The Research Behind DISC

DISC was first described by William Moulton Marston, a Harvard-trained psychologist, in his 1928 book Emotions of Normal People.³ While most psychologists of his era focused on pathology, Marston was interested in something different: how do healthy, functional people behave in everyday situations?

Marston identified four primary behavioral dimensions that he observed across normal populations:

• Dominance: How a person responds to challenges and asserts control over their environment
• Inducement: How a person influences and persuades others (later called Influence)
• Submission: How a person responds to cooperation and steady environments (later called Steadiness)
• Compliance: How a person responds to rules, standards, and structure (later called Conscientiousness)

What made Marston's work distinctive was its descriptive approach. He was observing how people do behave, not prescribing how they should behave. This non-judgmental foundation remains central to DISC today: no type is better or worse. Each represents a different, equally valid way of engaging with the world.

Marston was also a Harvard-trained lawyer, the developer of the systolic blood pressure deception test (a forerunner of the polygraph, published in 1917),² and the creator of the comic book character Wonder Woman. His background gave him a lens for observing human behavior outside clinical settings.

Marston described the four dimensions but never created a formal assessment instrument. That work fell to the researchers and practitioners who followed him.

Scale and Adoption

Commercial DISC instruments have been in continuous use since the 1970s. Wide use does not make an instrument accurate, so the rest of this page is about the method.

Why Forced-Choice Matters

Most personality questionnaires use a Likert scale: "Rate how much you agree with this statement from 1 to 5." The problem is social desirability bias. People can easily identify the "good" answer and select it, even unconsciously.

Forced-choice methodology presents groups of statements and asks respondents to pick which is most and least like them. When every option is about equally desirable, there is no obvious "right" answer to pick. Research finds this reduces faking, though it does not remove it.¹⁰ WiredType's scenarios go one step further and ask you to rank all four options, from most like you to least. A full ranking gives all six paired comparisons in every scenario, including the order of the two middle options that most and least picks leave unknown.

The scoring method changes what a score means. Here is the difference between the most/least counting of classic DISC instruments and the ranking model our server uses, and the limit the two share today.

What Our Scores Mean Today

Our scores are ipsative. Each scenario asks you to rank four options, one per style, so every answer is a comparison between your own styles. The model fits four numbers to those comparisons, and your answers identify only the differences between the four. In the fitted scale the four numbers add to zero: three degrees of freedom, not four. Two people who rank the same way get the same four scores, however much of each style either person has.

So your scores describe your own pattern: which styles lead for you, which are close behind, and which sit lower. They do not place you on a scale against other people. A D score of 80 means D leads strongly among your four styles. It does not mean you are more dominant than 80 percent of people, and it does not mean your D matches another person's 80.

The free result, the full result and the report all show each style on its own scale from 0 to 100, with 50 in the middle. A result saved before we moved to this scale shows the four scores as shares that add up to 100. Both views come from the same fitted pattern. Neither view adds information the other lacks.

Why We Use the Model Anyway

Thurstonian item response theory, formalized by Brown and Maydeu-Olivares in 2011,⁶ is the model that can recover between-person (normative) scores from forced-choice answers. It does so only under conditions we do not meet yet: item parameters estimated from a calibration sample, and blocks whose options load on their styles in different strengths and directions.⁷ Every option in our bank currently carries the same assumed parameters, so the model reduces to its simplest form, a ranking model with no between-person anchor.

We chose it because it uses all four ranks in every scenario instead of two picks, because it gives a standard error for each score (that is where the free preview's confidence comes from), and because the same model, once calibrated on our own sample, is the path to scores that compare across people. That calibration has not run. Until it has, and until we have published what it showed, every WiredType score is a description of your own ordering. Section 08 says how we check results today and what we measure next.

Reliability and Validity

Published research on forced-choice personality instruments reports the following:

• Mapping onto the Big Five: Peer-reviewed research has compared a four-quadrant DISC-style assessment with a Five-Factor measure and reported correlations between the two, so the DISC styles overlap with established personality dimensions without being the same thing.¹
• Modest but meaningful criterion validity: A comprehensive meta-analysis of forced-choice Five-Factor inventories reported operational validity coefficients generally in the low-to-moderate range for work-related outcomes.¹¹ These are real effects, not large ones.
• Forced-choice design reduces social desirability bias and faking compared to Likert-scale formats, though it does not eliminate them.¹⁰

Those findings belong to the instruments that were studied. Reliability and validity belong to a specific instrument. Our consistency checks run across completed assessments and our retest checks across signed-in retakes (section 08), and we publish a reliability, validity or accuracy figure only with its sample size and date; none is published yet.

