An awareness of factors influencing youth drinking outcomes is important for informed, evidence-based public policy. However, identifying models and strategies that advance responsible drinking or youth abstinence is only possible if literature reviews are comprehensive, systematic and accurate. Due to omissions and flaws, Casswell's [1] review of the research literature on alcohol marketing policy fails to deliver on its promised aims.

Casswell ([1], p. 479) begins by discussing the ‘received wisdom’ that early econometric research on advertising expenditures and alcohol consumption found ‘little evidence of effects’ (citing Smart's 1988 article [2]). She attributes this to an ‘early methodological emphasis’ on aggregate data, but this ignores recent reviews reaching the same conclusion over a wider range of methodologies [3, 4]. Casswell adds that ‘this line of research remains active’, and cites Nelson & Young [5]. However, our paper does not examine advertising expenditures, nor does it discuss the relationship with alcohol consumption. The abstract clearly states the modeling effort: ‘we study the effects of youth readership, price of advertisements, and audience size on alcohol advertising in 35 major magazines’ ([5], p. 551). Using a count data model, we concluded that ‘the results… fail to provide evidence that the alcohol industry is using magazine ad placement to target underage youth’ ([5], p. 555; see also [6]).

Casswell compounds this error by citing a brief critique of econometric studies by Saffer & Dave [7], but their critique ignored studies using quarterly or monthly data that contradict their claim of suppressed variation in expenditure data [4]. More important is Casswell's ([1], p. 479) claim that ‘investigation of naturally occurring regional differences in exposure… have achieved a major shift in the consensus’ on effects of advertising bans, citing Saffer & Dave [8] and Saffer & Chaloupka [9]. The first study ignores other alcohol policies, so the effect of advertising bans is statistically biased [10]. The second study ignores endogeneity of advertising bans and finds a sample that yields the desired result through the time-honored process of data-mining [11, 12].

An alternative approach to evaluation of marketing impacts is longitudinal surveys of young people and marketing practices, which includes studies of drinking beliefs, advertising awareness, marketing exposure and subsequent drinking outcomes (onset, frequency, bingeing). This literature has been reviewed previously [13, 14]. However, these reviews and Casswell's summary are seriously incomplete, as they ignore inconsistencies in the longitudinal results, including numerous null results [15] and publication biases [16, 17]. For example, my systematic review of 20 longitudinal studies, including 12 studies with comparable results, reports that ‘only 21 of 63 estimates (33%) are statistically significant at the standard 95% confidence level’ ([15], p. 885). In my meta-analysis of longitudinal studies, I found that ‘the empirical results are consistent with publication bias, omitted variable bias in some studies, and lack of a genuine effect, especially for mass media’ ([16], p. 191). I also demonstrate ‘reporting bias’ and ‘dissemination bias’ in the research and policy literatures.

A final comment deals with Casswell's policy proposals. The general thrust of her suggestions are either complete bans of alcohol advertising in all media ([1], p. 483) or severely restricted marketing of alcohol modeled after France's Loi Evin policy ([1], p. 480). Cross-national advertising bans are ineffective [10, 18]. Although Casswell mentions the Loi Evin several times, she also allows that, except for one expert opinion, ‘there are no other evaluation data available’ ([1], p. 480). It is difficult to see how a call for global governance of alcohol marketing can be based on ineffective bans, an under-evaluated policy for one country, and a flawed literature review.

Declarations of interest

The author has served as a consultant to law firms representing companies in the alcohol industry, specifically SABMiller and Anheuser-Busch. This letter was initiated by the author and was independent of the author's consulting committee.

  • Jon P. Nelson

  • Department of Economics, Pennsylvania State University, University Park, PA 16802, USA. E-mail:


  1. Top of page
  2. References