
Zippo: Hand Warmer
Point-of-purchase tobacco environments

Field observers collected information on the presence of various POP tobacco environment features including
1) tobacco product placement (self-service versus clerk-assisted);
2) promotions (multipack/cents-off discounts and gift-with-purchase offers);
3) tobacco-branded functional objects (free items provided by the industry such as counter change mats or shopping baskets displaying the sponsoring company's logo);
4) presence and extent of exterior and interior advertisements (such as those indicating special prices);
5) presence of low-height interior advertising or advertisements directly in the line of sight of very young children (at a height of <3>
6) tobacco-control signage (including Food and Drug Administration [FDA] (ss)-or industry-sponsored signage, health warnings, or messages indicating that identification is required to purchase tobacco products).
Data were weighted to account for both the community-level multistage sampling procedures and the store selection proba bilities.
The GENMOD procedure in SAS v.8 was used to run generalized estimating equations that accounted for community clustering while specifying a binomial distribution and a logit link function. Results were expressed as unadjusted odds ratios and 95% Wald confidence intervals. For all analyses, supermarkets were used as the referent category.

Exterior tobacco advertisements were observed in 58.9% of the stores, with 40.4% having high levels of exterior advertising (i.e., five or more advertisements or at least one advertisement >1 foot in any dimension).
Some form of tobacco advertising (interior or exterior) was present in 84.1% of the stores. Convenience/gas retailers were significantly more likely to have five of the six POP measures, convenience and liquor stores were significantly more likely to have four of the measures, and drug stores were significantly less likely to have two of the measures