I gathered from the readings that banding is used quite often in private businesses. Yet the literature clearly suggests major problems with the banding approach. This seems to be just another example of the disconnect between theory and practice in I/O. How can we help to bridge this gap, especially in personnel selection where we are making important decisions about the futures of individuals?
In a couple of articles it was mentioned that unions challenged SED banding in court. I’m not sure I understand the motivation for this.
In this set of articles I got a sense of the evolving science of selection and how something like banding that seems to have evolved from the social goal of wanting to select more women and minorities led to a statistical method of doing it (with all its accompanying critiques). Banding seems to be legally defensible. After all the articles, where do you stand on the value of SED and sliding banding?
•Don't banding procedures go against the sole purpose of testing, to discriminate applicants in order to detect different scores and thus performance? And aren’t we losing data/information when we group scores together? Despite SIOPS comment on it being psychometrically sound, as we learned in psychometrics this is actually not an accurate approach to testing and predicting. Why is it still used though?
It seems that we are only hearing one side of this argument, besides the previous Q and A paper with debating sides. Is it the case that we have been using banding procedures and there are no redeeming qualities to these techniques? What are some positives besides the case for diversity? What are some alternative solutions to banding beside the few proposed in the articles?
I’m currently convinced that the issue of increasing diversity in the workplace does have a psychometric solution (this stance could chance within the next day, however). What are some other ways you can think of to increase diversity in the workplace that is beyond selection?
Was anyone else disappointed in the attention that Ployhart et al. (2006) gave to the problems with banding? @ MariaTheresa, I actually (and surprisingly!) found the arguments against banding very persuasive. Schmidt's (1995) statements, "Contrary to Cascio et al., adverse impact is not a psychometric problem....It is a broad social problem: Members of some minority groups are less frequently developing high levels of the abilities and skills needed for job and career success" (p. 172.1) perhaps put the attention where it needs to be. There may not be any statistical or methodological answer to the problem of adverse impact. Perhaps, the problem is really related to deeper inequities in our social system such as culture, educational opportunities, healthcare access etc.
I gathered from the readings that banding is used quite often in private businesses. Yet the literature clearly suggests major problems with the banding approach. This seems to be just another example of the disconnect between theory and practice in I/O. How can we help to bridge this gap, especially in personnel selection where we are making important decisions about the futures of individuals?
ReplyDeleteIn a couple of articles it was mentioned that unions challenged SED banding in court. I’m not sure I understand the motivation for this.
ReplyDeleteIn this set of articles I got a sense of the evolving science of selection and how something like banding that seems to have evolved from the social goal of wanting to select more women and minorities led to a statistical method of doing it (with all its accompanying critiques). Banding seems to be legally defensible. After all the articles, where do you stand on the value of SED and sliding banding?
•Don't banding procedures go against the sole purpose of testing, to discriminate applicants in order to detect different scores and thus performance? And aren’t we losing data/information when we group scores together? Despite SIOPS comment on it being psychometrically sound, as we learned in psychometrics this is actually not an accurate approach to testing and predicting. Why is it still used though?
ReplyDeleteIt seems that we are only hearing one side of this argument, besides the previous Q and A paper with debating sides. Is it the case that we have been using banding procedures and there are no redeeming qualities to these techniques? What are some positives besides the case for diversity? What are some alternative solutions to banding beside the few proposed in the articles?
ReplyDeleteI’m currently convinced that the issue of increasing diversity in the workplace does have a psychometric solution (this stance could chance within the next day, however). What are some other ways you can think of to increase diversity in the workplace that is beyond selection?
ReplyDeleteWas anyone else disappointed in the attention that Ployhart et al. (2006) gave to the problems with banding? @ MariaTheresa, I actually (and surprisingly!) found the arguments against banding very persuasive. Schmidt's (1995) statements, "Contrary to Cascio et al., adverse impact is not a psychometric problem....It is a broad social problem: Members of some minority groups are less frequently developing high levels of the abilities and skills needed for job and career success" (p. 172.1) perhaps put the attention where it needs to be. There may not be any statistical or methodological answer to the problem of adverse impact. Perhaps, the problem is really related to deeper inequities in our social system such as culture, educational opportunities, healthcare access etc.
ReplyDelete