A company uses a cognitive ability test, a structured interview, and a work sample test to select software engineers. These are all examples of...
ACriteria — the outcomes the company wants to predict
BPredictors — measures used to forecast future job performance
CJob analysis methods — tools for understanding job requirements
DPerformance appraisal instruments — tools for evaluating current employees
In the predictor-criterion framework, predictors are measures collected before or during hiring that are used to forecast criteria (job performance outcomes). The cognitive ability test, interview, and work sample are all predictors — they generate scores used to predict how well candidates will perform on the job. The criterion is the actual job performance the company hopes to forecast.
Question 2 True / False
Using multiple predictors in a selection system always improves prediction over using a single predictor.
TTrue
FFalse
Answer: False
Adding predictors improves prediction only when the new predictor captures variance in job performance that is not already captured by existing predictors — that is, when it has incremental validity. If a new predictor is highly correlated with existing predictors, it adds redundancy rather than new information. For example, adding a second cognitive ability test to a battery that already includes one may add little incremental validity because both tests measure largely the same construct.
Question 3 Short Answer
What is the predictor-criterion framework, and why is it central to personnel selection?
Think about your answer, then reveal below.
Model answer: The predictor-criterion framework distinguishes between the measures used to evaluate candidates (predictors) and the outcomes those measures are supposed to forecast (criteria, typically job performance). It is central because it structures the entire validation process: a selection system is valid to the extent that predictor scores actually predict criterion scores.
This framework forces practitioners to be explicit about what they are measuring (predictors) and what they are trying to predict (criteria). Without this distinction, organizations may select on characteristics that feel relevant but have no demonstrated relationship to actual job performance. The framework also clarifies what validation means: empirically demonstrating a statistical relationship between predictor and criterion.
Question 4 Multiple Choice
An organization hires all applicants regardless of their test scores, then correlates test scores with subsequent job performance. This describes which validation strategy?
AContent validity
BPredictive criterion-related validity
CConcurrent criterion-related validity
DConstruct validity
Predictive validity involves collecting predictor data first, then waiting to collect criterion data later (after the person is on the job). By hiring all applicants regardless of test scores, the organization avoids restriction of range — a key methodological advantage. Concurrent validity, by contrast, collects predictor and criterion data from current employees at the same time. Content validity evaluates whether the test content representatively samples the job domain.