PSYC 3020 Chapter : Ch 7
Document Summary
Researchers began to point out problems with this. Does it actually measure a single concept if that"s what it"s supposed to. If we have different subgroups of test takers we want to make sure they all perform the same: -relations with other variables (formerly cirterion-related and construct validity. The people who did poorly didn"t get a job. Not valid in terms of how they were being used: content validity. Evidence based on test content: construct: an attribute, trait, or characteristic that is abstracted from observable behaviors. * the domain of interest was chapters 1-4 and want to cover each of them equally (on our test 1). We want to have an overlap as perfect as possible (1/4 information from each chapter= content valid: criterion-related validity. Evidence based on relations with criterion variables: criterion-related validity: evidence that test scores correlate with or predict independent behaviors, attitudes, or events, 2 types, predictive.
