Had a great time at #PolMeth2019. Now, as is my tradition, I'm going to go lock myself in a dark room and shun all human interaction for approximately two days. See you all on the other side! 👋

Schoolhouse Rock taught us that bills can change a lot as they become a law, but models using roll call/cosponsorship data treat bills as static. @jcrohsaine, @GeoffLorenz, and @zfurnas (👈 on the market) cleverly address this by leveraging interest group data #PolMeth2019

Really enjoyed seeing @jayrobw’s work using image analysis to map narrative clusters of extremism. Very cool, innovative work with a lot of potential applications #PolMeth2019

Very important public service announcement from discussant @burtmonroe: it’s pronounced “BEE-chum” #PolMeth2019

Nick Beauchamp presents an extension of our conceptual networks work, developing a Bayesian transition network topic model which captures topic shifting behavior #PolMeth2019

Patrick Miller discusses the pervasive discrimination against trans people: most Americans don’t know a trans person or understand what it means to be transgender. People who care about equality in other dimensions don’t necessarily support transgender rights #PolMeth2019

Fantastic work from Katherine McCabe looking at the political behavior of queer republicans. She also discussed more broadly the challenges of doing intersectional research and/or looking at subsets of the LGBTQ population #PolMeth2019

I particularly appreciated @Patrick_J_Egan’s point that a wide range of identities are fluid and change over time. We shouldn’t be methodologically surprised by the fact that gender identity is not fixed: it behaves similarly to other dimensions of identity #PolMeth2019

Really, really glad to see #PolMeth2019 include a panel on LGBTQ research and methodology. @Patrick_J_Egan kicks us off with a great talk on the challenges of accurately capturing these identities in surveys

Using millions of records from two reconstructed datasets, @roya_talibova examines the effect of repression on soldier performance in battle #PolMeth2019

Great work from @_LeslieHuang clustering questions that receive similar answers and identifying question-asking patterns. Prosecution and defense question their witnesses differently and there are a number of questions which result in the respondent unable to recall #PolMeth2019

Very cool approach using entropy to identify landmark legal case from @KristenRenberg and Joshua Lerner #PolMeth2019

In advance of elections, legislators actively shift policy strategy to focus on bills that will increase the chances of electoral success finds @LuciaMotoliniaC. Bonus points on this cool work for explicitly covering her topic model validation strategy #PolMeth2019

In her work looking at the effect of street vendors and informal markets on government revenue @janica_magat finds that we should stop considering poor people to be powerless: they can wield significant political power which is reflected in government tax policy #PolMeth2019

Love this work from Rachel Porter looking at candidate strategy by gender in the 2018 midterms. She scraped 1k candidate websites and used a structural topic model to identify different subjects raised in candidate biographies. Turns out: women are not a monolith #PolMeth2019

Excellent work from Erin Rossiter, presenting a framework for studying conversations online #polmeth2019

Shoutout to the brilliant @LauraK_Nelson during @arthur_spirling’s discussant comments at #PolMeth2019

Very important work from Luwei Ying, @Jacob_Montg and @b_m_stewart on validating topic models. Coming soon: an R package to help you run MTurk studies to validate the coherence and labels of your own topic models #PolMeth2019

Found #polmeth2019’s smallest methodologist. I think it was trying to ask a question?

Q: Do Gender Quotas Cause a Decline in Corruption? A:...not really

@JustinEsarey finds gender quotas can reduce corruption *if* women actually have power. Countries susceptible to international pressure are more likely to implement quotas but less likely to benefit #PolMeth2019

Very cool work from @brycejdietrich using vocal intonations to infer and emphasize the emotional valiance of respondents. He’s looking for data sets (especially phone surveys) and is developing an app that can record and analyze your audio data #PolMeth2019

What drives opinions on the Affordable Care Act? Cool work from @dhopkins1776 and @willrhobbs finds that the ACA had heterogeneous effects, making it hard to infer coherent public opinion on a mass level #PolMeth2019

All set up for today’s #PolMeth2019 poster session! Come find me in the corner at #52 to learn about my work inferring the network structure of reasoning. Come for the view and stay for the research!

Love this work from @LC_Sanford and @mollyeroberts: using a deep neural net, they identify and weight phrases of arbitrary length which are predictive of some outcome. For example: identifying posts which are likely to be censored #PolMeth2019

As a bonus: anyone who finds me in person at #PolMeth2019 can have one of these sweet, sweet #FeministMafia lanyards inspired by @prof_mirya.

At #PolMeth last year @JaneLSumner and @prof_mirya tweeted they would be friends with any attendees who didn’t know anyone or were otherwise anxious

Neither are here this year, so I’ll make the offer in their place: need a friend? Come say hello or DM! 👋🙌 #polmeth2019