Overall, it has become apparent the predictability and validity of having more than one selection measure. However, I am assuming that not all companies take this seriously and thus do not select their applicants well. With an increase in applications due to the economy, are we likely to see more discriminant complaints of selection procedures due to the ability to detect, for example, adverse impact by having a larger sample size?
Most of the articles suggested issues, in some shape or form, that arise when testing for adverse or disparate impact in selection procedures. While these articles provided guidance for issues such as small sample sizes, low power, predictor types, predictor reliability, etc., what do organizations actually do when faced with these issues? That is, do organizations really make an effort to fully investigate the possibilities of adverse impact?
What do you think of the point made in Ployhart et al., (p.339) that staffing researchers could broaden their scope of consideration of performance outcomes beyond task and contextual performance and also attempt to consider the attainment of broader societal goals?
If a company determines that there selection exhibits adverse impact, what do you think the best could of action is? Should individuals be investigated? Should the selection procedure be completely overhauled right away? Or should little changes be made to "hold them over" until something more large scale can be implemented? Should the company hire more minorities for awhile to try to restore balance? Really what I'm asking, what is the best way to right this wrong?
It still is not clear to me what should happen in a case when people score pretty much the same on a valid selection test or it is unclear who is the better candidate (because the sample size is too small for sign. testing). What is one to do in this situation to avoid adverse impact and demonstrate validity?
Are there potential issues with defining minority as anything other than white, male? As the work culture grow more diverse is this an unfair conceptualization? Shouldn't we be taking a more multicultural aspect of minority and majority? Would this depend upon what type of job we are talking about and who usually applies? As men enter into more typically female-dominated roles (e.g., nursing field) and vice versa (e.g., truck driving) should the conceptualization of majority and minority change?
So, it seems to me that the core issue is that our best predictors of performance (a desirable thing) tend to create disparate impact (an undesirable thing). Researchers appear to be working very hard to come up with better measures of prediction and to identify any problems with the analysis of adverse impact that might be leading to invalid results. My take-away, from a practitioner perspective, is one of concern, but ambivalence about what can practically be done that is not being done currently until further research produces a better solution. Anybody agree? Disagree?
Overall, it has become apparent the predictability and validity of having more than one selection measure. However, I am assuming that not all companies take this seriously and thus do not select their applicants well. With an increase in applications due to the economy, are we likely to see more discriminant complaints of selection procedures due to the ability to detect, for example, adverse impact by having a larger sample size?
ReplyDeleteMost of the articles suggested issues, in some shape or form, that arise when testing for adverse or disparate impact in selection procedures. While these articles provided guidance for issues such as small sample sizes, low power, predictor types, predictor reliability, etc., what do organizations actually do when faced with these issues? That is, do organizations really make an effort to fully investigate the possibilities of adverse impact?
ReplyDeleteWhat do you think of the point made in Ployhart et al., (p.339) that staffing researchers could broaden their scope of consideration of performance outcomes beyond task and contextual performance and also attempt to consider the attainment of broader societal goals?
ReplyDeleteIf a company determines that there selection exhibits adverse impact, what do you think the best could of action is? Should individuals be investigated? Should the selection procedure be completely overhauled right away? Or should little changes be made to "hold them over" until something more large scale can be implemented? Should the company hire more minorities for awhile to try to restore balance? Really what I'm asking, what is the best way to right this wrong?
ReplyDeleteIt still is not clear to me what should happen in a case when people score pretty much the same on a valid selection test or it is unclear who is the better candidate (because the sample size is too small for sign. testing). What is one to do in this situation to avoid adverse impact and demonstrate validity?
ReplyDeleteAre there potential issues with defining minority as anything other than white, male? As the work culture grow more diverse is this an unfair conceptualization? Shouldn't we be taking a more multicultural aspect of minority and majority? Would this depend upon what type of job we are talking about and who usually applies? As men enter into more typically female-dominated roles (e.g., nursing field) and vice versa (e.g., truck driving) should the conceptualization of majority and minority change?
ReplyDeleteSo, it seems to me that the core issue is that our best predictors of performance (a desirable thing) tend to create disparate impact (an undesirable thing). Researchers appear to be working very hard to come up with better measures of prediction and to identify any problems with the analysis of adverse impact that might be leading to invalid results. My take-away, from a practitioner perspective, is one of concern, but ambivalence about what can practically be done that is not being done currently until further research produces a better solution. Anybody agree? Disagree?
ReplyDelete