Cantrell et al. Respond

Abstract

We thank Sheehan for her comments. We are aware of the known correlation between racial segregation and neighborhood-level socioeconomic status (SES). However, the high correlation between these two measures can make it challenging to implement the analyses strategies recommended. A strong relationship between the two potential predictors can make their independent effects unidentifiable in regression analyses that include both.1 The ability to detect moderation effects is also limited because sparse data are often found in cross-tabulated categories (i.e., high African American and high SES). These issues are not always adequately addressed in published neighborhood studies. The 2010 American Community Survey census data demonstrated significant correlations between measures of block group-level SES (income, unemployment, education) and percentage of African Americans in the Washington, DC communities where outlets are based. Correlations ranged from 0.6 to 0.9, making it nearly impossible to separately evaluate whether point-of-sale (POS) tobacco marketing strategies were targeted based on SES status or the racial/ethnic composition of communities. Further, Short Form 1 census data, which were the data available for this study, did not include measures of SES at the block group.2 Finally, growing evidence suggests that cigar use is high among African American youths3,4 and adults.5 Accordingly, we focused our analysis on block group racial/ethnic demographic characteristics and did not include highly correlated SES measures. The previous literature on POS marketing has found both neighborhood-level minority composition and low SES to be associated with tobacco advertising.6–9 Yet again, many of these analyses were unable to unpack the separate influence of the two factors. This is a limitation of our study and similar studies in areas where community race/ethnicity and SES are highly correlated. Nonetheless, these patterns clearly indicate higher availability and advertising of little cigars and cigarillos (LCCs) in areas with high concentrations of African Americans, which also tend to be low SES communities. We appreciate Sheehan’s discussion of additional community structural characteristics. As she correctly notes, the lower availability of prevention programs and services found in African American communities may influence tobacco use. However, we obtained data and conducted analyses only on POS tobacco marketing and census demographics; thus, we refrained from making policy recommendations based on data that were not part of the study. We agree that neighborhood-level structural factors are likely critical in understanding LCC initiation, addiction, and cessation.10 We encourage researchers and policymakers to consider these factors when developing interventions and policies related to LCC use.