Sunday, October 21, 2018

Random observation

One of the questions in the General Social Survey is:  "If you were to get enough money to live as comfortably as you would like for the rest of your life, would you continue to work or would you stop working?"  While looking for something else in the GSS the other day, I noticed this question and wondered if there was any trend.   One thing led to another, and here are trends by educational level (college graduate or not):

The likely reason that more educated people would be less likely to stop working is that they have more satisfying jobs.   Job satisfaction has stayed about the same in both groups over the whole period, so that doesn't account for the difference in trends.  On the other side, more educated people presumably have more interests outside of work--maybe that gap has grown, although I'm not sure why it would.  Or maybe it reflects changes in the sense of moral obligation to work?

Wednesday, October 17, 2018

The fault is not in other democracies, but in ourselves

I had a post almost two years ago about the idea that support for democracy was declining.  That started as an example for a class I was teaching, and I'm teaching the same class now, so I was going to reuse it.  But on looking back it didn't seem very clear, so hear is a new version. 

I looked at seven well-established democracies.  The basic question is:  "Various types of political systems are described below. Please think about each choice in terms of governing this country and indicate if you think that it would be a very good, fairly good, fairly bad or very bad way of governing [your nation]:"

Here are the average ratings for "having a democratic political system"

There is no general pattern:  ratings increase in Spain and Australia, but decline in Japan and the United States.  At the beginning, Americans are third out of seven nations in their rating of democracy; at the end, we are eighth out of eight. 

Here are the average ratings for "having the army rule": 

The United States stands out here:  there has been a pretty steady increase.  In 1996, we were part of a group of three nations in the middle; in 2012, we had the most positive rating.  Having the army rule still gets a much lower rating than a democratic political system, but the gap has clearly narrowed.  This is unique to the United States--there is no clear trend in any of the other countries. 

[Data from the World Values Survey]

Friday, October 12, 2018

A new identity?

In my last post, I talked about perceptions of discrimination against blacks and whites.  I combined the two questions to get the numbers who said that there was more discrimination against whites, more against blacks, and equal amounts.   If you go back to the individual questions, which were asked in 2015:

Against blacks:

                    A lot       some    only a little   none at all
Whites             28%       45%      19%            6%
Blacks              63%      28%         7%           1%

Against whites:

Whites              9%       37%         32%       21%
Blacks              6%       27%         32%       30%

There is a big difference between blacks and whites in how much discrimination they see against blacks.  The difference in how much discrimination they see about whites is much smaller--whites see a little more, but not much.  For example, almost 80% of whites think that there is some discrimination against whites--but 70% of blacks think that there is some discrimination against whites.

The same question was asked in 2005.  At that time, the results were:

Against blacks:

                    A lot       some    only a little   none at all
Whites             22%       54%      14%            8%
Blacks              52%      34%         7%           5%

Against whites:

Whites              6%       39%         24%       27%
Blacks              7%       34%         20%       33%

The same pattern held then.  Between 2005 and 2015, perceptions changed a little.  Whites shfted from seeing "none at all" to "only a little" against themselves and and more saw "some" discrimination against blacks, with a decline in all other categories.  The number of blacks seeing "a lot" of discrimination against themselves grew from 52 to 63%.  However, nothing very dramatic.

I read a piece in the New York Times the other day (published a couple of months ago, but I missed it then), which said that there was "growing self-recognition among white people, prodded into being by demographic change and broader conversations about how racial identity works," which "could certainly lead toward self-acceptance and harmony . . . But we’re also staring at copious evidence of this self-recognition swinging in the other direction. . . . Some of us fixate on maintaining racial dominance, conjuring ethnonationalist states or a magical immigration formula that somehow imports half of Scandinavia. A majority of white Americans currently believe that their own race is discriminated against. News accounts fill with white resentment and torch-lit white-power marches. ...."  That would mean that white opinion was polarizing--there is no sign that is happening.  As I ,said last time, only a small minority of whites think there's more discrimination against whites than blacks--I haven't seen any survey results that would give an estimate of the number of "white nationalists," but I'm confident that it's even smaller.

[Data from the Roper Center for Public Opinion Research]

Saturday, October 6, 2018

What's the problem?

I saw a story in the New York Times the other day which reported on polls finding relatively tolerant attitudes towards immigrants and then quoted one of their reporters as saying she "had not expected voters to be quite so tolerant. . . since polls had previously found 'much higher support for people saying discrimination against whites had become as big of a problem as that against blacks and other minorities.'"  That reminded me that I had seen a number of stories mentioning that question .  The usual interpretation is summed up in this title "Why white people think they're the real victims of racism."  I had also noticed something that seemed to cast doubt on that interpretation.    In a 2012 survey, 53% of whites agreed with the statement that "today discrimination against whites has become as big a problem as discrimination against blacks and other minorities," but so did 27% of blacks.  So if we say that most white people think they are the real victims of racism, we have to say that 27% of blacks think so too.   That doesn't seem credible.

My interpretation is that the "yes" answers combined two kinds of people--those who think that whites are the real victims and those who think that there's a lot of discrimination against all kinds of people.  The second view could represent general cynicism or a kind of racial solidarity.  I couldn't think of any way to test the interpretations, so I looked for a survey that asked separately about discrimination against blacks and whites.  I found a recent one I had not seen before, from 2015.  The exact question was "now please tell me how much discrimination there is against each of these groups in our society today. How about ****? Would you say there is a lot of discrimination, some, only a little, or none at all?"  They asked about "African Americans" and "White Americans," so you can compare the responses to see how many people said there was more against African Americans, how many said it was equal, and how many said whites.  The results:
                     More vs. W        Equal       More vs.  AA
Whites                11%                39%           48%
Blacks                  2%                21%            72%

So not many whites think that they face more discrimination than blacks do.  There are a lot of people who think discrimination against blacks and whites is about equal--more among whites, but a significant number even among blacks.  Is that because they think that neither is discriminated against, or both are?  And does it matter?  I will discuss that in my next post.

