Research Questions

  • Can a serious game help undergraduate students achieve conceptual change about molecular emergence?

  • What role do the specific game mechanics play above and beyond the underlying interactive simulation?
  • How do players’ actions in the game world reflect their understanding?


Gauthier, A., & Jenkinson, J. (2016). Serious Game Facilitates Conceptual Change About Molecular Emergence Through Productive Negativity (RCT). In Connolly and Boyle (Eds.), Proceedings of the 10th European Conference on Games Based Learning (pp. 844–852). Paisley, Scotland: Academic Conferences and Publishing International Limited. (PDF)

Gauthier, A., & Jenkinson, J. (2015). Game Design for Transforming and Assessing Undergraduates’ Understanding of Molecular Emergence (Pilot). In R. Munkvold & L. Kolås (Eds.), Proceedings of the 9th European Conference on Games Based Learning (pp. 656–663). Steinkjer, Norway: Academic Conferences and Publishing International Limited. (PDF)


  • We performed a randomized controlled trial with 40 students, split randomly into control and gaming groups, who used the game or control simulation for 30 minutes
  • We compared misconceptions in these two groups to a baseline group of 486 students
  • It was found that both interventions facilitated conceptual change above and beyond the baseline, and that a trend suggested that the game worked better than the control
  • The game mechanics encouraged a greater number of productively negative events, the quality of which may be associated with lower misconceptions.
  • Game score was not related with misconception held by the student, suggesting that the intervention time frame, while allowing the stimulus to be effective, was not sufficiently long to permit replay of levels, thus capturing the students’ new understanding.

Serious game facilitates conceptual change about molecular emergence through productive negativity


Andrea Gauthier (PhD Student)

Jodie Jenkinson (Supervisor)

Cindi Morshead (Co-supervisor)

Nick Woolridge (Committee member)

Igor Jurisica (Committee member)


Undergraduate biology students lack an understanding of the emergent nature of molecular processes and systems and often think about molecular interactions in a teleological manner.

A game has potential to transform students’ misconceptions through immersion in an interactive environment and through challenges that encourage cycles of productive negativity — the player attempts a challenge and fails under their current misconception of directed molecular motion, then must reevaluate their understanding in order to complete the challenge and progress in the game. The purpose of this study was to test the pedagogical effectiveness our serious (i.e. educational) game, MolWolds, in facilitating conceptual change about molecular emergent systems through productive negativity. We were particularly interested in how game mechanics (e.g. resource management, immersed character, sequential level progression, and points systems) influenced interactions within the simulation and how this may have led to differential transformative experiences.

To do this, we compared interactions and learning outcomes from MolWorlds to an interactive simulation without game mechanics, MolSandbox, that was designed to be as similar as possible to the game in all other ways. We performed a randomized controlled trial with these two stimuli.


Participants were first-, second- and third-year biology undergraduate students. At the beginning of the semester, 526 students completed our Adaptive Molecular Concepts Survey and demographics questionnaire. Towards the end of the semester, 40 of these students volunteered to participate in our randomized controlled trial wherein they played with either the game or control app for 30 minutes. During this time, we screencast their interactions in their assigned app. All students then completed the survey a second time at the end of the semester.

Analysis and results


We performed a repeated measures mixed model ANOVA to determine the effects of intervention type (i.e. serious game, control simulation, or no intervention) and educational level (i.e. first-, second-, or third-year biology student) on the change in misconceptions from pre-test to post-test.

We found that intervention type had a significant effect (p < .001) on conceptual change from pre- to post-test. Both the control (p = .003) and game (p < .001) stimuli were significantly better than standard education (our baseline group). Furthermore, the game was trending toward being more efficacious than the control (p = .084).

Productive negativity and its relationship to misconceptions

We also coded all 40 30-minute screencasts for demonstrations of correct conceptual knowledge and instances of productive negativity. A demonstration of correct conceptual knowledge was identified as a series of actions wherein the user makes appropriate adjustments to the underlying simulation that helps him/her achieve the level’s goal. An instance of productive negativity was identified as some sort of negativity (i.e. delay of progress, level failure, frustration) that then prompts a demonstration of correct conceptual knowledge. We found that the control group elicited a far greater number of demonstrations of correct knowledge (p = .003), which we attribute to the lack of rules in the control simulation that allowed more freedom for experimentation. On the other hand, those exposed to the game intervention experienced a far greater number of productively negative events, we believe due to the presence of game mechanics which encourage cycles of negativity and reflection.

However, high levels of productive negativity does not necessarily suggest heightened understanding — we want to investigate the quality of these negative experiences by looking at how many demonstrations of correct conceptual knowledge were elicited per instance of productive negativity. While the numerical value of this quality metric is far greater for the control group (p < .001), the metric holds a trending relationship with post-test misconceptions amongst gaming participants (p = .066) but not control participants (p = .442). This suggests that playing a game offers a kind of learning guarantee; as you engage with game challenges, your actions are more strategic and therefore more representative of your true understanding, while the lack of rules and goals in the control app allows the student more freedom to ‘mess about’, resulting in actions that are not necessarily linked to their understanding.


We’d like to acknowledge those (other than the main contributors at the top of the page) who contributed to the creation of these apps.

  • Programming: Brendan Polley
  • Artwork: Cassandra Cetlin and Natalie Cormier
  • Level design: Derek Ng
  • Concept brainstorming: Derek Ng, Vijay Shahani, Shelley Wall, Dave Mazierski, Qingyang Chen, Priya Panchal-Banerjee, Megan Kirkland, and Erin Kenzie


SSHRC logo