統計学輪講 第2回
日時 | 2020年4月14日(火) 14時55分 ~ 16時35分 |
---|---|
場所 | Zoomオンライン開催(URLはシラバスまたは参加者メーリスをご確認ください) |
講演者 | 大森 裕浩 (経済学研究科) |
演題 | A multivariate randomized response model for binary data: An application for understanding drug administration policies |
概要 |
Sensitive topics are commonly involved in business, management, and social sciences studies. However, such topics are not easily investigated via empirical studies, especially when they concern matters that may have disciplinary or legal implications. This paper analyzes the implementation of drug administration policies using nurses' sensitive responses. From the employees' perspective, responding to questions that ask about whether they follow hospitals' guidelines in handling drugs is likely to be sensitive. In this paper, we propose a statistical method combining the randomized response technique, probit modeling, and Bayesian analysis to analyze a large-scale online survey of multiple binary randomized responses from a hospital cluster that serves approximately one-fifth of the population in Hong Kong. A statistical challenge is that nurses' true sensitive responses are unobservable because of a randomization scheme that protects their data privacy. From the heat map of the probabilities of not following hospital guidelines adequately, we can learn which hospital, which rank of nurses, or which level of nursing experience needs more support to reinforce the drug administration policies. Four main contributions of the paper are highlighted. The first is the construction of a generic approach in modeling multivariate sensitive binary data collected from the randomized response technique. The second is understanding drug administration policies that potentially involve sensitive topics. The third is studying the dependence of different types of administrative error. The last one is the calculation of an overall attitude score using sensitive responses. This particular healthcare example on drug administration policies presents a scientific way to elicit managerial strategies while protecting data privacy through analytics. This is a joint work with Amanda M.Y. Chu, Hing-yu So and Mike K.P. So. |