Smoking behavior prevalence in one's personal social network and peer's popularity: A population-based study of middle-aged adults in Japan.

Although previous social network studies have consistently shown the social influence of peers' smoking on one's (ego's) smoking, few studies have examined how the influence differs according to peers' structural positions in the network. Investigations are also lacking on whether vulnerability to the influence varies by ego's socioeconomic position. Thus, the present study aimed to examine how the association between peers' smoking and ego's smoking differs by peers' popularity in ego's personal network and ego's educational attainment. We used data from the third-wave Japanese Study on Stratification, Health, Income, and Neighborhood (J-SHINE) conducted in 2017, which targeted middle-aged (32-58-year-old) residents in four municipalities within Japanese metropolitan areas. Information on four close peers' characteristics and behaviors and their mutual relationships was collected by the name generator and name interpreter methods. Data on 1989 respondents and 7956 peers were evaluated. Peers' eigenvector centrality was used as their popularity index in ego's personal network. We set ego's smoking as an outcome, regressed on each peer's smoking, each peer's popularity, and ego's educational attainment adjusting for ego's age, sex, working status, marital status, spouse's/partner's smoking status, as well as similarity in socioeconomic backgrounds between peer and ego, using a logistic regression model with robust standard errors. We then added a three-way interaction term for these three explanatory variables to the model. Results showed that peer's smoking status was related to ego's smoking even more strongly when the peer was popular but only in the case of ego with lower educational attainment. The results suggested that the disparity in smoking behavior across socioeconomic positions may be partly explained by susceptibility to social influence from one's personal network among the socioeconomically vulnerable. This study proposes a plausible method for pinpointing the peer influencer in one's personal social network to close the socioeconomic gap in smoking.