Background The Framework Convention on Tobacco Control calls for the elimination of tobacco advertising, promotion and sponsorship. were no significant gender or ethnicity differences. Conclusions Tobacco packaging uses logos, colours and imagery to create desirable connotations that promote and reinforce smoking. By functioning in the same way as advertising, on-pack branding breaches Article 13 of the FCTC and refutes tobacco companies claims that pack livery serves only as an indentifying device that simplifies smokers decision-making. Given this evidence, signatories should see plain packaging policies as a priority consistent with their FCTC obligations to eliminate all tobacco advertising and promotion. and Three of these brands C (mid-price tailor made cigarettes), (lower price tailor made cigarettes) and (low price loose tobacco) C are available in New Zealand and represent brands with high ((which is available in New Zealand), respondents were unlikely to have encountered 470-37-1 manufacture the remaining four brands, which were US tobacco brands with varied penetration levels in markets outside New Zealand. The MMP2 brand, as its name suggests, is largely generic and has few brand elements; it functioned as a control relative to the other clearly branded packs. Respondents used 470-37-1 manufacture 15 adjectives selected to correspond to various brand personality dimensions  and that we had pre-tested in earlier studies assessing tobacco product positioning [9,30]. These included: young, mature, masculine, feminine, tough, cool, professional, classy, popular, plain, budget, traditional, relaxing, sophisticated, and trendy. Respondents were asked to associate as many or as few of these 15 attributes with each of the seven brands, depending on their perception of the brand concerned. The order of presentation of both the brands and the attributes was randomised to avoid question-order and item-order effects. Figure ?Figure11 outlines the question used and contains examples of the brand stimuli and attributes. Figure 1 Examples of Test Stimuli and Attributes. Analysis All analyses were undertaken using PASW(18). We initially used Principal Components Analysis to examine each brands underlying attributes and then ANOVA to test differences between gender, ethnicity and smoking status. Although we detected some differences for gender and ethnicity, these were not systematic. The only consistent differences occurred between non-smokers and smokers responses; the results section thus focuses on these groups, while noting differences by other variables where relevant. Results Most individual respondents selected between one and four attributes per brand; the mean number ranged from 1.7 to 2.5 and the median for all seven brands was two attributes. On average, each of the 15 attributes was associated with a particular brand 144 times, with a range between 0 and 705 associations. The first hypothesis posited that each tobacco brand would communicate different attributes to young adults. To test this hypothesis, we first examined the brand, which we used as a control. Given its generic appearance and name, it is not surprising that virtually the only attributes associated with were plain and budget. For this reason, we excluded from further analysis and discuss its evaluation separately. We then factor-analysed the brand descriptors for the remaining six brands. These analyses produced between three and five significant factors (Eigen values greater than 1.00) for each brand. Table ?Table11 contains an example of the factor analysis results for a familiar brand (as primarily traditional and mature, popular and relaxing, and masculine; very few regarded it as a sophisticated brand, though neither did they consider it budget or plain. Analyses by ethnicity showed that male Mori and Pacific respondents saw the brand as less traditional/mature than did other groups, particularly other male smokers, and female Mori and Pacific smokers. Most participants regarded as a primarily plain and budget brand, an association that reflects its lower price point. However, more than a third also saw it as professional and mature, and a quarter considered it relaxing, and popular and trendy. Mori and Pacific were more likely to make these latter associations than non-Mori and Pacific. Participants also saw as popular and traditional (particularly by women), and budget (again, a reflection of the brands cheap position). Participants linked fewer features with 9, a brandname variant not really 470-37-1 manufacture bought from New Zealand. Nevertheless, many viewed it being a youthful brand, apt to be marketed at a lesser price, and much more likely to become targeted at females than men. Respondents noticed as an ordinary or spending budget brand mainly, but linked it with trendy and great qualities also, and noticed it as missing a particular gender appeal. From the three new brands tested, acquired solid organizations with maturity and professionalism and reliability, and was regarded as a traditional, old brand. General, these results support our initial hypothesis: respondents linked different qualities with.