Criterion-related validity examines the actual relationship between variables that purportedly are related. For example, universities ask applicants to take national tests such as SAT or ACT because these tests are supposed to predict student success in college. Criterion-related validity checks this claim by measuring the relationship between students' test scores and their subsequent college grades.
Research is a special activity that is designed to yield accurate information on its topic. For this reason, researchers are careful to collect data systematically, analyze it logically and then check for various kinds of validity. If the conclusions of a study are not valid in multiple ways, the study is not useful for those who need its information.
Criterion-related validity allows researchers to look at the ultimate value of their data and its analysis. For example, test makers claim that their tests are effective predictors of certain things such as success in a professional school or the ability to do a certain skill. Criterion-related validity tests those claims. People get their license to drive after taking a written test and a driving test where they demonstrate their ability to handle a car and follow the law. These tests could be examined for criterion validity by comparing drivers' scores with drivers' actual records.