Investigating social network dynamics underlying ethnic inequalities in smoking among adolescents
Author(s): Márta Radó , Dorottya Kisfalusi, Anthony Laverty, Frank J. van Lenthe, Jasper V. Been, Károly Takács
Friday 16 | 11:40-12:00
Room: TP51
Session: Health inequalities
Background:
Despite the decreasing overall smoking rate, ethnic inequalities in smoking persist, underlying the importance of target social network intervention.1,2 This, however, requires a good understanding of the underlying social network mechanisms which is currently lacking.3,4 There is also a complete lack of evidence on whether negative networks, such as antipathies, contribute to the spread of smoking.3 Our study assessed whether (1) adolescents tended to form friendship or antipathy relationships primarily with peers of similar or different ethnicity, (2) smoking spread among friends and antipathies, and (3) the number of friendship and antipathy nominations of non-smokers differed between ethnic groups.
Methods:
We applied Stochastic Actor Oriented Models to the complete friendship network data of 263 adolescents in 9 classes with 2670 friendships and complete antipathy networks of 226 adolescents in 8 classes with 1234 antipathies over two waves. Data was collected between 2010 and 2011 in Hungarian secondary schools. We considered an adolescent as a Roma ethnic minority if this adolescent reported being Roma in any waves.
Findings:
Adolescents were most likely to nominate as friends (odds ratio [OR] = 1.45, 95% Confidence Interval [CI] = 1.08–1.94) and less likely to nominate as an antipathy (OR = 0.63, 95% CI 0.48–0.83) same-ethnic peers over cross-ethnic peers. Neither friends (OR = 3.56, 95% CI 0.18–70.05) nor antipathies (OR = 1.97, 95% CI 0.02–172.22) influenced each other's smoking behaviors. But smokers were more likely to receive friendship nominations than non-smokers (OR=1.21; 95% CI=1.01–1.44) and smokers and non-smokers did not differ in antipathy nominations (OR=1.17; 95% CI=0.98–1.40). Non-Roma adolescents were more likely to nominate as a friend a smoker non-Roma (OR=1.20; 95% CI=1.06–1.35) and less likely to nominate as a friend a non-smoker Rome (OR=0.66; 95% CI=0.51–0.87). Roma smokers did not nominate friends based on smoking status or ethnicity.
Conclusion:
We identified possible social network mechanisms potentially creating ethnic smoking inequalities. Adolescents tended to nominate as friends with same-ethnic peers, and smoking spread in friendship networks due attractiveness of smoking for friendship selection. Smoking was more attractive for friendship nominations for Roma adolescents. Non-Roma adolescents were particularly likely to nominate smoker Non-Romas as friends and exclude non-smoker Romas. Antipathy social networks did not contribute to the spread of smoking, therefore, they did not contribute to smoking inequalities.