Pre-Employment Surveys: A Brief Background

One Accord employment survey in the trucking industryCompanies have used “tests” in some form or another for as long as companies have been hiring. The test used may or may not have had anything to do with the applicants’ ability to do the job. After World War II testing got taken over by the industrial psychologists and went one of four basic ways:

  1. IQ
  2. Aptitude
  3. Psychological
  4. Skills or Skill / Knowledge

Each of these is better than just an interview. Interviews are notoriously poor indicators of job performance. They are far too subjective and frequently focus on criteria that have little or nothing to do with the person’s ability to do the job. In fact, the probability of the interviewer to predict job success is .50 or 50/50. It can actually be less than that since too often the interviewer is having to pick from between two people, neither one of whom is very good. Anything, therefore, that companies can do to increase predictability beyond pure chance is, of course, better.

IQ Tests – don’t even predict success in college very well and while all jobs require some level of intelligence, the relationship between IQ and job performance is not linear, i.e., the higher the IQ the greater the job performance. For example, a bank teller with an IQ of 140 is not necessarily a better bank teller than one with an IQ of 115. In fact, the relationship is more likely curvilinear.

Aptitude Tests – are too general. They may point you to very broad areas, but they don’t predict success in specific jobs.

Skills Testing – is good but it is best viewed as an indicator of minimum qualifications. Additionally, most jobs are a maximum 30% technical and even if one had a perfect measure of the 30% of the job that is technical (an unlikely event), one would still only be measuring 30% of the job and there is no way to predict overall job success based on 30% of the job.

Psychological Tests – are all over the place, but basically they are all looking for the “big six” or “big eight” personality traits (passive-aggressive, introvert-extrovert, etc.). They are measuring psychological traits or constructs. They measure the person in the abstract. Then someone has to make the logical leap from the abstract to the actual job requirements.

The use of psychological tests / assessments will yield a correlation in general terms, but the question we ask is why measure the person in the abstract? Why not measure the person for the job? Why not look at the actual job behaviors -all of them, not just the technical- which account for variances in productivity outcomes.

What do a high performing “extrovert” and a high performing “introvert” have in common? They both do the actual job behaviors which impact productivity, so why try to build a construct? Better (and more predictive) is to look specifically at the actual job behaviors that come from a job analysis with top performing employees.

Productivity is in the detail; the actual job behaviors, not in the trait or construct. The competitions’ model says that person x is an introvert and then they generalize to the behaviors. We don’t care what kind of personality a person is. We want to know if they do the specific job behaviors that superior performers tell us make differences in productivity outcomes on the job. We take their model and turn it right-side up.

Comments are closed.