Archive for the 'Methodology' Category

Page 4 of 11

Economists in a Blizzard

Economists are at their annual meeting in the teeth of a big snowstorm. In case you missed the conference, here’s the presidential address by Claudia Golden on the cause of the remaining gender gap:

A Grand Gender Convergence: Its Last Chapter
Claudia Goldin | Harvard University
American Economic Association Presidential Address [draft version]
January 4, 2014

More relevant to PSC is that Martha Bailey and Brad Hershbein were awarded the IZA Young Labor Economist Award for their paper The Opt-In Revolution? Contraception, Fertility Timing and the Gender Gap in Wages.

And via Twitter, Justin Wolfers points to some job openings for economists – internet economist.


[Link to Internet Economist article]

Human Behavior Trove Lures Economists to U.S. Tech Titans
By Aki Ito | Bloomberg
January 03, 2014

Death at the Summit: The story of Intrade

Death at the Summit
Graeme Wood | Pacific Standard Magazine
November 4, 2013

We have posted on prediction markets before, but this is a very nice, if long summary of the theory of prediction markets; use by scientists, mostly economists; the reaction to them by pundits; and the death of Intrade (and its founder).

Project Tycho: Historical Disease Data

Project Tycho is funded by NIH and the Bill and Melinda Gates Foundation. It has taken historical data from the National Notifiable Diseases Surveillance System (NNDSS) and created count data for diseases by location for the 125 year history of the surveillance system. Three levels of data have been made available to users from the Project Tycho website.

Other useful resources:

  • Materials and Methods: Digitication of US Weekly Surveillance Reports between 1888 and 2011
  • Preliminary State Reports (scroll down for access – here’s an example for Michigan)
  • What does Tycho stand for?
  • And, here is a paper published in the New England Journal of Medicine based on these data. It estimates that over 100 million cases of contagious diseases have been prevented in the U.S. since 1924 by vaccination programs against polio, measles, mumps, rubella, hepatitis A, diphtheria, and pertussis (whooping cough).

    Contagious Diseases in the United States from 1888 to the Present
    New England Journal of Medicine
    November 28, 2013
    html | pdf

    Do You Still Trust the Census Bureau?

    The New York Daily News had a typically provacative headline “Census ‘faked’ 2012 election jobs report” two nights ago. This is a serious charge and even more, it contributes fodder to those who do not trust or support the federal data infrastructure in the first place. The following is the banner above the comments section for the New York Post article – and this sentiment probably represents the early coverage of this story.

    trust census bureau logo

    The following is the coverage of this in chronological order (as much as possible). Note that there are some references to Jack Welch. He famously tweeted his disbelief of this particular jobs report back in 2012 [See previous coverage.]

    Census ‘faked’ 2012 election jobs report
    John Crudele | New York Post
    November 18, 2013

    If these claims by ‘reliable sources’ are proven true, the Obama administration will be dealing with another huge scandal
    Becket Adams | The Blaze (founded by Glenn Beck)
    November 18, 2013

    Census Bureau Statement on Collection of Survey Data
    November 19, 2013

    Here Are Some Issues With That Report About How The Unemployment Rate Was Faked Before The 2012 Election
    Joe Weisenthal | Business Insider
    November 19, 2013

    Was Jack Welch right? Jobs numbers under fire
    Jeff Cox | CNN
    November 19, 2013

    Did the Census Bureau Really Fake the Jobs Report?
    Jordan Weissmann | The Atlantic
    November 19, 2013

    Five questions about the New York Post’s unemployment story
    Erik Wemple | Washington Post
    November 19, 2013

    Census Sees No ‘Systemic Manipulation’ of U.S. Jobs Data
    Michelle Jamrisko | Bloomberg News
    Nov 19, 2013

    House panel to investigate unemployment data
    Annalyn Kurtz | CNN Money
    November 19, 2013

    House probes Census over ‘fake’ results
    John Crudele | New York Post
    November 19, 2013

