The popularity of field experiments that utilize some form of random evaluation seems to be correlated with increased methodological intolerance. Since correlation is not causation, it may be useful to examine what this intolerance is, why it seems to have developed and how it can be defused. The intolerance takes at least four, related forms. First, there is an identification of the notion of an experiment with the use of some randomization. This is actually just a simple semantic confusion, but colors debate on many other issues. Second, there is an aggressive disconnect from theory, whether it be economic theory or econometric theory. Third, there is unquestioned worship to a narrow concept of causality defined solely in terms of things that can be directly observed. Finally, there is a dismissal of the role of laboratory experiments. I argue against all the four positions.
