Linhda and Kyle review a New York Times article titled How Your Hometown Affects Your Chances of Marriage. This article explores research about what correlates with the likelihood of being married by age 26 by county. Kyle and LinhDa discuss some of the fine points of this research and the process of identifying factors for consideration.
Data Skeptic
[MINI] A Critical Examination of a Study of Marriage by Political Affiliation
[music] The Data Skeptic Podcast is a weekly show featuring conversations about skepticism, critical thinking, and data science. Well, welcome to another mini episode of the Data Skeptic Podcast. I'm here, as always, with my wife and co-host, Linda. Hello. So, Linda, you sent me this article from the New York Times. Could you give a quick summary of what the article's about? So, the article's titled, "How Your Home Town Affects Your Chances of Marriage." And the summary is that it looked, broke it down by county about where I assume they're defining hometown as where you grew up. Right. And then if you hover over the interactive map, it gives you a percentage point, whether you're above or below. The average, like you're saying, negative 2%, less likely than the average to get married by the age 26, if you grew up in a hometown in Area X or whatever. Yeah, so maybe we could cover some of the methodology first. Like you said, they looked at people based on where you grew up if you were married by age 26 or not, which I guess is a reasonable line in the sand. Do you think that's sort of an average age that people get married? Yeah, I think it's close to the average age of which women get married. And they also comment in this article that the researchers said if you did it at 30, they still see similar results. I'll take them at their word and say that that's a reasonable line of demarcation. The second question I think is really important to ask in cases like this is when there are differences from the mean, are those even interesting? Because there's always going to be one county, which is the highest rate of marriage, and one county which is the lowest rate of marriage, just by the nature that there's variation. What's important to ask is, are there aspects, features of those counties, that impact that score? Are there reasons why areas of young marriage or older marriage or non-marriage are they related to aspects of that county? So maybe we could talk about some of the factors we think are at play in why people from certain areas may or may not get married by age 26. What do you think? Well, I mean, just to summarize the article, they also break down the data in several different ways. They talk about by city, big population, small population. And then they also wrap it up by talking about partisanship as in Republican or Democrat, and then small towns, then specifically the South. So the article covers a few different topics. It does, yeah, and that's where we get into a gray zone, maybe. It's like, have they covered appropriate topics or have they covered too many topics? Meaning, if you look long enough, maybe you'll find something in each of those counties that correlates with the marriage rates. Like maybe if you have a comptroller who's very young, then the marriage rates are very low. I wouldn't expect that that actually influences anyone's marriage likelihood. But they looked at some more reasonable things, right? Like they looked at, as you mentioned, the voting records of those towns. They looked at the reality of the towns and some things like that. What kind of factors do you think impact the age of which people get married in general, county aside? Well, I mean, the article goes into a lot of them, which is education, income level, race, your religion, or in general, just your beliefs. The ratio of viable partners, this article assumes men and women it seemed like. You know, that's a really good point that I don't know if the researchers looked into this, but the article didn't cover. But you said this earlier, I hadn't thought of this, but yeah, the ratio of available mates is a significant consideration that I think it's overlooked a lot. I mean, other things to consider is whether contraception's available, whether people get married because they're pregnant, or what time they want to have kids, or accidentally have kids. Yeah, I've always wanted to look into like home prices and rental prices. Like, I know a lot of people in New York tend to move in together earlier in relationships than in other places, just because it's so expensive to live in New York. And maybe like, I don't know, people who feel that they can't purchase a home are less likely to be married. I don't know that for a fact, but there are aspects of rental and home prices that could impact the rates of marriage. So I think that one thing we can conclude is that the factors that go into why people in certain areas get married are probably pretty complex, would you agree? Oh, yeah. A large number of variables influence that. I mean, probably if you're in Vegas, people might get married more. Actually, I don't even know what county Vegas is in. We should check. Oh, yeah. Actually, Vegas, we should just stay away from that one because it's a weird outlier in that I suspect most of the marriages in Vegas in that county are for people not from that county. Right. Which is probably very atypical. I bet it's an outlier in that respect. Another interesting outlier they point out is counties in Utah, which has a strong Mormon presence, which highly correlates with republicanism, I