Canada Spends Logo

Explore our Budget 2026 analysis

Grant: Discovery Grants Program - Individual

University of Toronto

$50,000.00

Statistical Methods for Causal Inference from Large-scale Observational Studies

FY 2024-25

Summary
In the modern data landscape, extracting causal insights from large-scale datasets is imperative. A pervasive challenge in this realm is unmeasured confounding, which arises when unobserved variables simultaneously influence both the cause and the effect in a study, leading to skewed findings. For example, there is a correlation between the number of Nobel Prize winners per capita in countries and higher chocolate consumption. However, it is unlikely that one directly causes the other. Rather, an underlying factor, such as a nation's economic well-being, could be influencing both. In practice, modern observational datasets are typically marked by high dimensionality and complex dependence structures. While these attributes are often seen as challenges, my approach pivots towards innovation. I propose to develop methods that not only address but capitalize on these features to robustly tackle the enduring issue of unmeasured confounding. In the long term, my research aims to develop a suite of computationally efficient, theoretically grounded, and freely available analytical tools for drawing causal inference from large-scale observational datasets. This proposal develops a research program consisting of multiple interrelated themes, focusing on the fundamental challenge of unmeasured confounding. Specific aims include: (1) statistical methods for estimating the effect of a single exposure on a single outcome, while accounting for high-dimensional covariates; (2) statistical methods for estimating the effect of multiple exposures on a common outcome of interest; (3) statistical methods for estimating the effect of a single exposure on multiple outcomes of interest; (4) statistical methods for estimating the causal effects on a network. While the primary contributions from this work will be in causal inference, specifically from a statistical point of view, the proposed methods will facilitate reliable analysis of large-scale observational data that are prevalent in many scientific disciplines. Additionally, this project will provide opportunities for four PhD students, two MSc students, and three undergraduate students in statistical sciences to become actively involved in the scientific process and receive training in both foundational and applied statistics research. Methods developed through this grant will be coded in R and made publicly available along with publication of related research reports.
Program
Discovery Grants Program - Individual
Location
Ontario, Canada
Recipient Type
Academia
Type
Grant
Source
NSERC's Awards Data

As published

Application id
RGPIN-2024-04069
Name nom
Wang, Linbo
Department département
Statistical Sciences
Organization id
31
Institution établissement
University of Toronto
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2024
Competition year année de concours
2024
Award amount
50000
Program id
RGPIN
Program name en
Discovery Grants Program - Individual
Program name fr
Programme de subventions à la découverte - individuelles
Group en
DISCOVERY RESEARCH
Group fr
RECHERCHE AXÉE SUR LA DÉCOUVERTE
Committee code
1508
Committee name en
Mathematics and Statistics
Committee name fr
Mathématiques et statistique
Area of application code
1100
Area of application group en
Health, education and social services
Area of application group fr
Santé publique, éducation et services sociaux
Area of application en
Health, education and social services
Area of application fr
Santé publique, éducation et services sociaux
Research subject code
3011
Research subject group en
Statistics and probability
Research subject group fr
Statistique et probabilité
Research subject en
Multivariate analysis
Research subject fr
Analyse multivariée
Field of research list names en
Not available
Field of research list names fr
Non disponible
Socioeconomic objective list names en
Not available
Socioeconomic objective list names fr
Non disponible
Application title
Statistical Methods for Causal Inference from Large-scale Observational Studies
Keywords
causal inference; high-dimensional statistics; machine learning; unmeasured confounding; semi-parametric statistics; variable selection; biostatistics; missing data
Application summary
In the modern data landscape, extracting causal insights from large-scale datasets is imperative. A pervasive challenge in this realm is unmeasured confounding, which arises when unobserved variables simultaneously influence both the cause and the effect in a study, leading to skewed findings. For example, there is a correlation between the number of Nobel Prize winners per capita in countries and higher chocolate consumption. However, it is unlikely that one directly causes the other. Rather, an underlying factor, such as a nation's economic well-being, could be influencing both. In practice, modern observational datasets are typically marked by high dimensionality and complex dependence structures. While these attributes are often seen as challenges, my approach pivots towards innovation. I propose to develop methods that not only address but capitalize on these features to robustly tackle the enduring issue of unmeasured confounding. In the long term, my research aims to develop a suite of computationally efficient, theoretically grounded, and freely available analytical tools for drawing causal inference from large-scale observational datasets. This proposal develops a research program consisting of multiple interrelated themes, focusing on the fundamental challenge of unmeasured confounding. Specific aims include: (1) statistical methods for estimating the effect of a single exposure on a single outcome, while accounting for high-dimensional covariates; (2) statistical methods for estimating the effect of multiple exposures on a common outcome of interest; (3) statistical methods for estimating the effect of a single exposure on multiple outcomes of interest; (4) statistical methods for estimating the causal effects on a network. While the primary contributions from this work will be in causal inference, specifically from a statistical point of view, the proposed methods will facilitate reliable analysis of large-scale observational data that are prevalent in many scientific disciplines. Additionally, this project will provide opportunities for four PhD students, two MSc students, and three undergraduate students in statistical sciences to become actively involved in the scientific process and receive training in both foundational and applied statistics research. Methods developed through this grant will be coded in R and made publicly available along with publication of related research reports.