The author describes how coding errors can be avoided by cross-checking the code with the same data before publication and publishing the original data.
After noticing a possible error in a scientific study, a researcher repeatedly requests access to the study's original supporting data without success.