I think the solution is elegantly simple. However, the problem with elegantly simple solutions is that they may lead to metric tonnes of work. Release all of the data. If data was excluded, show why. Data exclusion is a real part of science. Happens all the time - but in the actual papers I read the data that was excluded in the calculations was still included in the data sets, was identified as something to be excluded, and reasoning was given as to why it was excluded.

The problem with doing this, if you're one of the faith, is that all sorts of "unqualified" folks will look at the data and make their own calculations and conclusions. They'd then have to spend time defending against that. It's much easier for the faithful to follow their patron saint AlGore and just tell us What We Need To Do.