Cross-sectional personal network analysis of adult smoking in rural areas

Abstract

Research on smoking behaviour has primarily focused on adolescents, with less attention given to middle-aged and older adults in rural settings. This study examines the influence of personal networks and sociodemographic factors on smoking behaviour in a rural Romanian community. We analysed data from 76 participants, collected through face-to-face interviews, including smoking status (non-smokers, current and former smokers), social ties and demographic details. Multilevel regression models were used to predict smoking status. The results indicate that social networks are essential in shaping smoking habits. Current smokers were more likely to have smoking family members, reinforcing smoking within familial networks, while non-smokers were typically embedded in non-smoking environments. Gender and age patterns show that women were less likely to smoke, and older adults were more likely to have quit smoking. These findings suggest that targeted interventions should focus not only on individuals but also on their social networks. In rural areas, family-based approaches may be particularly effective due to the strong influence of familial ties. Additionally, encouraging connections with non-smokers and former smokers could help disrupt smoking clusters, supporting smoking cessation efforts.

Keywords: data science; human behaviour; network science; smoking.

Figure 1.

Link-tracing sampling. The black squares represent the initial individuals, referred to as ‘seeds’, who were invited by researchers to participate in the study. The circles depict individuals who were recommended by the study’s participants. Those marked in red indicate individuals who participated in the study, while the ones in blue signify individuals who did not participate for various reasons, such as refusal or the need to reschedule. The directional arrows within the figure demonstrate the flow of the referral process, tracing the path from referees to their referrals.