The short answer
There is no standard a garment can be true to. The sizing standards that exist for children’s clothing measure children, not clothes, and every decision that turns a body measurement into a garment measurement is left to the brand. On top of that sits a manufacturing tolerance nobody prints, and after that a shrinkage allowance that changes the garment once you have already paid for it. Stack those three and the same label can legitimately produce garments a full size apart.
| What people assume | What is actually true |
|---|---|
| A size 3T means a defined chest measurement | It means the brand’s estimate of a three-year-old’s body, plus whatever ease the brand chose to add |
| Sizing standards are enforced | They are voluntary. No US or EU law sets apparel sizes |
| Two garments in the same size from one brand are the same | They are the same within a tolerance, commonly around ±1 cm on the chest |
| The measurement on the tag is what arrives | Knits are allowed to shrink several per cent in the first wash |
| “True to size” in a review is a property of the garment | It is a statement about one garment and one child, and only one of those is described |
None of this is a scandal, and none of it is a brand behaving badly. It is how apparel sizing has always worked. It just means the phrase parents rely on most is the one carrying the least information.
The standards measure children, not clothes
The US standards for children’s sizing are published by ASTM: D4910 for infant sizes from preemie to 24 months, D6192 for girls 2 to 20, D6458 for boys 2 to 20. They are titled, accurately, Standard Tables of Body Measurements.
The scope statement of D4910 is unusually candid about what that means. These tables list body measurements of infants’ sizes, it says, and although these are body measurements, they can be used as a baseline in designing apparel when considering such factors as fabric type, ease for body movement, styling, and fit.
Read that clause again, because the whole subject sits inside it. The standard tells a brand roughly how big a fourteen-month-old is. Everything between that number and the garment on the rail — how much room to add for movement, how much for the fabric, how much for the silhouette the designer wants — is a commercial decision the standard explicitly leaves open. Two brands can follow the same table faithfully and cut chests three centimetres apart, and both are correct.
Three further things follow, and they surprise people:
- The standards are voluntary. No law in the US requires a garment labelled 3T to measure anything in particular. There is no sizing regulator, and no penalty for a chart that flatters.
- Europe is no different. EN 13402 sets out how clothing sizes should be designated — in centimetres, by body dimension, with a pictogram — and it too is voluntary and describes bodies rather than garments. Which is why a German 92 and a French 2 ans and a US 2T do not reliably arrive the same width.
- Nothing obliges a brand to publish garment measurements at all. Most publish a body chart, which tells you which child the size is aimed at, not what will turn up in the parcel.
The one part of children’s clothing where a garment measurement is legally fixed is US snug-fit sleepwear, where maximum finished dimensions at seven points are written into federal regulation. That contrast is the sharpest illustration in the whole subject: where the law cares, it specifies the garment; everywhere else it specifies nothing. How that works is in the US sleepwear rules explained.
Tolerance: the plus-or-minus nobody prints
Every garment spec carries a tolerance, because cloth is not sheet metal and a cutting table is not a machine shop. Fabric relaxes, stretches on the roll, and behaves differently at the top and bottom of a spread. So a factory is given a target and a permitted deviation.
For a chest measurement on a knit, that tolerance is commonly in the region of half an inch, or about a centimetre, and on very stretchy knits it can be wider. The garment is passed if it lands anywhere inside the window.
That is a completely reasonable engineering allowance. The problem is its size relative to the thing next to it.
The arithmetic that kills the phrase
Take a toddler t-shirt. A common grade rule steps the chest by about 2.5 cm per size, so a 3T is roughly one inch wider than a 2T. Now add the other two numbers.
| Source of variation | Typical size, on the chest |
|---|---|
| One size step (the grade rule) | about 2.5 cm |
| Manufacturing tolerance | about ±1 cm, so a 2 cm window |
| Permitted shrinkage on a cotton knit | up to about 5 per cent, which on a 52 cm chest is around 2.6 cm |
The tolerance window alone is most of a size. Add shrinkage and the spread inside a single label exceeds a full size step. A garment at the top of tolerance that has not yet been washed, and a garment at the bottom of tolerance after three hot washes, are two different sizes wearing the same tag, and both passed quality control.
This is the honest answer to why a brand can feel consistent for a year and then produce one garment that feels wrong. It usually is not a change in the pattern. It is the same pattern landing on the other side of a window that was always there.
Shrinkage is a second size change, after you have paid
Dimensional change after home laundering is a measured property, tested under AATCC Test Method 135, and the commonly applied commercial limits are around 3 per cent for wovens and 5 per cent for knits. A garment that shrinks within that is compliant, not defective.
Five per cent does not sound like much until it is converted. On a 52 cm chest it is 2.6 cm. On a 60 cm body length it is 3 cm, which on a toddler is the difference between a t-shirt and a crop top.
