Tests B and C correlate 0.901 when given to the same pupils in grades III to VIII¶

Claim · All claims
Evidence · 1 study · design r2 · q2
Currency · newest study 1922 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? On 103 pairs of scores from pupils in grades III to VIII in a small city system, Tests B and C showed a positive coefficient of correlation of 0.901. → Wilson 1922

Evidence¶

Wilson 1922¶

Wilson, G. M. (1922). Language error tests. Journal of Educational Psychology, 13(6), 341–349. doi:10.1037/h0064444

q2 · i? · design · r2

Same-pupil administration of two parallel forms in a small city system; 103 paired scores yielded "a positive coefficient of 0.901", taken as further evidence of the forms' equality.

"Tests B and C were given at the same time to pupils in grades III to VIII, inclusive, in a small city system. On the basis of 103 pairs thus secured, the coefficient of correlation was figured. The results showed a positive coefficient of 0.901."

Discussion¶

Related Claims¶

• Percentage of pupils making each error runs surprisingly close across the three test forms — related
• The three Wilson Language Error Tests (A, B, and C) approach equality across grades III to XII — a broader claim this one bears on
• In one class of 13-year-old boys, memory scores were kept to a steady rank order by repeated practice (successive tests correlating +.87 to +.90) before fatigue testing began — related
• Spelling test scores correlate highly with teachers' year-long marks (average coefficient .69 across grades 3–8) — related
• Correlations between two series of mirror-reading tests on the same 16 subjects range widely depending on how errors are scored, from .087 to .700. — related
• The equal-group tests showed high test-retest and between-group correlations, supporting the reliability of the arithmetical-reasoning marking method. — related