Sunday, September 27, 2020

There you go again

 Journalists love to identify turning points--the gaffe that sank a campaign, or the snappy comeback that snatched victory from the jaws of defeat.  Academic studies regularly find no evidence that individual events make much difference, but the pundits are undeterred.  In today's NY Times Book Review (it came out last month online),  a review of Rick Perlstein's Reaganland by Evan Thomas talked about the 1980 election:  

 "At their final debate in late October, virtually tied in the polls, Carter started in on Reagan for having advocated, 'on four different occasions, for 'making Social Security a voluntary system, which would, in effect, very quickly bankrupt it.'   ..... Now, Carter warned, Reagan was trying to block national health insurance."

"As Perlstein tells it, Reagan looked at Carter smilingly, his face betraying 'a hint of pity.'  Then the old cowboy rocked back, and with an easy, genial chuckle, delivered the knockout blow. 'here you go again!' he said, beaming. The audience gave a 'burst of delighted laughter. … Jimmy Carter was being mean again.'

 "With one deft jab, Reagan had finished off his opponent. A few days later, the Republican candidate won in an electoral vote landslide."

What's wrong with this story?

First, although 1980 was an "electoral vote landslide", Reagan got only 50.8% of the popular vote, against 41% for Carter, a 9.8% gap.  That was a solid win, but closer to 1996, when Bill Clinton won by   49.2% to 40.7%, than to a real landslide like 1984 (58.8% to 40.6%).

 Second, even after the final debate, Carter was close to Reagan in the polls.  I found two that were taken between the debate and the election:

 Carter    Reagan    Anderson     

43%        44%        8%            CBS/NYT        Nov 1

43%        46%        7%            Gallup              Nov 1

I don't know whether the problem was a late swing, miscalculation of likely voters, or something else, but the pre-election polls were pretty far off that year, overestimating Carter's vote by 3-4% and underestimating Reagan's by a similar margin.  That's the difference between a close election and a substantial win. 

 Right after the election, several polls asked people how they had voted

 41%        49%       7%           CBS/NYT

43%        45%        6%           Gallup

38%        50%        6%            Roper

Most people who watched the debate thought Reagan did better than Carter, and "there you go again" may have contributed to that.  But the idea that it was a "knockout blow" is implausible to start with--it's hard to imagine how that would cause anyone to change a vote from Carter to Reagan.  If you want to stick with the boxing metaphors, you could say that he "ducked a punch."

 The review also recounted another popular idea that may not be true:  "In 1968, 17-year-old Patrick Caddell polled a working-class neighborhood in Jacksonville, Fla., about the upcoming presidential race for a high school project. He was surprised to hear, again and again, 'Wallace or Kennedy, either one.'" Thomas suggested that both appealed to a sense of anger in the working class.  A few years ago, I had a post about hypothetical races involving Kennedy, Eugene McCarthy, or Hubert Humphrey vs. Nixon and Wallace and there was no sign that Kennedy cut into Wallace's support.  An alternative explanation for what Caddell found is that most southern whites were still Democrats back then.  John F. Kennedy had been a popular president, and his assassination had added to his reputation.  So when people in Jacksonville were asked who they liked for president, they just went for the names they knew--a Kennedy and the governor of the neighboring state.  


[Data from the Roper Center for Public Opinion Research]

Friday, September 25, 2020

Breaking it down

At one time, knowing whether someone was a blue-collar or white collar worker helped you to predict how they voted--blue-collar workers were more likely to vote for the Democrats.  That's no longer true, or is true to only a slight extent.  But casual observation suggests that there are differences among occupations--for example, if someone is a university professor, it's a good bet that he or she voted for the Democrats (or at least didn't vote Republican).  It's hard to go beyond casual observation, since most opinion surveys don't ask about occupation.  However, the General Social Survey asks about exact occupation and codes it into detailed Census categories, and you can combine years to get a larger sample.  I did this for two periods:  1972-88 (which covered the 1968-1984 presidential elections) and 2000-2018 (1996-2016).  I computed the occupation effects on vote for president (omitting people who supported third-party candidates), controlling for race (black/non-black), education (have/don't have a college degree) and log of family income.  This let me get estimates for about 350 occupations.   The most Democratic and most Republican occupations in each period:

