Design experiments where you specify the falsification…

Design experiments where you specify the falsification criteria and record predictions before running the test, because memory reconstruction will otherwise align past beliefs with known outcomes.