skip to main content skip to footer

A Note on the Validity of Scores Corrected for Guessing in the Case of Certain Figure-Classification Items

Lord, Frederic M.
Publication Year:
Report Number:
ETS Research Bulletin
Document Type:
Page Count:
Subject/Key Words:
Guessing (Tests), Scoring Formulas, Spatial Ability, Test Validity, Visual Measures


The validity are compared of: 1) number right score (c); 2) score corrected for guessing (c-1/4w); 3) optimally weighted average of rights score (c) and wrongs score (w). Data was collected by C. Myers for his study of "The Factorial Composition and Validity of a Speeded Test." Myers gave four forms of a figure-classification test to first year midshipmen at the U.S. Naval Academy in Annapolis. He distributed the forms in rotation in the order that the men were seated. Each man took only one form. This study relates to three of these groups--group 2 (182 cases); group 3 (187 cases); and group 4 (170 cases) and to the 30-item subtests taken by each group. One of the subtests is the same for each group--sub- test x. The other is different for each group--subtest 2, 3, or 4. The subtests are highly speeded. First-year grades are used separately as criteria for determining the validity of the several scoring methods: 1) class ranking in Marine Engineering; 2) class ranking in Mathematics; and 3) over-all relative class standing. The validity of the score corrected for guessing was obtained by the formula shown. Validity coefficients are given in Table I. Table II gives scoring weights to be assigned to wrong answers, when right answers are given unit weight, in order to maximize the correlation of test score with grades. (SGK).

Read More