The issue of faking in personality assessments is widely studied and debated, both in academia and industry. Some, like Hogan, Barrett and Hogan (2007) argue that faking is not an issue in operational settings. Other researchers (and test publishers) acknowledge that faking does occur, but have different conclusions about what to do about it. I've received many questions during my career from concerned HR professionals, about specific situations when they suspected individuals of cheating and more generally how valid personality assessments are for hiring decisions.
The aim of this article is to discuss 1) how to define faking, 2) methods to detect faking, 3) how faking affects the validity of personality assessments and 4) how to deal with faking.
What is faking?
Let's start by differentiating between faking and other sources of uncertainty in personality scores. We know from research and experience that results from personality tests are affected to some degree by situational states, such as mood and motivation, and that results can change slightly over time (read more about these effects here and here). It is also reasonable to assume that individuals vary in their self-perception, which is how well they know and understand themselves and their behavior. These factors are often considered as part of the measurement error, which is captured by the reliability coefficient of the test.
In contrast, faking is commonly defined as 'intentional misrepresentation in self-report' (Holden & Books, 2011), which means that scores where faking is present would be different from scores where no faking is present.
This distinction both points to ways of studying the effects of faking, and potential challenges in detecting faking. First, we can conduct experiments where individuals take the same test under separate conditions and measure the difference in scores. This has been done in operational settings, for example by Hogan, Barrett and Hogan (2007), Landers et al. (2011) and Hausknecht (2010). In these studies, participants were candidates applying for a job, who where offered to retake a personality assessment after failing the test the first time. The observed difference between the first and second attempt was quite different in these studies, ranging from -0.02 (Hogan, Barrett & Hogan, 2007) to 0.86 (Landers, 2011) on average.
Second, we realize that in practice, it will be very difficult to distinguish between the effects of faking and other sources of uncertainty. Candidates will differ in their motivations for faking, depending on their goals with the assessment. According to Sackett, the most common goals are to impress, be credible and be true to one's self. Some candidates will care more about impressing their potential employer with high results on the test, while others will worry about being dishonest or being 'caught' by their supervisor or colleagues once employed.
Can faking be detected?
The oldest and most common method for detecting faking is to use specific scales designed to measure social desirability or lying. This method has been studied thoroughly and there is wide agreement among scholars that it is not useful (Sackett, 2011). Using social desirability scales to detect faking should therefore be avoided.
Some more recent methods have been proposed, such as the overclaiming technique, IRT approaches and the idiosyncratic item response approach. These methods have not been extensively studied, however, and caution is advised in implementing them in operational practice.
The most reasonable conclusion is that there are no methods for detecting faking suitable for operational use (Sackett, 2011). Test scores should not be adjusted on the basis of faking.
How does faking affect the validity of personality assessments?
The main interest in candidate selection is often the predictive validity of the assessment. An assessment with high validity will differentiate well between candidates who will perform well on the job and candidates who will perform poorly. In this context, faking may actually not be relevant as long as the assessment has high predictive validity.
There is evidence showing that the predictive validity of personality assessments is quite robust to faking (Sackett, 2011). In addition, hiring decisions are rarely based on personality tests alone. The highest predictive validity is attained when combining results from multiple measurements, such as tests of general mental ability (GMA), personality tests, interview ratings, work samples etc. In this setting, the effects of faking on the personality test on the overall assessment is quite small.
How to deal with faking?
Based on the discussion above, we can conclude that:
- It is likely that some candidates fake their responses to personality assessments to some extent. However, the reasons for faking are different depending on the individual and the situation, and how much faking affects the scores is unclear.
- There are no good methods for detecting faking. Social desirability scales should be avoided and test scores should not be adjusted on the basis of faking.
- Personality tests are valid for predicting job performance even when faking is present. The best way to include personality assessments in candidate selection is to combine it with other measurements, like a GMA test for example.
Landers, R. N., Sackett, P. R., & Tuzinski, K. A. (2011). Retesting after initial failure, coaching rumors, and warnings against faking in the use of personality measures for selection. Journal of Applied Psychology, 96, 202-210.
Hausknecht, J. P. (2010). Candidate persistence and personality test practice effects: Implications for staffing system management. Personnel Psychology, 63, 299-324.
Hogan, J., Barrett, P., & Hogan, R. (2007). Personality measurement, faking, and employment selection. Journal of Applied Psychology, 92, 1270-1285.
Holden, R. R., & Book, A. S. (2011). Faking does distort self-report personality assessment. In New Perspectives on Faking in Personality Assessment Oxford University Press.
Sackett, P. R. (2011). Faking in Personality Assessments: Where Do We Stand? In New Perspectives on Faking in Personality Assessment Oxford University Press.