What is social desirability bias, and does it break personality tests?

What is social desirability bias?

Social desirability bias is the tendency to answer a question the way that reflects well on you, instead of the way that is actually true. On a personality test it shows up as choosing the more flattering end of a statement: agreeing that you are dependable and calm, disagreeing that you get irritable, whether or not either is your ordinary pattern. Psychologists sometimes call the same thing impression management when it is deliberate, and self-deception when the person believes their own flattering answer.

It is not a flaw particular to dishonest people. Almost everyone shades their answers toward the desirable end a little, which is why test builders plan for it rather than trust instructions alone to prevent it.

Why does it matter more in hiring than in self-understanding?

How much a flattered answer costs depends on who reads the score. In hiring, a trait score is used to predict how someone will behave in a role: Conscientiousness relates to job performance across occupations and criteria, and both Openness and Extraversion relate to training performance (Barrick & Mount, 1991). Once a candidate knows which traits an employer is likely to value, the incentive to answer toward them is obvious, and the resulting score can distort a decision that affects someone's livelihood. Personality tests in hiring covers what these correlations can and cannot support.

A test you take to understand yourself carries none of that incentive. Nobody is deciding whether to hire you based on your Wellington bands, so there is nothing external to answer toward, which removes the main reason this bias becomes a serious problem in the first place.

How do personality tests reduce social desirability bias?

Test builders cannot stop anyone from wanting to look good, so the usual approach is to make that wish harder to act on cleanly. Several techniques do most of the work.

• Neutral wording. A statement describes a behaviour rather than praising or criticising it, so agreeing is not a boast and disagreeing is not a confession.
• Several items per trait. One statement is easy to answer strategically. A dozen covering the same trait from different angles are harder to game consistently, and averaging cancels some of the noise any single flattered answer adds.
• Reverse-keyed items. Mixing statements that score in opposite directions means agreeing, or disagreeing, with everything no longer produces a uniformly flattering profile. It produces an inconsistent one that is easy to spot.
• Honest instructions. Saying plainly that there are no correct answers, and that the report is for the person answering rather than a gatekeeper, removes some of the reason to perform.

Building a scale from several items, in more than one direction, so a lucky or motivated guess on one item cannot carry the whole score, is basic practice in classical test theory (Nunnally & Bernstein, 1994). Personality test questions shows real example statements from the item bank and why some are reversed.

Can a personality test be faked, and is it worth it?

A self-report questionnaire can be answered strategically. Whether that is worth doing is a different question, and the evidence points to no. Self-report is genuinely good at capturing internal, low-visibility states, such as how anxious or self-doubting someone tends to feel, better than an outside observer usually can (Vazire, 2010). What self-report cannot do is see itself from the outside, which is the gap that lets a flattered answer go unnoticed by the person giving it, and it is also why a second view is worth having: people who know you well judge evaluative traits, such as how capable you come across, more accurately than you judge yourself (Vazire, 2010), and their ratings predict outcomes such as job performance about as well as self-ratings, and better for some outcomes (Connelly & Ones, 2010).

Introspection alone has real limits, because much of what shapes behaviour and personality runs outside conscious awareness, and looking harder inward does not fix that; watching your own behaviour and hearing how others see you are the routes that help (Wilson & Dunn, 2004). Self-report vs observer report covers this evidence in full. And on a test you take for yourself, answering toward the flattering end only costs you: the report is written from your answers, so a profile assembled from the person you would like to be describes exactly that, an imaginary person, and hands you advice suited to someone who is not you. There is no prize for a better-looking score and no one grading it. The only loser is the reader, who is you.

What other response biases show up on personality tests?

Social desirability bias has close relatives. All of them move an answer for reasons that have nothing to do with the statement itself.

How do you answer a personality test well?

The instructions that beat all four biases at once are short.

• Describe your typical self over the last year, not your best week and not the person you are trying to become.
• Go with the first reading of a statement, since the three-second answer usually beats the thirty-second one.
• Use the middle option only when it is true, not as a way to avoid choosing.
• Remember there is no audience. Nobody scores you against a standard, so there is nothing to perform for.
• Answer in one sitting, without checking what a partner or colleague might expect you to say.

How does Wellington measure this?

Wellington is a science-based personality assessment from Therabot Labs LLC. It measures personality as continuous traits, built on the Big Five and HEXACO models, and writes the results back as a personal report rather than a type.

The free Snapshot asks 60 statements across 6 dimensions (Extraversion, Agreeableness, Conscientiousness, Emotional Stability, Openness to Experience, Honesty-Humility), about 7 minutes, with several statements per trait and reverse-keyed items mixed throughout, the same techniques described above. Every trait is reported as a percentile against the reference sample. The three bands (Quiet below the 30th percentile, Balanced to the 70th, High above) and the Portrait names exist to make the result easier to talk about; the percentile is the measurement, and it is always shown alongside. Our norms are provisional while the norming sample is completed. The first ten questions are answered before anything is asked of you; after question ten you sign in with an email address so your answers are saved and your results can be shown. There is no card.

The Wellington Membership adds a further 336 statements to reach 90 traits, including every Big Five facet. Wellington is a wellness tool for reflection, not a medical or psychological diagnosis, and it does not replace care from a qualified professional. It is not designed for hiring, and you can export or delete your answers at any time.

Questions people ask

Can a personality test tell if you answered dishonestly?

Not with certainty. An unusually flattering pattern across every dimension at once can stand out, and reverse-keyed items make some inconsistency visible. On a self-understanding test the more useful question is not whether it can be caught but why you would want to, since the report is written from your own answers and only describes you accurately if they do.

Does social desirability bias ruin personality test results?

It shifts them rather than ruining them. Building each trait score from several statements, keyed in both directions, is standard practice in classical test theory (Nunnally & Bernstein, 1994), and it limits how far any one flattered answer can move a score. Most people are only mildly pulled toward the flattering end rather than answering falsely throughout.

Should I answer as the person I want to be, or as I am?

As you are. A report built from aspirational answers describes someone who does not exist and gives advice suited to that imaginary person rather than you. If a result surprises you, treat that as useful information rather than a reason to have answered differently.

Is social desirability bias why people sometimes need a second opinion on their personality?

It is one reason among several. People who know you well judge evaluative traits, such as how capable you come across, more accurately than you judge yourself, while you are the better judge of your own internal states (Vazire, 2010). Comparing notes with someone who knows you closes part of that gap either way.