STATS ARTICLES 2012
Media Binging on Reports of Women Drinking
Rebecca Goldin, Ph.D., December 3, 2013
In a period when rates of alcohol use and abuse are relatively stable, the media has raced to report on claims that binge drinking among women is on a dangerous trajectory.
The trigger for this claim is a new book by the Canadian journalist and recovering alcoholic, Ann Dowsett Johnston, “Drink: The Intimate Relationship Between Women and Alcohol.” As part of the promotion for her book, she published a commentary last month in The Wall Street Journal—“The New Face of Risky Drinking is Female”—which claimed that binge drinking was increasing among young women, and that according to the Centers for Disease Control (CDC), it was a “serious under-recognized problem.”
But—as an editor’s note in the Journal partly explained—Johnston had misreported the survey data, combining two separate studies to come up with a propulsive but erroneous claim. Moreover, the idea that binge drinking among women is under-recognized is belied by a decade’s worth of headlines in the media all warning about the problem. There is nothing new about the phenomenon—the media have been telling us about it year after year.
This time, it wasn’t only Johnston who was promoting the idea of a boom in lady-binging. ABC’s Good Morning America, in a feature tied to the book, reported that “binging is up among all age groups,” with a link to the CDC ‘s fact sheet on the problem. But no where on the fact sheet does it actually say or suggest that “binging is up among all age groups,” and that’s because the CDC’s data doesn’t show such a trend—nor do the data from numerous surveys that track American drinking trends. Ironically, ABC News had to issue a correction last January—the last time there was a media surge in reports on women binge drinking—for making the very same claim about the CDC data.
If journalists can't even read the news put out by their own news organizations on the topic they're reporting on, or read what they are linking to as evidence, what hope has the public of understanding the data behind any social problem - particularly one where there are multiple sources of data trying to measure the same thing?
Comparing Survey Results
A basic tenet of Statistics 101 is that one should never directly compare survey results that come from different populations or which were provided from surveys taken with different methods, even if the surveys are angling for the same kind of data.
To illustrate the issue of bad comparisons, the 2011 National Survey on Drug Use and Health (NSDUH) found that 39.1 percent of full-time in college students aged 18-22 were binge drinkers (drinking 5 or more drinks in a sitting within the past month), whereas the 2011 Core Alcohol and Drug Survey reported that 44.8 percent of full-time students were bing drinkers (drinking 5 or more drinks in a sitting with the past two weeks.) Presumably the difference in reported binge rates--for the same populations the same year--would be even larger if both were measuring the same time period.
What might account for the difference? Surveys can return different results for a variety of reasons, from the time of year in which they are administered, to the population from which they choose participants, and even to how the questions are asked. The Core Alcohol and Drug Survey uses participating colleges (and the colleges administer the surveys), whereas NSDUH is an annual nationwide survey involving interviews with randomly selected individuals.
Johnston simply compared rates from these two surveys (in different years) to establish her claim (made in the original version of the Wall Street Journal--and subsequently redacted) that “The percentage of college students who binge drink—using the measure of five drinks in two hours for men, four for women—has risen in the past decade, nearing 45 percent.” She noted that the lower rate reported in NSDUH ‘s 1999 survey, to a high estimate of drinking in the Core survey, obtained in 2011. The magical increase resulted.
The problem with making such comparisons is that both of these surveys have systematic biases that come from their survey methods, which in turn makes direct comparisons problematic. It is reasonable to try to ascertain the robustness of trends using different surveys; for example, we can examine the trend of binging across the years in one specific survey and to see whether binging is increasing (in each survey). If the pattern is the same across several surveys, one can conclude that the results are more robust than if different surveys suggest different patterns.
The interest in whether the problem of binging is worsening or not certainly begs the question: are there other data that suggest that binging has a new female face?
The female face of high school drinking
First, let’s look at high school binging. The CDC does a biannual survey by gender of high school students. Though not yet graduated to college, these data are a broader pulse on abuse patterns among young people. According to the Youth Risk Behavioral Survey (YRBS), girls’ binging rates jump up and down by a few percentage points year to year, but have seen a modest reduction. In 2003, over 34 percent of female high school seniors reported binging in the previous month; this went down to 29.2 percent in 2005, and then 27 percent in 2011.(1*) The rates for 2013 are, as yet, unavailable, Thus the YRBS survey does not support the hypothesis that binging is increasing among girls attending high school.
Historical data from the NSDUH also fail to establish an increase in binging for women. In 2003, 16.5 percent of women aged 12-20 admitted to binging. The figure hovered around sixteen percent, and then trended downward to 14 percent in 2012. Again, this survey did not find a trend that women were increasing their binging in the past ten years.
Yet another survey to look at high school binge drinking is Monitoring the Future, a long-running survey run out of the University of Michigan. This study found that among 12th grade women, 21.2 percent reported binging in the previous month in 2003; by 2011, the figure was 19.7 percent with fluctuations of 1-2 percent each year. Yet again, the survey does not support the hypothesis that binging is up among women.
The female face of college drinking
Of course it is possible that college promotes a culture vastly different than high school. Perhaps the fact that binge drinking rates in high schools have been stable or slightly decreased over the past ten years is not reflected in college drinking culture. But yet again, the standard surveys give no hint of an increase. According to Monitoring the Future, 33.1 percent of women binged in the previous month in their 2003 survey. The percentage fluctuated by a couple percentage points each year, arriving at 34.5 percent in 2012. The percentage of women abusing alcohol this way was essential constant. In contrast, among women who were college age but didn’t attend college, there was a marked decrease in binging, from 33.1 percent in 2003 to 26.2 percent in 2012. Similarly, the National Survey on Drug Use and Health found that college students experienced an overall a decrease in binging over the past ten years. According to the 2012 NSDUH survey, “In 2012, male full-time college students aged 18 to 22 were more likely than their female counterparts to be binge drinkers (45.5 vs. 35.3 percent).” The rate in 2011 was slightly lower overall (not statistically significant) but it was not broken into men versus women.
The observation that binge drinking is not worsening among young women does not detract from the seriousness of the problem of excessive and binge drinking among college-aged women. Chronic alcohol abuse continues to affect both men and women, and consistently more men than women. And while 40 percent of students are getting drunk, evidence suggests about 10 percent of students will find themselves dependent on alcohol. Addressing the problem of underage, irresponsible, or unhealthy drinking is a worthwhile endeavor that universities and colleges across the country need to undertake. However, scare-mongering headlines suggesting that females are suffering worse than males or that the female trajectory is one that will invariably land them to rehab or to death are a misleading use of statistics.
1* Among male teens, there was also a minor reduction in binging, from 39.5% of 12th graders in 2003 (and 36.2% in 2005) to 35.7% in 2011.
Rebecca Goldin, Ph.D., Director of Research, December 3, 2013.