Saturday, February 16, 2019

The old reliable

This morning I was thinking that it was about time for a post, when I read an opinion piece by Christine Emba that mentioned an "epidemic of loneliness."  She linked to a column by George Will, which quoted liberally from a book by Senator Ben Sasse.  That book, and more generally the claim that loneliness is on the rise, has received a good deal of attention.  As I have pointed out before, there is no evidence that loneliness is rising, and some that it is falling.  (Claude Fischer has made the same point several times, most recently here). However, there are few repeated survey questions about feelings of loneliness.  If people are having less contact with family, friends, and neighbors, it might be reasonable to suspect that loneliness is on the rise despite the lack of direct evidence. 

Since the 1970s, the General Social Survey has had questions about "how often you spend a social evening with" various kinds of people:  "someone who lives in your neighborhood," "friends who live outside the neighborhood," and "relatives."  The questions are on a seven-point scale going from never to almost every day, and I recoded them so that higher means more often.  Because people often say that the negative trend is mostly among the "working class" I calculated the average separately for people with and without college degrees (the blue line is without and the red line is with a degree).  The means for spending an evening with neighbors:

A downward trend for both groups--maybe there is something to the idea of an epidemic of loneliness?

Friends outside the neighborhood:  maybe a slight downward trend for people with college degrees, but a more definite upward trend for people without college degrees.  Since there are more people without degrees than with degrees, the trend for the population as a whole is slightly upward.  

Relatives:  a pretty clear upward trend for both.  

So in addition to the lack of direct evidence of an epidemic of loneliness, there is evidence against the idea of a general decline in social contact.  

Loneliness is a real social problem, and claims of an epidemic seem to be more effective in getting people to pay attention than just saying that something is a problem and it hasn't been getting enough attention.  So in that sense, claims that loneliness is rising don't do any harm and may do some good.  However, they are often offered as an explanation of rising political polarization (as in Sasse's book).  In this respect, they are harmful because they divert attention away from the real causes of polarization, which I think involve political and media elites, especially on the conservative side.   

Friday, February 8, 2019

Red town, blue town

Last week I read a column by Timothy Egan about books by J. D. Vance and Tara Westover, which he said were "guides to a Trumpland that is terra incognita to most Americans."  Vance grew up in southern Ohio and Westover grew up in Idaho.  A couple of days later, I read a column by Jill Abramson saying that reporters should "report the story from the places where the pro-Trump and Trump-curious live." She didn't say precisely where those were, but referred to the "heartland" and "rural America."  These reminded me of an observation I have made before, which is that you don't have to travel all that far to find places where most people voted for Trump, even if you live in Manhattan or Washington, DC.  (There are parts of the country where you have to travel a long way to find places where most people voted for Clinton--see this map). 

For example, Connecticut is divided into 169 towns:  Trump got more votes than Clinton in 81 of those.   He won some of them by big margins, getting as much as 68% of the vote (see the map of election results here).  What sort of places were they?  They tended to be more rural:  here is a scatterplot of Republican share of the two-party vote by population density.

What about income?  At the state level, there's a substantial negative correlation between average household income and Republican support; among Connecticut towns it was negative but small (-.11).  The town data I downloaded also had median household income, and the correlation between with Republican support was only .04.  But in an idle moment, I divided mean by median income and found  a pretty strong relationship:

Mean divided by median is a measure of skewness, but given the typical distribution of income, it could also be regarded as a measure of inequality--that is, Republican support was lower in places with a bigger gap between the top and the middle.  I'm not sure what to make of this relationship, but it seems potentially interesting. 

In most accounts, urban/rural differences in Republican support are treated as a matter of economics:  rural areas are falling behind, and therefore people feel angry.  But Connecticut towns are small, averaging only 25 square miles.  Also, even the biggest metropolitan areas in Connecticut aren't all that big.  So regardless of where you work, you are in reasonable commuting distance of urban, suburban, or rural places.  This point suggests that the urban/rural differences are mostly about preferences--the sort of people who decide to live in low-density areas are the sort of people who vote Republican.  This possibility has sometimes been noted, but it hasn't received much attention.  I think that's because of the tendency represented by Egan and Abramson--exaggerating the geographical distance between "red" and "blue" America.  (I think they also exaggerate the social distance, but that's a topic for another post).   

