Do You Answer Honestly on Personality Tests?
By PsyCove Editorial · 11 min read · Updated 2026-10-27 · Checked against the sources listed below
- Faking works when someone is told to fake. Across 51 studies, instructed faking shifted scores on every one of the Big Five, and no dimension was harder to fake than another (Viswesvaran & Ones, 1999).
- It matters less than the warnings suggest. Social desirability has been called a red herring for selection because it barely changes what the scores predict (Ones et al., 1996).
- The bigger error is not knowing. People are poor judges of their own anxiety and mood, and friends beat them on intellect (Vazire, 2010). That is not dishonesty, and no honesty instruction fixes it.
- You are also a moving target. Repeated state reports put each person across a wide band on every dimension, so two sittings can honestly disagree (Fleeson, 2001).
What Does “Honestly” Even Mean Here?
The word usually means three different things, and only one of them is a real problem.
- Not lying. Deliberately choosing an answer you know is false. Rare outside a high-stakes setting.
- Not exaggerating. Choosing the flattering answer because it is flattering. Very common, mostly unconscious, and not the same as lying.
- Not knowing. Choosing the answer that seems right about you and is not. The most common of the three, and the one an honesty instruction cannot touch.
| What went wrong | How common | What it does to the result |
|---|---|---|
| Lying | Rare without a stake | Moves the score, but usually in a direction the person could have reached by other means |
| Exaggerating | Common, mostly unconscious | Shifts a position a little. The shape of the profile tends to survive |
| Not knowing | The default condition | Moves the score in a direction you cannot detect, because being wrong about yourself does not feel wrong |
Most advice about personality tests is aimed at the first item, which is the rarest, and silent about the third, which is the largest source of error in a self-report. The ranking would be wrong if faking turned out to be the largest of the three sources; the evidence reviewed here says it is not.
Can You Fake a Personality Test?
Yes, easily, if someone tells you which way to fake. The interesting part is that the ability is uniform across traits.
A meta-analysis pooled 51 studies that instructed participants to respond in a particular direction. Faking produced substantial shifts in scores, and the amount did not vary by dimension: every one of the Big Five was about as fakable as the others (Viswesvaran & Ones, 1999).
Two details from that analysis are more useful than the headline. The largest distortions appeared on social desirability scales rather than on the trait scales, and instructions to fake good produced smaller effects than instructions to fake bad. The scales built to catch faking were the easiest thing to fake.
Does Faking Ruin the Result?
Less than you would think, and the reason is worth understanding: faking moves scores, but it moves them in a way that partly preserves the ordering between people.
The standard review of social desirability in personnel selection concluded that it functions as a red herring. People differ in how much they want to look good, and that difference is itself a stable trait, but it adds little to what the personality scores already predict (Ones et al., 1996).
The distinction that makes this coherent comes from separating two components of socially desirable responding. Enhancement is claiming positive qualities; denial is disclaiming negative ones. They are not the same disposition, and they do not distort the same scales (Paulhus & Reid, 1991).
So the honest summary is narrower than the folklore. Deliberate faking works and will change your result. It also tends to produce a result that is still useful for ranking people against each other, which is why employers who use these tests seldom catch it and seldom need to.
So What Is the Real Problem?
The real problem is that a self-report asks you to be your own informant, and on some traits you are a poor one.
In the round-robin study that has become the reference point, 165 people rated themselves and were rated by four friends and up to four strangers, with behavioral tests as the criterion. The self was the best judge of anxiety and mood — a trait description rather than a diagnosis, and not a clinical assessment of anyone. Friends were the best judge of intellect. On extraversion, everyone did equally well (Vazire, 2010).
Read that as a map of where your answers are trustworthy. Emotional life is the one region where you have access nobody else has. Anything that shows up in how you perform is the region where other people have better data than you do (Vazire & Mehl, 2008).
- Trust your own report on anxiety, mood, and how often you feel overwhelmed. This is the region where you have access nobody else has.
- Trust it less on anything that shows up in how you perform — quickness, competence, how clever you are in a room. Observers have better data there.
