Analysing New Zealand Online Extremism

The document outlines a two-pronged research architecture designed to analyze New Zealand-specific online extremism and international extremist activity targeting New Zealand. It defines extremism based on 'social identity' and the pursuit of in-group supremacy, focusing on far-right, far-left, and Islamist ideologies. The methodology employs a combination of qualitative ethnographic research, API-based data collection, and Natural Language Processing (NLP) for thematic classification. The authors emphasize a strict ethical framework prioritizing anonymity, the use of public data only, and the production of aggregated outputs to balance individual privacy with public safety.

• The research defines extremism as the active pursuit of systemic political and societal change based on an ideology claiming the supremacy of one 'in-group' over 'out-groups' and the dehumanization of those out-groups.
• Consultations and ethnographic research identified three primary extremist ideologies with an established presence in New Zealand: far-right, far-left, and Islamist. While misogynistic narratives were present, the researchers did not find cohesive New Zealand online communities for phenomena such as Incel movements.
• The study utilized two distinct data pipelines: a 'New Zealand pipeline' to measure activity from domestic extremists and an 'international collection' to measure activity about New Zealand from extremists globally.
• For the New Zealand pipeline, the researchers developed an automated thematic classification system using NLP to categorize content into six themes: the economy, the environment, health, politics, race, and religion.
• The performance of the NLP classifiers varied by theme, with 'Religion' achieving the highest F-score (81.6%) and 'Race' the lowest (65.6%).
• The international pipeline processed over 26 million records, including 16 million from 4Chan's /pol/ board, using a geo-parsing architecture (incorporating the CLIFF tool and GeoNames database) to identify mentions of New Zealand locations with 96% precision.
• The research adhered to a strict ethical framework that prohibited the study of non-public spaces, such as password-protected areas or private Facebook groups, and ensured all outputs were aggregated and anonymous.
• The project was explicitly designed as a research effort to identify broad trends rather than a law enforcement investigation, though a referral process was established for credible, immediate threats to life or safety.

APA

Miller, C., Guhl, J., & Comerford, M. (2021). Analysing New Zealand Online Extremism. Institute for Strategic Dialogue. https://www.isdglobal.org/wp-content/uploads/2021/06/NZ-methodology-paper.pdf

Chicago

Miller, Carl, Jakob Guhl, and Milo Comerford. Analysing New Zealand Online Extremism. Institute for Strategic Dialogue, 2021. https://www.isdglobal.org/wp-content/uploads/2021/06/NZ-methodology-paper.pdf.

Wikipedia

{{cite report |last1=Miller |first1=Carl |last2=Guhl |first2=Jakob |last3=Comerford |first3=Milo |title=Analysing New Zealand Online Extremism |publisher=Institute for Strategic Dialogue |date=2021 |url=https://www.isdglobal.org/wp-content/uploads/2021/06/NZ-methodology-paper.pdf |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{miller2021analysing, author = {Miller, Carl and Guhl, Jakob and Comerford, Milo}, title = {{Analysing New Zealand Online Extremism}}, institution = {Institute for Strategic Dialogue}, year = {2021}, url = {https://www.isdglobal.org/wp-content/uploads/2021/06/NZ-methodology-paper.pdf}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }