Thursday, July 21, 2016

Not about Trump

This post was inspired by a story about the scarcity of Republicans among prominent rock musicians.   However, it;s not about Trump, but about an idea in the sociology of culture:  that more educated people tend to be "omnivores," who like all genres, rather than liking just a few things.  Of course, this doesn't mean that more educated people like every artist better, but that they like "quality" artists from all genres.  I recently wrote about a 2009 Pew survey that asked people how well they liked various popular musicians.  My impression is that all of the musicians are pretty well regarded by critics, so the "omnivore" idea suggests that education should have a positive correlation with feelings about each one.  The correlations, ranging from highest to lowest:

Frank Sinatra      0.157
Beatles            0.144
Bob Dylan          0.119
Bruce Springsteen  0.099
Coldplay           0.092
Madonna            0.089
Aretha Franklin    0.087
Jefferson Airplane 0.074
Grateful Dead      0.053
Rolling Stones     0.042
Michael Jackson   -0.009
Jimi Hendrix      -0.014
Elvis Presley     -0.020
Kanye West        -0.064
Nirvana           -0.093

Education has a positive correlation with views of ten, and a negative correlation with five.  Eight of the positive correlations and only two of the negative ones are statistically significant (the standard errors vary, but are mostly about 0.03).  So the omnivore hypothesis is right in a general way--more educated people tend to like a broader range of musicians.  But there are striking differences among the correlations for different musicians.  

Two factors that might affect the correlations are when the musician was most popular, since education might increase the chance of being familiar with the older ones, and whether the musician was a tabloid celebrity, which might make educated people regard them less favorably (even though the question asked about the music, it's hard not to think about the general image).  I did a regression with my ratings of when they had their musical peaks and whether they were a tabloid figure, and found some support for both ideas.  But there's still significant variation left unexplained.  The fact that education has a negative correlation with ratings of Nirvana seems particularly interesting, since although they sold a lot of records, they made a point of being outside the mainstream and not being focused on popular success, qualities that would be expected to increase their relative appeal to educated people.

[Data from the Roper Center for Public Opinion Research]

Saturday, July 16, 2016

Geography of police shootings

The Washington Post has compiled a database of fatal shootings by police in 2015 and 2016.  Breaking it down by state seems like an obvious thing to do, but I haven't seen anyone do it, so here are the rates (per 100,000) by state.


New Mexico     1.6
Wyoming        1.4
Alaska         1.1
Oklahoma       1.1
DC             1.1
Arizona        1.0
Nevada         0.9
Colorado       0.9


Mass.          0.25
Pennsylvania   0.23
Maine          0.23
Michigan       0.22
New Jersey     0.21
Rhode Island   0.19
New York       0.15
Connecticut    0.14

The general pattern is clear: fatal shootings by police are most common in the West and least common in the Northeast.  The South may be a little higher than average, but doesn't stand out:  southern states range from 10th (Louisiana) to 35th (Virginia).  Of course, there are racial and ethnic differences:  27% of those killed were black and 18% were Hispanic.  But controlling for population composition would make the West stand out even more, since the black population is relatively small in those states.

Monday, July 4, 2016

What's left?

I have seen several articles saying that Donald Trump is "to the left" of Hillary Clinton on trade (e. g., this one by Fareed Zakaria).  The idea is that support for free trade is the conservative position, and that protectionism is the liberal (or leftist) position.  This idea seems hard to square with the history of trade agreements.  For example, NAFTA was negotiated mainly by the George H W Bush administration, then passed with strong support from the Clinton administration.  It was not a party-line vote:  in the Senate, Republicans voted for it by 34-10, Democrats against by 27-28.  Supporters included Ted Kennedy and Phil Gramm; opponents included Paul Wellstone and Strom Thurmond.

But that's about political elites--what about the general public?  In 2009, the Pew Research Center asked "In general, do you think that free trade agreements like NAFTA, and the policies of the World Trade Organization, have been a good thing or a bad thing for the United States?"  Among people who said they were conservative or very conservative, 38% said "good thing" and 43% said "bad thing"; among those who said they were liberal or very liberal, 49% said good and 31% said bad (the rest weren't sure).  Democrats were more favorable (44% good, 33% bad) than Republicans (39% good, 40% bad).

