The quality of our research improves with each participant, so we're grateful for your involvement!

Temperament

When thinking about the ways that people "typically" behave, the term temperament often comes up. This is...

Tools for assessing cognitive ability were one of the first contributions of psychological science. Despite this...

Learn about the most widely-cited personality models.

• "Big Five" / Five-Factor Model of personality, HEXACO Six-Factor Model of personality , NEO Personality Inventory™ (NEO-PI-R™) , Eysenck Personality Questionnaire (EPQ) , RIASEC Model of Interests , Cattell-Horn-Carroll (CHC) Model , V-P-R Model, Theories of Emotional Intelligence

Looking for more info?

• Discover My Profile includes Big Five personality tests , Jungian typologies , a Well-Being assessment , several intelligence tests , and lots more.
• The Your Personality site includes several tests, including one for attachment styles and changes in personality over time .
• The IPIP NEO personality test gives scores for the public-domain representation of the Five Factor Model, with a long version (300-items) and a short one (120-items) .

Vocational psychologists focus on individual differences in career choice. But people pursue interests outside of work too -- we watch different shows, listen to different types of music, and "like" different content on social media. Time spent on...

Myers-Briggs Type Indicator® (MBTI®) | Official Myers Briggs Personality Test

Myers-briggs type indicator® (mbti®).

Results that engage and inspire

The Myers-Briggs Type Indicator ® (MBTI ® ) assessment is one of the world’s most popular personality tools—because it works. Used by more than 88 percent of Fortune 500 companies in 115 countries, and available in 29 languages, it has become the go-to framework for people development globally. With more than 70 years of science-based, research-based insight, the MBTI assessment is a robust tool for self-awareness and improvement. It provides positive language for understanding and valuing individual differences. With practical insight that’s easy to understand and implement, the MBTI assessment has helped thousands of organizations and millions of people around the world improve how they communicate, learn, and work.

Powering productivity

The MBTI assessment can transform how people work together. Only 14 percent of executives believe that the traditional model of a hierarchical organization is effective. Forward-looking leaders are moving to a flexible, team-focused model. * MBTI insights enhance personal development, supporting team and leadership training, conflict management, career change, and transitions. Multinational companies value the common language the Myers-Briggs assessment provides for discussing interpersonal differences, making it the preferred choice for training and development programs around the world.

*  2017 Deloitte Human Capital Trends: Rewriting the rules for the digital age

How the MBTI assessment works

MBTI Personality Preferences

• How do you direct and receive energy—by focusing on the outside world, interacting with people and taking action, or by focusing on your inner world and reflecting on ideas, memories, and experiences?
• How do you take in information—by focusing on what you perceive using your five senses or by seeing the big picture and looking for relationships and patterns? 
• How do you decide and come to conclusions—by logically analyzing the situation or by considering what’s important to the people involved?
• How do you approach the outside world—in a planned, orderly way or a more flexible, spontaneous way? 

Your natural preferences in these four areas sort you into one of 16 distinct MBTI personality types. Understanding these types gives you objective insight that you can use to enhance your professional and personal relationships, as well as your direction, focus, and choices. 

The Four Myers-Briggs Preference Pairs

Get mbti certified.

GET MBTI CERTIFIED

Take the MBTI Assessment

Take the MBTI assessment online and discover your personality type.

TAKE THE MBTI ASSESSMENT

What our customers are saying

We've leveraged the MBTI ® framework to enhance our organization's culture through team development and applying the MBTI concepts to common growth areas, i.e. dealing with stress, change, etc.

— Gary Kinser, HR Manager, Toastmasters International

At FedEx, we’re leveraging the MBTI ® assessment to impact organizational culture by championing deeper dives past just the MBTI Profile, using some of the more robust reports to provide participants with more insight.

— Sam Haskins, Learning and Development Consultant, FedEx Corporation

The MBTI ® framework is foundational to our Leadership Development, which has a significant impact on our organization’s culture.

— Tracy Burpee, Director, Organizational Development, Augusta Medical Center

We use MBTI ® assessment about ten times per year and it always sparks amazing conversation. It brings such self-awareness to our students that the MBTI certified trainers here find it priceless.

— Allison Gross-Ebert, Training Manager, US Department of State

Myers-Briggs

Reviewed by Psychology Today Staff

The Myers-Briggs Type Indicator (MBTI) is an assessment of personality based on questions about a person’s preferences in four domains: focusing outward or inward; attending to sensory information or adding interpretation; deciding by logic or by situation; and making judgments or remaining open to information. The MBTI was initially developed in the 1940s by Katharine Cook Briggs and her daughter, Isabell Briggs Myers, loosely based on a personality typology created by psychoanalyst Carl Jung.

When responses are scored, the assessment yields a psychological “type” summarized in four letters, one for each preference: Extraversion (E) or Introversion (I); Sensing (S) or Intuiting (N); Thinking (T) or Feeling (F); and Judging (J) or Perceiving (P). The results combined into one of 16 possible type descriptions, such as ENTJ or ISFP.

While the MBTI is used by many organizations to select new personnel and has been taken millions of times, personality psychologists and other scientists report that it has relatively little scientific validity. Psychologists who investigate personality typically rely on scientifically developed assessments of traits clustered into five (the Big Five ) or six ( HEXACO ) domains.

• Is the Myers-Briggs Legitimate?
• Why People Love Personality Tests

Why do experts take issue with the MBTI? One reason is that while the Myers-Briggs assigns people distinct types, scientific evidence indicates that personalities do not fit neatly into 16 boxes.

Traits are more accurately viewed not as categorical dichotomies—extrovert or introvert, thinker or feeler—but as continuous dimensions: For each trait, an individual can rate relatively high, low, or somewhere in the middle, and most people fall in the middle. Personality tests favored by scientists, such as the Big Five Inventory, describe each personality not in categorical terms, but rather based on how high or low a person scores on each of five (or six) non-overlapping traits.

The MBTI’s type-based feedback is also not especially consistent; a person who takes the test twice may well receive two different type designations. Moreover, the MBTI omits genuine aspects of personality that have negative connotations, such as neuroticism (emotional instability) or facets of low conscientiousness . It is untrue that the MBTI measures nothing at all, however. Research suggests that when MBTI preferences are evaluated as continuous dimensions, rather than split into categories, there is some correlation with scores on the Big Five traits.

The MBTI’s type for any one individual is often not consistent over time: People may take the test on multiple occasions and receive different personality types, even if they have not changed drastically in real life. Research has found that over a period of only a few weeks, up to half of participants received two different type scores.

Developers of the MBTI even acknowledged that in their sample, 35 percent received a different type after a four-week period. And despite the use of the MBTI in work settings, research does not suggest that the MBTI types are especially good predictors of job outcomes.

Forced choice fails to capture the dimensional nature of personality. The MBTI’s scoring format places individuals into one of each pair of categories regardless of how extreme their scores are. A person who scores a 53 percent on the introversion - extraversion dimension receives the same result as someone who scores 95 percent: Both are labelled “extravert.” The person who scores 53 percent, however, is probably much more similar to the “introvert” who scored just below the 50 percent mark.

Personality “types,” therefore, miss a lot of information; characterizing everyone as either an introvert or an extravert glosses over the reality that most people actually land somewhere in the middle of the spectrum.

The notion that personality is completely fixed from birth isn’t accurate, and it can be valuable to possess flexibility in how people view themselves and their ability to evolve. But when people take a personality test, they may adopt that label and incorporate it into their identity and life narrative. Labels can be limiting, which is why it’s important to acknowledge the limited nature of the Myers-Briggs itself .

The MBTI has been used in an array of domains. Companies have used it to hire and organize employees. Career advisors have used it to recommend which professions individuals should explore. Premarital counselors have used it to gauge the “compatibility” of couples. However, there is little evidence to support the connection between the Myers-Briggs and outcomes such as job performance or relationship success.

Despite its limits as a valid personality assessment, the Myers-Briggs can be a valuable tool for self-reflection. Taking a fun personality test can serve as a starting point for people to consider how they view their personality, how human behavior can vary, and how they relate to others in their life. The MBTI can provide an initial vocabulary from which to expand.

Personality tests have almost become ubiquitous. A high school guidance counselor may assign students to take a personality test to determine which colleges to apply to. Corporations may administer tests for hiring decisions or team-building activities. Personality tests may lead friends to bond over a shared “ personality type ,” find others like them, or put words to different dimensions of character.

The desire to understand ourselves better and categorize the world around us helps drive the popularity of the Myers-Briggs and others like it.

People are endlessly fascinated by personality tests. This may be because people seek hidden information about themselves, wanting to understand and access their true nature. People also have an inherent desire to belong; identifying with a “type” can help people feel normal and understood—there are similar people out there. People also appreciate simple ways to categorize and interact with other people.

The Myers-Briggs often delivers results that aren’t entirely reliable—so why do people trust them? One reason for this illusion of accuracy is confirmation bias : When people believe something is true, they begin to filter information based on that belief. People may also love the feeling of being recognized, like the test “gets” them. The results are fairly general which makes them widely applicable, and they skew positive so people are often happy to accept them.

The MBTI is perhaps the most well-known, but other popular personality tests include the Enneagram, which assigns personality descriptions based on nine primary types and often secondary types called “wings,” and the DISC, assessments that assign individuals one of four types, or a blend of the types: Dominance (D), Influence (I), Steadiness (S), and Conscientiousness (C).

Psychologists generally agree that tests of the Big Five personality traits are most valid. The current version of this assessment is the Big Five Inventory-2 (BFI-2).

Most personality psychologists use tests that measure the Big Five personality traits: extraversion, agreeableness , conscientiousness, emotional stability , and openness to experience . These five traits represent five categories of individual characteristics that tend to cluster together in people.

This framework has the benefit of 1) being developed with the scientific method 2) using continuums rather than categories 3) showing how people change over time 4) predicting outcomes that personality should predict, such as life satisfaction, education and academic performance, job performance and satisfaction, relationship satisfaction and divorce , physical health, how long people live, and more.

A nine-part animated series about the Enneagram.

Do you have ADHD and feel like you are always swimming upstream? Understanding yourself, prioritizing self-care, and focusing on your strengths can bring you more ease.

Thinking of using a personality test to help you sort out important career decisions? You may want to think again.

It is not unusual for the last pair of chatters in a room to include one introvert who secretly has been attempting to escape the conversation for some time.

A new assessment divides decision-making into four styles. Which works best for you?

Juliette Swann discusses her company's unique approach to online dating.

Would Santa Claus be an outgoing introvert or reclusive extravert? Is he energized by the smiling faces, or is a feeling that people drain him the reason he sneaks in the dark?

Have you ever taken a personality test? A new book reviews the scientific evidence showing how your personality actually changes, and then shows you how to change it deliberately.

Stereotyping and ignorance can often be a case of unchecked heuristics.

The next time someone tells you the Myers-Briggs has been debunked, ask them for a citation. Then hand them this one.

• Find a Therapist
• Find a Treatment Center
• Find a Psychiatrist
• Find a Support Group
• Find Teletherapy
• United States
• Brooklyn, NY
• Chicago, IL
• Houston, TX
• Los Angeles, CA
• New York, NY
• Portland, OR
• San Diego, CA
• San Francisco, CA
• Seattle, WA
• Washington, DC
• Asperger's
• Bipolar Disorder
• Chronic Pain
• Eating Disorders
• Passive Aggression
• Personality
• Goal Setting
• Positive Psychology
• Stopping Smoking
• Low Sexual Desire
• Relationships
• Child Development
• Therapy Center NEW
• Diagnosis Dictionary
• Types of Therapy

Understanding what emotional intelligence looks like and the steps needed to improve it could light a path to a more emotionally adept world.

• Coronavirus Disease 2019
• Affective Forecasting
• Neuroscience

• Create new account

The TypeFinder® Personality Test

Beyond briggs myers' 16 types, find your true strengths., 898,479 tests taken in the last 30 days.

This free personality test reveals who you really are. Discover the 16 personalities created by Myers & Briggs, test your personality type, and find your strengths.

To take the personality test, mark your answers based on how well each statement describes you.

Questions? Learn more about the personality test.

TypeFinder Personality Test FAQ

Q . what is this personality test based on.

A. This test is based on the personality theory created by Isabel Myers and Katharine Briggs. It measures your preferences on Myers and Briggs' four dimensions of personality type, as well as 23 more detailed facets of type to personalize your results.

Q. How long is this personality test?

A. The test consists of 130 questions and takes about 10-15 minutes to complete.

Q. Is this personality test really free?

A. You do not need to purchase or register to take this test and view an overview of your results. If you would like, you can purchase a more comprehensive full report for a small fee.

Q. Is this personality test accurate?

A. This test has been researched extensively to ensure it is valid and reliable, using a variety of statistical methods. These results are detailed in the TypeFinder Technical Documentation . Most of our users describe their results as both accurate and insightful. However, it is important to note that no test can determine personality type correctly for everyone—it's essential that you evaluate your results on your own to decide if they describe you well, and research other possible types if necessary.

Q. What will my results for this test look like?

A. You will first see a brief, free report showing the key points from your results. After reviewing your brief report, you then have the option to unlock your full report for a small fee. To see what you can expect from your full report, see this sample report .

Q. How can I access my personality test results?

A. After you take a test, you will have the option to create an account by entering your email address. If you create an account, you can view your test results at any time by returning to Truity.com and logging into your account. We do not email your results to you.

Q. Do I need to complete this personality test all at once?

A. If you’ve created an account and are logged in when you take the test, your responses will be saved as you go through the test. If you do not log in to a Truity account before starting the test, your progress will not be saved and you will need to complete the test all at once.

Q. Can I have my employees, team or group take the TypeFinder test?

A. Absolutely. Our Truity@Work platform is designed to make it easy to give a TypeFinder personality test to your team or group. See discounted group pricing and learn how to quickly and easily set up testing for your group on the Testing for Business page.

Q. Will this test tell me which careers are best for my type?

A. This test has brief information about the careers for your type, but if you main goal is to find the right career for you, then we recommend you take the TypeFinder for Career Planning , which is specifically designed to help you find the right career for your type as well as your individual interests and strengths.

Q. Is this personality test appropriate for children?

A. None of our tests are appropriate for children under the age of 14. Some of our tests may have mature content, and anyone younger than 18 should only take the test with parental guidance.

Q. Where can I find more information about the 16 personalities?

A. You can find comprehensive profiles of each of Myers and Briggs' personality types here: INFP • INFJ • INTP • INTJ • ENFP • ENFJ • ENTP • ENTJ • ISFP • ISFJ • ISTP • ISTJ • ESFP • ESFJ • ESTP • ESTJ

Q. Can my personality type change over time?

A. If you asked Isabel Briggs Myers and Katherine Briggs (the creators of the 16 personality types) or Carl Jung (the psychologist whose theories Briggs and Myers studied), they would say no, a person's personality type does not change over time. However, personality psychologists who study large populations have found that shifts in personality do occur over time. Research shows that age and individual life experiences can cause a shift in your personality. However, drastic shifts in personality are unusual, and most people find that changes are small and gradual.

Q. I'm looking for the official MBTI® assessment. Is this it?

A. The MBTI® is the original assessment developed by Isabel Myers and Katharine Briggs. The TypeFinder® is based on Myers and Briggs' theory, but is not the same as the MBTI® assessment. Some key differences:

The MBTI® Assessment

• Developed by Isabel Briggs Myers
• Based on theories of C.G. Jung, Katharine Briggs and Isabel Myers
• Measures 4 preferences of personality type
• Available through certified practitioners or online
• Results cost $49 (for MBTI® Online )

The TypeFinder®

• Developed by Truity
• Based on Myers and Briggs' theory and original empirical research
• Measures 4 dimensions and 23 facets of personality type
• Available online
• Results are free, or choose to purchase an expanded report

Q. Are you going to sell my data?

A. . We do not sell your email or other data to any third parties, and we have a zero-spam policy. We carefully comply with applicable privacy laws in handling your personal information. You can read more in our privacy policy .

Myers-Briggs Type Indicator, Myers-Briggs, and MBTI are registered trademarks of The Myers & Briggs Foundation in the United States and other countries. Truity has no affiliation with the organizations publishing or holding rights to the MBTI® assessment.

Get Our Newsletter

Recommended tests

Statistical "Which Character" Personality Quiz : This tool will compare your answers to a database of 2,000 fictional characters. The database is made by crowd-sourcing ratings of the characters, and the goal is to match people to characters they will agree are similar to them using techniques from recommendation engines. There also is a peer report verison, which is even more advanced. And a version for couples .

Other tests

Psychometric Media Recommendation Engine : Get reccomendations for TV, movies, and books based on personality questions.

Fisher Temperament Inventory : The FTI is general measure of personality that traces human behavior back to the function of the neurotransmitters in the brain. It categorizes people into one of four temperaments, each of which is associated with specific neuro-chemicals.

Inventory of Phonetic Associations (Experimental) : Humans are biased towards making associations between sound and meaning in specific ways that appear to be consistent across cultures. This test measures how strongly you show the typical bias, and explores what that means. Research has indicated that individuals with autism show less of this bias than neurotypicals.

Firstborn Personality Scale : This test was desgined to produce the maximum possible difference between scores of first-born (oldest) and later-born children. It correlates with birth order more than any other self-report scale, but the correlation is still extremely small because most of the common claims about the effect of birth order on personality are exaggerated and wrong.

Analog to Multiple Broadband Inventories : Most personality tests ask the same kind of questions, they just organize their results in different ways. This one computes all the scores you would likely get if you took 8 different well regarded personality tests, from just one bank of items.

Multidimensional Introversion-Extraversion Scales : The idea of introversion and extraversion is one of the oldest and most well known ideas in personality psychology. The evidence indicates that people can mean several different things when they describe themselves as an introvert or extravert, so the trait of introversion-extraversion should actually be broken down into a couple different, though related, traits.

OSPP Enneagram of Personality Scales : The Enneagram of Personality is a system of nine personality types organized by a geometric diagram. It has been promoted as a spiritual and self-help tool by many authors and there exist several different popular tests of Enneagram type. The OEPS was developed by this website and reflects the average idea of what each type is in the population of on-line Enneagram enthusiasts.

Zodiac-sign Associated Personality Scales : The ancient practice of astrology connects the way a person is to their date of birth. This survey will determine if your personality matches the stereotypes for your astrological sign.

Protestant Work Ethic Scale : There is sociological theory that Northern European countries developed faster in the industrial revolution than southern ones because of the additudes towards work promoted by Protestantism (versus Catholicism). This idea has been taken by some psychologists who believe that individuals can have different levels of Protestant work ethic.

Exposure Based Face Memory Test : Measure of face memory and face blindness.

Vocabulary IQ Test : Vocabulary test giving an IQ score like result.

