Winter Coats and Car Seats: What the Harness Test Assumes

Every winter the same advice circulates: no bulky coats in the car seat. It is correct, and it is almost always delivered as an assertion with nothing behind it, which is why it gets argued with in the school car park every December.

There is something behind it. The federal standard that every car seat in the United States is certified against does not merely test the seat. It specifies the crash, the dummy, the harness tension and, in a detail almost nobody quotes, the clothing the dummy wears. A winter coat sits outside two of those specifications at once. Once you have read the numbers, the advice stops being a slogan and becomes arithmetic.

What the standard actually specifies

FMVSS 213 is the federal standard for child restraint systems, set out at 49 CFR 571.213. It applies to seats manufactured before December 5, 2026, and FMVSS 213b takes over for seats manufactured on or after that date. The provisions that matter here are written identically in both, so nothing below changes when the handover happens on December 5, 2026.

The crash. Test Configuration I is a frontal impact simulation at a velocity change of 48 km/h, roughly 30 mph.

The pass marks. The head injury criterion, HIC36, must not exceed 1,000. Resultant acceleration at the dummy's upper thorax must not exceed 60 g, except for intervals whose cumulative duration is no more than 3 milliseconds. Head excursion must stay within 813 mm and knee excursion within 915 mm.

The harness tension. This is the number to hold on to. Under S6.1.2(d)(1), the harness that restrains the dummy is tightened until a 9 newton force applied to the webbing at the top of each shoulder pulls that webbing 7 mm away from the dummy. Nine newtons is about two pounds of pull. Seven millimetres is roughly a quarter of an inch. That is the condition under which every number on the certification label was produced: a harness so close to the body that a light tug lifts it a quarter of an inch.

The clothing. Section S9.1 is titled "Type of clothing" and it is prescriptive. The newborn dummy is tested unclothed. The 12-month-old dummy wears a tight-fitting cotton-polyester sweatshirt with long sleeves and ankle-length long pants, and the combined weight of the two garments must not exceed 0.25 kg. For the three-year-old dummy the shirt weighs 0.090 kg and the pants 0.090 kg.

A quarter of a kilogram is about nine ounces for a complete outfit. A toddler's insulated winter coat on its own routinely weighs more than that, and more to the point it is built from loft, which is to say from trapped air held in place by fabric.

Why loft matters more than thickness

The objection people raise is reasonable on its face. If the harness is tight against the coat, and the coat is tight against the child, where is the slack?

It is in the loft, and it appears at the moment it can do the most harm. Insulation works by holding air. Under the steady, gentle pressure of a harness being tightened by hand, that air stays put and the coat feels firm. Under the loads of a 48 km/h frontal impact, the same air is driven out of the fabric in milliseconds and the coat flattens. The webbing does not shorten to follow it. The distance the coat gave up becomes free travel between the child and the harness, and the harness only begins to restrain the child after that travel is used up.

Two consequences follow. The child's torso is moving before the harness engages, so it engages against a body already in motion rather than decelerating it from the start. And the excursion the standard measures to 813 mm now starts from further forward.

This is also why thickness alone is the wrong test. A close-fitting fleece of real thickness compresses very little because it holds little air. A thin, glossy down-filled jacket can compress a great deal. What you are judging is the loft.

The coat test

There is a check that settles it for a specific child in a specific coat, and it needs no equipment. Consumer Reports and the child passenger safety community both publish it in the same form.

1. Put the coat on the child and fasten them into the seat as usual.
2. Tighten the harness until you cannot pinch any webbing between thumb and forefinger at the collarbone. Your fingers should slide off the strap.
3. Without touching the harness adjuster, unbuckle and lift the child out.
4. Take the coat off, put the child back in, and buckle up again against exactly the same harness setting.
5. Try the pinch test again.

If you can now gather webbing between your fingers, that is the slack the coat was hiding, and the coat is too bulky to go under the harness. If the harness is still snug, the garment is compressing very little and it is fine.

The standard's own description of the target is quoted in every manual because S5.6.3 requires it to be: "A snug strap should not allow any slack. It lies in a relatively straight line without sagging. It does not press on the child's flesh or push the child's body into an unnatural position."

What to do instead

None of the alternatives involve a cold child, which is the fear driving the argument in the first place.

Layer thin and close. Long sleeves, a close-fitting fleece, thick socks. Fleece and knit are warm without much loft, and a jacket of that sort will often need a small harness adjustment rather than disqualifying itself.

Buckle first, coat second. For a child old enough to manage it, put the coat on backwards over the fastened harness, arms through the sleeves, with the back of the coat lying across the front of the child like a bib. The warmth is on the outside of the webbing where it cannot compress into slack.

Blankets go on top. Over the fastened harness, never under it. A blanket has no fastening and adds nothing to the restraint, so its position is entirely a question of where it interferes.

