Less Is More

November 11, 2012
Posted by Jay Livingston
Cross posted at Sociological Images

[Note, Nov. 13:  Thanks to the comment from “maxliving” directing me to the Chen-Rodden paper, I have substantially revised this post.  Hat tip to Max.] 

In a democracy, all votes are created equal - one person, one vote – but apparently some votes are more equal than others.  Obama won the electoral college vote 62% - 38%, though his margin in the popular vote was much smaller: 51% - 48%.


A similar discrepancy happened in the vote for Congressional representatives.  The Republicans control the House of Representatives, where they have 54% of the seats. But if you add up all the votes for those seats, the Democrats come out slightly ahead (by about 500,000 votes).  More votes but fewer seats. 

That discrepancy arises from the distribution of Democrats and Republicans in a state’s Congressional districts.  Take a hypothetical state with four districts, each with 200 people.  The popular vote splits evenly – 400 Democrats, 400 Republicans. Here are the election results:

District Dem. Rep. Total 1 180 20 200 2 70 130 200 3 70 130 200 4 80 120 200 Total 400 400

The Republicans have 50% of the popular vote but get 75% of the seats. 

Less hypothetically, in North Carolina, Democratic candidates outpolled Republicans 2.22 million to 2.14 million.  But Republicans won 10 of the 14 seats.  The Democratic votes were crowded into four districts.  In three of those four districts, the Democrats won big – by an average of 133,000 votes.  (If the 7th district, where Democrats now have a slim lead, goes Republican, that average margin will be 177,000.)  Had some of the Democrats from one of those districts been mapped into the neighboring district, they might have won both, though by smaller margins.  The Republican districts had secure but smaller majorities.  Republican winning margins averaged 50,000 votes, less than half the margin where Democrats won.

My first thought was that this was pure Gerrymandering.  State legislatures get to draw the maps of their Congressional districts.  And many more state legislatures are controlled by Republicans.  In fact, some of the North Carolina districts have unusual shapes.  The NC-12, the thin blue line along Interstate 85 stretching nearly to the border, was created as a “majority-minority” district so that Black votes would not be diluted.  The downside for Democrats is that it packs those votes into that narrow corridor.  So the Democrats take that district by over 180,000 votes.  The Republicans win the neighboring districts but by much smaller margins - 23,000, 25,000, and 53,000.  In those four districts, the Democrats got 53% of the vote, but Republicans took three of the four seats.

The Democratic district snaking down through the middle of the state is the 4th, which contains “the Triangle” to the north, but now has that tail stretching down.  Democrats carried the district  by 170,000 votes.  Surrounding it is the 2nd (in pink), which Republicans carried by only 45,000 votes. 

Similar differences crop up in Ohio and Pennsylvania. The popular vote is close, and in two of these states it goes to the Democrats.  But Republicans get most of the seats.  Republicans win their seats by less than half the margin of Democratic winners.  Here is a graph of the actual returns from Ohio, Pennsylvania, and North Carolina. (The Ohio total does not include the vote from the two uncontested districts, one Democrat, one Republican.  For the maps and election results, check out Politico.)

The Republican share of Congressional seats is far out of proportion to its share of the vote.  In Ohio and North Carolina, Democrats received more votes, but Republicans got 70-75% of the House seats.  It certainly is possible that Republican-dominated state legislatures drew the districts so as to cram Democratic voters into electoral ghettos.

I don’t know enough about the demography and geography of these states, but I do wonder why the districts are drawn this way.  A paper by Chen and Rodd (here) that uses 2000 election data argues that what looks like gerrymandering is in fact the result of “human geography.”  It’s not the legislatures that pack Democrats together, it’s the Democrats themselves.  They cluster in cities.  As for Democrats outside of cities,

many rural, small-town, and suburban precincts that lean Democratic are often subsumed into moderately Republican districts. . . . There are isolated pockets of support for Democrats in African-American enclaves in the suburbs of big cities and in smaller towns with a history of railroad industrialization or universities. However, these Democratic pockets are generally surrounded by Republican majorities, thus wasting these Democratic votes. As a result, the Democrats are poorly situated to win districts outside of the urban core.

Regardless of intent, the effect is to keep Democratic votes concentrated in the 4th.  If that blue tail of the NC-04 were subsumed into the pink NC-02, both districts might be blue.

In any case, Democrats have not always been on the wrong side of the seat/vote discrepancy.  John Sides at The Monkey Cage posted this graph showing the ratio for the last twenty-six elections. 

Sides quotes Matthew Green on the general trends:

• the winning party usually gets a “boost” in the number of seats
• that boost used to be much larger

That trend might fit with the deliberate-gerrymander explanation, provided that in the earlier decades more state legislatures were controlled by Democrats.  But I’m not sure how it fits with Chen and Rodden’s human geography idea of “unintentional gerrymandering.”

