A personality inventory measuring conscientiousness, administered to the same individuals once in a low-stakes research context and again under genuine job application conditions, produces measurably more favorable average conscientiousness scores under the job application condition — a well-documented pattern called faking good, where candidates motivated to present themselves favorably for a specific employment decision shift their self-reported responses in a socially desirable direction relative to how the same person would answer in a context without this specific evaluative pressure.
Why this happens even among candidates who aren't consciously lying
Faking good doesn't necessarily require deliberate, calculated dishonesty in every instance — candidates under genuine selection pressure often engage in a milder, less consciously deliberate form of favorable self-presentation, interpreting ambiguous personality inventory items through the lens of what a good candidate for this specific role would plausibly and reasonably say, rather than through purely candid self-reflection disconnected from the evaluative context they're actually in.
Why this specifically distorts the trait measurement the inventory is meant to provide
A personality inventory's usefulness for hiring decisions depends on scores accurately reflecting genuine, stable individual differences in the trait being measured — if candidates systematically shift their responses toward social desirability under selection pressure, the resulting scores partly reflect each candidate's skill at recognizing and presenting a favorable self-image, a genuinely different trait than the personality characteristic the inventory was actually designed to measure, potentially favoring candidates skilled at impression management over candidates with a more genuinely desirable underlying trait profile who simply answered more candidly.
What research comparing selection and non-selection contexts has actually found
Studies specifically designed to compare the same individuals' responses across genuine selection conditions and lower-stakes research or developmental contexts consistently find a measurable, systematic shift toward more favorable self-presentation under selection conditions, a pattern replicated across a range of specific personality traits and testing formats, indicating the effect is a general feature of high-stakes personality assessment rather than a rare or occasional occurrence.
What testing programs actually do to detect or reduce this distortion
Social desirability scales — additional items embedded within a personality inventory specifically designed to detect unusually favorable, implausibly consistent self-presentation patterns — can flag candidates whose overall response pattern suggests substantial faking, allowing for either adjustment or additional scrutiny of their results. Forced-choice item formats, requiring candidates to select which of several equally socially desirable-sounding statements is more true of them, rather than simply rating agreement with each statement independently, make faking good considerably more difficult to execute effectively, since the format removes the option of simply endorsing every favorable-sounding statement. Explicit warning statements informing candidates that response distortion can be statistically detected have also been found to measurably reduce the degree of faking observed.
What this means for organizations relying on personality assessments in hiring
- Recognize that personality inventory scores collected under genuine selection conditions are likely somewhat shifted toward social desirability relative to a candidate's more candid self-view
- Consider testing formats specifically designed to reduce faking, including forced-choice items and embedded social desirability scales
- Use explicit warnings about response distortion detectability where the testing platform supports it, given documented evidence this measurably reduces faking
- Treat personality assessment results as one input weighed against other information, rather than a fully candid, undistorted measure of a candidate's underlying trait profile
Faking good isn't primarily a story about dishonest candidates — it's a predictable, well-documented consequence of asking people to describe themselves candidly at precisely the moment they have the strongest possible motivation to describe themselves favorably instead, and good testing design has to account for this reality directly rather than assuming it away.