Erin Brockovich (2000): Does the Science Hold Up?

Erin Brockovich is a twice-divorced, unemployed single mother of three who loses a personal-injury suit and then talks her way into a job as a file clerk at the small law firm of Ed Masry. Sorting through a pro bono real-estate file, she notices medical records filed alongside property documents for homes in Hinkley, a small town in California's Mojave Desert, and begins asking why. Her digging uncovers that Pacific Gas and Electric had used hexavalent chromium as a corrosion inhibitor at its Hinkley compressor station, and that the company had been in contact with residents about buying their homes. Going door to door without formal legal training, she wins the trust of hundreds of residents and assembles their accounts of illness into the basis of a mass action. The case is forced into binding arbitration, and in 1996 PG&E settles with more than 600 Hinkley plaintiffs for $333 million. The film ends on Brockovich's vindication and a large bonus cheque, framing the outcome as proof that the contamination caused the town's illnesses.

Spoilers throughout. Discussing the cancer storyline means discussing how this ends.

The contamination is real, documented, and not seriously disputed. California's Lahontan Regional Water Quality Control Board records that PG&E used hexavalent chromium to fight corrosion in cooling-tower water between 1952 and 1966, that the wastewater was discharged into unlined ponds, and that some of it percolated into the groundwater — leaving a plume the Board describes as at least eight miles long and two miles wide. A 2000 public health assessment concluded that the site had posed a past public health hazard. The $333 million settlement figure is accurate. The corporate-conduct story the film tells is therefore substantially true: a utility discharged a recognised human carcinogen into unlined ponds, it reached drinking-water aquifers, residents were not adequately informed, and a legal assistant with no credentials drove the accountability.

The step the film takes beyond the evidence is the causal one — from that exposure to the specific cluster of cancers and other illnesses its residents describe. It presents a settlement as a verdict, and a verdict as proof. It also flattens the exposure route, which turns out to be the crux of the science, and it gives the audience no way to distinguish 'this company did something indefensible' from 'this company caused that person's cancer'. Those are different claims and only the first is established.

This page exists because the honest version of this story is more complicated than either side of it, and the film is not the only thing that gets it wrong.

The contamination is established. The carcinogen is established. What is not established — and what the film treats as its climax — is that the contamination caused the cancers.

The case never tested it. Anderson v. Pacific Gas and Electric settled in 1996 for $333 million on behalf of more than 600 residents, and it settled before any adjudication of general or specific causation. A settlement is a negotiated payment to end litigation; it is not a finding of fact. The size of the payment is routinely cited as though it were proof, and it is not — settlement magnitude reflects litigation risk, the cost of defending hundreds of claims, punitive-damages exposure and reputational exposure, all of which can be severe even where the causation evidence is weak.

The claim advanced in the litigation was also far broader than the toxicology supports: a wide range of cancers plus non-cancer conditions, where the ingestion evidence points mainly at the gastrointestinal tract.

Two separate scientific questions are tangled together here, and pulling them apart is the whole point of this page.

Is hexavalent chromium a carcinogen? Yes — and the route matters

IARC classifies chromium (VI) compounds as Group 1, carcinogenic to humans, and the US National Toxicology Program lists them as known to be human carcinogens. Both classifications rest on the same body of evidence: occupational cohorts — chromate production, chromate pigment production, chrome plating — and the cancer is lung cancer, from inhaling it.

That is not the exposure in Hinkley, which was drinking water. On the oral route, IARC found little evidence of an association with stomach cancer and dismissed the single ecological study of contaminated drinking water in China as insufficient on its own.

The picture has since moved, and in the direction of more concern rather than less. A National Toxicology Program drinking-water bioassay reported after the film found tumours in the oral cavity of rats and in the small intestine of mice. EPA's IRIS toxicological review, finalised in 2024, concluded that hexavalent chromium is likely to be carcinogenic to humans by the oral route — note the wording, deliberately weaker than the "is carcinogenic" it uses for inhalation. California's own health goal for chromium-6 in drinking water rests explicitly on ingestion carcinogenicity.

