BEGIN:VCALENDAR VERSION:2.0 PRODID:-//events.la.psu.edu//EN CALSCALE:GREGORIAN BEGIN:VTIMEZONE TZID:America/New_York BEGIN:STANDARD DTSTART:20201101T020000 TZOFFSETFROM:-0400 TZOFFSETTO:-0500 RRULE:FREQ=YEARLY;BYMONTH=11;BYDAY=1SU END:STANDARD BEGIN:DAYLIGHT DTSTART:20200308T020000 TZOFFSETFROM:-0500 TZOFFSETTO:-0400 RRULE:FREQ=YEARLY;BYMONTH=3;BYDAY=2SU END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT UID:3906-07dce6268942e787c3194a8c2398e7b7@events.la.psu.edu DTSTAMP:20260806T164343Z DTSTART;TZID=America/New_York:20220923T121500 DTEND;TZID=America/New_York:20220923T133000 SUMMARY:Political Science Colloquium: Yiqing Xu DESCRIPTION:\n&ldquo\;What To Do (and Not to Do) with Causal Panel Analy sis: Parallel Trends\, Heterogeneous Treatment Effects\, and Diagnostics &rdquo\;\n\nTwo-way fixed effects (TWFE) models are a popular choice for causal panel analysis in political science. Recent developments in the methodological literature repudiate the TWFE estimator in the presence o f heterogeneous treatment effects (HTE) and violations of the parallel t rends assumption\; to diagnose or alleviate these issues\, researchers h ave proposed a variety of new estimators and testing procedures. How muc h do these problems matter for empirical political science research and how do new proposals compare in practice? We answer these two questions by reviewing\, replicating\, and reanalyzing 34 papers recently publishe d in American Political Science Review\, American Journal of Political S cience\, and Journal of Politics that use observational panel data with binary treatments. Leveraging both visual and statistical tests\, we fin d that parallel trends violations are a much bigger threat than HTE and that four newly-proposed HTE-robust estimators yield qualitatively simil ar findings when a strong pretrend is not observed. Moreover\, underesti mation of uncertainties\, which leads to uncontrolled Type-I errors\, re mains a prominent issue in empirical panel analysis. Overall\, we find t hat claims in one-third of the studies we review are strongly supported by data. We provide recommendations for empirical researchers to improve practice.\n\nFor more details: https://events.la.psu.edu/event/plsc-col loquium_yiqing-xu/ X-ALT-DESC;FMTTYPE=text/html:

“What To Do (and Not to Do) with Causal Panel Analysis: Parallel Trends, Heteroge neous Treatment Effects, and Diagnostics”

Two-way fixed effe cts (TWFE) models are a popular choice for causal panel analysis in poli tical science. Recent developments in the methodological literature repu diate the TWFE estimator in the presence of heterogeneous treatment effe cts (HTE) and violations of the parallel trends assumption; to diagnose or alleviate these issues, researchers have proposed a variety of new es timators and testing procedures. How much do these problems matter for e mpirical political science research and how do new proposals compare in practice? We answer these two questions by reviewing, replicating, and r eanalyzing 34 papers recently published in American Political Science Re view, American Journal of Political Science, and Journal of Politics tha t use observational panel data with binary treatments. Leveraging both v isual and statistical tests, we find that parallel trends violations are a much bigger threat than HTE and that four newly-proposed HTE-robust e stimators yield qualitatively similar findings when a strong pretrend is not observed. Moreover, underestimation of uncertainties, which leads t o uncontrolled Type-I errors, remains a prominent issue in empirical pan el analysis. Overall, we find that claims in one-third of the studies we review are strongly supported by data. We provide recommendations for e mpirical researchers to improve practice.

For more details: https://e vents.la.psu.edu/event/plsc-colloquium_yiqing-xu/

LOCATION:302 Pond Laboratory END:VEVENT END:VCALENDAR