Shirt!
January 22, 2016 8:20 AM   Subscribe

Threadbase purchased 800 of the most popular Men's t-shirts in every size that they could get their hands on. They then measured the shirts, washed them repeatedly, and tracked their shrinkage/stretching over time. Notably, they observed that shirts get wider and shorter over time, but actually wearing the shirt undoes most of the shrinkage that happens in the wash. Also, sizing systems vary wildly across manufacturers.
posted by schmod (40 comments total) 43 users marked this as a favorite
 
I laughed so hard at this when it first posted - I appreciate their earnest research but my father and his buddies in the garment trade could have told them all of this decades ago, for the price of a martini (or maybe two). One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom...
posted by twsf at 8:28 AM on January 22, 2016 [36 favorites]


Interesting and useful!

I particularly like the bit about manufacturing tolerances. It would be great to have more data on that particular area, though I understand why it would be expensive to obtain.
posted by jacquilynne at 8:28 AM on January 22, 2016


This research deserved a Nobel-prize for T-Shirt Science.
posted by blue_beetle at 8:30 AM on January 22, 2016 [1 favorite]


blue_beetle: "This research deserved a Nobel-prize for T-Shirt Science."

What happened? Was it discredited? Did they forget to include a turtleneck control group?
posted by schmod at 8:33 AM on January 22, 2016 [3 favorites]


I obviosuly will nevr win the Nobel-prize for Spelling.
posted by blue_beetle at 8:35 AM on January 22, 2016


One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom...

There's a huge difference between subjective experience and controlled research.
posted by MisantropicPainforest at 8:39 AM on January 22, 2016 [31 favorites]


"I appreciate their earnest research but my father and his buddies in the garment trade could have told them all of this decades ago, for the price of a martini (or maybe two). One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom..."

So how exactly does one go about finding these experts and how does one verify that they are, in fact, experts?
posted by I-baLL at 8:41 AM on January 22, 2016 [11 favorites]


This research confirms my experience that Zara and Express shirts make me feel like a total heifer.
Interesting outlier on the t-shirt lengths: "long line" shirts that are in vogue among the man-bun set right now.
Thanks for this post! I think I might actually try some brands I'd never have considered based on the average measurements. I wish they would have tested variance across brands, but that would have required purchasing an immense amount of shirts. Anecdata: no two American Apparel shirts are the same.
posted by BuddhaInABucket at 8:44 AM on January 22, 2016 [2 favorites]


And I thought I was the one getting wider and shorter
posted by exogenous at 8:47 AM on January 22, 2016 [10 favorites]


I particularly like the bit about manufacturing tolerances. It would be great to have more data on that particular area

I used to work for a noted catalog retailer (who is on the list) and I know quite a lot about this.

One thing to remember is that for a lot of knitwear, things are batch cut. Instead of cutting around a single bit of fabric, imagine something more like a cookie cutter pressing down on a stack of 50 or 100 bits of fabric. As the cutter presses, the fabric moves around, and so what you end up with is the bits of fabric on the bottom of the stack being slightly larger (usually) than the ones at the top.

Also, a single garment may be cut and stitched in more than one factory, so despite best efforts the cutting at one factory may be marginally larger or smaller than spec - so you can end up with a 1/2 inch variance between the blue shirt and the white one if one factor is cutting 1/4 inch small and the other is cutting a 1/4 inch big.

And, of course, stitching plays just as much of a role - so if you're cutting 1/8 of an inch large, and stitching another 1/8 of an inch past the mark, you end up with a 1/4 variance, and so on.

