Please support these cartoons at Patreon.
One thing that continually bugs me is the “racists are unicorns” view, which I hear usually but not exclusively from the right. Basically, in this view, almost no one is racist, and no act is racist. Sure, KKK members are racist, but nothing else – not vote suppression, not claming a judge is biased “becuase he’s a mexican,” not the racial wage gap – can be called racism. Because naming racism is considered worse than racism itself.
Artwise, this is one of the cartoons I did with superdistorted figures – huge heads, tiny everything else. It’s always a fun challenge to try and make these anatomically impossible people work out well. I also decided that the mentor character should keep his nose held high, high in the air in all four panels.
TRANSCRIPT OF CARTOON
Panel 1
An older white man in a three-piece suit is lecturing to a younger white man wearing a short-sleeved shirt with a tie. We can see that they are indoors; there’s a window with curtains behind them. Throughout this cartoon, the older man has his head lifted veryhigh (i.e., “nose in the air”).
OLDER MAN: Liberals call anything they disagree with “racist.” But we Republicans are more serious.
Panel 2
The older man continues lecturing, holding out one hand in a “stop” or “slow down” gesture.
OLDER MAN: Racism is a serious accusation. Before calling anyone “racist,” we always ask, “is racial animus the only possible motive here?”
Panel 3
A close-up on the older man, who is now looking positively indignant.
OLDER MAN: And we NEVER call our enemies “racist” just to score a cheap political point!
Panel 4
The younger man asks a question; the older man looks pleased with himself as he answers.
YOUNGER MAN: Wow. So is there ANYONE we DO call ra-
OLDER MAN: “Black Lives Matter.”
Same old desire to use semantics to make a political point.
The dictionary “prejudice, discrimination, or antagonism directed against people of other racial or ethnic groups (or, more widely, of other nationalities), esp. based on such beliefs.” is probably the most common definition by a considerable plurality. It is also the fallback position for any racial debate.
The modern trend is basically an attempt to game the negative attributes of the common usage (reflexive avoidance; social disapproval; guilt, etc.) and apply them to a larger body of actions–without doing the work to make the associations accurate.
And it’s obviously tempting: by using the racist label, you conveniently get to skip over the “…and therefore it’s bad and you shouldn’t do it” part. It’s like any other argument-winning word, from “mansplain” to “SJW” to check your privilege;” if you can get a win and skip the actual argument it’s going to get overused.
Most folks have twigged to it, though, and I don’t think it is working so well. After all, if you mean “unconscious prejudice” or “prejudice plus the power to act on it,” you can always just use those terms. But people don’t.
Ampersand is correct, here is a wide quote:
The results seem to have been interpreted in a relative sense. So imagine if a non-Trump supporter indicated they “strongly agreed” that “black people have gotten less than they deserve” because they felt the plight of black people is an important political issue. Then image a Trump supporter, having just responded to the previous questions on how everyone has been screwed in the current economic situation, felt the statement focusing on black people was less poignant and hence indicated they merely “agreed”. That would be taken (if repeated as a statistical pattern) as evidence that Trump supporters have “racial resentment”.
Despis and others, you act like these surveys were just invented out of whole cloth and someone’s whim.
Maybe you don’t know much about the process such instruments undergo in development. Psychology and sociology have processes in place to develop the kinds of questions used in these surveys, and to make sure that they are both valid (testing for the thing they think they test for) and reliable (repeatable with the same or similar populations, getting the same results).
My grasp of my grad school statistics classes is unfortunately tenuous at this point, but one part I remember clearly is the section on factor analysis, in which you subject a proposed instrument (questionnaire) to various statistical tests to see if the constructs under them are sound. For example, I might think I’ve developed an instrument that tests respondents’ degree of religious conservatism, dislike of raisins, and preference for paper money vs. coins, but if I subject the results to a factor analysis and it suggests there are only two underlying “factors,” then one of my categories probably overlaps with another to such a degree that they’re indistinguishable (maybe there’s a high degree of correlation between being religiously conservative and disliking raisins, and my questions haven’t revealed any significant differences within that group). But if the results suggest there are four underlying factors, then one of my categories is too broad – maybe there are two sub-groups within religious conservatives, those who hang out with the conservatives because they like going to church a lot, and those who hang out with the conservatives because they prefer grape juice to real wine during communion.
In any case, while factor analysis doesn’t necessarily point you at the “ideal” label for a category (“should we call it ‘dislikes raisins’ or ‘prefers plain oatmeal cookies’?”), it tests the validity of your instrument and gives you a good idea of whether, in order to identify raisin-haters, you need to ask how they feel about fresh grapes and dried grapes, or if you just need to ask how they feel about dried grapes.
To come back to our survey here, then, the surveyors did not just brainstorm a bunch of questions that they thought “sounded kind of racist” and then ask them of people to “prove who’s really racist.” They took a great deal of sociological and psychological research on the phenomenon that has come to be known as “racial resentment” for the purposes of research, ran a bunch of data through various statistical tests in order to find out which questions were helpful in identifying a particular set of beliefs and attitudes and which were not, and then used that information to shape the final questionnaire administered on a regular basis to a subset of the American electorate. The results of which are reported in the study.
Of course, you only give weight to this if you think that racism, and racist attitudes, and beliefs that support racist action/inaction, is an actual thing worthy of study and concern, and not some kind of gamesmanship used for the purpose of seeking power. But there’s only one “side” in this “discussion” that believes that, as far as I can tell.
I disagree with this assertion about factor analysis. It doesn’t inform about whether the researchers are “testing for the thing they think they test for”. All it can identify whether there is a single common factor that is driving each of the individual the measures. What that factor is (i.e whether it’s “racial resentment”, “just-world belief”, “historical ignorance”, or something else) comes down to the author’s subjective interpretation of the set of questions. Statistics cannot do that.
Nor can factor analysis tell you whether that factor is actually a compound of other factors that simply happen to correlate with each other in the specific domain covered by the questions.
This study highlights the weaknesses of the “racial resentment” scale:
Desipis, thanks; that’s interesting stuff.
But according to your link, they found the racial resentment inaccurate only for conservatives. But two problems with that.
1) Would people who voted for Obama – who are the people under discussion – be conservatives?
2) The research I cited controlled for political ideology.
It’s difficult to believe that what’s going on here is unmeasured conservatism being misread as racial resentment, since they controlled for that, and since Obama voters are unlikely to be strong conservatives.
I think we’re working on a model where the expected correlation of political positions is based on preconceptions that are two generations out of date.
You know what I would like? I’d like a survey that asks all those political compass-type questions, but at the end, instead of plotting your position on the chart, it allowed you to self identify what you think you are. I’d also like to change the methodology of these surveys, to include a null answer for topics you don’t really care about.
One of the pitfalls of a two party system is that you need to prioritize your policy positions. People don’t support Trump because he grabs pussy, they support him in spite of it, or maybe they disliked Clinton more, and the grabbage didn’t overcome that dislike.
That’s the boat I would have been in, by the way. Were I an American, I probably would have left the top of the ballot blank, or maybe held my nose and voted for Johnson, but if you held a gun to my head and told me I had to pull that lever D or R? You might not like my answer.
Now let’s tie all of that together: Would I have voted for Obama?
In 2008? Absolutely. 2008 Jeff was disgusted by Bush, and inspired by rhetoric he didn’t yet know was absolutely empty. In 2012… Probably not. 2012 Jeff was more skeptical, and Romney-Ryan was a significant improvement from McCain-Palin.
Anecdotal? Sure. But I don’t think I’m special.
The problem with this thread being active is that every time I see it I see the art for the comic strip again and at some point I noticed that I accidentally drew the windows in panels 1 and 2 at different heights and NOW I CAN’T UNSEE IT.
So it’s a cartoon about racism in a surreal world with moving windows. It’s just a science fiction political cartoon. You’re fine.
Also, the window behind you isn’t moving.
I’m sure you can find a study that questions the construct of racial resentment. Not being a sociologist myself, I don’t know off the top of my head what the response has been to that study, but it’s from 2005 and I’d be quite surprised if other scholars haven’t engaged with it since then. It certainly has a lot of citations according to Google Scholar, so it’s provoked a good deal of discussion.
I asked Philip Cohen of Family Inequality Blog whether he knew of the state of the conversation in sociology regarding “racial resentment” but it’s not his area. And unfortunately I don’t have time to do a literature review. But I wouldn’t assume that the 2005 paper was necessarily conclusive.
Amp, maybe it’s just a room from one of the houses featured on McMansion Hell?
It’s difficult to get at precisely what that means though, as the author didn’t include any methodological information in the article. We don’t know what the measure were, how factors were controlled for, etc. That’s not to mention that it’s research that’s not published in an academic journal, not peer-reviewed, and has done by what seems to be a politically motivated “policy analyst”, writing for a politically motivated media outlet. That doesn’t mean it’s wrong, just that an approach of “I analysed the data and it backs my political opinions, just trust me” doesn’t provide a lot of confidence.
Well, often they were. Literally. Someone wrote them in the first place.
Actually I know quite a bit about the process, which is the root of my distrust.
BWAHAHAHAHA.
Seriously. Even for well known studies, a significant set of them don’t replicate, or they do but not at the claimed power level.
And even for the ones which do good work, there is no preregistration so we’re only seeing them after the researchers have massaged the data (hard to catch, though sometimes possible) and after they have filtered which results they choose to report (impossible to catch without preregistration.) If you don’t publish bad or impolitic results, and if you don’t account for that bad data, then even your facially reliable results fall into question. Because in fact they may not be reliable; it’s as if you did a ton of t-tests to go “p hunting,” and you know that doesn’t work.
I don’t think all psych studies are BS–I was a psych major and I’m related to a number of psychologists and psychiatrists. But the chances seem greater than not that the average study is shitty, than that it is good, reliable, and accurate. And the shit/good ratio sure seems to increase the closer that the study is to a political goal.
And even THAT doesn’t touch study design. Which is, of course, yet another thing that affects results–as has also been documented.
And… even that doesn’t touch the questions of which studies folks choose to write on.
And… even that doesn’t touch the issue of labeling. Like, you ask the question “Are Asians smarter than whites, on average?” You can code those responses in a shitload of different ways. Depending on your goals, perhaps the yes folks are coded as “racist,” or “anti-white,” or “pro-Asian,” or “high racial resentment.” OTOH perhaps they’re coded as “truthful,” or “accurate,”, or “well aware of the slight but significant differences in average IQ score reports of adults.”
So: This 1,200 person online-selected study of online voters, which was retroactively selected for analysis of a subset of that group? Maybe it’s good or maybe it’s not, but I certainly don’t assume it’s good. Nor do I assume the good faith or competence of folks reporting on it. YMMV.
Elusis –
The racial resentment measures have been the subject of a lot of discussion since 2005. I think the 2005 paper did have an impact – it’s seemingly become standard practice since 2005 to control for political ideology and for conservative beliefs. Which is a good thing.
That said, some more recent studies seem to disprove the theory that racial resentment among conservatives is actually just a measure of conservative ideology, and has no real connection to racism.
Prejudice or Principled Conservatism? Racial Resentment and White Opinion toward Paying College Athletes – Jan 05, 2017
This is an interesting one because the policy in question – should college athletes be paid? – pits two conservative beliefs against each other (being skeptical of change vs being in favor of free markets).
The questionnaire displayed a picture of college athletes; in some cases, the accompanying photo showed three white athletes, in other cases the photo showed a mix of white and black athletes. For those with low levels of racial resentment, which photo they were shown didn’t change their results. But people with a high level of racial resentment were significantly less likely to oppose paying college athletes if shown an all-white group of athletes, and significantly more likely to oppose paying college athletes if they saw a photo that included black athletes. A follow-up study, using the most common black vs white names rather than photos, found the same results.
So racially resentful people, randomly put into the “saw a black athlete” or “saw only white athlete” groups, had their opinion changed based on the race of the people they were shown as examples. But non-racially-resentful people did not. It’s hard to explain that result if racial resentment measures aren’t truly measuring racism.
Another peer-reviewed study, this one from 2009, looked at the same policy but with different beneficiaries.
This is a fair point. Given how long academic papers take to be reviewed and published, it will probably be a year or two before any peer-reviewed research directly addressing this question comes out.
No, he definitely moved the goalposts. At first he said to most people racism means “something like ‘KKK member.'” To me, “something like ‘KKK member'” means “someone who is a member of an organized explicitly racist group, or someone who commits violence in support of racism”; it refers only to extreme racism.
Then, after I pointed out a lot of examples of people using racism to mean much less extreme forms of racism than “something like ‘KKK member'”, and posted two dictionary definitions that obviously include much less extreme forms of racism than “something like ‘KKK member,'” Sebastian shifted his definition to “would refuse to attend a wedding,” a much milder form of racism than being something like a KKK member. That is shifting the goalposts.
If Sebastian was just being hyperbolic when he said “something like ‘kkk member,'” then he can say so. People misunderstand hyperbole all the time, and I’m certainly not immune to that; it’s a common thing to happen in disagreements. But otherwise, it’s fair to say he moved the goalposts.
* * *
What, specifically, is the understanding of racism you refer to? And how do you know it’s the most common understanding?
That first link has multiple studies and study #2 some interesting discussion:
Their justification for why study #2 failed to show a link between race prime and racial resentment, but why a later study did show a link was a bit handwavey and comes down to a “we didn’t to the priming quite right”. It’s also interesting that they acknowledge that their samples for the studies that support their conclusions about “racial resentment” (Mechanical Turk) were younger and more liberal than the samples for the studies that didn’t (YouGov – CCES). This seems consistent with the study I linked above that made the point that the “racial resentment scale” is only valid for liberals.
The conservative values part of that may just be down to the authors (mis)application of conservative values to the situation:
For starters, conservatives generally don’t like to change things that have been in place for a long time, and paying college athletes is a significant change to a system that has been in place for a long time. Also, in my experience, conservatives will focus on freedom of the (non-government) institutions and organisations, not just freedom of the individuals. That is, if the institution of college sports is successful in its current form, then the government should not interfere. In being asked a question like “Do you agree or disagree that college athletes should receive a salary in addition to their scholarship?“, a conservative is quite likely to read the “should” as implying the necessity of government interference.
However, probably the most telling part of that first link is that the results that they talk about in that abstract and in the bits I’ve quoted above were not statistically significant. Coupled with the difference with study #2 it very much seems to demonstrate the problem gin-and-whiskey talked about above. Researches just keep rolling the dice with small sample after small sample until they get one conveniently fits their hypothesis, and ignore or gloss over the studies that don’t fit their predetermined beliefs.
Desipis, they seem to say that it is statistically significant for whites with high levels of racial resentment. What am I misunderstanding?
Ampersand, I was referring to the lack of p<0.05 markers in the data tables.
I’ve had another read of the paper, and I think I’ve figured out what they’ve done. They’ve plotted at the confidence intervals of two different linear regressions, and observed that for part of the domain (0.6 – 1.0) they don’t overlap.
Firstly, this isn’t what is normally referred to by statistical significance. Secondly, it’s not a valid statistical test anyway.
The critical issue is that, like much of quantitative social science research, this analysis has assumed a linear relationship between the variables. Normally if the research just trying to discover the existence of a relationship, a linear model does the job since a linear model will generally capture the basic trend of a more complex relationship. However, a linear model can be quite inconsistent with the data at particular points in the graph.
In the study it’s quite conceivable that the relationship between the “racial resentment” measure and the likelihood of support for student athlete pay isn’t linear (in either the general, white prime or black prime). It could be that there’s threshold level of “racial resentment” that drives many people to flip from supporting pay to not It could be that less than 0.5 on the scale has no effect, but that higher than there there’s an escalating impact.
If the relationship isn’t linear, then modelling as various curved functions would greatly change where the confidence intervals of the black-prime and white-prime curves go. Given the low correlation coefficients, lack of statistical significant of the linear models, and without any attempt to model the data in a non-linear fashion, it’s not really possible to have any confidence in a linear model. As such it’s not possible to have confidence that any particular part of the linear model is representative of the actual relationship in the data, and so their claim of statistical significance doesn’t hold up.
It might be a good example of why many people in the hard sciences don’t have a lot of respect for social science research. You need to listen to the data, not torture it until it says what you want it to.
I can’t get to the full text (do you have a link?) but this sounds sort of like “p hunting”.
In any study of decent size you can usually find a few things which will match significance. If the “significant things” line up with your study groups then you’re all set. But if not then a typical researcher will go “p hunting.” They’ll run stats on various groups, looking for significance. If they find it, they’ll report it as if it was some sort of pre-considered or reasonable group. But even though such results LOOK reasonable, they are not reasonable. When someone goes p-hunting, not only are the groups suspect but the entire process is suspect; if you’re gong to do that then you also need other metrics which show base levels of variance across all groups, etc.
So I read “the group of people who scored 0.6 or higher on a 0-1 test” and I think: Why not 0.5? Why not 0.7? What happens to the significance at 0.5 and 0.7? Who chose 0.6? If you think it’s variable as degree of resentment increases, why aren’t you running a different test? Why not run the whole group against each other? IOW, why are they reporting an researcher-defined subset of a single group?
It might be flowers in the end, but claims like this should always set off the BS meter.
I felt hesitant about posting a link to the full paper, because it’s sci-hub (i.e., it’s pirated), and a previous time I posted a sci-hub link someone objected (in a very nice, non-judgemental way).
But since you asked, here it is (pdf file).
Thanks. Some of those stats are beyond my knowledge level so I won’t speculate on them. In particular I can’t entirely tell (anyone know?) whether the OLS analysis across multiple variable pairs, shown in places like Table 1, already includes an adjustment for the expected “if you run enough different comparisons on random data you’ll find significance” problem; I think the test inherently addresses that issue, but I’m not positive.
I can’t say that I entirely agree with the decision to omit the un-primed control population from their subsequent studies, though. Seems sloppy.
More to the point, it seems clear that they did not find that there was a significant effect of racial resentment. If you look at Table 4 on page 10 of the PDF (numbered page 214) you’ll see that the “black names treatment x racial resentment” did not get a star.
They only found significance (if you can call it that) by looking at populations with racial resentment > 0.6, which has no particular justification other than “the number above which there’s significance.” They call that “high” and they call the rest “low” but that’s just p-hunting, like I thought; if it were really an issue they’d have shown significance in Table 4.
IMO they have no real business reporting that as “significant” with a straight face, though I think it’s an interesting paper nonetheless. And it does seem likely that there’s an effect to be found with a better study.
As always with such things, though, it’s sort of interesting how folks decide to speculate on motives. I mean, here you are with a bunch of people who just answered a computer survey. Clearly you can write it to give them a prompt like “Please explain in a sentence or two why you oppose paying athletes”. I wish they had, it would be interesting to see.
“If Sebastian was just being hyperbolic when he said “something like ‘kkk member,’” then he can say so. People misunderstand hyperbole all the time, and I’m certainly not immune to that; it’s a common thing to happen in disagreements. But otherwise, it’s fair to say he moved the goalposts.”
I meant that on the scale of “racism” which goes from “would travel to NYC to kill a black man” on one extreme of racism to “might hesitate for an extra half second trying to recognize a person of a different race” on the academic extreme side of racism, common usage is no where near the academic usage.