Does the conventional wisdom have it backwards?  Pew also asked the question in 2008.  At that time, Republicans and conservatives were more favorable.  For example, 41% of Republicans said they were a good thing (44% bad thing), compared to only 30% of Democrats (55% bad thing).  (Overall support for "good thing" was considerably lower in 2008 than in 2009--I'm not sure why).

There were some consistent patterns--people with more education were somewhat more favorable at both times, as were people in the highest income category (over $150,000).  But whether support for free trade was a "liberal" or "conservative" position seems to depend on the party of the president.  My interpretation is that most ordinary people, even well-informed people, don't have a very definite position--almost everyone is in favor of "free trade" as long as it's "fair trade," and almost no one knows enough about the details of particular agreements to be able to say if they are "fair trade."  In recent history, presidents have consistently been in favor of free trade agreements, so the position gets associated with its most visible supporter--the incumbent president.

[Data from the Roper Center for Public Opinion Research]

Sunday, June 26, 2016

Another way to look at discrimination

It's sometimes said that most whites believe that America is a "colorblind society," or even think there is more discrimination against whites than against blacks. I've had several posts on this general issue, looking at questions that asked directly (see this post for links).  This time I'll take another approach and look at responses to distinct questions about how much discrimination there is against blacks and against whites.  In 2005, a survey sponsored by the National Conference for Community and Justice asked) "Would you say there is a great deal of discrimination, some discrimination, only a little discrimination, or none at all against. . ." various types of people.  Blacks and whites were among those asked about, but were not close to each other.  Among whites, 48% saw more discrimination against blacks, 10% against whites, and 43% chose the same category for both.  In 2106, the American National Election Studies Pilot Survey asked "How much discrimination is there in the United States today against each of the following groups?"  The possible responses were "a great deal," "a lot," "a moderate amount," "a little," and "none at all."  This time, 52% saw more discrimination against blacks, 18% against whites, and 30% chose the same category for both.

The drop in the number choosing the same category is at least partly a result of the larger number of categories in 2016.  However, there is a clear change in the ratio perceiving more against blacks to more against whites, from almost 5:1 in 2005 to about 3:1 in 2016.  There were some differences in survey procedures--2005 was a random-digit phone sample and 2016 was a weighted opt-in internet panel.  That could make a difference, although I'm not sure of the direction.  Also, in 2016, the questions about blacks and whites were separated by only one item (Hispanics).  My guess is that if this made any difference, proximity would reduce the number of whites saying there was more discrimination against whites, because some would regard it as something they "shouldn't" say.

The general result fits with others suggesting that white perceptions of discrimination against blacks fell somewhat in the early 21st century.  The general estimate of the proportion seeing more discrimination against whites (10-20%) is about the same as with the direct questions.

PS:  for blacks, the 2005 percentages were 64, 3, and 33 (more against blacks, whites, same); the 2016 percentages were 80, 4, 17.

PPS:  The 2005 questions also were asked in 1996, but unfortunately the original data don't seem to have been preserved, so it's not possible to do a parallel analysis for that year.

Saturday, June 18, 2016

World government

It appears that an overwhelming majority of Americans who have an opinion on the subject think that Britain should remain in the European Union. But how many would support the United States joining an organization like the EU? My guess is very few. But back in 1946, the Gallup Poll asked "Do you think the United Nations organization should be strengthened to make it a world government with power to control the armed forces of all nations, including the United States?" 54% said yes, and only 24% no, with the rest undecided. The question was asked again in 1946 and 1947, with similar results. In 1951, the margin was smaller, at 49-36%. In 1953 and 1955, there were narrow margins against the idea. That was the last time the question, or anything like it, was asked. Of course, opposition probably would have increased if anyone had seriously tried to implement a plan like this, but for a while many Americans were willing to at least contemplate the idea.

Thursday, June 9, 2016


Republicans are scarce at elite universities today.  A recent poll of Harvard seniors found that only 4% said they would vote for Donald Trump in a race against Hillary Clinton.  Of course, some of that is Trump (although it's not directly relevant, I can't resist mentioning that he's involved in a dispute with the Harvard Lampoon), but only 19% said that they voted for Mitt Romney in 2012.  As I observed in a previous post, things used to be different:  in 1932, polls at 17 elite colleges and universities found a median of 63% for the Republican (Herbert Hoover), and only 17% for the Democrat (FDR), with 19% going to the Socialist candidate.  In 1936, there was an even more extensive series of polls conducted at almost 100 universities.  Unfortunately, I haven't been able to track down the complete list, but I have been able to find figures for twenty colleges and universities, mostly from the Vassar Miscellany News (Oct 31, 1936), supplemented by the Princeton Alumni Weekly (Nov 6, 1936).  Here they are, ranked from biggest margin for Roosevelt over Landon on down:


