Get back

 A couple of weeks ago, the New York Times published an op-ed by Adam Grant, Marissa Shandell, and Courtney Elliott called "The Secret Reason Bosses Want Everyone Back in the Office, Every Day of the Week."  A few days ago, the Times promoted it in an e-mail, saying "Here’s what six years of data says is actually driving the push to get employees back to the office."  The reason they propose is narcissism--that many CEOs are narcissists, and narcissists want people to display deference to them and find displays of deference more gratifying when they are done in person.  The op-ed is based on a paper by Shandell, Elliot and Grant published in Organizational Behavior and Human Decision Processes with the title "Worship Me at the Office Altar."  They collected statements about remote work and several qualities that they regarded as measures of narcissism for a sample of about 250 CEOs of Fortune 500 companies and found that CEOs who ranked higher in narcissism were more negative in their comments about remote work.

Their argument for a connection seems unconvincing to me.  Fortune 500 companies are large organizations, so the CEO never has face-to-face contact with the average employee.  If the general claim about narcissists is correct, it would be relevant to the top management team--CEOs who rank higher in narcissism would want them to return to in-person work more quickly--but wouldn't have any implications for ordinary workers.  However, they do appear to have strong evidence of a connection (t-ratio of about 4), so how do I explain that?  

One factor is that the usual t-ratios assume that the errors are normally distributed.  In fact, the residuals from the regression aren't normally distributed:  

With an error distribution like this, the standard errors will tend to be underestimated so the t-ratios will be overestimated.  There is no perfect way to address this, but I tried a few reasonable alternatives and they produced t-ratios of about 2.5.  That's less impressive, but still reasonably strong evidence by conventional standards.  

The more important factor involves their measure of narcissism.   They propose four indicators:  size of signature on the annual reports, size of the CEO photo, cash compensation relative to the second-highest paid executive, and non-cash compensation relative to the second-highest paid executive.  If you have multiple measures of the same underlying concept, all of the measures should be positively correlated with each other.  In fact, the correlations among these variables range from -.113 to .160, with an average of .044.*  There are two possible interpretations:  either there is no underlying quality to measure, or that at least three of the four "indicators" are not good measures of that quality.  I'll go with the poor measures interpretation:  in particular, even CEOs don't have complete control over their own compensation or the compensation of other managers.   

 If you include the individual variables as predictors in the regression, relative cash compensation consistently is statistically significant, relative non-cash compensation sometimes is, photo size sometimes scrapes in as "significant at the 10% level," and signature size is never close.  So there's pretty good evidence that CEOs who are highly compensated (relative to the next highest paid employee) expressed more opposition to remote work.  How would I explain this?  My proposal is that managements generally wanted to get employees back to in-person work--maybe not all the way back, but faster and farther than the typical employee would want--and that more prominent CEOs took the lead in delivering the bad news.  "Prominence" could be understood as status within the profession--high compensation is a sign that someone is regarded as an outstanding leader.  Or it could be within the organization:  high compensation could indicate a hierarchical organization in which nothing was final until you heard it from the big guy.  In less hierarchical organizations, the CEO could stand back and let subordinates deliver the bad news.   

Although I disagree with their interpretation, Shandell, Elliott, and Grant deserve credit for collecting and compiling the data (the link is on p. 5 of their paper, which I think is open access).  Unfortunately, they don't include the names of the companies, which would let other researchers add to the data set.  Confidentiality isn't an issue, since all of the data are from public records.  

*If you're interested in Cronbach's alpha, it's about .125.  

Done them proud?