Updated: 15 October 2019

Feedback: [email protected]

Privacy policy

Sentino Personality API – Psychology NLP

Personality API, Big Five

Trait-based Personality Tests

Trait-based personality tests – validity, reliability, relevance.

The validity, reliability, and relevance of personality tests are all important considerations when evaluating their usefulness and accuracy. There are several different approaches to personality testing, but trait-based personality tests are currently the most widely-used. They focus on identifying an individual’s specific personality traits .

Trait theory is one of several theories of personality that are commonly employed in psychology and scientific research. It is based on the idea that individual differences in personality can be described by a set of stable, enduring traits that are relatively consistent across situations and over time. These traits are believed to be biologically and genetically influenced. Besides, they are thought to be relatively stable throughout an individual’s lifetime.

One reason why scientific tests are often based on trait theory is that it allows researchers to study personality in a systematic and objective way. By measuring traits, researchers can quantify individual differences in personality and compare them across individuals and groups. This can help researchers understand how personality traits are related to other psychological variables, such as cognition, emotion, and behavior.

Another reason why trait theory may be preferred in scientific research is that it has good predictive validity. This means that trait scores are often able to accurately predict an individual’s behavior in a variety of situations. Thus, research has shown that trait measures of conscientiousness and agreeableness are predictive of job performance. At the same time, trait measures of extraversion and openness are predictive of social and occupational success.

Personality Test Validity

Test validity refers to the extent to which a test measures what it is intended to measure. In other words, it refers to the accuracy and meaningfulness of the test results. There are several types of validity that can be evaluated for a test.

Validity Types

Content validity refers to the extent to which a test covers the full range of the construct that it is intended to measure. For example, a test of mathematical ability should include items that assess a wide range of mathematical skills, from basic arithmetic to advanced topics.

Construct validity refers to the extent to which a test is related to other measures of the same or related constructs. For example, a test of intelligence should be related to other measures of intelligence, such as an IQ test.

Predictive validity refers to the extent to which a test is able to predict important life outcomes. Among such outcomes, job performance and relationship satisfaction can be named. Thus, a test might be considered to have good predictive validity if it is able to accurately predict an individual’s job performance.

Criterion-related validity refers to the extent to which a test is related to a specific criterion, or outcome measure. A test of reading comprehension that accurately predicts an individual’s reading ability might be considered to have good criterion-related validity.

Validity Evaluation

There are several ways to evaluate the validity of trait-based personality tests. One way is to compare the scores on a test to scores on other measures of the same trait. If a test is measuring the trait of extraversion, researchers might compare scores on the test to scores on other measures of extraversion – self-report surveys or observer ratings. If the scores on the test are highly correlated with scores on these other measures, it suggests that the test is measuring the trait of extraversion in a valid way.

Another way to evaluate the validity is to examine the predictive validity of trait-based personality. If scores on the test are able to accurately predict these outcomes, it suggests that the test is measuring real, meaningful differences in personality.

Finally, researchers can also evaluate the validity of trait-based personality tests by examining their factor structure. This involves examining whether the items on the test cluster together in a way that makes theoretical sense. Consistency with other research on the trait being measured is equally important.

Personality Test Reliability

Test reliability refers to the extent to which it produces the same results over time and under different conditions. Scientists differentiate several types of test reliability.

Reliability Types

Test-retest reliability refers to the consistency of scores on a test when it is administered to the same group of people on two or more occasions. For example, the scores on a test of intelligence administered to the same group of people once and then a week later should be highly consistent if the test is reliable.

Internal consistency reliability refers to the consistency of scores within a single administration of a test. If a test has items that are all measuring the same construct, the scores on these items should be highly consistent if the test is reliable.

Inter-rater reliability refers to the consistency of ratings made by different raters. If multiple people are rating an individual’s performance on a task, their ratings should be highly consistent if the task is a reliable measure of the construct being assessed.

There are several ways to evaluate the reliability of trait-based personality tests.

Reliability Assessment

One way to assess reliability is to examine the test-retest reliability of the measure. This involves administering the test to the same group of people on two or more occasions. Thus, the consistency of the scores over time can be examined. If the scores are highly consistent over time, it suggests that the test is reliable.

Another way to evaluate the reliability of tests is to examine their internal consistency. This involves examining the consistency of scores within a single administration of the test. If a test is measuring the trait of conscientiousness, researchers might examine the consistency of scores on items that are related to conscientiousness. If the scores on these items are highly consistent, it suggests that the test has good internal consistency.

Researchers can also evaluate the reliability of trait-based personality tests by examining their inter-rater reliability. This involves having multiple raters score the same set of responses and examining the consistency of their ratings. If the ratings are highly consistent, it suggests that the test is reliable.

Personality Test Relevance

Test relevance refers to the extent to which a test is applicable to the specific situation/context in which it is used. In other words, it refers to the extent to which the test measures constructs that are meaningful and important in the particular context.

Test relevance is an important consideration in determining the usefulness and validity of a test. A test that is not relevant to the specific context may not accurately measure the construct it is intended to assess.

For example, a test of reading comprehension may be highly relevant and useful for assessing the reading abilities of students in a primary school, but it may not be as relevant or useful for assessing the reading abilities of adults who are working in a technical field. In this case, a test that is specifically designed to assess technical reading skills may be more relevant and useful.

Relevance Measurement

One way to assess relevance is to examine the content of the test and determine whether it covers the full range of the construct that it is intended to measure. For example, a test of mathematical ability should include items that assess a wide range of mathematical skills.

Another way to evaluate the relevance of tests is to examine their construct, predictive and criterion-related validity as described above. Highly-valid tests are considered ti be reliable.

The validity, reliability, and relevance of personality tests are important considerations when deciding whether to use them. It is critically important to choose a test that has been shown to be valid, reliable, and relevant for the specific purpose. There has been a great deal of scientific research conducted on trait-based personality tests.

As a result, trait-based personality tests are widely considered to be effective measures of personality. This is explained by the fact that they have good reliability and validity. There is extensive scientific evidence to support their use as a measure of personality in research and applied settings. Thus, research has consistently found that people tend to score consistently on trait-based personality tests over time and across different situations. Additionally, these tests have been found to be related to other psychological constructs, such as job performance, relationship satisfaction, and mental health. It suggests that they may be measuring something meaningful and valid. Moreover, the tests accurately measure the traits that they are intended to assess. Today trait-based personality tests are widely used in research and applied settings to assess individual differences in personality.

Like any psychological assessment, trait-based personality tests can be influenced by various factors, which can affect the accuracy of results. We can name test-taking motivation and the respondent’s mood as an example. It is important to note that trait-based personality tests should not be used in isolation. They are only one tool that can be used to assess personality. That is why they should be considered alongside other sources of information.

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
• Front Behav Neurosci

Reliability and validity analysis of personality assessment model based on gait video

1 School of Electronic, Electrical and Communication Engineering, University of Chinese Academy of Sciences, Beijing, China

2 Institute of Psychology, Chinese Academy of Sciences, Beijing, China

3 School of Computer Science and Technology, University of Chinese Academy of Sciences, Beijing, China

Deyuan Chen

Tingshao zhu.

4 Department of Psychology, University of Chinese Academy of Sciences, Beijing, China

Associated Data

To protect the privacy of the participants, the original datasets in the article cannot be made public. If necessary, feature datasets of gait are available from the corresponding author on reasonable request. Requests to access the datasets should be directed to TZ, nc.ca.hcysp@uhzst .

Personality affects an individual’s academic achievements, occupational tendencies, marriage quality and physical health, so more convenient and objective personality assessment methods are needed. Gait is a natural, stable, and easy-to-observe body movement that is closely related to personality. The purpose of this paper is to propose a personality assessment model based on gait video and evaluate the reliability and validity of the multidimensional model. This study recruited 152 participants and used cameras to record their gait videos. Each participant completed a 44-item Big Five Inventory (BFI-44) assessment. We constructed diverse static and dynamic time-frequency features based on gait skeleton coordinates, interframe differences, distances between joints, angles between joints, and wavelet decomposition coefficient arrays. We established multidimensional personality trait assessment models through machine learning algorithms and evaluated the criterion validity, split-half reliability, convergent validity, and discriminant validity of these models. The results showed that the reliability and validity of the Gaussian process regression (GPR) and linear regression (LR) models were best. The mean values of their criterion validity were 0.478 and 0.508, respectively, and the mean values of their split-half reliability were all greater than 0.8. In the formed multitrait-multimethod matrix, these methods also had higher convergent and discriminative validity. The proposed approach shows that gait video can be effectively used to evaluate personality traits, providing a new idea for the formation of convenient and non-invasive personality assessment methods.

Introduction

Personality is a characteristic set of behavior, cognition and psychological state, with stability and persistence ( Corr and Matthews, 2020 ). Personality affects the behavior, mental states and subjective well-being of individuals ( Costa and McCrae, 1980 ; Baumert et al., 2017 ). Studies have shown that personality traits are even related to physical health factors such as obesity ( Jokela et al., 2013 ) and the risk of death ( Iwasa et al., 2008 ). In addition, personality is closely related to academic achievements and learning styles ( Komarraju et al., 2011 ), career choices and satisfaction ( Seibert et al., 2001 ; Kern et al., 2019 ), love quality and marriage relationships ( Holland and Roisman, 2008 ). Based on these issues, the need for a more convenient and objective user personality evaluation approach has become increasingly urgent.

