Partial measurement limits the cumulative evidence for behavioural theories of climate adaptation

When the ground shakes or a river swells, the safety of a community often depends on what individual people do next. They might reinforce their homes, pack emergency bags, buy insurance, or move to safer ground. These protective actions are not just personal choices; they are the frontline of climate adaptation. For decades, scientists have tried to understand why some people take these steps while others do not. To do this, they rely on behavioral theories—structured maps of the human mind that list the specific thoughts, feelings, and beliefs that drive a decision. A theory might suggest that a person acts because they fear a disaster, believe they can fix the problem, or feel pressure from their neighbors. If researchers can measure these mental factors accurately, they can design better policies to help everyone stay safe.

However, a new analysis reveals a significant crack in this foundation. While scientists have been busy applying these theories to study floods, earthquakes, and wildfires, they have not been measuring the theories consistently. A team of researchers from universities and research centers across the globe examined hundreds of studies to see how well these mental maps were being used in practice. They found that most researchers are only looking at a small, easy-to-measure part of the picture, while ignoring the complex factors that often explain why people fail to act. Because the measurements are so inconsistent, the vast amount of evidence gathered over the years cannot be reliably combined to form a clear guide for policy. The field has built a large library of books, but the pages are written in different languages, making it impossible to read the story as a whole.

The researchers began by gathering 222 scientific studies that used surveys to ask people about their risk perceptions and their protective behaviors. These studies covered a wide range of natural dangers, from sudden events like earthquakes and floods to slow-moving threats like droughts and heatwaves. The team focused on the most popular theories used in this field, particularly two major frameworks: one that looks at how people weigh the threat of a disaster against their ability to cope, and another that traces the step-by-step decisions people make when facing danger. They checked every study to see if the researchers had measured every single part of the theory they claimed to use.

The results showed a striking pattern of omission. Even within the most widely used theories, researchers typically measured only a fraction of the required components. For the theory that focuses on motivation and fear, only 7% of the studies measured all the necessary parts. For the theory that tracks the decision-making process, the number was even lower, at just 4%. Instead of measuring the full set of ideas, scientists tended to pick the constructs that were easiest to ask about in a survey. They frequently measured simple concepts like "how scary do you think this is?" or "do you feel capable of acting?" These are the mental building blocks that are easy to turn into a few survey questions.

In contrast, the researchers consistently skipped over the more difficult, yet critical, factors. They rarely measured the cost of taking action, such as the money or time required to retrofit a home. They often ignored the level of trust people have in government institutions, or the specific barriers that stop people from acting even when they are afraid and feel capable. They also missed the ways people might cope in unhelpful ways, such as denying the risk exists. These missing pieces are exactly the factors that explain why a person might understand the danger and feel able to help, yet still do nothing. By leaving these out, the studies provide an incomplete story, one that overemphasizes fear and ability while hiding the real-world obstacles that block action.

This selective measurement creates a deeper problem for the entire field. Because every study measures a different mix of ideas, their results cannot be compared or combined. If one study asks about the cost of action and another does not, their findings about why people act are fundamentally different, even if they use the same theory name. The analysis found that only about 29% of the studies provided enough detail about their survey questions and testing methods to be reused by other scientists. This means that for the majority of the research, the tools used to measure human behavior are lost to future researchers. The field has accumulated a massive amount of data, but it is fragmented into thousands of isolated pieces that do not fit together.

The researchers also discovered a paradox between popularity and quality. The two most popular theories, which appear in the majority of studies, were actually the worst documented. The studies that used these theories often failed to report how they tested their survey questions or whether the questions actually worked as intended. Conversely, some less common theories were measured with much higher care and detail. This suggests that the sheer volume of research on the popular theories has not translated into better understanding, because the quality of the measurement has not kept pace with the quantity of the studies. The dominance of these theories seems to be driven by how easy they are to apply, rather than how well they are understood.

To move forward, the authors propose a practical path to repair this broken evidence base. They have created a map showing exactly which tools already exist that can be reused immediately. For the most common theories, there are already high-quality survey instruments that measure all the necessary parts, but researchers are not using them. Instead, they are inventing new, ad-hoc questions for every single study. The researchers argue that the field should stop reinventing the wheel and start using these existing, validated tools. This would allow different studies to speak the same language, making it possible to combine their results into a powerful, cumulative body of knowledge.

Beyond reusing old tools, the analysis points to specific gaps that need new tools. There are currently no reliable, tested ways to measure certain crucial concepts like the cost of adaptation or institutional trust in the context of climate hazards. Developing these new measures requires a coordinated effort that goes beyond what individual researchers can do alone. The authors also call for a change in how studies are reported. Scientists must be required to state clearly not only what they measured, but also what they chose to leave out and why. This transparency would prevent the hidden biases that currently skew our understanding of human behavior.

The ultimate goal is to build a measurement infrastructure that is as robust as the theories themselves. Just as engineers need precise instruments to build a bridge, adaptation scientists need precise, consistent tools to understand human behavior. Without these tools, policies designed to protect communities may be based on a fragmented and incomplete picture of reality. By fixing the way they measure the human mind, scientists can finally turn a scattered collection of observations into a clear, reliable guide for helping people adapt to a changing climate. The path forward is not to discover new theories, but to measure the ones we already have with the care and consistency they deserve.