[Data from the Roper Center for Public Opinion Research]

Tuesday, October 2, 2018


For a long time, going back to at least the 1950s, there seemed to be a growth in tolerance of people with unpopular opinions.  Recently there have been some claims that things are moving in the other direction.  Often these are about particular kinds of people, like liberals, millenials, or college students, but I'll start with people in general.  The General Social Survey has a series of questions going back to the 1970s about whether various kinds of people should be allowed to "make a speech in your community," "teach in a college or university," and if their books should be removed from your public library.   It asks about five types of people:  a Communist, "a man who admits that he is a homosexual," "a person who advocates doing away with elections and letting the military run the country" (or militarist for short), "a person who believes that blacks are genetically inferior" (racist),  and "a person who is against all churches and religion" (atheist).   I computed a score of tolerance for each one by adding the three items (which were all yes/no).  The means:

Support for the rights of a type of person depends on two things:  how you feel about what they say or do and how committed you are to the general principle of tolerance.  The more rapid increase for the "man who admits that he is a homosexual" can plausibly be explained by a trend towards acceptance of gays and lesbians.  On the other side, the lack of an increase for the racist can be explained by a trend toward stronger disapproval of those views.  The other three all have very similar upward trends.  Apart from the difference in trends, the year-to-year changes are very similar.  I thought there might be some distinctive movements at least for the Communist, as people might have seen it as less of a threat after the breakup of the Soviet Union, but there's no evidence of that.  The period 2004-10 saw a plateau or slight decline in tolerance, but then it started up again through 2016.  So unless you think that people happen to have become more sympathetic to Communists, militarists, and opponents of religion, it seems that there has been and still is a fairly steady growth in support for the principle of toleration.

Wednesday, September 26, 2018

Making a difference

I've written several times about opinion on the minimum wage.  Surveys almost always find strong support for increasing it.  For example, a poll in January 2014 asked "As you may know, the federal minimum wage is currently $7.25 an hour. Do you favor or oppose raising the minimum wage to $10.10?"  72% were in favor and 26% opposed.  Not many surveys have asked about the possibility of lowering it, but there was one at about the same time (December 2013):  "The federal minimum wage is now $7.25. Do you think the federal minimum wage should be raised, lowered, or should it remain the same?"  Including that option didn't make any discernible difference--71% said raised, 25% remain the same, and 2% lowered.  Some surveys have raised the possibility that an increase in the minimum wage would reduce employment, but that doesn't reduce support by much.  An example from December 2013:  "Some people say the minimum wage should be raised to help low-income workers get by. Others say raising the minimum wage will lead some businesses to cut jobs. Given these arguments, do you support or oppose raising the minimum wage? Do you feel that way strongly or somewhat?"  66% were strongly or somewhat in favor, 31% strongly or somewhat opposed.

But a Fox News poll in January 2014 included another argument:  "As you may know, the federal government sets the national minimum wage--the lowest rate in dollars per hour that most workers should be paid--which is now set at seven dollars and twenty-five cents an hour. Which of the following comes closest to your view on how the federal government should handle the minimum wage?...The government should raise the minimum wage because it would help lots of people pay their bills. The government should not raise the minimum wage because it would cause businesses to cut jobs. There shouldn't be a minimum wage because government shouldn't tell businesses what to pay their employees."  On this question 56% were in favor of raising it, 25% said it should stay the same, and 15% said there shouldn't be a minimum wage.  The question was repeated in September 2014 and only 49% favored an increase, with 21% saying there shouldn't be a minimum wage.  So the point about principle apparently had a lot more impact than the point about job loss.  It's also interesting that including it didn't just move people from saying it should not be increased to saying it should be abolished--it cut into the number supporting an increase.

[Data from the Roper Center for Public Opinion Research]

Thursday, September 20, 2018

Not sure I believe this

"How serious a problem do you think racial discrimination against blacks is in this country--a very serious problem, a somewhat serious problem, not too serious, or not at all serious?":  This question was first asked in October 1995 and repeated a number of times, most recently in September 2016.  The means, with "very serious" counted as 4, "somewhat" as 3, "not too" as 2, and "not at all" as 1:

There is a decline between 1995-6 and 2008-10, but then an increase.  The lowest value was in November 2008, just about the time that Barack Obama was elected president.  As the title suggests, I'm not sure I believe that there really was a large increase between 2010 and 2015-6.  It seems that answers to this question might be influenced by context--if you ask the question after questions about various kinds of racial inequality, ratings of seriousness would be higher than if it's part of a serious of miscellaneous questions.  Perhaps the surveys were different in this respect.  However, the possibility of a shift is interesting, since most recent commentary on racial attitudes has emphasized the lack of change--for example, a column by Thomas Edsall in the New York Times today said that Obama's health care proposals "hit the wall that often confronts Democratic policymakers: race" and drew parallels to the late 1960s. 

If there was a change between 2010 and 2015, and a further change between 2015 and 2016, why did it happen?  The most obvious possibility would be the publicity given to police violence against blacks, which in many cases was supported by video evidence.  Some people argue that this led to a backlash, with whites rallying around police, but maybe it had a straightforward effect--some people were persuaded that there was a real problem.

[Data from the Roper Center for Public Opinion Research]

Saturday, September 15, 2018

Very dishonest people

I had posts a couple of weeks ago about views of the "honesty and ethical standards" in various occupations.  The figures for "college teachers" had an unusual pattern, rising from the 1970s until about 2000 and then declining.  Breakdowns by party ID are available for some years and shown in this graph:

The average rating from Democrats rose after 2000, while the average rating among Republicans fell.  There was only a small gap in average ratings in the 1990s, but by 2012 Democrats rated the honesty and ethical standards of college teachers substantially higher. 

I looked at two other occupations for comparison.  I didn't have time to do all years, so I just took 1990, 2000, and 2012.  For journalists:

                R         D        Diff.
1990     3.09     3.28       0.19
2000     2.66     3.15       0.49
2012     2.57     3.21       0.64

Also divergence, although it grew more in the 1990s--most of the divergence on college teachers was after 2000. 

For business executives:

               R         D
1990     3.19    3.10      -0.09
2000     3.18    3.03      -0.15
2012     3.03    2.81      -0.22

Maybe a slight divergence, but nothing dramatic.  The substantial changes are for two groups that are favorite targets for conservatives.  It's also worth noting that there are obvious reasons that the ratings of executives would have declined between 2000 and 2012, just because of the state of the economy.  However, that doesn't apply to college teachers--the economy has ups and downs, but the amount and quality of what professors do doesn't change much from year to year.  So my guess is that conservative media outlets just started running more stories about outrages on college campuses, and ratings among Republicans declined as a result.  The rise among Democrats might have been a reaction to that, but I think it's more likely to be a continuation of the previous upward trend (which I would attribute to more people having attended college).   

[Data from the Roper Center for Public Opinion Research]

Sunday, September 9, 2018


In 1992, a poll asked "In your view, are most people who receive welfare payments genuinely in need of help or are they taking advantage of the system?"  The question was asked several times over the next couple of years, and then there was a long gap before it was asked in 2012, and then in 2016 and 2017.  The figure shows percent saying "genuinely in need" minus percent saying "taking advantage."

