The issue of climate change is controversial among policymakers and the public. Many opt for the wait-and-see approach, hesitant to take any drastic actions toward mitigation and response. Researchers are concerned that this cautious approach is a result of incorrect assumptions about how carbon dioxide (CO2) emissions relate to carbon dioxide concentration levels. In an effort to address these misconceptions, the authors of this paper developed and tested an interactive climate change computer simulation, where participants received continuous feedback about the results of their actions. They also looked at differences in learning between adults who had majored in science and technology (STEM) and those with non-STEM majors.
The study aimed to address two common misconceptions about climate change science. The first misconception is that a decrease in CO2 emissions directly and immediately leads to a decrease in atmospheric CO2 concentration. Though these two are certainly connected, the feedback loop is neither linear nor quick. The second misconception is that even if emissions are in excess of absorptions, CO2 concentrations can still stabilize (a violation of mass balance).
Two sets of variables were tested: (1) whether misconceptions decreased after study participants completed the computer-simulation “experience” condition versus the control “description” condition, and (2) whether misconceptions decreased for those with STEM backgrounds versus those with non-STEM backgrounds. The research study targeted 120 total participants, ages 18 to 55. Among the participants, 60 were randomly assigned to the experience condition and 60 to the description condition. Within each condition, 30 had STEM backgrounds and 30 had non-STEM backgrounds. In total, participants spent about 30 minutes in the activities.
The main difference between the experience and description conditions lay in whether the activity provided immediate feedback to the participants. The description condition was a paper-and-pencil CO2 stabilization activity. Participants were given some information about CO2 processes and historical CO2 concentrations over the last 150 years. Next, they were asked to draw an estimated line graph of future CO2 absorptions and emissions, corresponding to a given line graph of CO2 concentration from 2001 to 2100. In other words, they were asked to stabilize CO2 concentration at the given level for each of 100 years, based on absorptions and emissions.
The experience condition included an additional task, in which participants performed the same CO2 stabilization activity using the dynamic climate change simulator (DCCS). For every value of CO2 emission and absorption that participants chose per year, the DCCS responded with immediate feedback, showing the resulting effect on CO2 concentration levels. Following the simulation activity, participants in the experience condition also performed the baseline paper-and-pencil task. The overall CO2 stabilization activity, whether on paper or the computer, specified a goal of 938 GtC (gigatons of carbon), a CO2 concentration level that participants aimed to reach by 2100. All participants also explained their drawings and reasons after completing the tasks.
Generally, the authors found that participants—no matter what their background—tended to rely more on preconceived and incorrect notions of correlation and mass balance if they took part in the description condition. All participants who worked with the DCCS and received feedback on their decisions relied less on these assumptions, and also did better on the paper-and-pencil task that followed the simulation.
With regard to differences based on educational background, the results were interesting. Both sets of participants with STEM and non-STEM backgrounds relied on their flawed correlation heuristics in the description condition, but in the experience condition the STEM participants performed significantly better than those with non-STEM backgrounds. In looking at the concept of mass balance, the authors found that STEM participants generally did better than non-STEM participants in both description and experience conditions.
Based on their findings, the authors concluded that the computer simulator activity allowed participants to gain immediate and repeated feedback based on their decisions. This feedback corrected their errors and assumptions, leading them to better understand the relationships between CO2 emission, absorption, and concentration. The simulation learning experience also transferred to the paper-and-pencil task, particularly for those with STEM backgrounds.
The authors suggested that future research studies may focus on creating simulation tools that are effective for learners with STEM and non-STEM backgrounds, as well as designing dissimilar tasks that test knowledge transfer.
The Bottom Line
In communicating complex topics and relationships, feedback-based simulation activities do a better job of disrupting incorrect assumptions and teaching accurate cause-and-effect relationships than non-feedback-based tasks. Those with STEM backgrounds also perform significantly better in these experiential conditions, perhaps due to their prior experiences in forming connections and creating structures of comprehension.