A few days ago, the NY Times had an article on the increasing association between the prosperity and voting patterns of counties. It noted that in 1992, there was a small tendency for poorer counties to give more support to the Republicans, but that this tendency has become stronger, and that in 2016 "the Republican Party won almost
twice the share of votes in the nation’s most destitute counties — home
to the poorest 10 percent of Americans — than it won in the richest." There is a mistake in that sentence--the poorest 10% of counties are not home to the poorest 10% of Americans (although of course they are home to some of them). I point that out not just to be pedantic, but because it will play a role in this post.
The article also discussed some examples--Montgomery County in Ohio and Macomb in Michigan. It said that both had declined in relative income, and had moved towards the Republicans. It also mentioned Gallatin County, Montana, which had gained and moved to the Democrats. However, these examples don't shed much light on the issue. If there is a relationship between county affluence and party support, then it follows that places that change their relative standing will tend to change their voting. But in order to understand why the relation arose in the first place, you should look at poor places that shifted to the Republicans and rich places that shifted to the Democrats.
I don't have data on voting at the county level going back to 1992, but I do have data for 2016, so I decided to look more deeply at the relationship between economic conditions and party support in that election. My general though was that it might depend on the distribution as well as the average level of income. I tried to download statistics on average household income by county from the Census, but found that I had the percentage in sixteen different income categories. Since I had them, I regressed the Republican lead in 2016 (Republican share minus Democratic share) on the the percentages. Based on those results, I made five groups: less than $10,000 a year, 10-15,000, 15-45,000, 45-100,000, 100-150,000, and 150,000 and up. I will call those "very poor," "poor," "lower," "middle," "upper middle," and "upper." The regression results:
Constant 1.1
v. poor -3.8
poor -0.9
lower 0 [reference]
middle -0.4
u. middle -1.1
upper -3.2
So the predicted performance of Republicans is best when more people have moderate (or moderately low) incomes, and worse when more people are well-off or poor. That is, the Democratic vote increases not just general affluence, but also the degree of inequality.
In order to try to understand the relationship, I identified some counties with extreme predicted values and small residuals. Some pro-Democratic examples:
Predicted lead Actual lead
Westchester, NY -.29 -.32 (32%-64%)
Fairfield, CT -.27 -.20 (38%-58%)
Norfolk, MA -.22 -.28 (33%-61%)
Some pro-Republican examples:
Baxter, AR .55 .54 (75%-21%)
Mifflin, PA .54 .57 (77%-20%)
Wabash, IN .53 .51 (73%-21%)
On the Republican side, I restricted it to counties with at least 10,000 households. As it happens, I know something about all of the Democratic examples--affluent suburbs of large metropolitan areas. I didn't know anything about the others, but between the data and Wikipedia a few things emerge:
1. They are not "destitute": they're places where a lot of people have moderate incomes, but few have incomes over $100,000.
2. They are rural or small town; none of them are within commuting distance of major cities.
3. It doesn't seem that any of them are or were particularly dependent on industry--they seem to have fairly diversified economies.
I will say more in future posts, but that gives me enough to think about for now.
Saturday, February 1, 2020
Tuesday, January 28, 2020
You tell me it's the institutions
Last week Yuval Levin had a piece in the New York Times saying that there had been a "collapse of our confidence in institutions — public, private, civic and political." He proposed an explanation: "we lose faith in an institution when we no longer believe that it plays
this ethical or formative role of teaching the people within it to be
trustworthy" and that most institutions are no longer doing this. He mentioned politics, "the sciences, ... law and ... other professions meant to offer expertise," "the academy," religion, and "artists and athletes." He then said "the military is the most conspicuous exception and also the most
unabashedly formative of our national institutions — molding men and
women who clearly take a standard of behavior and responsibility
seriously."
I had a post a few years ago about confidence in various institutions, and it has declined for most since the 1970s, with the military as the one clear exception. So that seems to fit with Levin's account. But I also had a few posts about a lesser-known series of data about the "honesty and ethical standards" of various occupations and professions. There have been some declines, but more increases. For example, evaluation of the standards of funeral home directors and realtors have generally increased since the 1970s; lawyers declined from the 1970s to about 1990, but have increased since then. In this group, the military no longer seems so exceptional.
