Eharding----
I am a little confused about the exact original question that is being posed: "What were the best predictors of the White HRC Primary vote?"
Firstly, "White", is a bit problematic in terms of identity in an extremely multi-cultural society, where there are many individuals who have family backgrounds from a variety of different races/ethnicities that don't self-identify as part of only one "check box" on a Census form or voluntary self-identification form when applying for a job. Additionally, many Latinos might self-identify as White or African-American, while still identifying as Latino because of Country of origin, shared primary/secondary language, etc....
Secondly, Are we talking the word "predictor" --- One definition being- "something such as an event or fact that enables you to say what will happen in the future"
http://dictionary.cambridge.org/us/dictionary/english/predictorOr are we talking about a Predictor Variable, commonly known as an independent variable?---
https://en.wikipedia.org/wiki/Predictorhttps://en.wikipedia.org/wiki/Dependent_and_independent_variablesSo if we are talking about what results were the best predictors of the HRC White Primary Vote overall, than M&Ms answer certainly sees to show some directed correlations, although it should be noted that there significant regional/statewhite primary variations....
If look at the very first primary caucuses----
Iowa- indicated a significant weakness among Blue-Collar Whites in places like (Black Hawk County, Linn County, and the towns along the Quad Cities area in Eastern Illinois, and among University Students (Johnson County), but also strength among White Collar workers in Polk County (Des Moines).
NH- indicated that her weakness among White voters from a wide variety of economic backgrounds was not just limited to a Midwest caucus state Iowa, but was also the symptom of a much deeper problem among Independent voters as well...
SC- Indicated that she still held significant strength among Southern White Democrats, if we look at the far NW part of the state, as well as the wealthy suburbs of Columbia (Lexington County). So not only was Bernie getting slaughtered among Southern African-American Democrats, but also losing by significant margins among White SC Democrats.
Super Tuesday in the South, reinforced that pattern, even in counties/areas where one might have expected Bernie to perform better in heavily White Counties.
I'll let someone else speak to the Super Tuesday New England states, that can run down these numbers by Township and better versed to speak to that subject regarding the Clinton electoral primary coalition in MA & RI, and the implications for the GE performance...
Michigan- This should have been a giant warning sign to the Clinton campaign, both as a result of the collapse of her support in rural parts of the State, a less than spectacular performance in Oakland & Macomb, and Wayne County not delivering the types of numbers expected.... Trump winning MI in the open primaries should have drawn even more attention in the GE. Trump captured 48% of the Rep vote in Macomb and 41% in Wayne!!!
Wisconsin---- If Michigan was a warning sign, Wisconsin should have been a four story alarm fire, considering that it has voted Republican for Statewide and Senate races in recent years, and looking the the Wisconsin map is almost as ugly as the Oregon Primary Map....
So,
1.) If we look at the overall "predictors" of HRC 2016 Primary performance using the definition regarding the predication of the outcome of the "White" vote during the 2016 Primary, then clearly IA/NH/SC/ Super Tuesday, and MI/WI pretty clearly indicated her relative strengths and weaknesses (Coalition) among these voters during the primaries. One could say "but what about Ohio?"---- that was an unusual situation because of the surge of cross-over voters in OH to back Kasich, that likely reduced Bernie's performance in many parts of Ohio (NE Ohio, SE Ohio, and other parts of the state)
2.) There is a direct correlation between the collapse of HRCs rural and blue-collar support in Northern States (Midwest springs immediately to mind) and her extreme under-performance among White voters in those same regions/counties/cities/ and even precincts
So are we talking about the predictors in the '16 Dem primary specifically or the predictors for HRC performance in the GE?