Correlation or causation?: climate change and increased violence

Few recent scientific studies have received as much media coverage as one published recently in Science, arguing for a causal link between climate warming and violence.  The range of media sources covering it is as impressive as the number:  not just general news media at all levels, in many parts of the world, and not just science journals, but also those with particular perspectives, among them religious (or, at least Christian) environmental, police-based, medical and business.  Why should this particular scientific knowledge claim within shared knowledge make such an impact?

It is no news that the most sensational news is unqualified news. “Lock your doors” begin some of the news studies—and go on to claim that violence will unquestionably increase. Further, of course, all climate news is big news. Yet the very fact that so much attention has been given to the study makes it useful for a little TOK scrutiny.  The story provides a fine example of some running ideas in TOK: differences between correlation and causation, methodology of areas of knowledge, types of justification offered for knowledge claims, caution over accepting results, and implications of accepting knowledge claims. 

Methodology and Justifications

Even ignoring the actual content of this study, this example of science-in-progress illustrates much about the methods by which new knowledge can be sought:

 1. Based on the original research of 60 other studies, this is a strong example of a metastudy.

 2. The fact that so many different areas of knowledge were involved in the study points to the importance of the interdisciplinary nature of much contemporary work. In this case the investigation drew from studies in climatology, archaeology, economics, political science, and psychology.

 3. The fact that the study draws on many different cultures, across a broad swath of history likewise points to the range of evidence any such study might and perhaps should involve:

“Researchers noted examples including increased domestic violence in India and Australia, assaults and murders in the United States and Tanzania, ethnic violence in Europe and South Asia, land invasions in Brazil, police violence in the Netherlands and civil conflicts throughout the tropics.”  (Climate Change and Violence, Huffington Post)

 The most fundamental knowledge question, naturally, is whether the study is convincing.

As any TOK teacher will point out, the source is important. Berkeley and Princeton, of course, are reputable universities and Science is considered one of the two most authoritative science journals. Still, reputable sources are not enough.

Equally important is the way in which this new knowledge meshes with current shared knowledge. Most journalists echo the authors in pointing out two ways in which the results “make sense” (using reason as a way of knowing, and a coherence check):

1. Climate change can produce crop failure with major social and economic conflict resulting.

2. Even temporary high temperatures seem to precipitate aggressive behavior.

Also significant is the way the study makes its claims, providing evidence and details.

“We found that a 1 standard deviation shift towards hotter conditions causes the likelihood of personal violence to rise 4 percent and intergroup conflict to rise 14 percent,” said another author, UC Berkeley’s Marshall Burke, a Ph.D. candidate in agricultural and resource economics.

Caution in Accepting Knowledge Claims

Does the study, therefore, actually provide “knowledge”?

First, the scientists themselves involved in the study (as reported in some media—but certainly not all) voice the kind of caution appropriate to scientists — and rare in pseudoscience or some media:

“But they could not say why this might be the case. More studies would be needed to confirm the results and explain why such a correlation might exist, they said.”  (Climate Change Linked to Violent Behaviour)

Second, some journalists quote external authorities sceptical about the study. Dr Halvard Buhaug from Oslo’s Peace Research Institute, is quoted for his distrust of reaching such conclusions on the basis of the current study:

“Even if you found that conflict, defined in a particular way, appeared to be associated with climate, if you applied a number of complementary measures – which you should do in order to determine the robustness of the apparent connection – then you would find, in almost all cases, the two were actually unrelated.”

Dr Buhaug is deeply sceptical of the latest comprehensive study – which focuses entirely on quantitative analysis that excludes other complex social factors.

Third, some journalists voice their own scepticism about the study, as does commentator Nafeez Ahmed:

“Missing from the analysis is the reality that climate change is always refracted through the complex socio-political, economic and cultural relations of different societies.”

Implications of Accepting Knowledge Claims

Used as an example, this study thus highlights several issues any TOK teacher will find surfacing whenever there is a claim to increase the pool of shared knowledge.  It also suggests the role of generally applicable critical analysis:  evaluate the sources, assess the knowledge claims themselves, and consider counter arguments.

In addition, though, this particular example raises some further issues. After all, as noted at the outset, this particular study is having a major impact. Thus, we might ask of any assertion:

1. What are the practical implications and applications?

2. Can a correlation have practical applications even if it is not proved to be a causation?

3. If a causation is clear but the mechanism of causation is not, can it still justify taking action?

4. If a correlation has not been proven how much should cost-benefit calculations underlie precautionary action?

In this particular case, do the potential benefits of limiting carbon emissions outweigh the potential economic drawbacks?

“The result is alarming,” said Burke [a graduate student from Berkeley studying the link between climate change and food security]. “However, if we get our act together and we mitigate future climate change…the effects will be much smaller.


Clearly, this recent study could be used in class as an illustration of numerous knowledge questions under discussion at different points in the TOK course.  Bringing in examples drawn from current media in this way emphasizes the constant relevance of TOK to the world around us.  This study could also provide material for class presentations, with students zeroing in on specific knowledge questions to investigate.  We can hope that our students will increasingly find in comfortable to apply their TOK awareness and skills to topics such as this one.


Selected References

Nafeez Ahmed, “Will climate change trigger endless war?”, Earth Insight blog, August 2, 2013 :

Fiona Harvey, “Climate Change Linked to Violent Behaviour”, The Guardian, Aug 2, 2013.

Mark Kinver, “Climate Shifts ‘not to blame’ for African civil wars”, BBC News, September 6, 2010.

Robin Wilkey, “Climate Change and Violence Linked, Breakthrough Study Finds”, Huffington Post, August 1, 2013)

“Climate Change Conflict: Warmer Temperatures Lead to Increased Violence, Study Says”, International Science Times, August 2, 2013.

No Comments Yet.

Leave a Reply

Your e-mail address will not be published. Required fields are marked *