While I like Schmidt et al.'s discussion of utility, it seems their findings might be exaggerated because of the degree to which computer programmer's are specialized. Would they find this same sort of utility findings if we were talking about a less specialized skill set. That is, I would expect a computer programmer who was not actually proficient at programming would be more detrimental than a customer service person who is not actually proficient at customer service. The type of job we are selecting for seems to have a huge impact on the utility of the selection practices.
Maybe this is just me, but the conclusion that using valid predictors can lead to gains in productivity seems like common sense. Wouldn’t we expect that as the validity of a predictor increases, the productivity of an employee might increase? Schmidt et al. demonstrated this in their sample, but would we expect this same relationship in other types of jobs?
•Is the issue of not applying or using utility selection models in selection similar to the issue in not running training evaluations? Are companies who lack the understanding of utility models’ purpose and benefits, thus not seeing how it can be used, and maybe lacking the funds the ones not using any utility models? •In addition, aren’t conservative or meeting the minimum required productivity/utility (i.e., profit) best and safe for companies and those who present the predictions in forecasting or getting an estimate of selection utility models? •Companies with small # of applicants would not seem to benefit from this type of forecasting of valid selection procedures because of the "sample" size affecting SR in the formula, thus not receiving accurate predictions of utility from even valid methods. How do we fix this problem?
I thought it was interesting that they authors said that "personnel psychologists have typically failed to appreciate the magnitude of productivity gains that result from use of valid selection procedures" (p. 609). If personnel psychologists fail to see this, then why are we surprised when individuals who aren't educated in this area do not recognize the importance of valid tests? What are some reasons companies might have such difficulties seeing the importance of valid selection tests?
I wonder what the impact of an article like this (from respected scientists) was to the profession at the time (1979). Was it read widely? Has anyone researched the affect on hiring of I/O psychologists and other personnel psychologists after this article was published given the evidence that serious money could be saved by using valid tests in selection?
Schmidt et al. (1979) did a good job of confronting some of the concerns with utility calculations that may have been preventing their use by personnel psychologists. It seems from the subsequent articles that for the most part psychologists have largely been persuaded by the evidence for utility, but I wonder if criticisms of validity generalization (of which Schmidt is an advocate and which is foundational to their argument for the usefulness of utility estimates) have had any impact on the use of utility by selection specialists?
I’m not quite understanding this. Could someone please explain to me why the new calculation of SDy is better than the old and how this is related a better calculation of utility?
While I like Schmidt et al.'s discussion of utility, it seems their findings might be exaggerated because of the degree to which computer programmer's are specialized. Would they find this same sort of utility findings if we were talking about a less specialized skill set. That is, I would expect a computer programmer who was not actually proficient at programming would be more detrimental than a customer service person who is not actually proficient at customer service. The type of job we are selecting for seems to have a huge impact on the utility of the selection practices.
ReplyDeleteMaybe this is just me, but the conclusion that using valid predictors can lead to gains in productivity seems like common sense. Wouldn’t we expect that as the validity of a predictor increases, the productivity of an employee might increase? Schmidt et al. demonstrated this in their sample, but would we expect this same relationship in other types of jobs?
ReplyDelete•Is the issue of not applying or using utility selection models in selection similar to the issue in not running training evaluations? Are companies who lack the understanding of utility models’ purpose and benefits, thus not seeing how it can be used, and maybe lacking the funds the ones not using any utility models?
ReplyDelete•In addition, aren’t conservative or meeting the minimum required productivity/utility (i.e., profit) best and safe for companies and those who present the predictions in forecasting or getting an estimate of selection utility models?
•Companies with small # of applicants would not seem to benefit from this type of forecasting of valid selection procedures because of the "sample" size affecting SR in the formula, thus not receiving accurate predictions of utility from even valid methods. How do we fix this problem?
I thought it was interesting that they authors said that "personnel psychologists have typically failed to appreciate the magnitude of productivity gains that result from use of valid selection procedures" (p. 609). If personnel psychologists fail to see this, then why are we surprised when individuals who aren't educated in this area do not recognize the importance of valid tests? What are some reasons companies might have such difficulties seeing the importance of valid selection tests?
ReplyDeleteI wonder what the impact of an article like this (from respected scientists) was to the profession at the time (1979). Was it read widely? Has anyone researched the affect on hiring of I/O psychologists and other personnel psychologists after this article was published given the evidence that serious money could be saved by using valid tests in selection?
ReplyDeleteSchmidt et al. (1979) did a good job of confronting some of the concerns with utility calculations that may have been preventing their use by personnel psychologists. It seems from the subsequent articles that for the most part psychologists have largely been persuaded by the evidence for utility, but I wonder if criticisms of validity generalization (of which Schmidt is an advocate and which is foundational to their argument for the usefulness of utility estimates) have had any impact on the use of utility by selection specialists?
ReplyDeleteI’m not quite understanding this. Could someone please explain to me why the new calculation of SDy is better than the old and how this is related a better calculation of utility?
ReplyDelete