Views are considerably more favorable in recent years--in fact, the April 2017 survey is the first one in which more than half chose "genuinely in need."  Why the change?  Several possibilities occur to me:
1.  Welfare may be less strongly associated with race in the public mind.  At one time, blacks were overrepresented in news stories on poverty and welfare.  I don't know about more recent research, but it seems possible that news organizations became more aware of this and made efforts to avoid it.  My impression is that there is more attention to small-town and rural (that is, mostly white) poverty than there used to be.  If there has been a change in media coverage, public perceptions might have followed. 
2.  Animosity to blacks is declining, so even if welfare is still associated with race, people may not be as bothered by the idea that the government is spending to help blacks.  The idea that there has been real decline in anti-black prejudice is not popular in sociology now, but there's a lot of evidence for it.
3.  It may be a result of changes in anti-poverty programs.  There's been a big growth in the Earned Income Tax Credit and Child Tax Credit, which are popular, rather than AFDC/TANF, which are not.  Also, my guess is that welfare programs are administered more efficiently now--one thing that computers are good for is keeping track of people and money.

1.  See this post for a related question.
2.  I noticed a poll from 2013 which showed 30% "genuinely in need" and 56% "taking advantage"--more like the 1990s than the other surveys from the 2010s.  But on reading the fine print, I found that it was a sample of Hispanics only.  Hispanics have more negative views of welfare than non-Hispanics?  Apparently yes--I checked one of the other recent surveys and it was the case there.

[Data from the Roper Center for Public Opinion Research]

Monday, September 3, 2018

Who's the problem? (part 3)

   Changes in average ratings of the "honesty and ethical standards" of more occupations.  First, some professions:

Ratings of the clergy declined over most of the period, and ratings for lawyers might have declined a little.  Engineers increased, and there was no clear change for accountants.
    Ratings for professions related to medicine rose;

Then some occupations that don't fit into any of the previous categories.  They generally increased.

Finally, a group that's of particular interest to me.  There's a unique pattern for college teachers:  an increase until about 2000 and then a decline, with a big drop between 2012 and 2016.  Data for elementary and high school teachers don't begin until about 2000, but both of them seem to have a declining trend since then. 

That's a lot of data.  What conclusions can be drawn?  Here's what I notice:

   1.   There are more upward than downward trends.  And the trends are upward for almost all of the "ordinary" occupations--ones that are not in the news or the subject of political controversy.   General trust in people has been declining since the 1970s, and you might think that this would lead to a declining trust in most occupations.  But this is not what has happened.
   2.  Almost all of them rose in 2001.  Of the 21 occupations which were included in 2000 and 2001, 19 had increased, and the two declines were very small.  This was presumably because the 9/11 attacks (the 2001 survey was taken in late November) created more general solidarity.  I thought that the increases might be larger for occupations that could be seen as closer to the center of American life, but there was no obvious pattern.
  3.  There are differences among the political occupations.  Members of Congress and Senators declined a lot after 2000, governors and state officeholders declined by a smaller amount, and local officeholders didn't decline at all.   That is, people are making distinctions, at least to some extent:  turning against congress more than against "government" or "politics" in general.
4.  The decline for all types of journalists is pretty closely.  I thought it might track views of politicians (especially Congress) more closely--that is, that journalists would be blamed when they brought bad news.  But there's not much sign of that.

I also have some thoughts about the recent decline for college teachers, which I will discuss in a future post. 

Friday, August 31, 2018

Who's the problem (part 2)

Here are graphs showing changes in the average rating of the "honesty and ethical standards" in various occupations.  There were a lot of them, so I divided them into groups.  First, politics:

Until 200, there was little or no trend for any of them.  Since 2000, Senators and members of Congress have dropped substantially, governors and state officials have dropped by a more moderate amount, and local officeholders have stayed about the same. 

Then there are occupations that I classed as "related to politics":  lobbyists and several different types of journalists. 

For the journalists, it seems to be a pretty steady downward trend since the 1970s.  Lobbyists are rated much lower, and there is no change over the decade for which the question has been asked. 

Then some business occupations:

Bankers and stockbrokers declined in 2008 and have not recovered.  Business executives have a more steady downward trend.  But people in advertising, HMO managers, and nursing home operators show no change or maybe an increase.  Around 1980, stockbrokers were rated much higher than people in advertising--now they are about the same. 

Then justice and the military: 

They have only asked about judges and the military since about 2000--judges have declined a bit, while military officers have stayed about the same.  For police, there is an upward trend until about 2000 and little change since then.  Two individual years stand out--one is 2001, when a number of occupations had a jump, which was probably a consequence of 9/11 (the survey was taken in December).  The other is 2014, which was when the Michael Brown shooting and protests in Ferguson, Mo. took place. 

That's a lot of data, but there's more, so I will save the rest of it for my next post. 

Tuesday, August 28, 2018

Who's the problem? (part 1)

Some people argue that people are unhappy with the state of American economy and society because of things like slow economic growth and rising inequality.  I have had several posts arguing against that idea--this is the most recent and this is another one.  People are not especially discontented with their economic situation, or the state of the nation, or life in general.  However, confidence in most American institutions has declined since the 1970s.  Is that because people have become more negative about elites, or certain elites, or people in general?

Since the 1970s, the Gallup Poll has asked "Please tell me how you would rate the honesty and ethical standards of people in these different fields -- very high, high, average, low or very low? How about.."  The list of occupations changes--forty-three different ones have been included, some only once but others as many as 35 times.  They include some true elites (e. g., Senators), some professions that cover a wide range (e. g., lawyers, clergy), and some ordinary jobs (e. g., auto mechanics).  There's a lot of information there, so I'll break my discussion into several parts.  One of the first thing I did was to fit a time trend to each occupation.  27 were positive (that is, in the direction of higher honesty and ethics), and 10 of those were statistically significant at the 5% level.  Fourteen were negative, and ten of those were statistically significant.  (Two of them were asked just once, so no trend could be estimated).  The biggest statistically significant upward trends were:  nursing home operators, auto mechanics, funeral home operators, labor union leaders, and medical doctors.  That's a diverse group--I can't think of anything that they have in common.  The largest statistically significant declines were state governors, stockbrokers, members of Congress, TV reporters, and bankers.  Those could all be regarded as elite occupations.   

Sunday, August 26, 2018

Ideology and morality

About a week ago, Donald Trump advised people to "study the late Senator Joseph McCarthy."  It turns out that I have been thinking about Joe McCarthy, although not quite for the reasons Trump says we should.  I was looking at an essay by Daniel Bell, originally published in 1953 and reprinted in The End of Ideology (1960).  In it, he said "the tendency to convert concrete issues into ideological problems, to invest them with moral color and high emotional charge, is to invite conflicts which can only damage a society.  ... It has been one of the glories of the United States that politics has always been a pragmatic give-and-take rather than a series of wars-to-the-death."  The second sentence reflects a conventional view of American politics, but on reflection it doesn't seem convincing.  Of course, there has been a lot of pragmatic give and take, but compared to other countries Americans seem to have had a tendency to invest issues with "moral color and high emotional charge."  For example, alcohol had been widely used in American society for centuries, but was completely banned in 1920.  I don't think anything like this happened elsewhere--there was a strong temperance movement in Britain, but it never came close to achieving prohibition, even thought that would just have taken an ordinary act of parliament, while in the United States it required a constitutional amendment.  A lot of people must have felt very strongly to devote that much effort to the cause and not to be satisfied with anything less than complete prohibition. 

