Ben Radford & Rebecca Watson A Skeptical Debate: Eating Disorders, the Media, and Skepticism

Prior to reading this article, you must read the following, in order.  It will not take you long, and without doing so, none of which follows will make any sense.

1.       New TV Show Perpetuates Anorexia Myths, Ben Radford

2.       Does ‘Black Swan’ Encourage Anorexia, Ben Radford

3.        Eating Disorders, The Media, and Skepticism, Rebecca Watson’s criticism of 1 & 2 

4.       Mass Media, Eating Disorders, and Research, Ben Radford’s reply to 3

Throughout the following article, these posts will be referenced in parentheses.

Introduction

Ok.  So, we have this back and forth between two great skeptics: Rebecca Watson and Ben Radford.  What is the topic in question?  The skeptics ponder the impact of media on female self image and related eating disorders.  What are the positions?

Ben Radford:

“Nearly every woman in America regularly sees thin women in everyday life and the media, yet according to the National Institute of Mental Health, only about one percent of them develop [sic] the disease. If there a strong link existed [sic] between media exposure and anorexia, we would expect to see an incidence many orders of magnitude higher than is found.”

Rebecca Watson:

“So, I’m forced to continue to side with what appears to be the consensus opinion: the media’s portrayal of the thin ideal most likely negatively impacts the body image of those who process those images poorly.”

I have chosen the above quotes because they reflect the positions in the debate.  Ben Radford’s intent in writing the first two articles was not engage Rebecca Watson, as they preceded her criticisms. While there are many issues which can be tackled, here my focus is on the reason why Rebecca calls Ben out, not necessarily to tease out all the nuances of the debate or to review all the research which informs the debate.  Rebecca calls Ben out for how he uses the research, not the research he uses to support his thesis.

Research

                Rebecca outlines where she believes Ben quoted out of context.  We are at a disadvantage: two of the three articles are behind pay walls.  Unless we are willing to pay, we cannot read them.  Fortunately, though, Ben does not refute Rebecca’s appraisal of the research he cited to support his view “Eating disorders such as anorexia nervosa and bulimia are biological diseases, not voluntary behaviors.”(1)  In his response to Rebecca, he argues his use of the research was to make general statements about

“Rebecca consulted several of the references I provided in my short columns and states that, ‘If you care to read it, you will experience the same realization I did: this paper comes to the exact opposite conclusion Radford claims.’ This misunderstanding seems to lie at the heart of the matter, and I figured out what the problem is so I will address it at length. Rebecca is assuming that the quotes were selected as representing the conclusions of those particular studies from which they were cited. I made no such claim.

Rebecca is finding contradictions where none exist. We are both correct and accurate in our quotes, but we are talking about two different things. I was addressing the larger question of whether there’s good overall evidence that media images cause or bring about eating disorders. On this, Botta agrees with me that (as of 1999), the answer is no. The same is true for the Posavac article that she cites; the claim that (as of 1998) ‘experimental studies have not shown that exposure to media images increases women’s weight concerns’ is completely accurate, not taken out of context, and (like the Botta piece) is quoted merely as an accurate and concise summary of the literature up until that point from dozens of studies—not specifically reflecting either Botta or Posavac’s work.”

Accordingly, we can assume because he has no issue with Rebecca’s summary of the research, her summary is – at least as far as Ben is concerned – accurate.   Thus, we can examine the back and forth of our skeptics with at least some confidence in the findings of the research Ben has quoted without having to have read it ourselves.  This is certainly, by no means ideal, but because the antagonists agree, our examination can go forward.

This brings us to the admission on Ben’s part that he and Rebecca are using the research in different ways.  He writes he is using the introduction and the articles cited within the research to make the general point that the research into the relationship between media images and eating disorders is not causal.  And, true to his word, when he quotes Heidi Posavac’s research (the one article which is free to view),Ben takes it (1) directly from the Abstract:

“Despite the popular belief that the thin standard of female attractiveness currently presented in the media is a primary contributor to the high level of concern with body weight among women‚ experimental studies have not shown that exposure to media images increases women’s weight concern.Three experiments are reported demonstrating that exposure to media images does often result in increased weight concern among women‚ but that body dissatisfaction‚ a stable personality characteristic‚ is a moderator of vulnerability to this effect. Although most women reported higher weight concern when exposed to media vs. neutral images‚ women with low initial body dissatisfaction did not. In addition ‚ this research suggests that negative effects on weight concern may result from even passive exposure to media images‚ but that exposure to realistic attractiveness is less likely to cause increased weight concern. The ethnicity of the participants in these studies reflected that of the local population ‚ with over 90% white. The nonwhite participants primarily belonged to one of the following groups; Asian‚ Pacific Islander‚ Latino.”

The only other statement similar to this is found later in Posavac’s introduction where, as Rebecca suggests, she uses the failure of previous experimenters to find the link as a jumping-off point for her own research, ”The distinction between women with high vs. low trait body satisfaction may help account for previous failures to find effects of exposure to media images on weight concern.”        

Certainly, then we cannot deny what Ben has argued; namely, that he referenced Botta and Posavac not for their own research, but for their summary of other research.  Yet, it must also be noted Ben does not make this admission in the original articles (1 &2).  In fact, he argues (1) that Posavac supports the view, “Concrete evidence is still necessary…to show that this existence of thin images and ideals has an effect on viewers….At this point, the relationship between television exposure and body image disturbance remains in question.”  If one returns to the abstract above – specifically to the line following the highlighted, underlined sentence fragment Ben chose to quote – one sees that Posavac’s research does not support this view. 

Is this a big deal?  After all, Ben did say he was simply using their research to support his position that the balance of research failed to find a causal link.  Right?

Well, wrong.  Ben makes the point that science is littered with studies which run contrary to the bulk of the research.  (4)  While this is true, it in no way necessarily invalidates that contrary research.  If a new piece of research was methodologically matched (using the same methodology, with the same variables, etc.) and came up with wildly dissenting results, Ben would be correct to be skeptical.  However, if we look at the scope and the complexity of the Botta as Rebecca outlines it, and the nuanced, mediating variables of Posavac’s, it becomes clear that the researchers were doing something different.  When we look to their conclusions, we do not see a radical departure from consensus, but a more complex, nuanced understanding of the relationship of media images and body image disturbance (Botta) and weight concern (Posavac).  Interestingly, the Posavac research involved experimental manipulation, showing causality.  So, I ask again, is it reasonable for Ben to ignore the research from which he has taken his citations when that research may break with those citations; especially when a causal link – albeit a mediated causal link – is supported by the research?  No, it is suspect.

Confirmation Bias: The Skeptic's Bane

Ben has done exactly what Rebecca accused him of doing with elipses (…); but on a much larger scale.  Ben has completely ignored the contrary research.  Ben has completely de-contextualized the citations.  By being lazy, using someone else’s research to find the references he did not take the time to find, he shot himself in the foot.  I do not believe Ben did this purposefully; I believe Ben relied on these researchers to find his confirming evidence.  This is a trap we must try to avoid, at all costs, as skeptics.  This is a trap Ben is well aware of, yet by which he was victimized: the tendency to look for confirmation of our beliefs. 

Indeed, if Ben simply wanted a study which was nothing more than a meta-analysis of the research that is the study for which he should have looked.  Ben did do this, in his response to Rebecca.  Ironically, in his response (4) he provided a study from 2009 which reviews the literature and research.  Sadly, he once again, ignores the contrary evidence, in favor of that which supports his point- of-view.  This is what he chose to quote, “currently, engagement with mass media is probably best considered a variable risk factor that might well be later shown to be a causal risk factor.”  Guess where Ben took this from?  Yes, the Abstract.  However, if we look a little deeper, to the end of the paper, we find a more robust, nuanced conclusion.  The authors begin their Conclusions and Recommendations section with:

Are Mass Media a Causal Risk Factor? Based on the weight of the empirical evidence in relation to criteria 1, 2, 5, and 7, a defensible conclusion is that the content, use, and experience of mass media—in and of themselves, and in the context of synergistic messages from peers, parents, and coaches—renders them a possible causal risk factor. Further, although the strength of the relationship is relatively weak to modest, the extent of exposure to mass media in general and to thinness-depicting media (Harrison & Cantor, 1997) in particular is correlated concurrently and prospectively with negative body image and disordered eating (criteria 3 and 4). In fact, because the hypothesized paths of influence between media exposure (engagement) and the outcomes of interest are mediated and moderated by several variables, we should probably expect effect sizes that are modest at best. Stronger evidence (criterion 5) for a causal relationship between mass media, negative body image, and disordered eating comes from the extensive experimental literature (Grabe et al., 2008; Groesz et al., 2002).

