This is a perfect example of how many variables can work to change the pressure and velocities obtained with different components or even the same components and variations in powder charges. There are more things at work here. I'm sure that "Most" of the data was accurate for the loads that they tested and with the exact components they were using. Change a lot number of primers, powder, wads, or even hulls and shot, deeper or shallower crimp, etc, and you have a good explanation for the variations. Seating a wad deeper due to a change in the powder charge can also have an effect on the crimp, which could in turn affect pressures.
What ZZT stated about the depth of the wad seating is a reasonable explanation for light loads developing higher pressures than expected, but mostly in a tapered hull, or one such as the newer Winchester AA.
A few years ago some forum participants pointed out some suspected errors in the Alliant loading data. They were very quick to correct the data on their site. I'm sure that all of the previously printed data will automatically change itself to reflect those changes as well.

Always double check your data if you are in doubt. Tjis is the reason I buy so many different loading manuals. I use them for a reality check when selecting loading data.
This is why it's best to stick with established loads and to be conservative when selecting loads. I don't like to use loads that are listed at higher pressures than about 9500 PSI or 9000 Lups. I might go "slightly" higher, but not by much. Change anything and you have the potential to increase pressures dramatically. They can also drop dramatically, but you have no way of knowing which way they would go unless you have the proper testing equipment available. This is why I don't recommend swapping components around without some data to go on. I'd rather pass up a load than to risk damage to my firearms or personal injury.