This example shows a problem with Bell's first sentence.  An issue can have moral and emotional charge without being part of an ideology:  that is, "an all-inclusive system of comprehensive reality" (quoting Bell again, this time "The End of Ideology in the West").  Prohibition wasn't an ideology like socialism--it was a position on one issue.  I happened to run across a 1974 article by Samuel Huntington which made this distinction, observing that "highly systematized ideologies . . . have been notably absent from the American scene.  But it is a mistake to move from this truth to the assumption that political ideals have played a less important role in the United States than in Europe. . . . American politics has been characterized by less sophisticated political theory and more intense political beliefs than most other societies."

Bell concluded his essay on McCarthy by suggesting that the conflict would pass pretty quickly.  He was right about that.  In contrast, for at least the last decade the United States has been repeating the same conflicts, like those over immigration and health care, without coming closer to a resolution.   I wonder if what has made recent conflicts so enduring is that the traditional "moral color" of American politics has come to be combined with ideology. 

Sunday, August 19, 2018

It's the rich wot gets the gravy

"If ______ is elected President, do you think the policies of his/her administration will favor the rich, favor the middle class, favor the poor, or will they treat all groups the same?"

                                                   Rich          Middle         Poor      Same      DK
Aug 2007   John Edwards            30%          24%            9%         18%      19%
Mar 2008   Hillary Clinton           23%          29%          13%         28%       7%
Mar 2008   Barack Obama           13%          30%           18%        33%        6%
Mar 2008   John McCain              53%          16%           0             23%        8%
Oct  2008   McCain                       59%          11%            3%        21%         6%
Oct  2008   Obama                          8%          38%           22%       24%         8%
Jul   2012   Mitt Romney               53%          11%             2%       30%         4%
Sep  2012  Obama                          12%          26%          22%        30%       10%
Sep  2012  Romney                        53%            8%            1%        33%          6%
Sep  2012  Obama                            9%           27%          31%       26%          7%
Oct  2016   Donald Trump             57%           14%            1%       27%          1%
Oct  2016   Clinton                         37%           24%          14%       22%          3%

The figures for the Republicans--McCain, Romney, and Trump--are just about the same, with between 53 and 59 percent saying their policies would favor the rich, but there are substantial differences among the Democrats.  With Obama, between 9 and 13 percent said his policies would favor the rich; with Hillary Clinton, it was 23% in 2008 and 37% in 2016.  I had a post about a similar question that was asked in 2008 and June-July 2016, which also showed that McCain and Trump were just about the same but that Clinton was substantially different from Obama.  I said "One possibility is that it's a fixed part of her image--maybe people are thinking of the well-compensated speeches she's made to Wall Street firms.  Another possibility is that the contrast with Bernie Sanders made people think of her as more favorable to rich, and that as people start focusing on the contrast with Trump perceptions will change...."  We can rule the second one out given the results of the October 2016 survey (it was taken less than two weeks before the election).  However, it wasn't completely fixed--it was there in 2008 and stronger in 2016. 

The pattern doesn't fit the way that the different candidates have usually been depicted in the media.   In 2008, Clinton was usually presented as down-to-earth and Obama as a bit of an elitist (sometimes even "professorial".  And there have been many stories contrasting the conventional businessman Romney with the "populist" Trump, who sometimes talked about an infrastructure program or closing tax loopholes that benefited "the hedge fund guys." 

[Data from the Roper Center for Public Opinion Research] 

Wednesday, August 15, 2018


I had a post a few weeks ago about satisfaction with life in general, specifically:  "Here is a ladder symbolic of the 'ladder of life'. Let's suppose the top of the ladder represents the best possible life for you, and the bottom represents the worst possible life for you. On which step of the ladder do you feel you personally stand at the present time?"  That showed data from 1964 to 2014, and no trend was apparent.  But since the late 1960s is sometimes seen as a turning point, that means there wasn't much information on the "before" state (only surveys from 1964 and 1966).  I found that there was an earlier example of the same question, in 1959.  The results including that survey:

Again, no sign of a trend.  The average for 1959 is pretty much in the middle, higher than the figures for 2009 and 2001, but lower than 2005, 2006, and 2014.  There is a popular (and plausible) story which holds that after growth in average incomes slowed down in the 1970s, people became more discontented, which made them less generous and more inclined to look for someone to blame, and the longer it went on, the more discontented people became, making them even less generous and more punitive.  If this is true, the only way to change the public mood is to return to faster and more broadly based economic growth, which no one has any idea how to accomplish.  So in a way, the absence of a trend in perceived position on the "ladder of life" is encouraging.

[Data from the Roper Center for Public Opinion Research] 

Saturday, August 11, 2018

Love him or hate him

Only about 40-45% of people say they approve of the job Donald Trump is doing as President, but it's often said that his supporters are especially enthusiastic.  Some surveys that ask the question about approval follow with one about whether you approve/disapprove "strongly" or just "somewhat."  I looked at those for presidents back to Jimmy Carter.  To limit the amount of data, I picked the first time when the President in question had roughly the same overall approval as Trump does now.  The figure shows the percent who strongly approved and strongly disapproved, along with the predicted values from simple time trend.  

There is an upward trend for both, and it's about the same size for approval and disapproval.  That is, there's more strong approval and disapproval, and fewer "somewhat" answers.  Relative to the trend, Trump doesn't stand out in terms of either strong approval or strong disapproval.  Two who did stand out to some extent were George W. Bush, with more strong reactions, and Bill Clinton, with fewer.  I noticed that the date I had picked for Bush was early September 2005, just after hurricane Katrina, and thought that might have temporarily boosted his strong disapproves, so I added one from July 2005.  His overall approval rating was a bit better then, with fewer strong disapproves and about the same number of strong approves--he was still above the trend for both.

I was expecting more differences among individual presidents, but seem to have found one more example of the gradual growth of partisan polarization since the 1970s.

[Data from the Roper Center for Public Opinion Research]

Wednesday, August 8, 2018

By popular demand

I had a post a couple of years ago about the effect of height and weight on earnings for men and women.  Basically, the pattern seemed to be that for women, being thinner meant higher earnings; for men, earnings were highest in a middle range.  That is, for the purposes of earnings, women couldn't be too thin, but men could.  Recently, someone asked in a comment "What happens if you superimpose the plots for men and women?"  The literal answer is that it would be pretty much unreadable, since I had lines for the height/weight relationship at four different heights for each gender.  Another issue is that I have since decided that the way I constructed the lines (as a function of height, weight, and weight squared) was misleading.  So I went back and computed average earnings of men and women by BMI (rounded to whole numbers).  The numbers are small for some categories, so I show smoothed curves as well the the means.