 In 1999, Gallup asked "Overall, do you think the signers of the Declaration of Independence would be pleased or disappointed by the way the United States has turned out?"  The question was repeated in June 2001, 2003, June 2008, 2013, and June 2026.  In 2007 and 2010, CBS asked a similar question:  whether the "founding fathers" would be pleased or disappointed.  Finally, in 2010 (January and June), 2011, 2016 and 2017 Fox News asked if the founding fathers would be "proud of the country if they could see it today."  Here is the percent positive (pleased or proud) answers:

The breakdown by party identification (June 2010 and June 2016 are unavailable):*

Partisan differences were small and Democratic and Republican opinions moved in the same direction for the first decade or so, but then diverged in 2010.  The next figure shows the gap between supporters of the President's party and supporters of the opposition:

    For some evaluations, like economic conditions, partisan differences have increased steadily in the 21st century:  growing from Bush to Obama, Obama to Trump, Trump to Biden, Biden to Trump II.  This is different:  the partisan gap was bigger under Obama than in either Trump term (unfortunately the question wasn't asked in the Biden years).  Why is this question different?  My guess is that people focus on the political situation when answering, and evaluation of the political situation involves views of the opposition as well as the party in power.  That is, someone might think that the president was doing a good job but that the founders would be distressed to see the amount or kind of opposition that he was encountering.  Of course, Democrats were upset about Republican opposition during the Obama years, but I think they were still somewhat hopeful about winning a clear majority or achieving some bipartisan successes.  As a result, they were more optimistic in assessing the general political situation than Republicans are now.  

*I don't show independents in order to keep things simpler.  As you'd expect, they are generally in between supporters of the president's party and the opposition, and I don't see any clear changes in their relative position.  

[Data from the Roper Center for Public Opinion Research]

That's great

 In May, the Guardian published a list of the 100 best novels "as voted for by authors, critics and academics worldwide."  They just gave ranks, but I was interested in total scores--how much difference is there between first and 10th place, or 10th and 100th?  I also wondered if there were novelists whose votes were scattered among a number of different works, so they would do better in a ranking by authors.  This article by Jarrett Kobek  has a link to a spreadsheet with the choices of individual voters.  I calculated scores as 10 points for a first choice, 9 for second, .... 1 for 10th.*  

There were 172 voters, who selected 670 different novels.  The distribution was highly skewed--the leader was Middlemarch, with 446 points, and at the other end there were 67 novels that got a single point (ie, one voter placed them tenth).  To make the top 100 required only 18 points--twelve novels had exactly that total.  Turning to authors, 474 received votes.  The top 25:

                                        N    points Guardian

George Eliot                 60    469 83.45
Virginia Woolf                64    407 84.35
Jane Austen                    63 390 82.50
Leo Tolstoy                    47     379 65.95
Toni Morrison                 57    361 75.05

James Joyce                 39 295 53
Charles Dickens         44 269 57.45
Marcel Proust                 27 201 37.05
Gustave Flaubert         22 159 29.95
Vladimir Nabokov         27 157 34.85

Henry James                 29 151 36.55
Gabriel García Márquez 23 126 29.30
Charlotte Brontë         21 121 27.05
F Scott Fitzgerald         21 120 27.00
Herman Melville         18 116 23.80

Emily Brontë                 15 114 20.70
Fyodor Dostoevsky         17 110 22.50
Laurence Sterne         18 109 23.45
Franz Kafka                 18 107 23.35
George Orwell          20 104 25.20

Kazuo Ishiguro         19 97 23.85
Chinua Achebe         15 94 19.70
Salman Rushdie         15 86 19.30
Miguel de Cervantes 11 84 15.20
Muriel Spark                 16 83 20.15

On the other side, here are some well-known novelists who didn't get much support:

                                points
Balzac                        7
Stephen Crane            9
Richard Wright          10
Zola                            13

In 1998,  the Modern Library offered a list of the top 100 novels published in English in the 20th century.  Novels from that list that got no support in the Guardian survey include Brave New World (#5), Darkness at Noon (8),  An American Tragedy (16), Native Son (20--Wright's points in the Guardian poll were for Black Boy), Henderson the Rain King (21), Sister Carrie (33), All the King's Men (36), The Heart of the Matter (40), and Lord of the Flies (41).  

One pattern that seems clear is that realist novels of the 19th and early 20th century don't get much support from the Guardian voters.  Why not?  The basic realist plot is that an outsider tries to get somewhere but is defeated, partly by social injustice but partly by his/her flaws and bad decisions.  I think that accounts for their poor showing in the Guardian survey:  educated middle class people are increasingly uncomfortable with anything that seems to express a critical view of disadvantaged people.  Another pattern is that novels by Americans and Europeans which are set in the rest of the world don't do well in the Guardian survey.  In addition to the examples of Henderson the Rain King and The Heart of the Matter, Malcolm Lowry's Under the Volcano ranked 11th in the Modern Library list and got only 9 points from the Guardian, which puts it about 250th.  There's one big exception:  Heart of Darkness gets 42 points, for a rank of 40.  I think there's a reason for that:  it's about the damage done by colonialism.  Conrad gets only 9 points for all of his other novels (Modern Library had The Secret Agent ranked 46th, Nostromo 47th, and Heart of Darkness at 67th).  Finally, two novels with political themes do well in the Guardian ranking:  1984 at 16th and The Handmaid's Tale at 36th. But that's about it--several of the Modern Library picks that got no support in the Guardian survey seem at least equally relevant to contemporary politics.  Overall, the Guardian list seems to reflect the view that writers should "stay in their lanes."  

*The Guardian used a different formula:  1 point for being chosen and then a bonus of 0.5 for first, 0.45 for second, .... 0.05 for tenth.  That doesn't make sense to me--it makes the difference between 11th and 10th more important than the difference between 10th and first.   

Highly respected, part 2

 In a recent post, I found that people who believed that they were treated with "less courtesy or respect than other people" were more likely to vote for Trump in 2016 (and probably 2020).  This post will look at it from another direction:  who is more likely to think that they are treated with less respect?*  I considered race, sex, age, and education.  My expectations were that black people, younger people, men, and less educated people would be more likely to perceive disrespect.  Age turned out to be the most important factor, followed by race, with no clear differences by education and sex.  But although the overall average is about the same for men and women, there is an interaction with age:

Young men perceive more disrespect than young women, but middle-aged women perceive more disrespect than middle-aged men.  Or looking at it in terms of the relation with age, perceived disrespect declines steadily with age for men, but stays about the same or even increases until about age 40 or 50 for women and then declines.  There's also an interaction between education and sex:  education makes a difference among men (less educated men perceive more disrespect), but not as much (maybe not at all) among women.  There are also interactions involving race:  black men perceive more disrespect than black women.  In fact, the means for white men, white women, and black women are very close, and are consistent with the hypothesis that race doesn't make any difference among women.  Finally, there's an interaction between race and education:  among whites, people with college degrees perceive less disrespect; among blacks, there's no clear difference.  I looked at possible interactions between race and age, and education and age, but found no evidence of them.  I also looked at the possibility of a three-way interaction between race, sex, and education, but the numbers are too small to say anything definite.  

A belief that you are treated with less respect than other people could reflect experience--you really are treated with less respect--or greater sensitivity--you are more likely to interpret the same treatment as disrespect.  The group differences probably reflect a mix of both, but there's no way to distinguish them in these data.

*In my previous post, I asked "whether the association involves support for conservatism in general, or a specific kind of populist conservatism... I'll try to shed light on that by looking at the association with various political views."  I did that analysis, and found that among people with college degrees, feeling that you are treated with less respect is generally associated with more conservative views, but it's not possible to be more specific, partly because the number of cases isn't that large and partly because the GSS questions on politics aren't very well suited to identifying populist conservatism.

A distinction without (much) difference

Last year I had a post about the claim that men don't have as many friends as they used to.   I noted that the 2021 survey cited as evidence of a large decline in friendship among men showed a similarly large decline among women.   That survey used different procedures from previous ones, so it's not clear that we should accept it as evidence of a real decline in friendship, but if we do, it shows a decline among people generally rather than men specifically.  

