Jn. Farrell et al., THE MEASUREMENT OF PERFORMANCE VALENCE - AN EXAMINATION OF CONSTRUCT-RELATED EVIDENCE, Organizational behavior and human decision processes, 60(2), 1994, pp. 157-178
Garland's (1985) summated valence measure is an indirect measure of go
als, not an improved operationalization of the valence construct, as d
emonstrated in this paper. In Study 1, a meta-analysis showed that Gar
land's measure of valence correlated negatively with both personal goa
ls and performance while all other measures of valence correlated posi
tively with personal goals and performance. In Study 2, a protocol ana
lysis showed that the majority of subjects estimated valence based upo
n a comparison between personal goals and possible future performance
levels. In addition, subjects' use of second level outcomes in making
ratings sharply decreased across items, indicating the performance val
ence measure wasn't eliciting the information necessary to calculate v
alence according to Vroom's (1964) definition. Evidence from both the
meta-analysis and the protocol analysis support the assertion that Gar
land's measure of performance valence is best conceptualized as an ind
irect measure of goals. (C) 1994 Academic Press, Inc.