Classification of contrarian claims about climate change

The Argument Against Everything.

Image from here.

This recent paper examines the language used to argue against climate change. The authors developed a model to detect specific contrarian claims around climate change. To do so, they built a comprehensive taxonomy of contrarian claims. The process resulted in five major categories:

  1. It’s not happening.
  2. It’s not us.
  3. It’s not bad.
  4. Solutions won’t work.
  5. Climate science/scientists are unreliable.

Below is an example of how the sub-claims fell under the categories.

This figure displays the three layers of claim-making by climate change contrarian actors.

Next, the authors built a deep learning model to classify specific contrarian claims (Methods). From the paper,

The state-of-the-art pre-trained Transformer Language Model RoBERTA was employed to train another classifier using the Simple Transformers software package. RoBERTa is an optimized version of the popular BERT language model.

To examine the history of climate change contrarianism, the authors built a data set of 250+ thousand documents from conservative think tanks and contrarian blogs that span the last 20 years. This model was applied to the documents and annotated. The below shows the prevalence of the claims for the last 20 years.

So, the next step is very cool. Now, that the authors know who is talking about what topics. We can examine the funding of these think tanks and blogs and show what organizations are pushing what narrative.

Linear regression results show that the proportion of category 5 and category 1–3 claims are positively associated with the proportion of funding originating from these 10 key donors. Likewise, we find a negative association of category 4 claim prevalence with key donor funding. Figure 4d illustrates the sources of funding for 14 CTTs in our sample. Notably, prominent contrarian CTTs such as the Heartland Institute are heavily dependent upon these key donors and, in particular the “donor-advised” funding flows from Donors Trust and Donors Capital Fund, which ensure anonymous funding to conservative causes.




I like science, machine learning, technology, and start-ups.

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Hypothesis Testing | All a beginner needs to know

Favorite Query Service Tricks in Adobe Experience Platform (Part 4)

Using analytics to understand increasing motor accidents in India

COVID Update: Fire Suppression System Market is projected to reach a value of over USD 39.2

Introduction to Product Data Science & Analytics

Using R to Analyze Sports Data

RFM Customer Segmentation

QC Sampling - From a Business Perspective

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Angela Wilkins

Angela Wilkins

I like science, machine learning, technology, and start-ups.

More from Medium


AI machines as moral agents, Avoiding definitions and the Turing test. (part 8)

The Raven Paradox

Shoe store shelves

AI & Law: Legal Echo Chambers