by: Namoos Haider, Ariadna Sandoval & Westley Youngren
We sought to use Agent Based Modeling to examine how fear caused by co-rumination spreads and becomes extinct during varying levels of social distancing. In order to understand how fear spreads through social interactions, one must first understand the constructs of co-rumination and social contagion. Co-rumination refers to excessively and repetitively discussing one’s fears or worries with peers (Starr & Davila, 2009). Research has demonstrated that co-rumination can lead to the spread of fear, anxiety, or worry, from one individual to another (Boren, 2013). Much like co-rumination, the Social Contagion Theory also involves the transference of behaviors from one individual to another (Christakis & Fowler, 2012). Specifically, the Social Contagion Theory has demonstrated that social networks can directly impact individuals’ behaviors, with the size of effect being determined by the strength of attachment to the social network (Christakis & Fowler, 2012). Thus, Social Contagion Theory would infer the closer one is to a social network, the more likely their behaviors are to be influenced by the social network. In sum, both co-rumination and the Social Contagion Theory suggest behaviors such as fear, can spread through social interactions.
Another important construct to understand in regard to co-rumination and fear, is fear extinction. Fear extinction refers to the natural phenomenon where fear of a variable gradually reduces overtime if the variable is no longer threatening or becomes habituated too (Myers & Davis, 2007). For example, when a red light is paired with a shock, individuals begin to fear the red light; however, once the red light has been shown enough times in the absence of a shock the fear of the red light will naturally begin to reduce, and eventually become extinct. The concept of fear extinction can encompass many fears, such as fear of a disease or other worries and anxieties (Marks & Toben̂a, 1990).
Given the idea that co-rumination can lead to the spread of fear and considering our current real-world context (social distancing due to Covid19), we sought to explore how social distancing and fear extinction impact the spread of fear caused by co-rumination. We hypothesized that high levels of social distancing would lead to less spread of fear caused by co-rumination, whereas low to no levels of social distancing would lead to more spread of fear caused by co-rumination. Additionally, we hypothesized the speed of fear extinction would impact the spread of fear caused by co-rumination. In sum, our model used the current theories and real-world context to examine how fear spreads through co-rumination during times of social distancing, all while considering levels of fear extinction.
Agent properties
Our agents represent people within a social network and are all homogeneous. The model starts with one-hundred agents but can be altered to go as high as 280, in proportion to the University of Kansas student body (which is approximately 28,000). There are two types of agents; those afraid due to co-rumination will be represented in red while those not afraid will be represented by blue.
Interaction rules
The purpose of this model was to examine the impact of directly manipulated levels of social distancing along with fear extinction on the ability of an agent to co-ruminate and to what extent. Interaction rules are defined as set parameters when conducting an agent-based model. As well as represent the possible outcomes. Based upon the research model, the following interaction rules are contrived. Firstly, the agents would seemingly follow the beliefs and/or practices preached or demonstrated by the inner circle (close friends & family). This links back with the Social Contagion Theory principles that ties in with our proposed model. Furthermore, as the scenario of spread of disease worsens/ stricter social-distancing measures, it is more likely that agents would act in self-preservation and obey social-distancing practices more firmly-e.g., wearing masks/disinfecting methods. The increase in social-distancing would also in-turn cause fear-extinction. Thus, fear caused by co-rumination would decrease. Lastly, afraid agents (red) who co-ruminate will report significantly higher levels of fear (derived from rumination) on than other agents (blue) who do not co-ruminate.
NetLogo initial setup
There are four parameters to our model: number of people afraid (num-afraid), number of people (num-people), social distancing (social-distancing), and fear extinction (fear-extinction). The default values are determined by what model you are using. If using a higher level of social distancing the social-distancing variable will have a lower score, whereas if you are using a model with little to no social distancing the social-distancing variable will be higher. Another variable that can be manipulated is the fear-extinction variable, where higher speed scores infer people extinguish their fears more quickly, and vice versa. By manipulating the social distancing and fear extinction levels, one can examine how the interaction of fear and social distancing spreads throughout groups of agents. We also examined our model over a time of 28.8 in order to represent approximately a month of interactions. Thus, our NetLogo setup allows us to examine our previously proposed relationships.
Simulation procedure
When running the model, our plot reflects how long it takes for X amount of people to become afraid due to in-person rumination. For our model, X is determined by levels of social distancing and fear extinction. We will use both high and low levels of social distancing and fear extinction in order to simulate how fear caused by co-rumination spreads during times of high and low social distancing, while also considering the level of fear extinction. Thus, our model allows us to view the spread of fear caused by social interactions during high and low levels of social distancing.
Simulation Results
The results of our simulation demonstrate that an agent-based modeling approach is feasible and useful for exploring the impact of social-proximity behaviors/social-distancing on co-rumination derived fear outcomes. The combination of parameters of social-distancing and fear-extinction were essential in the outcome of experimental results showcased in Figures 1 & 2. As previously stated, we sought to explore how two models compare and contrast. The first model we examined looked at the spread of fear in environments with strict social distancing practices and moderate levels of fear extinction. This model portrayed a small and slow spread of fear, where over 28 days only 10 agents became afraid (as seen in Figure 1). Our second model sought to examine the spread of fear in environments with relaxed social distancing practices and low levels of fear extinction. This model portrayed a rapid and large spread of fear, where over 28 days 272 agents became afraid (as seen in Figure 2). Our results demonstrated that both social distancing and fear extinction impact the spread of fear caused by co-rumination.
Figure 1
Figure 2
Implications and future directions
As social distancing measures intensify, the spread of fear through co-rumination decreases. Our findings are consistent with existing science literature, such as the study previously mentioned by Christakis & Fowler (2012), meaning that a feeling of fear can also spread to others within a social network like stated in the social contagion theory. This suggests that expressing fear about a disease to one’s peers will spread fear throughout a community, especially when social distancing measures are loosened or not enforced. This implies the need for stricter social distancing measures to help prevent the spread of fear. Fear spreading could cause people to overreact in a way that is harmful to others. Bringing awareness of the fact that talking negatively or fearful to one’s peers has the possibility to spread to everyone if social distancing measures are not in place. Therefore, social distancing, especially now during a pandemic, could greatly benefit people from contracting both a disease and the feelings of fear. As a future direction, incorporating telephone or online co-rumination into our model would be more realistic in terms of social distancing. A good intervention would be to show people how fear can spread through co-rumination in both high and low social distancing measures to encourage people to take social distancing (in person & online) more seriously.
References
Boren, J. P. (2013). Co-rumination partially mediates the relationship between social support and
emotional exhaustion among graduate students. Communication Quarterly, 61(3), 253-267
Christakis, N. A., & Fowler, J. H. (2013). Social contagion theory: examining dynamic social
networks and human behavior. Statistics in medicine, 32(4), 556–577. https://doi.org/10.1002/sim.5408
Hankin, B. L., Stone, L., & Wright, P. A. (2010). Corumination, interpersonal stress generation,
and internalizing symptoms: accumulating effects and transactional influences in a multiwave study of adolescents. Retrieved from https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4031463/
Marks, I., & Toben̂a, A. (1990). Learning and unlearning fear: A clinical and evolutionary
perspective. Neuroscience & Biobehavioral Reviews, 14(4), 365-384
Myers, K. M., & Davis, M. (2007). Mechanisms of fear extinction. Molecular psychiatry, 12(2),
120-150
Starr, L. R., & Davila, J. (2009). Clarifying co-rumination: Associations with internalizing
symptoms and romantic involvement among adolescent girls. Journal of adolescence, 32(1), 19-37.