SUMMARY. This study examined factors that influence illegal tobacco sales to minors, such as buyer, clerk, and store characteristics. Thirty-seven youths made 314 attempts to purchase tobacco in 11 towns in Illinois. Purchase attempts were made from over-the-counter and vending machine vendors. Multilevel multivariate logistic regression analyses, which controlled for town clustering effects, were run to determine predictors of illegal sales to minors. Findings revealed that the strongest pre 64dictors of selling cigarettes illegally to minors were clerks' failure to ask a minor for age or identification. The implications of using multivariate methods to identify factors influencing illegal tobacco sales are discussed. [Article copies available for a fee from The Haworth Document Delivery Service: 1-800-HAWORTH. E-mail address: <getinfo@haworthpressinc. com> Website: <https://www.HaworthPress.com" xmlns:xlink="https://www.w3.org/1999/xlink">https://www.HaworthPress.com> © 2002 by The Haworth Press, Inc. All rights reserved.]

Recent estimates indicate that each day 5,500 children try smoking for the first time, and almost 3,000 more become established smokers (Gilpin, Choi, Berry & Pierce, 1999). In the United States, first use of a tobacco product almost always occurs before high school graduation (Centers for Disease Control [CDC], 1994). Easy availability of tobacco products among youngsters might contribute to alarmingly high rates of tobacco use (Rhodes & Jason, 1988). Despite laws prohibiting the sale of tobacco products to persons under the age of 18, most minors have little difficulty purchasing cigarettes (Dason et al., 1999; Jason, Biglan, & Katz, 1998). It is unclear which minor, clerk, or store characteristics might influence youths' ability to purchase tobacco products illegally.

Previous research examining illegal tobacco sales to minors has suggested that a minor's ability to purchase tobacco may be influenced by specific buyer characteristics, such as the minor's age and/or gender. DiFranza, Savageau and Aisquith (1996), for example, found that youth who appeared to be 16 and 17 years of age were more successful in illegally purchasing tobacco than youth who appeared to be 11 to 15 years of age. O'Grady, Asbridge and Abernathy (1999) found that older minors were the most successful and the younger minors were least successful in illegally purchasing tobacco. Clark, Natanblut, Schmitt, Wolters and Iachan (2000) also found that the success rate in purchasing cigarettes increased with each increasing year of age. However, one study by Arday, Klevens, Nelson, Huang, Giovino and Mowrey (1997) did not find a relationship between apparent age and sales outcome.

When gender has been examined as a predictor of illegal sales, the literature suggests that girls are more frequently sold cigarettes. Arday et al. (1997), DiFranza et al. (1996), and Clark et al. (2000) found that girls were more successful than boys during the purchase attempts. When 65two minors conducted a purchase attempt together, O'Grady et al. (1999) found that mixed dyads (males and females) were the most successful when attempting to illegally purchase tobacco followed by female dyads, with male dyads being the least successful.

When exploring what type of clerk behavior is a predictor of sales outcome, asking a minor for age or identification substantially reduces illegal sales. Landrine, Klonoff, and Alcaraz (1997) found that when age or ID was requested, illegal sales were significantly less likely to occur. If minors were questioned by clerks about their age during a tobacco purchase attempt, 96% of the time they were refused cigarettes, and if the clerk requested identification from the minor, 99% of the time the tobacco sale was refused. Clark et al. (2000) also found a significant relationship between a clerk's request for proof of age and refusal of sales. Minors were refused illegal sales 90% of the time when asked for proof of age.

Several studies have examined the relationship between gender and being questioned about age or identification. Arday et al. (1997) noted that female buyers were less likely to be questioned by clerks during a tobacco purchase attempt. DiFranza et al. (1996), however, found that boys and girls were equally likely to be asked for proof of age. Landrine et al. (1996) also did not find gender differences with respect to questioning a minor during a purchase attempt. Only one researcher, to date, has examined the relationship between a field agent's ethnicity and being asked about age or identification. Landrine et al. (1996) found that African American youth were asked their age and ID significantly more than white youth.

Several studies have also examined whether a successful purchase attempt may be associated with clerk characteristics, such as the clerks' apparent age and/or gender. DiFranza et al. (1997) found that younger clerks (estimated to be under 21 years of age) were more likely to sell to minors than older clerks (estimated to be over 21 years of age). However, Arday et al. (1997) did not find the estimated clerks' age to be related to a successful sales outcome. Forster, Wolfson, Murray, Wagenaar and Claxton (1997) noted that male clerks had a tendency to sell more frequently to minors than female clerks. Contrary to those findings, Clark et al. (2000) found that female clerks were more likely to sell to underage youth. However, Arday et al. (1997) did not find the clerks' gender to be related to a successful sales outcome.

Along with individual characteristics that pertain to purchase success, researchers have examined whether store characteristics play a role in determining sales outcome. Forster et al. (1997) found that gas 66stations and convenience stores were most likely to sell cigarettes to youth followed by grocery stores and restaurants. Clark et al. (2000) found that gas stations had the highest sales rates, but convenient stores that did not sell gas had the lowest sales rates. However, Arday et al. (1997) found no relationship between type of store and successful sales outcome when looking only at stores with over the counter sales. O'Grady et al. (1999) also did not find a relationship between type of store and purchase success rates.

When examining the type of tobacco product access, Forster et al. (1997) found that when tobacco products are locked up or behind a counter accessible only to the clerk, they are less likely to sell tobacco to minors. In addition, youth tend to have more success purchasing tobacco products from vending machines that are unlocked than from vendors with over the counter sales (DiFranza et al., 1996).

Other variables that may relate to sales outcome are tobacco advertising and tobacco sales promotions (Wakefield et al., 2000). Several studies have evaluated whether vendor-warning signs relating to the laws prohibiting the sale of tobacco to minors may have some influence on sales outcome. DiFranza et al. (1996) found that stores that participated in "It's the Law" programs, which involved posting vendor-warning signs were somewhat more likely to request proof of age from minors attempting to illegally purchase tobacco, but the differences in questioning did not result in a reduction in illegal sales. Arday et al. (1997) found that warning signs had no effect on vendors' compliance with the state minors' access law. Although clerks in stores with warning signs posted were somewhat more likely to ask questions of the minor during the purchase attempt, stores with warning signs were actually associated with a higher likelihood of selling to minors. Forster et al. (1997) indicated that the presence of tobacco-industry or other signs that made a reference to tobacco sales law were not associated with purchase sucess.

Finally, the presence or absence of store customers at the time of the purchase attempt may play a role in determining sales outcome. Forster et al. (1997) examined the environment of the store at the time of the purchase attempt and found that the presence of one or more customers in line behind a youth buyer positively predicted purchase success.

Clearly, the literature reviewed above is mixed concerning the role of minor, clerk, and store characteristics that might influence illegal sales of cigarettes. Many of the studies reviewed above examined certain aspects of these characteristics, but rarely have the comprehensive array of various youth, clerk, and store characteristics been examined simul67taneously using multifactorial analyses. It is possible that the discrepant findings reviewed above are due to past studies focusing on selective variables (i.e., minor characteristics only, clerk characteristics only, or store characteristics only). The purpose of this study was to examine multiple youth, clerk, and store influences that might affect illegal tobacco sales to minors.