Human behaviour experimentation and participation in scientific activities in the wild

Cooperation is one of the behavioral traits that define human beings — and other complex systems — and that has allowed us to evolve. However, even nowadays, after years of scientific advances we are still trying to understand why some systems, and particularly humans, cooperate. Human behavior lab experiments based on social dilemmas modeled as behavioral games try to shed light on these unknowns.

One of the recent approaches in this vein has been the move of human behavior experimentation from laboratories to public spaces, where behaviors occur naturally, to study the main behavioral traits — cooperation, trust, reciprocity, risk aversion or collectivity sense —. Citizen science practices have provided a perfect framework to experiment in the wild and to promote the participation of people that are usually not involved in science.

This dissertation is focused on advancing the field of behavioral experimentation in the open environments by performing experiments based on citizen science practices. The work is divided in two blocs, one oriented to the design of experimental platforms and another focused on the analysis of experimental results. Particularly in the first, we study how to design citizen science platforms that allow scientific activities in the wild and we present two of them that are applied to two different contexts. In the second, we perform experimental studies of social systems in order to analyze behavioural traits, to look for the emergency of behavioral patterns and also to evaluate the designs of the platforms.

With the first platform we investigate how citizen science systems can serve as a catalyst for the promotion of scientific thinking and to increase engagement in science. We introduce Natural Patterns, a platform that presents scientific activities based on the scientific research method. We introduce the results of a user experience study and, in light of the contributions of the participants, we describe a series of design principles that increase the motivation of the participants and promote scientific participation in the wild in this specific context.

The second platform is designed to study traits of human behavior and to help creating experiments in the wild, where real social phenomena emerge, by encouraging the participation within the framework of citizen science. To that effect, the platform is very modular and includes a broad range suite of behavior games. It is light, which allows experimentation in places with little infrastructure — public spaces, festivals or conferences —, and scalable, allowing to introduce new social dilemmas and interactions. We evaluate participation, robustness and quality of the collected data in all the experiments that have been carried out so far.

With respect to the experimental studies, we present the first two human behavior experiments that were carried out with the platform. In the first, participants have to predict market movements under different circumstances. Specifically, we study whether the strategies that emerge are robust in the replicas of the experiments that were performed in different locations with very different sociodemographic samples. The second experiment studies the behavior patterns — or phenotypes — that emerge from the decisions that participants make when they face a set of social dilemmas. Concretely, we focus on the data analysis by unsupervised learning techniques and conclude that the behavior of the participants facing this set of social dilemmas with different tensions can be summarized in a total of five phenotypes.

The last two experiments presented in this dissertation also tackle and fully delve into human behavior experimentation. The first is a collective-risk dilemma in which a group of participants has to contribute their assets to overcome a collective damage, specifically climate change. We study how resource inequalities cause unfair behaviors, provoking the most vulnerable participants to be more harmed than the most favored ones.

Finally, the last experiment is a collective-risk dilemma, but in this case there are no economic inequalities between the participants. Unlike the rest of experiments, this one is carried out within a given collective: the mental health ecosystem, formed mainly by people affected by mental disorders, caregivers and relatives. We study the tensions that exist when all of them collaborate collectively to solve a collective uncertain bad. From the results of the experiment it is emphasized that the cost of collective actions falls on those affected by mental conditions.

To conclude, the platforms that we have built help us to carry out participatory scientific activities within the framework of citizen science, which has plenty of applications, from learning to activism. This is especially so in the case of the experimentation platform that studies human behavior in vivo, which makes it possible to access a representative sample of the population and to integrate the participants in the research process. The platform will help set the basis for future behavioral experiments in the wild. In the light of the social experiments' results we can comprehend how we behave collectively, in society or within a certain group, when we face social dilemmas, and consequently evaluate behavioral traits and the emergence of behavioral patterns thanks to unsupervised machine learning techniques. These insights offer new ways to study the tensions that exist in a given community around a particular social issue.

This doctoral dissertation was performed under the supervision of Dra. Mercè Gisbert and Dr. Jordi Duch, and in collaboration with Dr. Anxo Sánchez, Dr. Josep Perelló or Dr. Haoqi Zhang, among others. This work was defended in Universitat Rovira i Virgili, on May 18, 2018, in the presence of the Supervisory Committee integrated by Dra. Marisa Ponti, Dr. Sergio Gómez and Dr. Frederic Guerrero. It was awarded an outstanding "Excellent with Honors, Cum Laude", as well as obtaining the International Mention and the Industrial Doctorate Mention.

Resources

Vicens, J. (2018). «Human Behavior Experimentation and Participation in Scientific Activities in the Wild»

Download Thesis