How we research

Each of the 50 products on this site went through the same five steps, and the dates of each check appear on the product itself.

1. 01

   Confirm it is currently sold

   A recommendation for a discontinued model is useless. Every product carries the date we confirmed it is still on sale, and a model that has been replaced gets swapped for its successor rather than quietly left in place.

2. 02

   Read the manufacturer's own specification page and manual

   Not a retailer listing, not a competitor's review. Every figure in a specification table names the page it was read from, which is what makes it checkable rather than merely asserted.

3. 03

   Check the recall database

   Against NHTSA's recall records, with the date recorded. Recalls apply to specific manufacture-date ranges rather than whole model names, which is why every product page tells you to check your own unit's date stamp.

4. 04

   Look for a published third-party rating

   NHTSA publishes an Ease-of-Use rating per seat and IIHS publishes booster belt-fit ratings. Both sites block automated access, so no per-model figure could be read from the primary source — see below for what we do instead.

5. 05

   Score it against fixed criteria

   Five criteria, the same for every product, so two products can be read against each other. The criteria and their reasoning are in the next section.

Every product carries a 0-10 score in five categories, weighted equally into an overall figure. They are published so you can disagree with the weighting rather than having to guess at it.

This is the most unusual thing about the site and it deserves a direct explanation.

There is no published consumer crash score for individual car seat models in the United States. What exists is FMVSS 213 — the pass-or-fail federal standard every seat must meet — plus NHTSA's Ease-of-Use rating and IIHS's booster belt-fit ratings. Those last two are genuinely useful and we point at them constantly.

What we cannot do is reprint them. Both NHTSA's and IIHS's rating databases return errors to automated requests, so no per-model figure could be verified from the primary source. We could copy numbers from secondary websites, and many sites do. On a page a parent may use to choose what protects their child, publishing an unverified or stale safety number is not a risk worth taking — so every product page links the official lookup tool for that exact model instead.

The cost is real: the pages look less authoritative without a number. We think that is the correct trade. Full explanation of what each rating measures is in car seat safety ratings explained.

Every product photograph on this site is a licensed stock image of the product CATEGORY, not a photograph of that specific model, and every one is labeled 'Category photo' in the interface.

The reason is the same as everywhere else here: we do not have manufacturer image feeds for this site and we do not own the units, so a photograph implying we did would be a small lie in service of a nicer page. Labeling it costs us polish and keeps the page honest. When a real product image feed is available, these will be replaced.

• A recall naming a model would be added to its page immediately, with the affected manufacture-date range.
• A specification change — manufacturers quietly revise shell heights and weight limits without changing a model name — would be re-read from the source and re-dated.
• A published third-party rating we could finally verify would be added, with its source.
• Discontinuation would move the product out of recommendations and its successor in.
• A reader correction with a source. These go to the front of the queue — see contact.
• A live price data layer, once available, would replace 'Check price on Amazon' with a dated figure through the normal price layer rather than hardcoded numbers.

What would NOT change a score: commission rate, brand relationship, or whether a product is in stock. Our editorial policy covers the independence rules, and the affiliate disclosure covers the money.