Lessons learned about the development of digital entertainment tools for experiments on resources distribution

Highlights

• Games are presented as tools to researches on resources distribution.

• Lessons learned during exploration, ideation, prototyping and development are listed.

• Recommendations are presented, especially for academic teams.

• Examples illustrate the translation of dilemmas into children oriented games.

Abstract

Resource Distribution is a topic from the field of distributive judgement, generally studied by cognitive psychology researchers, which make use of economic experiments to investigate aspects of the evolution of cooperation, altruism and selfishness in humanity. Researchers commonly set experiments to observe human behavior, seeking to increase the degree of engagement of the participants, in order to improve the quality and the amount of results. Digital entertainment artifacts, such as video games, are then used as a viable way of bringing such a high level of engagement. However, psychology researchers usually do not have the knowledge needed to develop digital applications. In addition, as far as we know, no scientific work have tried to describe the impact of using digital entertainment tools in experiments on decision making for resource division. This work aims to report a compilation of discoveries and knowledge learned when designing digital entertainment tools, tailored for scientific experiments on Resources Distribution. The task included interviews with specialists, a questionnaire sent to researchers, the conduction of ideation techniques to generate new concepts and, finally, a set of tools were prototyped and a subset was ultimately developed. Results were collected and six main lessons we learned are reported. We hope these lessons learned to be helpful to guide others researchers to conceive and develop their own digital entertainment tools for scientific experimentation.

Introduction

The advance of our understanding on the human decision making process has a long tradition in Psychology studies. Of particular interested are the topics related to the behavior of people when sharing valuable goods. This can be related to issues like inequality between rich and poor or even to causes of wars, but also to everyday situations such as how to split the bill in a restaurant with friends. Research on this subject by psychologists started by using hypothetical situations as a method to investigate moral judgments in children (Piaget, 1994). Hypothetical scenarios are still very common, however different results can be found when experiments inforce real consequences to participants, such as actually giving valuable goods (Blake, McAuliffe, & Warneken, 2014).

Giving can be seen as a type of human cooperation whose multidisciplinary nature is much discussed these days. From other perspectives, researchers such as Axelrod (2010) try to understand how cooperation evolved even though humans tend to be self-interested. Studies on distributive judgment and behavior are relevant because they help to understand how people act in social interactions that are important to survival. Researchers use social dilemma experiments based on Game Theory (Alencar & Yamamoto, 2009) to understand the dynamics and the evolution of human behavior when there are situations where it is necessary to cooperate or compete for valuable resources. Experiments are based whether on economic games like the Ultimatum Game (Bearden, 2001), the Dictator Game (Kahneman, Knetsch, & Thaler, 1986) and the Public Goods Game (Hauert, 2005) or on dilemmas like the Prisoner’s Dilemma (Salen & Zimmerman, 2012) and the Stag Hunt (Skyrms, 2004).

These Game Theory approaches can rely on human participants but also in abstract strategies conceived to maximize results (Axelrod, 2010). In experiments involving people not always participants are committed to good performance during tests (Kahneman, Daniel, Knetsch & Thaler, 1986). In order to overcome this problem some researchers use playful elements which bring fun to the participants and hence a more engaging experiments. According to Washburn (2003), the first psychological study in which computer games were mentioned was in 1976. Besides the lack of engagement problem, other factors can compromise a study, such as scalability, accuracy, stability, flexibility and ease of use of the research tools.

This work is surrounding the Game Design Thinking study. The Game Design Thinking subject is wide and browses in the intersection of game design (as a science) (Hunicke, LeBlanc, & Zubek, 2004), behavior design (Fogg, 2016), Engagement Design (Grace & Farley, 2016), and neuroscience (chemicals that effect our brains). This work draws on game mechanics, narrative design, psychology and economics to find ways that we can answer questions like: How to develop digital entertainment research tools for a scientific experiment on Resource Distribution (RD)? Which are the shortcuts and pitfalls in conceiving and implementing them, especially with an academic team, as usual in most researches? There are several game development methods, strategies, libraries, authoring tools, and engines. They vary according to the needs of each research. These needs can be economic (such as the cost and effort needed), technical (such which features are necessary to develop a particular application), or related to the nature of the research. Therefore, what is the complexity involved in creating this new tool? Researchers on RD are commonly from cognitive or social psychology and usually do not have a consistent software development experience. This paper aims to explore these questions through lessons learned during a set of experiments designed and developed by a multidisciplinary team.

The idea of creating a guide for using instruments for scientific research is not recent, even though first references focused on non-digital tools (Wilkinson & Birmingham, 2003). As far as digital tools are concerned, there are different approaches to data collection and analysis (Bryman, 2015, Paulus et al., 2013, Xenitidou and Gilbert, 2009). The tools already reported in scientific literature can be conceptually categorized as interactive digital narratives, also known as storytelling, frameworks to develop economic games, virtual reality environments and video games.

