Retail Data Analytics Is Key To Plus Size Fashion Success
Maybe you’ve heard that the average American woman is a size 14. Or size 12-14. Or size 16. Or something between sizes 16, 18 and 20W. Or none of the above. Why is it so hard to answer this seemingly simple question?
University of Washington researchers Deborah A. Christel and Susan C. Dunn of Washington State University explained some of the complexities of determining the clothing size of the average American woman in their new paper. There’s no public source of detailed body measurements, for one thing. Data collected for study of public health, for example, does not include all the measurements necessary to make garments that fit.
If you’ve seen headlines about women’s sizes in the news over the past couple of weeks, they probably referred to Christel and Dunn’s work. Tim Gunn, of Project Runway fame, mentioned it in an op-ed for the Washington Post, and many media outlets have picked up the story in one form or another. Some of the reports say the average size is now a 16, others between a 16 and 18. Christel and Dunn themselves concluded that the average size is something between a 16, 18 and 20W.
Something between a 16, 18 and 20W? Using the best currently available public data, waist measurements from the Centers for Disease Control and Prevention, Christel and Dunn compared the average waist measurement to ASTM standards for Misses size women’s clothing. They found that the average woman’s waist measurement (based on data from 2007-2010) of 95.2 centimeters (about 37”) would be something between a Misses size 16 and 18. The ASTM does not have a current standard for Plus size women’s clothing, so Christel and Dunn used the expired standard to conclude that the same waist measurement corresponds to a Plus size 20W.
So, let’s say that you’re a woman with a waist that measures exactly 95.2 centimeters. What size should you wear, a size 16, 18, or 20W? Maybe none of the above.
There’s the matter of shape. This just in: women come in different shapes.
Research by Zhiping Huang of the Harvard School of Public Health and others showed with data what we already know from observation: women's shapes vary. Among a sample of 840 women (nurses participating in a long term health study), 132 had waist to hip ratios of .73 or less, and 188 had waist to hip ratios of .84 or greater. A 34 inch waist with a waist to hip ratio of .73 would mean hips of about 47”, while a ratio of .84 would mean a hip measurement of about 40”.
That’s quite a wide range. Even with the same waist measurement, clothes that fit a woman with 40” hips don’t fit a woman with 47” hips. Yet more than 35 percent of the women in the study had waist to hip ratios at or beyond the .73-.84 range. And the variations for women with small waists may be different than those for women with larger waists. So it’s hard to define what’s “average”, as far as women’s clothing size goes.
Sizing standards don’t help. The most common sizing convention, Misses’ sizes, was designed for women who are adults, but have not given birth or changed shape with age. But the median age of an American female is somewhere in the late thirties, and that includes the roughly one in five females who are not yet adults. It’s more realistic to say that the “average” adult woman is a mother in her early forties.
That woman would, in theory, be better served by sizing known as Plus (or Women’s) sizes. But, as Christel and Dunn pointed out, there’s no current Plus size standard. What does that say about the women’s fashion industry?
Industry standards for clothing sizes are voluntary. Nobody is obligated to use them, and many manufacturers don’t. The same woman may find the best fit carries a very different size label from one garment to the next. This is true even when an industry standard is available. And lack of an industry standard doesn’t stop manufacturers who make Plus size clothing. They use the expired standard, or their own standards.
The result is so much variation in sizing that we couldn’t fully identify an average woman’s clothing size, even if we had a precise set of the average woman’s body measurements. That’s why, as a statistician, I can’t accept any estimate of an average American woman’s size. An average waist does not equate to an average woman, and neither the fashion industry nor the research community has yet developed data adequate to define an average, let alone a full range of sizes that properly meets the needs of adult women.
Media focuses on the growing size of American women and the lack of nice clothes to fit average and larger women, but misses a key conclusion of the research. None of the press coverage that I’ve seen mentioned that Christel and Dunn called for the development of new standards based on measurements made by body scanning. Suitable 3-D scanning devices are now available and widely used, yet there is no comprehensive and current source of data suitable for developing clothing size standards.
The analytics community and fashion industry should step up to serve American women by developing realistic sizing conventions and keeping them up to date. It could be done with a tape measure and some effort, but new technology makes measurement and analysis easier and more accurate than ever before. This isn’t a charitable effort, it’s a matter of survival. Do it right, or the customers will either go to a competitor or keep the money in their wallets.
I corresponded with Deborah Christel about the challenges of defining an average size. Here’s what she told me:
I understand that all research has its limitations. Time and page space in journals also provides limitations in research and reports. My co-author, Susan Dunn, and I would love to further examine our nation’s average clothing size as the data we used is 4 years old and I am sure there have been changes since then. A fully comprehensive study that takes into consideration all combinations of measurements is ideal, but those combinations vary by company and is proprietary information within each company. We wanted to use publicly available data that was collected in a reliable scientific manner. Also, every clothing company has its own sizing chart and we recognize that we are making a generalization based on two very specific data sets. We used the measurement that made the most sense to us and that was agreed as acceptable by peer reviewers in our discipline.
Meta S. Brown is author of Data Mining for Dummies and creator of the Storytelling for Data Analysts and Storytelling for Tech workshops. http://www.metabrown.com.