Friday, February 26, 2016

Stop the presses!

I was part way through a post on another topic when I heard a story about how Donald Trump was predicted to have a 97% chance of beating Hillary Clinton in November  and a 99% chance of beating Bernie Sanders.  The prediction, by Helmut Norpoth of Stony Brook University, is described in a lot of stories, but this one gives the most details.  He's been developing the model for a number of years, and this paper gives a prediction for the 2012 election.   At that time, the formula was:

y=48.2+.445PI-.138PO+.366y1-.338y2

y is the Democratic share of the two-party vote, while y1 and y2 are the Democratic shares in the previous two elections.  PI is the the share of the vote among the top two finishers in the New Hampshire primary for the candidate of the incumbent party, in this case the Democrats.  PO is the share for the candidate of the opposition (Republican) party.  The values of the primary variables are limited to 35-65--if you get more than 65% it's counted as 65; less than 35% is 35.  So the key idea is that candidates who do well in the NH primaries do better in the general election, which seems reasonable.

Trump is predicted to get 54.7% of the two-party vote against Clinton.  The other potential Republican candidates are predicted to get just under 50%.  Trump would get a 65 on PO, while any other would get 35 (Trump got 35%, and second-place finisher Kasich got 16%, meaning Trump got about 68% of the top two share).  So a 30 unit difference in PO gives about a 5 unit difference in y, which is consistent with the formula above.

But Sanders is predicted to do slightly worse than Clinton, losing narrowly to a non-Trump candidate, even though he won the NH primary easily, with about 61% of the top two vote.  With the formula above, he would be predicted to run almost 10 points better than Clinton, meaning an easy victory against Trump and a historic landslide against any of the others.

It sounds like the current predictions are based on a formula using the first few primaries, where Clinton and Sanders are running about even.  But the paper I linked to above investigated the issue and concluded that the New Hampshire primary was the only one with any predictive power.  So if you accept the Norpoth formula from 2012, you should not believe the Norpoth predictions for 2016.  Personally, I wouldn't put much faith in either one, but that's a subject for another post.


Thursday, February 18, 2016

Who Knew?

In the last few days, Ross Douthat and a news story in the New York Times have talked about a what they present as a surprising discovery--that a lot of Republicans are not "consistent" conservatives.  This is actually one of the main findings of early survey research from the 1940s and 1950s.  There are not many people who have liberal or conservative (or moderate) opinions across the board--most people have a mix of liberal, conservative, and moderate opinions, with a few extreme left or right positions thrown in.  This is true even of party loyalists.    Party loyalty is often based on a sense that one party is smarter, more honest, or cares more about "people like me" rather than an examination of the issues.  And most people don't pay that much attention to politics--they may not even know they have the "wrong" position on some topic, and even if they do they don't worry about it much.

I'll present a few figures from a 2014 Pew survey.  One of the questions asks "what do you think is more important--to protect the right of Americans to own guns, or to control gun ownership."  Another asks "do you think abortion should be legal in all or most cases, or illegal in all or most cases."  Both of these are issues which sharply divide the parties and have been prominent for a long time.  A cross-classification of opinions follows:

                                    Guns
                            C                    L

                    L     23%             28%
Abortion
                    C    33%              16%

That is, 61% (33+28) hold "consistent" positions--liberal or conservative on both.  39% are liberal on one and conservative on the other.  There is some association between the two opinions, so the number of "consistent" liberals and conservatives is higher than you would expect from chance alone. But this is just two different opinions--if you took the top five or six issues, or even "social issues," few people would be liberal on all or conservative on all.

The association between opinions is stronger among more educated people and people who follow politics more closely.  For example, the survey contains a question about how often people vote in congressional primary elections.  Among those who say they "always" do, the percentages are like this:

                               Guns
                            C                    L

                    L     20%             36%
Abortion
                    C    34%              10%

That is, 70% are consistent.  The association would presumably be even higher among people who write about politics, and of course they're the ones whose voices usually get heard.  So the rediscovery that consistency is low among the general public always comes as a surprise.

