Impact of Lifestyles and Social Attitudes on Happiness Based on CGSS Modular Data—A Bayesian Approach

摘要: A Bayesian forward procedure, based on the predictors common to the final models obtained from the Bayesian BRPN and FBRPN variable screening procedures, is proposed to identify the most efficient models containing important and meaningful protective and risk factors for happiness across the lifestyle, social attitude, and socio-demographic modules of the CGSS data. For the combined modules, in descending order of importance, the protective factors include perceived fairness of society, trustworthiness of most people, participating in physical exercises, ideal number of children, relaxation, socializing, gathering with relatives, TV usage, total individual income, female gender, browsing the Internet, watching TV or DVDs, equal sharing of household chores, lifelong learning, application for Party membership, political affiliation, overnight stays, education level, listening to music at home, shopping, magazine usage, and tolerance of premarital sex; the risk factors include acceptance of extramarital sex, being taken advantage of if not vigilant, prioritizing layoffs of females, prioritizing marriage over career, and assuming men surpass women in capability. The relationships among the RVs of the modules are discussed, concepts of ‘endogenous’ and ‘exogenous’ related to variable selection are introduced, and the origins of the Bayesian–Frequentist debate are traced to support the application of the proposed procedures.