by Neil Auwarter
On frozen pond: the Carnegie Mellon study
As detailed in this issue’s companion article, Climate change in black and white: The power of binary framing, recent work in cognitive science suggests people are more impacted by climate information framed binarily — this or that — than by incremental data. The example used in the Carnegie Mellon University study is a lake that in our warming climate has increasingly failed to freeze over in winter. One group of participants was shown a graph charting the rising average winter temperature at the lake yearly since 1940. Another group was shown a graph charting only whether the lake froze over or not each winter. The key finding was that participants who saw the frozen lake chart were more likely to perceive climate change as abruptly altering the environment.
Why binary climate data gets our attention
The beauty of binary data framing appears to be that it follows the “Goldilocks” principle, conveying a “just right” amount of information. Individual weather events may provide too little information — often even for those who experience them — because people can normalize individual events as just “weather.” Conversely, while aggregating weather events over time can show they amount to climate change and not just weather, this data may provide too much information to readily grab the public’s attention. The volume and creep of incremental data can fail to register, as with the frog in the pot. This is illustrated by the frozen lake study, where the participants who received more detailed temperature information were actually less impacted. When it comes to getting our attention, sometimes less is more.
Advanced case of boiling frog syndrome. Artist KC Greene, used with permission.
Other examples of binary climate data
The key to binary framing is to pose an impactful climate related yes-or-no query, and then sort climate data into two corresponding sets based on that query. A good query may depict a climate tipping point, a striking image, or some other compelling fact or event. This framing focuses on whether a critical event occurred, not on detailing incremental shifts. Here are some examples:
Heat and Health: Average temperatures in the United States have increased almost 2° F since 1979. But a more gripping statistic might be to divide temperature into two sets divided by the 90° threshold, where the risk of adverse health effects markedly increases. This binary sorting of data reveals, for example, that the number of yearly 90°+ days in Austin, Texas rose from 99 to 151 — more than 50% — during this same period. To illustrate, see Figure 1, a standard U.S. temperature graph, and Figure 2, a binarily-framed Austin days-at-90+ F° graph.
Figure 1. Average U.S. temperature (°F) from 1979-2023
Figure 2. Days at 90°F or above in Austin TX from 1979-2024
Tropical storms: Climate scientists have recorded troves of data on the frequency and intensity of tropical storms. But this vast and detailed data set might be condensed into the query whether in a given year the United States suffered catastrophic flooding from a tropical storm. A simple graph would show only three occurred between 1900 and 1998 (Galveston, Great Mississippi, Agnes), compared to six in just the 25 years from 1999 to 2024 (Floyd, Allison, Katrina, Sandy, Harvey, Helene).
Wildfires and smoke: Reams of data show incremental increases in the number and severity of wildfires since 2000. But this data might be distilled binarily, for example by the number of Americans who did or did not experience exposure to dangerous smoke levels. This framing could show that over the decade ending in 2023 the number of Americans exposed to dangerous smoke (3x the EPA health threshold) increased 27-fold, and that in 2020 alone 25 million Americans were exposed to at least one day of dangerous smoke.
Striking a balance between an impactfulness and detail
Climate scientists will rightly shudder at the thought of their research being reduced to scant binary statements. Science by its nature relies on broad datasets, averaged measurements, and incremental signals to separate noise from trend and establish causation. Still, anyone who desires to penetrate the public consciousness should heed the fact that while science demands detail and patience, public engagement thrives on thresholds and vivid turning points. The challenge is to report climate events in a way that bridges these competing demands. Perhaps the ideal piece of public-facing climate reporting would lead with a striking binary factoid and then set out detailed data to support and expand on it. In this way climate communicators may achieve both accuracy and resonance.