A populist puzzle

 A few days ago, Paul Krugman asked "why has [the Republican party] also remained the party of plutocrats and the enemy of any policy that might help its many working-class supporters?"  He drew a contrast with France, where  the National Rally "is if anything to the left of President Emmanuel Macron," pointing specifically to differences over the age for retirement benefits--Macron has said it needs to be increased, and Le Pen that it should be maintained and even reduced for some workers.  George Will unwittingly provided an answer on the same day.  Will complained that Macron was "retreating even from his vow to raise the pension-eligibility age from 62 to 65. Now, fighting populism with a dose of it, he suggests sending pension reform to almost certain defeat in a referendum."   That is, the National Rally is "to the left" of Macron only because Macron took a position in favor of changing the retirement system--the Republican position is similar to the National Rally's.  

To a large extent, views of retirement programs aren't part of general left/right ideology, but form a separate dimension.  One end, which you could call the "responsible" position, holds that broad-based sacrifices will be necessary--given rising life expectancy and low birth rates, people will have to work longer or pay higher tax rates to finance retirement programs.  The other end holds that no major changes are needed.  The "responsible" position is popular among elites, but not among the public, so neither Democrats or Republicans openly support it.  In the past, sometimes bipartisan groups of politicians have agreed to make cuts, but increasing partisan polarization has put an end to this (the last major effort was Barack Obama's fruitless attempt to reach a "grand bargain" with John Boehner).  

The interesting question is why Macron ever made a point of calling for an increase in the retirement age.  One possibility is simply because he didn't have much previous political experience and didn't realize that it was better to stay away from the issue.  Or perhaps there's more deference to elite opinion in France, so that the position was less unpopular than it would be here.  

 A 2018 Kaiser Family Foundation poll asked whether Congress should increase, reduce, or keep spending the same in a number of areas.  The percent saying increased minus the percent saying reduced, going from most to least support for an increase:

Education                      +64

Social Security             +46

Medicare                      +39

Medicaid                      +22

National defense          +22

Welfare programs           -8

Entitlement programs   -14

Foreign aid                    -34

There was a substantial partisan split, with Democrats more in favor of increased spending, on every item except Social Security.   On that, support was about equally strong among Democrats, Republicans, and independents.

[Data from the Roper Center for Public Opinion Research]

What's next?

 Recent columns by Paul Waldman  and Jamelle Bouie say that if Roe v. Wade is overturned, conservatives will take aim at Griswold v. Connecticut, the 1965 Supreme Court decision that struck down state laws against birth control for married couples (a 1972 decision extended that to unmarried people).  Recent surveys find that only about 10% of people say that birth control is "morally wrong," so it's safe to say that even if Griswold v. Connecticut were reversed, even the most conservative states wouldn't try to ban birth control.  But what about 1965--was the Supreme Court leading or following public opinion?  As far as I know, no surveys asked how people felt about the decision.  There weren't many relevant questions of any kind, but a Gallup survey from February 1965 asked "what situations can you think of where you would approve of the use of birth control methods, such as contraceptive pills?"  About 18% said there were none, another 13% said they didn't know or gave no answer, and 70% offered one or more reasons.  There was also a question that noted that in some states it was illegal to supply birth control information and asked if "birth control information should be available to anyone who wants it?"  82% said yes, 11% said no, and 7% didn't know.  There were also a few questions about the position of the Catholic church.  After one that mentioned that the church was opposed to the use of birth control, there was another asking "do you feel that most Roman Catholics would like to see the church change its position on the use of birth control methods, such as pills, or not?"  62% said yes, 11% said no, and 27% didn't know; among Catholics, it was 65% yes, 20% no, and 15% didn't know.  

It seems like the court was following rather than leading public opinion on this issue.  At the time, only two states, Connecticut and Massachusetts, had general laws against birth control.  Both of those had large Catholic populations, but even at that time, many Catholics didn't share the official church position, so the laws probably wouldn't have survived much longer even in those states.  

