Racist, sexist, and anti-Semitic jokes in Trump land

This post contains racist language.

Updated: See comment note and data caution at the end.

This is purely observational, not causal. People Google for racist, sexist, and anti-Semitic jokes more in states that are more favorable toward Trump in the presidential election.

The point of the exercise, as suggested by Seth Stephens-Davidowitz in a 2012 paper published here and discuss here, was to look for population traits that might skew votes in ways the polls did not predict. If people were racist, maybe they would not admit they opposed Obama, but they would still Google “nigger jokes” in their spare time. We don’t yet know whether the polls will accurately capture the vote outcome this year, but I’m interested in the underlying patterns anyway.

I use state data from Google Trends, which coughed up relative search frequencies for the past fives years by state. Each search term is scaled from 100 in the state with the highest search frequency of the term to zero for the lowest (except they don’t go down to zero). For example, West Virginia scores 100 on searches for “nigger jokes” and Oregon scores 17 (the lowest score). Trends does not report the actual number of searches, and some small states are not reported for some jokes, presumably because the data are too sparse.

So here I compare search frequencies for three offensive kinds of jokes, “blonde jokes” (N=48), “nigger jokes” (N=38), and “holocaust jokes” (N=29), with controls for two kinds of innocuous jokes “puns” (favored by Clinton supporting-states) and “knock knock jokes” (favored in Trump states). This might capture the general tendency to Google for jokes. I compared these relative search frequencies to the state polling summary from FiveThirtyEight, which has the Clinton lead from +32.8 in Hawaii to -30.4 in Wyoming (DC is not included here).

The bivariate correlations with the Clinton lead are -.67 for “blonde jokes,” -.61 for “nigger jokes,” and -.48 for “holocaust jokes.” Here are the scatters (click to enlarge):

Again, nothing causal claimed here. Just accounting for other joke telling (which is interesting in itself, here are the multivariate results:


Blonde provides the best fit but they all are still pretty good with the innocuous jokes controlled.

Incidentally, “puns” has no bivariate correlation with Clinton lead, but with “knock knock” controlled it’s very strong. Go figure!

OK, there you have it. Deplorable joke behavior is strongly correlated with Trump support. Nothing causal claimed here.

I put the data and Stata code, including code for the figures, on the Open Science Framework here.

For other relevant posts follow the Google tag and the Trump tag.


Thanks to the efforts of University of Wisconsin graduate student Nathan Seltzer (see the comment below), it’s come to my attention that the “past five years”data is unstable. Looking just at the “holocaust jokes” data, s/he found non-trivial noise comparing the downloads just a few hours apart. To check this, I just went and repeated the search: “holocaust jokes” for “past five years,” and this is that I got:


Yuck. Thanks for the free data, Google! I’m thankful for Nathan pointing this out. Good lesson in the benefits of sharing data so we can find problems like this — and the trouble with counting on non-open, private data providers like Google. When they’re good, they’re good, but they’re non-transparent and unaccountable when they’re not. It would be great if Google figured out what’s going on and fixed their public access tool. If anyone else can explain this I would be interested to hear.


Filed under Politics

Vasectomy reversal, divorce, and American optimism

That would be a good title for a longer essay (feel free to use it).

“Studies suggest that up to 6% of such men will request vasectomy reversal,” wrote the authors of a chapter in Clinical Care Pathways in Andrology. “Divorce with remarriage is by far the most common reason for vasectomy reversal.”

So, when in the divorce process do people start Googling “vasectomy reversal”? Is it men with younger girlfriends, considering leaving their wives? Women considering marrying a divorced man? Divorced couples considering another round of kids?

I don’t know, but Google does, or they could if they looked into it. I’ve only gotten as far as the strong relationship between searches for “vasectomy reversal” and state divorce rates:


I like to think of it as the optimism rooted in the American spirit. We always look forward to the next renewal, the next reboot, rebranding, or escape. Not because I really think it’s true, I just like to think of it that way.

