In my first diary in this series I used a technique used by FiveThirtyEight.com called Venn Diagrams to try and examine the Democratic primary.
In this diary I will look at the four early primary and caucus, and also adjust the circles in these states for the relative strength of each group. But 1st off let me state that this is data analysis and not an attempt to boost or disparage any candidate. The groupings aren't made to reflect, what "would happen if voters only knew what candidate's __ true position are and would do for them!!!1!!1!!" But rather what to the best of my abilities, I see as how the race shaping up as now. Feel free to disagree on the data, I would rather argue and discuss policy positions in a different diary (there are plenty of them). So yes this is a "horse race" diary.
Secondly, this diary isn't so much making predictions of who will win each of the 4 states, as it measuring the advantages of each candidates based on party composition. Candidates who have huge advantages have subsequently lost races, to better run campaigns and candidates.
Now let me recap my last diary, 538 used the Venn technique to try and understand the crowded GOP field, and developed the following chart:
This lead me to thinking could a similar diagram be developed for the Democratic party? Knowing the Daily Kos primary wars would eventually heat up, I decided to develop one (I petition 538 to do so numerous times) and then try and place candidates where I think they would sit. First of all here is my Democratic party Venn diagram chart:
The Democratic circles are:
Liberals: Self describe liberals, environmentalist (example: Sen. Russ Feingold)
Populist: Pro-union, fair traders, economic progressives (example: Sen. Sherrod Brown)
Moderates: Social moderates, economic moderates (example: Sen. Mark Warner)
Civil Rights Voters: Discrimination. immigration, LGBT issues, feminism (example: Rep. John Lewis)
Establishment: Major fundraiser, elected officials, etc (example: Rep. Chris Van Hollen)
Each circle represents a group of issues that is the "most important" to each group of voters. The reason the circles overlap is that most voters fall in more than one circle. For example many black church goers are big on civil rights, populism, but are social moderates (not socially liberal).. A big NYC fundraiser may be both a moderate and support of the establishment. I then made a subjective judgment and placed each candidate on the charts based on a few factors; where their primary support appeared to be coming from, where their campaign's appeared to be concentrating their efforts, my opinion on where each candidate's strength lie, etc. This isn't a measure on how I believe each candidates' positions, policies, or "heart" lies, rather it's a pure "pundit-like" analysis of where I see the candidate's campaigns are today. This also isn't an attempt to advocate or disparage any candidate within the primary and caucus races.
Lastly based on their national poll numbers I gave each candidate 1 font point for each percentage point in the national polls (yes this leaves most candidates with font sized that you would need a magnify glass to see). If you're a supporter of O'Malley, Chafee, or Webb, you can find their positions on my last diary. Other wise the font sized Venn shows this race as a two person contest.
Before I look at the four early primary and caucus states I want to make some updates to my model. I added environmentalist to the "liberal" circle. I also made a more advance chart that took in to account some of the concerns of where I place groups.
But I also should make clear, the relative positions of the circles vary from primary to primary. In some years 1984, the establishment backed a liberal (Mondale) in 1992 they backed a moderate (Bill Clinton). In some years (2008) both Obama and Clinton competed for Civil Rights (Obama blacks, Clinton Latinos, LGBT split), Establishment (Clinton had more, but Obama had a large portion) and to some extend populist. The graph would have looked something like this:
0
Obama 000 Clinton
0
Obama pulling anti-war liberals, Clinton pulling moderates.
But currently I would argue that Sanders is to the left of most of the establishment (not arguing that's good or bad, just that's what I think) and I believe the majority of the establishment Dems are backing Clinton.
Also different states can have different circles (and sized circles). For example Massachusetts and Maryland two very liberal Democratic dominated states would have circles of much different sizes. Maryland for example would have a much larger "Civil Rights" circle (Maryland is 25% black & 7% Latino vs Massachusetts 6% black & 8% Latino).
