We analysed the 1972–2006 cumulative General Social Surveys (GSS) data,7 which allows an examination of health status/behaviours and political ideology at a micro level. Health status was ascertained from a question, ‘Would you say your own health, in general, is excellent, good, fair, or poor?’ For analysis, we created a binary variable with poor health = 1, 0 otherwise. Smoking status was measured based on an affirmative response to the following question, ‘Do you smoke?’. Of those reported, 5% were considered as in poor health and 35% were smokers in the weighted, pooled samples. Political ideology was based on a question, ‘Generally speaking, do you usually think of yourself as a republican, democrat, Independent, or what?’, with respondents’ choosing one of the following categories: strong democrat (15%), not very strong democrat (22%), independent near democrat (12%), independent (15%), independent near republican (9%), not very strong republican (16%), strong republican (10%) and other party (1%). For ease of presentation and interpretation, we also grouped these into democrats (49%) comprising strong democrat/not very strong democrat/independent near democrat, republicans (35%) comprising strong republican/not very strong republican/independent near republican, and independent/others (16%). We included age, sex, race, marital status, religious service attendance, highest educational degree, total family income in the last year prior to the survey and survey year as covariates. After excluding missing values on the responses and predictors, the final analytic sample consisted of 32 716 individuals for the poor-health model and 14 803 individuals for the smoking model. In the GSS, the self-rated health question was asked over 23 survey years whereas smoking was asked for 13 survey years, which accounts for the large differences in the sample sizes. We used weighted, binary logistic regression model procedures as implemented in SAS v9.1.
In fully adjusted models, using democrats as a reference, the odds ratio (OR) for reporting poor health for republicans was 0.74 [95% confidence interval (CI) 0.65, 0.85] (Figure 1a). Similarly, republicans were less likely to be smokers (OR 0.85, 95% CI 0.78, 0.92) compared with democrats (Figure 1b). The ORs for reporting poor health or smoking for independent/others were not different from democrats.
Bookmarks