Many political scientists quickly concede that experimental research has high internal validity compared with research with observational data and they dismiss experimental research (especially laboratory experiments) as being low on external validity compared with research with observational data. Both opinions tend to understate and ignore the multitude of issues involved in establishing the internal validity in the senses we have mentioned – statistical conclusion, causal validity, and, if the empirical study involves theory testing, construct validity. If a result is not statistically significant, cannot be established to be causal in the population originally investigated, or is estimated from an empirical study that has little relevance to the theory being evaluated, then how can it possibly be considered externally valid or robust as a causal relationship? It makes no sense to say that some empirical research is low on internal validity but high on external validity.
Side 275 i Morton, R. B. & K. C. Williams (2010). Experimental Political Science and the Study of Causality: From Nature to the Lab. Cambridge: Cambridge University Press.