- Trust it about as much as anyone's on sociability and energy. A stranger who has watched you for an hour does as well as you do (Vazire, 2010).
None of this is dishonesty, and that is the point. You can answer every item in good faith and still be systematically wrong about one part of the map, because being wrong about yourself does not feel like being wrong.
Why Do Your Answers Change Between Sittings?
Because the thing being measured is not a fixed quantity, and because a single sitting samples one moment from a wide range.
When people report their current state repeatedly over weeks, each person occupies a broad band on every dimension rather than a point. The variation is not measurement error alone; it is the person responding to situations (Fleeson, 2001). A test taken on a Tuesday after a bad week samples the bottom of that band.
The instrument adds its own wobble, though less than people assume. A meta-analytic review of the indicator's score reliability found strong internal consistency and test–retest estimates across studies, with variation between them (Capraro & Capraro, 2002). The scores are steadier than the folklore, which means the change between your two results is more likely to be you than the test. And the ordering of people on a trait is more stable than the positions of individuals, so the population holds its shape while the people in it move (Roberts & DelVecchio, 2000).
And a small shift can flip a letter, because the letters are cut from continuous scores near a midpoint most people sit close to (Bess & Harvey, 2002). Two sittings that differ by four points can produce different four-letter results without either one being wrong.
What Actually Makes an Answer More Accurate?
Five adjustments, none of which involve trying harder to be honest.
- Answer for the last month, not for your best month. A specific window gives you something to recall instead of something to imagine.
- Answer for the median case. If you can name one memorable exception to an item, do not let that exception decide your answer.
- Do not aim for consistency. Trying to answer as the person you were last time imports a second task into the test, and it makes the result less informative rather than more.
- Take it twice, a few weeks apart, and keep both. The gap between them is a reading on your range, which is more useful than either number alone.
- Let someone who knows you well answer about you, then compare. Where the two of you disagree is where the interesting information lives, and the disagreement is usually about behavior rather than feeling.
Is There Any Point in a Lie Scale?
A lie scale is a set of items almost nobody would endorse honestly, used to flag people who endorse them. It catches something real, and that something is narrower than its name.
The problem is that the tendency to claim flattering things is a stable individual difference, so a lie scale mostly measures a trait rather than detecting a state. That is also why the scales designed to catch faking were the easiest ones to distort when people were told to fake (Viswesvaran & Ones, 1999; Paulhus & Reid, 1991).
- It would tell you how much a person wants to be seen well, which is already visible in the trait scores.
- It would not tell you whether this particular person answered this particular item falsely.
- It would add a fifth score to a page that reports four, and the fifth would be the least informative of them.
We do not put a lie scale in this test, for that reason. Our test reports a position on four dimensions and would mostly be recording how much a person wants to be seen well, which is not a useful fifth dimension to add. Our result page says the same thing in fewer words: a modest amount of self-flattery shifts a position. It does not turn the answer into a fiction.
What Happens to Your Answers After You Submit Them?
Most people answering a personality questionnaire think about the result rather than the record. In the European Union and the United Kingdom, the record is the part with legal weight.
Three rights do most of the work, and they are worth knowing before you start rather than after.
- Access and erasure. Under the GDPR, and under the UK GDPR as supervised by the Information Commissioner's Office, you can ask what is held about you and ask for it to be erased, subject to narrow exceptions. A site that cannot answer either request is not compliant, whatever its privacy page claims.
- Automated decisions. Article 22 restricts decisions taken about you solely by automated means where they produce legal or similarly significant effects. A pre-employment screen that rejects candidates on a score is the case that provision was written for.
- Employment context. A score held by an employer is employee data, and most EU member states attach extra conditions to it. In the United States the picture is a patchwork of state privacy statutes rather than one federal rule.
The useful habit takes one minute. Before you start a test, ask two questions out loud: "What exactly do you store about me, and how long do you keep it?" and "Who inside your company can see it, and how do I get it deleted?" A questionnaire that cannot answer both is asking for more than it needs.
How Should You Read Your Own Result, Then?
As one honest sample from a range, taken on a particular day, about a person who is not fully transparent to themselves.
That sounds deflating and is meant to be steadying. It removes two failure modes at once. You no longer have to worry that your result is a lie, because lying is the smallest of the three risks. And you no longer have to treat a letter as a verdict, because a sample from a range is not a verdict.
What the result is good for is the part you can check: whether a description matches what you actually did last week, and whether the people who know you would agree. Where they would not, you have found something more interesting than a letter. The one feeling to distrust is the sense that the whole description fits, since identical text handed to a room full of people is rated as personally accurate by most of the room (Forer, 1949).
Frequently asked questions
Can a personality test tell if you are lying?
Only weakly. Lie scales detect a general tendency to claim flattering things, which is a stable trait rather than a momentary lie (Paulhus & Reid, 1991). They also happen to be the easiest scales to distort deliberately (Viswesvaran & Ones, 1999).
Should I answer as the person I want to be?
No, but not for moral reasons. Answering as an ideal produces a result that describes an aspiration, and an aspiration does not predict how you will behave on a difficult day. Answer for the last month instead.
Why did I get a different type the second time?
Because the scores are cut into letters near a midpoint that most people sit close to, and because a single sitting samples one moment from a wide personal range (Bess & Harvey, 2002; Fleeson, 2001). A few points of movement is enough to change a letter.
Do personality tests work if people can fake them?
Faking shifts individual scores but has limited effect on what the scores predict about groups of people, which is why the review literature calls social desirability a red herring for selection (Ones et al., 1996). For describing one person accurately, the bigger problem is limited self-knowledge.
Is it better to have a friend fill it out about me?
For some traits, yes. Observers are as good as the self at extraversion and better at intellect-related traits, while the self remains the best judge of anxiety and mood (Vazire, 2010). Neither perspective is universally superior.
Further Reading (primary sources & authoritative references)
- Viswesvaran, C., & Ones, D. S. (1999). Meta-analyses of fakability estimates: Implications for personality measurement. Educational and Psychological Measurement, 59(2), 197–210. — DOI
- Ones, D. S., Viswesvaran, C., & Reiss, A. D. (1996). Role of social desirability in personality testing for personnel selection: The red herring. Journal of Applied Psychology, 81(6), 660–679. — DOI
- Paulhus, D. L., & Reid, D. B. (1991). Enhancement and denial in socially desirable responding. Journal of Personality and Social Psychology, 60(2), 307–317. — DOI
- Vazire, S. (2010). Who knows what about a person? The self–other knowledge asymmetry (SOKA) model. Journal of Personality and Social Psychology, 98(2), 281–300. — DOI
- Vazire, S., & Mehl, M. R. (2008). Knowing me, knowing you: The accuracy and unique predictive validity of self-ratings and other-ratings of daily behavior. Journal of Personality and Social Psychology, 95(5), 1202–1216. — DOI
- Fleeson, W. (2001). Toward a structure- and process-integrated view of personality: Traits as density distributions of states. Journal of Personality and Social Psychology, 80(6), 1011–1027. — DOI
- Capraro, R. M., & Capraro, M. M. (2002). Myers-Briggs Type Indicator score reliability across studies: A meta-analytic reliability generalization study. Educational and Psychological Measurement, 62(4), 590–602. — DOI
- Bess, T. L., & Harvey, R. J. (2002). Bimodal score distributions and the Myers-Briggs Type Indicator: Fact or artifact? Journal of Personality Assessment, 78(1), 176–186. — DOI
- Forer, B. R. (1949). The fallacy of personal validation: A classroom demonstration of gullibility. Journal of Abnormal and Social Psychology, 44(1), 118–123. — DOI
- Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of personality traits from childhood to old age. Psychological Bulletin, 126(1), 3–25. — DOI
Related reading
- How to find your true type — the practical version of the same question, minus the honesty framing.
- Two INFJs can be nothing alike — what a single sitting can and cannot tell you about one person.
- Where MBTI stops — the structural problems that no amount of honest answering can fix.
- Trait versus type — why a position survives a bad week better than a letter does.
- You can see where your own four positions sit on the test page.