It is a response bias: the tendency to answer in a manner that will be viewed favourably by others, independent of what is true.
Why it matters when the plan changes
Assessment at work is high stakes for the person taking it, and a rating scale invites them to describe the person they would like to be seen as. This is the ordinary pull toward a favourable self-account described as social-desirability bias: the tendency to answer in a manner that will be viewed favourably by others. The distortion is not usually deliberate deception, but it moves scores enough to change conclusions without anyone intending to mislead, because the bias grows exactly as the stakes do.
The tension is that instruments which resist the bias are harder to complete and harder to explain. Comparative judgment scales stimuli through pairwise comparisons rather than absolute ratings, which is why forced choice between two desirable statements removes the easy answer respondents would otherwise take. It feels uncomfortable to respondents precisely because it works. The comfort of a rating scale is bought with the accuracy of what it produces.
In practice
A candidate rates themselves highly on collaboration, decisiveness and attention to detail. All three are desirable and some of them pull against each other. The profile that results is flatteringly high across the board and distinguishes this candidate from nobody, which is what a rating scale under stakes tends to produce.
Evidence
Social desirability bias is a response bias in which people answer in ways that will be viewed favourably, over-reporting good behaviour and under-reporting bad.
Social-desirability bias, Wikipedia (2026)Ranking-based measurement scales stimuli through comparisons between pairs rather than absolute self-ratings.
Law of comparative judgment, Wikipedia (2026)
What it cannot tell you
Naming the bias does not tell you how much a given answer has been distorted, or which answers are affected. It identifies a direction, toward the favourable, without a way to measure it in a single response. Knowing the tendency exists does not recover the true score underneath it.
Questions
No, and treating it that way misses the point. Most of it is unconscious: an ordinary pull toward a favourable self-account that operates without any intent to deceive. Deliberate faking is a separate and smaller problem, and instruments that resist one generally resist the other.
No, which is what makes it damaging. If everyone inflated equally the ranking would survive. People differ in how strongly they respond to the pull, so the bias reorders candidates rather than simply shifting them all upward, and the reordering is invisible in the scores.
Mainly by removing the easy answer. Forcing a choice between equally desirable statements means a respondent cannot endorse everything, so their ranking carries information that a rating scale would have lost. Detection scales and consistency checks catch some of what remains.
Somewhat, and less than people hope. Respondents assess their own stakes regardless of what they are told, and in an employment relationship they are usually right to assume results may matter. Instrument design is a more reliable defence than reassurance.
Ask what the instrument does about response bias, whether it uses forced choice or rating scales, whether it carries detection measures, and what evidence exists that scores hold up under high-stakes conditions rather than only in research settings where nothing was at risk.