UPDATE (Nov. 15):  John Sides at The Monkey Cage has more data on the vote/seat discrepancy.  He calculates seats expected given the popular vote and compares that to the actual outcome.  Even in states where districts were drawn by a bi-partisan agency or the courts, the Democrats fell 7% below the expected number of seats.  In states where legislatures did the redistricting the differences were starker.

Republicans gained benefits across the board from controlling the redistricting process.  By contrast, Democrats exceeded their expected seat share only slightly in the three states where they controlled the process

Post Parade

November 10, 2012
Posted by Jay Livingston

The front page of the New York Post is frequently a moral struggle.  Well, not much of a struggle.  The newspaper stands some obvious sinner in the pillory for Post readers to pelt with rotten vegetables.  The parade includes the criminal thug, the celebrity or sports star, the politician (foreign leaders make especially good targets provided they’re someone Post readers might have heard of), and others.

But some days, the Post goes for the puerile pun, as though the headline-writing room had been staffed with boys from a middle-school playground. 

And when the stars are properly aligned, you get both.

Then there’s the Daily News version.  Remember the moral outrage at MoveOn.org for using “Gen. Betray Us” epithet? From Wikipedia:

Fellow Max Boot* accused Moveon.org of “desperate attempts to besmirch one of the most admired soldiers in the entire American armed forces” and argued that the ad will “backfire.”

That was five years ago, and now the Daily News (hardly a left-wing rag) uses the same shameful name pun on page one.  Some backfire.

(Previous Post posts are here, here, and here.)
--------------

* If you were trying to think up names for a right wing hawk, could you come up with anything better than “Max Boot”?

“I Cannot Tell a Lie”

November 9, 2012
Posted by Jay Livingston

The father of our country may have been scrupulously truthful, but at least one of his namesakes has been fudging the data.

From Inside Higher Ed

George Washington University on Thursday became the third private university this year to admit that it has been reporting incorrect information about its new students -- both on the university's website and in information provided to U.S. News & World Report for rankings.

In the case of GW, the university -- for at least a decade -- has been submitting incorrect data on the class rank of new students. For the most recent class of new students, George Washington reported that 78 percent of new students were in the top 10 percent of their high school classes. The actual proportion of such students is 58 percent.

The irony goes beyond the name.  Universities all have rules about plagiarism and cheating on exams, and in the Student Handbook, accompanying the penalties and procedures section, schools sometimes add a righteous explanation: going out in the world disguised with a phony GPA is not only morally wrong, it is ultimately self-defeating. 

As if.

Cheating is rational, and the conditions that make it rational are the same ones that make it rational for George Washington to tell a lie. To judge the quality of a student or school in some meaningful way would be just too cumbersome.  You would have to get to know the person or school in some deeper way.  But when you have many students to grade and many schools to rate, it’s just so much more convenient to reduce all that information to a couple of numbers.  And it’s much easier to manipulate the numbers than it is to change what those numbers supposedly represent.

Predictions and Results

November 7, 2012
Posted by Jay Livingston

“Well, Howard, my predictin’ days is over,” said Muhammed Ali, even though the champ had been fairly accurate at predicting how many rounds an opponent would last. 

My post yesterday quoted Peggy Noonan predicting a Romney win.  On Fox last night, when one of the anchors asked Noonan if she were “surprised” by the results, she dodged the question and said only that she was disappointed.  Are her predicting days over?  Unlikely.

I contrasted Noonan’s “all the vibrations are right” methodology with Nate Silver’s thorough and complicated Bayesian model.  Conservatives accused Silver of liberal bias (and effeminacy), and they offered their own unbiased, clear-eyed  predictions. 

How did the prognosticators do?  Well, if somebody had bet on the Bayes, they would have cleaned up.  Here’s a scorecard.  I filled in some of the blanks in the grid that Neil Sihnababu posted at VoteSeeing .  Right now, the electoral vote stands at 303-206, with Florida not yet called, though most experts have put it in the Obama column.  Assuming that Florida is blue, Silver’s electoral prediction is perfect.  As for the battleground states, they all went for Obama (again, assuming Florida).  So did Nevada, adding one extra wrong prediction to Steve Forbes’s clear-eyed predictions.

(Click on the table for a larger view.)

You’d think that Karl Rove, George Will, and the others who did about as well as P’lod would follow Ali’s example.  Or at least, they would adjust their models, if they had any.  Or they would stop scoffing at science when it doesn’t tell them what they want to hear.  But don’t bet on it.

UPDATE:  Gabriel Rossman dresses up in his St. Paul costume (left over from Halloween, presumably) to deliver the same message. [Note for non-stats people who click on the link to Gabriel’s blog:  CLT = Central Limit Theorem.]