think. If you see strong young marriage rates as they do, can we say that's because of the republican influence or the Mormon influence? I don't know if we can separate those entirely, but we could try. Well, I think in this case, it's probably Mormonism because in the certain counties where bring them young and such exist, they said you're 20% more likely to be married. They were more likely to be married by age 26, which I think was probably the highest of any county in the US. I think so. Yeah. Strong outlier. They also had an interesting comment about counties that vote strongly republican because this article looks at the percentage of votes, not just who won, but like this county voted 90% republican in the last US presidential election. Let's say that those strong republican leanings tend to also be very rural and we know people in rural areas get married young perhaps for independent reasons like lack of availability of options or things like that. But they did point out that there are some densely republican areas that are also dense in population and those fit this trend. So that's good supporting evidence that maybe partisanship is closely correlated with likelihood of marriage by age 26. In fact, they share a scatterplot in this article. Could you maybe describe that? What sort of general trend do you see where we have democratic leanings on the left of our x axis and republican leanings on the right side of our x axis and marriage rates above or low the mean value on the y. Yeah, I mean it definitely looks linear as in the more republican you are, the more likely you are to be married by age 26 and the more liberal you are, the less likely you are to be married by age 26. Yep. That's a pretty convincing plot. If you don't go too deep into it, and in fact if this were some sort of regression course and you handed in a result that said you did a regression and you found it was linear and it was negatively correlated. That would be the correct score from the mathematical point of view. But the question is, is this the right way to slice the data? Do you think it's as simple as saying the partisanship leanings of a county is strongly correlated with the marriage likelihood? I mean correlated meeting linked, I guess so. Yeah, it's definitely linked. But I guess the question for me is, can we reduce it down to just voter records? Can we set aside all those other factors we brought up earlier? Do they not matter enough? Is this the most pronounced feature? I'm not sure. That plot does seem to imply that. But then there's one aspect of the article that makes me pause a little bit and they say, "Oh, the south is a little bit complex." Do you remember that part? Yeah, so that was my favorite. I am from the south. So... The deep south? I don't know what the deep south is. I don't either. I just know true southerners are usually from the deep south. I don't know if you're from the south. You're from the south. But anyway, so they said that in the south, if you are of the high income population, the south pushes you to actually marry by the age 26. You're more likely to be married by age 26. And then if you're in the low income bracket, then the south pushes you to not be married by age 26. Yeah, which is sort of different from other parts of the country. They're kind of asking us to say that the south is an exception, which is a little bit of not a red flag, but like a yellow flag for me. As soon as you say, "Oh, my theory is this, but here's a footnote. Here's some exception that weakens your theory because it's now more complex theory and we have to kind of score it a priori a little bit less likely." Which is not necessarily meant to be a huge criticism of this work. I think that it's some really good analysis and investigation just to say that this is a pretty complex topic. And I think we can agree that there are a lot of factors that go into deciding whether or not people from a certain county are going to be married by age 26 or not. So I appreciate the analysis. If I had a criticism, I guess maybe it's that scatter plot we were talking about that seems to suggest the more densely Republican an area is, the more likely the marriage rates will be on the increase, which is probably fundamentally true, but it maybe over reduces the problem. Would you say that or no? I don't really know. What do you think impacted you and I getting married? I feel like you're putting our marriage up here to be discussed. I don't know. What do you think? I think our marriage is an anecdote, a cherry-picked example. But yeah, we're from the less likely to be married by 26 areas of the country. Well, we didn't get married by age 26, so... So it's certainly not support for the argument, but it is not a refutation either. We fit into the prediction that the naive prediction one would come away from with this model. But anyway, how did you get on to this whole article to begin with? I mean, I think the most interesting part is that the New York Times did a series of articles talking about upward mobility in the U.S. and how it can be tough and how where you grow up, your city, your education, the people you're around can influence how likely you are to change statuses, i.e. jump from low income to high income. Oh, interesting. Now that would be a wonderful series of topics for other episodes, maybe some interviews, maybe some mini-sodes, but I think we've pretty much tapped out for this mini episode. So thanks again for joining me, Linda. Thank you. Bye-bye. [outro music] [BLANK_AUDIO]