What does an EWG score actually mean?

It rates hazard — whether a substance could cause harm at some dose — rather than risk, which is whether it will cause harm at the amount actually present in the product and the way you use it. A high score means an ingredient has been linked to harm somewhere in the literature, or that data about it is missing. It does not mean the product in your hand will hurt you.

What is the difference between hazard and risk?

Hazard is the capacity to cause harm. Risk is hazard multiplied by exposure. A shark is hazardous; the risk depends on whether you are in the water. Salt is hazardous at a large enough dose and harmless on your dinner. Safety assessment is a risk question, because dose and exposure are what determine outcomes, and a rating system that leaves those out is answering a different question.

Why do some ingredients get a bad score just for lacking data?

Skin Deep applies a data-gap penalty: less publicly available data pushes the score up. The intention is precautionary, but the side effect is that a newer, well-formulated molecule with a thin public literature can score worse than an older ingredient that has simply been studied for longer. Dermatologists and cosmetic chemists have raised this particularly around preservatives.

What is EWG Verified and is it independent?

EWG Verified is a mark brands pay EWG to carry. EWG is also the body that sets the criteria and does the verifying, without a third-party certifier in between. EWG says the fees cover the cost of verification and that the programme drives transparency. Critics, including reporting in Asparagus Magazine and Fashionista, argue that a single organisation writing the standard, judging compliance and collecting the fee is a structural conflict worth disclosing whenever the mark is cited.

No. It is a fast, free index of what has been published about an ingredient, and it has made ingredient lists something ordinary shoppers look at. Read as a pointer to further reading it works. Read as a safe-or-toxic verdict it is incomplete, because the number does not include the dose.

What should I use for an actual safety assessment?

The Cosmetic Ingredient Review panel in the US and the EU Scientific Committee on Consumer Safety both publish exposure-based opinions that state the concentration considered safe, the exposure assumptions used, and the reasoning. They are less convenient than a score out of ten and considerably more informative, because they answer the question in the units that matter: how much.

What an EWG Score Actually Measures (Hazard, Not Risk)

Researched, not tested. We have not personally used this product or protocol. This page answers the question from published sources, every one of which is linked. Our first-hand testing lives in reviews.

Evidence: mixed Mixed evidence — the data are limited, indirect, inconsistent or formulation-specific.

Almost every argument about this database comes down to one distinction that takes about thirty seconds to learn and then never stops being useful. It isn't specific to skincare, it isn't controversial among people who assess chemical safety for a living, and once you have it, the scores become much easier to read for what they are.

The thirty-second version

Hazard is whether something can cause harm. Risk is whether it will, given how much of it you actually meet.

A shark is hazardous. Standing in a car park, your risk is zero.

Salt is hazardous — enough of it will kill you — and the risk from seasoning your dinner is nil.

An EWG Skin Deep score is a hazard rating. It tells you an ingredient has been associated with harm somewhere in the published literature, at some dose, in some model organism. It does not tell you what happens at 0.3% in a moisturiser you rub on your arms.

That’s the whole thing. Everything below is detail.

Why the distinction isn’t a technicality

Dose is not a footnote in toxicology, it’s the subject. Two vitamin A tablets are healthy and two hundred will damage your liver. Water is fatal in the right quantity. There is no such thing as a safe substance, only a safe exposure — which is why every regulator that assesses cosmetics works in concentrations rather than in verdicts.

When you strip dose out of the calculation, you get a system where an ingredient present at trace levels and the same ingredient at industrial concentration produce the same red number. Toxicologists and science communicators have made this criticism of Skin Deep for years, and it is the single most reliable thing to know about how the scores behave.

The data-gap problem

This one is less known and more consequential in practice.

Skin Deep raises an ingredient’s score when there is limited publicly available data about it. The logic is precautionary and defensible on its face: we don’t know, so assume the worst.

The side effect is not defensible. A well-characterised modern molecule with a short publication history can score worse than a long-established ingredient that has simply been sitting in the literature for fifty years. Absence of evidence gets converted into evidence of harm, and that conversion happens silently inside a number.

Dermatologists and cosmetic chemists have pushed back on this hardest around preservatives, which matters because under-preserved products are a real microbial safety issue, not a hypothetical one.

EWG Verified, and the part worth disclosing

EWG Verified is a certification brands pay for. EWG writes the standard, applies it, and collects the fee, without an outside certifier in the loop.

EWG’s position is that the fees cover the cost of a rigorous process and that the mark drives market transparency. That’s a reasonable defence and plenty of certification schemes are funded the same way.

The counter, made in Asparagus Magazine and Fashionista, is that one organisation occupying all three roles is a structural conflict, and that it should be named whenever the seal is used as evidence. We think that’s the right standard, and it’s the reason we mention it here rather than treating the mark as neutral.

What EWG is genuinely good for

We’re not interested in a takedown. Used for what it is, the database is useful.

It’s a fast, free index of what has been published about an ingredient, with citations you can follow.
It made ingredient lists visible to ordinary shoppers, which was a real shift.
If you’re trying to find whether anyone has raised a concern about an ingredient, it’s an efficient starting point.

The failure is only in the last step: converting the index into a verdict. As a pointer to further reading, it works. As a safe-or-toxic answer, it’s missing the variable that decides the outcome.

Where we’re being careful

Two honest caveats.

The hazard-versus-risk principle is settled toxicology. That part isn’t contested by anyone qualified.

But most of the published criticism of Skin Deep specifically is expert commentary and journalism, not a peer-reviewed audit of the database, and a meaningful share of it comes from people with cosmetic-industry ties. That’s why we’ve rated the evidence on this page mixed rather than strong. The underlying principle is solid; the specific quality judgement on EWG is an argument with participants on both sides.

The grown-up alternative

If you want to know whether an ingredient is safe at the amount in your product, read the bodies that answer in concentrations:

Cosmetic Ingredient Review (CIR) — US expert panel; publishes safety assessments naming the concentrations considered safe.
EU Scientific Committee on Consumer Safety (SCCS) — publishes opinions with exposure assumptions and margins of safety shown.

Both are slower to read than a number out of ten. They’re also answering the question you meant to ask.

Is the Yuka App Right About Your Skincare? — the same structural flaw in a different app, plus the concentration problem.
Parabens: What the 2004 Study Actually Found — a worked example of hazard, risk and one over-read study.
The Castor Oil Eye Trend — where the risk is real and no scoring app flags it.
Don’t Put Neosporin on Your Face — a documented top-tier allergen that scoring systems rarely treat as alarming.

Sources

Researched, not tested. This page assesses a rating method, not a product. Evidence rating is mixed: the hazard-versus-risk principle is settled, while the specific critique of Skin Deep rests on expert commentary rather than a published audit, with interested parties on both sides.