A popular method for assessing compliance-gaming behavior involves having people rate lists of preformulated message strategies for likelihood of use. This “selection procedure” has been employed much more frequently than the alternative “construction procedure,” a method that requires people to generate their own message strategies. The present article argues that the selection procedure is much less sensitive than the construction procedure to the effects of situational and individual-difference variables on compliance-gaining behavior. The article further suggests that the insensitivity of the selection procedure is due to a type of social desirability bias known as the item desirability effect. Seven studies were carried out testing the Marwell and Schmitt (1967) and Wiseman and Schenk-Hamlin (1 981) strategy checklists for the item desirability effect. These studies found that (a) the likelihood of use ratings prouided for the strategies on both checklists could be accurately predicted by the rated social appropriateness of the strategies, (b) likelihood of use ratingsfor preforrnulated strategies haue relatively poor reliability, and (c) the construction procedure is much less susceptible to social desirability biases than the selection procedure. On the basis of these and related findings, it is recommended that researchers eschew the use of strategy checklists in future research.