A recent investigation by Pulitzer Prize winner Inside Climate News has uncovered damning activity by fossil fuel company Exxon. Long before they supplied millions of dollars to conservative think-tanks who misinformed the public about climate science, Exxon’s own scientists informed them of the scientific consensus that fossil fuel burning would cause disruptive climate change.
This echoes past activity of the tobacco industry, who knew from internal research about the health consequences of smoking but nevertheless funded misinformation casting doubt on the link between smoking and cancer. The same misinformation tactics employed by the tobacco industry are used by the fossil fuel industry.
Even the same spokespeople defending tobacco have also attacked the science on climate change. Given the obvious parallels between the activities of the tobacco and fossil fuel industries, the New York Attorney General has issued a subpoena further investigating Exxon’s activities regarding climate change.
Measuring Scientific Consensus
Research published in 2010 by Uri Shwed and Peter Bearman offers further insight into one aspect of the tobacco and fossil fuel misinformation campaigns. Their study was concerned with the question of scientific consensus – how do we know if and when a consensus forms?
The authors looked at how scientific papers reference other papers, known as citations, in order to construct a network of scientific research. Scientists mostly cite papers they agree with. Consequently, a network of papers where there is scientific consensus will look quite different to a network where there is still ongoing debate.
When the scientific community is undecided about an issue, the published literature will show different “communities” of papers citing each other. In network terms, the community shows higher modularity (in other words, it’s “clumpy”). As scientific consensus forms on an issue, the structure of the community evolves from distinct groups into a single, united community (and the “clumpiness” smooths out).
With the assumption (based on empirical research) that papers mostly cite papers they agree with, Shwed and Bearman took a purely mathematical approach to quantifying consensus, without having to read the content of all the published research. This eliminated the need for domain experts to manually categorise scientific research, as well as removed any possible bias from people who are manually analysing the content.
They looked at the evolving consensus on several scientific issues but we’re interested in two in particular: the consensus linking smoking to cancer and the consensus linking human activity to climate change.
The consensus on climate change followed a “spiral” trajectory where scientific agreement on the major question of human causation was initially settled in the early 1990s. From that point, scientific investigation narrowed (or spiralled) onto more specific questions (for example, the behaviour of clouds in a changing climate or better understanding of regional climate change).
Measure of scientific disagreement about climate change Shwed & Bearman, 2010
This result matches what we found in our analysis of climate papers from 1991 to 2011. There was already a strong consensus that humans were causing global warming in the early 1990s that only strengthened over time.
In contrast, scientific research into the link between smoking and cancer took a cyclical trajectory. Initial evidence linking smoking to cancer emerged in the late 1950s. However, the debate in the published research ebbed and flowed for decades before scientific consensus eventually solidified.
Measure of scientific disagreement on smoking causing cancer Shwed & Bearman
Comparing Misinformation Tactics
The differing evolution of consensus between these two issues tells us about one difference between the misinformation campaigns waged by the tobacco and fossil fuel industries.
The tobacco industry cooperated closely to fund publishable, but biased research, prolonging the debate in the scientific literature for several decades. It is easy to design smoking studies to find no significant effects, and hard to prove errors in their work, so they were able to inject many such papers into the literature.
It is much harder to do that with climate science, given public measurements, conservation laws of physics, and well-established theory. Consequently, the fossil fuel industry placed more emphasis on PR campaigns through think-tanks and political leaders.
Our own analysis of 21 years of published climate research found that peer-reviewed research rejecting human-caused global warming has had a negligible presence in the scientific literature over the last few decades. The science consensus had already formed, but of course the public consensus had not.
Unfortunately, both approaches have effectively delayed action for decades. Both had excellent lobbyists. Both employed front groups and for-hire think tanks as those grew in the 1990s. Both used whatever media were available, starting with print and radio, and now the Internet. Both funded public misinformation campaigns long after their own scientists had discovered the truth.
Public antipathy to the tobacco industry increased strongly as it became clear that the industry knew about links to cancer decades earlier. It will be interesting to see if similar revelations about the fossil fuel industry yield similar changes in the public viewpoint.