That suggests another explanation, which I implied in my earlier posts: confidence has declined in institutions that are connected to politics, especially national politics. Most of the "big" institutions--education, business, religion, etc. are involved in politics, either because they taking sides or because they are subject of political controversies. People don't like political conflict, so as polarization has increased, confidence in these institutions gets pulled down as well. The military is unusual because it has a strong tradition of staying out of partisan politics and because both parties avoid criticizing it. Although Democrats tend to favor lower military spending, they rarely make negative statements about the military--they just say that we already have the strongest military in the world. As a result, the military doesn't get pulled down by association with political controversy--it may even gain because it seems to be above politics.
The larger issue here is that Levin holds that political polarization is a symptom of a general social crisis: "We can see that in everything from vicious partisan polarization to rampant culture-war resentments to the isolation, alienation and despair that have sent suicide rates climbing and driven an epidemic of opioid abuse." The alternative, which seems more convincing to me, is that political polarization is about politics, not a symptom of something else, and that there are specific social problems but no general social crisis.
I had a post a few years ago about confidence in various institutions, and it has declined for most since the 1970s, with the military as the one clear exception. So that seems to fit with Levin's account. But I also had a few posts about a lesser-known series of data about the "honesty and ethical standards" of various occupations and professions. There have been some declines, but more increases. For example, evaluation of the standards of funeral home directors and realtors have generally increased since the 1970s; lawyers declined from the 1970s to about 1990, but have increased since then. In this group, the military no longer seems so exceptional.
That suggests another explanation, which I implied in my earlier posts: confidence has declined in institutions that are connected to politics, especially national politics. Most of the "big" institutions--education, business, religion, etc. are involved in politics, either because they taking sides or because they are subject of political controversies. People don't like political conflict, so as polarization has increased, confidence in these institutions gets pulled down as well. The military is unusual because it has a strong tradition of staying out of partisan politics and because both parties avoid criticizing it. Although Democrats tend to favor lower military spending, they rarely make negative statements about the military--they just say that we already have the strongest military in the world. As a result, the military doesn't get pulled down by association with political controversy--it may even gain because it seems to be above politics.
The larger issue here is that Levin holds that political polarization is a symptom of a general social crisis: "We can see that in everything from vicious partisan polarization to rampant culture-war resentments to the isolation, alienation and despair that have sent suicide rates climbing and driven an epidemic of opioid abuse." The alternative, which seems more convincing to me, is that political polarization is about politics, not a symptom of something else, and that there are specific social problems but no general social crisis.
Monday, January 20, 2020
Coastal elites
A couple of years ago, I proposed that "American society has become a lot more socially egalitarian over the
last 60 years or so. Educated people don't want to be thought of as
snobs or elitists, and less educated people are less likely to think
they should 'improve themselves' by emulating the middle class. At one
time, you could say that Democrats thought of themselves as the party of
the common people, and Republicans thought of themselves as the party
of successful people. Now both parties think of themselves as the party
of the common people, plus the fraction of the elites who care about or
understand the common people." I have looked for evidence bearing on this point, and report two examples in this post: white people have become less likely to say that racial inequality is due to things that could be regarded as the fault of black people, and people have become more likely to say if someone is poor it's the result of "circumstances beyond their control" rather than "lack of effort."
My idea suggested that the change would be larger in the advantaged group--that is, that people would be less likely to be critical of people "below" them. That is the case for race--blacks became a little more likely to say that racial inequality was the result of lack of effort or ability. I didn't look at opinions about the causes of poverty then, but here are results for 1964 and 2018 broken down by education:
People at all educational levels have moved towards "circumstances," but the move is larger among more educated people. As a result, the relation between education and opinions has changed. In 1964, more education went with more chance of saying that poverty was due to lack of effort; in 2018, the correlation between education and opinions was essentially zero. There was a relationship in 2018, but it wasn't linear: people at middle levels of education were most likely to say that poverty was due to lack of effort. The 2018 survey also distinguished people with graduate education, and they were even more likely to say that poverty was due to circumstances.
Of course, these are just two groups, so you could say that maybe elites have more contempt for certain kinds of people, like those who didn't graduate from college or those who live in the "heartland." I have looked for more general survey questions, but haven't been able to find anything very relevant (although this one is interesting).