 

Before I conclude, I would like to make one more point on as far as what I think is Ben’s failure to recognize his own confirmation bias.  In 2007 Ben wrote a very detailed article which we can also read.  Ben lays out his position very well.  He supports his position with many, many references.  I leave it to you to judge by his words whether his current posts differ much from his article of a few years ago,

“I do not claim that all or most people are entirely satisfied with their weight or appearance. It is natural—and beneficial—to be dissatisfied with ourselves in some ways, and most of us are well aware of our imperfections. Yet studies show that while we recognize those flaws, most of us do not continually obsess over them, and certainly not to the point of endangering our health. While the body of evidence indicating significant genetic (e.g., Grice et al,. 2002) and biochemical (e.g., Frank et al., 2005) etiologies for eating disorders has grown more robust, evidence for the sociocultural model—which has dominated the discussion for decades—remains essentially stagnant. Researchers have spent 20 years studying the links between body image, “ideal” media images, and eating disorders. Yet the conclusions remain tentative and heavily qualified; the one thing that researchers agree upon is that more research is necessary. Perhaps what is truly needed is a reexamination of the premises underlying previous studies. Do thin fashion models, Barbie dolls, and Playboy models truly represent some sort of agreed-upon female ideal, as is often claimed? Do girls and women actually view fashion dolls and models as achievable, realistic body shape ideals? If the media influences are as pernicious and pervasive as often stated, why is the incidence of anorexia so low (about 1%), and why do most women and girls report being satisfied with their appearance? Body image, self-esteem, media influences, and eating disorders are important social, medical, and mental health issues. Tragically, because of political activism, poor journalism, and a failure of researchers to examine underlying premises, discussion of these issues has become muddled with misinformation. Researchers and the news media must do a better job of separating evidence-grounded, research-based facts from theory and speculation.”

The take-away is that Ben wrote 1 & 2 above after having fully committed to this position years ago.  Today he appears to be seeing evidence supporting his position on a subject which is important to him, as evidenced by the number of times he has returned to that subject: the relationship between media images and eating disorders and related phenomena.

Conclusion

Ben’s position may very well turn out to be correct.  The difficultly in conducting psychological research is a very real issue.  Unlike physics, chemistry, and even biology, psychology is extremely messy.  There are a great number of variables, most of which we may overlook.  Correlations between two or three variables are often very weak, suggesting multi-variable interactions.  This includes sociocultural variables as well as bio-chemical, neurological, and genetic variables.  No one of these variables alone completely explain the relationship between body image, ideal body image, and eating disorders.  Indeed, as Rebecca's dissection of the research shows, in psychology the variables are not easily divided into causal and effect.  Often we find our dependent variables can flip and become effective under a different experimental paradigm.  Progress in psychology is convoluted, and slow.  Twenty or thirty years, forty or fifty years…?  Looking for causes in psychology is...well...like looking for causes in psychology; there's nothing quite like it in the rest of science.  Sometimes it feels like we are reaching for the horizon.

Nevertheless, as we add in more variables from more research the answer will begin to emerge.  Are some avenues of study more plausible than others?  Of course they are.  However, it really is not for Ben Radford, Rebecca Watson, myself, or any layperson to say.  The researchers are the ones who need to make the decision where the research is headed, apply for grants, and further the research in light of years of study and ongoing research.  If you are not steeped in research, you are likely to misunderstand the research protocols, or confuse research designed to tease out details about body image with that designed to address correlates of a specific eating disorder, like anorexia.  Indeed, this is a trap Ben does fall into.  The following is taken from the comments section after his rebuttal of Rebecca. (4)

Comments:

#1 badrescher (Guest) on Tuesday December 28, 2010 at 3:21pm

“Did Botta find evidence in the study I quoted of some link between eating disorders and the media? Yes, and in fact I stated that quite clearly in the Facebook discussion.”

Actually, she didn’t, Ben. I’m working on a post on the topic myself. I’m not sure that I need to now, but perhaps I will finish it anyway.

