fine.  good to have separate sub-sections for the separate research questions.

It would be good to have a table of the fitted values, SD, and so forth for final model(s) in the main text.  This is something that most readers (even less technical readers) might expect to see.

references look good.  I would just say, put the years in parentheses.  Also it is not necessary to spell out first names; initials will do.

fine.  you might find different answers to question 2 if you first perform the variable transformations that were indicated by your initial exploration of the data.

generally fine.  A good way to organize this section would be to have separate sub-sections for each of the four separate research questions.  The title of each subsection can summarize the research question to which the methods being described in that subsection will be applied.  

generally fine.  A good way to organize this section would be to have separate sub-sections for each of the four separate research questions.  The title of each subsection can summarize the research question being addressed.