On the other hand, how much evidence is there for a rise in elite condescension? The go-to example is Hillary Clinton's "basket of deplorables"--in which she said that not all Trump supporters were racists, xenophobes, etc. She didn't say anything explicit about the social position of the two "baskets," but since her point was that many Trump supporters has legitimate economic grievances, that implies that working-class supporters were less likely to be "deplorables." So it was a foolish remark, but not an example of contempt for the working class. Then last week I saw the NY Times review of a new book by Nicholas Kristof. The reviewer (Sarah Smarsh) wrote "even well-intentioned, coastal, college-educated scribes commit obliviously condescending word choices ('flyover country')...." That reminded me that I recently read Paul Collier's The Future of Capitalism and noticed that he also cited the expression as an example of elite condescension. Although I am coastal (grew up in NJ and lived most of my adult life in Connecticut) and college-educated (at an elite coastal institution), I'm not sure I've ever heard or seen "flyover country" used in a straightforward way--just in an indignant or ironic way. Finally, a couple of weeks ago, I followed a link from Ross Douthat to an article written right after the 2016 election by Walter Russell Mead. He said "the Republican Party’s extraordinary dominance in this election demonstrates just how costly the Democrats’ scornful rejection of 'hillbilly populism' has been." Finally, a good example of elite contempt! I didn't remember anyone using this term, so I did a search, and the first thing that came up on Google was Mead's own article, then something from a book discussing Orville Faubus, then another book talking about an "older South," ..... nothing from a Democrat during the 2016 election. OK, it would have been a good example if anyone had actually said it.
[Data from the Roper Center for Public Opinion Research]
My idea suggested that the change would be larger in the advantaged group--that is, that people would be less likely to be critical of people "below" them. That is the case for race--blacks became a little more likely to say that racial inequality was the result of lack of effort or ability. I didn't look at opinions about the causes of poverty then, but here are results for 1964 and 2018 broken down by education:
People at all educational levels have moved towards "circumstances," but the move is larger among more educated people. As a result, the relation between education and opinions has changed. In 1964, more education went with more chance of saying that poverty was due to lack of effort; in 2018, the correlation between education and opinions was essentially zero. There was a relationship in 2018, but it wasn't linear: people at middle levels of education were most likely to say that poverty was due to lack of effort. The 2018 survey also distinguished people with graduate education, and they were even more likely to say that poverty was due to circumstances.
Of course, these are just two groups, so you could say that maybe elites have more contempt for certain kinds of people, like those who didn't graduate from college or those who live in the "heartland." I have looked for more general survey questions, but haven't been able to find anything very relevant (although this one is interesting).
On the other hand, how much evidence is there for a rise in elite condescension? The go-to example is Hillary Clinton's "basket of deplorables"--in which she said that not all Trump supporters were racists, xenophobes, etc. She didn't say anything explicit about the social position of the two "baskets," but since her point was that many Trump supporters has legitimate economic grievances, that implies that working-class supporters were less likely to be "deplorables." So it was a foolish remark, but not an example of contempt for the working class. Then last week I saw the NY Times review of a new book by Nicholas Kristof. The reviewer (Sarah Smarsh) wrote "even well-intentioned, coastal, college-educated scribes commit obliviously condescending word choices ('flyover country')...." That reminded me that I recently read Paul Collier's The Future of Capitalism and noticed that he also cited the expression as an example of elite condescension. Although I am coastal (grew up in NJ and lived most of my adult life in Connecticut) and college-educated (at an elite coastal institution), I'm not sure I've ever heard or seen "flyover country" used in a straightforward way--just in an indignant or ironic way. Finally, a couple of weeks ago, I followed a link from Ross Douthat to an article written right after the 2016 election by Walter Russell Mead. He said "the Republican Party’s extraordinary dominance in this election demonstrates just how costly the Democrats’ scornful rejection of 'hillbilly populism' has been." Finally, a good example of elite contempt! I didn't remember anyone using this term, so I did a search, and the first thing that came up on Google was Mead's own article, then something from a book discussing Orville Faubus, then another book talking about an "older South," ..... nothing from a Democrat during the 2016 election. OK, it would have been a good example if anyone had actually said it.
