Dear Society members and friends,
It is my pleasure to announce the best poster awards for posters presented
at this summer's annual (and virtual!) PolMeth Conference. Please join me in
congratulating Melody Huang (UCLA), Nuannuan Xiang (Michigan), Erin Rossiter
(WUSTL), Luwei Ying (WUSTL), and Jay Goodliffe (BYU).
My thanks to the award committee members John Londregan (chair), Michael
Bailey, Sarah Bouchat, Patrick Brandt, Charles Crabtree, Thomas Gschwend,
Erin Hartman, and Betsy Sinclair. The announcement from the committee
follows.
Best,
Suzie Linn
After a careful perusal of a very competitive field of 83 posters, the
prize committee is pleased to announce the following three awards.
The prize for the best graduate student poster in methods is shared by
Melody Huang and Nuannuan Xiang.
Huang's poster, ``Leveraging Observational Outcomes To Improve the
Generalization Of Experimental Results", takes an innovative approach that
leverages outcome data in the target population to improve the precision of
the population estimate by residualizing the sample data before doing
inverse propensity weighting. Melody goes on to reanylze a multi-site
experiment and finds her method delivers more precise estimates of a known
experimental benchmark.
Xiang's poster entitled ``A Gaussian Process Model for Causal Inference with
TSCS Data" compares her GP model with the state of the art GSC model. The
GP approach mostly returns tighter credible intervals, though it does not
track the March 2020 shock to unemployment. We expect that the paper
iteration of this paper will provide more guidance on what aspect of the
model's nonlinearity leads to it's impressive ability to outperform the GSC.
The prize for the best applied graduate student poster is shared between
Erin Rossiter and Luwei Ying.
Rossiter's poster, ``The Consequences of Interparty Conversation on Outparty
Affect and Stereotypes", examined why Americans increasingly dislike members
of the opposite political party, a vital issue as our democracy strains due
to increasing polarization. Erin combines two important innovations. First,
she implements a blocked cluster design that facilitates statistically
efficient analysis of experimental treatments. Second, she develops a chat
software that allows people to have real-time written conversations on-line.
This design allows her to analyze the effects of different types of
interactions. While those with no contact with members of the other
political party exhibit no change in their feelings toward the other party,
those who engage in a conversation exhibit clearly warmer feelings toward
members of the other party. The effects appear roughly similar for both
political and non-political conversations. These results provide new
evidence that interparty social interaction, regardless of whether the
conversation is politically-charged or not, can work to undo the negative
view of outparty members held by many Americans.
Ying's Poster entitled ``Religiosity and Secularism: A Text-as-Data Approach
to Recover Jihadist Groups' Rhetorical Strategies" tests the hypothesis that
as Jihadist groups become stronger their rhetoric moves along a continuum
from religious to secular, while it moves back again when events turn
against the Jihadis. To operationalize this Luwei gathered an impressive
multilingual corpus spanning decades of Jihadi literature, and hand coded
it, with separate filters for religious and secular vocabulary in each
language.
Ying shows that the log odds of using secular vocabulary does indeed rise
and fall with variables measuring the political success of the Jihadis.
Luwei goes on to provide confirmatory analysis using twitter.
The prize for the best faculty poster goes to Jay Goodliffe, for his poster
``Using Latent Transition Analysis to Explain Donor Behavior", in which he
analyzes why citizens start and stop donating to campaigns. This is an
important question because most donors give only occasional small amounts.
Jay develops a sophisticated latent transition model to identify patterns of
giving. One pattern, for example, is donors who give only to presidential
candidates, while another pattern is a donor who contributes medium-size
donations to out-of-state congressional candidates. Jay's approach allows
him to analyze transitions from one pattern to another finding, for example,
that donating in a presidential election does not generally lead people to
become donors in midterms. Goodliffe's work combines rigorous statistical
techniques with informative visualizations to help us better understand an
important political phenomenon.
We thank the authors of all 83 of the posters in this extremely strong field
for the opportunity to learn about their work!
------
Suzanna Linn
Liberal Arts Professor of Political Science
President of the Society for Political Methodology, www.polmeth.org
<http://www.polmeth.org/>
Penn State University
Department of Political Science
Pond Lab
University Park, PA 16802
Office: 320 Pond
Fax 814-863-8979
**************************************************************
Political Methodology E-Mail List
Editor: Erin Rossiter <[log in to unmask]>
**************************************************************
Send messages to [log in to unmask]
To join the list, cancel your subscription, or modify
your subscription settings visit:
https://www.cambridge.org/core/membership/spm/mailing-list
**************************************************************
|
