Tag Archives: methods/use of data

Just for Fun: The Folly of Small Sample Sizes

1 (3)

Saturday Morning Breakfast Cereal, by Zach Weiner.

Lisa Wade is a professor of sociology at Occidental College and the co-author of Gender: Ideas, Interactions, Institutions. You can follow her on Twitter and Facebook.

Sunday Fun: Correlation or Causation?

correlation

By xkcd.

Originally posted in 2009.

Lisa Wade is a professor of sociology at Occidental College and the co-author of Gender: Ideas, Interactions, Institutions. You can follow her on Twitter and Facebook.

Tuskegee Syphilis Study Recruitment Letter

Flashback Friday.

The Tuskegee Syphilis Experiment is one of the most famous examples of unethical research. The study, funded by the federal government from 1932-1972, looked at the effects of untreated syphilis. In order to do this, a number of Black men in Alabama who had syphilis were misinformed about their illness. They were told they had “bad blood” (which was sometimes a euphemism for syphilis, though not always) and that the government was offering special free treatments for the condition. Here is an example of a letter sent out to the men to recruit them for more examinations:

The “special free treatment” was, in fact, nothing of the sort. The researchers conducted various examinations, including spinal taps, not to treat syphilis but just to see what its effects were. In fact, by the 1950s it was well established that a shot of penicillin would fully cure early-stage syphilis. Not only were the men not offered this life-saving treatment, the researchers conspired to be sure they didn’t find out about it, getting local doctors to agree that if any of the study subjects came in they wouldn’t tell them they had syphilis or that a cure was available.

The abusive nature of this study is obvious (letting men die slow deaths that could have been easily prevented, just for the sake of scientific curiosity) and shows the ways that racism can influence researchers’ evaluations of what is acceptable risk and whose lives matter. The Tuskegee experiment was a major cause for the emergence of human subjects protection requirements and oversight of federally-funded research once the study was exposed in the early 1970s. Some scholars argue that knowledge of the Tuskegee study increased African Americans’ distrust of the medical community, a suspicion that lingers to this day.

In 1997 President Clinton officially apologized for the experiment.

Originally posted in 2009.

Gwen Sharp is an associate professor of sociology at Nevada State College. You can follow her on Twitter at @gwensharpnv.

Sunday Fun: Statistically Significant Others

boyfriend2By xkcd.

Lisa Wade is a professor of sociology at Occidental College and the co-author of Gender: Ideas, Interactions, Institutions. You can follow her on Twitter and Facebook.

On Women’s Comfort with Topless Sunbathing

What should we make of changes in fashion? Are they the visible outward expression of new ways of thinking? Or do fashions themselves influence our sentiments and ideas? Or are fashions merely superficial and without any deeper meaning except that of being fashionable?

It’s summer, and once again magazines and newspapers are reporting on beachwear trends in France, proclaiming “the end of topless.” They said the same thing five years ago.

1 (2) - Copy

As in 2009, no systematic observers were actually counting the covered and uncovered chests on the beach. Instead, we are again relying on surveys – what people say they do, or have done, or would do.  Elle cites an Ipsos survey: “In 2013, 93% of French women say that they wear a top, and 35% find it ‘unthinkable’ to uncover their chest in public.”

Let’s assume that people’s impressions and the media stories are accurate and that fewer French women are going topless. Some of stories mention health concerns, but most are hunting for grander meanings. The Elle cover suggests that the change encompasses issues like liberty, intimacy, and modesty.  Marie-Claire says,

Et en dehors de cette question sanitaire, comment expliquer le recul du monokini : nouvelle pudeur ou perte des convictions féministes du départ ?

But aside from the question of health, how to explain the retreat from the monokini: a new modesty or a loss of the original feminist convictions? [my translation, perhaps inaccurate]

The assumption here is that is that ideas influence swimwear choices.  Women these days have different attitudes, feelings, and ideologies, so they choose apparel more compatible with those ideas.  The notion certainly fits with the evidence on cultural differences, such as those between France and the U.S.

1 (2) 1 (3) - Copy

Americans are much more likely to feel uncomfortable at a topless beach. But they are also much less likely to have been to one. (Northern Europeans – those from the Scandinavian countries and Germany – are even more likely than the French to have gone topless.) (Data are from a 2013 Harris survey done for Expedia.)

This second graph could also support the other way of thinking about the relation between fashion and ideas: exposing your body changes how you think about bodies.  If people take off their clothes, they’ll become more comfortable with nudity. That is, whatever a woman’s original motivation, once she did try going topless, she would develop ideas that made sense of the experiences, especially since the body already carries such a heavy symbolism. She would not have to invent these topless-is-OK ideas all by herself. They would be available in the conversations of others. So unless her experiences were negative, these new ideas would add to and reinforce the thoughts that led to the original behavior.

This process is much like the general scenario Howie Becker outlines for deviance.

Instead of deviant motives leading to deviant behavior, it is the other way around; the deviant behavior in time produces the deviant motivation.  Vague impulses and desires … probably most frequently a curiosity … are transformed into definite patterns of action through social interpretation of a physical experience. [Outsiders, p. 42]

With swimwear, another motive besides “vague impulses” comes into play:  fashion –  the pressure to wear something that’s within the range of what others on the beach are wearing.

