They can test theoretically derived hypotheses that cannot be studied in the lab.

They tend to be better at proving causality.

They can determine if effects observed in the laboratory generalize to the real world.

They have some methodological advantages over laboratory experiments.