Currently, many personality assessment methods are available, and questionnaires and scales are the most widely used measurement tools ( Bing et al., 2007 ; Rammstedt and John, 2007 ). However, self-reported questionnaires are not applicable in some occasions. For example, in the psychological assessment of job hunting or enrolment, the participants may have more motives for answering deceptively when interests are involved ( Ones et al., 1996 ). In addition, questionnaires are not suitable for situations where multiple measurements are required, because filling out the same questionnaire multiple times leads to practice effects.

As a natural and easily observed body movement, human gait conveys much information about emotions, cognition, intentions, and personality ( Matsumoto et al., 2015 ). Previous studies have shown that it is possible to establish a relationship between certain qualities of body motion and personality ( Koppensteiner and Grammer, 2010 ). The walking speed of a person in adulthood reflects, in part, the individual’s personality ( Stephan et al., 2018 ). Higher degrees of extroversion and conscientiousness are associated with faster initial walking speeds and lesser walking speed declines, while high neuroticism is manifested by slow walking ( Tolea et al., 2012 ; Agmon and Armon, 2016 ). Hand movements can also effectively express personality traits ( Wang et al., 2016 ). For instance, the more open a person is, the more violent their vertical arm movements and the more obvious the changes in their movement directions ( Koppensteiner and Grammer, 2010 ). In addition, thoracic and pelvic movements and the coordination of limbs are related to personality ( Satchell et al., 2017 ). For example, individuals with high neuroticism and low extroversion show decreased mobility and poor limb coordination ( LeMonda et al., 2015 ).

Although much evidence has shown that personality can be reflected by gait, a personality assessment method based on gait has not yet been fully established. Previous research has mainly focused on the statistical correlations between personality traits and gait. Sun et al. (2019) performed preliminary research explorations regarding the modeling of gait and personality traits, but they lacked a comprehensive evaluation of the model’s performance. Since personality contains many traits, it is necessary to establish a multidimensional model. Existing model evaluation methods, such as accuracy- or error-based approaches, cannot evaluate the correlations between the dimensions of the model. Therefore, we apply the reliability and validity evaluation method used for scales to a machine learning model. We use the correlations between the prediction scores of each dimension of personality obtained from models and actual scores from scales to calculate the model validity and use the correlations between the predicted scores of models based on the two halves of the input gait data to calculate the model reliability. This method has been proven feasible in the field of affective computing ( Park et al., 2015 ).

In addition, a practical method should also include convenient tools to record gait data. The need for expensive and complex facilities in previous studies, such as motion capture systems ( Mündermann et al., 2006 ; Leardini et al., 2007 ; O’Connor et al., 2007 ), smart wearable devices ( Tao et al., 2012 ), and Kinect ( Springer and Yogev Seligmann, 2016 ; Sun et al., 2018 ), made them unusable as real-life solutions. In real life, due to the popularity of cameras, we can easily obtain gait video. Many studies have shown that gait video data can be used to achieve efficient gait recognition ( Singh et al., 2018 ; Wan et al., 2018 ), providing ideas for us to use gait video to establish a personality assessment model. Because gait video is convenient, easy to obtain and non-invasive, we can carry out large-scale gait experiments.

This study uses ordinary cameras to record two-dimensional gait video, builds a multidimensional machine learning model, and explores the criterion validity, split-half reliability, convergent validity, and discriminant validity of the developed personality assessment model. The purpose is to provide a new convenient personality assessment approach.

Materials and methods

Data collection, participants.

The personalities of adult individuals tend to be stable ( Briley and Tucker-Drob, 2014 ). For meaningful evaluation, datasets should contain at least 30 participants and possibly more ( Hofmann et al., 2014 ). We recruited 152 adult participants without mental illness or physical disability, including 79 males (52%) and 73 females (48%) with an average age of approximately 23 years (SD = 1.07).

Collection process

Participants walked back and forth for 2 min in a rectangular area with a size of 6 m × 2 m according to their daily walking conditions. During this period, a camera was used to record the participants’ gait videos. The experimental setup for gait data collection is shown in Figure 1 .

Experimental setup for gait data collection. The camera was fixed on the side of a rectangular footpath with a size of 6 m × 2 m. Participants walked back and forth on the footpath.

After completing the gait collection process, the participants immediately filled out the Big Five Inventory (BFI-44) personality scale. The BFI consists of 44 items and five subscales: extraversion (8 items), agreeableness (9 items), conscientiousness (9 items), neuroticism (8 items), and openness (10 items) ( John et al., 1991 ). Each item of the BFI-44 is assessed on a five-point Likert scale, ranging from 1 (“disagree strongly”) to 5 (“agree strongly”). This study used the Chinese version of the BFI-44 scale. The range of Cronbach’s alpha was 0.698–0.807, and the test-retest reliability was between 0.694 and 0.770 ( Carciofo et al., 2016 ).

The above protocol was performed with permission from the Institutional Review Board of the Institute of Psychology, Chinese Academy of Sciences (approval number: {"type":"entrez-nucleotide","attrs":{"text":"H15010","term_id":"879830"}} H15010 ).

Data preprocessing

Key point extraction.

We used OpenPose to extract skeleton coordinates from the gait videos. OpenPose is a human posture recognition system that can detect key points of human body, hands, face, and feet ( Cao et al., 2021 ). This study used OpenPose to extract the two-dimensional coordinates of 25 key points of the body, as shown in Figure 2 .

The 25 key points of the human body in OpenPose. In the key point labels, L and R represent the left and right sides of the human skeleton, respectively. The XOY in the upper left is the original coordinate system, and the X ′ O ′ Y ′ in the middle is the new coordinate system formed after coordinate translation.

Data unification

The experimental setup in Figure 1 involved walking back and forth, so the gait video contained the participants’ front and back gaits. Related studies have shown that gait skeleton evaluation based on the front view is more accurate than that based on the back view ( Fang et al., 2019 ). The differences in the amounts of training data available for the participants can greatly affect the performance of machine learning models ( Luyckx and Daelemans, 2010 ). Therefore, we kept at least four complete gait cycles based on the front view for each participant. One gait cycle represented the process of one foot from leaving the ground until landing ( Baker and Hart, 2013 ). In this study, the gait frames of all participants were unified to 75 frames. This method has been used in many studies and has proven effective ( Sun et al., 2017 ; Zhao et al., 2019 ; Yeye et al., 2020 ).

Coordinate translation

In the XOY original coordinate system ( Figure 2 ), the coordinates of key points were greatly affected by the body shapes and positions of the participants, and their coordinate sequences changed irregularly, as shown in A and B of Figure 3 . The movement of the human center of gravity can be used to assess the stability of a person’s gait ( Iida and Yamamuro, 1987 ). The movement changes between the center of gravity and the center of the pelvis are very similar, and the movement of the pelvis during the gait is obviously related to the movements of the limbs and torso ( Whittle, 1997 ). Therefore, this study used the key points of MidHip (No. 8) as the coordinate origin to establish a new coordinate system X ′ O ′ Y ′ ( Figure 2 ). The coordinate translation formula is as follows:

Coordinate sequence of key points of gait. (A,B) Before coordinate translation, in the XOY coordinate system. (C,D) After coordinate translation, before filtering, in the X ′ O ′ Y ′ coordinate system. (E,F) After filtering. The left figures show the x-coordinate sequence of the LAnkle key point. The right figures show the y-coordinate sequence of the RHip key point.

where i = 1,2,…,24 and x i and y i represent the horizontal and vertical coordinates of point i in the XOY coordinate system, respectively. After coordinate translation, the coordinate sequence obeyed an obvious motion law, reflecting the periodicity of gait movement, as shown in C and D of Figure 3 .

Due to the interference of the video background, high-frequency noise was contained in the key point coordinates, as shown in Figure 3 . We used a template with a one-dimensional convolution kernel ([1,4,6,4,1]) to smooth the coordinate sequence, as shown in Figure 4 . The filtering formula is:

Filtering. The one-dimensional convolution kernel was [1,4,6,4,1].

where j represents the coordinate value of the jth frame before filtering, and j ′ represents the coordinate value after filtering. The time series data after filtering (E and F in Figure 3 ) were obviously smoother than the original data (C and D in Figure 3 ).

Feature engineering

Feature construction, interframe difference.

Gait is the dynamic change of the body exhibited while walking ( Minetti, 1998 ). The movement information of a gait is contained in the frame-to-frame changes shown in the corresponding gait video. We used the forward interframe difference method to focus on the key point coordinate changes between two adjacent frames. The forward difference formula is:

where f j+1 and f j represent the skeleton coordinates of the (j+1)th frame and the jth frame, respectively, and △ f j represents the difference between two adjacent frames.

Distance between joints

Gait requires coordinated movement among the joints of the body ( Murray, 1967 ; Luyckx and Daelemans, 2010 ). Therefore, gait research cannot examine the movement rules of a certain key point or joint in isolation. The distance between a pair of joints is a local motion unit composed of two joint points. We used the horizontal and vertical distances between joints to characterize the changes between two joint points. We found that the distance between joints had a corresponding meaning in terms of gait movement. For example, the distance between LWrist (No. 7) and RWrist (No. 4) represented the swing of the hands at the associated moment ( Donker et al., 2002 ; Park, 2008 ), and the distance between LAnkle (No. 14) and RAnkle (No. 11) represented the stride at that moment ( Whittle, 2014 ). We proposed 13 distances between joints, including 26 coordinate distances. See Table 1 for specific indicators.

Distances between joints.

*In the name, dist_a_b_x and dist_a_b_y, respectively, represent the x-axis distance and the y-axis distance between joints a and b, where a and b are key point (or joint) labels.

