Well, that's very true in Physics, Biology, Chemistry, etc. But the issue is that a lot of climate study work is based around these data sets, which often take a long period of time, and lots of money to get. If we had to wait another 20 years for another data set (which would be insufficient), we would have already made a choice, possibly a bad one. Worse, it seems the only way to get the necessary money these days is to tow the AGW line before starting, or side with those for which AGW would be a disaster (oil companies), neither of which seems academically honest to me.
Besides, unless if you royally messed up the collection of data, collecting a new set would be a waste of time. What needs to be done is a testing (not verification as they say) of their models, and a checking of their statistics. Claim all you want, but I think that'd be a lot easier with raw data as well. I'd like to see proof that the data was actually destroyed in the 1980's, they were talking about it in their emails more recently.