Demystifying the Complex Science of Predictive Hiring
Reading Time:
Psychology has its own theories that explain human behavior in ways that seem impossible to disprove: ingroup-outgroup bias, cognitive dissonance resolution, and attribution error. Intuitively, these theories seem to perfectly explain why we prefer candidates who are similar to us, how we rationalize poor hiring decisions, and how we misattribute the causes of recruitment outcomes. Just like confirmation bias, these concepts become evident everywhere in recruitment management once you become aware of them.
Keep testing
But science never sleeps. We can't just accept a theory because it feels like common sense. Theories are never fully proven, even the good ones that have been around forever. Theories are tested, improved, refined, evolved, and sometimes disproved and replaced. Rival theories can crop up like seagulls around a French fry shop. The same applies to hiring software; it must constantly evolve. New algorithms and features are developed, tested, and refined to ensure they remain effective and relevant in the ever-changing landscape of recruitment.
There's science in hiring, too. The focus is to explain the most effective method to evaluate the candidate and predict their future job performance. Extensive research has been done on various assessment methods to identify the best performing predictors. The challenge for the HR manager who wants to use this scientific evidence is to understand the picture that all these results paint and make sense of it.
Now comes the scientific part:
A rich source of evidence for predictive hiring decisions is meta-analytic studies. Stay tuned. This is a form of research that combines all the data from numerous previous studies to develop an overall theory of what is going on, generally using sophisticated mathematical kung fu. In the case of predictive hiring, this means finding as much data as possible from previous studies, examining which recruiting tools best predict on-the-job performance, throwing it all into a big pot, and mathematically identifying the best-performing predictors (read our article “AI in Recruiting” to learn more about the benefits of using predictive hiring in your recruiting).
In 1998, Hunter & Schmidt published the results of their meta-analytic research based on decades of previous assessment research. The paper provides a handy ranking that the beleaguered, time-pressed hiring manager who wants their hiring process to actually work can consult to understand which assessment methods best predict job performance.
The study, or at least the ranking, is widely used to communicate which assessments to avoid and which to use if you want to make sound hiring decisions.
But relying on these results goes against the science! The theory needs to be tested. So the same researchers ran the analysis again in 1998, 2004, and 2016. The results varied between studies due to the different data sets and sophisticated mathematics applied. But multiple studies give us more confidence in the robustness of the results and in the differences in impact in different contexts.
Who can I trust?
When a company uses science to market its products, it often presents only a single study whose results support its commercial interests or sharpen the axe it wants to grind. In predictive hiring, a recruiter can present a scientific result that it thinks best fits its message. To navigate this jungle, here are three simple tips to help you understand the science and improve your predictive hiring processes:
- Never trust the results of a single study!
- Consider the broader research area, case studies, and industry best practices.
- Once you've made your choice, do your own science to monitor the effectiveness of your predictive hiring methods.
Understanding the science behind predictive hiring is crucial for effective recruitment. Tools like IceHrm can help implement evidence-based practices, ensuring you select the best candidates while continually refining your methods for better results.
