I would agree with you that people who do well during MSI and II would be likely to do well during clinicals because from what I have heard a lot of your marks during MSIII have to do with how hard you work....and I think we all know that grades from MSI/II most strongly correlate with hard work vs. raw "intelligence."
However, the numbers I was talking about were pre-med GPA and MCAT and the statistical correlation between these numbers and student performance breaks down in the 3rd year of medical school. This doesn't mean that people who have a good pre-med GPA, MCAT, and MSI/II performance won't do well in third year. It just means that those with good pre-med GPA and MCAT won't necessarily do well in 3rd year, whereas they are fairly likely to do well in MSI and MSII. These things are just generalizations and do not necessarily apply to individuals.
I think people are interested in these things for one of two reasons: 1)To try to provide reassurance that they will do well in 3rd year because either 1. They don't have good numbers, but during 3rd year this doesn't matter anymore or 2. They do have good numbers and this is good proof that they will continue to do well
The second reason people are interested in these things is to try to provide proof that a certain group of students with lower entrance MCAT/GPA's do or do not belong in medical school...because if lower numbers correlate with inferior clinical performance then this can be used to strengthen anti-affirmative action arguments whereas if lower numbers do not correlate with clinical performance, then this can be used to counter anti-affirmative action agruments.
In my opinion, it is all a little ridiculous and everyone should stop obsessing over it.