Assessments are simply a standardized means of collecting information about an applicant to help with hiring or promotion decisions. It is a broad category and can take the form of a standardized written assessment, a resume screen or an interview. There is not one assessment that is a gold standard to be used universally across all roles—it depends on what characteristics matter for the role in question (see our previous post for more on that process).

If there isn’t direct science linking the assessment to job performance or to the characteristic that matters for the job in question, don’t use it.

In this third and final part on hiring, we dispel four of the most pervasive, detrimental myths regarding the use of assessments in hiring.

[Here are the first two.]

1. The more assessments, the better. Despite what many think, using more assessments is not always better—and can often be worse than using too few. Using additional assessments that do not meaningfully distinguish who is best suited for the job will result in worse hiring decisions than if these additional data points were not used at all. Worse yet, too many lengthy assessments can negatively influence the candidate experience and decrease the size of the candidate pool. In one case, a staffing firm we worked with found that long screening assessments were causing otherwise promising candidates to drop out of the process, thereby driving down the quality of hires.

Using less scientific tools, like an unstructured interview, introduce noise to the decision-making process and may bias against certain types of applicants. Simple advice—if there isn’t direct science linking the assessment to job performance or to the characteristic you’ve determined matters for the job in question, don’t use it.

2. Assessments that use the latest machine learning/AI are better than long established assessments. While we share the enthusiasm for the potential of advanced technologies and techniques (e.g., gamified assessments), we shouldn’t throw the baby out with the bath water.

It is important that new approaches are held to the same standards of integrity to which we hold more traditional assessments. Decades of research has shown that traditional assessments of cognitive ability and personality traits (e.g., conscientiousness) are still some of the best predictors out there.

At a minimum, assessment vendors should have technical documentation describing the reliability, validity and prescribed uses of their assessments. Also pay attention to any information on test bias, administration requirements and potentially available databases for scoring and test interpretation.