Learn 2 R

I've been using R regularly, for W-NOMINATE and for other things. It's a free, open source programming language created to perform statistical analyses, and it's very popular in some academic circles.

I never got a good introduction to it, though, so I'm excited to see that Coursera is offering a course, Computing for Data Analysis, which is all about using R. I will definitely be taking this, and I recommend it to anyone else who does statistical analyses on a regular basis.

Data is online

I procrastinated a bit, but the W-NOMINATE data for the NH House from 1999 - 2012 is now online at Google Drive, if anyone wants to look through it. I just uploaded the entire file straight from my computer, so it also contains helpful notes that I kept for myself, the programs I wrote to make the process more manageable, and some other things.

Next on the agenda-

1) DW-NOMINATE, a program that allows direct comparisons between legislative sessions. I finally found the program online, but I'm having trouble using it. I'll have to email Keith Poole about it.

2) Party influence comparisons.
Seth Masket's fantastic book No Middle Ground, which inspired my NOMINATE project, includes graphs comparing legislators' ideological positions with the liberalness (or conservative-ness) of their districts. The idea is to see who legislators represent -- their party or their district?

I made a similar graph in 2011 using New Hampshire LDI's, though I didn't post it at the time.

You would expect to see a straight diagonal line if legislators were perfectly representing their districts. If they were perfectly representing their parties instead, you would see two horizontal lines, one for each party. In this graph, it looks like they might be representing their districts more than their parties, though it's hard to tell. They don't seem to be representing anyone very accurately.

It would be nice to compare this through the years, but it'll take some work to get the data.

3) Senate NOMINATE scores. I have the data, so I will get the older state senate W-NOMINATE scores sooner or later.

Will filibuster reform help working families?

Liz Iacobucci at The New Hampshire Labor News argues that it will:

Progress Massachusetts looked closely at Scott Brown’s voting record and came up with 40 bills that would have passed the Senate – if they hadn’t been killed by a Republican filibuster.  The list includes: 

• Restoring American Financial Stability Act of 2010 (the original financial regulatory reform bill);
• Creating American Jobs and Ending Offshoring Act;
• Emergency Senior Citizens Relief Act of 2010;
• American Jobs Act of 2011;
• Rebuild America Jobs Act;
• Middle Class Tax Cut Act of 2011; and
• The Buffett Rule (a 30% effective tax rate on income exceeding $1 million).

All of those bills would have passed the Senate – if they had ever gotten to the floor for a vote.

If you've ever studied game theory, though, that conclusion seems a lot less persuasive. Republicans are almost certain to change their legislative strategy in response to the new rules. The outcome in this scenario is impossible to predict without more information.

Sarah Binder at The Monkey Cage puts it this way:

It’s tough to turn on a new rule and  calculate the effects that are likely to follow because it’s hard to know how senators will react.  A new rules regime—particularly one curtailing the right of extended debate under Rule 22—could encourage senators to aggressively avail themselves of every procedural avenue in the Senate rule book for obstructing the Senate.  For instance, the minority could become less likely to agree to invoke cloture on the motion to proceed, preventing the majority from calling up bills high on its agenda.  Or senators could become more aggressive in the demands they make on a leader as a condition for signing onto consent agreements.   Both scenarios suggest that filibuster reforms could bring  unintended consequences.

If labor sympathizers want to help working families, they'll have to do it the old-fashioned way – by swaying public opinion and helping to elect more liberal politicians.

14 years of the New Hampshire House (in W-NOMINATE)

This is based on data taken from the General Court website, which only provides the most recent party affiliation of each legislator, so some legislators in earlier years may have incorrect party labels.

Watch as the middle slowly erodes. Independent and hybrid representatives are disappearing, too, another sign that party activists are gradually exerting more control over elections.

The moderates of 1999

I'm working through older voting data right now, from 1999 to 2010. Over the past week I've been laboriously scraping that data from the General Court website.

Right off the bat, the results are striking. This is the House of Representatives in 1999:

There are so many moderates! There's supposed to be a gaping void in the middle!

So the hypothesis that the House is growing more polarized? Mark that one down as "confirmed".

Stark choices in NH Congressional races

I haven't been following NH Congressional races very closely. And, even if I wanted more information, I'm not sure where I could get it. At least in NH2, both campaigns are full of utterly useless information: "[Ann Kuster says] 'Eff him!'" and "Charlie Bass voted for Congressional pay raises 3 times." (Oh, the horror!)

So I was happy to see that Boris Schor at the University of Chicago has put together a measure-- similar to NOMINATE-- of U.S. Congressional candidates' ideologies.

Unlike the scores I usually work with, these range from -3 (very liberal) to +3 (very conservative).

