Make your own free website on

Ivan Ehlers

Elizabeth Moses


Arthur, W. Jr., Woehr, D.J., & Graziano, W. (2001). Personality testing in employment

            Settings: Problems and issues in the application of typical selection practices.

            Personnel Review, 2001, Vol. 30, Issue 5/6, 657. 20. Retrieved October 8, 2003

            From Proquest database.



1)      Purpose

A)     This review article critically evaluates the traditional approaches to personnel selection and focuses on five problematic issues therein

B)     The point of the review is not to argue against the use or importance of personality tests or to provide an all-encompassing review of the literature 

2)      The five issues

A)     Appropriateness of linear selection models

1) Curvier linear models might be more appropriate

B)     The problem of personality-related self-selection effects

1)  Restriction of range

C)     The multidimensionality of personality

1) Finding an ideal employee is harder in real life

D)     The detection of biases related to social desirability, impression management and faking

1)  How to detect these and account for them when hiring

E)      The legal implication of personality assessment in personality

1)      Discussion of legal issues with personality tests and employment

3)      Conclusion

4)      Critique

5)      Interesting/important points


Personality Testing in Employment Settings


This review article critically evaluates the traditional approaches to personnel selection and focuses on five problematic issues therein. Being that personality testing in employment settings is such a complex issue, the authors stress the importance of viewing the evaluation as a whole, rather than in bits and pieces. The authors also seek to remind readers that the point of the review is not to argue against the use or importance of personality tests or to provide an all-encompassing review of the literature on the topic. Instead, the review is intended to identify and discuss issues that may be problematic in traditional approaches of personality testing in employment contexts.

            The first issue described is the appropriateness of linear selection models. These models assume that the higher the score is on the predictor scale, the more desirable it is. The review acknowledges how this model may work for variables such as knowledge, skill and ability, but it creates a distinction when considering personality variables on the linear model. They state that “various personality constructs and job performance may be better conceptualized . . . as being nonlinear” (Arthur, Woehr, and Graziano, 2001, p 658). Essentially, the authors’ point is that “more is not necessarily better” in regards to variables such as extraversion and conscientiousness (Arthur, Woehr, and Graziano, 2001, p 658). The curvilinear model proposes that the higher the variable, the more desirable it is to a certain point; after that point, the variable’s value diminishes. This is contrary to the linear model that assumes more to be better continuously.

            The second issue addresses the problem of personality-related self-selection effects. This argument bases itself on the idea that like-minded individuals tend to congregate in the same professions, and, in turn, create a “restriction of range” when their personalities are measured. Given that one profession harbors a tendency for people to be introverts, the “variability on the extraversion dimension will be restricted. This restriction of range will in turn limit the ability of the extraversion dimension to predict job performance” (Arthur, Woehr, and Graziano, 2001, p 659).

            The multi-dimensionality of personality is addressed next, along with the concept of generating a “composite” personality score. The authors consider this aspect to be problematic when “broad brush dimensions” are used to describe a whole person, all at once, in one, sweeping term. Another issue looks at tests that decide the compatibility between individuals and the organization itself. Once again, it is problematic when a few variables are used to emphatically determine the “totality of an applicant’s personality” (Arthur, Woehr, and Graziano, 2001, p 660). While acknowledging the difficulty in applying methods that appreciate and illustrate the “complexity and interconnection among personality dimensions,” the authors conclude that “it is highly unlikely that only one dimension will be important for successful performance and even more unlikely that simple main effects will provide a complete picture” (Arthur, Woehr, and Graziano, 2001, p 661).

            The next issue is the detection of faking and the use of top-down selection models. Researchers often have many different methods instituted to determine if the questions have been answered honestly. However, two different kinds of faking have been presented: “impression management,” where the person is consciously lying in order to please a prospective employer; and “self-deception,” where the test-taker is “less aware of the motivational processes leading to distortions” (Arthur, Woehr, and Graziano, 2001, p 660). As research generally goes, scores that are extreme are typically accounted for, and are otherwise tossed out as inaccurate. However, in employment settings, the top 5-10% are the ones that will be hired first, in a top-down model. According to the authors, those scores in the top 5-10% are most likely to be the ones that have been influenced by faking, whether deliberate or unintentional.

            The final issue is the legal implications of personality assessment in employment contexts. Test constructs are controlled by anti-discrimination laws and may be forced to provide neutral results that are not telling of anything relevant to the employer. While the difference between races has been erased, there still lies a difference between the genders. To counter this, within group scoring should be used to “interpret an individual’s score . . . within the context of their group” (Arthur, Woehr, and Graziano, 2001, p 661). However, Civil Rights Acts prevent such adjustments from occurring. Based on this “conflict between law and science,” writers have taken relative loopholes to take within-group cut-offs based on the ability to “increase accuracy of measurement or accuracy of prediction without increasing the adverse impact of the test,” as per the guidelines of the Civil Rights Act. Two other issues directly related are invasion of privacy laws and the Americans with disabilities act. (Arthur, Woehr, and Graziano, 2001, p 660). Both of these laws provide potential hindrances to the usage of personality testing in employment settings.

            In conclusion, the authors seek mostly to stress that personality testing within employment settings is not nearly as easy as it appears to be. While battling issues such as keeping costs down and time short, test makers/administrators must always remember that efficiency can never trump legal requirements.




1.    Non-Linear Model- Argument for problems with linear model depends on the existence of theoretical, individual extreme scores.
2.    Multi-Dimensionality- Argument states that testing for a few variables does not give a “complete picture” of the individual. The employer is not after any such complete picture. The employer wants to see the relevant variables and does not need to see the costly and time-consuming irrelevant details.
3.    Social Desirability- Besides assuming that the top 5-10% of scoring individuals are lying, this model also assumes that employers will automatically hire that same 5-10% based solely on the personality measure, without looking at other information (such as experience, education or ability).

Interesting Points:


1.Easier for person to accept they do not have the temperament for a job and not pursue it, than it is to accept that they do not have the skill for it.

2.Idea that employers hire based on the top 5-10% of the top-down model, knowing that those scores are most likely inaccurate.