Beyond that. it's always seemed to me that many people are suspicious of ideology and would be attracted to someone who makes a point of having some unexpected positions.  I hadn't been able to find any questions on that issue, but I gave it another try, and found one from ABC News/Washington Post:  "Do you think the Republican nominee for president should take only conservative positions on issues or do you think it is OK to have the Republican nominee for president take moderate positions on some issues? Do you feel that way strongly or somewhat?"  The question was asked of Republicans in 2011 and 2015:

                                                                 2011                 2015
Only conservative, strongly                      17%                  11%
Only conservative, somewhat                     9%                    8%

Some moderate, somewhat                        35%                  25%
Some moderate, strongly                           32%                  47%

I had expected the balance to be in favor of some moderate positions, but was surprised at how lopsided it was.

[data from the Roper Center for Public Opinion Researc]

Saturday, February 13, 2016

Dying for their party?

A few days ago, Nicholas Kristof wrote:

"This echo chamber [conservative media] deluded its believers to the point that it sometimes apparently killed them. During the 2009-10 flu pandemic, right-wing broadcasters like Limbaugh and Glenn Beck denounced the call for flu shots, apparently seeing it as a nefarious Obama plot. The upshot was that Democrats were 50 percent more likely than Republicans to say that they would get flu shots, according to a peer-reviewed article in The Journal of Health Politics, Policy and Law.  So when the pandemic killed up to 18,000 Americans, they presumably were disproportionately conservatives. "

  He didn't provide a link to the article, but I tracked it down (here it is).  It mentioned that "according to an October 2009 Pew Project survey ....  Democrats were nearly 50 percent more likely than Republicans (60 vs. 41 percent) to indicate that they would take the vaccine."  So the ultimate source of the information was not a "peer reviewed article," but a Pew survey.  That's actually a plus, since unlike most academic researchers, Pew has the resources to obtain a (hopefully) representative national sample for its surveys. The Pew survey apparently didn't ask if people had been vaccinated, but a November 2009 CNN/ORC survey asked "As you know, there are two types of flu today--swine flu, which has not occurred in recent years, and the regular, seasonal flu which occurs every year around this time. Thinking specifically about the vaccine against swine flu, which of the following best describes you personally?...You have been vaccinated against the swine flu. You have taken steps to get vaccinated against swine flu but have been unable to do so. You want to get vaccinated against swine flu but you have not yet taken steps to do so. You have not gotten vaccinated against swine flu and do not plan to do so."

The results:
                                       D              R               I
Vaccinated                     6%            8%           8%
Tried                             18%             9%         13%
Want, no steps           23%           28%        16%
No plans                      51%            54%       60%

There is an interesting difference between the proportion of Democrats and Republicans who said they had tried but were unable to get the vaccine, but there's little difference in the proportion who say that they don't plan to (and it's not statistically significant).  If you look at self-rated ideology, the percentages with no plans to get vaccinated are 52, 56, and 58 for liberals, moderates, and conservatives (and there's no difference in the percentages who said they tried but couldn't).

An ABC/Washington Post survey, also from November 2009, asked "Do you plan to get the swine flu vaccine this year, or do you think you probably will not get the swine flu vaccine?" 8% of Democrats, 6% of Republicans, and 6% of independents said that they already had.  On the other side, 59% of Democrats, 67% of Republicans, and 70% of independents said they probably wouldn't.  By political ideology, 7% of liberals, 8% of moderates, and 6% of conservatives said that they'd had the vaccine; 61%, 61%, and 73% said probably not.   

 So there was probably some difference between Democrats and Republicans intentions to get the vaccine, but it wasn't as large as the Pew survey suggested (which doesn't mean there was anything wrong with the survey, just that there's sampling variation).  Also, independents may have been the most likely to say that they wouldn't.  It's not clear if there was any difference in behavior--there might have been, but the numbers are too small to be sure. 

 There were big partisan differences on some issues related to swine flu.  For example, on "How much do you blame each of the following for problems with getting the swine flu vaccine to all those who would like to get it--a great deal, a moderate amount, not much, or not at all? How about...the federal government?" 29% of Republicans and only 14% of Democrats blamed the federal government a great deal, while 26% of Republicans and 43% of Democrats said "not much" or "none at all."  Republicans were also more likely to say that the threat had been exaggerated.  But there were only small partisan differences in opinions about whether the vaccine was safe (and independents were the least likely to think so).

