生物识别与生物统计学杂志

Joint Generation of Binary and Nonnormal Continuous Data

Abstract

Hakan Demirtas

The use of joint models that are capable of handling different data types is becoming increasingly popular in biomedical practice. Evaluation of various statistical techniques that have been developed for mixed data in simulated environments requires concurrent generation of multiple variables. In this article, I comprehensively evaluate the unified framework proposed by Demirtas et al. for simultaneously generating binary and nonnormal continuous data given the marginal characteristics and correlation structure. I conduct this assessment in three simulated settings with synthetic bivariate and multivariate data as well as real depression score data from psychiatric research. Considering close agreement between the specified and empirically computed quantities on average, as measured by some key bias- and precision-related quantities, the methodology appears to have prospects to address the need of generating intensive data that have binary and continuous parts for simulation purposes.