# Challenge: Advertised vs. actual waistline

September 30, 2010

### Topic

Contests, Statistical Visualization  /

Ever notice how pants seem to fit differently from store-to-store even though they’re labeled as the same size? Why does the 36-inch at Old Navy feel kind of loose but the same size at The Gap feels like you had too many fries at lunch? Here’s your answer from the Esquire Style blog. The actual size (from this über-scientific study, I am sure) tends to be bigger than the size as advertised. A 36-inch waistline actually means 41 inches in Old Navy units.

And just for kicks, here’s a simplified remake. I opted for a graphic sans belts and tape measures, but sorta less fun.

[Esquire via @demaws]

Update: It just occurred to me this would be a great challenge. You’ve seen my remake, now let’s see yours. Post your redos in the comments below by Friday 5pm PST, and I’ll randomly select a submission. Winner gets a free copy of The Visual Miscellaneum by David McCandless.

• Adjusted version is not just for kicks! It is good, sensible interpretation of the same data. Mean, Honest, Borderline dishonest is lousiest scale I have seen. Also using tape measure as a scale and putting ” numbers in to it… terrible. Not to speak about old navy appearing twice as long as they claim to be.

Now that simplified chart is the stuff.

• I think your graphic is actually more misleading. No ordinary human has a waist between 0 and 20 inches so by including that range you mislead more than the original graphic.

• Sorry Matt – what about Scarlett O’Hara’s 16 inch waist? Well – I guess she was extraordinary. ;)
(Good point about the challenge of choosing the appropriate range to depict.)

• One cannot choice a baseline when displaying a graphs that is judged by length. Any length must begin at zero. That is not true for judgments that are based on position such as line charts or dot plots.

• Actually, Scarlett O’Hara’s waistline was 18 and 1/2 inches…. Not 16.

• Except, the range in this case is not of pants sizes, but deviations from listed pants size (1 to 5 inches). The graphic should be shifted to the left, and Old Navy should be longer relative to H & M.

• Matt, I had the exact opposite reaction. The updated graphic has the correct proportions, and therefore it is less misleading. In the original graphic, the 36″ waist size of H&M is half the size of that of Old Navy. In reality, as well as in the updated graphic, H&M is 10% less than Old Navy.

This idea has been covered quite clearly by Tufte.

It’s very helpful to have the true 36″ indicator line in the updated graphic. Otherwise the graphic at first glance looks to be just a bunch of horizontal lines.

• Pretty tricky to check out your table if you’re not looking for the 36″ range.

• I would like to see a study of this for female clothing. The sizes are so all over the map that it makes it virtually impossible to shop for a specific-fit item without trying it on.

• For five years I have been the same weight but when I pick up a pair of pants the waist size has been smaller and smaller. Shirts too. I used to wear a large and now I medium, and sometimes small. So great data. The basic chart is clean and simple. With all the space you save you could add comparisons for other waist sizes too. Like 32!

• JFiedler

I wonder what the author thinks 36” waistline on pants means. It seems he thinks that the pants should have 36” of fabric at its waistline, when I’m fairly sure that is not at all what is meant. What I think is meant is that these pants are designed to fit a man with a 36” waistline. And if that’s the case then obviously the pants must have more than 36” to surround that waistline.

Now, we should expect the same amount of fabric in different pairs of 36” pants only if they all use the same thickness fabric and the same pattern. But of course neither is true. For instance, I know that Express pants never fit me right, but Calvin Klein and (loose-fitting) Banana Republic pants usually do, and it’s not because of the waistlines. The point is, we are not cylinders, and different ideas of how to fit a 36” waist will result in differing amounts of fabric. Was the author surprised to find that pants that are cut to “Leave yourself some wiggle room” have been cut to leave you some wiggle room?

I am agnostic about whether or not there really is widespread vanity sizing, but the author certainly hasn’t made his case. In fact, he doesn’t seem to have put much thought into any of what he’s saying, which makes his alarm seem silly.

“This is science, damnit.” Science? I don’t think so.

• Great comment.

• There is some truth to what the original author is saying and some truth to what JFiedler is saying.

It is true that pants measurements are designed to fit a true waist of 36″. Therefore if you buy jeans with a high rise and wear them at your true waist, the pant size should match your body measurement. If you buy low rise (and almost everything sold now is some sort of medium rise), the measurement is going to be larger in order to accomodate your ass and hips (which are going to be larger than your true waist).

It is also true that many brands (especially on the lower end like the ones shown) are known for “Vanity Sizing” where they specifically make a pair of jeans to fit someone with a 38″ waist but put a 36″ tag on it to make you feel better.

The end result is that there are some uncontrolled (potentially…maybe they very scientifically tried to choose only pants with 8″ rise so the waist is meant to be worn in the same place) variables that could be having a big effect on the data.

That being said, in my experiance, they are doing a reasonable job of showing vanity sizing…I can definitely imagine old navy pants fitting huge compared to H&M (just look at who the market to)

• Also in response to kcar1 below, a half inch to inch difference of vanity sizing is what we just like to call “quality control”

Sure, some brands are more precise than others (levis for instance are pretty random….make sure to try them on). Also, denim is not a very rigid weave (it is made out of fairly stretchy cotton with a twill weave) and as such it can stretch/shrink a lot. An inch of stretch could easily occur between taking your pants out of the dryer and when you throw them back in the wash (more if you bought them tight or don’t wear a belt).

Vanity sizing usually end up moving you at least a full standard (even numbered) size so you would go 36 to 34 or less.