[Data from the Roper Center for Public Opinion Research]
Friday, January 10, 2020
The Golden Age?
Ross Douthat had a looking-back-at-the-decade column in which he said "the 1990s, was a period of hubris,
when we half-believed that we were entering a new age of domestic
dynamism and global power — that our leaders deserved trust again ...." Or did we? General confidence in government rose in the first half of the 1980s, but then declined, hitting a record low in 1990, bouncing back in 1992, and then falling to new low in 1994. George Bush, despite success in the cold war and the first Gulf War, not only lost his bid for re-election in 1992, but got a smaller share of than Jimmy Carter had in 1980. Ross Perot got 19% of the vote in that election even though he had no political background and wasn't particularly well known before he ran for president. Bill Clinton had a fairly high approval rating when he took office, but it declined rapidly--by June 1993, it was under 40%. There was no dramatic negative event in that period--as I recall, there were just a number of relatively minor missteps but people judged them harshly. Then there was the 1994 election, when the Republicans gained 50 seats and got their first majority in the House of Representatives in 40 years. We've become used to Republican majorities over the years since then, so people have forgotten just how unusual this was--not only hadn't there been a Republican majority for 40 years, but there hadn't even been any close calls. For example, after Reagan's landslide re-election in 1984, the Democrats still had a 252-182 majority in the House.
Overall, the public mood was very negative in the first half of the 1990s--it didn't start to improve until about 1995. I don't have any explanation to offer--I just find it intriguing.
While thinking about this, it occurred to me that I've sometimes suggested that Ross Perot appealed to the same general sentiments of "America First" that Donald Trump did, but I haven't actually looked at the evidence. I looked at a survey in June 1993 that had questions about immigration, NAFTA, and a number of questions about economic competition with Japan. Opinions about NAFTA and Japan had little or no connection with reported vote in the 1992 election (Perot later became known for opposition to NAFTA, but at that time most people said they didn't know what his position was, and many said they didn't know what Bill Clinton's position was). Opinions about immigration did make a difference, though--support for Perot was 8% among people who who thought immigration levels should be increased, 11% among those who thought they should stay the same, and 16% among those who thought it should be reduced. 72% of the people who reported voting for Perot thought that immigration should be reduced, compared to 62% of Bush supporters, 62% of non-voters, and 55% of Clinton supporters.
That's not overwhelming, but it gives some support to my ideas about Perot's appeal.
[Data from the Roper Center for Public Opinion Research]
Overall, the public mood was very negative in the first half of the 1990s--it didn't start to improve until about 1995. I don't have any explanation to offer--I just find it intriguing.
While thinking about this, it occurred to me that I've sometimes suggested that Ross Perot appealed to the same general sentiments of "America First" that Donald Trump did, but I haven't actually looked at the evidence. I looked at a survey in June 1993 that had questions about immigration, NAFTA, and a number of questions about economic competition with Japan. Opinions about NAFTA and Japan had little or no connection with reported vote in the 1992 election (Perot later became known for opposition to NAFTA, but at that time most people said they didn't know what his position was, and many said they didn't know what Bill Clinton's position was). Opinions about immigration did make a difference, though--support for Perot was 8% among people who who thought immigration levels should be increased, 11% among those who thought they should stay the same, and 16% among those who thought it should be reduced. 72% of the people who reported voting for Perot thought that immigration should be reduced, compared to 62% of Bush supporters, 62% of non-voters, and 55% of Clinton supporters.
That's not overwhelming, but it gives some support to my ideas about Perot's appeal.
[Data from the Roper Center for Public Opinion Research]
Monday, December 30, 2019
Nationalism and internationalism
In 1964, a special Gallup survey asked people for their reactions to the following statement: "We shouldn't think so much in international terms but concentrate more on our own national problems and building up our strength and prosperity here at home." 55% agreed and 34% disagreed (11% had no answer). But there were large differences by education and age--the chance of disagreeing was less than 10% for people in their 80s with a grade school education, but about 2/3 for college graduates in their 20s--so it would have been reasonable to expect a gradual move away from agreement. The question has been asked a number of times since then; the figure shows the percent who disagree:
There's a lot of short-term variation, but if there is any trend it is towards agreeing that we should "concentrate more on our own national problems." In the most recent surveys, the educational differences are still there (and about as strong as they were in 1964), but the age differences have disappeared.
