So the excitement of last week's announcement has died down a bit, but I've finally gotten together the polling numbers by state, with decent numbers for 43 states, thanks to the relentless work of Public Policy Polling. Of course, everybody still wants to know about the impact on the November election, so the states most likely to be close have been scrutinized in the press. Conclusion: they don't favor gay marriage. Which is true - when the question asked is "Do you think gay marriage should be legal or illegal?"
But the three part question - asking whether somebody favors gay marriage, civil unions, or no legal recognition at all - is more useful here in my opinion. If we assume that those who feel most strongly about the issue are those for whom a candidate's position is most likely to affect their vote, then what we want to know is which group is bigger: those who support gay marriage or those who are against all legal recognition? I call the difference between these two groups the Intensity Difference in the chart below, which includes all the Toss-Up and Lean states as defined by that arbiter of conventional wisdom, the New York Times:
What we see is that the Intensity Difference is very small in some of the biggest Swing State prizes: VA, MI, FL, WI, PA. Meanwhile, it is clearly positive in the purple western states, IA, and NH. If this Intensity Difference actually relates to political behavior, Obama's position is an obvious help for his reelection efforts. I'm guessing Obama's campaign was well aware of this prior to last week.
Below the fold, I will give you numbers for 43 states, and hints for the rest. The states all seem to be following a steady march towards equality, yes, even in the Deep South. Meanwhile, opinions on gay marriage are surprisingly constant within a region. Local culture, broadly defined, seems to have a large effect on views, even larger than respondent's race, which is unusual for political opinion in this country.
First, I would like to thank the commenters from last week's diary summarizing the national polling picture on marriage equality. Every portion of this diary grew out of an idea raised by a commenter, and I also took advantage of several suggestions for improvement and links to new data.
State by State
Now, here's the numbers: on the left, a list sorted by state, on the right, a list sorted by support for gay marriage. These are the aggregated results of the DailyKos/SEIU/PPP national polls separated out by state, and PPP's state polls, based on 58,405 respondents in 2011 and 2012.
Note there are now two states with outright majority support for gay marriage - Massachusetts and Vermont. Together, support for some sort of union is now at 86% in MA and 79% in VT. Amazing!
There, is, sadly, one state with majority support for No Legal Recognition: Mississippi. Five more southern states have nominal majority support for this position, but within the margin of error.
But look what's NOT the matter with Kansas: smack in the middle of the list, with support similar to the national split of one-third for each position. It's not a fluke, either; Nebraska has similar numbers.
There are seven more states and DC left out of the list because of lack of data. Of them, ID, WY, and SD show generally low support for gay marriage, RI shows moderate support, and AK, DC, and DE show high support. ND has only 6 respondents on the issue.
The Arc of the Moral Universe
We can plot all the states with decent data on a graph like last week's:
How to read this graph: To find out the percent that support gay marriage, follow the blue line up and to the left towards the left axis. To find out the percent who think there should be no legal recognition, go follow the red line down and to the left towards the bottom axis. For the middle positions, go straight right along a grey line to the right axis. This axis is labeled Civil Unions, but also includes those who chose Not Sure as a response, as numbers on this kind of chart must add up to 100 and Not Sure is a "mushy middle" response that fits far better with Civil Unions than pro-marriage or anti-everything.
Nationally, opinion sits almost exactly in the middle of this chart, and has been moving to the left slowly but steadily. In fact, because of changes within demographics and generational differences, it seems that individual states are on a path following the green arrow. (Time will tell for sure).
It's interesting to note that no state has support for Civil Unions at more than about 40%. In other words, that opinion appears to be an unstable intermediate position - eventually the change of heart becomes complete.
Deep in the Heart of Dixie
But what about the most recalcitrant of opposition? Will Mississippi ever change? Yes. Even in the most conservative areas, young people are more liberal than their elders. Below is a graph by age for three regions with decent size sample of 18-29 year olds. Regions are defined here, and you can see maps of the Northeast Corridor, Rural Heartland, and Old South.
Although the magnitude of generational change (length of colored arrow) is much larger in the Northeast Corridor, the Old South is nevertheless on its way, as is the Rural Heartland. This is consistent with polls showing that even young white evangelical christians are more supportive of gay marriage than their grandparents. Specifically, 27% of 18-24 year-old white evangelical christians favor same-sex marriage, compared to 19% of white evangelical christians of all ages. (Note that in the second report linked, there were really too few respondents in the youngest age group for a decent number; the first report is a survey of 2000 people age 18-24.)
It's Place, Not Race
If we look at these same three regions, but this time separate out racial categories, we see something unusual: a political question where African-Americans and whites in the South are in close agreement. In fact, we see more variation across different regions than we see between races within a region.
What is going on here? It must be regional cultural differences. There are three things I can think of that may be contributing: religion, familiarity with gay families, and public laws. Obviously, these three factors are not independent, leading to a bit of a vicious cycle.
Religion: More specifically, the states with lowest support for gay marriage closely track this map of Baptist adherents. Support in the Mormon Triangle of UT, ID, and WY is lower than neighboring states too.
Gay Couples: Census data show a lower concentration of same-sex households in the South in general, especially AL and MS. (However, this is also true of the Plains and parts of the Midwest.)
Laws: If Massachusetts hadn't legalized gay marriage first, would it still be leading in acceptance of gay marriage? What about Iowa? Views in Iowa are not particularly different from neighboring states, but gay marriage has only been legal for three years so far.
So two of these factors, at least, can change over time. Religion - well, Mormons at least have changed their doctrine with regards to marriage in the past... I suppose they could do it again...