Intercorrelations among analyzed trait ratings show a spread of the general impression, and their size varies extremely with the rating method, from central tendencies of 0.3 to 0.9¶

Claim · All claims
Evidence · 1 study · associational r1 · q2
Currency · newest study 1921 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q2 i? Three rating methods for the same fact yielded central tendencies of trait intercorrelations of 0.3 to 0.9, making the general factor impossible to diagnose and partial out. → Knight 1921

Evidence¶

Knight 1921¶

Knight, F. B., & Franzen, R. H. (1921). Pitfalls in rating schemes. The Journal of Educational Psychology. https://doi.org/10.1037/h0073817

q2 · i? · associational · r1

Comparisons across a Boice Teachers Rating Card (over 100 teachers, central tendency +0.5), a rank-order card in a Massachusetts system (100 teachers, many correlations 0.9), and competent judges rating 100 university instructors (average +0.32) show method-dependent variability.

"Using three different methods for obtaining the same fact, we obtain central tendencies of inter-correlations of 0.3 to 0.9, which means that inter-correlations of traits is a highly variable function of the method of obtaining them."

Discussion¶

Related Claims¶

• IAR and CDR factor scores intercorrelate too weakly to confirm the two instruments measure the same locus of control construct — related
• Card-slip ranking yields a wider spread of ranks than list-based ranking typically produces