How to engage different audiences with the same graph
TLDR: Despite the scientific consensus on climate change, there is a growing partisan divide in the support for climate policies over the last decade within the United States. Given the same climate evidence, why do some people become concerned while others remain unconvinced? Here we argue that people look at different parts of the same graphical evidence of global temperature depending on their prior motivations and beliefs, therefore arriving at different conclusions and actions. We further demonstrate that drawing attention to specific parts of the graph that are consistent with people’s prior motivations can encourage desired actions.
How do people interpret graphs? In a perfect world, everyone will view the same graph and come to the same conclusion. In the human world, our interpretations are often colored by our prior beliefs and motivations. A key challenge in data visualization, therefore, is figuring out how to present the same data to draw different audiences with diverse motivations and backgrounds.
My PhD student Yu Luo and I decided to tackle this challenge using a climate change dataset from NASA. The scientific evidence for climate change has been unequivocal: 97% of actively publishing climate scientists agree that human activities are causing global warming. But climate change is one of the most contentious topics in the public mind, polarizing along party lines in America. A poll in 2006 showed that 79% of democrats versus 59% of republicans say there is a solid evidence that the average temperature on Earth has been getting warmer. This divide has not only endured, but widened over time. In 2017, 92% of democrats versus 52% of republicans say there is solid evidence that the average temperature on earth has been getting warmer. Such growing divide has significant implications for setting policy agendas. For example, 77% of democrats versus 36% of republicans in 2017 say stricter environmental laws and regulations are worth the cost.
What is driving the partisan divide?
Have democrats and republicans been viewing different kinds of evidence on climate change? No. When the same climate evidence is presented by NASA or IPCC, why do political disagreements continue to grow? To answer this question, we proposed a new framework that suggests that people’s political motivations shape their visual attention to climate change evidence, which influences their perception of the evidence and subsequent actions to mitigate climate change. These altered perceptions and actions can reinforce their initial motivations, further entrenching the divide. To put simply, what you believe in determines what you see, which guides your future actions.
How do political motivations influence attention to climate change evidence?
We set out to examine the first research question: How do people with different political motivations view the same climate change evidence? To answer this question, we conducted an experiment in the lab where we presented participants with a graph showing the global temperature change from 1880 to 2013.
Participants were instructed to look at the graph and estimate the average global temperature change. We tracked participants’ eye gaze when they were viewing the graph using an eyetracker. Eyetracking allowed us to measure where on the graph people were paying attention. After viewing the graph, participants provided a numerical estimate of the average global temperature change from 1880 to 2013, and also rated their political orientation on a scale from -5 (very liberal) to 5 (very conservative).
We found that the more liberal people were, the more attention they paid to the rising phase of the temperature curve (1990 to 2013) relative to the flat phase of the curve (1940 to 1980). Liberals who focused more on the rising phase of the curve relative to the flat phase provided higher estimates of global temperature change, but this is not the case for conservatives. This is the first time that we evidenced varying attentional biases with political orientation.
How to track attention online? The BubbleView technique
Since the first experiment was conducted with undergraduate students in the lab, we wondered if the results would apply to a broader sample more representative of the public. Thus, our next research question is: How do liberals and conservatives from the US public view the same temperature graph? In the second experiment, we implemented the BubbleView technique to track people’s attention on an online platform. Instead of tracking their eye gaze, we tracked their mouse location. In this technique, the graph was first covered by a black mask, only a small circular area around the mouse was transparent. Participants had to move their mouse to reveal the underlying graph. You can try it below!
https://yuluo.psych.ubc.ca/studies/spotlight/spotlight_2/estigraph.php
We again found that liberals who focused more on the rising phase relative to the flat phase tended to provide a higher estimation of global temperature. This result not only replicated our previous finding in the lab, but also validated our BubbleView technique. So it seems that the more liberal you are, the more attention you pay to the rising phase of the curve which is consistent with your prior beliefs of climate change, and the higher global temperature you conclude at the end.
How does drawing attention to motivationally consistent evidence influence subsequent actions?
The previous experiments established a correlation between political orientation and visual attention, but was this relationship causal? To seek causal evidence, we manipulated attention by highlighting different parts of the curve as a simple visualization technique. In other words, our research question in this experiment is: How does drawing attention to motivationally consistent evidence influence subsequent actions on climate change? In this experiment, we manipulated attention by coloring different parts of the temperature curve in red to deliberately bias attention to stronger or weaker evidence of climate change. Specifically, we highlighted either the rising phase of the curve in red (i.e., rising red condition), or the flat phase of the curve in red (i.e., flat red condition). As a manipulation check, we found that people in the rising red condition looked more at the rising phase, and participants in the flat red condition looked more at the flat phase. We have successfully directed people’s attention to different parts of the graph!
Now the important question is how biasing attention influences people’s actions to mitigate climate change. After viewing the graph, participants were asked whether they were willing to sign a petition to stand with the Nature Conservancy to call on U.S. leaders to stand strong on climate change, and whether they were willing to donate to Natural Resource Defense Council.
We found that liberals were more likely to sign the petition or donate when the rising phase was highlighted in red than when the flat phase was highlighted. In other words, when attention was drawn to motivationally consistent evidence, people were more likely to act. In contrast, conservatives were less likely to sign the petition or donate when the rising phase was highlighted than when the flat phase was highlighted. This shows that when attention was drawn to motivationally inconsistent evidence, people were less likely to act. In summary, drawing attention to evidence in the graph that is consistent with people’s prior motivations can encourage desired actions.
What are the implications?
Our findings have several implications for climate change communication that uses data visualization to engage the public. First, providing climate change evidence alone is likely to be insufficient since people pay attention differently depending on their prior motivations. Second, we cannot use the same communication strategy for liberals and conservatives. Climate visualization needs to align with audience’s ideological motivations. Aside from highlighting specific parts of the graph, another approach can frame climate change consistently with people’s values, such as framing mitigation efforts as promoting a warmer society and economic or technological development. Or, we can provide information on peer group norms to shift attention, since people may hold incorrect beliefs of how their peers view a controversial issue. Tailored visualization can increase the efficacy of communication, and also bring an awareness in consumers of the potential misuse of these techniques.
Take home message
When visualizing scientific data on a controversial tropic, we should emphasize different parts of the data to align with audience’s prior motivations, to increase the chance of successfully conveying the message and reaching the desired impact. Also, the new online attention-tracking tool, BubbleView, is free! This means that researchers interested in visual attention can save costs and carbon from a conventional eyetracker.