At least at our retailer, every service rep has access to both the baseline product spec and the acceptable variance range for that spec. So, if you ask about a garment measurement on a shirt, we're able to give you fairly complete information. Unfortunately, most people out there have no idea what to do with that information, and you end up with people saying "Oh, my chest measures 40", but the large is spec'd at 46" in the chest - that will be way too big. Give me the medium that is spec'd at 40" -- not really thinking through that you generally don't want your garment measurement and your body measurement to be exactly the same, because you expect your clothing to hang away from your body.
posted by anastasiav at 8:52 AM on January 22, 2016 [24 favorites]


Interesting data, and the actual site is really useful... I wear extra long t-shirts and am always looking for the best bang for my buck, although it turns out that I was already buying the heaviest and almost least expensive shirts (Haynes Beefy T).
posted by Huck500 at 8:53 AM on January 22, 2016 [1 favorite]


I laughed so hard at this when it first posted - I appreciate their earnest research but my father and his buddies in the garment trade could have told them all of this decades ago, for the price of a martini (or maybe two). One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom...

posted by twsf at 11:28 AM on January 22


I'm right there with you to some extent. The utility of this though for me is as a consumer I now know two new brands of T-Shirts that might be long enough for my torso. Now as for the manufacturer, their sell-in story may certainly be "T-shirt Shrinkage doesn't really matter since the customer can stretch it appropriately" - and that may well be true and what was used to maintain distribution with certain retailers, but the democratization of this insight among the edge case switchers may force retailers to tailor their behavior (pun clearly intended).

As an approaching middle aged data junkie, I get that the old guard which doesn't wrap their heads around data may have had anecdotal knowledge and understanding of this reality, their ... pig headedness on this is a slight variation on the "F-ck you I got mine attitude" - unwilling to move the iceberg of their business towards legitimate technological changes in the products. This is why you saw this massive swing and development of something like Under Armor - they arrived late and built up their own specifications - it isn't an actual innovative push from a pre-existing brand. Net impact: existing brands of sportswear apparel manufacturers saw market share erosion - even though UA also brought in a metric ton of new/additional customers to the category.

So this old guard attitude of "I already know what the data says, we figured this out years ago" works for now - and it may work for the length of time the old guard are still employed for their company, but in 10-15 years if we all aren't prepared to be silk screening our own custom shirts on a nigh-daily basis I'd be somewhat surprised. And the reality is the old guard aren't looing at that end of the future... their future is best served by preventing a new company which

Now with that said, I don't find a garment research analytics company to be the best place to expect to make my claim (it'll be an uphill battle with retailers, manufacturers, and distributors), but this is a bit of a start. Sadly I see this information as about as useful to the industry as America's Test Kitchen is to ConAgra - any mention of food does keep it front of mind, but then hey... we all need to eat, and wear T-shirts*.

*Clearly not all, but I've got to assume that nudist colonies do not represent a targeted growth market for T-shirt manufacturers.
posted by Nanukthedog at 8:58 AM on January 22, 2016 [7 favorites]


There's a huge difference between subjective experience and controlled research.

My grandfather grew daffodils which have a problem with a nematode that just loves to burrow deep inside, over time he and that industry learned how to kill them (boil 12 hours with formaldehyde). At one point he saw that the scientists at the ag-station had a report, pretty expensive but he sent for it. Confirmed exactly what they had been doing for years.

Also see cowpox

Science should learn from everything. And we do need more citizen research projects with red wine and dark chocolate.
posted by sammyo at 8:58 AM on January 22, 2016 [2 favorites]


This is absolutely perfect! I have short, stubby little legs and a long torso so short shirts are an absolute no-go. Gap tends to be fine but apparently there are some other shirts I should check out. Nice to know the places I should avoid as well. I have been hoping to stumble across something like this for a while, thanks!
posted by Neronomius at 9:12 AM on January 22, 2016


my father and his buddies in the garment trade could have told them all of this decades ago, for the price of a martini (or maybe two). One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom...

So not to put too fine a point on it, but where's the link to your father's and his buddies' blog? Do they have an email list archive somewhere? Have they written a wikipedia page?