[Data from the Roper Center for Public Opinion Research]

 

 

 

 

 

 

 

 

 

 

No-knock raids

 In February, Amir Locke was killed in a "no-knock" police raid in Minneapolis.  No-knock raids seem to have started in the 1970s, and to have become considerably more common over the decades, although there's no systematic data collection.  According to a recent CNN story "there's growing consensus among policing leaders that the risks of the tactic . . . far outweigh any potential rewards."  The only information on general public opinion I could find was from a survey in June 2020 (shortly after the killing of George Floyd), which asked whether you supported or opposed "Banning no-knock warrants that allow police to enter a person's residence unannounced."  52% supported a ban and 43% opposed.  This figure shows the relationship between education and support for ban, controlling for party.  ( I show the observed levels of support and the regression lines for Democrats and Republicans.)

 

There's a good deal of support for a ban, and it's higher among more educated people, who are more politically influential.   Moreover, although there are substantial party differences, there's significant Republican support for a ban.  So it would seem like this is an issue on which you might get bipartisan support for reform.   However, 46 states allow no-knock raids, and there hasn't been a strong movement to ban (or substantially restrict) them, despite growing concern with issues of policing in recent years.  More generally, the popular and media attention hasn't produced many changes in policing or criminal justice.  Why not?  One factor is that people who are interested in politics tend to focus on the national level, but most of the relevant laws or policies are set at the state or local level.  Another possible reason is that people with influence in politics and the media are  likely be from middle and upper-class backgrounds.  Of course, that was always the tendency, but it's become stronger.  They also may be more likely to have followed a straight path rather than having to bounce around for a while before establishing themselves--I don't have systematic evidence on that, but it's my impression.  Many people say that these qualities make contemporary elites less sympathetic to ordinary people, but I don't think that is the case--today's elites probably are more sympathetic than those in the past.   But their range of experience is narrower, and that makes a difference for outcomes.  The general sense that we have to do something about racial inequality wound up being directed mostly at things that elites were familiar with, like universities and media organizations, rather than policing and criminal justice.  

[Data from the Roper Center for Public Opinion Research]

Nothing plus nothing

At the end of 2020, John Lott wrote (and Donald Trump tweeted about) a paper that purported to find evidence of fraud in the 2020 election.  Subsequently Andrew Eggers, Hartiz Garro, and Justin Grimmer tried to replicate his analysis and found a number of errors.  I also had a blog post arguing that his method of analysis produced a biased estimate of the key parameter.   The other day Andrew Gelman wrote that Lott's paper is going to be published in the journal Public Choice (at least Lott says it is--it's not on their list of forthcoming articles).   The published version is different from the 2020 version, and Eggers and Grimmer argue that it makes new errors.  However, I'll set all of that aside and assume that the basic method of analysis makes sense and the tables are correct.  Even given those assumptions, his results don't support his conclusions.

Lott's approach is to compare Trump support in absentee and in-person ballots.  His idea is that it's easier to fabricate absentee votes, so that unusually strong support for Biden in absentee ballots is evidence of fraud.  But what counts as unusual depends on the political leanings of the district, so he matches precincts in two cities where fraud is alleged to have occurred (Atlanta and Pittsburgh) with neighboring suburban precincts.  He estimates the regression equation da=p1*dv+p2*c.  da is the difference in proportion supporting Trump in absentee ballots in the two precincts, dv is the difference in proportions supporting Trump in in-person votes, and c is a dummy variable for the precinct with alleged fraud (that is, in Fulton or Allegheny counties).  So the parameter of interest is p2:  p1 is just a control for the real political differences between the precincts.  The results (from Lott's Tables 2 and 5)*:

         Atlanta                      Pittsburgh

p1     .6059***                  .3068***

        (.15)                            (.07)

p2    -.00282                   -.0025

         (.02)                         (.0075)

Standard errors are shown in parentheses (Lott shows p-values of the estimates--I used those to calculate the standard errors).  

In both cities, p2 has a negative sign--Trump did worse than expected in absentee ballots--but the estimates are small (0.3 in percentage terms) and nowhere near statistical significance (t-ratios of about 0.15 and 0.3, giving p-values of about 0.9 and 0.77).   That is, his results show no evidence of fraud in  Atlanta and no evidence of fraud in Pittsburgh.  But when Lott analyzes both together (his Table 9), things are different:  his estimate for p2 is -.034725 (3.5%), with a standard error of about .011.  Of course, the estimate when cases from both groups are considered together should be somewhere in between the estimate for either group separately.  The reason that it's not is that in analyzing the combined sample, Lott assumed that p1 was the same in both--the estimate was about 0.45, in between the Pittsburgh and Atlanta values.  If dv and c are correlated, anything that affects p1 will affect p2.  But looking at the estimates above, you clearly aren't justified in assuming that p1 is equal in both groups--a combined analysis should allow p1 to be different in Pennsylvania and Georgia.  In that case, the estimate for p2 would be about -.0026  or -.0027 and not significantly different from zero.