Hats, gloves, scarves and boots are not the issue. The harness does not pass over any of them. Only what lies under the webbing at the shoulders, chest and hips changes the fit.

Preheat the car. Two minutes of preheating does more than any garment, and it removes the trade-off completely.

For infants, the practical version is a warm sleepsuit with feet, the harness snug against it, and a blanket or a manufacturer-supplied cover over the top once the buckles are done.

Aftermarket covers, inserts and strap pads

This is where the winter question turns into a purchasing question, and the answer is less reassuring than the packaging.

Aftermarket accessories sold for use with a car seat are not covered by FMVSS 213. NHTSA has said so directly in its interpretation letters: there are no federal motor vehicle safety standards that apply to aftermarket accessory items such as strap-holder devices used with child restraint systems. The standard certifies the restraint as the manufacturer built and submitted it. A product added afterwards by somebody else has not been part of any certification test.

That matters for three categories in particular.

Bunting bags and full inserts that go behind the child, adding a padded layer between the child and the shell. This is the same compression problem as a coat, moved to the other side of the body, and it is untested in that seat.

Harness pads and strap covers bought separately from the seat. There is no federal performance test for them, so a claim of being crash tested refers to a protocol the seller chose.

Shower-cap style covers that stretch over the outside of an infant carrier without passing under the harness or between the child and the shell. These sit in a different position entirely, because nothing about the restraint path changes. Many manufacturers sell one for their own seat, and the manual is the authority on whether yours permits it.

The rule that covers all three: if the accessory did not come in the box with the seat, and the manual does not name it, then nobody has crash tested that seat in that configuration. That is not the same as proof of harm. It is an absence of evidence where the packaging implies the presence of it.

The fit checks worth doing at the same time

Winter is when most families revisit the harness anyway, so three adjustments are worth making in the same trip.

Harness slot height. Rear-facing, the straps come from at or just below the shoulders. Forward-facing, at or just above. Children grow through slots in autumn without anybody noticing.

Chest clip height. Armpit level. It is a positioner that holds the two straps at the correct width, not a restraint, and low on the abdomen it does neither job.

Where you pinch. At the collarbone, on the strap, never at the hip. A harness can be tight at the lap and slack at the shoulder, and it is the shoulder that governs how far the torso travels.

Seat fit in the vehicle is a separate question and it is covered in will this car seat fit my car. How long each stage lasts is in how long rear-facing. The seats themselves are ranked in the car seat roundup. The first federal side-impact requirement, which shares its December 5, 2026 date with the 213b handover, is in what FMVSS 213a changes, and the date moulded into the shell is explained in car seat expiration dates.

Questions people actually ask

Is a fleece jacket acceptable under the harness?

Usually yes. Fleece and knit hold far less air than an insulated shell, so they compress very little. Run the coat test with the actual garment rather than assuming: the test takes two minutes and settles it for that child in that seat.

What about a snowsuit for a baby in an infant carrier?

A padded snowsuit has the same loft problem as a coat and it adds it at the hips as well as the chest. A warm footed sleepsuit under the harness, with a blanket or a manufacturer-supplied cover over the fastened buckles, gives the same warmth without changing the harness fit.

Does the standard really say what the dummy wears?

Yes. Section S9.1 of 49 CFR 571.213 is titled Type of clothing. The newborn dummy is tested unclothed, and the 12-month-old dummy wears a tight-fitting sweatshirt and long pants weighing no more than 0.25 kg combined. The clothing is a controlled test condition, not an afterthought.

How tight is the harness supposed to be?

In the compliance test it is tightened until a 9 newton force pulls the webbing 7 mm off the dummy's shoulder, about two pounds of pull for a quarter of an inch of movement. At home the equivalent is the pinch test at the collarbone: if your fingers slide off the webbing rather than gathering it, the harness is at the tested condition.

Can I just tighten the harness harder over the coat?

No, and this is the most common workaround. You can only tighten to the point the coat allows, and the coat still compresses in a crash. The slack is created by the material giving way under load, so it is not removable by pulling harder against the uncompressed garment.

Are car seat ponchos and covers safe?

It depends entirely on whether anything sits between the child and the harness or between the child and the shell. A cover that goes over the outside of the seat without touching the restraint path leaves the fit unchanged. Anything the harness passes over, or that the child sits on, has not been tested in that seat unless the seat's own manual names it.

Sources

• eCFR, 49 CFR 571.213, FMVSS No. 213 Child restraint systems
• eCFR, 49 CFR 571.213b, mandatory applicability beginning December 5, 2026
• NHTSA, keep your little ones warm and safe in their car seats
• NHTSA interpretation 07-007038, magnets and strap holders
• Consumer Reports, the dangers of winter coats and car seats
• Car Seats for the Littles, non-regulated products
• Safe Kids Worldwide, right fit for your child's car seat
• Michigan State University Extension, why winter coats and car seats are a dangerous combination