My previous conclusions about the relationship among women need to be revised:  now it appears that weight doesn't make much difference up to a BMI of about 25 (which is where the official "overweight" range starts); after that it goes with lower earnings.  For men, my previous conclusions were about right--men who are in the official "overweight" range (25-30) earn more than men in the "normal" range (18-25).  Men earn more than women over most of the range, but the difference disappears and is even reversed at the low end.  That is, skinny men appear to earn low incomes, compared with men who weigh more and also to women. 

There are not many men in the lower ranges of BMI--for example, for someone who is 70 inches tall, a BMI of 20 means a weight of 139 pounds.  Still, it is interesting that deviating from the "ideal" weight seems to matter more for men than for women, and that for men being in the "normal" weight range is worse than being in the "overweight" range. 

Friday, August 3, 2018

Tough enough?

In September 2017, an ABC News/Washington Post poll asked "Before (Donald) Trump became president, do you think the US was too tough in enforcing immigration laws, not tough enough or was enforcement about right?"  and then "How about now, under (Donald) Trump, do you think the US is too tough in enforcing immigration laws, not tough enough or is enforcement about right?"  The results:

                                      Before                    Now
Too tough                         6%                        45%
About right                     44%                        30%
Not enough                     49%                        22%
People clearly saw Trump as having made enforcement tougher, and appear to have been about equally divided on whether that was a good thing:  slightly over half (52%) said that things were either about right or still not tough enough; 50% said that things had been about right or too tough.  Of course, the assessment of "before Trump" might have been influenced by experience under Trump.  In 2004, there was a very similar question:  " How do you rate the federal government on immigration? Is it too tough, not tough enough, or about right?" 10% said too tough, 61% not tough enough, and 26% about right.  That's considerably less favorable than the 2017 "before Trump" assessment, but opinions could have changed quite a bit in the 10+ years before Trump was elected.  The most similar question I could find in those years was in 2013, when another ABC News/Washington Post Poll asked "Overall, do you support or oppose...stricter border control to try to reduce illegal immigration? Do you feel that way strongly or somewhat?"  63% said support strongly, 17% support somewhat, and 10% oppose somewhat and 7% oppose strongly.  Despite the difference in the questions, it seems reasonable to take the "strongly support stricter controls" as roughly equivalent to "not tough enough."  So as I suggested in my last post, a solid majority supported tougher enforcement in principle before the rise of Trump; after getting it, support fell, either because of what people saw and heard about the effects of the policy, or because the policy was associated with Trump.  But with about 50% support, it was (as of Sept 2017) still more popular than Trump was overall. A lot has happened since that time, but I can't find any comparable survey questions.  My guess is that opinions have not changed much, since general opinions about Trump have been very stable.

[Data from the Roper Center for Public Opinion Research]

Monday, July 30, 2018

Populism and popularity

Ross Douthat had a column on Sunday in which he asked why Donald Trump is pressing for tariffs, even though they are unpopular with Republicans in Congress and not very popular in the country as a whole.  He says it's because Trump has broken all of his other populist promises (like an infrastructure program or tax cut directed to the middle class), so this is all he has left.  Oddly, he missed another example, the crackdown on illegal immigration.  That was one of Trump's central issues, and he's definitely sticking to it.  As with trade, the public does not rate his performance highly:  in a Washington Post/Schar School poll in late June 41% said they approved of the way he was handling trade, and 39% said that the approved of the way he was handling immigration.

Douthat's explanation for why Trump wants to keep some "populist" elements is that it's a way to hold onto "working class" voters who put him over the top in 2016, even at the cost of driving away middle-class suburbanites.  The idea seems to be that those voters are located in Midwestern swing states, so they are more valuable.  I think this explanation attributes an implausible amount of strategic thinking to Trump.  In my view, there are two reasons that he's sticking to these policies.  The first is simply that Trump has strong beliefs on them:  they were major themes in his tweets from the beginning, while the other "populist" elements didn't show up until he started his campaign.  The second is that the ideas of "getting tough" with foreign countries and illegal immigration had been popular, and are still fairly popular. 

  Between 2005 and 2010, the Washington Post asked the following question nine times:  Do you think the United States is or is not doing enough to keep illegal immigrants from coming into this country? .... Do you feel that way strongly or somewhat?"  The distribution of answers barely changed, so I'll just show the figures from 2010:  10% doing enough (strongly) 13% doing enough (somewhat) 17% not doing enough (somewhat) 58% not doing enough (strongly). 

Of course, the actual policies involved in "doing more" have been less popular, just like actual cuts in government spending are less popular than the general principle of "cutting government spending.   Still, when the question was asked this June,  46% still said "not doing enough," versus 50% who said "doing enough."  (Unfortunately this question didn't have a "too much" or "going too far" option, but I'll discuss one that did in my next post).  The Democrats have traditionally benefited from the image of being more interested in the average person, apart from any specific issue positions. There may be a parallel advantage in having a "get tough" image on immigration and trade. 

[Data from the Roper Center for Public Opinion Research]

Friday, July 27, 2018

The s-word

There has been a lot of talk about socialism since the upset win of Alexandria Ocasio Cortez (a member of the Democratic Socialists of America) in a New York congressional primary.  After the 2008 and 2012 elections, there were brief flurries of attention to socialism, with some people saying that support was on the rise.  There's some of that now, but most of the attention now seems to come from Republicans thinking that it's something that they can use against the Democrats.  In general, the public does have a negative view of socialism:  in the most recent question I could find, a 2016 Gallup poll, 35% said they had a positive image and 58% said they had a negative one.  That leads to a question of whether it's just a label, unconnected to other political views, or part of a pattern. 

I addressed this question using a 2011 Pew survey that asked people if they had a positive or negative reaction to socialism and five other political terms:  liberal, conservative, libertarian, progressive, and capitalism.  I computed correlations between each pair.  Views of "socialism" had a positive correlation with views of "liberal" and "progressive," and negative correlations with views of "conservative," and "capitalism."  More surprisingly, they had a positive correlation with views of "libertarian"--it was the second largest correlation, behind "liberal."  Since familiarity with ideological terms increases with education, I then did the correlations separately for college graduates and everyone else.  Of course, you can expect the correlations to be larger for college graduates, but there's a question of whether they are just stronger and weaker versions of the same underlying pattern, or two distinct patterns.  Among college graduates, there are positive correlations with "liberal" and "progressive," and negative ones with "conservative" and "capitalism."  The correlation with "libertarian" is essentially zero (.03), which you might regard as strange, but I'll leave that aside for the moment.  Among non-graduates, there are positive correlations with liberal, progressive, libertarian, and conservative.  The only negative correlation is with "capitalism" and that's not significantly different from zero.  There's a striking difference between liberal, where the correlation is .38 for educated people and .33 for less educated, and conservative, where the correlations are -.29 and +.085. 

