Dataset Information

Causal simulation experiments: Lessons from bias amplification.

ABSTRACT: Recent theoretical work in causal inference has explored an important class of variables which, when conditioned on, may further amplify existing unmeasured confounding bias (bias amplification). Despite this theoretical work, existing simulations of bias amplification in clinical settings have suggested bias amplification may not be as important in many practical cases as suggested in the theoretical literature. We resolve this tension by using tools from the semi-parametric regression literature leading to a general characterization in terms of the geometry of OLS estimators which allows us to extend current results to a larger class of DAGs, functional forms, and distributional assumptions. We further use these results to understand the limitations of current simulation approaches and to propose a new framework for performing causal simulation experiments to compare estimators. We then evaluate the challenges and benefits of extending this simulation approach to the context of a real clinical data set with a binary treatment, laying the groundwork for a principled approach to sensitivity analysis for bias amplification in the presence of unmeasured confounding.

SUBMITTER: Stokes T

PROVIDER: S-EPMC8721560 | biostudies-literature | 2022 Jan

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Causal simulation experiments: Lessons from bias amplification.

Stokes Tyrel T Steele Russell R Shrier Ian I

Statistical methods in medical research 20211123 1

Recent theoretical work in causal inference has explored an important class of variables which, when conditioned on, may further amplify existing unmeasured confounding bias (bias amplification). Despite this theoretical work, existing simulations of bias amplification in clinical settings have suggested bias amplification may not be as important in many practical cases as suggested in the theoretical literature. We resolve this tension by using tools from the semi-parametric regression literatu ...[more]

PMID: 34812681

Similar Datasets

Project description:Herbicides are detected year-round in marine waters, including those of the World Heritage listed Great Barrier Reef (GBR). The few previous studies that have investigated herbicide persistence in seawater generally reported half-lives in the order of months, and several studies were too short to detect significant degradation. Here we investigated the persistence of eight herbicides commonly detected in the GBR or its catchments in standard OECD simulation flask experiments, but with the aim to mimic natural conditions similar to those found on the GBR (i.e., relatively low herbicide concentrations, typical temperatures, light and microbial communities). Very little degradation was recorded over the standard 60 d period (Experiment 1) so a second experiment was extended to 365 d. Half-lives of PSII herbicides ametryn, atrazine, diuron, hexazinone and tebuthiuron were consistently greater than a year, indicating high persistence. The detection of atrazine and diuron metabolites and longer persistence in mercuric chloride-treated seawater confirmed that biodegradation contributed to the breakdown of herbicides. The shortest half-life recorded was 88 d for growth-regulating herbicide 2,4-D at 31°C in the dark, while the fatty acid-inhibitor metolachlor exhibited a minimum half-life of 281 d. The presence of moderate light and elevated temperatures affected the persistence of most of the herbicides; however, the scale and direction of the differences were not predictable and were likely due to changes in microbial community composition. The persistence estimates here represent some of the first appropriate data for application in risk assessments for herbicide exposure in tropical marine systems. The long persistence of herbicides identified in the present study helps explain detection of herbicides in nearshore waters of the GBR year round. Little degradation of these herbicides would be expected during the wet season with runoff and associated flood plumes transporting a high proportion of the original herbicide from rivers into the GBR lagoon.

Dataset Information

Causal simulation experiments: Lessons from bias amplification.

Publications

Causal simulation experiments: Lessons from bias amplification.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets