Ds. Kiker et Sj. Motowidlo, EFFECTS OF RATING STRATEGY ON INTERDIMENSIONAL VARIANCE, RELIABILITY,AND VALIDITY OF INTERVIEW RATINGS, Journal of applied psychology, 83(5), 1998, pp. 763-768
This study experimentally manipulated various rating strategies to tes
t effects on interdimensional variance, reliability, and validity of i
nterview ratings. Undergraduate research participants (N = 180) rated
60 interview transcripts under 3 experimental conditions that represen
ted different rating strategies. Interdimensional variance was higher,
internal consistency reliability was lower, and interrater reliabilit
y was higher in the rating condition that gave raters the least opport
unity to form and use general impressions of interviewees. Rating stra
tegy had no discernible effects on validity. To interpret these result
s, the authors argued that the rating strategies used in this study di
ffered in the degree to which they allowed raters to form idiosyncrati
c general impressions of interviewees. These idiosyncratic halo effect
s explain why rating strategies that raise internal consistency reliab
ility simultaneously lower interrater reliability.