Now to my method on how to draw different sized circles. I will be relying on two sources. One is the following data chart by 538.com it shows the number of white liberals (as well as whites and liberals) in each state:
The size of my circle for liberals will come from this. The circle for Civil Rights voters will come from a combination of 3 numbers. The inverse of whites gives me the number for minorities and since 2/3 of POC (people of color) still view racism as a major obstacle in life, 32% of Dems view themselves as feminist I'll assume 2/3 vote on this, and since 8% of Democrats think of themselves as LGBT I will use estimate that 2/3 of this group vote strongly on this isuue.
In NV, Iowa, and NH I will use union membership as a proxy for populist voters (in SC I will have to take an educated guess, using John Edwards 2008 numbers), now remember since voters can (and usually are) in more than one category the populist numbers will include white, liberals, and minorities (people of color) . I will also make some assumption based on past performances of candidates to measure establishment numbers (Kerry in Iowa in 2004 for example), once again, establishment numbers can and do include whites and people of color.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This is a great state for Sanders. Liberals are a majority of caucus goers (the only category that is a majority). Populist which I estimated by using union membership and adding an additional amount based on polls of resistance to TPP (they broke this down by region not state so I'm using an average of the Midwest) is also large. Hillary only has a large circle of Moderates. This should also be a good state for Bernie to get "Civil Rights voters". Iowa is predominately white, and the mix of CR voters will be more heavily feminist and LGBT issue voters. I think Hillary will have a stronger pull on self described feminist, but I don't believe either candidate will have more sway with LGBT voters. Based on my positioning I think Bernie has a slightly greater than 50% advantage in Iowa, as he has a larger group predispositioned to his positioning. If there was a third candidate more to "the center" of the Democratic field I would actually have made him the prohibitive favorite, I expect Sanders to continue to gain momentum in Iowa.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
At first glance this state looks almost as good for Sander's as Iowa does. Slightly more liberals, although slightly less populist than Iowa, but the overall ratios are quite close. There is also a similar demographic of Civil Rights voters. In fact 538 considers NH "Bernie country" much like they do Iowa. But on second look you see a much larger number of Establishment Democrats. In fact NH has a history of backing Establishment supported Democrats (Dukakis, Clinton, Gore, Clinton). My guess is that we could see a reprise of 2008. Sander will leave Iowa with momentum, polls will show a virtual tie, but Clinton will leave with a NH narrow victory (similar to her 2008 victory of Obama). The presence of this Establishment circle, to me is the best fit for the string of victories by candidates listed above. I'm predicting a close race, but with a narrow, but solid advantage for Clinton. (Note: because I have worked in this state for so many years, I did step slightly outside "my box" and make a prediction)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This is the first state where Moderates will outnumber Liberals. This would prove much more problematic for the Sanders campaign if not for the heavy union presence, boosting the populist numbers. Nevada is one of the nation's most heavily unionized states (it may actually now be THE most unionized state). There is also less of a presence of Establishment Dems even with it being the home state of Harry Reid (note: I made an estimate of their numbers based on 4 Nevada contacts I made at NetRoots Nation, but this number really is a best guess, not enough historical data points). But team Clinton currently has two large advantages. Civil Rights voters swell to nearly 1/2 of Democrats. Nevada is one of 5 states with double digit Latino and African-American registered voters (also substantial Asian and Jewish communities), this combined with self ID'ed feminist and LGBT voters makes this a large circle. Also as mentioned above Moderates outnumber Liberals 5 to 4. Unless Sanders makes inroads with Civil Rights voters by the time of the Nevada caucus, Hillary will have a decisive advantage.
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Where to begin? This is the first state where people of color will outnumber white Democrats. Combined with a party that skews more female (white Southern males are scarce here) and your feminist numbers are slightly above average. This combination means 2/3 of voters here will be civil rights voters. Moderates outnumber Liberals 5:4 in South Carolina. There is a decent populist vote despite low unionization (John Edwards has done well) but there is also a good sized establishment presence (John Kerry has also done well here). If Sanders hasn't made inroads with civil rights voters by this time, this state will give Hillary a decisive advantage.