    Rep. Issa gets involved in alleged Census data fabrication, demands documents: ‘These allegations are shocking’
    Becket Adams | The Blaze
    November 19, 2013

    Monthly jobs numbers from Census Bureau may have been manipulated since ‘10 – report
    RT USA
    November 19, 2013

    Republican House leaders to look into report on faked jobs data
    Reuters News Service
    November 20, 2013

    Political Questions About the Jobs Report
    Nelson Schwartz | New York Times
    November 20, 2013

    Census Bureau: No systematic manipulation of jobs data
    Paul Davidson | USA Today
    November 20, 2013

    Count your blessings; you could live in Canada

    The following are articles, mostly from the Canadian press about the (a) the quality of data in the National Household Survey (NHS); and (b) the politicization of funding for basic science research. Much of the poor quality of the NHS data has to do with design changes at the behest of the prime minister’s office, rather than the statistical experts at Statistics Canada.

    [Criticism of the National Household Survey]
    To restore faith in Statscan, free the Chief Statistician
    Munir Sheikh | The Globe and Mail
    October 24, 2013
    This op-ed is written by the former Chief Statistician who resigned amid the changes in the design of the National Household Survey. He could not agree with the statements coming from the Prime Minister that a voluntary survey can be a substitute for a mandatory survey. Here’s his resignation letter with the famous “It can not” sentence:

    And that’s all he wrote. . . Munir Sheikh resigns as Chief Statistician
    Kady O’Malley | CBC
    July 21, 2010
    [Resignation letter]

    Canada’s voluntary census is worthless. Here’s why
    D. Hulchanski, R. Murdie, A. Walks, and L. Bourne | Globe and Mail
    October 4, 2013
    Data from the NHS show that Canada’s income inequality has dropped. But, this may have more to do with the flawed NHS than reality. The authors compare tax receipt data to NHS data to illustrate the problem.

    Canadian income data ‘is garbage’ without census, experts say
    Tavia Grant | The Globe and Mail
    October 4, 2013

    [Politicization of Science Funding]
    Blinded to science: The plight of basic research in Canada
    Josh D. Neufeld iPolitics Insight
    October 21, 2013
    This piece is a good summary of the move by the Canadian government towards funding applied research instead of basic research. This statement summarizes the issue:

    Basic research is the seed corn of the economy, generating the applications and economic benefits of tomorrow … Trouble is, it’s very difficult to predict which basic research programs and projects will lead to the innovations of tomorrow.

    Others from the series of posts on science policy in Canada can be found here:

    Series of Posts on Science Policy in Canada
    to be published in iPolitics

    Quantitative Text Analysis: Michael vs Jacob

    Most of the data demographers use are numeric and are easily handled via statistical packages. Text data via Google NGrams or names from the Social Security Names Database are more commonly analyzed using Python.

    CSCAR and ARC are sponsoring free Python training Friday, November 8th. Space is limited.

    In case you miss the workshop, here’s a link to some Big Data Tutorials by Neal Caren at the University of North Carolina, Chapel Hill.

    And, to get back to the title of this blog entry, below are three data visualizations on names. The first two are the most common name by state & gender from 1960 to 2010.

    Click on the images to activate the gifs.

    us_map_gnames us_map_bnames

    Notice that for the girls, Lisa dominated the US in 1965, which means I was born 10+ years too early to have that name. And for the boys, watch the epic battle for Michael vs Jacob. Also note that Jose is the dominant male name in Texas in 1996. Arizona also has two Hispanic names (Jose and Angel) in the recent past.

    The third data visualization explores unisex names:
    unisex names

    Finally, think of these as data. We have a link to research on black first names as well as a post on the declining popularity of Mary.

    Big Data Tutorials, Neal Caren (University of North Carolina, Chapel Hill).

    Google NGrams Viewer

    Google NGrams Data
    Note, we have downloaded quite a bit of this. See Data Service before you download another copy.