The Google data is from their Trends tool, the divorce data is from the ACS via IPUMS.org.)

(This is a return to an old post, in which I first noticed this relationship, with new data. Now I think divorce per population makes more sense for Google correlations, rather than divorce per married population, which I used before.)


Filed under In the news

Don’t think economic anxiety is rational and racial anxiety is not


Photo by Patrick Feller (and check out his essay about it: https://flic.kr/p/6voQ7g)

Here’s a very quick thought, over which I’m happy to hear objections.

You know how people say “inner city” or “urban” or “low-income” instead of Black because they don’t want to seem racist by mentioning race? This debate over whether Trump supporters are motivated by economic anxiety or racism reminds me of that.*

The debate seems to divide class leftists from race leftists. Some class leftists want to emphasize the economic anxieties of racist Whites (with which they are sympathetic), and the race leftists want to emphasize the racism of economically anxious Whites (which they want to expose). In the regular liberal media, at the same time, there is the common tendency to treat class anxiety as rational and material while any racial motivations are by definition irrational and emotional (which makes “I don’t see race” a moral high ground.)

But Whites losing their race privilege is a real, material concern. And losing class privilege motivates ugly hatred and animosity too. What if “China” threatens your jobs, and your emotional response is to support a dictator who opposes imports and immigration (which will do nothing for you economically)? What if a Black president and anti-discrimination laws threaten your privileged access to relatively high status social recognition, and your rational response is to support school segregation and oppose affirmative action (which might actually protect your privileged status)?

Of course, defense of both class or race privilege can be emotional and ugly and vitriolic, and they always are coming from the mouths of Trump supporters. And you could reasonably argue that all people would be better off embracing a more open and inclusive politics even if it cost them some ill-gotten gain. But those defenses can also be rational and material, and the privileges they protect may need to be forcefully degraded rather than just reasoned away.

Why should Trump supporters motivated by economic anxiety be any more deserving of respect than those motivated by racial anxiety? That’s the politics of our time; don’t treat it as fixed or essential.

Now, you wouldn’t excuse rich bankers supporting a dictator because they had anxiety about their economic position, even though they might really have a lot of economic anxiety. So Why would you excuse White working-class people from supporting a dictator because of their economic anxieties? Is it because they’re actually poor or economically insecure? Well, they’re poorer and more economically insecure than the people having this conversation, but not compared with actual poor people in this country or — shudder — in most of the world that disgusts them.

It is an empirical question whether the anxieties around race are more or less rational than the anxieties around economics for White working-class Americans. Losing your race privilege might mean getting worse service from schools and emergency services and police, and not seeing your people in high status and visible cultural positions, and not hearing your music all the time, and so on and on. These are all the things minorities want. Of course I don’t feel sorry for people losing them like I do for people who never had them, but the issues are the same. You can’t say minorities are rational for demanding these things and then say Whites are irrational for trying to hold onto them. You could (and should) argue it’s not a zero-sum game, of course, but that’s an empirical question and a matter to be worked out through politics and cultural change.

People who support Trump definitely are anxious about losing things of value as the world changes, and their response is deplorable and must be opposed — regardless of the relative mix of economic and racial components in their minds.

* Not doing a full lit review, but to get a sense of it, read and follow the links in pieces by Dylan Matthews, Mike Konczal, Derek Thompson, or Michael Tesler, Brian Beutler, and there must be some other White men I’m missing. Feel free to recommend readings you prefer in the comments.


Filed under Politics

The liberalization of divorce attitudes proceeds apace

The 2016 Gallup poll results on what is morally acceptable versus morally wrong came out over the summer, and they show that U.S. attitudes toward divorce continue to grow more positive. The acceptable attitude has gained 5 points in the last 5 years:


This parallels results from the General Social Survey, which asks, “Should divorce in this country be easier or more difficult to obtain than it is now?” The latest GSS is still 2014, but it also shows a marked increase in the liberal easier view over the same time period:


See more under the divorce tag.