Since Piaget (1994) narratives are used in experiments about distributive justice, especially with children. In experiments conducted by Tschudy, Dykstra-Erickson, and Holloway (1996) narratives are told in a graphic way through cartoon cards as cameras to capture participants’ responses. Similarly, researchers report (Kirova & Emme, 2008) the use of photos by children to create comic books in order to translate thoughts and capture their social experiences. Therefore, storytelling could be an effective means of doing experiments in which the participant may intervene in the story flow while it is told (Thorpe, 2004). When they are digital, narratives can be interactive and collect data, still being very simple to understand (from the user’s point of view) and to develop (from the researcher point of view). In fact, the simplicity is its strength, and the digital forms of storytelling may allow long-term experiments in a very scalable manner.

These frameworks are authoring tools intended to facilitate the creation of games as research instruments. It is especially used in game theory investigations and in the economy and psychology fields. These frameworks allow researchers to configure experiments in which they can test hypotheses about the behavior of individuals, simulate the behavior of a population using artificial intelligence scripts and study the interaction between virtual actors and real people. Thus, these frameworks integrate a diverse set of digital mechanisms in a single engine. One example is the Colored Trails framework, which consists of a parametrized digital game. This framework allows specifying different reward structures, enabling the examination of trade-offs in bilateral economic experiments like the Prisoner’s Dilemma, Stag Hunt and the Ultimatum Game (De Jong, Hennes, Tuyls, & Gal, 2011). Another important tool is the INVITE Framework (Baptista et al., 2013) which allows researchers to set up experiments such as the Public Good Game, the Prisoner’s Dilemma, the Stag Hunt Game and more complex situations involving cooperative behaviors. It is also possible to simulate behavior of robots with artificial intelligence and study interactions between virtual agents and real participants. The framework provides a digital game to run the experiments that has a 3D interactive environment. The authors argue that, being a 3D environment, it becomes the most immerse experiment possible, bringing the necessary engagement of the participants.

3D virtual environments with social interaction, such as the social networks Second Life (Linden Research, 2016), ActiveWorlds (ActiveWorlds, 2016) and the digital game World of Warcraft (Blizzard Entertainment, 2016) have been regularly used as tools for scientific experiments. People simulated in this environment are avatars, supposedly expressing the identity of their human owners. The author (Bainbridge, 2007) argues that any avatar is subjectively a second “I”, then the reputation starts to become important for the owner even if his actions in the virtual environment cannot be tracked in the real world. Therefore, data captured in these virtual environments can be relevant for scientific studies. The article “Virtual worlds - past, present, and future: New directions in social computing” (Messinger et al., 2009) makes a detailed study, creating a taxonomy and listing the virtual environments with social network bias. ActiveWorlds consists of a 3D social virtual environment, created in 1995, supports voice communication, modification through scripts and commands and navigation via the web browser. Works such as “Using the multi-user Virtual Environment to Research Approaches to Ethical Dilemmas” (Cram, Dick, Gosper, & Hedberg, 2009) report using that system for experiments with ethical dilemmas. Already the Second Life is a more recent project, created in 2003, in 2013 obtained the mark of 1 million regular users (Linden Research, 2013). Second Life is widely used for scientific research precisely because it offers an interaction with other participants or with virtual agent, has the concept of virtual property and you can enter changes by scripts.

Virtual reality technology consists of a system that allows the viewer to interact “naturally” with objects and other individuals in an experience similar to the “real world”. In mid-1965, Ivan Sutherland (Loomis, Blascovich, & Beall, 1999) already designed and built a virtual screen system that encouraged the feeling of immersion in virtual environments. The researcher argued than that virtual reality immersion technologies are applicable for the study of basic knowledge in psychology as they provide ecological validity without compromising experimental control. After decades of improvements, this kind of technology recently became commercially available. For example, the Oculus Rift (Oculus, 2016) is a virtual reality device, to which manufacturers have provided development kits. Researchers have begun to test them as research tool to study decision-making. For example, Pereswetoff-Morath (2014) proposes to assess the immersion factor and how participants felt during critical decision-making situations in social dilemmas. They were divided in two groups, one using game with Oculus Rift and other using the same game and situation, but without using the device, and results were compared.

According to a study by ESA (2013), 58% of Americans play video games. According to ISFE (MediaCT, 2012), a survey involving 16 European countries found that 25% of individuals who like to play video games play at least once a week and for 36% of people the most common word associated with game is “fun’’. The use of digital games as a research tool is not new. According to Washburn (2003), the first psychological study in which computer games are mentioned was in 1976. Since then, games became part of day-to-day life. There are several types of games that can be used as “Serious Games” (a term assigned to games with scientific purpose), or as research tools.

Games are often reported to bring people to a flow state (Csikszentmihalyi and Csikszentmihalyi, 1992) (Rogers, 2012), thus games are frequently seen as a good candidate to bring engagement and help participants to endure longer experiments. The use of digital games facilitates the extraction of dynamic information (behavior over time) as well as detailed information by clipping (as a precision gain). Peysakhovich and Rand (2014) reported the evolution of behavioral conduct of some players in the online game League of Legends (Riot Games, 2016). The researchers studied the players’ behavior dynamics, analyzing the evolution of virtual society and conducting controlled experiments to obtain information about groups of individuals who had not inappropriate behavior within the game.