Is there any statistically relevant way to combine multiple p-values measuring different hypotheses but that are all trying to measure the same thing "in spirit"? Eg, if I wanted to measure some teaching method's effectiveness, and there are a bunch of studies that test people before and after a particular class and present p-values. Each value is technically for a different hypothesis as the particular test and classes are different, but they're all trying to measure the effectiveness of this method.

Is there a way to numerically summarize the results of such studies? Eg, with a mean or median of the p-values? Or would something like that not really make much sense?