Angle between joints

The angle between joints is also an important indicator for measuring human gait movement ( Davis et al., 1991 ). The angle between joints is a local motion unit composed of multiple joints. We used the angle formed by 3 joint points to characterize the relative motion between multiple joints. For example, the angle between Nose (No. 0), Neck (No. 1), and RShoulder (No. 2) represented the tilt of the head at the associated moment, and the angle between RHip (No. 9), RKnee (No. 10), and RAnkle (No. 11) represented the bending movement of the right knee at this moment ( Seel et al., 2014 ). We proposed 10 angles between joints, as shown in Figure 5 . See Table 2 for specific indicators. It is worth noting that ∠ RNeck and ∠ LNeck both represented the tilt angle of the neck. However, when participants shrugged or slanted their shoulders, the two angles did not constitute a supplementary angle. As MidHip (No. 8) was eliminated during preprocessing, for ∠ LHip and ∠ RHip , we used angle _9_12_13 and angle _12_9_10 instead of angle _8_12_13 and angle _8_9_10, respectively.

Angle between the joints of the human body. Some unused key points have been omitted.

Angles between joints.

*In the name, angle_a_b_c represents the ∠abc composed of joints a, b, and c, where a, b, and c are key points (or joints) labels.

Wavelet transform

In frequency domain analysis, the Fourier transform is often used to observe signal spectra, but this transform is not suitable for analyzing signals whose frequencies change with time ( Oppenheim et al., 1996 ). The short-time Fourier transform developed on this basis realizes time-frequency localization by adding a moving window function, but a problem remains: the window function cannot change with the frequency ( Kwok and Jones, 2000 ). The wavelet transform overcomes the above shortcomings and can realize multiresolution analysis ( Daubechies, 1992 ). In the gait video, through observing the raw gait data, we found that there are differences in the movement amplitude and frequency of different key points. For example, RWrist (No. 4) has a larger movement amplitude than REye (No. 17), but REye has a faster movement frequency. In order to express the difference in frequency, we chose the wavelet transform and used the “haar” wavelet basis to decompose the gait skeleton coordinate sequence with 5 layers of wavelets. The source signal X is decomposed into:

where D 1 , D 2 , D 3 , D 4 , and D 5 are the high-frequency signals (or detail coefficient arrays) decomposed from the first to fifth layers, respectively, and A 5 denotes the low-frequency signal (or approximate coefficient array) decomposed from the fifth layer.

Feature extraction

For feature construction, we constructed a feature data pool, including gait skeleton coordinates, interframe differences, distances between joints, angles between joints, and frequency domain arrays of wavelet decomposition (left side of Figure 6 ). We extracted the time-frequency domain features based on the feature data pool. In Figure 6 , we used different colors to distinguish the changes of data. Different colors from left to right represented the original coordinate sequences, feature construction, feature extraction functions and extracted features, respectively.

Flow chart of feature extraction.

Time domain feature extraction

Since the gaits included both simple linear motions and complex non-linear processes ( Iqbal et al., 2015 ), we used 10 linear and non-linear functions for time domain feature extraction, as shown in Table 3 . Based on the time series data in the feature data pool, we obtained 1320-dimensional time domain features.

Time domain feature extraction functions.

*x represents time series data.

Frequency domain feature extraction

Due to the complexity of gait, some gait patterns cannot be distinguished in the time domain, but some laws can be reflected in the frequency domain ( Orović et al., 2011 ). We calculated the absolute maximum values, mean values, variances, and absolute energy of the 6 coefficient arrays after wavelet decomposition in the feature data pool and obtained 1152-dimensional frequency domain features.

Compared with the number of participants, the dimensionality of the 2472-dimensional time-frequency features obtained after feature extraction was too high, and this easily led to serious model overfitting ( Bishop, 2006 ). We used principal component analysis (PCA) to reduce the dimensionality of the standardized features. Then we used Sequential Forward Selection (SFS) to select 40 features accordingly based on the properties of different machine learning algorithms in the modeling. SFS added features to form a feature subset in a greedy fashion. At each stage, the estimator (machine learning algorithm) chose the best feature to add based on the cross-validation score of an estimator. While selecting features, the performance of the model was continuously improved. Finally, SFS selected a feature subset containing 40 features for each model. This subset enabled the model to achieve optimal performance.

Personality assessment is a regression task. Different regression algorithms have different characteristics. Linear regression (LR) is often used as a baseline model because of its fast calculation speed, low complexity and easy interpretation. Gaussian process regression (GPR) is widely used in time series analysis, and gait videos are time series data. Random forest regression (RFR) is a typical ensemble algorithm, which can deal with errors caused by imbalanced gait data. Support vector regression (SVR) is effective in high-dimensional gait time-frequency feature space. Therefore, we selected 7 typical machine learning regression algorithms for modeling, namely GPR, LR, RFR, SVR, where the SVR algorithm contains 4 kernel functions: “linear,” “poly,” “rbf,” and “sigmoid.” The kernel function directly determines the final performance of the SVR algorithm, but the selection of an appropriate kernel function has always been an unsolved problem ( Zhou, 2021 ), so we made 4 attempts with the SVR algorithm.

As shown in Figure 7 , the modeling process included three data streams, which contained 75 frames of complete data and odd-even split-half data. We used “all frames” to train the standardization, PCA, SFS, and algorithm models and applied these models to “odd frames” and “even frames.”

Modeling flowchart. The solid arrows indicate the data flows, and the dotted arrows indicate the applications of the model. “Odd frames” and “even frames” were divided from the first 74 frames among “all frames.” PCA, principal component analysis; SFS, sequential forward selection; GPR, Gaussian process regression; LR, linear regression; RFR, random forest regression. SVR-linear, SVR-poly, SVR-rbf, and SVR-sigmoid represent the SVR model using the “linear,” “poly,” “rbf,” and “sigmoid” kernel functions, respectively.

We trained and tested the model with 10 times of 10-fold cross validation and used criterion validity, odd-even split-half reliability, convergent validity, discriminant validity and root mean square error (RMSE) as the model evaluation indicators. RMSE is the most commonly used performance metric in regression tasks ( Zhou, 2021 ) and is defined as:

where Predicted n and Actual n represent the personality prediction score and personality scale score of the nth participant, respectively.

Criterion validity and split-half reliability

Using the BFI-44 scores as ground truth, criterion validity was assessed by examining the Pearson correlation coefficient between the model and scale scores for each personality trait. We chose the odd-even split-half reliability as the reliability indicator. The Pearson correlation coefficient between model prediction scores for the “odd frames” and “even frames” was used to evaluate the odd-even split-half reliability of the model. That is, the correlation coefficient between “odd predicted value” and “even predicted value” in Figure 7 was calculated to indicate reliability.

Convergent and discriminant validity

This study used a multitrait-multimethod matrix to explore the convergent and discriminant validity of the personality assessment models. The matrix was filled by Pearson correlation coefficients and included five personality traits (extraversion, agreeableness, conscientiousness, neuroticism, and openness) and two measurement methods (the BFI-44 scale and personality assessment model). Convergent validity was numerically the same as criterion validity. However, when evaluating convergent validity, the BFI-44 scores were no longer used as the calibration, but the BFI and the model were regarded as a method of measuring personality, respectively. Discriminant validity was assessed by comparing the magnitude of between-trait correlations (e.g., between extraversion and conscientiousness) within models with those within BFI-44.

Statistical analysis

To explore the relationships between the features used for modeling and the gait joint points, we mapped the optimal feature combination selected by SFS to 24 key points (MidHip was the coordinate origin and was eliminated during data preprocessing). For feature engineering, features were constructed based on the gait skeleton coordinates, so the mapping process was the inverse process of feature construction. In statistical analysis, we counted each key point with different weights according to the rules of feature construction, so as to avoid deviations in the results of statistical analysis due to different key points used in constructing features.

In Table 4 , r 1 represents the criterion validity, and r 2 represents the odd-even split-half reliability. Among the 7 algorithm models, the GPR and LR models had the best criterion validity (the mean r 1 values were 0.478 and 0.508, respectively, and their RMSE values were lower than other algorithms), and r 1 was above 0.4 for all personality traits. The other models exhibited unbalanced performance across different personality traits. In addition, except for that of the SVR-poly and RFR models, the split-half reliability of the other models was good, and the mean values of r 2 were above 0.8. In general, the GPR and LR models had the best performance with good criterion validity and split-half reliability, as shown in Table 4 . The results of the remaining models are listed in Supplementary Appendix A of the Supplementary Materials . (The appendices are in the Supplementary Materials , the same below).

Criterion validity and split-half reliability of the GPR and the LR personality assessment models.

RMSE, root mean squared error. r_1 represents the criterion validity, and r_2 represents the split-half reliability. All correlation coefficients are highly significant (p < 0.001).

In the multitrait-multimethod matrix (as shown in Tables 5 , ​ ,6), 6 ), the bold numbers represent the correlations between different methods for measuring the same trait, the italic numbers represent the correlations between different traits measured by the same method, and the numbers in the rectangular area (except those in bold) represent the correlations between different methods for measuring different traits.

Convergent and discriminant validity of the GPR personality assessment model (GPR-PAM).

E, extraversion; A, agreeableness; C, conscientiousness; N, neuroticism; O, openness. All correlation coefficients are highly significant (p < 0.001). Bold numbers represent the correlations between GPR-PAM and BFI-44 for measuring the same trait, the italic numbers represent the correlations between different traits measured by GPR-PAM or BFI-44.