With the exception of Charlie Bass, all NH candidates are more extreme than the average of their parties. Bass scores a surprising (to me) +0.2, putting him near the center. Here are the scores:

Candidate Scores

NH1	D	Carol Shea-Porter	−1.33
NH1	R	Frank Guinta	1.64
NH2	D	Ann Kuster	−1.73
NH2	R	Charles Bass	0.2

Now this is information we can actually use!

Visit Boris' blogs here, here, and here for more details.

New NOMINATE data!

It's been a long time since I've updated the NOMINATE data stored on this site. In fact, I've never updated it, because the memory of copying responses to roll call votes hundreds of times and pasting them into a spreadsheet turned me off to the idea.

Fortunately, I've been learning some programming (in Ruby and in Python), so I wrote a program to build the voting list for me instead. It's much pleasanter this way! I've also tried to improve the graphs by getting to know R better (as you can tell from above). Now the whole process takes only 5 or 10 minutes.

Within the next week I expect to release the Ruby code for my vote sorting program (I already posted it at Google Drive, but I want to clean it up a bit), along with simple instructions for getting W-NOMINATE data yourself. And hopefully I will be able to get data from earlier years, to add another dimension of fun to this.

Can't wait for the next legislative session to start! It will be much more interesting, examining this data in real time! It'll be possible to use this data to try to predict the outcomes of House and Senate votes in advance, among other things.

--The new data, including more graphs of the House and Senate and a list of legislators, is over here.

PS: William O'Brien didn't have a score before, because he was practically always "Presiding" instead of voting. He now has enough votes to get a score -- (0.72, -0.69).

Don't OD on polls

I've gotten a chilly reception from poll enthusiast Steve Vaillancourt both times I've mentioned Intrade in comments on his blog. He'll probably like me less after I post this blog by John Sides at The Monkey Cage.

The Constant Attention to Individual Polls Is Hurting America

A short excerpt:

Poll results vary for random reasons—that is, because of sampling error. ... we don’t actually know the true proportions of Obama and Romney supporters in the public.  There is no way to know definitively which poll is “the truth.” ... not only don’t we know Obama’s or Romney’s true share of the vote at this moment, but we don’t know with much confidence how much any pollster might be systematically overestimating or underestimating Obama’s or Romney’s share of the vote.  For that reason, if you just want to know where the horserace stands, look at the average of the polls and ignore the individual polls.

The logic of third party politics

My friend Matthew Richards has an essay posted at Progressive Action NH criticizing nearly every conceivable voting strategy this November. I agree with him about a lot of things, but mostly I'm using our disagreements as a springboard for defending third party politics.

In regards to Jill Stein, Matthew summarizes the classic argument against third parties:

Why not vote for Jill? We’ve all heard it. Jill knows it herself. She’s not going to win.  She’s somehow “taking votes away” from Obama.

He continues:

But is that a reason for not voting for her?  I think it’s a terrible reason not to vote for her, because people not following their consciences is part of what has gotten the world into the huge mess it’s in right now.

I agree that it's a terrible reason, but I don't trust my conscience as much as Matt does, so I rely on a more strategic argument.

In a sense, votes for Jill Stein do take votes away from Obama, and they could throw the race to Mitt Romney. And practically everyone voting for Jill Stein would prefer Obama to Romney. Combine this with the impossibility of Jill Stein winning (the prediction market Intrade estimates less than a 0.1% chance, as of Oct. 7th), and voting for Obama appears to be the logical choice. Only a masochist or an ignoramus would vote for Stein.

But if we take a longer view, the strategy can flip.

Winning the war on war

I have been reading Joshua Goldstein's fantastic book Winning the War on War: The Decline of Armed Conflict Worldwide, which summarizes the state of warfare today and suggests strategies for eliminating it.

The book begins by demonstrating that war has been declining by nearly every measure since World War II. The case is capably made with a cautious use of statistics, at times in grisly detail. Media coverage notwithstanding, no conflict has risen to the level of WWII in recent history. Since the collapse of the U.S.S.R., warfare has declined from the levels of the proxy wars between the U.S. and the Soviet Union in Vietnam, Korea, Afghanistan, etc. There is no direct war between nations in the world today, and various civil wars are ending faster than they are starting. (In May, he updated the argument on his blog, noting that "Warfare is literally shrinking across the face of the earth.")

Goldstein focuses primarily on UN peacekeeping efforts as the force behind this decline. Peacekeeping forces — though often underfunded and disorganized — can act as impartial observers to help enforce peace agreements, or to oversee elections. In some cases, a conflict will continue even when both sides would gain from its end, simply because neither side can guarantee its safety if the other decides to renege on a peace agreement. UN forces can provide that guarantee and circumvent this frustrating impasse.

