BSU Virtual Seminar: 'Using auxiliary variables in mediation analysis to address unmeasured confounding'
Duration: 56 mins 36 secs
Share this media item:
Embed this media item:
Embed this media item:
About this item
| Description: |
Speaker: Professor Richard Emsley, NIHR Research Professor, Institute of Psychiatry, Psychology and Neuroscience, King’s College London
Title: 'Using auxiliary variables in mediation analysis to address unmeasured confounding' |
|---|
| Created: | 2020-11-10 15:13 |
|---|---|
| Collection: | BSU Virtual Seminars 2020 |
| Publisher: | University of Cambridge |
| Copyright: | A.S. Quenault |
| Language: | eng (English) |
| Distribution: |
World
(downloadable)
|
| Keywords: | biostatistics; seminar; data science; |
| Explicit content: | No |
| Aspect Ratio: | 4:3 |
| Screencast: | Yes |
| Bumper: | UCS Default |
| Trailer: | UCS Default |
| Abstract: | Causal mediation analysis aims to estimate natural direct and natural indirect effects under clearly specified assumptions. Traditional mediation analysis based on ordinary least squares (OLS) estimation assumes an absence of unmeasured causes to the putative mediator and outcome, even in randomised studies. When these assumptions cannot be justified, instrumental variables (IV) estimators can be used to produce an asymptotically unbiased estimator of the mediator-outcome link. However, this bias removal comes at the cost of variance inflation and requires additional parametric assumptions about the validity of the instruments. I will describe the statistical approaches for identification and estimation of parametric regression and instrumental variable analyses, including examples of multiple instruments. I will introduce a new estimator that strikes a natural trade-off between the unbiasedness of the IV procedure and the relatively small variance of the OLS estimator. |
|---|---|