Convergent and discriminant validity of the LR personality assessment model (LR-PAM).

E, extraversion; A, agreeableness; C, conscientiousness; N, neuroticism; O, openness. All correlation coefficients are highly significant (p < 0.001). Bold numbers represent the correlations between LR-PAM and BFI-44 for measuring the same trait, the italic numbers represent the correlations between different traits measured by LR-PAM or BFI-44.

Among all the models, the average convergence correlation of the optimal LR model was 0.508, and the mono-trait Pearson correlations between the assessment methods were extraversion: r = 0.427; agreeableness: r = 0.487; conscientiousness: r = 0.474; neuroticism: r = 0.580; and openness: r = 0.574 ( Table 6 ). The average convergence correlation of the GPR model ( r GPR _ BFI = 0.478, see Table 5 ) was close to that of the LR model, while that of the other models is poor ( r RFR _ BFI = 0.145, r SVRlinear _ BFI = 0.137, r SVRpoly _ BFI = 0.286, r SVRrbf _ BFI = 0.333, r SVRsigmoid _ BFI = 0.316, see Supplementary Appendix B ). In LR and GPR models, the bold numbers were significantly larger ( p < 0.001) than the values in the same column or row of the rectangular area, which showed that our models had good convergent validity.

The discriminant validity coefficients for each method were shown in italics. The average magnitudes (absolute value; the same below) of the discriminant validity coefficient of models were significantly lower than that of the BFI-44 scale ( r GPR = 0.245, r LR = 0.251, r RFR = 0.069, r SVRlinear = 0.050, r SVRpoly = 0.093, r SVRrbf = 0.164, r SVRsigmoid = 0.098, in the upper left triangle; r BFI = 0.448, in the lower right triangle; p < 0.001) in Tables 5 , ​ ,6 6 and Supplementary Appendix B . Because italics indicated the correlation between different traits, a small correlation coefficient indicated good discriminant validity. This showed that the models were relatively better than the BFI-44 at discriminating between traits. In addition, the average magnitudes of the convergent validity coefficients of the models were significantly greater than the average magnitudes of their discriminant validity coefficients ( r Model _ BFI = r Model , p < 0.001), indicating the models had good discriminant validity. However, the RFR, SVR-linear, SVR-poly, SVR-rbf, and SVR-sigmoid models had poor convergent validity, that is, poor predictive performance, which may lead to large deviations in the above discriminant validity. In summary, the LR and GPR models had relatively good convergent and discriminant validity.

We mapped the features of the GPR and LR models with good reliability and validity to the key points of gait (B and C in Figure 8 ). A dotted line divided 50% of the key points according to their statistics. In the top 50%, the intersection of the two models contained 11 key points, namely, No. 1, No. 3, No. 4, No. 5, No. 6, No. 7, No. 9, No. 10, No. 11, No. 12, and No. 13. We converted the ladder diagram into a heatmap (A and D in Figure 8 ). The larger the statistical value was, the darker the color, which meant that the corresponding key point had a higher contribution rate to modeling. Except for those of the head and feet, the key points of other body parts presented high heat values.

Feature and key point mapping. (B,C) Statistical ladder diagrams of the GPR and LR models. The horizontal axis represents the labels of the key points, the vertical axis represents the statistical values of the key points, and the dotted line represents the 50% key point division line. (A,D) Statistical heatmaps of the GPR and LR models.

The results of this study show that personality assessment models based on gait video have good performance in terms of criterion validity, split-half reliability, convergent validity, and discriminant validity. We constructed static features (i.e., gait skeleton coordinates) and dynamic features (including interframe differences, distances between joints, and angles between joints). Interframe differences reflect the changes between different frames, and the distances and angles between joints reflect the states between the joints in the same frame. This gait feature construction method integrating static and dynamic information has been verified in many studies ( Pratheepan et al., 2009 ; Tsuji et al., 2010 ).

In machine learning, the RMSE is generally used to measure the performance of a model. We found that the mean value of r 1 was inversely proportional to the mean RMSE value for all models (see Table 4 and Supplementary Appendix A ), which shows that it is reasonable to use the criterion validity as an evaluation indicator of model performance. Cronbach’s alpha is the most commonly used reliability statistic. Usually, Cronbach’s alpha is above 0.7, which is considered acceptable reliability ( Cronbach, 1951 ). The alpha coefficient of each trait in BFI-44 is close to 0.8 ( Hongyan et al., 2015 ). For the split-half model containing 74 frames of data, it is impossible to calculate all split-half cases. So we used the odd-even split-half reliability ( VandenBos, 2007 ) approximation as the evaluation index. Except for those of the SVR-poly and RFR models, the mean r 2 values in each dimension of the other models was higher than 0.8 (see Table 4 and Supplementary Appendix A ), indicating that the reliability evaluation of the model was reasonable.

The correlations between the traits in BFI-44 (the lower right italicized areas in Tables 5 , ​ ,6 6 and Supplementary Appendix B ) showed that extraversion, agreeableness, conscientiousness and openness had positive correlations with each other, and they had negative correlations with neuroticism. A similar pattern was found in the correlations between the dimensions of the GPR and LR models (the upper left italicized area in Tables 5 , ​ ,6), 6 ), but no similar patterns were found in the other models. Both reliability and validity studies of the BFI-44 scale ( Carciofo et al., 2016 ) and model-based personality assessment studies ( Park et al., 2015 ) have confirmed the existence of this pattern. Therefore, it has been proven again that the GPR and LR models have good performance in many aspects.

Personality traits are closely related to body movements ( Koppensteiner and Grammer, 2010 ; Thoresen et al., 2012 ). Some studies have shown that agreeableness and pelvic motion, as well as extraversion and thoracic motion, are positively correlated, and conscientiousness and thoracic motion are negatively correlated ( Satchell et al., 2017 ). This is consistent with our results; that is, Neck (No. 1), RShoulder (No. 2) and LShoulder (No. 5) on the thorax and RHip (No. 9) and LHip (No. 12) on the pelvis all yielded high heat values ( Figure 8 ). In addition, during walking, certain differences are observed between arm swings and leg strides ( DeVita et al., 1991 ). In our results, there were differences between RWrist (No. 4) and LWrist (No. 7) on the hands and between Rankle (No. 11) and LAnkle (No. 14) on the feet ( Figure 8 ). We found that the key points of the head and feet presented low heat values, which might be due to the fact that the changes in these parts (relative changes between internal key points) were not obvious compared with the limbs and trunk, with little individual differences. In previous studies on the relationship between personality and gait ( Tolea et al., 2012 ; Wang et al., 2016 ; Satchell et al., 2017 ; Stephan et al., 2018 ), gait mainly focused on the limbs and trunk with a large range of motion. The above shows that our models learned some of the kinematic characteristics of gait. However, we have only initially explored the interpretability of personality assessment models from one perspective. Further research is needed in the future.

Our research purpose was not to find a machine learning algorithm but to explore the feasibility of predicting personality based on gait video through machine learning modeling and to provide a new idea and method for automatic personality assessment. According to a literature exploration and our knowledge, this study is the first to measure the reliability and validity of machine learning models in the field of automatic personality assessment using gait. In multidimensional studies, measuring the reliability and validity of machine learning models helps to ensure that a model can truly discover the patterns of corresponding traits; this could not be achieved by previous machine learning evaluation methods.

Our study is still in its infancy and cannot completely replace a personality scale, but it has good prospects. Our method needs only 3 s of effective gait video to efficiently and conveniently realize personality assessment. Personality traits are highly stable after adulthood ( Cobb-Clark and Schurer, 2012 ) and are closely related to health statuses ( Jokela et al., 2013 ), occupational tendencies ( Kern et al., 2019 ), and academic achievements ( Komarraju et al., 2011 ). In future job searches, enrollments and other occasions, it will be necessary to make personality assessments for the examinee. By using our method, an evaluation result can be quickly obtained without interference and used to assist the examiner in decision making.

This study exhibits some limitations. First, the educational levels of the participants in the experiment were concentrated in the postgraduate stage, and the relatively concentrated cultural level of the group may have caused their personalities to be similar. Second, deep learning algorithms may improve model performance to a certain extent, but the number of participants in our study was small. Third, we used a single camera to obtain gait videos of participants’ fronts, backs and turns while they walked back and forth. However, only the frontal gait was used in the experiment, which led to data waste. In future studies, we will expand the scope and scale of recruitment, invite participants with large individual differences, and improve our research algorithms and data collection methods.

This study moves one step forward toward a non-invasive and low-cost personality assessment solution, which will have potential value in personality-related psychological intervention and behavioral decision making. Our experiments show that Big Five personality assessment models based on gait video have good criterion validity, split-half reliability, convergent validity, and discriminant validity. Our preliminary research provides new ideas for evaluating the performance of machine learning models with multidimensional psychological characteristics and points out a possible direction for constructing convenient personality assessment methods.

Data availability statement

Ethics statement.

The studies involving human participants were reviewed and approved by the Institutional Review Board of the Institute of Psychology, Chinese Academy of Sciences (approval number: {"type":"entrez-nucleotide","attrs":{"text":"H15010","term_id":"879830"}} H15010 ). The patients/participants provided their written informed consent to participate in this study.

Author contributions

YW, BL, TZ, and DC contributed to conception and design of the study. TZ collected and provided the data and directed the writing and research process of this manuscript. YW analyzed the data and wrote the first draft of the manuscript. DC and BL proposed improvements to the research method. All authors participated in the editing and reviewing of manuscripts, contributed to the article, and approved the submitted version.