Private expertise is valuable to those who have it, but it's not preserved well across time, nor, by it's secretive nature, is it generally well-known or understood. Indeed, often that's the intent. Many industries have a guild mentality, where the secret knowledges of the insiders are deemed highly valuable. Trade secrets are valuable, but they're antithetical to social advancement---I may be a bit of an extremist here, but still.

I'm in science, which is another of those sectors that's frequently accused of poor communication. For me to take the attitude that those who want answers on my disciple should simply do the work of seeking me out, that's the sort of arrogance that's called out daily in the media, and is a significant thread of US national politics even.

There's an increasing sentiment within science that one is not a scientist at all, if all one does is lab/theory work. We have to publish. But it's not enough to just publish, we have to be able to teach. Increasingly as well, it's not enough to teach, you have to be able to communicate with the public, appear in the media, write articles for magazines, have a popular blog, to really be successful.

So just being an expert isn't enough. I've had this experience in my field as well, something that's very well known and understood gets replicated and "discovered" (sometimes repeatedly) by a member of the public. That's not their failure to do their homework (though it would be nice if they did), it's our failure as researchers to make the results known to people who need them, in a way they can manage to use them. And that doesn't just mean journal publications.
posted by bonehead at 9:37 AM on January 22, 2016 [24 favorites]


For you long-torso types, husband has had good luck with Duluth longtail shirts. Pricier, but wear like iron.
posted by emjaybee at 9:54 AM on January 22, 2016 [2 favorites]


Since they didn't test ASOS, a note for shorter skinny dudes. I am about 110 lbs and 5'0" and a small fits me fine.
posted by desjardins at 9:58 AM on January 22, 2016 [2 favorites]


My father and his buddies in the garment trade could have told them all of this decades ago, for the price of a martini (or maybe two). One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom...

Would your father and his buddies have actually had the detailed measurements of a number of different brands of shirts readily available to them to be shared over a martini or two? Because the value here isn't really knowing 'some shirts run big, and some run small' because everyone already knows that. It's having actual comparisons all on one page so you can start to pick your sweet spots.
posted by jacquilynne at 10:12 AM on January 22, 2016 [4 favorites]


Aw c'mon gang - of course the research is precise and useful - I was just laughing (apparently along with some other folks upthread - no pun intended) at the breathless tone with which they announce their findings, especially some of the underlying, long-understood PRINCIPLES that drive much of the variation in sizing and quality. Of COURSE the best results always come from combining ol' hard-earned wisdom with rigorous research. Although talking to a few geezer experts first might have saved them some time or even enhanced their research design...
posted by twsf at 10:25 AM on January 22, 2016 [2 favorites]


Now that I see this, I am reminded of an experiment that I ran several years ago. My friend and I started a screenprinted t-shirt business just to have some fun and we had to decide which shirts to use as blanks. I, being a scientist, bought black t-shirts in my size from dozen of manufacturers and ran them through hundreds of washes in hot water and dried on high heat and put them on and pulled them off too many times to count. I know they get a lot of hate, but the American Apparel poly-cotton blend shirts were amazing. Barely any stretch or shrinkage, very little pilling and virtually no fading. Surprisingly, AA's 100% cotton shirts were very ho-hum.

I am fascinated by the application of rigorous testing to even the most mundane of topics. Little good it did us, I think we sold a couple of dozen shirts to people we weren't either related to or friends with.
posted by roquetuen at 10:29 AM on January 22, 2016 [1 favorite]


It is just nice to have v necks given a fair place in the ranks in this great tool they've made.
posted by monopas at 10:36 AM on January 22, 2016


I think we sold a couple of dozen shirts to people we weren't either related to or friends with.

Sure, there are a lot of days like that, but you can also sell about 500 shirts on a game day weekend with about 10 guys at a 12,000 student school sporting event when your biggest rival is there and your shirt is the right level of offensive and tacky. After costs get removed, there might actually be enough money for a few months rent for everyone and a decent sized party. ... Just avoid the campus police...
posted by Nanukthedog at 11:05 AM on January 22, 2016


I know they get a lot of hate, but the American Apparel poly-cotton blend shirts were amazing

roquetuen, I still wear some of my samples from pre-retail American Apparel, which have held up much better than many other shirts bought more recently.