The Botta study you quoted doesn’t measure eating disorders. The study is about body image. Yes, body image is directly linked to disordered eating, but the causal nature and, if there is one, the direction of cause is not established at all.

A direct link of media to eating disorders is certainly not established and the evidence suggests that such a link is unlikely. The complexity of this topic is such that answers to one specific question cannot possibly be used to support another, different specific question.

#2 Ben Radford on Tuesday December 28, 2010 at 3:28pm

Yes, you are quite correct. In my rush to respond I fell into the familiar trap of lumping together eating disorders, body image, thin ideals, and the like. When I’m writing for a knowledgeable audience I’m much more specific, but you’re correct. Thanks!

 

Unfortunately, Ben's response, "When I'm writing for a knowledgeable audience, I'm much more specific" does not excuse the sloppiness of his thinking and writing.  I charge that the euphemism "specific" is a transparent attempt  hide the appropriate adjective accurate.  Who deserves accuracy, those who will call Ben on his writing (Rebecca, the actual researchers, those who Ben calls knowledgeable) or the people who trust what you write, and do not research beyond what you have written?  Probably neither is more deserving, we all deserve the most objective, accurate information. 

Certainly skeptics can comment, but when we do, we must be willing to take the time to consult the research fairly, with an open, skeptical mind.  When returning to the research, we must recognize we have not been working in the field since the last time we posted an article.  That means we cannot trust our memory, but must reacquaint ourselves with the research.  In truth, we probably need to re-read our own work on the topic...our memories are just that malleable.  Some areas of research are much less ambiguous, of course.  That makes our job that much easier, and less prone to error.  Some, though, like the current topic, are much more complicated than they might first appear.  We need to suppress our beliefs and prejudices and look at each piece of research for what it is, and then in light of previous research.

In the foregoing exchange Rebecca does this, while Ben falls short.  Rebecca focuses on the papers in question, drawing her conclusions from the research.  Undoubtedly she had preconceptions, otherwise she would not have questioned Ben's posts in the first place.  The difference, Rebecca was able to focus on the research, setting aside the larger question.  If we scroll up to their quotes I presented at the outset, we can see the difference in approach.  Nowhere in her criticism of Ben does Rebecca state a thesis.  Whatever her position, she reserves that position.  She goes as far as to refer to herself as a "layperson;" recognizing, as any self-aware skeptic should, her own limited depth of subject knowledge.

In contrast, Ben's quote is a thesis statement.  Actually, it could be a hypothesis.  It is certainly testable.  As with Rebecca's, I chose this quote because it is reflective of Ben's position, and the tone of the entire first (1) article.  It was important to choose this article, because it was the article on which Rebecca focused in her criticism (3), and it reflects Ben's approach prior to reflection.  It must be said, Rebecca was always in the critics role in this back and forth.  Ben, though, did not get this opportunity until later.  This is important, because, Ben does eventually come to a very different conclusion in his rebuttal (4) than he did in his earlier articles.

So, as I conclude, I offer these final two quotations.  The first is Ben's conclusion from his first article. (1)  The second is from his response (4) to Rebecca.

Anorexia is a tragic disease; some young women (and men) do diet to excess and have body image issues. But the scientific research shows that they are the exception, not the rule. The first step in solving a problem is correctly understanding it, and TV shows like “The Price of Beauty” may actually end up doing more harm than good.

Since research suggests that the causes of anorexia have more to do with genetics than thin fashion models, efforts to educate young girls about the artificiality of airbrushed media images won’t do anything to treat or cure anorexia. Girls and young women deserve facts and truth instead of myths and misinformation.

Contrast this with:

Just to summarize, there is indeed evidence both for and against the claim that the mass media play a role in causing eating disorders. Like everything in science, some studies show an effect but others don’t. Neither I nor anyone else would claim that it is impossible that mass media might somehow cause (or bring about) eating disorders. The real question is this: After decades of research and hundreds of peer-reviewed studies, has science found a strong link between images of thin women in the media and the onset of eating disorders? The answer has remained the same for thirty years: possibly, but we don’t have good evidence. 

As the commenter quoted above points out, Ben is still confusing research into causes of eating disorders and the science investigating the media's influence on body image, nevertheless, Ben has become much more subtle and nuanced in his thinking.  Ben sounds much more like the skeptic he is known to be.