Katches had said specifically that he wanted to know what Rust thought about BPA. She was inclined to believe vom Saal’s claims that BPA was dangerous to humans, and that only industry-funded studies found otherwise. But how could she prove it? As a first step, she decided to return to the PubMed digital archive, where she searched through existing studies on BPA. It yielded some 7,000 studies—too broad a sweep for her to feasibly tackle on her own.

Criteria . By skimming through several papers, however, Rust began to identify possible parameters for her investigation. Any studies she examined closely, she decided, should be as applicable to humans as possible. So she set two main criteria: First, they should have been conducted in a controlled, laboratory environment as opposed to in the wild; Second, they should have been performed on vertebrates, or animals with spines. Rust compiled an electronic spreadsheet of her “short list.” The database kept track of each study’s title and its author, and linked to the abstract for each paper. Importantly, she also included a column on the funding source for each study.

The new limits dramatically scaled down the number of papers for Rust to wade through to a still significant, but manageable, 258 studies. Through her husband, a University of Wisconsin doctoral student, Rust had access at the university to a range of online, by-subscription-only academic search engines and specialty science publications. The Journal Sentinel did not subscribe to such special-interest publications. In early October, Rust hunkered down on campus for two weeks, pulling up studies and reading through documents for hours on end. As she worked, new research details and caveats began to emerge.

It became clear, for example, that to compare studies directly was difficult because their methodologies varied. Different studies used different animals, administered different-sized doses to their subjects, and used a variety of dose-delivery techniques. While some tests gave the animals doses in food, for example, others had used water, injections or even small pumps beneath the skin. Nor did all of the studies mention the type of feed, bedding, and cages used to house the animals. Rust kept track of these elements as she built her spreadsheet so that she would be in the best position possible to compare like with like.

By mid-October, Rust was starting to think that vom Saal was in fact right. Most of the studies did seem to find BPA harmful. Moreover, those that found it benign were almost universally funded by the chemical and plastics industries. Academic studies indicated that even small amounts of bisphenol A could cause serious damage in humans. Of the 258 studies that Rust reviewed, 168 examined low-dose effects of BPA. Of these, 132 found health problems. All but one had been funded by sources outside the chemical industry. By contrast, nearly three-quarters of the studies that concluded BPA was safe were industry-funded.

But there was more. In mid-November, a government program issued a draft report on bisphenol A. Rust hurried to read it.