Some effects diminish when tests are repeated. Jonathan Schooler says being open about findings that don't make the scientific record could reveal why.