Sometimes an estimate that's not statistically significant in either group analyzed separately is statistically significant when all cases are analyzed together.  In Lott's words, then "the failure of some estimates to reach statistical significance arises from the small numbers of observations for each state."  But combining the samples has an effect by reducing the standard error, not by increasing the parameter estimate. 

 

 

Going to extremes

 Since 2000, some surveys have asked "Overall, would you describe the views and policies of the [party] as too extreme, or generally mainstream" for the Republican and Democratic parties.  The results, summarized as "mainstream" minus "too extreme."

The Democrats (blue line) have a consistent advantage on this--they were perceived as more mainstream than the Republicans in nine of the eleven surveys.  There is a negative trend for both parties.  It may seem strange that the last two surveys (September 2020 and Sept 2021) don't show any unusual movement for the Republican party--you might think that four years of Trump and the aftermath of Jan 6, 2021 would have made a difference.  But one of the most important principles for understanding public opinion is that most people don't pay much attention to politics, and that the people who ado tend to be committed partisans who will rationalize the actions of their side.  So unless something is at the top of the news, it won't have much impact on the party image.  Sophisticated observers tend to forget this and suggest that perceptions are strongly affected by subtle shifts in policy or "messaging."  The one unusually low point for the Republicans was during the 2013 debt ceiling standoff, which was a big news event.

[Data from the Roper Center for Public Opinion Research]

Forced to be fair

 In my last post, I said "Most people either favor free speech for their side only, or think that controversial ideas shouldn't be discussed, or should be discussed only with restrictions."  A few days, the New York Times had a major editorial titled "America has a free speech problem."  The critical responses to that editorial from "progressive" Twitter provided many examples of the first outlook.  The second one--not discussed, or discussed only with restrictions--is not as visible, because it is found mostly among people who are less interested in politics, but it is widely held.  Back in 1949, a Roper/Fortune survey asked "should colleges have courses that take up the subject of [topic], or should [topic] be discussed in class only when students ask about it, or would it be better not to discuss it at all?"  The topics were "racial and religious prejudice," "sex education," "how to be a good parent," "religious beliefs," and "Communism."  The results:

                                 Classes      When asked    No discussion     DK

Prejudice                         38%        23%              23%                  15%

Sex                                  54%        16%              16%                  13%

Parent                              62%        17%              10%                  11%

Religion                           36%       23%              28%                  12%

Communism                    35%       24%              21%                  19%

Note that support for classes was lower for the two political issues (prejudice and Communism), and higher for two of the personal issues (sex education and parenting).  

There was also a question about how colleges should "deal with the subject of socialism versus capitalism in class discussions with students."  The options were to allow professors to give only arguments in favor of capitalism (2%); require professors to give arguments in favor of both sides but allow them to express their own opinions only if favorable to capitalism (6%);  require professors to give arguments for both sides and allow them to express opinions, but try to have professors with different opinions (27%); and require professors to give arguments for both sides, but don't let them express opinions of their own (38%).  There were also 8% who said they didn't know what capitalism and socialism were, and 18% who didn't know.  None of the options represented the standard idea of academic freedom--let professors say whatever they think--but only 27% picked the one that came closest.  The most popular one was that professors shouldn't be allowed to express their own opinions.  For the questions on whether different subjects should be discussed, there were substantial differences by education--more educated people were more likely to say that they should.  For the question on capitalism and socialism, there were some educational differences in the relative popularity of the top two answers, but they were small.  Among college graduates, 44.4% favored allowing professors to express their own opinions, 43.9% favored giving arguments on both sides and not allowing opinions.  

