First page Back Continue Last page Overview Graphics

4) Evidence based on Relations to

4) Evidence based on Relations to

Other Variables

* Criterion-related evidence

(concurrent, predictive)

* Convergent and discriminant evidence

Notes:

The fourth source of validity evidence is based on relations to other variables. Evidence of relationships to other variables is commonly found by using correlations. If observed relationships match predicted relationships, then the evidence supports the validity of the interpretation.

A criterion is a dependent variable about which we want to make a statement. Criterion–related evidence takes two forms: concurrent or predictive, based on how far into the future the criterion variable is measured. Criterion–related concurrent evidence requires that both variables are captured at one point in time. For example, two measures of fourth grade reading are collected at the same time, and then a correlation is observed. On the other hand, predictive criterion–related evidence is based on two sources of data collected at different points in time. For example, in the case of fourth and eighth grade reading performance, if one is good in reading at the fourth grade, to some extent, one’s performance at eighth grade should also be good. A positive relationship would show evidence of predictive validity of fourth grade reading scores.

The second grouping is convergent and discriminant evidence. Convergent is when two measures are converging on the same construct and thus, should be strongly related. A fourth grade reading test, for example, should correlate very strongly with another fourth grade reading test. Discriminant is when two measures claim to be measuring different things and thus, should not be highly related. For example, the fourth grade reading test should correlate moderately with a fourth grade mathematics test and very poorly with a personality test measuring levels of introversion and extroversion.