Back to search

University of British Columbia

Evaluating counterfactuals : case studies of modeling choices and their implications

Abstract

dc:description

Individuals and organizations often turn to counterfactual questions for actionable insight. The answers provided to such questions can have important consequences, therefore it is critical to examine their reliability. In this dissertation, I demonstrate that examining the reliability of these answers forces us to closely evaluate the choices and assumptions that are embedded in the lifecycle of a counterfactual query -- from its construction to its application. I argue that such an evaluation must proceed in a context-sensitive manner, as it must take into account the intended aim of pursuing counterfactual questions, and also the context wherein the answers to such questions will be used. I highlight the different manners in which our choices and assumptions shape the epistemic and ethical reliability of the answers we provide to counterfactual queries by examining three case studies: (1) the selectivity of imagination-driven counterfactual thought; (2) the use of agent-based models for simulating counterfactuals, specifically in the context of investigating the potential impact of diversity on group performance; and (3) the use of machine learning models for answering counterfactual queries, and the use of counterfactual metrics for assessing the ethical reliability of those models. I argue that making choices and assumptions is inevitable, and that the reliability of our dealings with possibilities depends on being transparent about the influence of these assumptions and choices. Such transparency enhances the epistemic quality of the endeavour wherein counterfactuals are put to use by allowing us to explicitly evaluate when a given choice is justified and why. It also improves our capability to formulate informed mitigation strategies in the face of epistemic and ethical complications.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy - PhD
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Philosophy
Grantor dc:publisher
University of British Columbia
Year dc:date
2019

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Fazelpour, Sina

Rights

dc:rights
Statement dc:rights
  • Attribution-NonCommercial-NoDerivatives 4.0 International
Language dc:language
eng

Identifiers

dc:identifier.*
Handle dc:identifier
http://hdl.handle.net/2429/71779

Chain of custody

source
Harvested from
University of British Columbia
Base URL
circle.library.ubc.ca/oai/request
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
related terms
citation

Fazelpour, Sina. Evaluating counterfactuals : case studies of modeling choices and their implications. doctoral thesis, University of British Columbia, 2019. http://hdl.handle.net/2429/71779