After crunching numbers and playing with some simulations this morning, it is clear that even if North Carolina's African-American population were more liberal on the issue of gay marriage than the residents of Massachusetts, Amendment 1 still would have passed. Even if African-American opposition were unanimous - 100% voting no - it would have passed in about 3/4 of the simulations.
Can we Stop Blaming the Blacks yet?
After reading Denise's post last weekend, I started wondering if Amendment 1 could have passed without any support from African-Americans whatsoever. I went scrounging around for the numbers, and here is what I found:
I used PPP's polls of 4338 likely North Carolina voters to estimate likely ranges for voter turnout and Amendment 1 support by race. I then constructed 139 scenarios where the combinations of turnout and support added up to the actual election results.
Now for the hypotheticals:
Scenario 1: More liberal than liberal.
In 2011 DailyKos/SEIU/PPP polls, 13% of Liberals said there should be No Legal Recognition for gay couples, while 17% supported Civil Unions. In Massachusetts the numbers were 12% and 31%, respectively. (Keep in mind that about half of voters in NC who support civil unions nonetheless voted for Amendment 1.)
So if we set support for Amendment 1 among African-Americans to just 10% - more liberal than Liberals, more supportive than Massachusetts - in how many scenarios would Amendment 1 have passed anyways? Every damn one of them.
Scenario 2: Loyal Democrats
But what if Blacks in North Carolina were as committed to marriage equality as they are to the Democratic Party? Shouldn't this be an issue Democrats are united on? (In which case, what about the white Democrats? But never mind, they couldn't be at fault.) So let's set support to just 5% - that's the share of the NC African-American vote McCain got in 2008. Now Amendment 1... loses in 4 of the 139 scenarios - all four have an electorate that is 19% Black and 77% white.
Scenario 3: Total Unanimity!
And what if every single last Black voter who voted had voted against Amendment 1? It still would have passed in 102 of 139 scenarios.
Opposition to the formal recognition of gay relationships has very little to do with race in this country. Nationwide in 2011, it was 34% among whites and 36% among African-Americans. Everybody has a lot of work to do on this issue.
I used PPP's four polls of likely voters and their previous performance to generate estimates of voter turnout and support for Amendment 1. In the 2008 Democratic primary, PPP polls did a very good job estimating the composition of the electorate, coming within a point or two of exit polls. I also looked at early voting returns and registration numbers to confirm the range. Here's the ranges I used:
Percent Black: 15-19%
Percent White: 77-80%
Pecent Other: 4-7%
For support for Amendment 1, it appears just about all the undecideds voted yes. Here's the ranges I used:
Black Support: 60-66%
White Support: 59-63%
Other Support: 53-61%
To construct the scenarios, I varied composition and support by whites and others in discrete 1 pt steps, then calculated Black Support necessary to reach the actual result. Scenarios with Black Support outside the 60-66% range were removed. This left 139 'reasonable' scenarios for what the electoral composition and support for Amendment 1 may have been on election day. They do not have equal probability.
This is really a very rough simulation, with many faults. However, even this simple exercise made it abundantly clear that Amendment 1 would have passed no matter how African-Americans voted - assuming you don't believe Scenario 3 is very reasonable.