Filed under In the news

The new Wilcox thing is complete bologna and/or just dishonest

Update: after Wilcox updated his report with the complete data, I now conclude the report is just dishonest, not complete bologna. See below. 

Brad Wilcox and Nicholas Zill have a new report on Brad’s Institute for Family Studies website, about Arizona: “Stronger Families, Better Schools: Families and High School Graduation Across Arizona.” It bears a strong resemblance to a previous report, about Florida, “Strong Families, Successful Schools: High School Graduation and School Discipline in the Sunshine State.” Together these two give you a feeling like taking the first two bites of a leftover giant burrito that might have gone a little bad, and realizing there are probably about 48 more bites to go.

Anyway, this is about the Arizona one. I’ll first raise the possibility that it’s complete bologna – as in, fraudulent or error-ridden – and then discuss how it’s conclusions are dishonest at best even if the analysis is not technically wrong but rather just presented terribly.

Update: with the report corrected to show the complete data, the analysis now replicates fine. So I set aside the bologna issue. I leave this section here just so you can see the research design, but the main argument is in the next section.

First, the bologna issue

The report uses demographic data from 99 Arizona school districts to model graduation rates, and the gender gap in graduation rates. Their conclusion, based on two regression models using districts as the units of analysis and demographic indicators as the predictors, is this:

In Arizona, public school districts with better-educated and more married parents boast higher high school graduation rates. Gender equity is also greater in districts with more married parents. That is, boys come closer to matching the high school graduation rates of girls in districts with more married-parent families. Moreover, married parenthood is a better predictor of these two high school graduation outcomes than are child poverty, race, and ethnicity in public school districts across the Grand Canyon state.

To pad out the report, they also include appendix tables, so it’s theoretically possible to replicate their regressions. Unfortunately, unless I’m missing something, they don’t replicate. I wouldn’t normally bother rerunning someone’s regression, especially when the argument they’re building is so wrong-headed (see below), but just because we know from long experience that Wilcox does not behave honestly (in methods, ethics, and ethics) what the heck.

The report says, “Graduation rates and male/female graduation ratios for the 99 Arizona school districts in our study are shown in Table A1 in the Appendix.” Table A2 then lists the districts again, with the demographic variables. Unfortunately, table A2 only includes 83 districts – and the 16 missing are exactly those from Indian-Oasis to Paradise Valley in the alphabetical list of district names, so apparently an error handling the data. So I could only use 83 of the 99 for the regressions. Since I don’t know when they lost those 16 districts, I don’t know if it was before or after running the regressions (there are no Ns or standard errors on their regression tables).

For each of their dependent variables – graduation rate, and the male/female ratio in graduation rates – they list bivariate correlations, and adjusted betas from as multivariate regression. Here are their figures, with mine next to them. The key differences are highlighted:


If they’re using 99 cases and I have 83 (actually 81 for the gender cap because of missing data), you would expect some difference. But these are very similar, including the bivariate correlations and the R-squareds for the models.

The weird thing is that the biggest difference is exactly on their biggest claim: “married parenthood is a better predictor of these two high school graduation outcomes than are child poverty, race, and ethnicity…” That is based on the assertion that .29 is larger than -.28 (very luck for them, that tiny, insignificant difference in  magnitude!). In my model the minority-size effect is more than twice as large as the marriage-parenthood effect. So, huh. It’s definitely possible Brad simply lied about his results and made up a few numbers. (And I’m just using the data they include in the report.) But now let’s pretend he didn’t.

