The Case for a Robust Pre-Hire Assessment Process (Versus a Fast and Cheap one)
The market has been inundated with fast and low-cost psychological assessments for companies to use before hiring. While their ease of use of these may be attractive, there are several concerns to consider.
Beware of the fantasy that short and cheap can be effective.
Unfortunately, what we know about human behavior is that more information is almost always better. While many of these “quick and easy” assessments proport to be highly predictive of job behavior, the research simply would not agree with that. The small amount of information they collect will not be nearly as predictive as a longer battery. The science behind many of these is also lacking. Many of the organizations entering the market invest much more in marketing than in test development. Consequently, you use a questionnaire with good marketing but bad predictability. Importantly, the legal defensibility of many of these instruments is yet untested, and this puts the burden of defensibility on the employer using it.
Predicting behavior is much more difficult than describing personality.
Many short questionnaires can reasonably describe someone’s personality in broad terms. These are usually best for team building events, however, not hiring. Most employers don’t want a simple, horoscope-level description of a candidate’s personality. They want to know, “Will this person do a good job in the role we’re hiring them for?” and “Will they be a good addition to our team and its culture?” While a short questionnaire may be able to provide some basic information about the person, there is a large leap between knowing someone’s personality and predicting their ability to be a positive addition to the team and to positively fulfill their job responsibilities.
AI is evolving.
On the cutting edge of many of these questionnaires is artificial intelligence (AI). Again, there is more money being spent on marketing here than on developing these questionnaires. Although AI offers the lure of highly accurate assessments through limited questions, these instruments remain highly controversial. One of AI’s inherent problems is that it is only as good as the programming that goes into it. The history around this suggests that AI has a habit of dramatically skewing the people that it creates pass/fail scores for among job applicants. For these reasons, there are laws in some states that restrict the use of AI from key employment assessments. In those states, the burden of proof that the instrument is unbiased is on both the test publisher and the employer.
In summary, the promise of highly reliable assessments with single instruments or questionnaires with only a few items remains highly suspect. These test publishers usually are very good at hyping their product but tend to have poor track records over the long-term. Better and more legally defensible results are achieved through utilizing a more robust, evidence-based battery of assessments.