• grnoonan

I was thinking along the same lines. The rise also matters, as the lower on your hips the pants sit, the larger the waist size needs to be, since, as you pointed out, we aren’t cylinders.
I like the graphic–if not completely kosher, it is very entertaining. And since entertainment is the point of the article, the graphic should be fun, too.

• I agree with JFiedler, I do not find this at all enlightening because it raises all sorts of questions about what is being compared. Honestly, compare the 36″ and 40″ versions of the same pants — this is not even remotely like the comparing a 6 and 8 for women’s clothing, it is like comparing an 8 and 18. Keep in mind what the effect of increasing by one inch in circumference has on the area of a circle.

So, while I will buy an half inch to inch for vanity sizing, I think the difference in the actual waistband measurement has much more to do with fit/cut than actual sizing — a jean cut to sit lower will necessarily have a wider waistband than a jean cut to sit at the natural waist on the same person.

• Seth Tucker

Just an opinion, but this visualization is not Flowing Data worthy. It belongs on the Junk Charts blog, perhaps. It’s interesting — something most clothing consumers can appreciate, yes. But, at the end of the day, it’s a simple, mediocre infographic. :)

• Okay, I’ll be honest – I had scheduled this for a Friday post, but realized late last night that I didn’t have anything for today. At least I supplemented it with a redo, and there’s some nice conversating going on. Thanks for the high expectations :).

• Seth Tucker

See, Nathan… this is what you get for having a damn good blog: high expectations. ;) Thanks for the reply, don’t let my whining make you take yourself too seriously, and keep up the good work.

I’ll be buying the book. The least I could do for all things I’ve come across here.

• That’s alright by me!

• Just upgraded this to a challenge. Winner gets a free copy of The Visual Miscellaneum. See details at the end of this post.

• yulich

Second one is much, much better. It’s simple and easy to read. But the first one… I spent couple of minutes before I’ve managed to understand what it says.

• Here is my attempt
http://i360.photobucket.com/albums/oo50/ottothecow/size.gif

It is kind of ugly as I didn’t have a ton of time or tools available where I am right now but it sort of accomplishes my vision.

Basically from the commentary above I realized that the number if inches can be misleading as to how a pair of pants would actually look or fit. For example..a meager 2-inch difference would only appear as a 1″ differance when you fold the pants flat and that doesn sound like that much (and it is how they suggest measuring jeans here: http://blueingreensoho.com/site/index.php?Itemid=41&id=30&option=com_content&task=view )

I thought it might be more interesting to see potentially what the waist of someone who fit these pants might look like so I plotted a series of basic ellipses. What this shows me is that while everybody except H&M vanity sizes and old navy vanity sizes a LOT, most of the manufacturers cluster in a similar sized group–I would bet that if you saw people wearing the non outlier pants, you wouldn’t really be able to tell their sizes apart.

• I did a few modifications(still on the low-tech side)
http://i360.photobucket.com/albums/oo50/ottothecow/size2.gif

I thought it was a little misleading to hide the difference by splitting it into both sides (though I am still ok with the original comparison).

With my new chart, you can see how many inches “wider” *(from a frontal view) a person would have to be to properly fit the pants. It ends up being about a 2″ width differance between a true 36 and the old navy pants.

• I’m going to defend the Nigel Holmesian infograpic from Esquire. I’ll actually remember the information from it, as opposed to the spare Tuftian graph.

• I agree. If anything, the Esquire graphic has a certain charm to it that my remake lacks.

• Ken Pierce

I’ll agree with Cody. I actually liked the Esquire graphic. However I think it might have been more interesting to code the x-axis with the percent of stated length over for each pair of pants. If stated this way the Old Navy pants have a 5x larger deviation than the H&M pants. Its a given their all at least 36″. The graphic is concerned with the deviation. Great blog.

• o has the right idea. We should remember that the familiar 36″ waist ‘size’ is a convenient linear measure of a property (circumference) that is only related to true waist size (area).

We can’t compare waist size circumference measures on a linear scale as they are proportional to radius whereas the waist area is proportional to radius squared.

For example, Old Navy’s 41 inch ‘Size 36’ represents 130% of the *area* of an accurately measured Size 36 rather than 114% as one might imagine from comparing the circumferences of 41″ and 36″.

I think a simple table showing the percentage variances on both circumference and area bases is the clearest illustartion:

http://public.myjungledisk.com/9F8ZBAXB23Z446ED/waist.png

• I think a trimmed version is more effective at pointing out the deviation from the norm than a chart that starts at 0; so here’s my attempt:

http://www.flickr.com/photos/54461398@N06/5041715674/

• Never mind… here’s one with the right labels on the axis:

http://www.flickr.com/photos/54461398@N06/5041116637/

• agtz The photo is private. flickr tells me I don’t have permission to view it.
(I did see the deleted one.)

• Go figure, it’s a Friday… I just made it public, so you should be able to see it. The only difference is that the axis now doesn’t begin at 36.

• A quick attempt: http://bit.ly/bLsT4f

• richard k
• Did someone win the book?

• Sorry, behind on this. Will pick a winner soon!

• I agree with o. Most pants don’t sit at the waist, so showing an oval cross-section at the waist with different pants surrounding it is misleading. A more useful infographic would be a frontal view of a man’s waist-hip area with a line showing the waist, and different lines showing where the different pants would sit. But then you have to find out the relation between waist and hip sizes, and it’s probably all over the map for men of different body types.

If you’re a skinny man with a 36″ waist, the 36.5″ pants might sit just above your hips, and the 41″ pants would fall off.

If you’re fat, the 36.5″ might sit almost at your waist, and the 41″ right at the hips.