Two practical consequences:
- Judge the fit after the first wash, not out of the parcel. This conflicts with return windows, which is genuinely awkward, and the resolution is to try the garment on before washing and only then decide. That sequencing advice is in the baby sleepwear sizing guide.
- Most of the shrinkage is avoidable. Heat causes it, in the wash and far more in the dryer. A cool wash and line drying keeps most of the size you bought, and the rest of that routine is in how to wash baby clothes.
Fabric decides what a measurement means
Even where two garments measure identically flat, they do not fit identically, because a measurement is not a fit until you know what the fabric does with it.
| Fabric | What the same chest measurement means |
|---|---|
| Woven cotton, no elastane | The measurement is the limit. Nothing gives, so ease has to be designed in |
| Single jersey, no elastane | Some give, mostly one way, and it does not fully recover |
| Jersey with 5 per cent elastane | Substantial give that recovers, so one measurement covers a wider range of children |
| Rib | The most forgiving. A ribbed cuff or body reads as two sizes of stretch |
| Interlock | Denser and more stable, so it feels closer to its measurement than jersey does |
This is why the same brand can feel true to size in leggings and small in shirts. The leggings are 5 per cent elastane and stretch to meet the child; the shirt is a woven and does not. Nothing about the sizing changed.
It also explains a familiar and confusing experience: a garment that fits at first and then stops, without shrinking. Jersey without elastane grows at the neck and cuffs with wear, then relaxes unevenly, and the garment ends up both looser and worse fitting than it started.
What “true to size” actually means in a review
The phrase is doing something specific and it is worth naming, because once you see it you can use reviews properly instead of distrusting them.
“True to size” is a two-variable statement reported as one. It means this garment fitted this reviewer’s child in the size the reviewer expected. The garment is unnamed in measurements, and so is the child. Transferring the conclusion to your child requires both to match yours, and the review tells you neither.
Which reviews are worth reading, then:
- Ones that state the child. “My 3-year-old is 97 cm and 14 kg, the 4T fits with room” is usable data. “Runs small” is not.
- Ones that name a comparison brand. A reviewer saying a garment is wider than the equivalent from a brand you already own has given you a measurement in disguise.
- Ones that mention the wash. “Fitted perfectly, then shrank” is a different fault from “arrived small”, and only one of them is a sizing problem.
And the pattern to discount: a cluster of “runs small” on a snug-fit sleepwear listing, which is usually the fit working as designed rather than a sizing error. Why that happens specifically, and where PatPat’s own sizing does genuinely cause trouble, is set out in does PatPat run small.
The one number that transfers between brands
Body measurements are useful for choosing a size within one brand’s chart. They are almost useless for comparing two brands, because the ease is invisible and it is the ease that differs.
What transfers is the flat garment measurement, and the cheapest source of one is a garment you already own that fits.
- Find the best-fitting garment of that type your child owns. Type matters: a t-shirt tells you nothing about a coat.
- Lay it flat, smooth it, and measure across the chest from armpit to armpit. Do not double it and do not stretch it. Write the number down.
- Measure the body length from the highest point of the shoulder straight down to the hem, and the sleeve from shoulder seam to cuff.
- Compare those three numbers to the new brand’s garment chart, if it publishes one. If it publishes only a body chart, ask customer service for the flat chest measurement of the size you are considering. Many brands will give it; the question marks you as someone who will not be returning three sizes.
- Allow for the fabric. A ribbed or elastane-blended garment can measure 2 cm narrower than your reference and still fit.
Three numbers, taken once, and they stay useful for months. It is far more reliable than any amount of chart reading, because it compares a garment with a garment instead of a garment with a child.
Body chart or garment chart: the distinction that causes most of the confusion
Almost every sizing argument between a parent and a brand comes down to two different documents being called the same thing.
| Body chart | Garment chart | |
|---|---|---|
| What it lists | The measurements of the child the size is intended for | The finished measurements of the garment itself |
| What it is for | Choosing which size to order | Knowing what will arrive |
| Who publishes it | Nearly everyone | A minority |
| Comparable between brands? | No. Identical body charts, different garments | Yes, directly |
| Typical reader error | Reading it as the garment, and concluding the brand runs small | Reading it as the body, and buying a size too big |
A body chart with a 52 cm chest for size 3 does not mean the shirt measures 52 cm. It means the shirt is cut for a child whose chest is 52 cm, and the shirt itself will be wider by whatever ease the brand chose. If you read that number as the garment, every brand on earth will look like it runs small.
PatPat’s charts, and which columns are body and which are garment, are on the size chart page. Which size to order for a given age and weight, and how long it will last, is a separate question answered in clothing sizes by age and weight.
When a garment is genuinely wrong, and what to say
Everything above describes normal variation. Some garments fall outside it, and the difference is measurable rather than a matter of opinion.
A garment is worth raising when:
- It differs from the brand’s own published garment measurement by more than a centimetre or so, measured flat and unstretched.
- It differs noticeably from another garment you own in the same size and style, which is the clearest evidence available because it removes every other variable.