Thursday, January 31, 2019

A leader without followers?

A week ago, I didn't know who Howard Schultz was.  Now I do, and have read a substantial number of opinion pieces saying that he shouldn't run for president because he has no chance of winning.  Some of the reasons are good, like the barrier the Electoral College creates for third-party candidates and the scarcity of real independents, but I want to talk about a bad one:  the claim that few people want the social liberal/economic conservative combination that he seems to be offering.   Several cited a report by Lee Drutman which estimated that 3.8% of voters are in this group, as compared to 44.6% liberal on both, 22.7% conservative on both, and 28.9% liberal on economics, conservative on social issues.  That is a total of 73.7% liberal on economic issues, and 48.4% liberal on social issues.

Saying that about half of people are liberal on social issues seems reasonable, but over 70% liberal on economic issues?  Drutman's measure of economic liberalism includes questions on how important Social Security and Medicare are to the respondent, whether the economic system is biased in favor of the wealthiest Americans, and whether the distribution of money and wealth in this country is fair.  These are all questions on which large majorities are on what he counts as the "liberal" side.  However, they are not really liberal/conservative questions:  that is, they don't involve a choice between different policies.   The person who says that the distribution of income is unfair doesn't necessarily support liberal policies for doing something about it--he or she might think that nothing can be done, or that those policies wouldn't do any good, or just be expressing general cynicism.

David Leonhart says "large majorities of Americans oppose cuts to Medicare and Social Security and favor expanded Medicaid. They favor higher taxes on the wealthy and corporations. They favor a higher minimum wage and more aggressive government action to create jobs. "  That's basically right, although the majority in favor of higher taxes on the wealthy and corporations is not very large (e. g., in 2012, 50% said the Bush tax cut for people with incomes over $250,000 should come to an end, 43% that it should not).   However, there are also economic issues on which people are conservative:   in 2016 43% said we were spending too much on welfare, and only 22% said we were spending too little; there is generally strong support for things like drug tests and work requirements for public assistance.

In general, people are more favorable to programs that help "deserving" people (like retirees or people with jobs), that give people things they need (like medical care) rather than just giving money, and that require employers rather than taxpayers to provide benefits.  Most people also like tax cuts, even if they know they are skewed towards the rich, and are wary of tax increases, even tax increases on the rich.  That doesn't mean people favor spending cuts--they seem to think that the government can find the money somewhere or other.   Overall, majority opinion doesn't fall neatly into a left-right spectrum on economics.   It's a distinctive combination that isn't really represented among political elites,  mostly because both business interests and professional economists (of all ideological types)  are opposed to it.  Both liberals and conservatives can appeal to some elements of it, but not others.

Returning to Schultz, I think there is a substantial constituency for the combination he's offering--people who are socially liberal, and more interested in cutting taxes than in expanding programs.

PS:  I think liberals do have some advantage on economic issues, not because of specific policy issues, but because of the image of caring about the middle and working classes.  Conservatives have an advantage on social issues because of the image of being committed to patriotism, religion, and family.   One of the barriers to an independent candidate is that they can't draw on either of these.

Friday, January 25, 2019


In 2017, I had a post about opinions of Martin Luther King.  In surveys taken in 1964 and 1965, he had more positive ratings than negative ones, but not by much; in 1965 and 1966, he had more negative ratings than positive ones.  In the last survey taken during his lifetime (August 1966), only 33% gave him a positive rating, and 39% gave him the most negative rating possible (minus five on a scale of -5 to +5).