Most Democratic

1968-84                                                                                                1996-2016

Maintenance workers, machinery                                            Biological Scientists

Reservation & transportation agents                                       Judges and magistrates

Actors                                                                              Medical records technicians

Meter readers, utilities                                                            Writers and authors

Weighers, measurers, checkers                            Social and human service assistants

Pumping station operators                                                     News analysts, reporters

Procurement clerks                                                      Tax examiners and collectors

Other education, training, and library              Public relations & fundraising managers

Crushing, grinding, mixing workers                               Psychologists*

Therapists, all other                                              Food service workers, non-restaurant


Most Republican


Optometrists                                                       Credit authorizers, checkers & clerks

Health practitioner support techs                                          Engineers, all other

Printing, binding, and finishing workers                              Logging workers

Computer & info. systems managers        Office machine operators, exc. computer

Architectural and engineering managers                 Print binding and finishing workers

Pharmacists                                                                                 Pharmacists

Chemical processing machine operators               Dining room & cafeteria attendants

Industrial engineers                                                                   economists*

Insurance underwriters                                                             Parts salespersons

Software developers, applications & systems                      upholsterers


Because of small numbers, there's a good deal of uncertainty in the estimates for many of these occupations (or you could say that to a large extent being in the extremes is a matter of chance), but there's still a noticeable pattern:  the most Democratic occupations of the early period were mostly blue collar or lower white collar, while the most Democratic ones of the later period are mostly professional; the most Republican go from mostly managers and professionals to a wide-ranging mix. 

To get a more general picture, I used the top-level census categories.  There seemed to be a lot of variation within "professional and related occupations", so I broke that one into the next level.  Here is a figure showing the relative positions of the groups in the early and recent periods:

The ones above the line moved towards the Democrats (in relative terms).  The biggest moves were legal; education, training, and library; "life, physical, and social science"; and computer and mathematical occupations.  On the other side, resource, construction and maintenance, architecture and engineering, production, and social service (which includes clergy) moved towards the Republicans.  The changes led to a good deal of reshuffling of relative positions--resource, construction, and maintenance was the 10th most Republican (out of 14) group in the early period and third most Republican in the late one; on the other side, computer and mathematical occupations went from second to tenth.  Legal occupations were the second most Democratic in the early period, and became even more Democratic in the late period. 

*Professors of psychology or economics are classified as "post-secondary teachers," so these are psychologists and economists who work outside of universities. 

Saturday, September 19, 2020

The Story of the Decade, part 3

 This post will consider the claim that "elites" have angered ordinary people by treating them with contempt.  I've looked for data on "elite" views of working class people, or less educated people, and haven't found anything directly relevant.*  I've also mentioned the weakness of the evidence for elite contempt.  Now I'll take a glance at history.  In the late 19th and early 20th centuries, a number of middle-class people went out and observed the lives of the poor and working classes.  I recently ran across (I forget how) an example published in 1905, by Elsa Herzfeld, a student at Barnard College, who reported on twenty-four families in Manhattan.  The families were all white and that the husbands all had jobs, mostly in factories or construction.  Her research was supervised by Elsie Clews Parsons, who was then a lecturer in sociology and went on to a distinguished career as an anthropologist.    Before her detailed discussion of individual families, Herzfeld offered some general observations:  "even those members of the family who have had a fairly good amount of schooling possess small reasoning powers.  They show some small curiosity, but it is rather that of a child, than the intellectual desire to know.  They lose patience when they are unable to comprehend a thing in the beginning.  The questions they ask are frequently childish ones."  There were also critical comments on their work habits and standards of child care.  Parsons provided an introduction to the book, and asked "what are the schools . . . the churches, the settlements, etc., doing to improve home life?  Do they teach boys and girls anything about their obligations in marriage and child-rearing?"  That is, both Herzfeld and Parsons wanted to help the poor, but they seemed confident that the best way to help them was to teach them to emulate the middle class, at least to the extent that they could given their intellectual limitations.  There was no suggestion that their situation might be the result of bad luck or systemic injustice, and certainly not that the middle class had anything to learn from them.  This is just one example, but it was consistent with other things I've read from that period, and totally different from anything you'd see today.  Rather than increasing condescension and contempt, there has been increasing social egalitarianism.   

