Hoppa direkt till innehållet

Information till studenter och medarbetare med anledning av covid-19 (Uppdaterad: 18 januari 2021)

printicon

Statistiskt Seminarium

tis
3
nov
Tid Tisdag 3 november, 2020 kl. 01:15 - 02:15
Plats Zoom

When studying causal relationships, balancing observed covariates is a critical part of the process to unbiasedly estimate a causal effect. For this, weighting techniques are often used by scientists of different fields, one of the most common techniques being inverse propensity score weighting. When applying propensity score methods, specifying the model to be used is one of the main challenges and the resulting weights usually does not target neither balance nor stability. In contrast, the recently proposed stable balancing weights directly target covariate imbalances and allow the researcher to explicitly set the desired balance constraints. In a time-to-event outcome setting, common in medical and epidemiological studies, the finite sample properties of the marginal hazard ratio estimator based on stable balancing weights are evaluated through Monte Carlo simulations. The results are compared to inverse propensity score weighting methods under diverse data generation settings.

 

Join Zoom Meeting using the link:

https://umu.zoom.us/j/64121073388?pwd=QnJWN3ZIQ0JaRmJCb1VVK3dRZTlKUT09

Meeting ID: 641 2107 3388

Passcode: 690004

Please join zoom 5-10 min in advance to make sure it works for you. Please keep your computer muted during the talk. Questions will be organized at the end.

 

.

Evenemangstyp: Seminarium
Personalbild saknas
Talare
Guilherme Barros
Doktorand
Läs om Guilherme Barros
Kontaktperson
Patrik Rydén
Läs om Patrik Rydén