None of the surveys asked people to give reasons for their judgment, but some other ones shed light on the issue.  In May 1961, a Gallup poll asked "Do you think 'sit-ins' at lunch counters, 'freedom buses,' and other demonstrations by Negroes will hurt or help the Negro's chances of being integrated in the South?"  57% said hurt, and only 28% said help.  In June 1963 (a couple of months before the March on Washington), one asked "DO YOU THINK MASS DEMONSTRATIONS BY NEGROES ARE MORE LIKELY TO HELP OR MORE LIKELY TO HURT THE NEGRO'S CAUSE FOR RACIAL EQUALITY?," with similar results:  27% help and 60% hurt.  The "mass demonstrations" question was asked again in 1964 and it was down to 16% help and 74% hurt.  So it seems like most people thought that even nonviolent demonstrations did more harm then good.    A 1965 Harris survey asked "In the recent showdown in Selma, Alabama over Negro voting rights, have you tended to side more with the civil rights groups or with the State of Alabama?"  48% said civil rights groups, 21% said State of Alabama, 19% neither, and the rest not sure.  So there seem to have been a significant number of people who favored the civil rights groups when forced to make a choice, but didn't want to be forced to make a choice. 

There were two survey questions in the second half of 1973 that addressed these issues:  "Do you feel the protest marches led by Martin Luther King in the 1960's speeded up civil rights legislation, slowed it down, or didn't make much difference one way or the other?"  67% said speeded up, only 3% slowed down, and 23% said not much difference.  A different survey asked about whether different kinds of people did more good than harm or more harm than good--one was "Blacks who demonstrate for civil rights."  40% said more harm than good, and 60% said more good than harm, neither harmful nor helpful, or weren't sure (unfortunately, the report doesn't distinguish among those groups).  In retrospect, a lot of people seem to have decided that demonstrations had been effective, and even to have accept them as justified in the present.

My previous post was inspired by an article by Ta-Nehisi Coates which pointed to the surveys about Martin Luther King but concluded  the protests "affected the attitudes of the children of those white Americans who scorned them. . . . The point is the future."  The data from the 1970s suggest that he was right, but the future came faster than he thought.  It didn't take a generation--attitudes changed substantially in less than a decade.

[Data from the Roper Center for Public Opinion Research]

Saturday, January 19, 2019

Inequality and public opinion

In my last post, I questioned Thomas Edsall's claim that "study after study shows that as the gulf between rich and poor widens, voters become increasingly mean spirited and hostile to the welfare state, progressive taxation and regulations designed to protect consumers, workers and the environment."   The gulf between the rich and poor has widened substantially since the mid-1970s in the United States, but for the questions I looked at, opinions have either become more liberal or stayed about the same. 

There are strong reasons for expecting the demand for redistribution to increase as inequality increases.  One is self-interest--people will gain more from redistribution when the rich have more to redistribute.  The other is a sense of fairness--the bigger the gap between the actual distribution and what people think is a fair distribution, the more demand to do something about it.  The studies that Edsall cited point to effects going in the other direction--ways in which growing inequality might reduce support for redistribution.  For the most general one (and the only one to mention taking from the rich), there is no trend, suggesting that the effects in different directions pretty much balance out.  But it doesn't seem that people have become more mean-spirited towards the disadvantaged:  support for the welfare state has increased, not declined.   

Friday, January 11, 2019

Study shows?

In a column the other day, Thomas Edsall says "study after study shows that as the gulf between rich and poor widens, voters become increasingly mean spirited and hostile to the welfare state, progressive taxation and regulations designed to protect consumers, workers and the environment."  I glanced at the studies he cited, and none of them seemed to show that, so I turned to the General Social Survey.  It has a number of questions about whether spending on various national problems should be increased, reduced, or kept at about the current level.  The ones that seem relevant are "improving and protecting the environment," "solving the problems of the big cities," "Social Security," "improving and protecting the nation's health," "improving the conditions of blacks," "welfare," and "assistance to the poor."  There are also two more general questions.  One asks people to place themselves on a scale between "the government in Washington ought to reduce the income differences between the rich and the poor, perhaps by raising the taxes of wealthy families or by giving income assistance to the poor" and "the government should not concern itself with reducing this income difference between the rich and the poor."  The other is a scale between "the government in Washington is trying to do too many things that should be left to individuals and private business" and "the government should do even more to solve our country's problems."  The gap between rich and poor started growing at about the same time the GSS started (the mid-1970s), so Edsall's hypothesis implies that opinions should move steadily in a conservative direction.  I coded all of the variables so that high values mean more conservative.  Here are the average opinions by year.  First, welfare and aid to the poor (which were asked of randomly selected halves of the sample).  No obvious trend for either one.

