After being incapable of determining why I was getting a 25% bonus to research rate with every race in every game, I tried loading up a map with every research rate to test and found the following values:
Very slow: -50%
Slow: -25%
Normal: +25%
Fast:+33%
Very Fast:+50%
This makes no sense to me whatsoever, and should logically by something like:
Very slow: -50%
Slow: -25%
Normal: 0%
Fast:+25%
Very Fast:+50%
I assume that these values are bugged, but I would also ask that the tool tip be updated in the galaxy generation page to reflect what each of these levels actually *does*, rather that requiring the player to figure it out from hover tool tips in game. I also think that as a system for adjusting research this particular solution is terrible. Since the modifier is an additive penalty on research rather than a multiplicative penalty it unjustly harms civs with research maluses while harming civs with research bonuses less. It should be a static multiplier on all research, not an additive bonus/malus.
For example if a civ has +20% research and another civ has -20% research, and they play on a 'very slow' research rate galaxy, at the start of the game their effective research rates are (100%+20%-50%) = 70% vs (100%-20%-50%) = 30%. So the faster research race now has a 70/30% advantage, or researches 2.33x as fast.
On very fast research rate the math is: (100%+20%+50% = 170% vs (100%-20%+50%) = 130%. Now the disparity is 170/130, or 1.31x as fast.
Alternatively, if this global research modifier were a static multiplicative bonus/malus to research it could be:
(100%+20%)*.50 = 60% vs (100%-20%)*.50 = 40%, which is 1.5x as fast, which is exactly what you would *expect* the difference to be between a race with 120% normal research rate versus a race with 80% normal research rate.