[The Sunny Dog]


My overarching research focus is on understanding how interests are represented in the American political system. I am particularly interested in the coalitional behavior of societal groups and how the public interest is administered by the efforts of politicians. My research on these topics investigates parties, ideology, political leadership, public opinion, political participation, and political socialization.


My dissertation work is my primary project at the moment. Please refer to my dissertation web page for information about that project.

Effect of the Voting Age on Turnout Over the Life Course. Austria lowered its voting age to 16 in 2007. Many other countries have considered following suit. Perhaps most notably the UK Parliament recently took the matter under serious consideration, but ultimately decided against lowering the voting age largely because its Electoral Commission reported that there was insufficient understanding of the long-term implications: "Ultimately, arguments about the voting age and turnout in the UK are largely based on speculation and hypothesis, with strong views -- but little conclusive evidence -- on either side as to whether or not the longer-term effect of lowering the voting age on turnout would be positive." I investigate this question using voter list data adjusted for mortality and migration to estimate the effects of the 26th Amendment on turnout over the life course in the United States. Below is a graph of monthly birth cohort turnout rates in the 2004 general election estimated using the 2005 California voter list and making adjustments for mortality and migration. Note the prominent increase in the slope that takes place where cohorts were eligible to vote at 18. This indicates that cohort turnout at, say, age 40 is negatively affected by cohort enfranchisement at age 18 rather than 21.

Ideological and Congressional Polarization. Most of the congressional polarization developing in the past fifty years has evolved since 1990. However, substantial ideological polarization -- as measured by real ADA scores -- had already developed when the congressional parties slowly began polarizing in the 1980s. I am working on an explanation that relates these two developments. Below is a plot of the cumulative increase in congressional polarization (DW1) and MC ideological polarization (real ADA scores) that developed between 1961 and 2004. I argue that the race issue caused change in how liberalism and conservatism were organized, and that this ideological change ultimately caused the congressional coalitions to reorganize. Institutionalized congressional parties made it difficult for legislators to quickly reorganize around the emergent ideological cleavage, which is why DW1 polarization develops much later that ADA polarization.


Partisan Generational Effects: How the National Mood During a Cohort's Adolescent Year Shapes Its Lifelong Macropartisanship (download pdf). Paper presented at the annual meeting of the Midwest Political Science Association, Palmer House Hotel, Chicago, IL, Apr 12, 2007
Explaining cohort macropartisanship was my first project in graduate school. My MPSA paper uses NES and GSS survey data. But the plot that started me thinking about this question I generated using the 2005 California voter list. Here's a plot of the Republican-to-Democrat ratio of party registration in California by birth date cohort based on the 2005 California voter list -- the variation across years is striking!

Matthew D. Atkinson and Anthony Fowler. The Effect of Social Capital on Voter Turnout: Evidence from Saint's Day Fiestas in Mexico (download pdf). Under Review.
Figure 1 of our paper nicely captures how turnout is affected by the proximity of the election date to a community's saint's day festival. This figure is based on a kernel regression that accounts for variation associated with different election years and the mean turnout levels in each state.

Atkinson, Matthew D., Ryan D. Enos, and Seth J. Hill. 2009. Candidate Faces and Election Outcomes: Is the Face-Vote Correlation Caused by Candidate Selection? (download pdf). Quarterly Journal of Political Science 4(3): pp 229-249.

Some of the neat things we did with our faces data didn't make it into the article we submitted to QJPS. For example, Ryan made this graphic showing how the distributions of challenger and incumbent facial competence scores differ. The shaded region in each plot represent the distribution of challenger facial competence.