Copyright © 1995, 1997 by Steven J. Milloy. All rights reserved. First edition. Published by the Cato Institute, 1000 Massachusetts Avenue, N.W., Washington, D.C. 20002. Library of Congress Catalog Number: 95-72177. International Standard Book Number: 0-9647463-2-8.
The Lexicon Attributable risk n. The big risk number that puts you in the national spotlight.
Background risk n. The risk caused by birth. Currently estimated to be 1 in 3 for cancer. To be ignored because it likely dwarfs any risk you could possible find.
Benign tumor n. A noncancerous tumor that toxicologists pretend is cancerous.
Bioassay n. A laboratory experiment in which the risk of premature mortality for the experimental animals is 100 percent.
Biological plausibility n. A fantasy ostensibly based in biology that explains why a statistical association represents real risk.
Case-control study n. A type of epidemiologic study conducted by type A personalities... those who need results NOW!
Cohort study n. A type of epidemiologic study conducted by type B personalities... those who have got lots of time to kill, maybe 20 years or more.
Confounding risk factor n. A competing risk factor to ignore or belittle.
Data dredge v. To analyze a set of epidemiologic data every conceivable way to identify a positive statistical association between an exposure and a disease. Takes advantage of the fact that if enough analyses are conducted, a positive statistical association will turn up by sheer chance.
Epidemiology n. A methodology in which randomly occurring statistical associations are elevated to a cause-and-effect status.
Linear nonthreshold model n. A risk assessment theory that assumes if something poses a risk at high levels of exposure there is a risk at any level. Why risk assessors compare a dental X-ray to Hiroshima.
Meta-analysis n. An epidemiologic technique for turning a lot of nothing into something. compar. alchemy.
Negative data n. Conflicting research you'll need to explain away, preferably by belittling it.
Peer review n. Pre-publication congratulations from your colleagues. obs. An impartial, pre- publication review of the scientific merits of research.
Public health n. What we have now that the average American's life span exceeds 75 years. orig. What they didn't have in the 14th Century when one-third of all Europeans died from bubonic plague. Profession responsible for ensuring the U.S. public doesn't connect these facts.
Publication bias n. The tendency in the public health community not to publish studies inconsistent with the goal of finding risks.
Recall bias n. A phenomenon of case-control studies where subjects remember what epidemiologists want them to remember — even if it never happened.
Relative risk n. A type of statistical association routinely used in epidemiology. Actually has nothing to do with health risk. What you get when you merge a statistical association and ambition.
Risk assessment n. A recipe for success in public health. obs. a four-step process to identify and quantify hazards to human health.
Statistical association n. A highly prized numerical relationship between the exposure you're interested in and a disease. An omen of success.
Statistically significant adj. Confidence that a statistical association did not occur by chance. The standard level of confidence is 95 percent, but the ambitious may resort to lower levels of confidence (like 90 percent).
Strong association n. A relative risk greater than 3.0. The Holy Grail of public health research.
Toxicology n. The study of poisons. In risk assessment, a process where animals are poisoned and then sacrificed to see if the poison worked.
Weak association n. A relative risk between 1.0 and 2.0. Requires a snow job of unbelievable proportions to sell this as a genuine risk.
Click here to return to the home page
Copyright © 1995, 1997 Steven J. Milloy. All rights reserved. Site developed and hosted by WestLake Consulting.