Test valuation: Test valuation means that test scores rate significantly to job performance or to another job-relevant criterion. The stronger the relationship between test result and performance, the more effective the test is as a selection tool.
Steps to validate a test: There are two steps to validate a test. These are discussing below –
Empirical approaches to test validation attempt to relate test scores with a rob-related criterion, usually performance. If the test actually measures a job-related criterion, the test and the criterion exhibit a positive correlation between 0 and 1.0. The higher the correlation, the better the match.
- Predictive validity: Predictive validity is determined by giving a test to a group of applicants. After these applicants have been hired and have mastered the job reasonably well, their performance is measured. This measurement and the test score are then correlated.
- Concurrent validity: Concurrent validity allows the personnel department to test present employees and correlated these scores with measures of their performance. This approach does not require an interval between hiring and mastery of the job.
When the number of subjects is too low to have a reasonable sample of people for the test, rational approaches are used. These approaches are considered inferior to empirical techniques but are acceptable validation strategies when empirical approaches are not feasible.
- Content validity: Content validity is assumed to exist when the test includes reasonable samples of the skills needed to successfully perform the job. A typing test for an applicant who is being hired simply to do typing is an example of a test with content validity.
- Construct validity: Construct validity seeks to establish a relationship between performance and other characteristics that are assumed to be necessary for successful job performance. Tests of intelligence and scientific terms would be considered to have construct validity if they were used to hire researchers for a chemical company.