A few days ago, I read an interview of Laurie Santos, a professor of Psychology at Yale, by Derek Thompson.  Santos said there was a decline in friendship, and "that decrease is much worse for men. One study found that if you look at what’s standardly considered a good level of friendship -- do you have six close friends you could talk to? -- men have shown a decrease in that number by about half in the last couple of decades. And if you ask how many men say they have no close friendships at all, you see around 15% of American guys in midlife saying exactly that. That’s a fivefold decrease since folks have been running this survey."  Then there was this figure, summarizing the same 2021 survey I wrote about:

The share of men who reported six or more close friends indeed declined by more than half, from 55% to 27%.  The share of women who reported six or more friends declined from 41% to 24%.  That's a decline of 51% among men and only 41% among women, but it's just one possible comparison:  the share who report 10 or more declined by 62% among men and 61% among women, and the share who report five or more declined by 40% among men and 38% among women.  And the share of women reporting no close friends rose from 2% to 10%:  the same ratio of 5 as among men.  That is, the table doesn't support the general claim that any decline has been much worse among men.*  

That raises the question of why this misinterpretation is so prevalent (Santos followed this report, which is the source of the table).  The obvious factor is that people just like to talk about differences between men and women.  A second possibility is the influence of political ideology:  the idea of an especially large decline among men appeals to both liberals and conservatives, for different reasons.  Liberals are inclined to think that the traditional model of masculinity is unsuited to modern society, and we're seeing a symptom of that; conservatives are inclined to think that we've paid so much attention to the problems of women that we're neglecting the problems of men.  And moderates think that if both liberals and conservatives agree that there's a problem, we have a welcome opportunity for cooperation across ideological lines.  

*Even if you think that the comparison between six or more vs. five or fewer is especially important, you need to consider sampling error.  The 2021 survey had about 2000 people, and the 1990 survey had about 1200, so the figures for men and women are based on samples of 1000 and 600.  Then you have ratios, and differences between ratios, increasing the margin of error.  I estimate that the standard error of the difference between the ratios is about 5.5% assuming simple random sampling, but even without doing calculations it's clearly substantial.  

Highly Respected?

 It's often said that a large part of Donald Trump's support represents a reaction to the disdain that ordinary people receive from the educated middle classes.   For example, in a recent opinion piece in the New York Times, Dylan Gottlieb writes "yuppies and their arrogance bred new resentments. In the 2010s, a brand of populist conservatism opposed nearly every tenet of the yuppie dream, from racial and gender diversity to educational meritocracy to frictionless finance and globalization to gourmet culture and the very idea of urban living itself."  I don't think this works as a historical account:  rather than growing arrogance in the educated middle classes, there's been a growth of social egalitarianism.  But there's still a question about the individual-level relationship between resentment and politics:  are people who feel like they are looked down upon more likely to support "populist conservatism"?  There aren't many survey questions on the subject, but since 2018 the General Social Survey has had one on how often you are "you are treated with less courtesy and respect than other people" (six categories, from "almost every day" to "never").  A few years ago I looked at it and found "no clear connection to choices in the 2016 election or to opinions on a variety of political issues," but more data has accumulated since then, so I thought it was time for a new look.

The average (never=1.....almost every day=6) is 2.98 among people who say they voted for Trump in 2016 and 2.92 among people who say they voted for Clinton, for a difference of .06, but the standard error of the difference is about .08:  that is, no evidence that it's anything more than sampling error.  But feelings that you are treated with less respect differ by race and gender (higher among black people, especially black men), education (higher among less educated people), and age (higher among young people), and these factors also make a difference to voting choices.  In a logistic regression of Trump vs. Clinton voting with controls for age in years, college degree, black race, sex, and an interaction of race and sex, the estimate for feeling you are treated with less respect is .092 with a standard error of .023.  That is, people who think they are treated with less respect are more likely to vote for Trump.  It's not a huge difference, but it's large enough to be of interest (e. g., an increase of 4 points on the scale is about equivalent to the gender gap among whites).