DISC is built on two independent behavioral axes. Where you fall on each axis creates your primary behavioral style.

The vertical axis captures pace and assertiveness. People toward the top (Outgoing) tend to be fast-paced, assertive, and comfortable taking action. People toward the bottom (Reserved) tend to be moderate-paced, thoughtful, and more deliberate.

The horizontal axis captures focus. People toward the left (Task-Oriented) prioritize results, logic, and accuracy. People toward the right (People-Oriented) prioritize relationships, harmony, and collaboration.

Where these axes intersect, four quadrants emerge:

• D (Dominance): Outgoing + Task-Oriented. Drives toward results with directness and confidence.
• I (Influence): Outgoing + People-Oriented. Energizes groups with enthusiasm and optimism.
• S (Steadiness): Reserved + People-Oriented. Creates stability through patience and supportiveness.
• C (Conscientiousness): Reserved + Task-Oriented. Ensures quality through analysis and precision.

The four primary types are the starting point, not the full picture. Every person has all four dimensions present in different proportions. Your behavioral profile is the mix of how strongly each dimension shows up in your behavior.

The interaction between your primary type (your strongest dimension) and your secondary type (the next strongest) creates behavioral patterns that DISC practitioners have described for decades.

Why Subtypes Matter

Consider two people who both score highest on Dominance. One has Influence as their secondary trait (Di), while the other has Conscientiousness as their secondary (Dc). Both are driven and results-oriented. But the Di is more expressive and persuasive in their approach, while the Dc is more analytical and exacting. They share the core drive but express it in different ways.

Our platform names 25 combinations: 4 pure types, 12 two-style blends, 4 three-style types, 4 types in which two opposite styles are both strong (written with a small p, such as SpD), and 1 balanced type (all four close together). They are the shapes one scoring rule produces, and they are common patterns practitioners describe. They are not categories we have validated in a study of our own (section 08).

How the 25 Are Assigned

We use one rule on your four scores. Each score runs from 0 to 100, and 50 is the middle. A style is in your type when it is above 50 and at least three tenths as far above 50 as your highest style. When no style reaches 60, your type is the balanced type (DISC). When one style is in, you have a pure type. When two are in, you have a blend, written with the higher style first (Di, Sc). Three styles make a three-style type (IDS), and all four make the balanced type (DISC). The free test and the 28-question assessment use the same rule. A result saved before this rule began keeps the rule it was typed on, under which a style joined the type at two thirds of the highest score.

Three-Style Types Versus Two Opposite Styles

D and S sit in opposite corners of the map (fast and task-focused against steady and people-focused), and so do I and C. A careful reader will notice that a three-style type such as IDS also contains a pair of opposites. Here is the distinction the rule draws.

A type written with a small p, such as SpD, has exactly two styles high, and those two are opposites with nothing between them. The D and the S of a DpS share no corner of the map, so the profile pulls in two directions with no bridge. The rule writes the small p only in that case, and only when the lower of the two is at least three quarters as far above 50 as the higher. When the second opposite is weaker than that, the result is a blend (Ds).

A three-style type has a bridge. In IDS the I shares pace with D and people focus with S, so the three form a connected arc across the map rather than two poles. The same holds for DIC (D bridges I and C), CDS (C bridges D and S) and SIC (S bridges I and C). The code starts with the style whose corner of the map the type sits in, the one between the other two, and the other two follow in D, I, S, C order.

The point of a behavioral model is not to place people in boxes. It is to give people a shared language for the differences in how they communicate, decide, and work together.

Any credible assessment platform should be clear about its boundaries. DISC is a powerful tool for understanding behavioral tendencies, but it is not a universal measure of personality, and it was never designed to be.

Key Principles

• DISC describes HOW you tend to behave, not WHO you are. Behavior is one layer of personality, alongside values, experiences, skills, and motivations.
• Behavior is context-dependent. You may be more Dominant in work settings and more Steady at home. Both are you.
• No type is better or worse. Every profile has distinct strengths and distinct growth areas.
• DISC describes how you tend to act and decide. It exists to increase self-awareness and improve communication. It should never be used to stereotype, limit, or exclude.