A few things this study doesn't record: body taper (delta between shoulder width and waist width), sleeve length, sleeve taper. These are all make-or-break for fit and it's why I will never wear a Beefy-T.
posted by a halcyon day at 11:18 AM on January 22, 2016 [1 favorite]


I would love to see something like this for pants. Kids these days and their tight pants...
posted by kevinbelt at 11:35 AM on January 22, 2016 [1 favorite]


Those Pima Basic tees from Banana Republic tho
posted by oceanjesse at 11:44 AM on January 22, 2016


When I do laundry I have t-shirts that:

1) can be washed but can never go near a dryer and must hang dry
2) can be washed but must never go near a dryer and must be stretched out in every direction before it is fully dry
3) can be washed but must never go near a dryer and must be stretched in length before it is fully dry
4) can be washed and dried on low for a few minutes to get wrinkles out
5) can be washed and must be dried on high heat to shrink it back down to size
6) can be washed and dried any which way because they're probably too big and/or men's boxy shirts that only get worn around the house

I'm jealous. I really want one of these for women's t-shirts so I can find a shirt that fits like AA women's L but with about six inches extra length.
posted by Room 641-A at 12:29 PM on January 22, 2016 [5 favorites]


I wonder why they did men's T-shirts but not women's. In my experience, women care more about the way their T-shirts fit than men do.
posted by The corpse in the library at 1:20 PM on January 22, 2016 [4 favorites]


Fewer measurements to take, what with the width on men's t-shirts generally being the same from the bottom of the sleeves to the hem. I can see how the data visualizations would be trickier. Which isn't to say it shouldn't be done, obviously -- but it's gonna be more work to design the experiment and more work to take more measurements for each garment.

It wouldn't be any more difficult to repeat the experiment with women's t-shirts with straight cuts, but in most of those cases I'd just look at these datapoints and buy the men's version in the appropriate size -- because it's probably cheaper for no good reason.
posted by asperity at 3:06 PM on January 22, 2016


I skimmed down the chart to see which ones run long, and I had to google the outlier (the "longline" shirt). That isn't a shirt, that's a dress for men who wouldn't buy it if it was called a dress.

I'm not sure this is relevant to my life (since I know what fits and I'm happy where I am at), but anything that makes comparing brands and finding clothes that fit easier is a good thing in my book.
posted by Dip Flash at 8:15 PM on January 22, 2016


All my t-shirts, over time and repeated washing, get a weird twist-stretch so that if the sleeves and chest are lying flat, the hem end is several inches from being straight. It's hard to describe, but on the ones with side seams, one seam curves down from armpit toward belt buckle and the other 'round the back. I have never known what causes that, but it's been happening as long as I remember.
posted by ctmf at 8:27 PM on January 22, 2016 [2 favorites]


Always the same direction of twist, too. It's my t-shirt mystery.
posted by ctmf at 8:32 PM on January 22, 2016


twsf: "I laughed so hard at this when it first posted - I appreciate their earnest research but my father and his buddies in the garment trade could have told them all of this decades ago, for the price of a martini (or maybe two). One of the downsides to the current tech/youth mindset is the utter ignorance of, and lack of respect for, venerable experts and their wisdom..."

Old Scout: Oh yes, it passes the eye test—definite five-tool t-shirt: wears for comfort, wears for style, washes, dries, folds. I'm telling you, real arms on this shirt. Got that long, lean build like a Fruit of the Loom crew-neck. Could even be the next Hanes Tagless one day. Sky's the limit on this one.

Young Nerd: Okay, but how many standard deviations does the chest size run after the 16th wash?

