Van Iddekinge and Ployhart discuss computing the relative contribution of each predictor when multiple selection tests have been used, but state that in a MR where predictors are correlated the sum of the R2s will sum to more than the total variance accounted for due to overlap in the predictors. Wouldn't computing tolerances give the unique contribution (variance accounted for) of each predictor?
Could the use of in situ performance techniques (i.e., internships, contract work), as discussed by Cascio and Aguinus (2008), facilitate the use of predictive validity in the workplace? How so?
Although they aren't perfect the authors suggest using peer ratings as a performance rating. What are some of the advantages and disadvantages of using peer ratings?
These authors talk about the positive aspects of using multiple predictors for selection purposes. I would assume there is an optimal number of predictors one should use or that is feasible to use for these purposes. How do we know where this optimal level is? Should our multiple predictors focus on job performance as understood by total output? Or does it NEED to include predictors designed to assess things like potential absenteeism or OCB's?
There's so much here. What would be a good way of explaining differential prediction analysis (predictive bias) to company decision-makers? Why should we care about this?
In their review of citizenship behavior as an important criterion that needs to be better examined/is a trend in recent research, I think they missed the mark on how this type of behavior likely influences the work climate, which could in turn influence evaluation norms. I think the literature is in dire need of a multilevel approach to better understanding this construct. What do you think?
• Yes research and studies have shown that over time performance levels off and predictions are still accurate, but how can this account for world, life, and organizational changes that may occur over time? I feel that performance is only accounted for or predicted on some ability and psychological factors yet excluding the organizational criteria along with the fact that we can’t predict the future of economy and cultural changes.
Van Iddekinge and Ployhart discuss computing the relative contribution of each predictor when multiple selection tests have been used, but state that in a MR where predictors are correlated the sum of the R2s will sum to more than the total variance accounted for due to overlap in the predictors. Wouldn't computing tolerances give the unique contribution (variance accounted for) of each predictor?
ReplyDeleteCould the use of in situ performance techniques (i.e., internships, contract work), as discussed by Cascio and Aguinus (2008), facilitate the use of predictive validity in the workplace? How so?
ReplyDeleteAlthough they aren't perfect the authors suggest using peer ratings as a performance rating. What are some of the advantages and disadvantages of using peer ratings?
ReplyDeleteThese authors talk about the positive aspects of using multiple predictors for selection purposes. I would assume there is an optimal number of predictors one should use or that is feasible to use for these purposes. How do we know where this optimal level is? Should our multiple predictors focus on job performance as understood by total output? Or does it NEED to include predictors designed to assess things like potential absenteeism or OCB's?
ReplyDeleteThere's so much here.
ReplyDeleteWhat would be a good way of explaining differential prediction analysis (predictive bias) to company decision-makers? Why should we care about this?
In their review of citizenship behavior as an important criterion that needs to be better examined/is a trend in recent research, I think they missed the mark on how this type of behavior likely influences the work climate, which could in turn influence evaluation norms. I think the literature is in dire need of a multilevel approach to better understanding this construct. What do you think?
ReplyDelete• Yes research and studies have shown that over time performance levels off and predictions are still accurate, but how can this account for world, life, and organizational changes that may occur over time? I feel that performance is only accounted for or predicted on some ability and psychological factors yet excluding the organizational criteria along with the fact that we can’t predict the future of economy and cultural changes.
ReplyDelete