Update: with the complete data I can report that those two betas are actually .2865 (.29!) versu .2847 (.28!). The idea that one is a “better” predictor than the other is clearly not serious. Further, for some reason (we can only guess), they combined percent Black, Hispanic, and American Indian together into “minority,” which produced the .28 result. If they had entered them into the model separately, they would find that Hispanic and American Indian effects are each bigger than the married parent effect, as I show here:

So much for the headline result. Anyway, back to the argument…

Policy, shmolicy

The point of the analysis is to make policy recommendations. They conclude:

If the state enjoyed more stable families, it might also see better educational outcomes among its children. It’s for that reason that Arizona should consider measures designed to strengthen and stabilize families.

Their recommendations to that end are vocational education and marriage promotion.

Private and public initiatives to provide social marketing on behalf of marriage could prove helpful. Campaigns against smoking and teenage pregnancy have taught us that sustained efforts to change behavior can work.

First, I’m not an education specialist (and neither are they), but shouldn’t there be some kind of policy variables in this analysis, like per-pupil spending, or teacher salaries, or something about curriculum or programming? It’s unusual to use only demographic variables and then conclude that what we need is a policy to change the demographics. It’s just not a serious analysis. (Please also please remember that “controlling for income” is not an adequate control for economic conditions and status.)

But second, given the first billion dollars of money spent promoting marriage produced absolutely no increase in marriage, is there any possible way Brad legitimately thinks this is the best way to improve graduation rates?

These are just two ideas. More should be explored. The bottom line: policymakers, educators, business leaders, and religious leaders in Arizona need to address the fragile foundations of family life if they hope for the state’s children to lead the nation in academic achievement.

Does this report really support that “bottom line”? Would it be better to spend money promoting marriage than to spend the same amount of money on some effort to improve schools? That’s obviously a dumb idea, but is it possible he really believes it? These are the only policies proposed. Maybe I’m wrong, but I doubt he believes it. I think he wants to promote marriage promotion programs for other reasons: to fund him and his compatriots, to support pro-marriage ideology, and so on. Not to improve graduate rates in Arizona schools. But, maybe I’m wrong.

And a laptop

I think what Brad is really doing is noise noise statistics statistics marriage-is-good expertise trust me fund me. The details clearly aren’t that important.

Meanwhile, not coincidentally, things are looking up for Brad at the Institute for Family Studies (IFS), the organization he created to handle the foundation-money rake. He started in 2011 as president / director of IFS at a salary of $35,000. After paying himself a paltry $9,999 and in 2012, he started improving his productivity, paying himself $50,000 in 2013, and then $80,400 in 2014 as a Senior Fellow, the last year for which I found a 990 form. Much of that money is coming from the Bradley Foundation (which also funded the Regnerus/Wilcox study) — their 2015 report lists $75,000 for IFS, so projections are good for next year. This is, of course, on top of what he gets for his service to the public at the University of Virginia.

The IFS disclosure forms also show purchase of a MacBook Pro. Which might or might not have been for Brad.


I do not make this case, and make it personally, because I disagree with Brad about politics. There are lots of people I disagree with even more than him, and I don’t spend all day criticizing them. The dishonesty offends me because it’s work and issues I care about, it hurts real people, I’m well situated to expose it, and his corporate-Christian-right megaphone is big, so it shouldn’t go unchallenged.


Filed under In the news

Gender on the Diane Rehm show in September

The last Media Matters report on the Sunday TV talk shows reported that 73% of guests were men in 2015, a little less than the 75% recorded for the previous two years. (That includes journalists as well as politicos.) I expect my local NPR station, with its liberal audience, to have a better showing for women, and it does. The Diane Rehm Show, which is produced at WAMU but distributed nationally as well as podcasted, had 129 guests in September, and 80 of them (62%) were men, by my manual count. (I’m not counting the hosts, who changed over the month.)

But what has been striking me lately, and the reason I did the count, was how rarely it seems that women are in the majority among the guests, and especially how often there is one woman and more than one man. Without a whole conversation analysis, you can imagine the kind of dynamic that made me think, “sure is a high male/female word-count ratio in this discussion.”

