Quantitative Social Sciences
Title : | Advancing research methods in causal inference using latent factor models |
Area of research : | Quantitative Social Sciences |
Focus area : | Quantitative Social Sciences |
Principal Investigator : | Prof. Souvik Banerjee, Indian Institute Of Technology Bombay (IITB), Maharashtra |
Timeline Start Year : | 2023 |
Timeline End Year : | 2026 |
Contact info : | banerjee.souvik@iitb.ac.in |
Details
Executive Summary : | In this study, I seek to estimate the causal treatment effect of an endogenous latent continuous variable (e.g. depression) on multiple outcome measurements (e.g. labour market outcomes), whereby the latent treatment variable is generated from varied indicators (e.g. symptoms of depression, and anxiety disorders) and underlying causes (e.g. demographic and socio-economic factors). In addition, a latent (unobserved) factor will be included as an independent variable in the multiple outcomes equations and the endogenous treatment equation. The significant advantage of this modeling approach is that one is able to potentially address the endogeneity of treatment effect in two ways: (i) ameliorating omitted variable bias through the shared latent factor, and (ii) reducing measurement error in the latent continuous treatment variable by utilizing multiple observed indicators of the latent treatment variable. |
Total Budget (INR): | 6,60,000 |
Organizations involved