Empirical evidence in the behavioral sciences often comes from experiments human participants. Researchers can generate better evidence by conducting more studies and using bigger samples. But when participants are paid, collecting evidence requires an investment of scarce resources. How do researchers balance these monetary and scientific considerations? In one study we use data from Amazon MTurk to show how a sudden rate increase affected data collection. Many researchers appear to have exited the platform in the face of higher costs. In a second study, we conducted a field experiment on another platform. Researchers who were given a discount used far more participants than those who did not. Researchers respond to the price of science by collecting more or less evidence. This work highlights a factor—financial resources—that is prescriptively irrelevant, but practically integral to the current debate on research practices.