Barry Haldiman has been litzing (digitizing) old puzzles for many years. In fact, he supplied the Shortz Era puzzles from 1993 to 1996. The NYT did not have any electronic versions of them until we supplied them to the Times.
In 2012, David Steinberg began a project to organize, digitize, and proof as many pre-Shortz NYT puzzles as possible. He gathered a team of volunteers and started The Pre-Shortzian Puzzle Project to accomplish this.
All the pre-Shortz data are kept separate from the Shortz Era stats.
Thanks to changes introduced by Will Shortz, modern puzzles are much different than older ones. That doesn't mean older crosswords are uninteresting; there are many examples of creative ingenuity as clever as you'll see anywhere, but stats based on these pre-Shortz puzzles are less relevant to modern solvers and constructors.
Maybe. Constructor names are notoriously difficult to keep straight. Pre-Shortz constructor statistics aren't always reliable.
Those are available and are often fascinating. Rules about valid clues were much stricter back then. This resulted in some complex cruciverbal acrobatics just to find something suitable for common words like OREO or ESSO. When you look up clues for a word or words to fit a pattern on the Finder page, you'll see results separated into daily, variety, pre-Shortz, and dictionary entries.
Contact David Steinberg at his blog.