Composition scores based on only two compositions are too unreliable for measurement, requiring roughly 17 compositions to match the Terman test's reliability¶

Claim · All claims
Evidence · 1 study · associational r1 · q3
Currency · newest study 1921 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? The reliability of Composition based on two compositions was 0.25 ± 0.10; applying Brown's formula, 17 compositions would be needed to reach 0.85. → Wyman 1921

Evidence¶

Wyman 1921¶

Wyman, J. B., & Wendle, M. (1921). What is Reading Ability? The Journal of Educational Psychology, 518-531. https://eric.ed.gov/?q=%22What+is+Reading+Ability%22

q3 · i? · associational · r1

Grade VIII reading study reliability analysis: composition (two written topics) yielded the lowest reliability coefficient in the battery. The general conclusion adds that "the average score on from 15 to 20 compositions must be taken" if compositions measure ability. Reliability coefficients are printed, but no standardised effect size, so no magnitude is coded.

"If compositions are to be used, by applying Brown's formula it can be seen that in order to have a relia- bility coefficient comparable to the Terman (0.85) it would be necessary to give 17 compositions"

Discussion¶

Related Claims¶

• Rate tests that mechanically control comprehension yield more reliable results than uncontrolled rate tests — related
• Comprehension strategy instruction shows a large overall effect, but reliability is undermined by few standardized measures — related