                 Dem    Rep    Soc Comm
NYU              66% 20% 3%   1%
Columbia         57%    26%     6%  10%
Chicago          55% 29% 8%   8%
Johns Hopkins    54%    33%     7%   5%
Barnard         49%    37%     7%   5%
Radcliffe        52%    43%     2%   3%

Michigan         45.7%  46.3%   5%   3%
California       42%    46%     7%   5%
Harvard          45%    51%     3%   1%
Cornell          40%    52%     4%   3%
Bryn Mawr        39%    55%     5%   1%
Yale             33%    61%     4%   1%
Brown            33%    61%     3%   1%
Vassar           28%    61%     7%   4%
Smith            31%    61%     4%   1%
Dartmouth        28%    64%     5%   1%
Princeton        25%    70%     3%   0.5%
Amherst          24%    70%     5%   0.6%
Williams         19%    74%     4%   1%
Sarah Lawrence   13%    78%     7%   3%

Mean             39%    52%     5%   3.5%
National         60.8%  36.5%   0.4% 0.2%

The two small left-wing parties did much better among students at elite universities than among the public as a whole.  But a majority went for the Republicans, in a year when that party lost by what is arguably the biggest landslide in modern American history. 

It's sometimes said that class differences were sharper in 1936 than they had been in 1932.  To quote Archibald Crossley, one of the pioneers of opinion polling, "In 1932 there was a countrywide wave of protest against Hoover, reading into all income levels.  In 1936 anti-Roosevelt feeling ran high in the upper-income classes." Presumably students at these institutions were mostly from the upper income classes, but of the 14 that had conducted polls in 1932, Republican support fell in 11 of them, and fell by more than 5% at seven of them.  Of course, college students don't necessarily reflect the opinions of their parents, but I recall that one of the earliest Gallup polls asked about vote in 1932, and reported class differences were not noticeably weaker than they were in 1936. I'm not aware of any definite evidence that class differences increased.

The Socialist vote, which fell from 19% to 5% among elite college students between 1932 and 1936, fell from 2.3% to 0.4% among the general public (the socialist candidate was Norman Thomas in both elections).  So even if there wasn't a change in class alignments, it seems that there was a change in ideological alignments, in the sense that many progressives who had been skeptical of Roosevelt in 1932 were won over in 1936.  In the general public, the socialist vote was too small for this to make much difference, but it is something that people who were interested in politics would have noticed.

Wednesday, June 1, 2016

Music has charms....

One of the many controversies around the Donald Trump campaign involves the music at his campaign rallies.  A number of artists, notably the Rolling Stones, have asked him to stop playing their songs.  Of course, Trump made his selections based on what he wanted, not on market research, but a Pew survey from 2009 provides some insight into what his audience might like.  It asked people whether they liked or disliked the following musicians:  Bob Dylan, Madonna, the Grateful Dead, the Rolling Stones, Coldplay, the Beatles, Bruce Springsteen, Jimi Hendrix, Kanye West, Michael Jackson, the Jefferson Airplane, Nirvana, Elvis Presley, Frank Sinatra, Aretha Franklin.   The favorites among people who said they were conservative or very conservative and that they considered themselves a Republican or leaned Republican (lower numbers mean liked better):

Beatles                  1.64
Elvis Presley         1.66
Frank Sinatra        1.72
Aretha Franklin    1.73

And the least favorite:

Madonna              2.42
Grateful Dead      2.35
Kanye West         2.21
Nirvana                2.20

Among liberal Democrats, the favorites were:

Beatles                    1.34
Aretha Franklin      1.49
Rolling Stones        1.56
Jimi Hendrix           1.59

and the least favorite:

Madonna                 2.06
Grateful Dead         2.02
Kanye West            1.92
Jefferson Airplane  1.80

As you can see from the numbers, liberal Democrats generally liked all of the musicians better, but the rank order was similar among the two groups--the biggest exceptions were Elvis and Frank Sinatra. Those may seem obvious, but if you want something to puzzle over, the Jefferson Airplane is  16th out of 19 among liberal Democrats and 6th among conservative Republicans.