Comment by peterevans

You're right, and they do acknowledge this to some degree in section 5 ("Threats to Validity"). The projects they monitored were unaware of having been so, therefore the authors would have had no way to observe much nuance. They could only what was publicly visible.