Finally, there's the idea that people are turning to politics as a substitute for religion or something else that is missing in their lives.  The problem with this is that there isn't much evidence of growing discontent with one's own life, except in the specific sense of economic dissatisfaction among people towards the lower end.  

 That's my summary of what's wrong with some popular accounts of the rise of political polarization.  I'll break for a few posts on other topics, but return to proposing my own explanation.  

*I put in in quotes because many accounts use "elites" very loosely, more or less as college-educated professionals, or even just people with a college degree. 

Thursday, September 17, 2020

The story of the decade, part 2

 As I said in my last post, another popular explanation of the rise of political polarization is:

"For many years, elites ignored the wishes of ordinary people, or showed contempt for them.  Finally the people rose up in anger, with the Tea Party movement and then with the Trump insurgency.  "

The idea of a revolt against an elite consensus is another case of the right explanation at the wrong time.  In the 1990s, most of the public favored a harder line against immigration, especially illegal immigration, while political elites of both parties were willing to accept high levels of illegal immigration and consider a chance for citizenship.  If you didn't like that, you needed to turn to an outsider--someone like Pat Buchanan or maybe Ross Perot.  But over the next couple of decades, public acceptance of immigration grew, while Republican elites moved in the opposite direction.  By the time Trump came around, there was no elite consensus on the issue--if you wanted a hard line against immigration, there were plenty of Republicans to choose from.  That's just one issue, although it was Trump's signature issue, but the partisan divergence among political elites, and consequent decline of elite consensus, was a general process.  

 Then there's the idea of elite condescension or contempt.  In my view, this is not just wrong, but the opposite of the truth--in fact, there's growing disapproval of anything that can be interpreted as "elitist."   I've looked for survey evidence bearing on this point and haven't found much.   However, I think that my position can be supported by historical evidence, which I'll discuss in my next post.  I'll offer one anecdote here.  The NY Times had a story about the first Kansas City Chiefs game in which it mentioned calls to do away with the "tomahawk chop" practiced by Chiefs fans.  They quoted some fans who wanted to keep it, including a 19-year-old woman who said "how is it OK for the past how many years, but now that it's 2020 it's apparently offensive."  The most-liked reader comment said:  "Note to Hannah and her all-white sorority sisters and fraternity brothers: It wasn't 'OK' for the past 50 years, but fortunately society is capable of evolving and recognizing some of its offenses and is trying to correct them."  The interesting thing is that the story didn't say that she was a sorority member, or even a college student.  In fact, it also quoted her mother and gave her age as 38, meaning that she had her daughter at 18 or 19.  So it's a pretty good bet that the young woman's mother didn't graduate from college, meaning it's less likely that she's in college herself.  I think that illustrates the current tendency to assume that "elites" or people with "privilege" are the source of any problem. 

 Finally, while looking for evidence about "elitism" I ran across this question:  "In your opinion, do all students have the ability to reach a high level of learning, or do only some have the ability to reach a high level of learning?"  This doesn't directly bear on the question--either "all" or "only some" could be interpreted as the elitist answer, depending on how you look at it--but it's interesting in its own right.  Opinions differed by age--older people were more likely to say "only some"--but education, sex, and race made little or no difference.  There was a difference by partisanship--60% of Republicans and only 48% of Democrats chose "all," with Independents right in the middle.  Maybe this is because the survey was done in 2001, when the No Child Left Behind plan was in the news.  That was based on the idea that we could bring all students up to proficiency by educational reform, so it's possible that the pattern was just because of Republicans supporting the plan.  It would be interesting to repeat the question to see whether the pattern still exists.