The same survey contained a question about immigration, which I've written about before. In 1964, only seven percent said that immigration should be increased and 42 percent said it should be reduced; in the most recent survey, 28% said increased and 29% should be reduced.
What I think this shows is that a growth in "cosmopolitanism" doesn't necessarily mean a decline in nationalism, in the sense of a belief that you have stronger obligations to members of your own nation than to humanity in general. A few days ago Pico Iyer had a piece in the New York Times in which he said "I’m delighted to return to a newly open and creative London where the average person was born in another country. My four grandparents, all born in India, came of age in a richly multicultured society, but one in which they had little chance of encountering neighbors from Cambodia or Haiti or Ethiopia, as so many New Yorkers or Angelenos can today." That's the sentiment that has grown--that ethnic diversity is one of the good things about your nation, or that more ethnic diversity would make it better. Of course, not everyone feels that way--there are people who liked London the way it used to be and are sorry to see changes--but that is a difference between two kinds of nationalism.
[Data from the Roper Center for Public Opinion Research and Pew Research Center]
There's a lot of short-term variation, but if there is any trend it is towards agreeing that we should "concentrate more on our own national problems." In the most recent surveys, the educational differences are still there (and about as strong as they were in 1964), but the age differences have disappeared.
The same survey contained a question about immigration, which I've written about before. In 1964, only seven percent said that immigration should be increased and 42 percent said it should be reduced; in the most recent survey, 28% said increased and 29% should be reduced.
What I think this shows is that a growth in "cosmopolitanism" doesn't necessarily mean a decline in nationalism, in the sense of a belief that you have stronger obligations to members of your own nation than to humanity in general. A few days ago Pico Iyer had a piece in the New York Times in which he said "I’m delighted to return to a newly open and creative London where the average person was born in another country. My four grandparents, all born in India, came of age in a richly multicultured society, but one in which they had little chance of encountering neighbors from Cambodia or Haiti or Ethiopia, as so many New Yorkers or Angelenos can today." That's the sentiment that has grown--that ethnic diversity is one of the good things about your nation, or that more ethnic diversity would make it better. Of course, not everyone feels that way--there are people who liked London the way it used to be and are sorry to see changes--but that is a difference between two kinds of nationalism.
[Data from the Roper Center for Public Opinion Research and Pew Research Center]
Friday, December 27, 2019
The New Age
David Brooks made an interesting observation in his column: "We’ve seen gigantic events like impeachment, the Kavanaugh hearings, the
Mueller investigation and the 'Access Hollywood' tapes They come and
go and barely leave a trace on the polls, the political landscape or
evaluations of Donald Trump." According to the records kept by the Roper Center, Trump's lowest approval rating has been 32% and his highest has been 49%, and those are individual polls which are subject to sampling error--the averages of polls taken at about the same time have only ranged from about 35% to 45%. That's a lot less variation than any of his predecessors (starting with Roosevelt). But if you look more closely, it seems that the decline in volatility began with Obama. He started with very high approval ratings, around 70%. Those declined pretty steadily and by Fall 2009 he was at a little over 50%. After that, they only ranged from the low 40s to the low 50s--about the same range of variation as Trump, at a higher level of approval. All presidents from Roosevelt through GW Bush had more substantial ups and downs during their whole time in office.
So what changed with Obama? In a previous post, I suggested that the Republican strategy of complete opposition meant that he couldn't have any real political successes--even when he got his way, as with the Affordable Care Act, it looked ugly. In contrast, all previous presidents had some achievements that got bipartisan support. Trump has been like Obama in this respect--Democrats have been united in opposition and he hasn't made much effort to get anyone to break ranks.
This reminded me of a piece by John Sides in the Washington Post a few days ago. He had a figure showing that from Kennedy to GW Bush, presidential approval was related to "consumer sentiment" (ratings of economic conditions)--with Obama and Bush, the connection has disappeared.
[Data from the Roper Center for Public Opinion Research]
So what changed with Obama? In a previous post, I suggested that the Republican strategy of complete opposition meant that he couldn't have any real political successes--even when he got his way, as with the Affordable Care Act, it looked ugly. In contrast, all previous presidents had some achievements that got bipartisan support. Trump has been like Obama in this respect--Democrats have been united in opposition and he hasn't made much effort to get anyone to break ranks.