- It shrank well beyond a few per cent, having been washed as the care label instructs.
- The two sides are not symmetrical, or one sleeve is visibly longer, which is a cutting or sewing fault rather than a sizing one.
What makes the message land is measurements rather than adjectives. “The 3T tee measures 24 cm flat across the chest; your chart says 27 cm; my other 3T from you measures 27 cm” is a quality report. “It runs small” is a preference, and preferences do not get escalated internally.
And there is a version of this worth knowing for your own sake: if a garment fits but a second one in the same size does not, keep the one that fits and measure it. You have just generated your own reference number for that brand, which is more accurate than anything published.
What this guide does not claim
The tolerance and shrinkage figures here are typical commercial practice across the industry, not PatPat specifications, and not published requirements. Individual brands set their own tolerances, and some are tighter than the numbers used above.
We have not measured competitor garments for this article. The comparison method described is the one we would use ourselves, and it is offered so you can check us as easily as anyone else.
ASTM and EN standards are cited to establish what they cover, which is body measurement and size designation. Nothing here claims PatPat garments are built to any of them, and no sizing standard is mandatory in either market.
This is not a promise that following the method will make every order fit. It reduces the number of returns; it does not remove the fact that children are shaped differently from the average the charts describe.
Frequently asked questions
What does “true to size” mean in children’s clothing? Nothing measurable. There is no standard a garment can be true to, because the sizing standards describe children’s bodies rather than finished garments.
Are children’s clothing sizes regulated? No. The ASTM tables in the US and EN 13402 in Europe are voluntary, and no law requires a garment labelled 3T to measure anything in particular.
Why do two brands label the same size differently? Because each one chooses how much ease to add to the same body measurement. Both can follow the same standard and still cut chests several centimetres apart.
How much can two garments in the same size differ? Manufacturing tolerance on a knit chest is commonly around a centimetre either way, so the window inside one label is roughly two centimetres before any shrinkage.
Is shrinkage a defect? Usually not. Commercial limits sit around 3 per cent for wovens and 5 per cent for knits, and a garment that stays inside that is compliant.
Should I judge fit before or after washing? Try it on before washing so a return stays possible, then judge the real fit after the first wash. Most shrinkage happens in that first cycle.
What is the difference between a body chart and a garment chart? The body chart lists the child the size is meant for; the garment chart lists the finished garment. Reading a body chart as a garment chart is why brands appear to run small.
How do I compare sizing between two brands? Measure a garment your child already wears well, flat across the chest, and compare that number to the new brand’s garment measurement rather than to its body chart.
Why does one brand feel true to size in leggings but small in shirts? Fabric. Leggings usually carry elastane and stretch to meet the child, while a woven shirt has only the ease the pattern was given.
Are reviews that say “runs small” useful? Only when they describe the child. A height and weight with the size that worked is data; the phrase on its own is one family’s fit reported as a fact about the garment.
Why do so many reviews say snug-fit pajamas run small? Because they are supposed to fit closely. The snug fit is the safety mechanism, so the garment measuring correct and feeling tight is the design working.
Does a European size 92 equal a US 2T? Roughly, and not reliably. EN 13402 designates sizes by body height in centimetres while US sizes are named by age, so the two only overlap approximately.
What three measurements should I keep? Flat chest armpit to armpit, body length from the shoulder to the hem, and sleeve from shoulder seam to cuff, all taken from a garment that fits.
When is a garment genuinely out of tolerance? When it differs from the brand’s own published garment measurement, or from another garment you own in the same size and style, by more than about a centimetre.
Does buying a size up solve the problem? For daywear often yes, for snug-fit sleepwear no. Sleepwear is the one category where the garment measurement is a legal maximum rather than a commercial choice.
About this article
Written and fact-checked by the PatPat Editorial Team. How we research, source and review what we publish is set out on our editorial standards page.
Method. No competitor garments were measured for this article. Tolerance and shrinkage figures are typical commercial practice rather than published requirements or PatPat specifications, and are given as ranges for that reason. Standards are cited for what they cover, not as a claim of conformity.
Sources. Children’s body measurement tables and the scope language quoted above: ASTM D4910/D4910M, infant sizes preemie to 24 months, ASTM D6192/D6192M, girls 2 to 20 and ASTM D6458/D6458M, boys 2 to 20, maintained by ASTM subcommittee D13.55 on body measurement for apparel sizing. European size designation: EN 13402-1, size designation of clothes. Dimensional change after home laundering: AATCC Test Method 135. Fibre content labelling, which is regulated where sizing is not: FTC, Threading Your Way Through the Labeling Requirements.
Related. Which size to order now and how long it lasts is in clothing sizes by age and weight; the brand-by-brand comparison is in does PatPat run small; and the one category where garment measurements are fixed by law is in the US sleepwear rules explained.
Last updated 17 September 2026 · Next review September 2027
