Home › MCQs › Education › Question #6774
Back to Questions
Education QUESTION #6774
Question 1
A test developer wants to use the Split-Half method to estimate reliability. After splitting the test into odd and even items and correlating the halves, they get r = 0.70. However, this underestimates the reliability of the full test. Which formula corrects for this, and why is the correction necessary?
  • The KR-20 formula; because internal consistency must account for item difficulty
  • The Spearman-Brown prophecy formula; because reliability increases with test length, and the correlation between two halves reflects reliability of a test only half as long✔️
  • The inter-rater reliability coefficient; because two independent scorers are effectively two test halves
  • The KR-21 formula; because it handles non-dichotomous scoring
Correct Answer Explanation
Split-half reliability correlates two halves of the test, but a half-test is less reliable than the full test. The Spearman-Brown formula corrects for this by estimating the reliability of the full-length test from the split-half correlation. This is a fundamental principle: longer tests, all else equal, are more reliable because they sample the domain more broadly.