"Casting ballots" panel documents research progress in recent years

During yesterday’s MPSA conference (which I wrote about recently, and which continues through Sunday), there was one very interesting panel (nominally about methodology, but largely about the substance of election administration). Paul Gronke (who writes the Early Vote blog) was there with me, and we both found ourselves taking a lot of notes and finding the presentations very interesting. Unfortunately, the hotel facility where the conference was located did not have public wi-fi, so neither Paul nor I could engage in live blogging …

I’ll write more about some of the specific papers later, and I’m sure that both Paul and Thad will also write about their observations from this conference. However, there were a few general points that the papers made, and some specifics, that I thought I’d write about this morning before heading to the Illinois State Board of Elections hearing later this morning.

The papers fell into three areas: early voting, ballot design, and provisional balloting. The breadth of topics, and the sophisticated use of data and methodology, were impressive. It is clear that there is engagement now in the political science research community to study election administration, and this engagement promises to yield a healthy research literature in our journals soon.

Specifically, two of the presentations were about provisional balloting: Mcdonald’s and Andersen’s papers. Andersen relied largely on the EAC Election Day Survey data, and engaged in a sophisticated analysis of those data to study how many provisional ballots were cast, and how many were counted, in the 2004 election. Mcdonald, though, moved the analysis of provisional balloting in 2004 further, as he also used the Election Day Survey data, but he then provided new analysis of individual voter history data from North Carolina which allows an individual-level analysis of who votes provisionally and whether their vote was counted. While both Mcdonald and Anderson talked about problems with the Election Day Survey data (which Thad and I of course have written about recently), despite those problems there were a few important results that were discussed in their presentations. When I can link to these papers I’ll provide a more detailed discussion of their research.

Two of the other papers, one by Stein and Vonnahme, the other by Sinclair and myself, were about different topics (early voting and ballot design, respectively), but they were methodologically innovative, both using new techniques that take advantage of natural experimental situations. When I’ve got the time, I’ll discuss the substance and method of these papers as well.

Off to the hearings …