Title: Estimating the dependence of sensitive responses of mixed types in randomized response models
Authors: Mike So - The Hong Kong University of Science and Technology (Hong Kong)
Amanda Chu - Hang Seng Management College (China)
Thomas Chan - Hong Kong University of Science and Technology (Hong Kong) [presenting]
Abstract: The aim is to estimate the dependence of sensitive responses of multiple types obtained from randomized response techniques. Although only the respondents know which question, either sensitive or unrelated, they answer, by dividing the whole sample into two parts, we propose to estimate the summary statistics based on the method of moment. This approach is applicable to mixed response type data, namely having dichotomous and Likert sensitive responses in the survey. With the method of moment estimates, we calculate the conditional mean of continuous sensitive responses given a discrete response and the partial correlations among continuous sensitive responses. As a medical application, we study the dependence structure among the responses of a survey concerning health and pressure on college students by our proposed approach.