Which reliability statistic is specifically designed for dichotomous test items?

Prepare for the American Board of Professional Psychology (ABPP) Exam with flashcards and multiple-choice questions. Each question includes insights and explanations to help you excel. Get ready for your certification journey!

When test items are dichotomous—scored as 0 or 1—the reliability statistic designed for that data type is Kuder-Richardson. This family, especially KR-20, is built specifically for binary scoring, taking into account how many respondents get each item right and how those item scores relate to the overall test score to yield a single reliability estimate.

Split-half reliability looks at the correlation between two halves of the test and then adjusts with Spearman-Brown; Cronbach’s alpha is a general measure of internal consistency that applies to items on a broader scale, not restricted to binary data; Spearman-Brown is just a correction for length changes in a reliability estimate. Because KR-20 directly targets the dichotomous scoring structure, it’s the best fit for this scenario.