Is your Selection Procedure Committed to Diversity?

Is your Selection Procedure Committed to Diversity?

The Spring 2008 edition of Personnel Psychology contains three important articles on what is termed the diversity-validity dilemma. This refers to the relatively common finding that the selection process with the greatest predictive validity can be one that disadvantages already under-represented groups. A good example is tests of cognitive ability. These can be good predictors of success but racioethnic minority groups typically score lower than majority groups. The magnitude of the difference in scores is conventionally reported by the d statistic - the difference in mean scores for two groups divided by their pooled standard deviations. From US studies, the difference between white and black groups on cognitive ability is one standard deviation. On the other hand, differences on the Big Five personality dimensions are, for the most part, negligible.

The journal examines two responses to the diversity-validity dilemma:

Use strategies to reduce subgroup differences
Have affirmative action preferences

Strategies to reduce subgroup differences

Robert Ployhart and Brian Holz consider a total of 16 strategies to reduce subgroup differences. They put these into five categories as follows:

Use predictors with smaller subgroup differences
Combine scores across predictors, diluting the impact of the predictors with high d statistics
Remove irrelevant variance (e.g. language and cultural differences) from predictor scores
Allow practice prior to testing
Enhance applicant perceptions of the process by ensuring under-represented groups are in the applicant pool

Ployhart and Holz consider the first two categories the most powerful but the strategies mainly carry the problem that validity is lowered. On the other hand, the other three categories are less powerful but do not lower validity. However, within the first two categories the strategy of ensuring the full range of predictors is included (and not just cognitive ability) both reduces differences and increases validity. The authors also single out the use of structured interviews in place of unstructured as both decreasing subgroup differences and increasing validity. They also note the power of reducing reading requirements by replacing multiple choice questions with 'open-ended' responses. These reduce White-Black differences substantially but obviously carry the burden of having to review and mark open-ended responses.

Affirmative Action

David Kravitz separates three categories of affirmative action:

Opportunity enhancement such as targeted recruitment or training
Preferring the under-represented group when their credentials are equivalent to those of the non-target applicants
Preferring the underrepresented group even when their credentials are weaker than the non-target group

In the UK and EU legislation, positive discrimination is illegal (apart from provisions necessary to make disability and age legislation work). On the other hand positive action via balancing measures is within the law and last years consultation by the UK Government was aimed at making this clear as well as removing impediments to and encouraging such measures.
www.communities.gov.uk/documents/corporate/pdf/325332
In short the UK Government aim is to maximise affirmative action that falls within the first category.

Although preferences are illegal in Europe, Keith Tyburn from US law firm Fisher and Phillips and his co-writers give a clear summary of US law that allows preferences in the private sector if:

There has been substantial historical exclusion of a given group
The non-preferred groups rights are not unnecessarily trammelled
The programme has limited duration

In the US public sector preferences must:

Further a compelling government interest - e.g. to have a diverse police force
Be narrowly tailored and not unnecessarily trammel on the interests of others.

That said, David Kravitz makes clear that preferences are usually illegal in the US. He goes on to observe that, quite apart from its legal status, an affirmative action programme (AAP) can create resentment amongst the non-target group. First, he notes that many white Americans do not believe discrimination exists and therefore do not support actions to eliminate it. He describes "the most consistent finding" being that "preferences cause problems" that can include stigmatization and self-stigmatization of target groups.

Kravitz's recommendations are to put in place all the elements of a non-preferential AAP that will eliminate discrimination, increase applications from underrepresented groups and ensure that recruits feel included and are retained, summarising his recommendations in terms of an Attraction-Selection-Inclusion-Retention (ASIR) model. Actions to promote diversity can be taken at all stages in the model. At the inclusion and retention stage he advocates a general increase in management skills because underrepresented groups may interpret poor management as discrimination. He also recommends mentoring and affinity groups, as well as benefits that are of particular use to underrepresented groups, such as family friendly policies.
On the other hand, Kravitz notes that actions that do not appear open to all employees can backfire and he notes that there is research showing that diversity training - one of the most expensive diversity practices - has, under some conditions, been associated with a decrease in minority representation.

Overall, there is much in these articles for the HR practitioner who is concerned with increasing the diversity of their workforce. The articles present well the dilemmas that need to be confronted and upon which policy needs to be agreed.

How can Human Assets Help?

First of all we can help you determine your diversity strategy. We can help you consider the dilemmas, and find the best way forward. Next we can help implement the chosen solutions, such as an enhanced selection process, or the promotion of inclusion and retention. Finally, we can monitor the success of your strategies - both by gathering and analysing hard data but also by obtaining the views of target and non-target groups to gauge the impact of your initiatives.

To find out more about the Human Assets approach to choosing and promoting diversity, please contact our consultants by email at enquiries@humanassets.co.uk or telephone +44(0)20 7434 2122.

References

Pyburn, K. M., Ployhart, R. E., and Kravitz, D. A. The diversity-validity dilemma: Overview and legal context. Personnel Psychology, Vol 61 No 1, pps 143-151.

Ployhart, R. E., and Holtz, B. C. The diversity-validity dilemma: Strategies for reducing racioethnic and sex subgroup differences and adverse impact in selection. Personnel Psychology, Vol 61 No 1, pps 153-172.

Kravitz, D. A. The diversity-validity dilemma: Beyond selection-The role of affirmative action. Personnel Psychology, Vol 61 No 1, pps 173-193.

Newsletter: April 2008