ALERT! Warning: your browser isn't supported. Please install a modern one, like Firefox, Opera, Safari, Chrome or the latest Internet Explorer. Thank you!
Log inLog in to ACORN Wiki. | RegisterSign on to ACORN Wiki.

Calculating validity

There are a variety of methods of arriving at a coefficient of correlation for validity. Arguably though, the most critical element of validity is face validity, which requires no calculation at all, but lies in the eye of the beholder.

Statistical concepts of validity rest on the premise that a test score should predict something. The type of validity cited is dependent on the type of prediction being made.

A test, for example, which discriminates between individuals seeking mental health services and individuals not receiving such services can be said to have criterion validity. A test is said to have concurrent validity when it correlates with some other related measure taken currently. Concurrent validity is closely related to the idea of constuct validity, where a test is presumed to be valid if the items are shown to load on a known factor (i.e. global distress) measured by similar tests. The concept of predictive validity is closely related to other measures of validity, except that it is generally used to refer to the use of a test to predict some criterion variable which is measured in the future, rather than concurrently.

For example, a test may have good criterion validity in that it distinguishes between individuals in a mental health inpatient unit as opposed to people receiving treatment in the community, yet have poor predictive validity with regard to predicting which outpatients will be admitted to the hospital in the future.

-- JebBrown - 06 Jan 2007
This site is powered by FoswikiCopyright © by the contributing authors. All material on this collaboration platform is the property of the contributing authors.
Ideas, requests, problems regarding ACORN Wiki? Send feedback
Syndicate this site RSS
^