Grad students might do that. Actual scientists have had that tendency beat out of them by their Advisors.
I think the tendency returns when the effort becomes collaborative, especially when an entire community goes along. If climatologists are going to do anything concerning the study of climate change, then the community as a whole must accept that certain computer modeling is reliably predictive, "I'll accept the validity of your modeling, if you'll accept mine." Without this, climatology goes nowhere. Once this is done, it is professionally suicidal to start attacking the reliability of the method.
How else do you explain the inaccuracy of short term hurricane modeling from the same scientists that state that they can reliably predict a 1 degree average change over 100 years? The North Atlantic hurricane season is simply a matter of retained heat (or so we believe), and predicting 1-5, 5-6, 6-8, etc. storms shouldn't be too difficult if the modeling is good. Yet, forecasts are not historically any better than throwing darts.
When hurricane season forecasts become better than random chance, I'll start to believe that the modeling is on track.