What does the Kuder-Richardson formula measure?

The Kuder-Richardson formula specifically measures the reliability of tests that yield dichotomous outcomes, meaning the responses fall into two categories, such as correct/incorrect or yes/no. It is particularly useful for determining the internal consistency of these types of tests, assessing how closely related those items are as a group. The formula provides a means of quantifying the extent to which different items measure the same underlying construct or trait, thereby offering insights into the reliability of the test as a whole.

In contrast, while the overall score reliability of a test could be related, the Kuder-Richardson formula is designed explicitly for dichotomous tests. The variability of test answers relates to how scores differ among respondents rather than how consistently the test items measure a concept. Lastly, the validity of test constructs pertains to whether a test is accurately measuring what it purports to assess, which is outside the scope of what the Kuder-Richardson formula evaluates. The focus here is firmly on the reliability aspect, particularly for tests with binary response options.