This reminded me of a piece by John Sides in the Washington Post a few days ago. He had a figure showing that from Kennedy to GW Bush, presidential approval was related to "consumer sentiment" (ratings of economic conditions)--with Obama and Bush, the connection has disappeared.
[Data from the Roper Center for Public Opinion Research]
Saturday, December 21, 2019
Old news
Sometimes I start on a post but abandon it, usually because it no longer seems interesting or gets too complicated. One of these times was in August, about an article that was part of the New York Times 1619 project. I was reminded of it by a letter to the Times criticizing some aspects of the project, and after digging up my analysis, I decided it was worth writing about, although it is kind of complicated.
The article, by Linda Villarosa, was called "How false beliefs in physical racial differences still live in medicine today." Specifically, it was about the belief that blacks don't feel as much pain as whites. It started with an account of a 19th century physician who believed that blacks had thicker skin and conducted brutal experiments to try to find evidence for his hypothesis. Then it moved to the present and cited a review of studies that concluded that "black and Hispanic people .... received inadequate pain management compared with white counterparts."
Then came the part that caught my attention. It said that a 2016 survey found that "when asked to imagine how much pain white or black patients experienced in hypothetical situations, the medical students and residents insisted that black people felt less pain." I was curious about how big the differences were, so I read the paper. It described a study that gave each medical student a pair of hypothetical cases, one black and one white, and asked them to rate how much pain the patient was likely to be feeling and how it should be treated (opioids vs. something weaker). There were several scenarios, and they were rotated so that each respondent got a different one for his or her cases but the total number in each was the same for the black and white examples.
It didn't report the mean pain ratings for hypothetical black and white cases, but showed this figure:
Figure A shows the average pain rating for the black and white case by number of false beliefs about physical differences between the races. Medical students who had a high number of false beliefs rated the white cases as experiencing more pain; medical students who had a low number of false beliefs rated the black cases as experiencing more. High and low were defined relative to the mean, so that implied that medical students with average numbers of false beliefs rated the black and white cases about the same.
The authors included their data as a supplement to the article, so I downloaded it and calculated the means. The average rating for the black cases was 7.622, on a scale of 1-10, while the average rating for the white cases was 7.626--that is, almost identical. The study also asked how the different cases should be treated--135 gave the same recommendation for both of their cases, 40 recommended stronger medication for their white case, and 28 for their black case. Since the total distribution of conditions was the same for the black and white cases, this means that in this sample, treatment recommendations were different for black and whites. However, the difference was not statistically significant at conventional levels (p is about .14)--that is, the sample difference could easily have come up by chance.
So you could conclude that, in this sample, there is no evidence that medical students rate the pain of blacks and whites differently, but perhaps some evidence that they treat white pain more aggressively. (If you just went by statistical significance, you would accept the hypothesis that they treat hypothetical black and white cases the same, but a more sensible conclusion would that you should collect more data). The paper, however, didn't do this. They used pain ratings to predict treatment recommendations for black and white cases, and then looked at the predicted treatment differences for students with high and low numbers of false beliefs: "participants who endorsed more false beliefs (+1 SD) were less accurate in their treatment recommendations for the black target compared with the white target [β = 0.15, SE = 0.06, t(192) = 2.47, P = 0.014]. Conversely, participants who endorsed fewer false beliefs (−1 SD) did not differ in their treatment recommendation accuracy [β = −0.06, SE = 0.06, t(192) = −1.05, P > 0.250]."
The problem with this analysis is that, to quote the title of an article by Andrew Gelman and Hal Stern "The difference between 'significant' and 'not significant' is not itself statistically significant." That is, if you tested the hypothesis that β had equal magnitude and opposite signs for black and white cases--that treatment recommendations were affected by ratings of pain but not by race--you would not be able to reject it.
So to summarize, the statement that "the medical students and residents insisted that black people felt less pain" is false: they rated black and white pain as virtually equal. I don't blame Villarosa for that--the way it was written, I could see how someone would interpret the results that way. I don't really blame the authors either--interaction effects can be confusing. I would blame the journal (PNAS) for (1) not asking the authors to show means for the black and white examples as standard procedure and (2) not getting reviewers who understand interaction effects.
