Consider the concept of "problem drinking." Imagine that the researcher defined it as having four (female) or five (male) or more drinks in a row during one drinking episode. What are some potential problems you can foresee in terms of the validity of such a measure for the concept of problem drinking? What are some ways that you could test the reliability of such a measure?

Question

What will be an ideal response?

Answer 1

I could see that a potential problem with the validity measures is that if the concept is "problem drinking" such a question may not really measure that. For instance, is the four or five drinks over an entire day or just in a few hours? Is the problem drinking occurring when a person is drinking water regularly? Such concerns would reduce the validity of the measure because they would not adequately cover the entire concept of "problem drinking" (content validity question) nor could it really be substantiated by observations from friends who might have also been drinking (criterion validity). Also, maybe someone would intentionally lie or could forget how many drinks they had (face validity). Is problem drinking measured in this way connected to a variety of other measures (construct validity)? Finally, we could test the reliability of such a measure by asking a question that is similar, but different to separate respondents and see if there are similar results.