Really, I always thought the scoring was normalized to everyone taking that version of the test on that date?
There are certain repeat questions (which is why AAMC does not allow us to discuss real MCAT questions) that show up between tests and among test versions to help normalize the tests.
For example, let's say the question was "1 + 1". The first 10,000 people got it 50% right. If it shows up in another test, and this set of 10,000 people get it 60% right, it may be assumed that this group as a whole is smarter than the first group, all other variables being the same.
The fact that people score higher in April means that, overall, people who take it in April are smarter or there is some temporal condition that helps them study better. And in conjunction, the people who bombed April try again in August and most of them bomb that as well (more retakers in August, retakers tend to score worse, and voila)
This is NOT to say that if you take it in April, you'll magically be smarter, and score better. It's possible that most people taking it in April are doing MCAT coursework in their normal lectures, and this helps. But you have to remember, a lot of people taking in April also know that they can juggle school, work, volunteer, ECs AND studying for MCAT.
So the best advice is to figure out your own strengths, and take it accordingly. For me, I knew I did not want to do this test twice, and I would not do at my peak if I took it in April, so I devoted the summer to studying and took it in August.