Objective Measurement of Subjective Phenomena

8. Validity

Validity involves, in its broadest construal, whether a scale assesses the construct it was intended to assess (Cronbach & Meehl, 1955).

Traditional Tripartite Delineation of Validity – The 3C’s of Validity

Content Validity:

Judgment by experts that domain of interest has been properly sampled

Criterion-Related Validity:

Correlation of test score with one or more criteria it should predict:

  • Predictive Validity – test score correlates with criterion obtained at a later point in time

Example 26

Predictive validity example:

SAT scores and their correlation with college grades.
  • Concurrent Validity – test score correlates with contemporaneous criterion

Example 27

Concurrent validity example:

Score on Stanford Binet test of intelligence correlates with score on the Wechsler Adult Intelligence Scale.

Construct Validity:

Validity of the use of a score for an intended purpose – the core of the validation of a measure.

For example, does it:

  • Correlate with other measures of the same purported construct?
  • Correlate with criteria it is supposed to correlate with?
  • Vary appropriately across contexts (e.g., anxiety scores increase when confronted with anxiety provoking situation)?
  • Yield appropriate factor analytic solution? If one hypothesizes a single dimension is being assessed, does a factor analysis show that a one-factor model is optimal?