I'm not sure exactly what to make of this, but it seems that less educated people have a different understanding of "socialism" than more educated people, and that it's not just a matter of being less familiar with ideological terms. 

Another interesting point is that the term "libertarian" apparently hasn't made much impression on the public--the three strongest correlations are with liberal, progressive, and socialist among non-graduates.  All of those are in what would conventionally be regarded as the wrong direction (positive).  The correlations with "capitalism" are small in both non-graduates and graduates (.07 and .08).

[Data from the Roper Center for Public Opinion Research]

Monday, July 23, 2018

Tip of the iceberg?

I had several ideas for posts that turned out to require more work than I expected, so here is a short one.   A few weeks ago, the New York Times had a story about the First Amendment that said "in 1977, many liberals supported the right of the American Nazi Party to march among Holocaust survivors in Skokie, Ill. Far fewer supported the free-speech rights of the white nationalists who marched last year in Charlottesville, Va." The links just connected to stories about those events, not to information about public opinion.  However, I found a 1978 Roper survey that asked about the Skokie march (which had not yet taken place--1977 was the Supreme Court decision that found in favor of the right to march):  15% said it should be allowed, 73% said it shouldn't, 7% gave answers described as "yes, but it would be unfortunate if they did," and 6% didn't know.  Support among liberals was somewhat higher, but only 30% said yes or "yes, but" and 66% said no.  I guess you could call 30% "many," but the normal way to describe the distribution would be that a solid majority of liberals opposed the American Nazi Party's right to march.  As far as Charlottesville, I couldn't find any surveys that asked about the white nationalists' right to march, although there were many that asked about whether people approved of how Donald Trump handled it (of course, most did not). So there seems to be no evidence that liberal opinion moved against the right of extremists to hold marches. 

In fairness, the story was mostly about judges and legal scholars--the claim about liberals in general was just made in passing.  It's possible that progressive legal intellectuals have become somewhat less enthusiastic about free speech over the last 40 years--my not very well informed view is that they probably have.  And in the long run, major changes in opinion are roughly parallel in elites and the general public--for example, it's safe to say that there's more support for gender equality in both than there was 50 years ago.  However, it's not safe to say that the general public follows smaller or more subtle shifts in elite opinion. 

[Data from the Roper Center for Public Opinion Research]

Saturday, July 14, 2018

Education and redistribution

  As a general rule, more educated people are more liberal than less educated people on most "social" issues and more conservative on most economic issues. I wondered if this pattern has changed, so I looked at a question the General Social Survey has asked since 1978:  "Some people think that 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. Others think the government should not concern itself with reducing this income difference between the rich and the poor." People are shown a card with numbers from 1 (should do something) to 7 (should not concern itself) and pick the number that best represents their views.  If we limit things to people who are not black and compare those who graduated from college to everyone else, here are the means:

There's little or no trend among people without college degrees, but a downward trend--that is, more support for redistribution--among college graduates.  There are some year-to-year ups and downs that apply to both groups (partly related to the party of the president), so if you look at the difference the trend is even clearer:

The drop in 2016 is unusually large--if you fit a trend from 1978-2014 and extrapolate to 2016, the standardized residual is -2.63--but the trend is clear even without it.  That is, there has been a gradual decline in the the difference between the opinions of more and less educated people.  Why?  One possibility is that it has to do with economic trends--the income gap between college graduates and other people has been growing since the mid-1970s, so educated people feel more generous or more guilty, and more inclined to do something.  Another is that it's about politics--more educated people have been shifting towards the Democrats over the same period.  Regardless of the original reason for the shift, once you start voting for a party, you'll tend to have more trust in its leaders, and adopt more of the positions usually associated with that party.  I'm sure there are others, but those are the ones that occur to me.  

Saturday, July 7, 2018

The turning point?

On the 4th of July, the New York Times had a quiz about American history.  One of the questions was "What question, posed to Senator Joseph McCarthy by the Army lawyer Joseph Welch in 1954, is often cited as the unraveling point of McCarthy’s anti-Communist campaign?"  The correct answer was "have you no sense of decency?"  They add, "McCarthy’s national popularity disappeared overnight, and he died three years later" and link to the official website of the Senate, which says almost the same thing:  "Overnight, McCarthy's immense national popularity evaporated." 

The Gallup Poll had a number of questions about whether people had a favorable or unfavorable opinion of McCarthy.  This figure shows the percent favorable minus percent unfavorable (there were two different versions of the question, which I indicate by different colored dots):

The third vertical line is the date of the Welch/McCarthy exchange.  There's not much sign that it made any difference.  Moreover, McCarthy's popularity was not "immense" before it happened--in the last survey before it happened (late May), 33% were favorable and 43% unfavorable. 

Just looking at the numbers, it seems that there is one event that might have caused a lasting drop in McCarthy's popularity--a critical TV documentary by Edward R. Murrow that ran on March 9.  There seems to have been a downward trend even before that, but there was a large decline between March 2 (+10) and March 24 (-8 and -15) and support for McCarthy never bounced back. 

Why is the idea that "have you no sense of decency" was decisive so popular (I have heard it before)?  It's a satisfying story--people saw the exchange and recognized McCarthy for what he was.  If you say that public opinion was influenced by elites, like Murrow or the Senators who decided to have hearings on accusations against McCarthy, that raises questions.  What if Senate Republicans had stuck together behind McCarthy?  What if Murrow hadn't decided to do the program, or if the network executives had refused to let him run it?  Should journalists express a point of view rather than just report the facts?  It's more comforting to believe that people spontaneously saw the truth than to think about those kinds of things.

[Data from the Roper Center for Public Opinion Research]

Wednesday, July 4, 2018

Been down so long

A book review in the New York Times started off with a discussion of how discontented people are today.  It cited a Gallup question on "In general, are you satisfied or dissatisfied with the way things are going in the United States at this time?" Answers to this question don't actually show a clear trend since it started in the late 1970s.   But maybe the comparison shouldn't be now vs. a few years ago, but the 1970s and after vs. the 1960s and before?  In fact, the book in question refers to " America’s Fifty-Year Fail " in its subtitle, and the 1970s was a turning point in terms of economics--from a long decline in economic inequality to a long rise.