Becker was writing about deviance.  But when the behavior is not illegal and not all that deviant, when you can see lots of people doing it in public, the supportive interpretations will be easy to come by.  In any case, it seems that the learned motivation stays learned.  The fin-du-topless stories,  both in 2009 and 2014, suggest that the change is one of generations rather than a change in attitudes.  Older women have largely kept their ideas about toplessness. And if it’s true that French women don’t get fat, maybe they’ve even kept their old monokinis.  It’s the younger French women who are keeping their tops on. But I would be reluctant to leap from that one fashion trend to a picture of an entire generation as more sexually conservative.

Jay Livingston is the chair of the Sociology Department at Montclair State University. You can follow him at Montclair SocioBlog or on Twitter.

Sunday Fun: Confirmation Bias for Everyone!

1By David Malki at Wondermark.  H/t to @annettecboehm.

Lisa Wade is a professor of sociology at Occidental College and the co-author of Gender: Ideas, Interactions, Institutions. You can follow her on Twitter and Facebook.

Overweight Americans Have the Lowest Risk of Premature Death

Last year the Journal of the American Medical Association released a study aiming to determine the relationship between body mass index and the risk of premature death. Body mass index, or BMI, is the ratio between your height and weight. According to the National Institutes of Health, you are “normal weight” if your ratio is between 18.5-24.9.  Everything over that is “overweight” or “obese” and everything under is “underweight.”

This study was a meta-analysis, which is an analysis of a collection of existing studies that systematically measures the sum of our knowledge.  In this case, the authors analyzed 97 studies that included a combined 2.88 million individuals and over 270,000 deaths.  They found that overweight individuals had a lower risk of premature death than so-called normal weight individuals and there was no relationship between being somewhat obese and the rate of early death. Only among people in the high range of obesity was there a correlation between their weight and a higher risk of premature death.

Here’s what it looked like.

This is two columns of studies plotted according to the hazard ratio they reported for people.  This comparison is between people who are “overweight” (BMI = 25-29.9) and people who are “normal weight” (BMI = 18.5-24.9).  Studies that fall below the line marked 1.0 found a lower rate of premature death and studies above the line found a higher rate.

25-29

Just by eyeballing it, you can confirm that there is not a strong correlation between weight and premature death, at least in this population. When the scientists ran statistical analyses, the math showed that there is a statistically significant relationship between being “overweight” and a lower risk of death.

Here’s the same data, but comparing the risk of premature death among people who are “normal weight” (BMI = 18.5-24.9) and people who are somewhat “obese” (BMI = 30-34.9).  Again, eyeballing the results suggest that there’s not much correlation and, in fact, statistical analysis found none.

30-34.9

Finally, here are the results comparing “normal weight” (BMI = 18.5-24.9) and people who are quite “obese” (BMI = 35 or higher). In this case, we do see a relationship between risk of premature death in body weight.

35

It’s almost funny that the National Institutes of Health use the word normal when talking about BMI. It’s certainly not the norm – the average BMI in the U.S. falls slightly into the “overweight” category (26.6 for adult men and 25.5 for adult women) — and it’s not related to health. It’s clearly simply normative. It’s related to a socially constructed physical ideal that has little relationship to what physicians and public health advocates are supposed to be concerned with.  Normal is judgmental, but if they changed the word to healthy, they have to entirely rejigger their prescriptions.

So, do we even have an obesity epidemic? Perhaps not if we use health as a marker instead of some arbitrary decision to hate fat.  Paul Campos, covering this story for the New York Times, points out:

If the government were to redefine normal weight as one that does not increase the risk of death, then about 130 million of the 165 million American adults currently categorized as overweight and obese would be re-categorized as normal weight instead.

That’s 79%.

It’s worth saying again: if we are measuring by the risk of premature death, then 79% of the people we currently shame for being overweight or obese would be recategorized as perfectly fine. Ideal, even. Pleased to be plump, let’s say, knowing that a body that is a happy balance of soft and strong is the kind of body that will carry them through a lifetime.

Cross-posted at Adios Barbie.

Lisa Wade is a professor of sociology at Occidental College and the co-author of Gender: Ideas, Interactions, Institutions. You can follow her on Twitter and Facebook.

U.S. Jobs Are Back, but They’re No Match for Population Growth

Last week CNN triumphantly reported that the job market has recovered to its 2008 peak.  Here’s the headline:

1

Not so fast, though.

Sociologist Philip Cohen observes that the real news is hidden in the fourth paragraph. There the author of the piece acknowledges that the job data are numbers, not proportions.  The numbers have bounced back but, because of the addition of almost 12 million people to the U.S. population, the percent of Americans who have jobs or are in school remains lower than it was in 2008.

From CNN:

Given population growth over the last four years, the economy still needs more jobs to truly return to a healthy place. How many more? A whopping 7 million, calculates Heidi Shierholz, an economist with the Economic Policy Institute.

Using the Bureau of Labor Statistics, Cohen offers us a clearer look at where we’re at:

1a

Lisa Wade is a professor of sociology at Occidental College and the co-author of Gender: Ideas, Interactions, Institutions. You can follow her on Twitter and Facebook.