This dissertation examines the motivational effects of a mathematics serious game called the Number Navigation Game (NNG) amongst upper elementary school students in Finland and Mexico. The aims of the studies concern the NNG’s impact on motivation as measured through the expectancy-value framework, with special emphasis on the development of students’ situational interest during gameplay and individual interest toward mathematics. The role of factors such as game experience and the voluntary vs. compulsory nature of different play contexts is also explored. As this research was undertaken in the context of a larger project (CUMA) investigating the game’s effectiveness in enhancing mathematical skills, some attention is also paid to the game’s impact on mathematical learning outcomes. The results of three original empirical studies in which the game was implemented at the upper primary school level are reported. 
Study I presents the results of a large-scale experiment carried out in Finland amongst fourth-to sixth-grade students. The students were randomly assigned by class to either an experimental or control group. The experimental group (n= 642) played the NNG as part of their regular mathematics class for a ten-week period, while the control group (n= 526) continued with traditional textbook-based learning. The students completed various mathematical tests and questionnaires on their motivation and game experiences both before and after the intervention. The results revealed the game had a positive effect on learning outcomes, but that there was a slight decrease in the motivation expectancy values of the experimental group. Game experiences were mostly negative, which indicates room for improvements in game design. There was some variation in these experiences by gender, specifically regarding students’ feelings of competence and challenge, but in any case, game experiences did not play a role in either learning outcomes or motivation expectancy values. 
Study II concentrated on a subsample of participants from the first study, namely n= 212 fifth-grade students who, although they served as a control group for Study I, later played the NNG for a six-week period. These students’ individual interest toward mathematics was measured before and after playing the game, while their situational interest toward the game was measured on-task throughout five sessions. The results indicate that prior interest toward mathematics predicts initial situational interest. Growth curve mixture model analyses revealed a three-class model of situational interest trajectories, showing that the students’ interest develops differently throughout the intervention. The results indicated that the game was able to trigger and maintain the interest of most (73.9%) although not all participants. Some students’ interest was never triggered by the game (26.1%).In cases in which interest was triggered but not maintained by the game (15.9%),this had a negative impact on students ´ individual interest toward mathematics. At the moment it seems the game is beneficial to students who already have an interest toward the subject matter. 
In Study III, fifth-grade students from Mexico were randomly sorted by class into one of two play context groups: the voluntary group (n= 579) or the school group (n= 482). The school group played the NNG as part of their regular mathematics lessons. Students in the volunteer group received a copy of the game and were instructed that they could play in their free time if they so desired. Pre-and post-tests and questionnaires were completed before and after the intervention. The aim was to find out the effects of play context on game experiences, game performance, learning outcomes, and motivation expectancy values, and to explore to what extent students would play voluntarily, and how those who chose to play differed from those who did not by gender, pre-test mathematical skills, and motivation expectancy values. The results revealed that students in the voluntary group who played had higher prior mathematics interest and advanced mathematical skills than students in the voluntary group who did not play; interest toward digital games did not play any role in terms of whether students in the voluntary group played or not. While some students in the voluntary group did not play the game, their interest toward mathematics did not decrease after the intervention. Voluntary play had a positive effect on advanced mathematical skills compared to students from the school group. As for game performance, students in the school group played for longer, completed more in-game mathematical tasks, and had more enjoyable game experiences than students in the volunteer group did. Playing the NNG had a positive effect on mathematical skills regardless of play context. Motivation expectancy values remained mostly unchanged regardless of play context. 
The results from the three studies provide further evidence that motivation, as measured through the expectancy-value framework, is largely stable, and that serious games are not particularly successful in increasing student motivation toward a subject. Improvements in game design could result in improvements in experiences when playing the NNG, although it seems that game experiences do not play a role in either motivation or learning outcomes. The context in which a serious game is implemented, or the amount of freedom students have in playing does not seem to make a difference to motivational gains, either. The game is, however, able to trigger and maintain the situational interest of most students, although it seems that students who have a high prior interest toward mathematics are the ones who benefit. On the other hand, the game mechanism is successful at enhancing students’ mathematical skills. In this sense, serious games offer many possibilities as additional tools for teaching, but it is important that games be carefully selected for their proven learning outcomes rather than because they are assumed to be motivating for all students.