Public opinion surveys were not as common before the 1970s as they are today, so it's hard to address the possibility that there was a lasting drop after the 1960s.  In 1952, Gallup asked "As you look to the future, do you think life for people generally will get better, or will it get worse?"  That question was repeated in 1962, 1979, 1989, and 2009.  As discussed in this post, answers don't show a decline in optimism.  However, the 2009 survey was taken in January, and there might have been a short-lived spell of optimism accompanying the inauguration of Barack Obama.  Unfortunately, that question has not been repeated since 2009, so I looked for other possibilities.

In 1964, a special survey by the Gallup poll asked "Here is a ladder symbolic of the 'ladder of life'. Let's suppose the top of the ladder represents the best possible life for you, and the bottom represents the worst possible life for you. On which step of the ladder do you feel you personally stand at the present time?" and this has been repeated a number of times since then.  The means:

The last three times the question was asked were 2009, 2011, and 2014.  The mean was low in 2009 and 2011, but much higher in 2014.  That suggests that assessments might respond to economic conditions and in fact the mean has a substantial correlation (-0.73) with the unemployment rate.  There is no apparent trend or lasting one-time drop.

I've had several posts (e. g., this one) arguing that people are not all that discontented with general economic and social conditions--they are discontented with politics.  I think these figures give further support for that position.  In fact, you might wonder why people haven't become more discontented, given the slow growth (some would say absence of growth) in average family incomes over the last 40 or 50 years.  I would say that it's because people mostly compare themselves to people around them and to their own past--whether they are better off than they were and are keeping up with other people they know.  Whether average income growth was faster for your parents' generation, or whether rich people are getting bigger gains, are too remote to have much impact on how people rate their own lives.

[Data from the Roper Center for Public Opinion Research]

Friday, June 29, 2018

Vacation from facts, 3

I am back from vacation, but in my previous post I proposed that "a pure two-party system promotes ideological politics" and said that my next post would consider the question of why ideological differences between American parties were small until about 50 years ago.  My answer is "tribalism"---at one time, many people voted purely on the basis of ethnic, religious, or regional loyalties, without paying much attention to ideology.  As more people started to think in ideological terms, the tendency towards divergence started to take effect. 

For comparative purposes, the key facts are that "tribalism" was an unusually strong force in the United States because of size, ethnic diversity, and other historical factors, and that we have an unusually strong two-party system, probably because of political institutions.   That combination produced a unique path in the ideological differences between parties. 

Wednesday, June 20, 2018

Vacation from facts, 2

  My last post finished by asking why ideological politics has grown in the United States, in contrast to almost all other countries.  A distinctive feature of American political institutions is the dominance of two parties.  There have only been a handful of members of Congress who belonged to other parties; some third-party candidates for President have received a significant share of the vote, but none of those parties have lasted.

In a two-party system, ideological competition drives parties to the center, as Anthony Downs argued.  But there's another way to compete:  if you convince people that the other side is totally unacceptable, then they have no choice but to vote for you.  It seems to be easier to motivate people by fear rather than by a positive vision, so focusing on the negative may be a more attractive strategy than moving to the center.*  I have mentioned in several posts that negative feelings about both parties have grown.  In 2016, many people who weren't enthusiastic about Donald Trump voted for him anyway because they couldn't bear the idea of Hillary Clinton as President.  Some never-Trumpers voted for Gary Johnson or Evan McMullin or wrote someone in.  Others voted for Clinton but didn't publicly support her.   In contrast, if the British Conservatives chose someone like Trump as a leader, party members who were unhappy could turn to the Liberal Democrats.  Those votes wouldn't be wasted--by winning a small number of seats, they could produce a hung parliament, which has happened several times.  So people who strongly objected to the leader would not just vote for the Liberal Democrats, but publicly advocate voting for them, further strengthening the movement away from the Conservatives.

So my suggestion is that a pure two-party system promotes ideological politics.  If true, that raises the question of why the ideological differences between the American parties were small until about 50 years ago.  I will consider that in my next post.

*Why not do both--move to the center and promote negative feelings about the other party?  For example, you could try to convince people that they are incompetent or corrupt.  However, it seems to be easier to create strong negative feelings when the charges have some ideological content--the other party will take us down the slippery slope to a Soviet-style planned economy, or a Handmaid's Tale society.  In order to make those kind of charges seem sincere, you have to stake out an extreme position yourself--no compromise on X.

Friday, June 15, 2018

Vacation from facts, 1

I will be on vacation when this post appears, so it seems appropriate to take a break from facts and engage in speculation.

"Tribalism" has become a favorite word in writing about contemporary politics.  It seems like the wrong word to me--the key thing about a tribe is that you don't choose it, you are born into it.  A second feature is that tribal leaders have a good deal of freedom in conducting relations with other tribes (see this paper, p. 141)--if they say that we've traditionally been allied with group A, but now it's in our interests to make an alliance with group B, the members will go along.  Tribal politics can involve intense conflict, but it can also involve toleration and coexistence--you can't blame someone for being born a member of a different tribe, and there's a chance of winning them over by making a deal with their leaders.

What we have now is ideological politics.  where people choose a side because it represents the right principles.  Ideological politics necessarily involves conflict.  You can definitely blame someone for choosing the wrong principles; also, leaders have less freedom, because the members may revolt if they seem to betray those principles.  It's sometimes said that Republicans have abandoned their principles to follow Trump, but when those principles are specified they turn out to be things like free trade, concern about budget deficits, and the rule of law, which aren't traditional Republican or conservative principles--they cut across party and ideological lines, and are probably strongest in the "good government" center.  If Trump did something that really went against conservative principles--e.g., proposed a program of infrastructure spending financed by closing tax loopholes that benefit high earners--there would be a revolt.  Of course, I can't give evidence of that, because Trump has conformed to conservative orthodoxy on everything that's important to conservatives--you don't have to take my word for it, you can take Mitch McConnell's. 

That raises a question of why ideological politics grew in the United States.   In almost all other affluent democracies, it has been declining for a long time, and the decline seems to be continuing.  I will turn to that in my next post. 

Friday, June 8, 2018

Measuring racial resentment

My last post discussed a scale that is usually referred to as "racial resentment."  The questions (all with responses going from strongly agree to strongly disagree) are:
  1. Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.
   2. Irish, Italian, Jewish, and many other minorities overcame prejudice and worked their way up.  Blacks should do the same without any special favors.
   3. It's really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.
  4. Over the past few years, blacks have gotten less than they deserve.

I'd say that the first question simply measures perceptions of how much racial inequality there is. It showed a strong trend towards "disagree" between 1986 (when it was first asked) and 2012.  Even with the move towards "agree" in 2016 the correlation is about 0.8.  That is, people see less racial inequality than they used to, which is reasonable given actual changes in society. 

