Criticize the following statement: The cause of tuberculosis is infection with the tubercle bacillus.
A trait in chickens called yellow shank occurs when a specific genetic strain of chickens is fed yellow corn. Farmers who own only this strain of chickens observe the trait to depend entirely on the nature of the diet, specifically whether they feed their chickens yellow corn. Farmers who feed all of their chickens only yellow corn but own several strains of chicken observe the trait to be genetic. What argument could you use to explain to both kinds of farmer that the trait is both environmental and genetic?
A newspaper article proclaims that diabetes is neither genetic nor environmental but multicausal. Another article announces that one half of all colon cancer cases are linked to genetic factors. Criticize both messages.
Suppose a new treatment for a fatal disease defers the average time before onset of death among those with the disease for 20 years beyond the time when they would have otherwise died. Is it proper to say that this new treatment reduces the risk of death, or does it merely postpone death?
It is typically more difficult to study an exposure-disease relation that has a long induction period than one that has a short induction period. What difficulties ensue because the exposure-disease induction period is long?
Suppose that both A and B are causes of a disease that is always fatal, so that the disease can occur only once in a single person. Among people exposed to both A and B, what is the maximum proportion of disease that can be attributed to either A or B? What is the maximum for the sum of the amount attributable to A and the amount attributable to B? Suppose that A and B exert their causal influence only in different causal mechanisms, so that they never act through the same mechanism. Would that change your answer?
Adherents of induction claim that we all use this method of inference every day. We assume, for example, that the sun will rise tomorrow as it has in the past. Critics of induction claim that this knowledge is based on belief and assumption and that it is no more than a psychological crutch. Why should it matter to a scientist whether scientific reasoning is based on induction or on a different approach, such as conjecture and refutation?
Give an example of competing hypotheses for which an epidemiologic study would provide a refutation of at least one.
Could a causal association fail to show evidence of a biologic gradient (ie, Hill’s fifth criterion)? Explain.
Suppose you are studying the influence of socioeconomic factors on cardiovascular disease. Would the study be more informative if (1) the study participants had the same distribution of socioeconomic factors as the general population or (2) the study participants were recruited so that there were equal numbers of participants in each category of the socioeconomic variables? Why?