To buttress his arguments, Goldstein cites a study by political scientist Page Fortna:

[According to Fortna] The presence of peacekeepers reduced the risk of renewed war by somewhere between 55 and 85 percent. "The statistical evidence is overwhelming. In short, any way you slice the data, peacekeeping works." Quantitative studies by other researchers "have reached a consensus, and it is an optimistic one." Although these studies find that peacekeeping is less effective in reaching a cease-fire in the first place than in making it last once reached, the success of the latter "has emerged as a strongly robust result in the quantitative literature." Thus, the answer to whether peacekeeping works is "a resounding yes."

This research will create problems for libertarian critics of the UN. How does a person oppose state aggression while simultaneously denouncing an institution which has been remarkably successful at ending warfare— the most direct and brutal form of state aggression? Given the information in this book, libertarians should be much more supportive of UN peacekeeping. (But don't hold your breath.)

In the second half, Goldstein's history of American peace movements, and his analysis of their weaknesses, makes his book an indispensable resource for serious peace activists. He argues that, while it may be an institutional necessity, the tendency of organizations to focus on peace and justice (or freedom, etc.) undermines their activism. The two are inseparable, goes the peace-and-justice reasoning, and this allows groups to survive on justice issues when peace is less popular. In fact, research suggests that the two issues are largely unrelated, and peace organizations, to perpetuate the myth, must promote solutions that have little real-life impact.

Modern peace groups, for this and other reasons, fail to provide relevant information about UN efforts to American voters. If Winning the War on War can correct the oversight, it will do much to promote peace in the world.

Experimental politics

Slate has a nice article about how the Obama campaign (and some Republican campaigns) is using experiments to determine the effectiveness of particular approaches. I approve.

If these forays seem random, it’s because at least some of them almost certainly are. To those familiar with the campaign’s operations, such irregular efforts at paid communication are indicators of an experimental revolution underway at Obama’s Chicago headquarters. They reflect a commitment to using randomized trials, the result of a flowering partnership between Obama’s team and the Analyst Institute, a secret society of Democratic researchers committed to the practice, according to several people with knowledge of the arrangement.

I'm still hoping for experiments to catch on in policy-making circles.

The central location of NH's capital helps prevent corruption

According to new research from Filipe R. Campante and Quoc-Anh Do:

We show that isolated capital cities are robustly associated with greater levels of corruption across US states. In particular, this is the case when we use the variation induced by the exogenous location of a state’s centroid to instrument for the concentration of population around the capital city. We then show that different mechanisms for holding state politicians accountable are also affected by the spatial distribution of population: newspapers provide greater coverage of state politics when their audiences are more concentrated around the capital, and voter turnout in state elections is greater in places that are closer to the capital. Consistent with lower accountability, there is also evidence that there is more money in state-level political campaigns in those states with isolated capitals. We find that the role of media accountability helps explain the connection between isolated capitals and corruption. In addition, we provide some evidence that this pattern is also associated with lower levels of public good spending and outcomes.

Here are some cool graphs:
(Corruption is on the y-axis, and a measure of the centrality of the capital is on the x-axis. Click to enlarge.)

The paper is here.

Hat tip to John Sides and Seth Masket.

Deliberative defense budgeting

Most voters don't know very much about politics. This has been a consistent finding in public opinion polling for decades, a finding that some commenters find vexing. But the finding doesn't surprise economists, because economists know that people respond to incentives, and there's little incentive to stay informed. The chance of a single vote affecting an election is vanishingly small, so why bother? In this situation, rather than research the issues, most people will prefer to spend time with friends and family, do homework, work overtime, or watch a movie. Economists call this rational ignorance.

Some political scientists have argued that voter ignorance doesn't matter, because voters can use shortcuts to reach reasonably accurate decisions. For example, they may vote along party lines, or they may mimic the views of a favorite politician or talk show host. In reality, sometimes this seems to work, and sometimes it doesn't.

Looking for a better system, another group of political scientists have been experimenting with alternative forms of democracy, in particular deliberative democracy. In a deliberative democracy, voters are provided with relevant information to help them deliberate before voting. Researchers use a variety of approaches. Sometimes they have participants deliberate together, among themselves. Other times participants decide on their own. They select participants and the information to present in many different ways. (My favorite variant is the policy jury.)

Last Thursday, the Program for Public Consultation released the results of a deliberative democratic poll focusing on the defense budget. The participants, 665 Americans, were selected using random telephone numbers and street addresses. They were presented with arguments over the internet for and against cutting defense spending. The arguments were vetted by experts at non-partisan, non-profit organizations. Participants without internet access were provided with laptops and internet. Throughout the process, participants answered questions about their opinions.

