Privacy concerns are the most important reason. A more cynical reason might that the odds that a papers is deemed 'ok' is far larger if the raw/anonymous data i

Privacy concerns are the most important reason. A more cynical reason might that the odds that a papers is deemed 'ok' is far larger if the raw/anonymous data is not provided. Remember, the authors of the suspect studies provided their data voluntarily, and in the end it only hurt their reputation/impact.