The count confirms this. The show averaged 2.8 guests per episode. So how are the men and women distributed? Of the 46 shows aired in September, 12 featured just one guest, 8 of whom were male. Male guests outnumbered female guests overall in 29 episodes, or 63% of the shows. Female guests outnumbered men in only 8 shows (17%), with the remaining 9 (20%) being gender balanced. What accounts for my annoyance, maybe, was that in those male-dominated shows, more than half (16 of 29) featured just one woman and more than one man. The reverse – one man and more than one woman – happened just three times. Details in the figure.


The most common configuration is one woman and two men. 

My point is just that a 62% / 38% gender split leads to a lot of small-group discussions where men outnumber women, and especially solo-women versus multiple men, which is its own kind of gender situation. I imagine you get this pretty often in cases – say, at an academic conference – where there is some effort to reach gender balance on most panels, but women are less than half altogether. (You can see they were paying attention because there were no all-male panels of four or five.)

I’ll leave it to Media Matters to do their annual report again next year, but I did take a quick look at some of the Sunday shows for September. On Meet the Press I found 62% men, and 75% of the shows were male-dominated. On Fox News Sunday 71% of guests were male, and every show was male-dominated. Face the Nation had 72% male guests but also every one male-dominated. (Incidentally, Face the Nation has a convenient list of every guest so far for the year, so I was able quickly tally the gender of their 348 guests, 73% of whom were men, counting multiple appearances. That’s a tiny bit better than their 2015 total of 76%.)

Related on gender composition:

1 Comment

Filed under In the news

Cause and effect on myopia

It’s funny for a non-eye specialist to read articles about myopia, which in my line of work rarely means myopia, literally, which is nearsightedness. Takes some getting used to.

Anyway, in my book I use as an example of misleading correlations the link between night lights and myopia in children. Checking it to make sure it is still a good example to keep for the second edition, I was glad to see that it holds up well.

Here’s the story. In a 1999 paper (paywalled | sci-hub), Quinn and colleagues reported a “strong association between myopia and night-time ambient light exposure during sleep in children before they reach two years of age.” That is, kids who slept with night lights were more likely to be nearsighted. This was potentially big news, because we actually don’t fully understand why people become nearsighted, except we know it has to do with reading a lot and spending a lot of time indoors as a kid. They had some idea that light penetrating the eyelids at night might do something, but no real mechanism, just an association over a few hundred kids.

The paper didn’t have some important variables controlled, notably parents’ nearsightedness. Since the condition is also genetic, this was acknowledged as a problem. Still, they wrote:

Although it does not establish a causal link, the statistical strength of the association of night-time light exposure and childhood myopia does suggest that the absence of a daily period of darkness during early childhood is a potential precipitating factor in the development of myopia.

As I stress ad nausem in this post, the “strength” of an association is not an argument for its causal power. And neither is the number of studies in which the association is found. Real spurious findings can produce very strong, easily-reproducible results. And when researchers have a story to fit the rationale can seem strong. Also, the prospect of publishing in a top journal like Nature has to figure in there somewhere. (This problem is endemic in studies of, for example, family structure and child outcomes, among many other subjects.)

In this case there is a very nice explanation, which was reported less than a year later by Zadnik and colleagues (paywalled | sci-hub), who found no association between night lights and myopia – but they did report a very strong relationship between night lights and parents’ myopia. The same pattern was reported in another response to the Quinn paper, in the same issue, by Gwiazda and colleagues. It appears that nearsighted parents like to leave night lights on. Alternately, some other factor causes parental nearsightedness, child nearsightedness, and night light preference, such as education level (e.g., more-educated people read more and use night lights more).

Several other studies have also failed failed to confirm the night-light theory, and now the thing seems to have blown over. It’s not a perfect example, because the bivariate correlation isn’t always found, but I like it as a family-related case. So I think I’ll keep it in.


Filed under Uncategorized