#electionsciences2017 starting this week!

The first Election Sciences, Reform, and Administration Conference is happening this week in Portland, OR!

I’d like to thank Phil Keisling and Paul Manson of the Center for Public Service at Portland State University for helping to organize, and the Reed College Department of Political Science, the MIT Election Data and Science Lab, the National Science Foundation, and the Elections Team at the Democracy Fund for making this event possible.

Follow the link above, or point your browser to electionsciences.net for more information.

This announcement from Jay Lee, Matthew Yancheff, and Mia Leung, three Reed students who were in the Data and Election Sciences course that I taught along with Prof. Andrew Bray this spring.  They have released the results of their work to CRAN.

Thanks to Rob Richie and Theo Landsman of FairVote for helping push this forward.


Just wanted to let you know that the first version of our RCV package is now submitted to CRAN, the R package archive! Going forward we’ll be updating this work, so if you have any comments or bug fixes please feel free to submit a pull request or issue to our GitHub repository, or just email us directly.
I’ve included a few lines of code at the bottom of this email to install the package locally and go through an example election (San Francisco Board of Supervisors, District 7). You’ll need at least version 3.3 of R installed to run these. If you don’t have this installed and don’t want to, some of the examples are available at our GitHub repo (scroll down to the README).
Again, thank you so much for your interest in our work and any help you’ve given us in regards to this project. We look forward to hearing any comments or critiques you might have on your experience using our package.
Thank you,
Jay Lee
Reed College
sf_cleaned <- clean_ballot(ballot = sf_bos_ballot, b_header = T, 
                        lookup = sf_bos_lookup, l_header = T, 
                        format = "WinEDS")
results7 <- rcv_tally(sf_cleaned, "Board of Supervisors, District 7")

The results table for this election is stored in the `results7` object. It can be printed in the console with the first line of code provided below, or viewed in the RStudio window with the second line:
We also have a functionality for producing an interactive type of flowchart called a Sankey diagram. This is done with the networkD3 package, which you must install separately to produce the visualization. The code for that is again provided here, but if you don’t want to install it we have an example on our GitHub repo.
d3_7 <- rcv::make_d3list(results = results7)
networkD3::sankeyNetwork(Links = d3_7$values, Nodes = d3_7$names,
                         Source = "source", Target = "target",
                         Value = "value", NodeID = "candidate", units = "voters",
                         fontSize = 12, nodeWidth = 20)

Maricopa County, AZ is the second largest election jurisdiction in the United States (after Los Angeles County) and is contemplating a move to all-mail ballot delivery, with ballot returns by mail, drop box, or use of a “ballot center.”

This story from the Arizona Republic is lengthy, and it illustrates a lot of the concerns that will be raised in other localities who may contemplate the switch:

  • Is it secure?
  • It is efficient?
  • Is it fair to everyone?
  • I like voting in person, can’t I continue to do so?

Here’s the link: http://www.azcentral.com/story/news/local/phoenix/2017/06/20/county-recorder-adrian-fontes-changes-voting-system-fall-2017/391279001/

EVIC (or at least a report we worked on) is in the news!


I’ll let the OPB story speak for itself, since I was one of the co-authors of the report. 

Early Voting: An Advantage for Republicans?






The research team at the Elections Research Center at the University of Wisconsin, Madison, have a new paper analyzing the partisan impact of early voting laws, in combination with a set of other election reforms. The abstract is provided below; the piece is gated at the Political Research Quarterly but may be available from the authors. 

I would still caution against overinterpreting these results as providing a roadmap for election law gamesmanship. Burden et al. spend a bit too much time, in my judgment, opining about how partisan actors may or may not misestimate the political impact of reforms to election laws, without acknowledging the highly contingent and dynamic nature of the legal and administrative environment. 

For example, it’s almost certain than when a new voting method is made available, strategic political actors from both parties look at these changes, look at what groups opt for one or another method, and start to change their campaigns accordingly. Capturing this kind of institutional dynamic is nearly impossible to do in a national study like this, and can easily make gamesmanship seem a lot simpler than it actually is. 




I posted this query on the Political Methdology listserv:

Hello all, I have some students in an election sciences class who want to do some visualizations using CCES data. I’d like them to use the survey weights if possible, but don’t know an easy way to do this in R.

I have come across this package that claims to support graphics and complex survey weights, but can’t find a reference or vignette that uses any graphics: http://r-survey.r-forge.r-project.org/survey/


Paul Gronke
Reed College

And have provided a review of the answers:


Thanks to Jay Lee of Reed College for helping me assemble this. 

Thanks to those participants in the Political Methodology listserv who responded to a query that I posted a month ago about how to produce survey “toplines” using either Stata or R. The attached document provides a detailed summary of the responses; I have posted the most useful reply here.

Original Question

From: Paul Gronke <gronkep@reed.edu>

Quick question for the list: Lisa Bryant (CU Fresno) and I are preparing  some “top lines” and “tabs” for a client with whom we conducted a survey.

If you have seen these before, they are usually organized so that  categorial survey responses are reported on the rows, and the columns  report the overall responses, then responses “tabbed” or “crossed” by  various demographic and political categories. Roll your eyes if you will  that this is just a big set of exploratory cross tabs, but a lot of folks  expect to see them to help digest the survey results.

A typical “tab” looks like this:

VARIABLE                        Total   GOP     IND             DEM MEN     WOMEN    …

Category 1                             N  %    N  %    N  %            N  %    N %    N  %

Category 2                      N  % …


Stata Special

From: Jonathan Mendelson

Hi Paul,

I posted a response to the list, but it hasn’t gone through yet, so I thought I’d reply directly.  I encountered the same issue as you and wrote a Stata package that essentially creates “tabs” in spreadsheet form.  You can install it in Stata via “ssc install tabsheet” or view information at https://ideas.repec.org/c/boc/bocode/s458128.html; there are examples in the documentation so you should be able to get started fairly quickly.

The program is not particularly flexible, but it is easy to use, and some colleagues at my survey firm have found it very useful.  Although it doesn’t currently output to anything other than tab-delimited file (which can be opened in a spreadsheet), with some clever formatting in Excel, you could print the resulting spreadsheet to PDF for something nicer looking.

If you need something more flexible in Stata, I’d recommend tabout, although that may require more work to set up.  If you find out about any R packages that do something similar, I’d be interested in hearing about it.


The complete list of responses, including various R and Stata solutions, is provided in this PDF: polmeth-question-survey-tabs