In 2016, the relationship between marriage and voting declined
Although you wouldn’t know it based on which elements of my research agenda get media attention, the subject I have worked on more than any other since grad school is the relationship between family formation and partisan politics. Specifically, I argued in my dissertation, my first peer-reviewed article, and my first two books that the GOP benefits when people form family units at an early age.
But since the 2016 presidential election upended so many of the normal rules of politics, I thought it was worth checking if this rule also no longer applied.
It turns out that marriage age and voting were still correlated in 2016, but the relationship was much weaker.
In political science, it is a good idea to not get hung up on the R-squared (the amount of the total variance explained by the model). Even in models with a huge number of independent variables, it is pretty common to have a low R-squared. But that doesn’t mean that the relationship you are studying isn’t substantively important.
However, when it comes to the median age at first marriage and vote choice at the state level in presidential elections, you haven’t recently needed to include that caveat. To my knowledge, no variable has ever had such a strong, linear connection to state-level vote choice in presidential elections than the median age of first marriage for women. In 2012, this simple two-variable regression had a ridiculous R-squared of 0.72. That’s a number you simply don’t see in political science with such a simplistic model.
The 2016 election was a little different, however. The relationship was still there (if it totally disappeared, then the world really had been turned upside-down), but it was weaker. The R-squared dropped to 0.57. That’s still really good, but not amazing.
Part of this is due to Trump’s poor performance in Utah, which has, by far, the youngest median marriage age. Trump under-performed in that state compared to recent GOP nominees, especially the Mormon Mitt Romney.
But looking at it more broadly, compared to other recent GOP presidential candidates, Trump did less well in many states that have traditionally had a very young median marriage age, and better in many states where people do not get married, on average, at an early age.
People who know my work also know that I support pro-family policies and politics – which, in my view, also means rejecting many mainstream conservative principles about government. But based on electoral politics, it is better for the GOP if it is not totally dependent on people who form stable family units at an early age. Although I lament this trend, such families are becoming increasingly rare.