This is the issue using a word like knowing - it has all sorts of baggage philosophers will argue about endlessly.
The more precise answer I will give, that avoids such issues, is physics is a mathematical model. In QM we know the why of that model - it has to do with what are called generalized probability models - QM is just the next simplest one after ordinary probability theory. Now what does it mean. Physicists have been trying to answer that one for a long long time without success. We have all sorts of possible answers, like Many Worlds, but unfortunately no way to experimentally tell the difference. This is hardly surprising since they were all concocted to be the same as the formalism which everyone agrees on, and as I explained we have a rather good idea of why it is that way - mathematically.
So - where does that leave your query - sorry to say - nowhere - its bogged down in so much philosophical baggage the exact answer is often unsatisfactory (ie a generalized probability model) or when we try to be more specific - as the other poster said - we don't know. Its maddening - but is not the only area like that. Although not as often discussed ordinary probability theory is exactly the same - but I will let you investigate that one yourself. In fact,
John Baez thinks arguments about it are often simply the same arguments about probability in a different setting:
http://math.ucr.edu/home/baez/bayes.html
This is hardly surprising since we know QM is itself a generalized probability model so carries exactly the same issues with it - plus some others like Bell inequalities peculiar to QM.
Thanks
Bill