I deal with the collective, not cases-in-point. I've seen way too many posts here from people were doing well on practice tests and then scored worse.
As a statistic describing retakers overall, yes, but you can't make the assumption that it's actually the case for a given test taker. If an individual averages over 40 (for example) on all AAMC practice exams, then some extraordinary circumstance out of their control occurs that causes them to underperform and score a 32, I would argue it is in fact far more likely that they would increase their score drastically upon retaking. Regardless, there's no data specific to these circumstances so there's no way to say. The only information available for such situations (and the tester's decision) is that their practice average was much, much higher while their score was an outlier in comparison. If anything it's encouragement in deciding to retake.
I get paid to judge. A score in the low 30s is still competitive. A retake for a better score, while we can understand the rationale, still means the applicant is a risk taker who takes the risk on ill-founded assumptions. You assume that people are actually doing something intelligent like practicing and prepping. Yet, I see post after post here from people who simply retake because they expect magic to occur and get a higher score. I've interviewed scores of people like this with lower scores, who do the same thing, and I have no reason to think that judgement rises with a first-time MCAT score.
Of the people I've interviewed with 32+ scores who retook, the answer is always the same: "I thought I could do better". These are people who are not happy with a 90th+ percentile score! We're scared of these people. They're perfectionists, and in a bad way. They actually do poorly in medical school because they try to learn everything, and end up learning nothing. Either that, or they're constantly in our offices arguing over why their 95% on the anatomy exam should really be a 97%. Don't believe me? Ask gyngyn.
The poor choice being retaking? Or choosing to sit for the exam (or not void) despite extraordinary circumstances? If the former, I've addressed that above, and if the latter, without knowing the circumstances you can't judge the choice.
Why is such an assumption necessary and why does it matter? If an applicant scored a 33, and retook and scored a 25, you would have confirmation that not only was the 33 not a low estimate of their true score (whatever score perfectly represents what the test is trying to measure in them) but it was in fact a high estimate and their true score is statistically more likely to be below 33 and closer to 25.
Again, someone not happy with a 33 is someone who gives us pause. My Dean likes high MCAT scores, so he'll take these people, because he doesn't have to teach them. But at some MD schools, gyngyn's for one, a retake is a strike-out.
If the applicant retook a 33 and scored a 40, all this tells you is the 33 was an outlier and a low estimate of their true score, which is clearly closer to 40 (I argue this specifically because 40 is outside the 99% CI for a 33 and so the first score can be assumed an unrepresentative outlier when considering the scaling of the exam (i.e. you're more likely to get a lower score by chance than a very high score such as a 40)). You have no way of knowing A) whether the student's choice to sit for or score the first exam was a poor choice or B) whether the students choice to retake was a poor choice because you have no contextual knowledge about the situation.