    Social Security Names Database

    A Wondrous GIF Shows the Most Popular Baby Names for Girls Since 1960
    Rebecca Rosen | The Atlantic
    October 18, 2013

    America’s Most Popular Boys’ Names Since 1960, in 1 Spectacular GIF
    Megan Garber |The Atlantic
    October 24, 2013

    The most unisex names in US history
    Data Underload | FlowingData Blog
    September 25, 2013

    Visualizing Births and Deaths in Real-Time

    Data visualizations are becoming more and more popular and sometimes they include demographic concepts. The following are two simulations of births and deaths – one for the US and the other for the world.

    Click on the images to start the simulations. To read more about how these were made see references below:

    us_map world_map

    Watch This Anxiety-Provoking Simulation of U.S. Births and Deaths
    John Metcalfe | The Atlantic Cities
    December 11, 2012

    This Map Shows Where in the World People Are Dying and Being Born
    John Metcalfe | The Atlantic Cities
    October 14, 2013

    World Births/Deaths Simulation – Adding World Cities
    Brad Lyon | Nowhere Near Ithaca Blog
    October 9, 2013

    Reproducibility Initiative: It’s not just for cancer

    Reproducibility Initiative logo

    The following are links to related efforts in Open Science. The first is about funding for a “Reproducibility Initiative” to validate 50 landmark cancer studies. Frankly, this can/should apply to population research as well. Included are links from The Economist and Nature about the importance of replication.

    In general, there is a move towards “Open Science” across all disciplines. In fact, a different initiative, “The Reproducibility Project” is an effort to identify the predictors of reproducibility among published studies in psychology – a field that contributes far too much to the “Retraction Watch” website.

    Reproducibility Initiative
    Science Exchange News
    October 16, 2013

    Initiative gets $1.3 million to verify findings of 50 high-profile cancer papers
    Richard Van Noorden | Nature News Blog
    October 16, 2013

    Unreliable research: Trouble at the lab
    The Economist
    October 19, 2013
    Scientists like to think of science as self-correcting. To an alarming degree, it is not.

    The governments of the OECD, a club of mostly rich countries, spent $59 billion on biomedical research in 2012, nearly double the figure in 2000. One of the justifications for this is that basic-science results provided by governments form the basis for private drug-development work. If companies cannot rely on academic research, that reasoning breaks down. When an official at America’s National Institutes of Health (NIH) reckons, despairingly, that researchers would find it hard to reproduce at least three-quarters of all published biomedical findings, the public part of the process seems to have failed.

    If a job is worth doing, it is worth doing twice
    Jonathan Russell | Nature
    April 3, 2013

    Reproducibility Project
    Large-scale open collaboration to estimate the reproducibility of a sample of studies in psychology

    Retraction Watch
    Tracking retractions as a windo into the scientific process

    Center for Open Science
    A non-profit organization, which provides infrastructure tools for open science.

    Time for a Legal Prohibition on Data Re-identification?

    This is a very thorough blog post on respondent re-identification issues. The author takes to task the re-identification rainmakers, who have made careers out of exposing re-identification risks – often overstating the risks. He calls for a well-designed legal prohibition on data re-identification.

    In fact, may of the restricted data contracts PSC users operate under have an “inadvertent discovery” clause. Here’s the language from LAFANS, which prohibits broadcasting the “find” to others.

    Ethical Concerns, Conduct and Public Policy for Re-Identification and De-identification Practice: Part 3 (Re-Identification Symposium)
    Daniel Barth-Jones | Columbia University
    October 2, 2013

    Counting Prisoners

    The New York Times had another editorial on this issue:
    Prison-Based Gerrymandering
    Editorial Board | New York Times
    September 26, 2013

    A search on its site shows that this has been a common editorial/story topic
    [Counting Prisoners Editorials/Stories]

    The PSC Infoblog has had a previous post on this topic as well, which included the Census Bureau’s response to the issue. The Census Bureau released group quarters data in time for redistricting.

    Another excellent source on this topic is the National Academy of Sciences book, which is available in the PSC library:
    Once, Only Once and in the Right Place: Residence Rules in the Decennial Census
    Daniel Cork and Paul Voss, Editors | The National Academies