Reliability coefficients computed between parallel forms tend to be conservative estimates.

Question

Quizplus · Accepted Answer

Parallel forms reliability involves creating two different versions of a test that are designed to be equivalent, measuring the same constructs in the same way. The reliability coefficient computed between these parallel forms tends to be conservative because it not only captures the true score variance but also includes the variance due to the specific items of each form, which might not perfectly overlap. This can lead to underestimation of the true reliability of the construct being measured.

Reliability Coefficients Computed Between Parallel Forms Tend to Be Conservative