Academic journal article Contemporary Economic Policy

Reply to Siegel et Al.: Alcohol Advertising in Magazines and Disproportionate Exposure

Academic journal article Contemporary Economic Policy

Reply to Siegel et Al.: Alcohol Advertising in Magazines and Disproportionate Exposure

Article excerpt

I. INTRODUCTION

Siegel et al. (this issue) offer four general comments on my previous article (Nelson, 2006), which dealt with count data analysis of alcohol advertising in a sample of 28 magazines and possible "targeting" of underage youth. First, they report a technical problem in my data, which included overlapping estimates of 18-19 yr olds in the "youth" and "adult" audience variables. This was an oversight on my part that I have corrected, but the changes do not alter my results or conclusions. Second, they claim that correlation between two of my variables, percent youth and adult median age, creates a multicollinearity problem0 in my empirical results. However, their main evidence for this claim is a simple correlation coefficient. This is insufficient in the context of a multivariate problem, and I report a comprehensive statistic, the variance inflation factor (VIF), that refutes their claim. Third, Siegel et al. provide an alternative set of regression estimates that substitutes a new variable for young adult readership in place of my regressor (adult median age). However, they limit their analysis to the proportion of readers between the ages of 21 and 34 yr, which also is correlated with youth readership. The implicit claim is that their empirical model applies equally to magazines with diverse audience demographics, such as Rolling Stone and Popular Mechanics. I provide evidence that their empirical results are fragile, which supports my conclusion that targeting of underage youth is not occurring. Fourth, Siegel et al. claim that their results and other studies that they cite support a policy that would impose a rule of "disproportionate exposure" for alcohol advertising. Incredibly, Siegel et al. cite several public health studies that show that magazine advertisements have null or negative effects on youth drinking behaviors. In this reply, I will demonstrate that their results and citations do not support advertising policies based on a simple rule of "disproportionate exposure."

II. READERSHIP ESTIMATES AND AUDIENCE DEMOGRAPHICS

The print media industry distinguishes between a magazine's circulation and its audience or readership. Circulation is defined as the total number of copies sold through all channels of distribution (subscriptions, newsstands, bulk). Circulation numbers are relatively "hard" data and, among other things, are used in the form of minimum circulation guarantees for published advertising rates (Standard Rate and Data Service, 2003). In contrast, audience or readership estimates are based on surveys conducted by research organizations, such as Mediamark Research, Inc. (MRI). In MRI's adult survey, respondents (aged 18 yr and older) are shown a large set of magazine logos and asked if they have read or looked into any of the magazines within the last publication cycle (e.g., weekly magazines within the past 7 d). The survey is conducted annually using a combination of in-home interviews and self-administered questionnaires, and the sample size is about 26,000. MRI expands the sample responses to obtain population-weighted readership estimates for each age cohort. However, sampling error can lead to a disclaimer that an estimate is unreliable. My use of a composite average age variable reduces sampling error, but focusing on a narrow age group is less likely to have this effect (e.g., number or percent of readers aged 21-34 yr as used by Siegel et al.).

Some media experts claim that readership numbers are largely a marketing myth and the number of true adult readers is unknown (Dobkin, 2007; Media Dynamics, 2003). Survey-based estimates for youth readers will be even less reliable. First, MRI uses a self-administered mail survey for teens (aged 12-19 yr) that contains a simple listing of magazines. This is a less sophisticated procedure than the visual recall method used for adults (Mediamark Research, Inc., 2007). Second, in 2004, the teen sample size was 4,419 (Mediamark Research, Inc. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.