With the other three, the end that is scored as "resentment" can include two kinds of people--those who think that things are reasonably fair, and that blacks haven't taken advantage of opportunities, and those who think that blacks are getting some kind of unfair advantage.  In terms of the last question:   people who think that blacks have gotten about what they deserve and those who think they've gotten more than they deserve.  So they are basically just measures of general liberalism versus conservatism on the causes of racial inequality.  None of those showed a strong trend through 2012. 

In 2013, I had a post about a question that seemed like a pretty good measure of resentment "For each of the following groups, please tell me whether you feel that they are receiving too many special advantages, receiving fair treatment, or are being discriminated against?" The groups included blacks/African Americans.  I remarked then that unfortunately the question hadn't been asked since 2008, and it still hasn't.  It also goes back only to 1990, so I took another look for questions that might be regarded as measuring racial resentment. 

There is one question that was asked several times in the 1970s, and then reappeared in a very similar form in the 2010s.  The 1970s version asked if they agreed or disagreed with the statement "the government has paid too much attention to the problems of blacks and other minorities."  The 2010s version:  "Over the past couple of decades, the government has paid too much attention to the problems of blacks and other minorities," with "completely agree," "mostly agree," "mostly disagree," and "completely disagree" as possible responses.  If we collapse the 2010 categories into agree and disagree, the percent agreeing minus disagree is:

The figures for 1976, 1978, and 2012 are averages of multiple surveys (5, 2, and 3). Although the question wording differs, I don't think that could plausibly account for the difference in responses. It's reasonable to conclude that there's less racial resentment now than there was in the 1970s. 

[Data from the Roper Center for Public Opinion Research]

Tuesday, June 5, 2018

The Era of Good Feelings?

The paper I discussed in my last post showed changes in an index of "racial resentment" from the American National Election Studies, which rose in 2008 and 2012.  The authors interpreted this as evidence that whites felt the presidency of Barack Obama as a threat to their status.  That reminded me that I had a post mentioning the index several years ago.  The index is the sum of responses (strongly agree .... strongly disagree, reverse coded for #2 and #3) to the following statements:

   1. Generations of slavery and discrimination have created conditions that make it difficult for blacks to work their way out of the lower class.
   2. Irish, Italian, Jewish, and many other minorities overcame prejudice and worked their way up.  Blacks should do the same without any special favors.
   3. It's really a matter of some people not trying hard enough; if blacks would only try harder they could be just as well off as whites.
  4. Over the past few years, blacks have gotten less than they deserve.

At the time, I said "this scale certainly measures something of interest, but 'resentment' doesn't seem like the right term."  That was just my feeling based on reading the questions--does the data shed any light on the issue?   Let's start with looking at changes (among whites) over the entire period covered by the ANES:

So either racial resentment fell dramatically among whites in 2016, or the index doesn't really measure racial resentment.  The first interpretation doesn't seem very plausible, but the ANES survey has two parts, one of which takes place after the election, and that's the part in which these questions were asked.  So you could say that perhaps white fears of threats to their status fell after Trump was elected.  If we break it down by the party that people voted for (non-Hispanic whites only):

From 1988 through 2012, the means for Republican voters gradually rose and the means for Democrats fell, which is to be expected given general ideological polarization over the period.  The elections of 2008 and 2012 don't stand out as unusual.  Then in 2016, the mean fell a little among Republican voters, and a lot among Democratic voters.  The first change is consistent with the idea that people who would otherwise have felt threatened were reassured because they had a president who would look after them, but the second is puzzling from that point of view.  You would have to say that Democratic voters were the ones who were secretly yearning for a protector, and felt most reassured by Trump's win. 

So I stand by my earlier thought that this index doesn't measure resentment.  My next post will offer some thoughts on what it does measure.   

Sunday, June 3, 2018

Explaining too much

In the last few days, I have seen several stories saying that "a new study ... found that opposition to welfare ... has grown among white Americans."  The study (by Rachel Wetts and Robb Willer) didn't actually show any figures on opposition to welfare--it focused on changes in the gap between white and non-white opinions--but I can see how journalists would have that interpretation.  The abstract says:  "we find that whites’ racial resentment increased beginning in 2008, the year of Barack Obama’s successful presidential candidacy and a major economic downturn, the latter a factor previously shown to amplify racial threat effects. . . . These findings suggest that whites’ perceptions that minorities’ standing is rising can produce periods of 'welfare backlash' in which adoption of policies restricting or curtailing welfare programs is more likely."

Has opposition to welfare grown?  Here are the means for whites and blacks in a question from the General Social Survey about whether we are spending too much [3], too little [1], or about the right amount [2] on welfare.

White opinion moved in the "too much" direction after Obama's election, but so did black opinion.  Looking over a longer period, white opinion has moved up and down, and is more favorable to welfare spending than in the 1970s and most of the 1990s.  Black opinion has also gone up and down, but there seems to be a gradual shift towards "too much."

If you look more closely, it seems that the ups and downs are related to the party of the president:  people tend to say "too much" when a Democrat is president, "too little" when a Republican is.  Presumably this is because, rightly or wrongly, they perceive the government as doing more when a Democrat is in office.  This is a well-known pattern that has been documented in research on a variety of issues.  So what mattered was Obama's party, not his race (see this post for another example).

If you regress average opinions on party in power and a time trend, you get the following predicted values:

There is a clear trend towards less opposition among whites and more opposition among blacks, so the racial gap in opinions is gradually declining.  The effects of party control are almost the same among blacks and whites.  I didn't investigate systematically, but it doesn't appear that general economic conditions have any effect among either blacks or whites.

The paper proposed that whites would also regard the rising share of non-whites in the population as a threat.  Since this changes gradually from year to year, that would lead to gradually rising opposition to welfare among whites.  The actual trend is in the opposite direction from the predicted one.

The hypothesis that underlies the paper is that when "relative advantage in the racial status hierarchy" is threatened, whites turn against programs that are seen as helping minorities.  I think the general hypothesis is probably right--the other part of their paper provides pretty convincing experimental evidence that whites express more negative views on welfare when they are made to think about the prospect of America as a "majority minority" nation.  But although perceived threat may help to explain differences among people at a given point in time (e. g., between different places), when looking at historical change it is overwhelmed by the effect of a general decline of racial prejudice. 

Technical note:
1.  Wetts and Willer use a question on welfare spending from the American National Election Studies.  I used a very similar question from the General Social Survey, mostly because the question has been asked for a longer period of time and the GSS has a convenient cumulative file.
2.  My "white" category includes people who also report that they are Hispanic.  That probably accounts for some of the trend among whites.  If I were writing a paper, I would distinguish between Hispanic and non-Hispanic whites and maybe add some other controls, but since I am writing a blog post I will just say that I don't think it accounts for much of the trend.