Methodological ignorance: A comment on field experiments and methodological intolerance

Abstract

Glenn Harrison [Journal of Economic Methodology, 2013, 20, 103–117] discusses four related forms of methodological intolerance with respect to field experiments: field experiments should rely on some form of randomization, should be disconnected from theory, the concept of causality should only be defined in terms of observables, and the role of laboratory experiments is dismissed. As is often the case, the cause of intolerance is ignorance, as it is here. To acquire knowledge about potential influences, which we need for both the evaluation of internal and external validity of experimental results, we cannot do without theory. A purely empiricist (inductive) methodology will be unable to give us sufficient understanding of the validity of these results. An account of causality only based on directly observed things, is an account based on factual influences only. This account will be too restricted, because it will not deal with the unobserved potential influences, which we need – again – for the evaluation of the internal and external validity of the experimental results. Every investigation in a laboratory that is feasible regarding a specific potential influence may lead to deep knowledge. It is simply scientifically irrational to dismiss this kind of knowledge.