Causal Modeling of User Data from a Math Learning Environment with Game-Like Elements

We have created a math learning environment with game-like elements such as narrative, visual feedback, personalization, collection, etc. We made a study with four different versions of the tutor with different degree of 'game-like' and found that students preferred more 'game-like' tutor but we were not able to detect any conclusive difference in learning. Based on the data we collected through survey, logs and tests, we also built a causal model to understand the interrelationships between different student and tutor variables.