Assignment instructions – Complete "Conducting Empirical Validation and Adverse Impact Analysis" (Heinemann, p. 361).

Assignment Questions: Using the data above, calculate:

• Average PS scores for the whole sample, males, females, non-minorities, and minorities.
• The correlation between PS scores and performance ratings, and its statistical significance (r = .37 or higher is needed for significance at p < .05).
• Adverse impact (selection rate) statistics for males and females, and nonminorities and minorities. Use a PS score of 7 or higher as a hypothetical passing score (the score that might be used to determine who will or will not be promoted).
• Average performance rating scores for the whole sample, males, females, nonminorities, and minorities. For each group, evaluate whether the performance rating scores are different for subgroups of employees. Also, evaluate if the magnitude of these differences is sizable enough to warrant concern for Yellow Blaze.

PS Scores	Mean	Number
Total Sample	6.70	30
Males	6.80	15
Females	6.60	15
Minorities	6.89	9
Non-Minorities	6.62	21

2. With a cut score of 7 on the PS, would its use lead to adverse impact against women? Minorities? If there is an adverse impact, does the validity evidence justify the use of the PS anyway?
  
The overall selection rate equaled 17/30 with the breakdown of the four categories as follows: 

For minorities and non-minorities, the minorities selection rate was higher. Minorities had a 6/9 average, with a PS score of 7 or greater, which equals a selection rate of 66.7%. Non-minorities had an 11/21average and a PS score of 7 or higher like the minorities, which equaled a selection rate of 52.4%. 

As for the male and female averages, the male rates were higher. Males had a 9/15 average, with a PS score of 7 or higher equaling a selection rate of 60%. Whereas females had an 8/15 average and a PS score of 7 or greater, which equaled a selection rate of 53.3%. 

If you apply the 80% rule (66.7% x .80 = 53.4%), the non-minority group’s rate (low group) should equal at least 80% or the minority group’s rate (high group). Since the non-minority rate (52.4%) is less than the 80% of the minority rate, the sample size indicates there would be no problem with an adverse impact. 

If you apply the 80% rule (60% x .80 = 48%), the female group’s rate (low group) should equal at least 80% of the male group’s rate (high group). Because the female rate (53.3%) is higher than 80% of the male rate, the small sample size has no validity that there is an adverse impact even with the evidence indicates a higher rate. When using the PS score is a selection test, there may be rationalization because the focus has been placed on the PS score.    
  

• 3. What limitations do you see in the current study design? Do you think that the conclusions you would reach based on this sample of individuals who were promoted to store manager would generalize to the population of all salespeople who are being evaluated for promotion potential? Do you think that the method of rating performance is sufficient as a criterion, and if so, why? If not, what additional steps would you take to ensure that performance is measured adequately?
  
  One key limitation is the fact the sample size is small, “n” equals 30, which creates a reduction in the validity of the results. 
  
  Would you recommend that Yellow Blaze use the PS score in making future promotion decisions? Why or why not? If you do think the company should use the PS score system, can you think of anything they could do to make these measures even better than they are already? If you do not think the PS score system should be used, can you think of any ways that this system might be improved? 
  
  Based on the overall results reported above, it appears that the PS score can be used as a reasonably valid predictor of managerial performance. However, further research should still be conducted.
  
  
• One employee has raised questions regarding whether the performance ratings themselves are biased. This employee has not made a formal legal complaint against Yellow Blaze yet, but the organization wants to evaluate whether there is reason for concern. Based on the calculations you made regarding the differences for performance evaluation ratings for women relative to men, and for minorities relative to non-minorities, do you believe that there is reason for the organization to be concerned regarding this issue? In other words, do the data suggest that there is, in fact, a substantial difference in performance evaluation ratings for different groups of employees? How should the organization respond to this individual employee’s concerns?