Climate Change Representation in IPCC Reports and Wikipedia: A Comparative Analysis Through Natural Language Processing

Master’s Thesis : Climate Change Representation in IPCC Reports and Wikipedia: A Comparative Analysis Through Natural Language Processing

This thesis presents a comparative Natural Language Processing analysis of the
Intergovernmental Panel on Climate Change (IPCC) Working Group III Summaries for
Policymakers and corresponding versions of Wikipedia’s "Climate Change Mitigation" (CCM)
article. Using a comprehensive range of NLP techniques, including lexicometry, stylistic,
readability assessments, modality analysis, semantic similarity (Sentence BERT), topic
modelling (BERTopic), sentiment and emotion detection, and Named Entity Recognition, the
study explores how CCM is portrayed across these two sources between 1990 and 2022.
The findings show that IPCC SPMs consistently maintain a high level of technicality and
Wikipedia, while initially more accessible and focused on events and personalities, has
gradually aligned both semantically and stylistically with the IPCC. The analysis reveals no
evidence of deliberate bias in Wikipedia’s representation of climate change mitigation.
Instead, differences in framing and focus reflect its role as a publicly edited resource
intended for a general audience. The study concludes that Wikipedia plays an important and
evolving role in supporting public understanding of climate change, increasingly reflecting the
IPCC’s scientific assessments.