Some of the factors involved...

The big challenge in dealing with all the data I have collected is to figure out what factors are relevant.  In analyzing the recorded speech of my volunteers, I have worked out how to pull out the following pieces of information: duration of the vowel (in milliseconds); duration of the vowel + /r/ for those words ending in /r/ (it can be tricky to separate the /r/ from the vowel); the first three "formants" (or harmonics) of the vowels at nine different points throughout the vowel, giving a "snapshot" of the vowel at 10% of the way through, 20%, 30% ... 80%, and 90%.  That's the raw data, plus general characteristics of the words (what word is it, what is the vowel, what is the consonant) and of the speakers (they are all Manitoba-born English-speaking females, of slightly varying ages).

To all of that I have been able to add some information about lexical frequency.  When I first developed the idea of this topic, I suspected that lexical frequency might be an important factor in whether or not raising would occur for a given word or not, so I thought I need to include it.  For some time I wasn't sure of the best way to do that, but I recently found a great website developed by Prof. Mark Davies that allows anyone to go and search through some fantastic corpus databases.  All told, the English language corpora have an astounding 560 million words.  By searching through and recording the hits, I am able to add "hits per million" to the information I have for each word in my data, and hopefully I will be able to correlate this with all the other information I have.

That's pretty much where I'm at right now, playing with the numbers, creating graphs, and trying to see which of these factors are important, and how they are related to and affect one another.