Selection Methods and Validation
3
Validity Types
Criterion content construct
2
Criterion Subtypes
Predictive and concurrent
1
Highest Validity
Work samples predict best
The Three Validity Types
Criterion
Statistical correlation
Compares test scores to actual job performance data. Predictive (test before hire, measure later) or concurrent (test current employees, compare to current performance).
Content
Expert judgment
SMEs confirm test items match actual job tasks. No statistical analysis required. A typing test for a typist position is a classic example.
Construct
Measures a trait
Confirms the test measures an abstract quality (intelligence, leadership, emotional stability) that is linked to job success. The hardest type to establish.
Interviewer Bias Types
the exam gives you a scenario and asks you to name the biasHalo
One positive trait overshadows everything
Horn
One negative trait overshadows everything
Primacy
First impression dominates the evaluation
Recency
Last thing said dominates the evaluation
Contrast
Rating influenced by the previous candidate
Similar-to-Me
Favoring candidates who resemble the interviewer
ready to test your knowledge of validity?
Practice questions on validity types, interview bias, and selection methods
Practice Selection QuestionsExam Traps
Content validity does not require statistical analysis
Content validity is established by expert judgment, not statistics. Subject matter experts confirm that test items represent the actual job duties. Criterion validity, on the other hand, requires statistical correlation between test scores and job performance. The exam will ask which type needs data and which relies on expert review.
Structured interviews have higher validity than unstructured
Structured interviews use the same questions, same order, same scoring rubric for every candidate. This dramatically reduces interviewer bias and produces higher predictive validity. The exam will always favor structured over unstructured when asking about best practices.
Work samples have the highest predictive validity
Of all selection methods, work samples and job tryouts consistently show the strongest correlation with actual job performance. Assessment centers rank second. Unstructured interviews rank among the lowest. The exam tests this hierarchy.
Valid vs Reliable
Valid = accurate (measures what it claims to). Reliable = consistent (same results each time). A test can be reliable without being valid, but not valid without being reliable.
Content = No Stats
Content validity = expert judgment. Criterion validity = statistical correlation. Never confuse the two.
A weather forecast is valid if it actually predicts the weather. A hiring test is valid if it actually predicts job performance. If the forecast says sunny and it rains, the forecast was not valid.
A thermometer is reliable if it gives the same reading every time in the same conditions. A hiring test is reliable if candidates score similarly when retested. But a thermometer that always reads 72 degrees is reliable and completely useless (not valid).
Content = experts review the forecast model and confirm it uses the right data points. Criterion = compare past forecasts to actual weather outcomes (statistical). Construct = prove the model measures "weather" and not something else entirely.
Work samples are like looking out the window. Structured interviews are like using a consistent forecast model. Unstructured interviews are like guessing based on a feeling.
Ready to test your SHRM-CP knowledge?
1,200+ practice questions written by certified professionals.
Start Practicing SHRM-CP