The common (non-specialist) usage of racism is much closer to someone who would take direct action against someone based on race (i.e. subject them to violence, or refuse to attend an inter-racial wedding).
My argument is that by (in my view) over-using ‘racism’ when its explanatory power is low (i.e. people whose levels of racism is low enough that they are willing to vote for a black person for President), we end up missing clearer reasons for why things are happening in the world.
Racism is a problem.
It is a big problem.
It isn’t the only problem.
It seems to me that at a lot of the more left-liberal places I read (i.e. here, crookedtimber, lawyersgunsandmoney) racism is being used some of the time (not all of the time) to deflect from examining how coastal, white, middle-upper class people have been screwing over non-coastal, often but not always white, lower class people.
Chris Bertram has a classic of the genre here in response to the Brexit vote. I responded in the comments at 134 and 135 with what I think applies here with Trump as well (I had all sorts of other comments but I think I crystallized it best in those two).
Racists have been voting for Republicans for a few decades now. When you see a CHANGE in voting patterns, it probably isn’t ‘about’ racist voting for Republicans because they were already voting for Republicans. The change is very likely ‘about’ something else. Now it may very well involve people who care more ‘about’ whatever the something else is than they care about the fact that Trump is racist and/or putting him in power will empower racists on various levels. You can argue that they should care more ‘about’ the racist things that Trump represents than the other things that they care about. But at that point you are studiously avoiding talking about the things they are actually voting in response to.
My read of the Rust Belt is that lots of people there have been strongly screwed over by globalism. We (in the sense of Democratic/powerful left political people) have been telling them for decades that it all shakes out, that it makes the country stronger, and that after a short adjustment period things would be fine. That hasn’t happened. In the 90s that could be construed as a mistake. By 20 years later I’m sure that feels to them more like a lie.
They got hit by the global recession harder than most people, and voted against the Republicans who had their hands in that and for a black man who said he was going to do things differently.
They then saw 8 years of the country doing better, but just like before it didn’t come to them.
Then Democrats nominate essentially the most symbolically linked person to globalism possible. She was the wife of the man who told all of those globalism promises in the 90s. She is the very woman who was supporting even more global trade acts well into her primary. She gave speeches to Goldman Sachs suggesting that you have to tell the rubes one thing while the global markets know the real things.
Their other choice was someone who said that he was done with trade treaties, and was going to fight for American jobs.
Now that man is a liar.
But we have made it very easy to be portrayed as liars on globalism too. So they chose the lies that if they had any truth at all might make their lives better, rather than the lies that had been proven over 30 years to not make their lives better. They took a chance.
It was a horribly foolish and stupid chance to take. But it wasn’t ‘about’ racism. And telling them that it was ‘about’ racism, isn’t going to win them back to voting for Democrats the way they did for the previous 30 years.
With regards to the election, I think Sebastian H is correct that most overt racists already voted Republican. But this time around, many people who DON”T consider themselves racist, and don’t want to be called racist, saw that Trump promised them jobs, but came along with a side of racism/sexism/homophobia/Islamophobia/antisemitism and said I’m OK with that and then got upset when called on it.
White, rust belt, and rural working class folks are not the only ones who have been losing out economically over the last 20 years. Nonwhite, coastal, and urban working class folks have ALSO been losing ground, but many of them are nonwhite and/or have experience with nonwhite friends and coworkers, and are therefore less likely to overlook the racism on the side of Trump’s populist pantomime. Ironically, rust belt cities are now highly dependent on import/export, and will be badly hurt by any attempts to impose tariffs (or whatever Trump is calling them).
Both Democrats and Republicans have done working class folks a disservice by pretending that there is even a chance that manufacturing and/or mining jobs will ever come back in their old form. Enabling that fantasy is what gets us the angry resentment we see now. The US is STILL one of the largest manufacturing economies in the world, but just like with agriculture, 1 person can now do the work of 100s, thanks to automation. We need to figure out something to do with/for people who live where it is unlikely that there will be many new jobs created and soon, before the next wave of automation throws millions more out of work.