 These questions were never repeated, but a few later ones point in the same direction.  In 1985, one asked "do you think newspapers should be allowed to take sides in their editorial pages during election campaigns?"  44% said yes, 48% no.  In 2008, there was "the government should be allowed to require newspapers to allot an equal amount of time to liberal and conservative columnists":  62% agreed and 34% disagreed.   So there is substantial popular support for requiring educators and the media to stick to facts, or limiting the opinions that they can express.  

[Data from the Roper Center for Public Opinion Research]

Speech on campus

 I've had a number of posts based on GSS questions about whether certain kinds of people should be allowed to teach in a college or university.  These questions are valuable because they have been asked over a long period, but they are pretty general and, at least for some of them, the content is not that relevant to contemporary controversies over free speech.  Also, the GSS is a survey of the general population and just records the highest level of education, not where you got it.  Therefore, it can't tell you about opinions at elite colleges and universities, which are of interest partly because people who attend them are more likely to reach positions of power, and partly because they may indicate the direction of future change.   I recently discovered that the Foundation for Individual Rights in Education and College Pulse have conducted a survey at about 150 colleges.  The samples from each college are fairly large (100-200) and they seemed to make reasonable efforts to ensure that they were representative.  They also asked questions that represented the kind of issues that come up today.  These questions started with "Student groups often invite speakers to campus to express their views on a range of topics.  Regardless of your own views on the topic, would you support or oppose your school ALLOWING a speaker on campus who promotes the following idea?"   Four of the ideas would usually be characterized as left-wing, and four as right-wing.  The left-wing ones are "looting is a justifiable form of protest," "the police should be abolished because they are racist," "white people are collectively responsible for racism and use it to protect their privilege," and "religious liberty is used as an excuse to discriminate against gays and lesbians"; the right-wing ones are "abortion should be completely illegal," "Black Lives Matter is a hate group," "the lockdown orders issued in response to the coronavirus have infringed on our personal liberties," and "transgender people have mental disorders."  Answers were on a scale of 1-4:  strongly oppose, somewhat oppose, somewhat support, strongly report.  You can add them together to get an index of tolerance of left-wing and right-wing groups. 

 Unfortunately, I couldn't download the table, so I had to enter data by hand.  I started by looking at the colleges that were unusually high and low in support for allowing speakers, then adding some in the middle, then adding some others that I was interested in, and then quit when I had about half of them.  There was a pattern emerging, which you can see here:

 

I show the indexes for both left-wing and right-wing speakers in the same figure.  The report also gave the admission rate for the college--that's on the horizontal axis.   Support for allowing right-wing speakers has little or no relation to selectivity; support for left-wing speakers is higher in more selective colleges.  Another way to look at it is that in the least selective colleges, support for allowing left and right wing speakers was about equal; in the most selective, support for allowing left-wing speakers was much higher.  This pattern could be explained by two things:  students at more selective colleges are farther to the left--that is, they will find the left-wing ideas less objectionable and the right-wing ones more objectionable; and they are more committed to the general idea of free speech.  These two things work together to increase support for allowing left-wing speakers, and offset each others to keep support for allowing right-wing speakers about the same.  As far as why students at more selective colleges are more committed to the general idea of free speech, it could be because they are more liberal, and liberals tend to be more favorable to the principle of free speech, or because they are more intellectually sophisticated, and more intellectually sophisticated people are more favorable to the principle (or some mix of both).  Both of these have historically been true, but many people say that the link to ideology has disappeared or reversed.  

It's also worth noting that the overall level of support for allowing either kind of speaker is not that high.  The possible values range from 4-16, so 10 is the middle value--mixed support and opposition.  Only one college ranks over 10 for right-wing speakers (Hillsdale).  Just about half are over 10 for the left-wing examples.  This is consistent with survey evidence going back to the 1940s--general support for free speech is an "elite" position.  Most people either favor free speech for their side only, or think that controversial ideas shouldn't be discussed, or should be discussed only with restrictions.  

If actions directly reflected opinions, less selective colleges would have more cases of "cancellation" or "no-platforming" and at those colleges, they would affect left and right-wing speakers about equally.  In reality, cases seem to be more common at elite colleges, and almost always are directed against conservative speakers.*  I may talk about why that pattern exists in a future post--at this point, I just wanted to note the discrepancy. 

*I mean incidents initiated by students or faculty.  There are some recent attempts at legal restrictions, and those have come from the right.