[Data from the Roper Center for Public Opinion Research]

Wednesday, September 9, 2020

The story of the decade

 I started this blog ten years ago today, so I wanted to do something to mark the anniversary.  I decided to write about one of the most important trends of the decade--growing political polarization.  Of course, that's too much for one post, so I will just start it here and continue it in future posts.  Many people have talked about this development, but I'd say that most of the explanations fall into three main groups.

1.   Many white people feel threatened because their traditionally dominant status is threatened.  Their determination to hold on at all costs is the source of polarization.  

2.  For many years, elites ignored the wishes of ordinary people, or showed contempt for them.  Finally the people rose up in anger, with the Tea Party movement and then with the Trump insurgency.  

3.  People have experienced a loss of meaning in their lives, and have turned to politics to provide what is missing--politics has become a substitute for religion or community.  

The first explanation is popular on the left, while the second and third are more popular on the right and center.  In my view, none of them are right.  I've had a number of posts that criticize them, but haven't really offered an alternative, although I have hinted at it.  (There's also some relevant material in my new book). I'll start this series of posts by summarizing my criticisms, and then suggest an alternative.  

The first one is essentially the right explanation, but for the wrong time. It explains a lot about the poltics of the 1970s and 1980s, but not the politics of the 21st century.  The reason is that the opinions of white Americans on race have become substantially more progressive.  Forty years ago, most whites said that discrimination was bad, but didn't think it was a major problem, and a significant number thought that whites were the ones who were really discriminated against.  Today, whites are more likely to see discrimination against blacks as a real problem, and also to see ethnic diversity as a positive good. 

 I'll add one small but important bit of information that suggests that most people are OK with diversity, and even welcome it.  Here are the approval ratings of presidents immediately after they took office and (about) 100 days in, as recorded by the Gallup Poll.  (I just consider newly elected presidents, not ones who took over after death or resignation.

                          Initial        100 days

Obama               67%           63%

GW Bush          57%           53%

Clinton              54%           45%

GHW Bush       51%           56%

Reagan             51%            68%

Carter               66%           63%

Nixon               59%           62%

Kennedy           72%          83%

Eisenhower       68%         74%

Obama had higher approval ratings than other recent presidents when he first took office.  That means that many people who had voted against him switched to taking a favorable view, which doesn't fit with the idea that they regarded him as a threat.  There is a version of the threat argument (I forget who has offered it) that says that there was a short era of good feeling when people thought we'd become a post-racial society and could forget about race, but whites turned against him once he started governing--that is, he was all right when he was just a symbol, but was seen as a threat when he took action.  That's why I also show approval ratings 100 days on, when presidents have nominate cabinet members and started trying to pass legislation.  Obama was still popular, with a higher approval rating than any president since Reagan.*

Obama's popularity declined to about 50% in the fall of 2009, and then stayed within a narrow range.  But in the first six months or so, being black didn't prevent him from being popular.  


*Reagan's approval jumped after an assassination attempt on March 30 (it was 60% in the last survey before then). 




Sunday, September 6, 2020


 Recently there has been a lot of discussion of the possibility of "hidden Trump voters"--people who intend to vote for Trump but aren't willing to admit that in a survey.  For example, a story in the NY Times said "while the effects of a hidden Trump vote are certainly overstated by the president’s allies, that does not mean that no evidence exists that polls are missing some of his voters. A small percentage of his support is probably being undercounted, and has been in the past, public opinion experts said. And in states like North Carolina, where the margin of victory could be narrow, the undercount could make a difference between a poll being right or wrong."  The idea is that they think support for Trump will meet with disapproval or hostility, so they are unwilling to admit it in public.  

You can make objections against this argument--for example, Nate Silver points out that there are "many places in America where you would face no social stigma whatsoever for showing outward support for Trump (often the opposite, in fact)."  But you could defend it by saying that it isn't just local climate that matters, but the general hostility to Trump in the mainstream media--if people don't know a person they assume that he/she is or might react negatively to an expression of support for Trump.

