, is the alternate-choice item. It consists of an incomplete statement of a
proposition along with two or more alternative completions, only one of which
makes the statement true . For example:
An eclipse of the sun can only occur when the moon is:
(I) full (2) new.
Items of this kind do not call for absolute judgments of truth or fa lsity. The
critical element in the statement they make should be quite clear. Their indices of
di scrimination should be higher on the average than the indices of comparable
true- fa lse items given to the same examinees. The test scores therefore should be
more reliable. A recent study has verified this expectation . Students (N = 28)
enrolled in a class in educational measurement took parallel 25-item true- fa lse
and alternate-choice tests on each of eight units of instruction in the course. The
Kuder- Richardson 20 reliability coefficients for the true-false tests ranged from
. 13 to .7 1, with a mean of .47 . Those for the alternate-choice tests ranged
from .56 to .76, with a mean of .66 (Ebel, 1982).