I first propose to distinguish between two kinds of external validity inferences, predictive and explanatory. I then argue that we have a satisfactory answer to the question of the conditions under which predictive external validity inferences are good. If this claim is correct, then it has two immediate consequences: First, some external validity inferences are deductive, contrary to what is commonly assumed. Second, Steel’s requirement that an account of external validity inference break what he calls the ‘Extrapolator’s Circle’ is misplaced, at least when it comes to predictive external validity inferences.

John Worrall and Nancy Cartwright have both argued that randomized controlled trials (RCTs) are “testing the wrong theory.” They claim that RCTs are designed to test inferences about the causal relationships in the study population, but this does not guarantee a justified inference about the causal relationships in the more diverse population in clinical practice. In this article I argue that the epistemology of theory testing in trials is more complicated than either Worrall’s or Cartwright’s accounts suggest. I illustrate this more complex theoretical structure with case studies in medical theory testing from Alzheimer’s research and anticancer drugs in personalized medicine.

The standard philosophical definition of placebos offered by Grünbaum is incompatible with Cartwright’s conception of randomized clinical trials. I offer a modified account of placebos that respects this role and clarifies why many current medical trials fail to warrant the conclusions they are typically seen as yielding. I then consider recent changes to guidelines for reporting medical trials and show that pessimism over parsing out the cause of “unblinding” is premature. Specifically, using a trial of antidepressants, I show how more sophisticated statistical analyses can parse out the source of such effects and serve as an alternative to placebo control.