So: drinking it is a real hazard, it justifies regulation, and it is a different and weaker evidence tier than the inhalation-lung link. Both of those things are true at once.

Did it cause Hinkley's cancers? Nobody has actually answered that

The California Cancer Registry assessed cancer incidence in the census tract containing Hinkley three times. The first, covering 1988–1993, found 114 cancers against 91.2 expected — an apparent elevation that did not reach statistical significance, and which the authors attributed largely to undercounting the population during a period of growth. Extending the period to 1998 brought it to 173 against 167.2. A third assessment covering 1996–2008 found 196 against 224.2, slightly below expectation, and concluded that the findings did not identify a generalised cancer excess.

Those results have been criticised by other epidemiologists on three grounds worth knowing about, because they apply to almost every cancer-cluster investigation you will ever read about:

• Exposure misclassification. The analysis pools everyone in a large census tract, without separating the heavily exposed from the barely exposed. If the worst contamination affected a small number of homes, diluting them into a whole tract can hide a real effect.
• Statistical threshold. The first assessment's elevation is significant at the 95% level that most epidemiologists use, and not at the 99% level the registry applied.
• Study period. The third assessment starts in 1996 — the year of the settlement, after affected families had been relocated.

The registry's own author has said his purpose was to address the misconception that all Hinkley residents had high cancer rates. That is a fair aim, and a null result at tract level is genuine evidence against a large community-wide excess. It is not evidence that no individual's cancer was caused by the water, and it is not the same as a study of the exposed cohort. No study assigning individual exposures and following those individuals appears to exist — and that absence, rather than any null result, is the honest reason this is unresolved.

Why this page is written this way

Cancer clusters are one of the hardest things in epidemiology and one of the easiest things to write a story about. Small populations produce noisy numbers. Clusters appear by chance and are found because someone went looking. And the two available narratives — corporate villainy proven, or worried residents debunked — are both more satisfying than the truth, which is usually that the exposure was real, the harm is plausible, and the study that would settle it was never done.

If that pattern is useful to you, how to read cancer statistics and news headlines covers the general version. For exposures where the human evidence is strong, asbestos is the contrast case — and air pollution and cancer risk shows what a well-characterised environmental exposure looks like.

Where the regulation stands

There is still no federal drinking-water limit specific to hexavalent chromium; EPA regulates total chromium at 100 parts per billion, on the basis that the forms interconvert. California adopted its own chromium-6 limit of 10 parts per billion in April 2024, effective from October 2024, with compliance phased in by system size through 2028 — an earlier California limit had been invalidated in 2017 on economic-feasibility grounds rather than on the science. California's non-enforceable health goal is far lower still, because a health goal need not account for what treatment costs.

A film is not a diagnosis, and a dramatised illness is not a guide to your own. What a story like this is genuinely good for is lowering the barrier to a conversation you might otherwise put off.

If something in your body has changed and stayed changed for a few weeks, that is worth raising with a doctor — not because it is likely to be cancer, but because most of the time it will not be, and finding that out is the point. Where screening exists for a cancer, it is the only thing on the list you can do before anything is wrong at all. See screening and possible warning signs.

Cancer Explained is free and ad-free. Nothing is sold on this page and no sponsor sits behind it.

If this helped you understand something, or moved you to ask a question you had been avoiding, you can help keep clear cancer information free for whoever needs it next. Support our work.

Erin Brockovich is a great film and a bad citation. The contamination was real, the chemical is a genuine carcinogen, and the causal claim at the film's centre was settled around rather than proven. Holding all three of those at once is harder than picking a side, and it is the only honest position available.

This page describes a work of film or television for education. It is not medical advice, and nothing here is a judgement of anyone's real medical care. Spotted an error? Please email [email protected].