The internet is filled with comical examples of this type of misconception. Some are deliberately over the top, while others are real mistakes made by different people of various levels of society. For instance, a popular example to demonstrate false causation assumptions is the idea that the gradual decline in the number of pirates during the late 18^th and 19^th centuries caused climate change.

The idea is simple: as the number of pirates has decreased over the last 130 or so years, the effects of global warming have increased. So, we obviously need to reach for our cutlasses and black flags if we want to save the planet. This relationship has even been turned into a nifty graph, so it must be correct, right?

Here’s another one, did you know that America’s cheese consumption correlates with UFO sightings in Wyoming? So, as more people across the country eat cheese, they’re either bringing extraterrestrials to one specific US state or they’re causing Wyomingites to hallucinate visitors from space.

Obviously, these two examples are deliberately ridiculous, but they illustrate the type of logic that can take place when looking at data and making connections between two variables that are not otherwise linked. The old adage “correlation does not equal causation” has probably crossed your mind by now, but this offers a challenge when we think about the scientific process. If this is true, how do scientists “prove” anything they’re investigating?

Inductive reasoning […] is often the root of all scientific knowledge.

The confusion here is that scientists do attempt to establish causal relationships, but the process relies on a methodology of rigorous, controlled experimentation combined with statistical methods and logical reasoning.

The combination of these elements produces the most likely explanations for how X affects Y, yet this type of relationship can be easier to identify in some scientific disciplines than others. For instance, it is easier to conduct controlled experiments for subjects like physics and chemistry, but this is not always the case for biology, medicine, psychology, or the social sciences due to the complexities of living organisms.

Making a case for possible causation 

When it comes to establishing causality (whether in science or anything else, really), researchers rely on a mix of approaches to make their case. A core element of this process involves inductive reasoning, whereby someone makes generalizations based on the evidence they have from specific observations. This is often the root of all scientific knowledge.

For example, a marine biologist studying coral reefs notices that certain reefs experience widespread bleaching. The biologist records data from multiple sites and finds that the bleaching seems to be more prevalent in areas where water temperatures are higher. This leads them to hypothesize that there may be a correlation between these two variables (coral bleaching and surrounding water temperatures).

Additional research indicates that regions with higher ocean temperatures seem to have more bleaching, whereas regions with stable temperatures have little to no bleaching. As such, the researcher can hypothesize that “if ocean temperatures rise beyond a specific threshold, coral bleaching will increase”. It’s simple, and most importantly, it’s testable.

For instance, the biologist could design a controlled experiment where coral is grown in tanks of water, some of which have their temperatures raised while others do not. If the corals in the warmer water bleach, then it adds strength to the original inductive conclusion. They could then test the response of different coral species to see if they react differently, or collect more long-term data on sea temperature and bleaching events. This additional information will help refine their understanding.

If any causal explanation is correct, then it should be possible to make accurate predictions based on it.

This is a very simplistic hypothetical, but it demonstrates how scientists approach establishing causation for a phenomenon, especially one predicated on the observation of initial correlations. However, even this would not be sufficient to ascribe causation of coral bleaching to rising ocean temperatures. The researcher would also have to consider other confounding factors, such as ocean pollution, ocean acidification, the presence of disease, or the intensity of sunlight. Thankfully, these factors can also be tested within the laboratory or through field experiments.

In addition, other researchers may pay closer attention to the coral itself by delving into the question of how rising temperatures disrupt its biology, perhaps by exploring how heat stress can destabilize coral-algae symbiosis, which can lead to their bleaching. This adds valuable weight to the effects heat has on coral by explaining how bleaching can occur, rather than simply relying on an observed correlation on its own.

At this point, the plot is thickening, but more can be done. Researchers could carry out natural experiments that investigate the relationship between naturally occurring temperature variations and coral bleaching, to see if temperature changes are a key factor. Such studies could compare coral reefs in places with different thermal histories or more localized temperature anomalies.

For instance, researchers could examine how historical variations in temperature have influenced coral in the past. This could indicate that corals from regions with a history of higher temperature variability have greater resilience to thermal stress compared to those in waters that have remained more stable across time.

If any causal explanation is correct, then it should be possible to make accurate predictions based on it. Scientific models are often tested by trying to predict future trends or experimental outcomes. So, in this instance, climate models forecasting future rises in global temperatures could be used to predict the extent of coral bleaching under different scenarios. Similarly, epidemiological techniques could be used to predict the outbreak of a bleaching event before it occurs.

Confidence is key

The examples above all represent existing research using different scientific methods to address the subject of coral bleaching, but they are by no means the only examples. There have been numerous studies, all contributing pieces of evidence that, when taken together, allow us to say yes, rising ocean temperatures are causing corals to bleach.

Ultimately, evidence is the key to ascribing a causal mechanism for any phenomenon. The more complementary evidence you have from different sources, the stronger your assertions can be. 

The need to understand how scientists approach causation is more significant than ever.

In the above example, the idea that rising water temperatures may cause coral bleaching is not reliant on the findings from a single study, but rather multiple studies on related aspects of the same problem, resulting in greater confidence. 

If further research produces conflicting evidence, then data can be examined by others who can try to replicate the findings. If the results cannot be replicated, then perhaps it was due to methodological errors, which leaves the dominant explanation intact (though it should be noted that replication errors can also indicate deeper complexities that need additional research). But if these results can be replicated, then it is time to rethink the situation.

Contrary to the views of some, this is actually a strength for science. It provides flexibility that allows ideas and explanations to be updated if better evidence becomes available. The history of science is filled with this iterative process of review and reassessment. At the end of the day, it’s all contingent and there is no absolute proof, but rather levels of confidence based on the available evidence. 

And even when uncertainties remain, the approach to investigation that has been developed and honed over centuries of collective inquiry by scientists, philosophers, and other scholars has provided us with a way to make informed decisions based on the most likely explanations, be it in the realm of medical treatment, climate policies, or things like engineering and infrastructure, and so on.

Science sceptics and conspiracy theorists the world over often misuse the idea of “uncertainty” to cast doubt on well-supported theories. Given the state of the current troubling political situation in the US, where known conspiracy theorists have been given prominent positions within the government, the need to understand how scientists approach causation is more significant than ever.