Predicting Employee Performance: The Predictor Problem Flashcards
In the context of selecting predictors of job performance, _________________ refers to the extent to which performance on a measure is unaffected by measurement error, while ____________ refers to the degree to which the measure assesses what it was designed to asses.
- Reliability
- Validity
There are three types of validity: _______________, __________________, and __________________; the type of greatest concern depends on the nature and purpose of the predictor. In I/O, a predictor’s _____________ validity is usually of particular interest.
- Content
- Construct
- Criterion-related
- Criterion-related
There are ___ steps involved in evaluating a predictor’s criterion-related validity.
7.
Step 1: ___________________: Identify the knowledge, skills, abilities, and other characteristics required for successful job performance.
Conduct a job analysis.
Step 2: ________________ - Select/develop a predictor that measures the attributes identified by the job analysis.
Select or develop the predictor.
Step 3: ____________________ - Administer the predictor to a sample of job applicants (predictive validity) or current employees (concurrent validity); obtain criterion information for all individuals in the sample.
Administer the predictor and assess the criterion.
Step 4: ___________________ - Calculate the criterion-related validity coefficient to determine if there is a statistically significant relationship between predictor and criterion scores.
Correlate the predictor and criterion scores.
Step 5: __________________ - Determine if the predictor will have adverse impact - i.e., if use of the predictor will discriminate against a legally protected class of individuals.
Check for adverse impact.
Step 6: ____________________ - Determine if use of the predictor will increase the proportion of correct decisions.
Evaluate incremental validity.
Step 7: _________________ - Perform steps 3 through 6 with a new sample. Because all of the chance factors that maximized the original validity coefficient are not present in the second sample, the cross-validation coefficient is usually smaller than the original coefficient. This phenomenon is referred to as _______________.
- Cross-validate
- Shrinkage
Title ___ of the _______________ Acts of 1964 and other legislation prohibit discrimination in hiring, placement, training, promotion, and retention on the basis of race, religion, gender, national origin, age, or disability. The provision of these laws are enforced by the ____________________ (EEOC), which issued the Unified Guidelines on Employee Selection Procedures.
- VII
- Civil Rights
- Equal Employment Opportunity Commission
As defined by the Uniform Guidelines, a selection test or other employment procedure is discriminating against members of a legally protected group - i.e., is having an ______________ - when use of that procedure results in a substantially different selection, ____________, or promotion rate for members of that group.
- Adverse impact
- Placement
The ________________ is often used to determine if a procedure is having an adverse impact. The hiring rate for the majority group is multiplied by 80% to determine the minimum hiring rate for the minority group. As an example, if use of a test results in a 50% hiring rate for African Americans and 90% rate for whites, the test would be having an adverse impact for African Americans because 90% times 80% is 72%, which is greater than the hiring rate for African Americans.
80% (Four-Fifths) Rule.
Bonus: The 4/5 Rule requires that for a finding of adverse impact, a pattern of discrimination must be established.
______________________ occurs when a measure is valid for one group but is not valid (or is significantly less valid) for another group. The characteristic that distinguishes between the two groups - e.g., gender, race, or age - is referred to as a ____________ variable.
- Differential validity
- Moderator
_____________ occurs when members of one group consistently obtain lower scores on the predictor than members of another group, but the difference in predictor scores is NOT related to differences in scores on the criterion.
Unfairness.