Who believes in meritocracy?

This is a return to one of my regular topics:  the claim that successful people feel more disdain (or at least less concern with) ordinary people than they used to.  Michael Sandel provided an example in the NY Times a couple of weeks ago:  

"Meritocracies also produce morally unattractive attitudes among those who make it to the top.   The more we believe that our success is our own doing, the less likely we are to feel indebted to, and therefore obligated to, our fellow citizens....

          These attitudes accompanied the market-driven globalization of the last 40 years......

Meritocratic hubris and the resentment it provokes are at the heart of the populist backlash against elites. They are also potent sources of social and political polarization. One of the deepest political divides in politics today is between those with and those without a four-year college degree."

A few weeks ago, I wrote about a GSS question:  " Some people say that people get ahead by their own hard work; others say that lucky breaks or help from other people are more important. Which do you think is most important?"  I noted that in recent years, college graduates seemed to be shifting towards "lucky breaks" while people who didn't have a college degree were shifting towards "hard work."  I said that this seemed to start in 2008.  However, there's a good deal of uncertainty about the estimates for individual years, so it's hard to be sure about the timing.  After I wrote the post, it occurred to me that it might be more informative to look at cohort differences:  it seems that opinions on this question might be established early in life and not too sensitive to current events.  After some experimentation, I distinguished three cohorts:  born before 1960; born 1960-78; and born in 1980 and after.  I limited it to people aged 25 and over, and omitted blacks, since the views of different cohorts might be shaped by different experiences with racial discrimination.  The results for non-graduates:

              work      both          luck

oldest         67%       21%          12%

middle         72%       18%          10%

youngest       76%       14%          10%

A small movement towards "hard work" from the first two the second cohort, which continued from the second to the third.  Now the results for graduates:

              work      both          luck

oldest         62%       27%          11%

middle         66%       25%           9%

youngest       58%       27%          15%

From the first to the second cohort, there was a small movement towards "hard work," as there was among non-graduates.  But from the second to the third, there was a movement towards "lucky breaks or help from other people"--that is, away from "meritocratic hubris."  In all cohorts, college graduates were less likely to believe in "meritocracy" but the gap is biggest in the youngest cohort.   

Of course, everyone in the youngest cohort is young (25-38), and maybe educated people move towards a belief in meritocracy as they age.  I don't have time to investigate that now--maybe I'll look at it later.  For now, I'll just say it's one more piece of evidence that supports something I said a few years ago (I called it "speculation" then):   that American society has become (and is continuing to become) more socially egalitarian. 

Kicks in a plane

In 2016, a paper came out saying that the presence of a first-class cabin led to more incidents of "air rage" and that front boarding--where the economy passengers had to endure the humiliation of walking through first class before getting to their seats--led to even more.  It got a good deal of media attention at the time, and has 56 citations in Google Scholar, which is pretty good for for a 2016 publication.  I was reminded of it when looking at an old post on Andrew Gelman's blog yesterday.  He mentioned that he, Marcus Crede, and Carol Nickerson had a letter in the journal that published the article (PNAS), and that the authors, Karen DeCelles and Michael Norton, had a reply.  After reading those, I think I see where the analysis went wrong.

The dependent variable was whether or not an "air rage" incident happened on the flight.  Two important influences on the chance of an incident are the number of passengers and how long the flight was (their data apparently don't include the number of passengers or duration of the flight, but they  do include number of seats and the distance of the flight).  As a starting point, let's suppose that every passenger has a given chance of causing an incident for every mile he or she  flies.  Then the chance of an incident on a particular flight is approximately:

p=knd

p is the probability of an incident, k is the chance per passenger-mile, n is the number of passengers, and d is the distance.  It's approximate because some incidents might be the second, third, etc. on a flight, but the approximation is good when the probabilities are small, which they are (a rate of about 2 incidents per thousand flights).  When you take logarithms, you get

log(p)=log(k) + log(n) + log(d)

DeCelles and Norton used logit models--that is, log(p)/log(1-p) was a linear function of some predictors.   (When p is small, the logit is approximately log(p)).  So while they included the number of seats and distance as predictors, it would have been more reasonable to include the logarithms of those variables.  What if the true relationship is the one I've given above, but you fit a logit using the number of seats as a  predictor?  

