Meta-analyses and research reviews have provided valuable insight into the validity and utility of a wide range of selection methods. There is a particularly large amount of research surrounding the following five forms of personnel assessment:
Work Sample
These tests require the candidate to perform tasks that are similar to those that are performed on the job. Their close relationship with the job means that they have high content validity, are less likely to discriminate, and have higher face validity than personality measures. In addition, they been reported to have a predictive validity of .54 for overall work performance (1). This makes work samples the best predictors of future performance out of the 19 selection methods studied in the meta-analysis.
Unfortunately, work samples encounter practical problems in that they are costly, difficult to administer to groups, and restricted to samples of work which can be completed in a short amount of time. They are also limited in the way that they can only be used with applicants who already know the job, and fail to measure aptitudes for more difficult tasks that could be encountered in the future.
Structured Interview
In this type of interview the applicant is asked a fixed format of set questions with individual ratings for each item. Questions are usually developed through careful job analysis. Schmidt & Hunter (1) report that the average validity of the structured interview is .51 for overall job performance, and .38 for trainability. In his review of selection methods, Salgado (3) describes how highly structured interviews have low negative impact upon minorities and are resistant to interviewer bias.
Disadvantages include the fact that structured interviews are costly and that they often contain job knowledge components. This can make them unsuitable for inexperienced applicants (1). Salgado (3) suggests that structured element can make the method unpopular as it takes away interviewer autonomy and can be perceived as unfair by the applicant.
Job Knowledge Test
Some jobs, such as welder or chemist, require specific skills and knowledge. Employers often conduct an analysis of the tasks that make up the job and construct a job knowledge test. Alternatively, they can buy off the shelf tests that measure knowledge. Schmidt & Hunter (1) report that job knowledge tests obtain an average correlation of .48 with overall work performance, demonstrating high predictive validity. However, Salgado (3) describes how their validity is dependent upon complexity of the job and job-test similarity. This can mean that job knowledge tests are somewhat restricted to jobs that require greater levels of knowledge and judgement. They can also only be used to evaluate experienced applicants and their construction can be costly and time-consuming.
Personality
Meta-analyses have reported validity coefficients of up to .31 for predictions of job performance from personality tests (1, 2). Extensive research has led to the conclusion that Conscientiousness and Emotional Stability are valid predictors for all jobs and criteria (3). However, evidence also shows that care should be taken when using personality measures, as other Five Factor Model traits are limited to predicting performance in specific occupations and roles (2). This suggests that job performance is dependent upon a range of factors, and points towards the use of personality tests in conjunction with other methods. In addition, personality measures can lack face validity due to their perceived lack of relevance to jobs, and are vulnerable to faking.
General Mental Ability (GMA)
Schmidt & Hunter (1) claim that GMA occupies a “special place” in the pantheon of personnel measures. Tests have a predictive validity of .51 for overall job performance and .56 for trainability, giving GMA the highest validity and lowest application cost of procedures that can be used on candidates with a range of experience. However, the same study also found that GMA tests were much less predictive of performance in unskilled jobs than managerial and professional jobs, with validities of .23 and .58 respectively. Furthermore, the inclusion of methods such as job knowledge test, job sample test or structured interview alongside a GMA test further improved the prediction of job performance. This suggests that GMA tests only assess one component of the many abilities and skills that determine work performance.
1. Schmidt, F. L. & Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology: practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124, 262-274
2. Barrick, M. R., Mount, M. K. & Judge, T. A. (2001). Personality and performance at the beginning of the new Millennium: What do we know and where do we go next? International Journal of Selection and Assessment, 9, 9-30
3. Salgado, J.F. (1999). Personnel selection methods. International Review of Industrial and Organizational Psychology, 14.
|
