Thursday, November 7, 2024

Collider Bias in Empirical Authorized Analysis (HKLJ)

Knowledge Nonetheless Wants Principle: Collider Bias in Empirical Authorized Analysis
Benjamin Chen and Xiaohan Yin (PhD candidate)
Hong Kong Regulation Journal, Vol. 53, Half 3 of 2023, pp.1241 – 1258
Summary: Huge knowledge is characterised not solely by the quantity but additionally the varieties of data that may be created, saved, and processed. This explosion of information, accompanied by the capability to analyse them, has catalyzed giant n, quantitative approaches to the research of legislation and authorized establishments. However neither dimension nor high quality ensures the validity of causal inferences drawn from observational knowledge. For instance, though the inclusion of management variables will help isolate causal results, not all variables are good controls. Unhealthy controls will not be innocent and might create the impression of a causal relationship the place none exists. This spurious affiliation known as collider bias. We introduce the idea of collider bias and provides motivated examples of the way it can come up in empirical authorized analysis. The collection of good controls requires data and assumptions about causal buildings. Principle and area data are important for quantitative evaluation, even within the period of huge knowledge.

Please click on right here to view the complete article on SSRN.

Related Articles

LEAVE A REPLY

Please enter your comment!
Please enter your name here

Latest Articles