Monthly Archive for August, 2013

Natural Experiments: In Education

A former PSC student, Mike Lovenheim, has several recent papers that have made the news because they shed light on educational policy. I’m only going to highlight two of these, but link to several of his working papers for those who want to see the creative ways he is examining the economics of education and issues in local taxation. And, a new paper has just been added to this entry. It uses lottery-data to evaluate the NY small school movement.

Small High Schools and Student Achievement: Lottery-Based Evidence from New York City
Atila Abdulkadiroğlu, Weiwei Hu, Parag A. Pathak | NBER
October 2013
Abstract | Paper

Early Retirement Incentives and Student Achievement
Maria D. Fitzpatrick & Michael F. Lovenheim | NBER Working Paper, No. 19281
August 2013
Abstract | Paper

[Popular Press]
Student Test Scores Rose When Teachers Retired Early
Khadeeja Safdar | Wall Street Journal
August 19, 2013

Does Federal Financial Aid Affect College Enrollment? Evidence from Drug Offenders and the Higher Education Act of 1998
Michael F. Lovenheim and Emily G. Owens | NBER Working Paper, No. 18749
February 2013
Abstract | Paper

[Popular Press]
Financial Aid for College Students With Drug Convictions
February 7, 2013

Scientific version of insider trading

This is not within demography, but here’s a snippet that every researcher shudders to think about:

[From the Retraction Watch website]

Case Western dermatology department hit with second ORI sanction within 6 months

The charge:

“engaged in research misconduct by plagiarizing significant portions from research grant application R21 AR061881 that she had reviewed for NIAMS, NIH, and inserting that text into her submitted grant application R01 AR062378-01. Respondent also plagiarized significant portions of text from the following scientific articles and one U.S. patent application available on the Internet.”

Big data and the death of polls

The rumors of the death of polls might be greatly exaggerated. Recent coverage of a Twitter-based study ignores the weak effects in the original paper. [“For instance, being an incumbent predicts almost a 50,000 vote contribution to the Republican margin in their statistical model, whereas receiving 100 percent (all!) of tweet-mentions gets you only 155 votes”]. But, one of the authors of the papers even goes so far as to say “In the future, you will not need a polling organization to understand how your elected representative will fare at the ballot box. Instead, all you will need is an app on your phone.”

How Twitter can predict an election
Fabio Rojas | Opinions, Washington Post
August 11, 2013

Original Paper
More Tweets, More Votes: Social Media as a Quantitative Indicator of Political Behavior
J. DiGrazia, K. McKelvey, J. Bollen and F. Rojas | SSRN
February 21, 2013

How Twitter can Predict Elections: A Rebuttal
Rob Santos | Washington Post
August 16, 2013

Can Twitter Predict Elections? Not so Fast
Mark Blumenthal & Ariel Edwards-Levy | Huffington Post
August 16, 2013

Let’s Calm Down about Twitter Being Able to Predict Elections, Guys
Jason Linkins | HuffingtonPost
August 14, 2013

Popular Press
How Twitter can help predict an election – in one eye-catching study
Sean Sullivan | Washington Post
August 14, 2013
Want to figure out who is going to win a congressional race? Find out which candidate received the lion’s share of tweets in the lead-up to Election Day.


Some high-profile misses are also illustrative of the challenge of using tweets to reliably project elections. Anthony Weiner’s nearly 250,000 mentions on twitter (according to are unlikely to revive his downward spiral in the New York mayoral race – current front-runner Bill de Blasio has received barely 10,000 mentions in the same period. And while then-Rep. Ron Paul (R-Tex.) received wide recognition on Twitter during the 2012 Republican presidential primaries, he failed to win a single contest.

A New Study Say Twitter Can Predict US Elections
Robinson Meyer | The Atlantic
August 13, 2013

[Unrelated Visualization of Tweets before the 2010 Election]

Measuring Morality Data

The first phase of the Measuring Morality project is complete and the data are freely available once the user has registered on the website (and agreed to send citations to the project).

Measuring Morality website
Technical Details | Codebook

In the second phase of this project, selected items from the above survey will be included in the fourth wave of the National Study of Youth and Religion.

The Perils of Administrative Censuses

Some who are against a mandatory census argue that the government already has this information and is wasting money re-collecting data. Of course, not all information on individuals is tied to their residence and the census needs to know the location of the population for reapportionment purposes. Others who are against the census are also against big government so probably are not in favor of administrative records as a data collection device.