  
That is, there are systematic discrepancies between the predicted and actual values.  That's relevant to the estimates of the other predictors.  E. g., suppose that small planes don't have first class, large planes have first class and boarding in the middle, and medium size planes have first class and front boarding.  Then a model that adds variables for those qualities will find that first class with front boarding has higher rates than expected given the number of seats, which is exactly what DeCelles and Norton appeared to find. 

The authors didn't release their data, so I don't know the shape of the actual relationship. But I would be willing to bet (and give long odds) that a model using logs would fit better and would produce substantially different estimates for their variables of interest.  The general point is that when a control variable is a strong predictor, it's not enough to include it--you have to include it in the right form.  Fortunately, this usually isn't hard to do--in addition to trying x as a control variable, try log(x) too, especially if you're using a logit, Poisson regression, or other model for binary or count data.

PS:  This is the same problem that led to spurious result in the hurricane names study.

A day to remember?

During every presidential campaign, candidates make "gaffes" which occupy the attention of the media for a few days.  Most of them are soon forgotten, but Hillary Clinton's "basket of deplorables" has had unusual staying power.  People not only remember it, but often say that it had a big impact on the campaign.  A few days ago, Thomas Edsall had a column in which he quoted Charles Murray, in a 2017 podcast interview:  "the ‘deplorables’ comment by Hillary Clinton may have changed the history of the world....all by itself, [it] might have swung enough votes" to elect Trump. Edsall didn't present this as a controversial claim, but as illustrating the "animosity to elite liberalism that Trump has feasted on." 

Clinton made that remark on Sept 9 (a Friday), and it made the news almost immediately.  According to  the polling averages in 538.com, Clinton led Trump by 42.1% to 38.9%  on Sept. 8.  Her lead had been narrowing since the middle of August--on September 1,  it was 42.5 to 38.4.  What happened after the "basket of deplorables"?   On September 15, her lead was 41.6% to 39.9%.  Her lead continued to narrow, reaching a low of 41.7% to 40.6% on September 20th, before widening again. So nothing dramatic happened--she lost a little ground after the remark, but she'd been losing ground before the remark, and her lead recovered later.  If you looked at the figure without knowing anything about the campaign, you wouldn't identify September 9th as notable in any way. 

The 538 figures are a smoothed and weighted average of many polls--Real Clear Politics gives unadjusted figures.  They also show Clinton's lead as narrowing after 9/9, continuing a trend that had started in August.  Those figures show a slight increase in her vote share after 9/9--Trump increased faster, and support for third party candidates declined. 

Of course, the election was close and you can't rule out the possibility that it changed enough votes to alter the outcome, but there's no evidence that her remark had much influence on voters.  Its staying power is partly because of her peculiar choice of words, and partly because it fits with an idea that's popular among both conservative and liberal elites:  that elites show more disdain for ordinary people than they used to.  I've had several posts suggesting that this idea is not true, and I'll have another one soon. 

Deaths of despair?

In my last couple of posts, I've looked at how people with and without college degrees feel about their lives.  People without college degrees have become relatively less happy with their financial situation, but there's no evidence of a general sense of despair.  But what about the rise of "deaths of despair," which appears to have been mostly among people without college degrees?  I found that the Social Capital Project of Senator Mike Lee has compiled data on historical changes in what are called "deaths of despair" (suicide, alcohol, and drugs), which I use here.   The data go back to 1900, but I just use the years since 1948, so that I don't have to consider the effects of Prohibition, the Depression, and World War II.  Suicide and alcohol-related deaths:

They follow a similar course:  rising until about 1970, then declining until the late 1990s, then rising again. 
 
Now add drug related deaths:

 They follow a different trend:  basically just upward.  For the first part of the period, they were substantially less common than suicide or alcohol related deaths.  They passed alcohol in about 2000, and suicide in 2015--by 2017 (the last year in the data) they were about 50% more common than suicide and twice as common as alcohol-related deaths.

    So drug deaths don't follow the same course as the others.  If you look at them on a logarithmic scale, it's almost a straight line--that is, a steady rate of increase--and if you regress the log of drug deaths on a time trend, adding suicides and alcohol-related deaths doesn't significantly improve the fit.  That is, the three kinds of death don't act as if they are all indicators of the same thing.  