Their responses suggest that the shortcut theory fails in the case of defense budgeting. Old-fashioned polls repeatedly find that large majorities of Americans oppose cuts to defense spending. After being presented with arguments from both sides, however, participants in this deliberative poll—Democrat, Republican, and independent alike—overwhelmingly supported significant spending cuts. Among Republicans, about two thirds supported spending cuts, with an average preferred spending level 15% lower than the 2012 budget. Nine out of ten Democrats supported spending cuts, with an average 28% cut, and about two thirds of independents called for cuts, with an average 26% reduction.

Unfortunately, since politicians work within a system of old-fashioned, non-deliberative democracy, the results of this poll are unlikely to provoke policy changes. In order to keep their jobs, politicians must satisfy rationally ignorant voters, even if that leads to poor outcomes. But the first step to recovery is admitting that we have a problem.

You can read the official release here.

HT: The Economist.

Free Staters are taking over New Hampshire. Or not.

Diane Lacy, President of the NH State Employee Association (via Skip Murphy):

let’s be clear; the speaker [of the NH House, Bill O'Brien] is not exactly our typical Republican. He is a Free Stater. These people, these sponsors that are going out there with the legislation are Free Staters. They are not the typical Republicans.

...

We’ll have a conversation with our neighbors, all of our communities, about the values that are important to NH because one thing is clears: NH is being taken over by the Free Staters and we are not going to stand for it.

I don't know what Diane Lacy is smoking.

A Free Stater is a person who has moved to New Hampshire as part of the Free State Project. The Free State Project is "an effort to recruit 20,000 liberty-loving people to move to New Hampshire. We are looking for neighborly, productive, tolerant folks from all walks of life, of all ages, creeds, and colors who agree to the political philosophy expressed in our Statement of Intent, that government exists at most to protect people's rights, and should neither provide for people nor punish them for activities that interfere with no one else." (According to the website.)

Bill O'Brien is not a Free Stater. And it's pretty easy to figure this out, since he's been living in New Hampshire since before the Free State Project was founded in 2001.

(There is a small group of people who lived in New Hampshire at that time, who are technically Free Staters. They did this by signing the statement of intent before New Hampshire was chosen in 2003. Bill O'Brien did not do this. And most people who did this are Libertarian Party activists and hardcore libertarians, which Bill O'Brien is not.)

If Lacy had stated that Bill O'Brien was a Tea Partier, on the other hand, that would be plausible.

And what about sponsoring legislation and taking over New Hampshire? I've seen persuasive arguments that Free Stater state representatives are having some influence on NH politics. They certainly seem to sponsor plenty of legislation. But with only 12 or 15 representatives, out of 400, that influence is very limited.

Not only that, but many Free Staters actually oppose the right-to-work legislation Lacy is concerned about. Here's the 2004 platform of the Libertarian Party on this issue:

The Principle: We support the right of free persons to voluntarily establish, associate in, or not associate in, labor unions. An employer should have the right to recognize, or refuse to recognize, a union as the collective bargaining agent of some, or all, of its employees.

Solutions: We oppose government interference in bargaining, such as compulsory arbitration or the imposition of an obligation to bargain. Therefore, we urge repeal of the National Labor Relations Act, and all state right-to-work laws which prohibit employers from making voluntary contracts with unions. We oppose all government back-to-work orders as the imposition of a form of forced labor. [Bold added.]

If Diane Lacy wants to return to reality, I recommend this article about changing Republican politics by Betsy Russell in the Idahoan Spokesman-Review. It draws on the views of four political science professors.

Can you guess what they say?
Hint: it starts with "T", and ends with "Party".

Fallacy of composition, W-NOMINATE version

I've been neglecting my blog lately. Much of the reason is that I've been working to understand the political science program NOMINATE, so that I can gather information about the ideological make-up of the New Hampshire legislature.

The data is posted on the W-NOMINATE page of my blog. Here's the cool graph (click to enlarge):

Keith Poole and Howard Rosenthal, the main developers of the program, claim that the program measures economic attitudes on the x-axis, and, in recent years, social attitudes on the y-axis.

At first blush this is what it appeared to be doing for New Hampshire legislators. However, after taking a closer look at the votes, I discovered that it's not doing this. At least, not very well.

The x-axis measures economic issues well, and correlates beautifully with the House Republican Alliance's scorecard ratings. (See this impressive graph, courtesy of Seth Cohn.) The y-axis, on the other hand, has issues. Apparently, because there have been so few votes on social issues this year, there is little basis for the y-values. Mostly they reflect positions on gun rights and criminal justice issues, with pro-gun and tough on crime on the bottom.

This is yet another reminder that politics at the state level does not necessarily imitate politics at the federal level.

Thanks to Rep. Seth Cohn, Scott Pauls at Dartmouth, and professor Jason Sorens for criticism and help. Peer review fixed my mistake.

Why are state legislators so crazy lately?

Betsy Russell at the Idahoan Spokesman-Review offers some informed speculation, drawing on a few high-profile political scientists.

HT: Seth Masket.