[Kelley argued that OS's judgment of an inverse relation between inducement magnitude and attitude inference] is probably associated with assumptions (unchecked in Bern's work, as far as I know) that there is a distribution of opinion toward the task, and only the more favorable subjects complied in the $1 case and almost all, favorable or not, complied in the $20 case.

p. 226; as cited in: Yaacov Trope, "Inferential processes in the forced compliance situation: A Bayesian analysis." Journal of Experimental Social Psychology 10.1 (1974): 1-16. - "Attribution theory in social psychology." 1967