Interpreting a validity coefficient (Chapter 9)

Squared correlations and “variance explained”

Guidelines or Norms for a field

Statistical significance

Estimating practical effects

BESD

Taylor-Russel Tables

Utility analysis

Sensitivity/Specificy

How much of the variance in the criterion can be explained by a test

Limitations

In some cases it is technically incorrect to square the correlation

Gets interpreted on a quadratic scale while being measured on a normal scale

Limitation

Displays the number of successful and unsuccessful predictors made based on a correlation

Assumed equality of successes on a test and on criterion

Present success rate for a given baserate, validation coefficient and test rate

Whether the gains of a measurement outweight it's costs

Sensitivity: identify people with disorders

Specificy: identify people who do not have a disorder

Inreliable correlations might occur due to small sample sizes

The product of the sample size and the size of the validity coefficient in the sample provides confidence, that a test is correlated with a criterion in the population

validity coefficients get interpreted with regard to a particular field -> information about the existing research in that field is helpful

A=D

B=C