One reason this matters: the large longitudinal studies cited here find that personality traits stay moderately stable and still change over a lifetime. A meta-analysis of longitudinal studies found rank-order stability of personality traits is real but imperfect, and it rises gradually with age before leveling off in mid-adulthood.¹² A 50-year follow-up study found meaningful rank-order stability in the Big Five from adolescence into later adulthood, alongside real mean-level change.¹³ Those studies measured Big Five traits in whole populations. They did not measure DISC, and they do not predict how any one person will change. Three different questions hide in "will my type change": whether a trait score correlates with itself over time (what those studies measured), whether a type label comes out the same on a retake (a stricter test, since a label can flip on a small score change), and how WiredType's own assessment behaves on retest, which we measure on signed-in retakes and have not yet published (section 08). So we make no prediction about what your profile will look like years from now. The research supports one general point: behavior is moderately stable and still changes, so treat a profile as a description of now.

What every result rests on

• A model with a long history. The four styles come from Marston's 1928 model, and DISC assessments have been built on it since the 1940s (sections 01 and 02).⁴
• A full ranking in every scenario. The classic DISC format, used by the Personal Profile System of 1972, asks for two picks per block: the response most like you and the one least like you.⁵ Two picks leave the order of the other two responses unknown. You rank all four, so the scoring model gets all six comparisons between them in every scenario. Research on forced-choice formats, most of it on hiring tests, finds they reduce faking compared with rating scales, though they do not remove it.^{10, 15}
• Scores that show how settled they are. A published statistical model⁶ fits your four scores to all of your rankings and gives each one a standard error, an estimate of how far that score could move. The free result says in words how clear your lead is and marks where you sit on the DISC map. Section 07 explains what that estimate rests on.
• Your numbers, explained. Every result shows your own four numbers, and your report explains how they were calculated from your rankings, including the rule that names your type.
• Your result first. The free result appears on the page when you finish, and you do not need to give an email address to see it.
• Research only with your consent. Your results join our research only if you turn research sharing on, and then under a random research ID.

Checks across all results

• Consistency. Across completed assessments, we check how consistently each style's rankings hold together from scenario to scenario, and how each scenario lines up with the rest. These checks look at results together, with no names, to improve the questions and the scoring, as our privacy policy describes.
• Retest. When a signed-in person takes the assessment again at least a day later, we measure how much each score moved and whether the type stayed the same.
• What people tell us. We are adding short questions in two places. After the free result we ask, "How well does this describe you?" After the Full Profile we ask how accurately it describes you, how helpful it has been and how likely you are to recommend WiredType. Everyone who reaches the questions is asked, whatever their result, and nobody is rewarded for answering. People who work for WiredType or tested it for us are left out of the figures. Once 100 people have answered a question, its result will appear here with the question, the number of people asked, the number who answered and the dates. For the first three questions that is the share who chose one of the top two answers, and for the recommendation question it is the Net Promoter Score. These answers show how people see their results, which is a separate measure from the consistency and retest checks above.

Any consistency or retest figure we publish will use only results people chose to share for research. A consistency figure also needs the answers to individual questions, so it waits for the agreement described below.

What we measure next

Two studies come next: calibrating the questions so that scores can be compared from one person to another,⁷ and comparing our results with an established Big Five measure, a check that published DISC research has used.¹ Both need information research sharing does not collect today, so we will ask for your agreement before we collect it.

We publish a figure only with the number of people it comes from and the date.

Under a free result we ask "How well does this describe you?" After someone has read their Full Profile we ask how accurately it describes them, how helpful it has been, and how likely they are to recommend WiredType. We ask everyone who reaches the question, whatever their result, and we offer nothing for answering. Each person counts once, at the first result on which they were asked. We publish a figure only when at least 100 people have answered it, for the 12 whole months that end with the last complete month, and we update it once a month.

We leave out WiredType staff and test accounts, automated test browsers, sample results made for testing, invited first-round testers, people who joined through links from WiredType's own accounts, members of client teams, and partners, coaches and affiliates. For the Full Profile figures we also leave out anyone whose Full Profile was free or came through a private code or card. No one is left out because of how they answered, and each figure below lists how many people each group removed.

