Investigating the simulation elements of environment and control: Extending the Uncanny Valley Theory to simulations

Recent technological advancements have enabled the widespread application of simulations in organizations, particularly for training contexts. Two important simulation elements, environment and control, have often been shown to improve trainee outcomes. I argue that environment and control are reliant on each other, and their combined effects are explained by extending the Uncanny Valley Theory. The Uncanny Valley Theory proposes that individuals are comfortable with experiences that are very dissimilar or similar to reality, but are uncomfortable with experiences that fall between these conditions. In simulations, perceptions of realism are created through observations (environment) and interactions (control). Users are comfortable with experiences when these elements are in agreement; however, an Uncanny Valley effect may occur when these elements are in disagreement. In the current article, two studies analyze the realism of environment and control in predicting trainee reactions and learning outcomes. Both studies support the extension of the Uncanny Valley Theory to simulations. Simulations with only “low” or only “high” environment and control produce the greatest outcomes, and those with mixed “low” and “high” elements produce the worst outcomes; however, trainees did not differ in reactions to the simulations, indicating that the Uncanny Valley phenomenon in simulations may operate subconsciously. © 2017 Elsevier Ltd. All rights reserved. The popularity of digital simulations has greatly expanded with the advent of powerful computing technologies, and researchers have repeatedly applied simulations for training and education purposes (Kapp, 2012; Pasin & Giroux, 2011; Sitzmann, 2011). In 2012, the simulation training market generated $2.4 billion in revenue, and this figure is expected to grow to $6.7 billion in 2017 (Adkins, 2013). Further, many business schools continue to use simulations for MBA education purposes, providing effective and uniform instruction (Avramenko, 2012; Ben-Zvi, 2010). Given the widespread nature of simulations, several authors have devised methods to improve student and trainee2 reactions, learning, and transfer. A recent development is the integration of video game elements into simulations, spurred by the serious game and the gamification movements (Arnab et al., 2013; Botella et al., 2011). Business, University of South Alabama, Mobile, AL 36688, United States. a for their feedback on ideas that contributed to the manuscript. t to both students and trainees, but only trainees are referenced henceforth to improve the flow and Fig. 1. Visual representation of the uncanny valley. M.C. Howard / Computers & Education 109 (2017) 216e232 217 These movements propose that adding game elements to training programs may increase trainee enjoyment and motivation, and several typologies of game elements exist to identify (a) the individual elements, (b) their effects, and (c) relevant theories (Bedwell, Pavlas, Heyne, Lazzara,& Salas, 2012; Djaouti, Alvarez, Jessel,& Rampnoux, 2011; Marne, Wisdom, HuynhKim-Bang, & Labat, 2012; Wilson et al., 2009). Among the most popular typologies is Bedwell et al. (2012), presenting nineteen separate serious game elements determined through a literature review and empirical study. Despite the existence of many game elements, the current article performs a focused investigation into two: environment and control. Environment is the user's digital surroundings. In general, small and/or simple environments are considered “low” amounts of environment, whereas detailed and/or expansive environments are considered “high” amounts of environment. Alternatively, control is the permitted amount of user interaction with these surroundings. Basic forms of control, such as observation alone, represents “low” amounts of control, whereas advanced forms of control, such as physical movement and object manipulation, represents “high” amounts of control. Together, the nature of environment and control can greatly vary from simulation to simulation. Further, these two elements possess certain properties that are dissimilar to other elements, causing them to be extremely important for research and practice. Many game elements only appear in certain typologies, raising concern about their existence as a substantive element; however, every game element typology includes environment and control, providing large support for their existence and potential impact in simulation programs (Bedwell et al., 2012; Djaouti et al., 2011; Marne et al., 2012; Wilson et al., 2009). Also, environment and control are required elements of all simulations. Whereas other game elements, such as mystery, may not be included at all, consideration must always be given to environment and control, and any inferences about these two elements would impact all simulations. Lastly, as I propose, environment and control are reliant on each other to improve learner outcomes, and their combined effects may be explained by the Uncanny Valley Theory (Gray & Wegner, 2012; Mori, MacDorman, & Kageki, 2012; Saygin, Chaminade, Ishiguro, Driver, & Frith, 2011). The Uncanny Valley Theory suggests that a positive relationship exists between animatronics’ humanness and observer favorability; however, a severe “valley” in observer favorability appears slightly before animatronics become indistinguishable from humans, resulting in feelings of uneasiness and disgust (Fig. 1). Although this theory has largely been applied to animatronics, I argue that it is integral to many other human-computer interactions e particularly simulations. In extending the Uncanny Valley Theory, I propose that trainees have certain expectations of simulations’ realism, largely created through observations (environment) and interactions (control). When these elements are in agreement (i.e. both “low” or both “high”), then users are comfortable with their experiences; however, when the elements are in disagreement (i.e. one “low” and one “high”), an Uncanny Valley effect may occur, resulting in burdens to cognitive resources and reduced learning outcomes. This paper reports two studies that aimed to test these claims andwhether these effects occur consciously or unconsciously. The current investigation has many implications for simulations and other human-computer interactions. First, although many authors have created game element typologies, these typologies are relatively new and largely theoretical (Bedwell et al., 2012; Djaouti et al., 2011; Marne et al., 2012; Wilson et al., 2009). Research is needed to validate the importance of individual elements, and this paper contributes towards filling this research gap. Second, many analyses of game elements compare multiple simulations together, each with varying aspects of multiple game elements. In these cases, the most effective simulation is identified, but the exact effective elements are largely unknown. Authors have called for more focused empirical investigations to understand particular elements (Bedwell et al., 2012; Landers & Callan, 2011), and this research heeds this call. Third, authors have likewise called for more sophisticated integration of theory into serious game research to understand the dynamics of each element (Marne et al., 2012; Wilson et al., 2009). The studies reported in this paper M.C. Howard / Computers & Education 109 (2017) 216e232 218 investigate the mechanisms that cause the joint influence of environment and control, furthering the understanding of these elements. Fourth and lastly, authors have called for new theories to be applied in simulation research, as many investigations are atheoretical or often apply the same theories (Djaouti et al., 2011; Hatala, Cook, Zendejas, Hamstra, & Brydges, 2014). Through extending the Uncanny Valley Theory, this paper presents a new theoretical perspective that can be applied in all human-computer interaction research. Therefore, the current article has many implications for research and our understanding of simulations and human-computer interactions.