A study on alcohol advertising and underage drinking by associate professor Leslie Snyder of the University of Connecticut’s Communications Department is seriously flawed in its methodology, findings and conclusions, according to leading experts.
Critiques published by Drs. Reginald Smart and Don Schultz, world renowned academic experts who each have more than 30 years experience analyzing advertising effects, detail a few of the major problems with the Snyder study, which appears in the Archives of Pediatric and Adolescent Medicine.
Both experts concluded that Dr. Snyder's study, touted by her and by advocacy groups as supposedly showing a causal link between alcohol ads and underage drinking, is scientifically flawed and contradicts decades of scientific research which have found no causal link.
In reality, the study’s data actually showed that seeing more alcohol ads was associated with less drinking.
Dr. Smart, a frequent consultant to the World Health Organization on alcohol-related problems and who spent more than 40 years with the Addiction Research Foundation and the University of Toronto, pointed out in his published critique that the study results show that those who saw the most advertising over time actually decreased their drinking, calling into question the entire hypothesis of the study. Importantly, Dr. Snyder did not refute this point in her rebuttal letter, which was also published in the Archives of Pediatrics & Adolescent Medicine, but instead resorted to personal attacks against the authors who critiqued her study.
Dr. Smart also noted major problems with people dropping out of the study, which had only 31 percent of the original sample staying in until the end. He noted that while the study is called "longitudinal," over two-thirds of the sample was not followed through the four interview periods.
The alcohol research expert also pointed out that the measurement of alcohol use in the study is not a standard one stating, "The reliability and validity of the measure have not been assessed and must be suspect." Further he stated, "There is no reference to experimental, economic, and advertising-ban studies, predominantly showing negative effects on alcohol consumption. This suggests a biased, 1-sided view of alcohol advertising effects."
Dr. Smart concluded, "The study has limitations that make it impossible to understand or interpret the results. It will not change thinking in the alcohol-advertising debate."
Dr. Don Schultz, a professor at Northwestern University who has authored 18 books and over 100 research papers on advertising and marketing, challenged the study in his critique to the Journal stating that it is inconsistent with "decades of very sophisticated advertising, marketing communication, and consumer behavior research" and relies instead on a "Hierarchy of Effects model (circa 1961)....this conditioned-response, behaviorist model has been challenged for years, most recently by new studies in cognitive science."
Dr. Schultz also cited the startling attrition rate stating, "Starting with 1872 respondents and ending with 588 respondents results in a massive attrition rate (68 percent). One can only wonder at the representativeness of the final group compared with the original sample."
Further, he commented, "Assuming that certain levels of marketplace advertising expenditures result in certain levels of in-market advertising media weight; that then result in certain levels of message distribution; that then result in certain levels of consumer exposure; that then result in actual consumer behavioral changes is tenuous at best." Dr. Schultz concluded that, "These linkages simply cannot be made and no amount of statistical magic can make them appear."
Dr. Schultz emphasized that the authors ignore the essential distinction between mere correlation and actual causation.
Decades of research by governments, health agencies and universities do not show that alcohol causes someone to drink. This study is no exception.