optionally standardize and/or center raw data before qfwe

[maxheld83]

most factor scoring methods standardize raw data before multiplication (we currently don't – we standardize only after multiplication).
This doesn't matter so as long as the raw data are forced and symmetric, in which case the sd and means are all the same anyway.
When that isn't the case, pre-multiplication might matter.

Also, we might want to offer non-centered results (because that can be meaningful in free distros).

Consider also Thomspon 2004b (who has a suggestion for non-centered standardization) and DiStefano et al.