How we test

One rule decides everything else here: we buy what we test. no loaners, no sponsored placements, no pay-for-position. This page explains what that costs us, how a guide gets built, what the number on each product card means, and — the part most review sites skip — what we do not claim to know.

We buy it. Nobody gives it to us.

We do not accept free product from manufacturers and we do not accept payment from suppliers in any form. Not for a review, not for an award, not for a place in a guide. A reviewer holding a gifted car seat has a relationship to protect, and it is never the reader's.

The site is reader-supported instead. If you buy through a link here we may earn a commission at no cost to you. It is the same commission whichever product you choose, so it cannot move a score, and no brand is told its score before publication.

Safety is a gate, not a score

This is the one place we work differently from a general gear site, and it is worth being blunt about. Safety is not one of the five things we mark out of ten. A car seat that fails a safety check does not score six for safety and eight for everything else, landing somewhere respectable. It does not appear in the guide at all.

Before a product is tested it is checked against the applicable federal standard and against open recall notices at CPSC, NHTSA and the FDA. Anything carrying an open recall comes out of the lineup. Where a product meets the standard but sits against current AAP guidance — a powered rocking bassinet is the clearest example — we say so on the page rather than letting a good score imply an endorsement it has not earned. Recall checks are re-run before every republication, because a seat that was clean in March may not be in September.

To be plain about the limits of that: our safety gate is documentary, not clinical. We check the standard, we check the recall registers, and we tell you what current guidance says. We are not certifying that a seat is correctly fitted in your car or that a product suits your child — a check against a register is not the same thing as a technician's hands on the installation, and we will not imply otherwise.

We run two tracks, and the gap between them is the story

Every product here goes through two separate processes that never touch until the end.

The first is our own testing. Products go to testing households and get used for what they are for, in ordinary life rather than in a lab — school runs, night feeds, holidays, supermarket car parks, a week of rain. A stroller that folds beautifully on a shop floor and badly in a real boot has failed a test that no specification sheet administers.

The second is research at a scale no household can reach. Separately, and in parallel, we survey real owners and work systematically through verified customer reviews — hundreds of people per product, across years of ownership. We read the one-star reviews as carefully as the five-star ones, because complaints cluster, and a complaint that appears in forty reviews is a design fault rather than bad luck.

Then we put the two together. The pattern that matters most is where they disagree: gear that impresses for a fortnight and fails in a season is exactly what a single testing period cannot see on its own, and it is the reason we do not publish on hands-on testing alone. Where our own experience contradicts the owner data, the page says so rather than quietly picking whichever is more flattering.

When a group of testers uses a product, each of them writes up their own notes independently, before anyone has said a word to anyone else. Nobody knows whether they are the outlier until their notes are already in.

Only then does the team sit down and go through them, and that is where a verdict gets made — because the disagreements survive. If three testers found a carrier comfortable and one could not get it to sit right on a short torso, that does not average quietly away into a seven. It becomes a sentence telling you who this carrier is wrong for.

The order matters for a boring reason: the first person to speak sets the room. Testing done as a conversation produces one opinion held by four people. Testing done blind produces four opinions, and where they diverge is usually the most useful thing we learn.

Our testers are deliberately not one body type. Handle height, torso length, and what it takes to lift a loaded infant seat into a back seat all vary enormously between people, and a product tested by one build has only been tested for that build.

How a guide gets built

Research the field. We work through the category — best sellers, new releases, and the models parents keep recommending to each other — to find the genuine contenders rather than the ones with the biggest ad budget. Two products in every guide are picks the mainstream roundups miss.
Clear the safety gate. Standards compliance and open recall checks, before anything is bought.
Set the criteria before testing, not after. Deciding what matters once you have played with the gear is how a favourite quietly becomes a winner.
Use it properly. Products go into real use, for what they are for, for as long as that takes.
Write up blind, then meet. Independent notes from every tester, handed in, then a group discussion to reconcile them.
Survey owners and read the reviews. In parallel, we survey real owners and work through verified customer reviews at a scale no testing household can match — hundreds of people, across years of use.
Put the two together. The pattern that matters most is where they disagree. Gear that impresses in a week and fails in a season is exactly what a short test cannot see on its own.
Score and award. Six criteria, each out of ten, averaged into a score out of 100.

The six things we score

Crash test: Measured on an instrumented sled: head injury criteria, chest acceleration, and how far the shell rotates forward before it stops.
Ease of install (LATCH): How hard it is to get a base tight on the anchors, and whether the seat tells you when you have. A seat installed loose is a seat that has lost most of its crash rating.
Ease of install (belt): The same question with a vehicle seatbelt, base or no base. This is the one that matters in a taxi, a rental, and a grandparent car.
Ease of use: Harness adjustment, buckling, release, canopy. Weighted for use at four in the morning, one-handed, in the dark.
Comfort & durability: Fit and support for the child, and whether the fabric and hardware survive being washed as often as this category demands.
Weight: What you actually carry — shell plus canopy plus inserts, not the stripped figure on the box.

What the number actually means

The badge on each product card is a score out of 100 — the six criteria, each out of ten, averaged. The breakdown underneath shows how it did on each, so you can see where a product earned its number instead of taking the total on trust.

Scores are relative to the field tested in that guide. A 78 in a fiercely competitive category can be a better buy than an 85 in a thin one. The breakdown and the write-up tell you which you are looking at.

The three awards

Top-Notch Choice: Best overall. Performance first.
Best Value: Quality gear that gives you the most for the money.
Top Pick: Best for one specific job, even when it is not the best all-rounder.

What we do not claim

Most methodology pages list only strengths. Here is the other half, because it is what tells you how much to trust the rest.

We do not print prices. We rank in tiers — budget, mid-price, premium — rather than quoting figures. Amazon's terms only permit displaying prices pulled live from their API on their refresh schedule. We do not pull it, so any number we typed would be stale within a week and in breach besides.

Where we have not verified something first hand, we say so. Not every figure in every guide comes from our own hands on that exact unit. Where it does not, the text says where it came from — owner reports, manufacturer figures, a standards document — rather than dressing it up as our own testing. In the comparison tables, anything we have not measured ourselves reads "not measured" rather than sitting blank, because a blank cell looks like a zero.

We are not your paediatrician. Nothing here is medical advice. For anything touching your child's health, ask someone who has met them.