The Social Capital Project report also looks at survey indicators of unhappiness and finds some evidence of a rise since 2000, but it's not very clear. It concludes that drug deaths are different: "apart from the question of whether or why despair may be on the rise, we clearly remain within the grip of a national opioid crisis that requires the attention of policymakers."  As far as why drug deaths would be increasing so steadily, the most plausible answer is technological "progress":  people keep thinking of ways to provide them in more convenient and powerful (and therefore dangerous) forms. 

Lives of despair?, part 2

This continues my last post, where I was looking at changes in answers to general questions about life and expectations for the future.  The basic question is whether the views of people without college degrees (which journalists like to call "the working class") have become more negative relative to the views of people with degrees.

Continuing where I left off, views on whether most people can be trusted vs.. "can't be too careful."

Declining among both, but somewhat faster among people without a college degree.  The GSS also has questions about whether people try to be helpful and treat others fairly, and they show similar trends. 

Happiness of your marriage (for married people)

Basically parallel--maybe some divergence in the 1990s and convergence since then.   For people without a college degree, there has been little change since 2000, maybe a slight increase.

Finally, overall happiness:  a clear divergence, from essentially no difference in the early 1970s.  Over the last 20 or so years, it's stayed about the same or declined slightly among people with college degrees, and declined pretty steadily among those without.  There is also a question about whether you find life exciting, pretty routine, or dull, and that doesn't show a trend for either educational group--people with college degrees are consistently more positive.

What can we conclude from all this?  As often happens, things are complicated--there are some questions which less educated people are becoming relatively more negative, some on which there's little change, and one on which less educated people are becoming more positive.  The strongest cases of less educated people becoming more negative are those which either involve current economic circumstances or are strongly influenced by current economic circumstances (overall happiness).  For things involving expectations for the future, there's no clear class divergence.  There's only one question directly involving relations with other people you know (happiness of marriage), and for that there's no sign of divergence.  However, there is divergence in assessment people in general.  And then there's the divergence in the "wrong" direction for hard work versus lucky breaks or help from others.  Overall, rather than generalized despair among the "working class," it seems like there's been a turning inward--feeling that you can count on yourself, your family, and your friends, but not on people in general. 

What about the rise in "deaths of despair"?  I will write about that sometime in the future, but basically my view is that it's a catchy but inaccurate label:  by and large, those deaths aren't direct expressions of despair.  

Lives of despair?

In the last few years, there's been a lot of discussion of the rise of "deaths of despair" (drugs, alcohol, and suicide) among white people without a college degree.  Often it is assumed that they are an indicator of the feelings of the group as a whole--"deaths of despair" are an extreme manifestation of a widespread malaise among the "white working class."  But I haven't seen any systematic attempt to look at this issue, so I turned to the General Social Survey, which contains a number of questions on general feelings about people and life.  I have coded each question so that higher numbers represent more positive or optimistic feelings.  There were sixteen; I show figures for seven of them, and briefly mention the trends for the others.

Satisfaction with your financial situation:

 

People with a college degree have become a little more satisfied, those without less satisfied, so the gap between them has increased.  That is reasonable given the increasing gap between the earnings of the two groups.  Assessment of how your income ranked relative to others showed a similar pattern.

Change in your financial situation over the last few years:

Pretty much parallel.  Looking more closely, the gap between people with and without college degrees seemed to grow a bit in the 1970s and 1980s and then decline a bit.

For job satisfaction, beliefs about the chance of losing your job or how hard it would be to find a comparable job if you lost yours, there was not much change.  

Beliefs about whether your children will be better off than you are:

This one only goes back to 1994, but there may be some tendency for people without a college degree to become more optimistic relative to people with a college degree.  There was a question about how your standard of living compared to your parents', and the trends were similar.

Then there was a question on whether people get ahead by their own hard work or lucky breaks and help from others.  I counted hard work as the optimistic answer. 

A clear divergence after about 2008:  people without college degrees becoming more positive and people with degrees becoming more negative.  Well, this is unexpected. 

This post is getting pretty long, so I will stop here and finish it in a day or two.