They don't embarrass easy.

We have written harshly in the past ("The Week," July 19) about the EPA's manipulation of science to justify its imperial agenda. As if control over rivers, wetlands, trash, chemicals, and outdoor air were not enough, the agency seems determined to gain authority to regulate indoor air as well. It is revealing that some of the harshest criticism of the EPA's classification of environmental tobacco smoke (ETS) as a Group A carcinogen has come from scientists within the EPA.

Representative Tim Valentine, chairman of the environmental subcommittee of the Science, Space, and Technology Committee, has released a bunch of internal EPA documents showing how the political operators there tried to ignore the criticisms. An April 27, 1990, report from EPA's Environmental Criteria and Assessment Office (ECAO) says that studies of the health effects of second-hand smoke reflect limited evidence of human carcinogenicity and recommends that it not be classified as a Group A carcinogen. The ECAO scientists recommended major revisions of the EPA report, given the inherent limitations of the data and the comparative novelty of the approach used to interpret the data. That novelty consisted of a huge array of statistical manipulations including exclusion of major studies that inconveniently indicated no health effects, suspension of usual standards for avoiding chance association (confidence intervals), and failure to consider confounding factors.

A March 23, 1992, review by ECAO epidemiologist Patricia A. Murphy characterizes the EPA report as poorly organized and badly argued, and says that the case for a causal relationship between lung cancer and ETS is overstated. She notes there is no consideration of what epidemiologists call the "file-drawer problem" (the tendency of researchers to leave in the file drawer reports of research that finds no-relationship). Thus, according to Dr. Murphy, the process of "meta-analysis," which tries to weigh the results of all the published reports, is inherently biased in the direction of finding a relationship. The report on ETS was entirely meta-analysis of a selection of old studies that had found no more than a weak relationship between ETS and disease.

One ECAO reviewer says part of the draft EPA report is plagiarism from an article in a 1990 issue of the journal Risk Analysis. He adds: "This could be a source of embarrassment, as there are surely many people familiar with this article who will also be reading this EPA document." But the EPA apparently is beyond embarrassment.