Old Scout: Look, kid, this t-shirt has a great-looking girlfriend.
posted by Conrad Cornelius o'Donald o'Dell at 8:49 PM on January 22, 2016 [2 favorites]


This was my life last week. But for kids t-shirts, with a non-US manufacturer whose L was a western M, maybe and started at SSS, except all the online sizing info I could find was so widely varied, it was hopeless. And then the manufacturer said the sizing variance would be +/- one inch but they agreed to add an inch to the length without increasing the price. Basically we're just crossing our fingers.
posted by TWinbrook8 at 9:24 PM on January 22, 2016


Ctmf: my shirts do the same thing. They're all slowly creeping around to the right.
posted by The corpse in the library at 9:49 PM on January 22, 2016


ctmf: it must be the coriolis force. Are you in the northern or southern hemisphere?
posted by otherthings_ at 11:07 PM on January 22, 2016 [3 favorites]


In regards to the side-seams-wrapping around the body-- this is called torque (twist). Lots of research and engineering goes into reducing torque.

Perhaps you are familiar with hand-knitting. The knitter can use two needles, going back and forth. Or the knitter can use circular or double-ended needles, which simply go round and round.

Knitting machines knit the fabric in a continuous spiral, which forms a tube. At reaching the end of the original stitches, the row of knitting jumps up one step and forms a row on top of the original stitches. That jump forms a tiny diagonal pull in one direction. As the spiral goes round and round, that diagonal pull distorts the tension in the stitches.

The tube can be knit to size. In the cheapest shirts, the size "Small" tube is cut to length, slashed at the top to add sleeves, and the neck and shoulders added.

In more expensive shirts, the tube is layed flat, and the front and back are cut (with cookie cutters) and sewn together with side seams. If the side seams are shaped, more fabric has to be thrown away.

In more expensive shirts, the armholes are scooped out and the sleeves are sculpted so that the garment is three dimensional.
posted by ohshenandoah at 8:23 AM on January 23, 2016 [5 favorites]


So not to put too fine a point on it, but where's the link to your father's and his buddies' blog?

Note: Personal observations about daffodil tech originally made prior to the transistor. No Blogs available.

Bazingoblasty (to coin a swear word) science is important. Recording every last datapoint about a silly study of teenage girls preference of scrunchy in an overpriced difficult to access journal for the inner circle is vital, or at least the references to the study in the next grant request. What's really important in science? A unified theory or everything that no one can discover or another study on the health benefits of coffee? Gotta be hard to be a scientist these days. I expect that there were no trade secrets for my grandfather and data was shared by the highest technology available at the time (handwritten letter). We don't need the Singularity AI to solve all our problems just to help us craft a useful google filter to find a bit of valid information mixed with all the legitimate (as well as falsified) cruft.
posted by sammyo at 8:26 AM on January 23, 2016


I wonder why they did men's T-shirts but not women's.

My guess would be that women's shirts and fits differ even more vastly between makers and designers. Not that the endeavor isn't worthwhile, but it probably made sense for them to look at the items with smaller variation first to test their methods and so on before moving on to hard mode.
posted by BlackLeotardFront at 7:23 PM on January 23, 2016


I wish someone would investigate crew-neck collar expansion in t-shirts. I have very broad shoulders and have a problem with t-shirt collars stretching and sagging in the front after just a few wearings. With white t-shirts, I end up buying a small size, even though I'm a medium, just to keep the collar from sagging.

It's worse with printed tees; the only kind I have ever found that doesn't sag is American Apparel.

I can't be the only one with this problem.
posted by yellowcandy at 4:02 PM on January 24, 2016


You are not alone, yellowcandy! The struggle is real. I don't have broad shoulders but I have mobility issues in my arms so there is some contortion involved in putting on a t-shirt. I get the saggy collar effect after a few wears unless they are thicker material.
posted by desjardins at 8:07 PM on January 25, 2016


« Older “Your house is a crime scene and you two are...   |   'Starring Willie W, Bas Bron and Vieze Fur.' Newer »


This thread has been archived and is closed to new comments