This research was funded by the Key Research Program of the Chinese Academy of Sciences (No. ZDRW-XH-2019-4) and the Fundamental Research Funds for the Central Universities (No. E0E48922X2).

Conflict of interest

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

Publisher’s note

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

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fnbeh.2022.901568/full#supplementary-material

• Agmon M., Armon G. A. (2016). Cross-sectional study of the association between mobility test performance and personality among older adults. BMC Geriatr. 16 : 105 . 10.1186/s12877-016-0272-8 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Baker R., Hart H. M. (2013). Measuring Walking: A Handbook of Clinical Gait Analysis. London: Mac Keith Press. [ Google Scholar ]
• Baumert A., Schmitt M., Perugini M., Johnson W., Blum G., Borkenau P., et al. (2017). Integrating personality structure, personality process, and personality development. Eur. J. Pers. 31 503–528. 10.1002/per.2115 [ CrossRef ] [ Google Scholar ]
• Bing M. N., LeBreton J. M., Davison H. K., Migetz D. Z., James L. R. (2007). Integrating implicit and explicit social cognitions for enhanced personality assessment: a general framework for choosing measurement and statistical methods. Organ. Res. Methods 10 136–179. 10.1177/1094428106289396 [ CrossRef ] [ Google Scholar ]
• Bishop C. M. (2006). Pattern Recognition and Machine Learning. New York, NY: Springer. [ Google Scholar ]
• Briley D. A., Tucker-Drob E. M. (2014). Genetic and environmental continuity in personality development: a meta-analysis. Psychol. Bull. 140 1303–1331. 10.1037/a0037091 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Cao Z., Hidalgo G., Simon T., Wei S. E., Sheikh Y. (2021). Openpose: realtime multi-person 2d pose estimation using part affinity fields. IEEE Trans. Pattern Anal. Mach. Intell. 43 172–186. 10.1109/tpami.2019.2929257 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Carciofo R., Yang J., Song N., Du F., Zhang K. (2016). Psychometric evaluation of Chinese-Language 44-item and 10-item big five personality inventories, including correlations with chronotype, mindfulness and mind wandering. PLoS One 11 : e0149963 . 10.1371/journal.pone.0149963 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Cobb-Clark D. A., Schurer S. (2012). The stability of big-five personality traits. Econ. Lett. 115 11–15. 10.1016/j.econlet.2011.11.015 [ CrossRef ] [ Google Scholar ]
• Corr P. J., Matthews G. E. (2020). The Cambridge Handbook of Personality Psychology. Cambridge, MA: Cambridge University Press. [ Google Scholar ]
• Costa P. T., Jr., McCrae R. R. (1980). Influence of Extraversion and neuroticism on subjective well-being: happy and unhappy people. J. Pers. Soc. Psychol. 38 668–678. 10.1037//0022-3514.38.4.668 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Cronbach L. J. (1951). Coefficient alpha and the internal structure of tests. Psychometrika 16 297–334. 10.1007/BF02310555 [ CrossRef ] [ Google Scholar ]
• Daubechies I. (1992). Ten Lectures on Wavelets. Philadelphia, PA: Society for Industrial and Applied Mathematics. [ Google Scholar ]
• Davis R. B., Õunpuu S., Tyburski D., Gage J. R. A. (1991). Gait analysis data collection and reduction technique. Hum. Move. Sci. 10 575–587. 10.1016/0167-9457(91)90046-Z [ CrossRef ] [ Google Scholar ]
• DeVita P., Hong D., Hamill J. (1991). Effects of asymmetric load carrying on the biomechanics of walking. J. Biomech. 24 1119–1129. 10.1016/0021-9290(91)90004-7 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Donker S. F., Mulder T., Nienhuis B., Duysens J. (2002). Adaptations in arm movements for added mass to wrist or ankle during walking. Exp. Brain Res. 146 26–31. 10.1007/s00221-002-1145-2 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Fang J., Wang T., Li C., Hu X., Ngai E., Seet B. C., et al. (2019). Depression prevalence in postgraduate students and its association with gait abnormality. IEEE Access. 7 174425–174437. 10.1109/ACCESS.2019.2957179 [ CrossRef ] [ Google Scholar ]
• Hofmann M., Geiger J., Bachmann S., Schuller B., Rigoll G. (2014). The tum gait from audio, image and depth (Gaid) database: multimodal recognition of subjects and traits. J. Vis. Commun. Image Represent. 25 195–206. 10.1016/j.jvcir.2013.02.006 [ CrossRef ] [ Google Scholar ]
• Holland A. S., Roisman G. I. (2008). Big five personality traits and relationship quality: self-reported, observational, and physiological evidence. J. Soc. Pers. Relat. 25 811–829. 10.1177/0265407508096697 [ CrossRef ] [ Google Scholar ]
• Hongyan L., Jianping X., Jiyue C., Yexin F. A. (2015). Reliability meta-analysis for 44 items big five inventory: based on the reliability generalization methodology. Adv. Psychol. Sci. 23 755–765. 10.3724/sp.J.1042.2015.00755 [ CrossRef ] [ Google Scholar ]
• Iida H., Yamamuro T. (1987). Kinetic analysis of the center of gravity of the human body in normal and pathological gaits. J. Biomech. 20 987–995. 10.1016/0021-9290(87)90328-9 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Iqbal S., Zang X., Zhu Y., Saad H. M. A. A., Zhao J. (eds) (2015). “ Nonlinear time-series analysis of different human walking gaits ,” in Proceedings of the 2015 IEEE International Conference on Electro/Information Technology (EIT) , (Dekalb, IL: IEEE; ). [ Google Scholar ]
• Iwasa H., Masui Y., Gondo Y., Inagaki H., Kawaai C., Suzuki T. (2008). Personality and all-cause mortality among older adults dwelling in a japanese community: a five-year population-based prospective cohort study. Am. J. Geriatr. Psychiatry 16 399–405. 10.1097/JGP.0b013e3181662ac9 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• John O. P., Donahue E. M., Kentle R. L. (1991). Big five inventory. Berkeley, CA: University of California, Berkeley, Institute of Personality and Social Research. 10.1037/t07550-000 [ CrossRef ] [ Google Scholar ]
• Jokela M., Hintsanen M., Hakulinen C., Batty G. D., Nabi H., Singh-Manoux A., et al. (2013). Association of personality with the development and persistence of obesity: a meta-analysis based on individual-participant data. Obes. Rev. 14 315–323. 10.1111/obr.12007 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Kern M. L., McCarthy P. X., Chakrabarty D., Rizoiu M. A. (2019). Social media-predicted personality traits and values can help match people to their ideal jobs. Proc. Natl. Acad. Sci. U.S.A. 116 26459–26464. 10.1073/pnas.1917942116 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Komarraju M., Karau S. J., Schmeck R. R., Avdic A. (2011). The big five personality traits, learning styles, and academic achievement. Pers. Individ. Dif. 51 472–477. 10.1016/j.paid.2011.04.019 [ CrossRef ] [ Google Scholar ]
• Koppensteiner M., Grammer K. (2010). Motion patterns in political speech and their influence on personality ratings. J. Res. Pers. 44 374–379. 10.1016/j.jrp.2010.04.002 [ CrossRef ] [ Google Scholar ]
• Kwok H. K., Jones D. L. (2000). Improved instantaneous frequency estimation using an adaptive short-time fourier transform. IEEE Trans. Signal Process. 48 2964–2972. 10.1109/78.869059 [ CrossRef ] [ Google Scholar ]
• Leardini A., Benedetti M. G., Berti L., Bettinelli D., Nativo R., Giannini S. (2007). Rear-foot, mid-foot and fore-foot motion during the stance phase of gait. Gait Posture 25 453–462. 10.1016/j.gaitpost.2006.05.017 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• LeMonda B. C., Mahoney J. R., Verghese J., Holtzer R. (2015). The association between high neuroticism-low extraversion and dual-task performance during walking while talking in non-demented older adults. J. Int. Neuropsychol. Soc. 21 519–530. 10.1017/s1355617715000570 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Luyckx K., Daelemans W. (2010). The effect of author set size and data size in authorship attribution. Lit Linguist. Comput. 26 35–55. 10.1093/llc/fqq013 [ CrossRef ] [ Google Scholar ]
• Matsumoto D. E., Hwang H. C., Frank M. G. (2015). Apa Handbook of Nonverbal Communication. Washington, DC: American Psychological Association. [ Google Scholar ]
• Minetti A. E. (1998). The biomechanics of skipping gaits: a third locomotion paradigm? Proc. Biol. Sci. 265 1227–1235. 10.1098/rspb.1998.0424 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Mündermann L., Corazza S., Andriacchi T. P. (2006). The evolution of methods for the capture of human movement leading to markerless motion capture for biomechanical applications. J. Neuroeng. Rehabil. 3 : 6 . 10.1186/1743-0003-3-6 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Murray M. P. (1967). Gait as a total pattern of movement. Am. J. Phys. Med. 46 290–333. [ PubMed ] [ Google Scholar ]
• O’Connor C. M., Thorpe S. K., O’Malley M. J., Vaughan C. L. (2007). Automatic detection of gait events using kinematic data. Gait Posture 25 469–474. 10.1016/j.gaitpost.2006.05.016 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Ones D. S., Viswesvaran C., Reiss A. D. (1996). Role of social desirability in personality testing for personnel selection: the red herring. J. Appl. Psychol. 81 : 660 . [ Google Scholar ]
• Oppenheim A. V., Willsky A. S., Nawab S. H., Hernández G. M. (1996). Signals and Systems. London: Pearson. [ Google Scholar ]
• Orović I., Stanković S., Amin M. A. (2011). New approach for classification of human gait based on time-frequency feature representations. Signal Process. 91 1448–1456. 10.1016/j.sigpro.2010.08.013 [ CrossRef ] [ Google Scholar ]
• Park G., Schwartz H. A., Eichstaedt J. C., Kern M. L., Kosinski M., Stillwell D. J., et al. (2015). Automatic personality assessment through social media Language. J. Pers. Soc. Psychol. 108 934–952. 10.1037/pspp0000020 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Park J. (2008). Synthesis of natural arm swing motion in human bipedal walking. J. Biomech. 41 1417–1426. 10.1016/j.jbiomech.2008.02.031 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Pratheepan Y., Condell J. V., Prasad G. (eds) (2009). “ The use of dynamic and static characteristics of gait for individual identification ,” in Proceedings of the 2009 13th International Machine Vision and Image Processing Conference , (Dublin: IEEE; ). [ Google Scholar ]
• Rammstedt B., John O. P. (2007). Measuring personality in one minute or less: a 10-item short version of the big five inventory in English and German. J. Res. Pers. 41 203–212. 10.1016/j.jrp.2006.02.001 [ CrossRef ] [ Google Scholar ]
• Satchell L., Morris P., Mills C., O’Reilly L., Marshman P., Akehurst L. (2017). Evidence of big five and aggressive personalities in gait biomechanics. J. Nonverbal. Behav. 41 35–44. 10.1007/s10919-016-0240-1 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Seel T., Raisch J., Schauer T. (2014). Imu-based joint angle measurement for gait analysis. Sensors (Basel) 14 6891–6909. 10.3390/s140406891 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Seibert S. E., Kraimer M. L., Crant J. M. (2001). What do proactive people do? A longitudinal model linking proactive personality and career success. Pers. Psychol. 54 845–874. 10.1111/j.1744-6570.2001.tb00234.x [ CrossRef ] [ Google Scholar ]
• Singh J. P., Jain S., Arora S., Singh U. P. (2018). Vision-based gait recognition: a survey. IEEE Access. 6 70497–70527. 10.1109/ACCESS.2018.2879896 [ CrossRef ] [ Google Scholar ]
• Springer S., Yogev Seligmann G. (2016). Validity of the kinect for gait assessment: a focused review. Sensors (Basel) 16 : 194 . 10.3390/s16020194 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Stephan Y., Sutin A. R., Bovier-Lapierre G., Terracciano A. (2018). Personality and walking speed across adulthood: prospective evidence from five samples. Soc. Psychol. Pers. Sci. 9 773–780. 10.1177/1948550617725152 [ CrossRef ] [ Google Scholar ]
• Sun B., Zhang Z., Liu X., Hu B., Zhu T. (2017). Self-esteem recognition based on gait pattern using kinect. Gait Posture 58 428–432. 10.1016/j.gaitpost.2017.09.001 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Sun J., Wang Y., Li J., Wan W., Cheng D., Zhang H. (2018). View-Invariant gait recognition based on kinect skeleton feature. Multimed. Tools Appl. 77 24909–24935. 10.1007/s11042-018-5722-1 [ CrossRef ] [ Google Scholar ]
• Sun J., Wu P., Shen Y., Yang Z., Li H., Liu Y., et al. (eds) (2019). “ Relationship between personality and gait: predicting personality with gait features ,” in Proceedings of the 2018 IEEE International Conference on Bioinformatics and Biomedicine (BIBM) , (Madrid: IEEE; ). [ Google Scholar ]
• Tao W., Liu T., Zheng R., Feng H. (2012). Gait analysis using wearable sensors. Sensors (Basel) 12 2255–2283. 10.3390/s120202255 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Thoresen J. C., Vuong Q. C., Atkinson A. P. (2012). First impressions: gait cues drive reliable trait judgements. Cognition 124 261–271. 10.1016/j.cognition.2012.05.018 [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Tolea M. I., Costa P. T., Jr., Terracciano A., Ferrucci L., Faulkner K., Coday M. M., et al. (2012). Associations of openness and conscientiousness with walking speed decline: findings from the health, aging, and body composition study. J. Gerontol. B Psychol. Sci. Soc. Sci. 67 705–711. 10.1093/geronb/gbs030 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Tsuji A., Makihara Y., Yagi Y. (eds) (2010). “ Silhouette transformation based on walking speed for gait identification ,” in Proceedings of the 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition , (San Francisco, CA: IEEE; ). [ Google Scholar ]
• VandenBos G. R. (2007). Apa Dictionary of Psychology. Washington, DC: American Psychological Association. [ Google Scholar ]
• Wan C., Wang L., Phoha V. V. (2018). A survey on gait recognition. ACM Comput. Surv. 51 : 5 . 10.1145/3230633 [ CrossRef ] [ Google Scholar ]
• Wang Y., Tree J. E. F., Walker M., Neff M. (2016). Assessing the impact of hand motion on virtual character personality. ACM Trans. Appl. Percept. 13 : 2 . 10.1145/2874357 [ CrossRef ] [ Google Scholar ]
• Whittle M. W. (1997). Three-dimensional motion of the center of gravity of the body during walking. Hum. Move. Sci. 16 347–355. 10.1016/S0167-9457(96)00052-8 [ CrossRef ] [ Google Scholar ]
• Whittle M. W. (2014). Gait Analysis: An Introduction. Oxford: Butterworth-Heinemann. [ Google Scholar ]
• Yeye W., Deyuan C., Baobin L., Xiaoyang W., Xiaoqian L., Tingshao Z. (2020). Predicting personality based on self-introduction video. IFAC PapersOnLine 53 452–457. 10.1016/j.ifacol.2021.04.217 [ CrossRef ] [ Google Scholar ]
• Zhao N., Zhang Z., Wang Y., Wang J., Li B., Zhu T., et al. (2019). See your mental state from your walk: recognizing anxiety and depression through kinect-recorded gait data. PLoS One 14 : e0216591 . 10.1371/journal.pone.0216591 [ PMC free article ] [ PubMed ] [ CrossRef ] [ Google Scholar ]
• Zhou Z.-H. (2021). Machine Learning. Singapore: Springer. [ Google Scholar ]

