Safely simulating risks in a pandemic
Study tests behavioral interventions in a gaming experiment
Is it possible to conduct experiments about risk-taking during pandemics without endangering participants? To safely investigate the effectiveness of public health interventions and simulate the dynamics of virus transmission, researchers from the University of Plymouth, UK, the IESE Business School in Spain and the Max Planck Institute for Human Development have developed a game-based experiment. Their findings have been published in the journal Science Advances.
The world has relied on adherence to non-pharmaceutical interventions, such as ventilation, mask-wearing and physical distancing, to increase our safety throughout the ongoing pandemic. With vaccines completing the toolbox, these measures and the accompanying public health messaging continue to play an important role. Scientific evidence shows that physical distancing and mask wearing help limit the spread of the virus. However, these measures pit individual self-interests – the wish for a social life, the discomfort of wearing a mask – against the common good.
An international study from Germany, the United Kingdom, and Spain has shown that it is possible to test the effectiveness of interventions designed to foster safer behavior in a pandemic. A total of 700 participants from the US played variants of an online game in groups of 100 players each that simulated the transmission of a virus. The researchers designed a neutral framework without the mentioning of masks, mandates, or medical terminology. They developed a game that captured three key elements of a pandemic: the difficult to predict potential for exponential growth, the social dilemma between individual benefits and collective risk, and the compounding of small transmission risks.
In the game, 100 players start out blue (healthy), and then 8 randomly selected players turn purple (infected) in an initial outbreak. From the start, no one knows their color as they move forward. In each of 25 rounds of the game, players then choose between two actions: G (the safer alternative, representing mask wearing or distancing, for a possible reward of 8 points) and H (for the riskier alternative with a possible reward of 40 points). The potential rewards stand for consequences for personal well-being or self-interest.
Players who are randomly paired with a purple player then have a chance of turning purple themselves, with a transmission probability ranging between 0.05 and 0.25. This means that on average, between one and five encounters out of 20 lead to transmission. The specific probability depends solely on which actions the two players chose for the round (G or H). After 25 rounds, only the players who remained blue (or “healthy”) throughout the game receive monetary rewards. When more players are “infected” overall, fewer receive payouts. The payout size increases with the number of accumulated points.
The study tested the effectiveness of five interventions in preventing risk-taking behavior against a condition without intervention. The intervention that included a message with a moral component was the most effective. The simulation of multiple different games or the graphical depiction of the potentially wide-ranging consequences of early infections were also successful. In contrast, presenting information on expected case numbers in the current round was not effective: participants did not seem able to anticipate the exponential growth in transmissions. Perhaps surprisingly, communication that focused on the actions of others (e.g., describing how often others were wearing masks) even prompted greater risk-taking and led to negative consequences in the experiment. One should note that the neutral framing prevented this outcome from spilling over into real-world health consequences.
The results show that nothing is as effective as a clear rule combined with compelling moral reasons for following it. The results also suggest some will always opt to risk themselves and others in pursuit of potential personal gain. That said there is also a significant swing group that can be dissuaded from taking risks, given the appropriate intervention. Knowledge gained from observing behavior in the simulation experiment is applicable to real-life conditions and result in slowing transmission and reducing the burden on health systems.
“Nonpharmaceutical interventions – such as wearing masks, maintaining physical distance, and reducing contacts – require behavior change on a large scale. The behavioral sciences offer tools to foster individual compliance, but the relative effectiveness of methods has rarely been tested in controlled, experimental scenarios that still reflect the dynamics of infectious outbreaks. What is so important about this framework is that it allows testing interventions safely before implementing them with potential health consequences for participants,” explained Jan K. Woike, Associate Research Scientist in the Center for Adaptive Rationality at the Max Planck Institute for Human Development.
“The next COVID-variant or next pandemic may arrive in the future. Policy-makers need to know which interventions are most effective in promoting socially advantageous behavior. This research project is a step forward to answer this question without endangering participants.”