Psychometric truths in the absence of psychological meaning: A reply to Zuckerman and Knee.

Abstract

M. Zuckerman and C. R. Knee (1996) contend that the relationship we reported between selfenhancement and psychological maladjustment (C. R. Colvin, J. Block, D. C. Funder, ! 995) is "not warranted" because a difference score was used to operationalize self-enhancement. In this article we discuss the issues raised by Zuckerman and Knee, the warnings historically issued about the use of difference scores, and the problems associated with alternative statistical procedures. We conclude that difference scores can provide statistically valid and psychologically meaningful results.

Testing discrepancy effects: A critique, a suggestion, and an illustration

“…As was previously noted, Colvin et al (1996) also argued that a test of the interaction is equivalent to the problematic residual difference procedure (see also Asendorpf & Ostendorf's [1998, p. 956] concurrence with this position). But the residual difference procedure and the interaction are very different analyses.…”

“…The solution, according to Zuckerman and Knee, is to replace the simple difference procedure with a test of the interaction. In their response, Colvin, Block, and Funder (1996) argued that the use of a simple difference is justified if it is an appropriate definition of the construct. Given that the construct is self-enhancement, the difference between self-and other-favorabilityindices provides "a direct, easily understandable, and readily analyzable index" of that concept (p. 1253).…”

“…According to J. Cohen and P. Cohen (1983, p. 416), (1) When a and b are standardized, sd 5 1, and the equation can be simplified to (2) In their response, Colvin et al (1996) provided some of the correlations that make up Equations 1 and 2. Correlations of self-favorability with the criterion (r ya ) ranged from .31 to .47, with a median of .42; correlations of otherfavorability with the criterion (r yb ) ranged from .42 to .56, with a median of .48; the difference correlations ranged from 2.23 to 2.33, with a median of 2.24.…”

A Critical Review of Self-perceptions and the Positive Illusory Bias in Children with ADHD

“…Researchers endorsing this viewpoint contend that discrepancy scores represent constructs that are theoretically distinct from the constructs represented by the component variables (Tisak and Smith 1994a, b). Colvin et al (1996) asserted that for this reason, construction of difference scores facilitates a different type of parsimony than that advocated by opponents of the difference score construct. They noted that difference scores provide researchers with the opportunity to investigate the concepts that might otherwise be ''awkward to assess, comprehend, and evaluate'' (Colvin et al 1996(Colvin et al , p. 1253.…”

Applying Depression-Distortion Hypotheses to the Assessment of Peer Victimization in Adolescents

“…Indeed, there is no current consensus regarding an analytic approach for examining informant discrepancies that is appropriate for all types of datasets (cf. Cohen & Cohen, 1983;Colvin, Block, & Funder, 1996;Edwards, 1994;Rogosa et al, 1982;Zuckerman, Gagne, Nafshi, Knee, & Kieffer, 2002).…”