Back to search

University of British Columbia

Surfacing social norms in language with machine learning to better plan for sustainable transitions

Abstract

dc:description

Facilitating a successful just and sustainable transition requires buy-in and support from affected publics, but often planning efforts are met with a slow uptake of new behaviors or direct opposition to policy changes. This pushback or apathy is not always rooted in objective accounting of the outcomes, but from a perceived threat to a place’s identity and the social norms that sustain it. At the same time, noisy information environments present a complex challenge for planners–competing narratives and “information overwhelm” complicate the development and implementation of equitable planning solutions. Further, the volume of data generated on a daily basis that comprises these environments is too large for traditional qualitative methods to handle. In response, the goals of this research are two-fold: 1) to explore the role social norm communication plays in upholding or shifting place identity in response to sustainable transitions and 2) to explore ways in which machine learning techniques can aid in processing the scale of information required to study norms. Leveraging advances in natural language processing, this dissertation relies primarily on open-source, large-scale text data to surface normative communication across multiple transition domains, source types, and communication aspects. Towards the first goal, I find that norms communicated at scale by groups with perceived expertise reflect core aspects of identities for a place, and these can be at odds with necessary infrastructure development. Towards the second goal, I find that fully automated methods of topic modeling struggle to capture nuance and context, while methods explicitly informed by qualitative thematic coding methods can keep pace with added complexity while providing efficient processing. I also show how this additional human interaction early on in the modeling pipeline can serve as a checkpoint to reduce unintended consequences that can stem from historical, cognitive, and statistical biases. The findings from this dissertation highlight the need for planners to both account for the impact of entrenched norms on uptake of new behaviors or support for policies as well as entrenched historical inequities that can be encoded into large-scale data and models.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy - PhD
Level thesis:degree_level
doctoral
Discipline thesis:degree_discipline
Planning
Grantor dc:publisher
University of British Columbia
Year dc:date
2025

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Lore, Madison
Contributors dc:contributor
  • Caggiano, Holly

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/2429/91716

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

Lore, Madison. Surfacing social norms in language with machine learning to better plan for sustainable transitions. doctoral thesis, University of British Columbia, 2025. http://hdl.handle.net/2429/91716