The reason he was writing about this was that he was proposing that Trump was sort of a Frankenstein's monster produced by the conservative media.  The problem with that analysis is that Trump's major source of support is not hard-core conservatives--they have plenty of other options (everyone else except maybe Kasich).   His appeal seems to be his positions on immigration and protection for American industries and his outsider status.  As a result, his supporters probably have more chance of being persuaded to vote Democratic than Cruz's or Rubio's.  

[data from the Roper Center for Public Opinion Research]

Monday, February 8, 2016

Change and Decay

Marco Rubio got a lot of attention for repeating a talking point during Saturday's debate, but few people seem to have noticed how feeble the talking point was (although Ross Douthat acknowledged that it was 'not all that great to begin with').  The obvious flaws are that Barack Obama is not going to be a candidate in November and that all of the other Republican candidates have been critical of him too.  But it did make me wonder how Americans feel about "change" in general--there's a long tradition holding that we're attracted to change and novelty.  That reminded me that the World Values Survey asked people to place their own views on a ten-point scale running from "ideas that have stood the test of time are generally best" to "new ideas are generally better than old ones."  I wondered how the United States ranked on that question.  I took the 1995-6 wave, which was the last time it was asked in a wide range of nations.  The ranking of nations, with high numbers indicating more support for new ideas:

Colombia        8.21
Bangladesh 8.09


Dominican Rep. 6.19
Macedonia 6.12
El Salvador    6.05
Bosnia         6.01

Puerto Rico 5.95
Peru           5.92
Chile          5.86
Mexico         5.76
Spain          5.59
Norway         5.58
Sweden         5.54
Germany        5.51
Uruguay        5.47
Philippines 5.46
Venezuela 5.43
Croatia        5.42
South Africa 5.38
Romania        5.32
Bulgaria 5.31
Turkey         5.29
Argentina 5.23
New Zealand 5.20
China          5.16
Albania        5.12
Australia 5.08
South Korea 5.07
Poland         5.03
Finland        5.01

United States 4.97
Armenia        4.94
Estonia        4.92
Switzerland 4.84
Slovenia 4.80
Latvia         4.79
Moldova        4.78
Nigeria        4.72
Slovakia 4.62
Serbia         4.61
Lithuania 4.56
Czech Rep. 4.52
Hungary        4.43
Belarus        4.37
Georgia        4.37
India          4.22
Ukraine        4.21
Japan          4.16
Taiwan         4.15
Russia         4.11
Montenegro 4.11
Azerbaijan 3.77

The United States is in the middle, although a bit more towards the traditionalist side.  Beyond that, it seems like poorer nations tend to be more favorable towards new ideas, while nations of the former Soviet Union and eastern Europe give more support to tradition.  But nothing is very clear, which might be why the WVS dropped the question.  

I also ranked the nations by standard deviation:

Turkey          3.53
El Salvador 3.41
India           3.36
China           3.33
Nigeria       3.29
Venezuela 3.25
Moldova         3.09
South Africa 3.08
Peru         3.06
Dominican Rep. 3.04
Azerbaijan 3.00

Serbia         2.92
Puerto Rico 2.91
Montenegro 2.90
Macedonia 2.88
Georgia         2.88
Romania         2.80
Lithuania 2.74
Ukraine         2.73
Philippines 2.72
Bulgaria 2.70
Belarus         2.70
Bangladesh 2.67
Argentina 2.66
Mexico         2.65
Russia         2.60
Uruguay         2.59
Armenia         2.56
Chile         2.54
Slovakia 2.52
Latvia         2.50

Bosnia         2.49
Poland         2.47
Slovenia 2.47
Albania         2.45
Hungary         2.41
Croatia         2.38
Estonia         2.35
Spain         2.34
Switzerland 2.34
South Korea     2.27
Colombia 2.20
Czech Rep. 2.20
United States 2.19
Taiwan         2.15
Finland         2.11
Australia 2.06
New Zealand 2.03

Germany         1.97
Norway         1.86
Japan         1.85
Sweden         1.81


Now there's a very clear pattern:  affluent nations, including the United States, have less variation.  The highest-ranking nations have a bimodal distribution in which the most popular choices are 1 and 10.  In the United States, the most popular choice is 5, followed by 6 and 4.