Comparing Contemporaneous Laboratory and Field Experiments on Media Effects

Experimenters often conduct their studies in the laboratory or in the field. Each mode has specific advantages (e.g., the control of the lab versus the realistic atmosphere of the field). We develop two hypotheses concerning the relationship between treatment effects in lab and field settings that are tested in contemporaneous experiments. In particular, registered voters in a medium-sized city were assigned to a laboratory or a field experiment involving newspaper treatments. The analyses show significantly larger treatment effects in the laboratory experiment, especially for public opinion outcomes in which the content of the treatment could be readily linked to the dependent variable. There also is suggestive evidence that as lab and field experiments become similar on one particular dimension (the temporal distance between stimulus and outcome), differences in the size of treatment effects moderate.