There are valid reasons and there are invalid reasons. That's why being open about the raw data is important.
[Edit] I should have also mentioned that it's also important to be open about how the data is adjusted. Explaining your reason for adjusting the data is not enough.
Which reminds me. Someone commented that scientists don't share code. That's not quite true. In many areas, it suffices to share data because the analysis routines are already shared.
I fully agree! But both sides are guilty of subjective analysis. As to which side is most guilty... well, choosing either/or gives away your position on the issue.
[Edit] I should have also mentioned that it's also important to be open about how the data is adjusted. Explaining your reason for adjusting the data is not enough.
Which reminds me. Someone commented that scientists don't share code. That's not quite true. In many areas, it suffices to share data because the analysis routines are already shared.