But even though MWI is "correct" in some sense, it has a similar problem with its pedagogy as the more common Copenhagen interpretation (measurement is special, wavefunction collapse is a real physical phenomenon) which causes almost as much confusion: MWI is usually presented as "When you make a measurement, what was previously one universe before the measurement splits into two universes after the measurement." That's wrong. Nothing "splits" during a measurement. There are exactly "as many universes" (I put this in scare quotes because the line between universes is not sharp and so the number of them is not well defined) before a measurement as after.
For a more detailed critique of MWI see http://blog.rongarret.info/2009/04/on-shadow-photons-and-rea...