Decision Making

1
Decision Making
2
Lecture Outline

• Choice of Assessment Method
• Validity, Correlation with other predictors,
  Adverse impact, Utility
• Determining Assessment Scores
• Compensatory, Multiple hurdle
• Determine Cut Scores
• Minimum competency, Top-down, Banding

3
Choice of Assessment Method

• Validity coefficient
• Correlation with other predictors
• Adverse impact
• Utility

4
Validity Correlation With Other Predictors

• Validity
• Criterion-related validity
• Extent to which predictor adds value to
  prediction of job success
• Face validity
• Extent to which a test looks like it measures
  what it is supposed to measure appears to be
  job-related
• Correlation with other predictors
• To add value, a predictor must add to prediction
  of success above and beyond other predictors

5
Discussion Question

• When making a decision about which methods to
  use, what matters more criterion-related
  validity or face validity?

6
Adverse Impact Utility Analysis

• Adverse impact
• When a predictor discriminates by screening out a
  disproportionate number of minorities and women
• Utility analysis
• Expected economic gain from using a predictor
• Focuses on the monetary impact of using a
  predictor

7
Discussion Question

• Would you rather use a predictor with high
  validity and high adverse impact or one with low
  validity and low adverse impact?

8
Determining Assessment Scores

• Single predictor
• Multiple predictor methods
• Compensatory model
• Scores from one predictor are added to scores on
  another predictor to create some overall score
• Check out pages 537-542 in the textbook
• Multiple hurdles model
• An applicant must earn a passing score on each
  predictor before advancing in the selection
  process

9
Determining Cut Scores

• Cut score
• The minimum score an applicant must reach to not
  be eliminated from the selection process
• Methods for determining cut scores
• Minimum competency
• What are the minimum qualifications necessary to
  do the job?
• Top-down
• Once cut score is set, who are the best
  candidates?
• Banding
• Can we statistically figure out a group of
  applicants that are basically equal even though
  they have slightly different scores?

10
Discussion Question

• Do you think it is ethical to use banding to
  increase diversity?

11
Discussion Question?

• All models we have discussed thus far take into
  account KSAs as a basis for hiring. If we know
  that fit is also an important factor that
  impacts performance, turnover, etc., how should
  fit be included in the selection process and
  decision?

12
Take-Home Points

• Aware of the factors to consider when determining
  which assessment methods to use
• Understand different methods of combining
  predictors and creating cut scores
• Realize that fit is also a big part of the
  equation