On the more general 1619 project, my thoughts are:
1. Most of the articles I read weren't very convincing
2. but it's a perspective that deserves to be heard
3. and it was published in the NY Times Magazine, which doesn't claim to be a straight news section, but to give perspectives and interpretations
The article, by Linda Villarosa, was called "How false beliefs in physical racial differences still live in medicine today." Specifically, it was about the belief that blacks don't feel as much pain as whites. It started with an account of a 19th century physician who believed that blacks had thicker skin and conducted brutal experiments to try to find evidence for his hypothesis. Then it moved to the present and cited a review of studies that concluded that "black and Hispanic people .... received inadequate pain management compared with white counterparts."
Then came the part that caught my attention. It said that a 2016 survey found that "when asked to imagine how much pain white or black patients experienced in hypothetical situations, the medical students and residents insisted that black people felt less pain." I was curious about how big the differences were, so I read the paper. It described a study that gave each medical student a pair of hypothetical cases, one black and one white, and asked them to rate how much pain the patient was likely to be feeling and how it should be treated (opioids vs. something weaker). There were several scenarios, and they were rotated so that each respondent got a different one for his or her cases but the total number in each was the same for the black and white examples.
It didn't report the mean pain ratings for hypothetical black and white cases, but showed this figure:
Figure A shows the average pain rating for the black and white case by number of false beliefs about physical differences between the races. Medical students who had a high number of false beliefs rated the white cases as experiencing more pain; medical students who had a low number of false beliefs rated the black cases as experiencing more. High and low were defined relative to the mean, so that implied that medical students with average numbers of false beliefs rated the black and white cases about the same.
The authors included their data as a supplement to the article, so I downloaded it and calculated the means. The average rating for the black cases was 7.622, on a scale of 1-10, while the average rating for the white cases was 7.626--that is, almost identical. The study also asked how the different cases should be treated--135 gave the same recommendation for both of their cases, 40 recommended stronger medication for their white case, and 28 for their black case. Since the total distribution of conditions was the same for the black and white cases, this means that in this sample, treatment recommendations were different for black and whites. However, the difference was not statistically significant at conventional levels (p is about .14)--that is, the sample difference could easily have come up by chance.
So you could conclude that, in this sample, there is no evidence that medical students rate the pain of blacks and whites differently, but perhaps some evidence that they treat white pain more aggressively. (If you just went by statistical significance, you would accept the hypothesis that they treat hypothetical black and white cases the same, but a more sensible conclusion would that you should collect more data). The paper, however, didn't do this. They used pain ratings to predict treatment recommendations for black and white cases, and then looked at the predicted treatment differences for students with high and low numbers of false beliefs: "participants who endorsed more false beliefs (+1 SD) were less accurate in their treatment recommendations for the black target compared with the white target [β = 0.15, SE = 0.06, t(192) = 2.47, P = 0.014]. Conversely, participants who endorsed fewer false beliefs (−1 SD) did not differ in their treatment recommendation accuracy [β = −0.06, SE = 0.06, t(192) = −1.05, P > 0.250]."
The problem with this analysis is that, to quote the title of an article by Andrew Gelman and Hal Stern "The difference between 'significant' and 'not significant' is not itself statistically significant." That is, if you tested the hypothesis that β had equal magnitude and opposite signs for black and white cases--that treatment recommendations were affected by ratings of pain but not by race--you would not be able to reject it.
So to summarize, the statement that "the medical students and residents insisted that black people felt less pain" is false: they rated black and white pain as virtually equal. I don't blame Villarosa for that--the way it was written, I could see how someone would interpret the results that way. I don't really blame the authors either--interaction effects can be confusing. I would blame the journal (PNAS) for (1) not asking the authors to show means for the black and white examples as standard procedure and (2) not getting reviewers who understand interaction effects.
On the more general 1619 project, my thoughts are:
1. Most of the articles I read weren't very convincing
2. but it's a perspective that deserves to be heard
3. and it was published in the NY Times Magazine, which doesn't claim to be a straight news section, but to give perspectives and interpretations
Subscribe to:
Posts (Atom)