Construct Validity – Whether your measure actually measures what it is suppose to measure
- Worry about systematic error, other conceptual variables that affect measure
Face Validity – Whether content of measure appears to reflect content being measured
- Useful measure to see whether test measures what it is suppose to measure
- Not always useful, for example, when measuring undesirable trait
e.g., Racism, people might not answer items accurately - Some valid measure lack face validity
e.g., REM measure dreaming but has no face validity - Easiest way to assess construct validity
Content Validity – Degree to which measured variable appears to have adequately sample from potential domain of question that relate to conceptual variable of interest
e.g., Exam with one chapter; IQ test with only math questions → both have low content validity
∴ Face and Content Validity should be used in initial stages of question development; the ones below are superior
Criterion Validity (more objective)
Predictive Validity – Extent to which measure predicts future behaviour
e.g., GRE has predictive validity
- Correlate score on measure with future behaviour that measure should predict
Concurrent Validity – Extent to which measure correlates with behaviour at the same time as measure of interest
- Be able to distinguish one group with another who are similar
e.g., Distinguish between Bi-Polar and Depression - Involves correlation measure with behaviour
Convergent Validity – Score related to score on other measure of same or similar concepts
Discriminant Validity – Extent to which measures are unrelated to scores of measure on different concepts
0 Responses to “Validity In-Depth”