Scientific Evidence Portal
Can observed associations with health be artefacts of measurement error? | P.N. Lee
Article Published: 1983/11/10
Type: Articles and Dissertations
Further Information These notes by P.N. Lee are about 25 years old. Then as it is now, it was clear that studies on ETS are plagued with all kind of problems, and they are completely unreliable. In only eight type-written pages, Lee provides several quite educational descriptions of how an epidemiologist examining ETS can come up with a variety of false and inconclusive results even with the best of intentions.
“In many situations in epidemiology, error in measurement of exposure variables has no effect on the direction of an association, serving only to make the magnitude of an observed association less marked than the association that actually exists. This is not always so, however, and the purpose of this note is to highlight some situation's relevant to the passive smoking story in which, as a result of error in measurement of exposure variables, an apparent association can occur when no real association exists at all.”
In his examples on how uncertain conclusions can be reached, however, Lee does not seem to zoom into the one and fundamental flaw of all and any epidemiological study on passive smoking (and not only passive smoking), regardless of their results: self-reporting based on . A self-reporting that is not in the least verifiable, or measurable, and thus open to all kind of bias in all directions. questionnaires
The first incredibly dishonest point is, therefore, that un-measurable assertions that no one can prove or disprove – including the persons doing the self-reporting! – are entered as
hard data and precise numbers. The second incredibly dishonest point is the mass-production of this statistical garbage, and the worrisome belief by many “experts” that, when hundreds of junk studies with the same fatal flaws are produced, the errors somehow magically disappear, or cancel each other out. For example, Lee brings the case of many smokers who, victimized by the stigma induced by the antismoking propaganda, misrepresent themselves during the interviews as non-smokers exposed to ETS – a problem common to ALL epidemiological studies on passive smoking. According to either incompetent or dishonest “public health” institutions, however, by having hundreds of studies once again the problem disappears, or becomes negligible, and the invalid results become collectively valid. The logical conclusion of this reasoning is the all-Nazi concept that
a lie is a lie if spoken by just one thousand people, but it becomes the truth if it is repeated by one million. The third incredibly dishonest point is that those hundreds of flawed “studies” constitute the so-called “unquestionable mountain of evidence” that is at the base of smoking bans and social demonization of harmless smokers all over the world.