In my previous post, I suggested that feelings of resentment might be particularly strong among more educated conservatives.  That is, conservatives are a minority at most colleges and and in many professional jobs, so they probably will encounter occasions when they are treated with less respect than their colleagues.  If you add an interaction between college degree and feelings of being treated with less respect:

                                           B       SE          T        P

disrspct	.057	.027		2.094	.037
black	-2.407	.202	.	-11.926	.000
female	-.402	.064		-6.301	.000
age	.009	.002		4.599	.000
black*female	-.761	.330		-2.306	.021
college	-.413	.206	.	-2.010	.045
college * disrspct	.106	.047		2.250	.025

That is, the estimated effect of perceived disrespect among people who didn't graduate from college is .057; among those who did, it's .163.  This is just the 2016 election:  the 2020 estimates are in the same direction and of similar size, but not statistically significant.  

Overall, it appears that feeling that you aren't treated with respect is associated with support for Trump, but that the association is stronger among more educated people:  that is, it's more relevant to middle-class support than to working-class support.  This raises the question of whether the association involves support for conservatism in general, or a specific kind of populist conservatism.  In a future post, I'll try to shed light on that by looking at the association with various political views.  

Why not?

 In January, I had a post about why Trump was renominated in 2024.  I argued that it wasn't because he of strong support among ordinary Republicans, but because of support from Republican elites.  Andrew Gelman recently discussed my post in his blog, which led me to some further thoughts.  My post concluded by saying "'the base' didn't impose Trump on Republican elites; Republican elites asked for him."  In retrospect, saying that they "asked for him" was going too far--it would be more accurate to say that they didn't put up much opposition even though Trump had shown that he was a drag on the party.  He underperformed in both 2016 and 2020, and he hurt Republican chances in Congress by promoting weak candidates like Herschel Walker and Dr. Oz.  So even if they liked some of Trump's policies, they had a good reason to turn to a new candidate in 2024.  

Why didn't Trump encounter significant resistance from Republican elites?  One possibility is that they they realized had so much support from ordinary Republican voters that resistance would be futile (and would expose them to retaliation).  In my earlier post, I looked at questions about whether people would like Trump to run for the 2024 nomination and compared them to earlier questions on whether they wanted Gore, Kerry, and Romney to run for re-nomination after their defeats.  Those questions are usually asked in the first couple of years after the election; once it gets closer to the next election, you get questions about who people would choose out of the declared or potential candidates.  The figure shows the percentages of Republicans favoring Trump and DeSantis in surveys between July 2022 and August 2023.

Trump was ahead in the first few surveys, but his support declined and DeSantis's rose until sometime in early 2023.  After that, Trump steadily gained, but he didn't consistently break 50% until late May 2023.  In the last few surveys, the average was 54% for Trump, 17% for DeSantis, and 29% for all others.

So there was a period when Trump seemed vulnerable and other candidates were starting their campaigns (Trump declared in November 2022, Haley in February 2023, and others in March-June 2023).   During this time, endorsements might have been helpful in getting opposition campaigns off the ground.  But only a few prominent Republicans offered them:  among Republican Senators and Governors, by the middle of May ten had endorsed Trump and no one had endorsed any of his opponents.  There were six for various opponents by the end of June, then one for Trump in July and one more in August.  After that Trump pulled ahead steadily.  That is, very few prominent Republicans endorsed other candidates when their endorsements might have made a difference.  To consider a few examples, Susan Collins and Lisa Murkowski both endorsed Nikki Haley, but not until early March 2024, when it was too late to matter.  A few days ago, the New York Times had a story on a new memoir by Lamar Alexander (who left the Senate in January 2021) in which he "says President Trump committed an impeachable offense on Jan. 6."   But Ballotpedia doesn't record an endorsement from Alexander in the 2024 campaign.  Mitch Daniels didn't offer one either.  

 Why did prominent Republicans who didn't support Trump and had no particular reason to fear him stay on the sidelines?  As I've mentioned before, the rise of partisanship means that the perception of party unity has a larger impact on the fortunes of the party--when a party is seen as divided, it does worse.   If a large number of prominent Republicans had come out against Trump, that might have stopped him from getting the nomination, but it would also have hurt Republican chances in 2024 (especially since Trump would not have been a gracious loser).  As an alternative, they hoped that he would fade away on his own, and by the time it became clear that wouldn't happen, it was too late,

[Data from the Roper Center for Public Opinion Research]