What type of instrument is the Kuder-Richardson (KR-20) coefficient primarily used for?

The Kuder-Richardson (KR-20) coefficient is specifically designed to measure the internal consistency reliability of instruments that use dichotomous response formats, such as true/false or yes/no questions. This coefficient evaluates how consistently different items on a test measure the same underlying construct.

In this context, response formats that yield only two possible outcomes are essential for the KR-20 calculation, as the formula computes the covariance between all pairs of items that have these binary answers. By assessing how well these items correlate with one another, the KR-20 provides insight into the reliability of the instrument as a whole.

In contrast, instruments with multiple-choice responses may require different reliability measures, such as Cronbach's alpha, which is better suited for items with more than two response options. Numeric scoring instruments may also use Cronbach’s alpha for testing reliability, while qualitative data instruments typically do not apply these coefficients as they focus on non-numeric, descriptive information and not on the consistency of responses.