It would be easy to say lot-to-lot discrepancies account for the differences in data, but if that was true, the discrepancies were often indeed huge ones.

Different pressure equipment, perhaps calibrated differently, with results read and interpreted by operators with varying degrees of skill may have something to do with all this. That might mean actual lot-to-lot differences were much smaller than what they may seem. I think acceptable allowable differences in powder lots sold for handloading are somewhere on the order of around 3% maximum. Perhaps that hasn't always been the case.

As I recall, it's only been common knowledge for the last 10 -15 years that HP-38 and 231 were the same.

One can look up old HP-38 and 231 data in handloading manuals or such excellent works as Ken Waters PET LOADS series. Waters, though long-retired, was among the very best of the handloading writers during the past fifty years or so. However, his data was not pressure tested. He used traditional, accepted methods ("eyeballing") for determining safe / maximum loads. Some such methods are still recognized today, some aren't.

About the only way for handloader to make a safe determination is through real education (not "Internet instant results") and the use of a chronograph.

Pressure-tested data in recent and fairly recent manuals should be more accurate than some of the older stuff. I'd be inclined to go with such data and verify by chronographing. To further add to the confusion, look at the current Hodgdon data for HP-38 / 231: charge weights identical, but much lower than Lyman. Perhaps coincidentally, perhaps not, my old notes show powder charges of 231 .38 Special loads and #358311 are midways between Hodgdon and Lyman data.

Most importantly, remember this is the Internet and I'm no expert. Good luck-