An Empirical Comparison of Direct Questioning, Scenario, and Randomized Response Methods for Obtaining Sensitive Business Information*
通过全国制造业CEO邮件调查，比较直接提问、情景法和随机应答法在估计敏感商业行为普遍性上的表现，发现基于他人的直接提问或情景法估计值最高，而基于个人的直接提问最低。
ABSTRACT Certain business practices include legal but ethically questionable activities. Surveys intended to determine the nature and extent of such activities must employ questioning methods which mitigate the inherent threat of sensitive questions and account for social desirability effects. This study uses a national mail survey of chief executive officers (CEOs) of manufacturing firms to compare the performance of direct questioning, scenario, and randomized response methods for estimating the prevalence of several sensitive business practices. The direct questioning and scenario versions used self‐reporting (individual‐based) questions, as well as the CEO's perceptions of the extent to which others engage in questionable activities (other‐based). In general, the estimates of the prevalence of selected questionable activities were lowest when the individual‐based direct questioning was used and highest when other‐based (either direct questioning or scenario) methods were used. The individual‐ based scenario and randomized response estimates represented intermediate estimates. Suggested guidelines for using the three methods for eliciting sensitive information are discussed.