Analogs in Luce's global psychophysical theory of Stevens's psychophysical regression effect.

This article is based on the magnitude production representation in Luce's (2004, 2011) global psychophysical theories for subjective intensity of both binary and unary continua. It is shown that the slopes of certain predicted linear relationships for magnitude production and estimation differ when, in the two procedures, the respondent-generated reference signals differ. This contrasts with Stevens's (1975) regression effect, which arises because the exponents of the psychophysical scale differs with procedural changes. Four experiments are suggested to evaluate these predictions, which, importantly, should be analyzed for individual respondents, not data averaged over respondents.