Persistence of Educational Differences in

Objectives. Concerns about SES disparities in smoking and prospects that current anti-tobacco policies will reduce disparities suggest the need to examine recent trends. This study describes changes in smoking prevalence by education level for race and gender groups from 1976 to 2005. Methods. The analysis uses a measure of self-reported current smoking gathered in 23 National Health Interview Surveys. Logistic regression models test for differences in smoking changes across four education groups (0-11, 12, 13-15, and 16+ years of schooling) and use quadratic coefficients for year to generate predicted values and smoothed trends in smoking for each education group (with race-gender groups both combined and treated separately). Results. With controls for age, race, and sex, the results for the combined sample reveal that the gap in smoking between the lowest and highest education group declined by 4 percentage points from 1998 to 2005. However, this decline stemmed largely from trends among Hispanic men and women; in contrast, smoking disparities among white and African-American men and women show no evidence of narrowing. Further, smoking of high school graduates and those attending some college relative to college graduates widened rather than narrowed in recent years. Conclusions. Although educational disparities in smoking are no longer widening as they did decades ago, their persistence for most race-gender groups suggests that policies and programs more directly target low SES smokers.