Category Archives: EAVS

PCEA research white papers

The Presidential Commission on Election Administration’s report is getting a lot of attention and praise following its release on Wednesday. One aspect of the report I want to highlight is the degree to which the Commission aimed to ground their findings in the best available research, academic and otherwise.  It renews my faith that it may be possible to build a field of election administration that is more technocratic than it currently is.

The report’s appendix, available through the supportthevoter.gov web site, is a valuable resource on the available research about each aspect of the commission’s charge.

I want to lift up an important subset of that appendix, which is a collection of white papers written by a collection of scholars, drawn from a variety of fields and perspectives, that summarized the large literatures that were relevant to the commission’s work.  A collection of those papers has been assembled in one place, on the VTP web site, so that others might have easy access to them.  Here are the authors and subjects:

Much of this research effort was assisted by the Democracy Fund, though of course, the research is all the work and opinions of the authors. Speaking personally, I greatly appreciate the support and encouragement of the Fund through these past few months.

An easier solution to the leading zero problem

  1. Start with the Excel file
  2. Add in 3 new columns (D, E, and F)
  3. Highlight column C and select “data / text to columns” from the menu
  4. Select “fixed format” and then use the cursor to separate columns 1-2, 3-5, and 6-9
  5. Finish

The result should look like this and allows a match/merge by State, County, and Township.

State Jurisdiction FIPSState FIPSCounty FIPSextra QA1a
AK ALASKA 02 0 0 560146
AL AUTAUGA COUNTY 01 1 0 34727
AL BALDWIN COUNTY 01 3 0 114952
AL BARBOUR COUNTY 01 5 0 16450
AL BIBB COUNTY 01 7 0 12239
AL BLOUNT COUNTY 01 9 0 31874
AL BULLOCK COUNTY 01 11 0 7650
AL BUTLER COUNTY 01 13 0 12898

Leading zero problems in the NVRA Dataset

In the EAC’s NVRA 2010 dataset, the FIPS code is stored as a numeric value, and this will cause problems with anyone who is trying to match / merge this with other data because 10 state FIPS codes start with a “0”.  Most statistical programs will strip that leading zero, making it hard to identify the states properly.

For instance, lines 2-4 of the Excel file look like this:

AK ALASKA 200000000 560146 only active voters
AL AUTAUGA COUNTY 0100100000 34727 active and inactive registered voters
AL BALDWIN COUNTY 0100300000 114952 active and inactive registered voters

The third column contains the FIPS code; columns 1-2 contain the state code, columns 3-5 the county code, an additional columns refer to smaller jurisdictions (mainly townships in NE and the Midwest).

The first problem you’ll notice is that Alaska’s FIPS is already artificially shortened to “2” not “02” an users will need to fix this manually.  Next, additional FIPS codes will come across a “10010000” (for example) without the leading zero.

The solution for now is a kludge; insert a dummy line into the Excel file and include a string value such as an “x” in column 3.  This will trick Excel into outputting the values as text, and statistical programs will input the data as character strings.  Then the process is straightforward; capture substrings (01, 02, etc) and match / merge away!