Figure 1.

Kernel density plots illustrating the balance improvement obtained by matching the propensity scores for use of any pharmaceutical aid to aid a quit attempt and for use of buproprion, nicotine and varenicline among adults who made quit attempts in the Tobacco Use Supplement to the Current Population Survey, United States, 2002–2003 and 2010–2011. Each left-hand panel represents the full data sets, while the right-hand panels are the propensity scores for only the matched data. Nearest-neighbor matching was used to balance samples for every medication assessed. A 1:1 nearest-neighbor matching ratio was used for matching the any pharmaceutical aid and nicotine groups, and a 2:1 nearest-neighbor matching ratio was used for matching the bupropion and varenicline groups.