We took the world’s most scientific personality test—and discovered unexpectedly sexist results

• Personality
• Personality Traits
• Personality/Social

Personality tests are both incredibly popular and largely bogus. BuzzFeed made its name in part by publishing quizzes telling readers which ‘90s kid they are , which Friends character they are , which Disney princess they are , and…well… which Disney princess they are, really . None of these have any scientific basis. Then there’s the somewhat more reputable Myers-Briggs test, inspired by Jungian theories about personality types. Some 2.5 million people take it every year, and 88% of Fortune 500 companies use it. Despite its reputation, however, the Myers-Briggs has poor scientific validity .

There is one personality test that is far and away more scientifically valid than any of the others: the “Big Five.”

The Big Five evaluates personality by measuring—as the name suggests—five personality traits: openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism, each on a continuous scale. Studies have shown it that it effectively predicts behavior , and the test is often used in academic psychological personality research. People who score higher in conscientiousness tend to work harder , for example, while more neurotic personalities are more prone to anxiety and depression .

Despite its scientific validity, and even with the contemporary fascination with personality tests, the Big Five is relatively unpopular outside of academia. A recent FiveThirtyEight article on the subject suggested that personality scientists haven’t effectively marketed the one credible personality test.

But there are serious concerns not just with the marketing of the test, but with how it’s presented to a public audience. Despite the scientific rigor around the Big Five in academia, many online versions of the test are designed to give sexist results.

Read the whole story (subscription may be required): Quartz

I like doing personality test for fun however logically I find they are only accurate for how how the person is feeling at the moment they take the test or where they’re at in life We are constantly changing so I think our personalities are constantly changing like somebody going and taking a test for depression if they are going though a major life change the test will most likely say they have major depression however if that person taking the test on a “good” day with little stress could come up with the result of dysthymia. I think it would be better to come up with a test that shows what a person’s morals or values are as I find this would give us a better understanding of who the person is are you a person who is empathetic are you a person who selfish a narcissist that’s what I want to know about a person’s personality and I find that those are things are core personality traits and never change no matter what a person’s going through or where they’re at in life.

APS regularly opens certain online articles for discussion on our website. Effective February 2021, you must be a logged-in APS member to post comments. By posting a comment, you agree to our Community Guidelines and the display of your profile information, including your name and affiliation. Any opinions, findings, conclusions, or recommendations present in article comments are those of the writers and do not necessarily reflect the views of APS or the article’s author. For more information, please see our Community Guidelines .

Please login with your APS account to comment.

Does Psychology Need More Effective Suspicion Probes?

Suspicion probes are meant to inform researchers about how participants’ beliefs may have influenced the outcome of a study, but it remains unclear what these unverified probes are really measuring or how they are currently being used.

Science in Service: Shaping Federal Support of Scientific Research  

Social psychologist Elizabeth Necka shares her experiences as a program officer at the National Institute on Aging.

A Very Human Answer to One of AI’s Deepest Dilemmas

Imagine that we designed a fully intelligent, autonomous robot that acted on the world to accomplish its goals. How could we make sure that it would want the same things we do? Alison Gopnik explores. Read or listen!

Privacy Overview