Abstract

Researchers often aim to make correct inferences both about that which is actually studied and about what the results generalize to. The language of internal and external validity is not used by everyone, but many of us would agree that intuitively the distinction makes a lot of sense. Two claims are commonly made with respect to internal and external validity. The first is that internal validity is prior to external validity since there is nothing to generalize if the findings obtained in, for instance, the experimental setting do not hold. The first claim is explicit in many writings. See for instance Francisco Guala’s influential book The methodology of experimental economics. And it is often implicitly relied on. The second claim is that researchers have to make a trade-off between internal and external validity. When one is increased, the other will decrease. The second claim was made already from the start by D.T Campbell in his classic Factors relevant to the validity of experiments in social settings. There is a certain tension between the first and the second claim. It has been argued before that it might be difficult to combine them. We intend to make the stronger point that both claims are misconstrued. Our hypothesis is that the relationship between internal and external validity has to be re-conceptualized, and we will briefly indicate how