How we change what others think, feel, believe and do
Personality tests seek to identify - guess what - aspects of a person's personality that are correlated in some way with job performance.
Personality tests are often administered as self-completed sets of questions about preferences and behaviors each of which contributes towards a score or position along a number of personality dimensions, e.g.:
Any given score may be correlated with a particular job. For example, jobs that require significant interaction with people may have a correlation of extraversion with job success.
Personality tests are hard to validate and so are developed over a long period of hypothesis, test and observation.
There are several actions a test developer can use to minimize faking:
Personality tests are very commonly used, although often from a viewpoint that (incorrectly) perceives them as very strong predictors of behavior. Personality is a complex concept and whilst personality tests can give useful indicators, the world is not divided up into 16 (or less!) types of people who are unable to see or act outside of their personality profiles.
There is often a belief that personality is fixed and does not change. In practice there are three types of instability:
Much research shows value of personality tools and their links to job needs, for example the best pilots have emotional stability and extraversion. Many people will self-select jobs based on their perceived via this.
Bandwidth can be an issue, where the breadth of cover by each instrument is insufficient. However, lots of factors becomes unwieldy, too few are criticised as simplistic (eg. 16PF vs. Big Five arguments).
The jangle fallacy occurs where same trait name used by two or more questionnaires. This can be confusing.
The most predictive personality factors of job performance are conscientiousness and general intelligence (but what is ‘job performance’?). Sub-factors of ‘conscientiousness’ in studies also varied (competence, order, dutifulness, etc.). Combined traits are finding favour, such as consciousness and agreeableness. Extraversion is important in some situations, such as sales - but high agreeableness may result in lower sales and in some settings, managers do less well if they are conscientious.
Overall, though, personality tests have low predictive validity of job performance, but they are used often for this, for example people may be de-selected solely on test results. People have even been made redundant from jobs based on personality tests (and giving them a biased report to show this).
Work is often done in teams and personality tests often do not cover this (or do so only in a limited way).
Distortion and faking
Distortion and faking can be a problem where people may deliberately or subconsciously bias their self-reports (where social desirability bias can have an undesired effect). There may be a central tendency where people take the safe choice. Acquiescent people tend to use 'yes' and 'agree' to answers more than they should.
Despite concerns, faking does not affect validity that much.
Faking good (Impression Management) can be useful in the target job and is itself an indicator of personality.