The following is the record of the German administrative census as compared to a population count. Some of the sources are US-based research from the Census Bureau, which is looking to use administrative records to supplement its address-based census.

Germany Counts Heads and Finds 1.5 Million Fewer Residents Than It Expected
Press Release | Statistisches Bundesampt [German Federal Statistical Office]
May 21, 2013

Lessons from the German Census
D’Vera Cohn | Fact Tank: Pew Research Center
June 20, 2013

When the results of the 2011 German census were announced recently, they included an embarrassing error – at least in the demographics world. It showed the German population was 1.5 million people short of what the government had expected. The news dealt a blow to Germany’s reputation for efficient record-keeping, and it’s also relevant to how the next U.S. Census is conducted.

2010 Census Administrative Records Use for Coverage Problems Evaluation Report
Sheppard, Dave, | Census Bureau
March 18, 2013

2020 Census: Local Administrative Records and Their Use in the Challenge Program and Decennial

February 21, 2013
Highlights | Full Report

And Now for Something a Little Different. . .
Bob Groves | Director’s Blog: Census Bureau
June 27, 2012

Toward a Vision: Official Statistics and Big Data
C. Capps and T. Wright | AmStatNews
August 1, 2013
This piece even references Herman Hollerith:

The Census has a long history of innovation. Herman Hollerith invented the punch card for the 1890 Census; the first civilian computer was used for the 1950 Census. The first official sample survey was used by the Census Bureau to measure unemployment in 1937. Some of the basic technology for GIS was developed in the Dual Independent Map Encoding/Graphic Base Files efforts for the 1970 Census and TIGER for the 1990 Census.

Each of these innovations was done to reduce escalating cost and to preserve official statistical integrity. For these same reasons, the Census Bureau will continue to explore the possibility of using the explosion of Big Data to reduce cost, reduce reporting burden, and increase the effectiveness of national statistical estimation.

These benefits will accrue only if the Census Bureau can continue to preserve individual and corporate confidentiality, working to earn and preserve the public’s trust.

Home Mortgage Disclosure Act Data Series: 1997-2011

These data are available via the Urban Institutes’s National Data Repository.

These data include information about mortgage loan applications, including the outcome of the application, information about the loan and applicant and location of the property (census tract). The Urban Institute has summarized the loan-level data into indicators on the racial and income distribution of borrowers, denial rates by race and income and loans from subprime lenders by race.

Users must register to download the data and provide attribution to the Urban Institute.

Urban Institute Data Repository
Click on UI Home Mortgage Disclosure Act Data Series and then the Data and Analysis tab.

Additional Resources:
A Guide to Home Mortgage Disclosure Data
K. Pettit and A. Droesch | Urban Institute
December 2008

The original source for these data comes from the FFIEC website

Friday News: Bad

Here is some bad news for scientific research funding, which pretty much missed the news cycle. This will be updated with the reactions and responses which will hit the media cycle next week.

NSF cancels political-science grant cycle
US funding agency said to be dodging restrictions set by Congress.
Beth Mole | Nature
August 2, 2013

The NSF’s decision removes one of the main financial lifelines for political-science research. “This is somewhere between devastating and crippling,” says Henry Farrell, a political scientist at George Washington University in Washington DC and an author of the Monkey Cage, a widely read political-science blog. But Farrell blames the political climate rather than the funding agency for the cut. “The NSF is in an extremely awkward situation,” he says.

. . . . . Avoiding the August funding round may be a strategic move by Hume to see whether the constraints disappear when the next spending bill is passed, says Aldrich. “If he can save the money and spend it later when there’s more clarity, that would be helpful,” Aldrich says.

Other researchers agree. “I think they’re probably worried about upsetting Congress,” says Rick Wilson, a political scientist at Rice University in Houston, Texas, and editor of the American Journal of Political Science. “So why not pull the plug rather than risk it?”

Friday News: Good

The Senate voted to approve John Thompson as the new Census Bureau director. As a reminder, the term for the Census Bureau director is now a fixed 5-year term, which would normally have started in 2012 and ended in 2016. [See this post, for clarification.]

Thompson Confirmed as Census Bureau Director
Census Bureau
Friday, August 2, 2013

Acting Director, Thomas Mesenbourg closed out his tenure with some thoughts about the importance of the federal statistical system for democracy

Farewell from Tom Mesenbourg
Director’s Blog | Census Bureau
August 1, 2013

“Democracies require statistics that are credible, trusted, nonpartisan, and relevant”

Convincing Congress of that will be one of the many challenges facing Thompson.