Looking back over previous American elections since surveys began, there's one case in which it seems that these kind of pressures would have been particularly strong--George Wallace supporters in 1968.  Mainstream opinion was overwhelmingly against Wallace, and this was when the media was dominated by a few "gatekeepers"--TV networks, newspapers, and news magazines.  He didn't have a major party behind him, and his supporters were a minority in most of the country.  Also, at that time the Gallup poll still conducted their surveys in person, and you might expect social pressure to be stronger in a face-to-face interview than over the phone. Nevertheless, the polls in 1968 did not underestimate support for Wallace--the last two polls before the election showed 15% support for Wallace among registered voters, and he actually got 13.5% of the vote.  

One of those polls had an experiment which suggests that they were concerned about the possibility of underestimating the Wallace vote.  A randomly selected half of the sample was not asked to say who they would vote for, but given a "secret ballot" which they marked and put in an envelope.  It wasn't truly secret, since it was connected to their survey, but that technique had been found to produce more accurate predictions of the vote in France, where standard methods underestimated the Communist vote.  Comparing the "secret ballot" to the standard question:


All respondents:

                    Nixon    Humphrey     Wallace        Undecided

Standard        41.1%        34.3%            15.9%            8.8%

Secret            42.9%        35.4%            18.0%            3.7%

 The biggest difference is that more people made a choice when they were asked to mark the ballot.  Adjusting for that, support for Wallace was a little higher with the ballot, although the difference is not statistically significant.  


Registered voters:


                    Nixon    Humphrey     Wallace        Undecided

Standard        43.1%        34.3%            15.6%            6.9%

Secret            44.2%        36.5%            15.7%            3.6%


The difference in estimated support for Wallace disappears.  That means that most of the difference in the first table involved people who weren't registered--30% of them chose Wallace with the ballot, and only 18% in the standard question (that difference is statistically significant).  

 I broke it down by region--Southern states (defined as states that were part of the Confederacy) versus others--but no differences in the effect of the ballot were visible for either registered or non-registered respondents.  

It seems reasonable that people who weren't registered voters would be less informed about politics and less confident in their choices, and therefore more influenced by what they think other people might think.  This gets at what I think is the major problem with the "hidden Trump voters" argument.  Survey interviewers are instructed to be neutral and just record your answers without expressing surprise or disapproval.  So for someone who's in an environment where Trump is unpopular, it's a rare opportunity to speak your mind without catching flak for it.   Similarly, for someone who thinks the media is biased against Trump and sees the polls as part of the media, it would be a chance to tell them what he really thinks.

[Data from the Roper Center for Public Opinion Research]

Tuesday, September 1, 2020

When the looting starts.....

 There is currently some discussion of the possibility that Donald Trump will gain support as a response to the protest-related violence in places like Kenosha and Portland.  People who think this will happen point to Richard Nixon's victory in 1968 as a parallel.  However, there's not much evidence that Trump is actually gaining ground.  There are several possible reasons for this, like  the general decline in the number of swing voters and the simple fact that there isn't as much violence today as their was then.  But one additional reason is suggested by answers to a Gallup poll question from May 1968 (the month after widespread riots following the assassination of Martin Luther King):


Whites thought it was the best way by a 57%-39% margin, with the rest undecided (only 11% of blacks thought it was the best way).  Less educated people were more likely to favor shooting on sight, but even among white college graduates, 45% thought it was the best way.  

No similar question has been asked this year, as far as I know, but that is revealing in itself--"a shoot on sight" policy is no longer a part of serious public discussion.  

The survey also had a number of other questions on race relations.  One was "how well do you think that negroes are being treated in this community":  72% of whites and only 29% of blacks chose "the same as whites are."   Another was "who do you think is more to blame for the present condition in which Negroes find themselves--white people or Negroes themselves.  Among whites, 24% said whites, 57% "Negroes themselves," and 19% are recorded as no opinion; among blacks, it was 45%, 13%, and 42%.  I think that "no opinion" is combining a number of different responses--later surveys that asked the question recorded a substantial number who volunteered "both," and a smaller number who volunteered "neither."  Then there was a question about whether "most businesses in your area discriminate against Negroes in their hiring practices":  among whites, 19% said yes and 67% said no; among blacks it was 64% and 32%.  Then there was a question about whether labor unions in your area discriminated, which showed a similar split.  

[Data from the Roper Center for Public Opinion Research]