I don't know what a proper sample size would be, but those results aren't necessarily out of line with what you might expect. Probability is really unintuitive.
Here's the block of code that controls escape chances that a dev posted awhile back:
It's still possible that there are bugs in it that could end up affecting the probability. For instance, if the DailyTick triggers more than once in a day for some reason, it would lower the time to escape by about half. Or, if the game doesn't properly classify the prisoner's state, then it might use the wrong numbers. Nothing really stood out to me as unusual from my own anecdotal testing though. Here's a sample of some npc lords that were captured by other npcs and placed in npc dungeons in my game:
I didn't cherry pick these examples, they were just the cases that I found in Vlandia within a short time period. There could still be some selection bias though, because I would be more likely to find imprisoned lords if they had been in prison for a longer period of time, which could skew the imprisonment time toward the high side.
Edit: I will say this though, I donated my own prisoners to npc settlements a few times and they
did seem to escape faster than you might expect. A couple of them at the very next daily tick, so there
might actually be an issue with that scenario.