I was shocked by an implicit claim in an HBR article.

enter image description here

The most shocking part to me there is that job experience came dead last among those predictors of job performance. Alas the graph is from a talk, so it won't be that easy to see exactly what it means by job experience; the HBR article does not detail that because its topic is mainly personality tests (which I don't want to get into this question).

Obviously "job experience" could be measured in a silly way, like just having a prior job, say a as waiter, doesn't predict well job performance as a neurosurgeon. However this may turn out to be more or less how HR departments actually measure job experience, i.e. in an irrelevant way.

So is there more research elaborating or refuting job experience measures as predictors for job performance? And if so, is there a difference between how job experience ought to be measured and how it is, i.e. are there any studies on this potential theory-practice discrepancy?

  • 2
    The phrase "correlation between tests scores and predicted job performance" suggests a poor understanding of data science. You should compare a metric with actual performance, not predicted. If you have a linear model, then there should be 100% correlation between a metric and what's predicted based on it. – Acccumulation Dec 3 '17 at 18:10
  • 50% of employees will be below average. Why would you expect that to be banded by years of employment, such that people with 2 years of experience and are below average, would be better than the people that have 1 year of experience and are above average? I find the 13% to be supiciously high. – jmoreno Dec 9 '17 at 19:14

I managed to find however the 1998 paper of Schmidt and Hunter. In that paper, job experience is simply the number of years in the workforce, so there's little surprise it doesn't correlate well with performance.

Judging by a quick search, it seems more recent research has been using "prior related work experience" (PRWE), which makes more "duh"/intuitive sense. Alas with even this latter measure there's been wide variance in the degree of association with job performance as reported by various studies, according to the brief summary I found in a Uppal et al. (2014).

  • Only doing this because the mods apparently undeleted my question... unless I've done it and forgot about it. – Fizz Dec 3 '17 at 0:19
  • 2
    If you go to this link, you can see which mod undeleted the question. I found that by finding a revisions page (I used the one for this answer, as it was right here) and replacing the post ID with that of the post that you want to see. The post ID (40059) appears in the URL. – Brythan Dec 3 '17 at 2:35

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .