Are papers using experimental data trustworthy? since you can never know whether