Next, the conditions of blacks and the problems of the big cities.  On spending to improve the conditions of blacks, it looks like a trend towards support for more spending, with an unusually large move from 2014 to 2016.  On "the big cities" there are ups and downs with no obvious trend.

People generally support spending more on the environment and social security (in the scale on the left, 3 would mean everyone favors spending less, 1 would mean that everyone favors spending more, and 2 would mean equal numbers favor spending more and spending less).  It looks like there might be some trend towards favoring more spending on social security.

Opinions about spending on health show a distinctive pattern:  a trend towards support for spending more through 2008, and then a sharp move in favor of spending less.  The reason for that seems pretty obvious:  "Obamacare" was passed in 2010. 

Finally, the two general opinions:
It looks like there was a more in a conservative direction in the first few years, from the early 1970s to the early 1980s, and no obvious trend since then.

Opinions on these questions generally are more conservative when a Democrat is president, and more liberal when a Republican is, which makes sense since all of them refer to the way things are as a baseline.  If you regress each on a dummy variable for the party of the president and a time trend, the estimate on the time trend is negative and statistically significant for spending on the environment, social security, conditions of blacks, welfare, and aid to the poor, and not significantly different from zero for the rest.  So in general, it seems that people have become more generous and accepting of the welfare state during the period of rising inequality.  

This raises a number of questions which I don't have time to address now, but will get to soon.


Saturday, January 5, 2019

The twelfth day of Christmas

A Washington Post story had a quote from Lindsey Graham that sums up the last decade of American politics:

"It’s resonating with our base for sure.  .  .  . This is not really about immigration now. . . . When they’re told by the people they hate that this makes no sense, it makes them more determined to get the wall.”

The idea of a wall along the US-Mexican border has been around for a while--there were survey questions as far back as the early 1990s.  A 2006 NBC/Wall Street Journal survey asked "Would you favor or oppose building a wall or a fence all along the border between Mexico and the United States, from Texas to California?"  Opinions were split evenly:  32% strongly in favor, 15% somewhat in favor, 16% somewhat against,  33% strongly against.  Conservatives and Republicans were more likely to favor one.  There was little or no difference between blacks and whites--support was lower among Latinos, but a substantial number (35%) were in favor.

                                       SF            F            A        SA
Democrats                     25%        13%       18%     43%
Independents                 37%        11%        15%    38%
Republicans                  42%         21%       16%    21%

Liberal                            24%        11%      15%    48%
Moderate                        29%        18%      17%    32%
Conservative                  40%        14%      13%    27%

White                             33%        15%      16%     32%
Black                             39%         12%     18%     28%
Hispanic                        21%         14%     13%     46%

In January 2018, an ABC News/Washington Post poll asked "Do you support or oppose building a wall along the U.S. border with Mexico?"  Support was somewhat lower than in 2006:  27% strongly in favor, 10% somewhat, 10% somewhat against, and 49% strongly against. 

                                      SF            F            A        SA
Democrats                        7%        5%         11%    75%
Independents                 20%        12%        15%    45%
Republicans                   54%        18%          9%    17%

Liberal                              9%          3%       8%    79%
Moderate                        18%        12%      12%    55%
Conservative                  53%        14%      10%    23%

White                             33%        12%      10%     41%
Black                             11%           3%     10%     75%
Hispanic                          9%           7%     16%     66%

Support has grown among Republicans and conservatives, but fallen among Democrats and liberals, and also among independents and moderates.  In fact, moderates in 2018 were less favorable than liberals were in 2016.  The most interesting thing to me is that support has collapsed among blacks and Latinos.  Part of that is party identification, but the changes are so big I don't think that's the whole story.

Graham was thinking of one side, but neglecting the other--if Trump and his base insist on the wall, then people who are not part of his base oppose it more strongly. Since most people aren't part of his base, that means that support for a wall has fallen. 

[Data from the Roper Center for Public Opinion Research]