1. Jones, C.S., & Hartley, N.T. (2013). Comparing correlations between four-quadrant and five-factor personality assessments. American Journal of Business Education, 6(4), 459-470. DOI: 10.19030/ajbe.v6i4.7945. A peer-reviewed comparison of a four-quadrant DISC-style instrument with a Five-Factor measure.
2. Marston, W.M. (1917). Systolic blood pressure symptoms of deception. Journal of Experimental Psychology, 2(2), 117-163. DOI: 10.1037/h0073583. The deception test that made Marston a forerunner of the polygraph.
3. Marston, W.M. (1928). Emotions of Normal People. London: Kegan Paul, Trench, Trubner & Co. The foundational work establishing the four behavioral dimensions that became the DISC framework. Public-domain edition on WiredType.
4. Clarke, W.V. (1956). The construction of an industrial selection personality test. The Journal of Psychology, 41(2), 379-394. DOI: 10.1080/00223980.1956.9713011. The published construction of the Activity Vector Analysis, the first assessment instrument tied to Marston's model.
5. Geier, J.G. (1972). Personal Profile System. Minneapolis: Performax Systems International. The forced-choice DISC instrument that set the pattern for commercial DISC products.
6. Brown, A., & Maydeu-Olivares, A. (2011). Item response modeling of forced-choice questionnaires. Educational and Psychological Measurement, 71(3), 460-502. DOI: 10.1177/0013164410375112. The paper that formalized Thurstonian IRT for scoring forced-choice instruments.
7. Brown, A., & Maydeu-Olivares, A. (2013). How IRT can solve problems of ipsative data in forced-choice questionnaires. Psychological Methods, 18(1), 36-52. DOI: 10.1037/a0030641. The conditions under which forced-choice scores stop being ipsative: calibrated items and blocks keyed in different directions.
8. Maydeu-Olivares, A., & Brown, A. (2010). Item response modeling of paired comparison and ranking data. Multivariate Behavioral Research, 45(6), 935-974. Foundational statistical methodology for forced-choice scoring.
9. John, O.P., & Srivastava, S. (1999). The Big Five trait taxonomy: History, measurement, and theoretical perspectives. In L.A. Pervin & O.P. John (Eds.), Handbook of Personality: Theory and Research (2nd ed., pp. 102-138). Convergent validity reference for relating DISC dimensions to the Big Five.
10. Christiansen, N.D., Burns, G.N., & Montgomery, G.E. (2005). Reconsidering forced-choice item formats for applicant personality assessment. Human Performance, 18(3), 267-307. DOI: 10.1207/s15327043hup1803_4. Peer-reviewed review of forced-choice formats and their effect on faking and social desirability.
11. Salgado, J.F., & Tauriz, G. (2014). The Five-Factor Model, forced-choice personality inventories and performance: A comprehensive meta-analysis of academic and occupational validity studies. European Journal of Work and Organizational Psychology, 23(1), 3-30. DOI: 10.1080/1359432X.2012.716198. Meta-analytic evidence on the modest but real criterion validity of forced-choice personality instruments.
12. Roberts, B.W., & DelVecchio, W.F. (2000). The rank-order consistency of personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126(1), 3-25. DOI: 10.1037/0033-2909.126.1.3. Meta-analytic evidence on the stability (and imperfection of stability) of personality traits across the lifespan.
13. Damian, R.I., Spengler, M., Sutu, A., & Roberts, B.W. (2019). Sixteen going on sixty-six: A longitudinal study of personality stability and change across 50 years. Journal of Personality and Social Psychology, 117(3), 674-695. DOI: 10.1037/pspp0000210. A large 50-year longitudinal study of Big Five stability and change from adolescence into later adulthood.
14. Wright, A.G.C., Ringwald, W.R., Vize, C.E., Eichstaedt, J.C., Angstadt, M., Taxali, A., & Sripada, C. (2026). Assessing personality using zero-shot generative AI scoring of brief open-ended text. Nature Human Behaviour, 10(3), 541-555. DOI: 10.1038/s41562-025-02389-x. Seven AI models scored Big Five traits from people's writing, and the average across models agreed best with their questionnaires. Our writing-sample analysis follows its practice of averaging several AI readings; the study did not measure DISC.
15. Cao, M., & Drasgow, F. (2019). Does forcing reduce faking? A meta-analytic review of forced-choice personality measures in high-stakes situations. Journal of Applied Psychology, 104(11), 1347-1368. DOI: 10.1037/apl0000414. A meta-analysis that found much less score inflation on forced-choice measures than on single-statement measures, most of all when the statements in a block are balanced in social desirability.