Back in mid-March, a couple of weeks after the primaries in Texas, Ohio etc, I wrote a set of four posts
analysing "the return of the class gap". I dug into the exit poll data by education, and included a table, maps and a graph.
Now that PA, NC and IN have voted, I wanted to post an update of the data.
First thing that's noteworthy is how the results of this round of primaries compare with those of previous rounds. Mind you, it's odd sets of states that one is combining, of course. IN, PA and NC form a relatively cohesive set, but the March 4+11 states, ranging from the Hillary-friendly states of Ohio and Rhode Island to Texas to the very different but both very Obama-friendly states of Vermont and Mississippi, are a varied set. On the other hand, you could argue that because of that, they were pretty representative of the US as a whole, just like the Super Tuesday states.
So how does the comparison work out?
Noteworthy: the results from the last three states, broken down by education, are extremely similar to those from the March states. Down to just one point difference! That makes you marvel, again, at how static this race has been. And just how little effect, despite all the endless hyping by the media, passing affairs like Wright, Tuzla, "Bittergate" and the debates have had.
Basically, what we're seeing here is that on the whole, Obama continued to keep a modest edge among college grads, and Hillary continued to keep a narrow lead among those without college degree. Compared to Obama's mid-February surge on and around the Potomac primaries, this is a setback for him. But compared to Super Tuesday, a more favourable pattern holds: he remained stable among college grads, but minimized his deficit among those without college degree.
Re the old debate about whether Obama's initial disadvantage among working class voters was a question of his message not appealing to them or their being low-infomation voters who only caught on as the race became everpresent, this is something of a stalemate. He did caught on among non-college educated voters more as time went by, even when the brief mid-February surge faded, but still only closed the gap halfway. So a bit of both - plus, of course, Obama also adjusted his campaign focus and tone over time.
Now let's break it down by state, because there's a fair amount of randomness to these collections above. Here's the updated, full state-by-state table:
The latest three states are highlighted in yellow; the March primaries are highlighted in green. The states are sorted by Obama's performance among non-college educated voters.
The listing confirms the impression from the
Nation and
Enik Rising charts two posts up, that Obama's made some progress, if rather minor progress.
In North Carolina, Obama did a little worse among both college grads and non-college educated voters than in Mississippi. But considering that blacks made up half the voters in MS and just a third in NC, the difference is quite small. (This is because he did much better among whites than in in MS.)
As for the Midwest, you see here what Melber and Enik Rising showed above: Obama did a little better among non-college educated voters in Pennsylvania than he'd done in Ohio, and a little better still in Indiana than in PA. Among college grads there's no such neat trend: he did actually a little worse among those in PA than in OH, but improved significantly on both in IN.
All in all, Obama did about average among non-college educated voters in Pennsylvania and Indiana. His performance there ranks 15th and 18th in the list of 31 states we have exit poll data for. In North Carolina he did very well, but this is probably largely because of the high number of African-American Obama voters there, who tend to be less likely to be college grads. He did less well than in Maryland and Virginia though, which have comparable black populations; this is probably because he did less well among white voters than he'd done in those two states in the Potomac primaries.
Take a look at what I dubbed the "class gap": the difference between the lead or deficit that Obama had on Hillary among those without college degree, and that which he had among those with a degree.
This "class gap" was non-existent in North Carolina, just like it had been very low or even inversed in other states with a large number of black voters (South Carolina, Georgia, Mississippi, Virginia and Maryland). In these states, Obama's two core groups of support, higher-education whites and African-American voters, weigh in to equal effect on the two sides of the education divide.
More interesting is that the "class gap" was also relatively small in PA and IN. Lower than average in PA, and just about average in IN. The four states with the highest class gap were all Super Tuesday states. That's an illustration of how Obama did close his deficit among lower-education voters at least half the way.
