Save_The_Money
Description
After the pedagogical intervention, a quantitative assessment was carried out, with the intention of verifying cognitive latency (Likert-type scale). The instrument used was the questionnaire for evaluating educational games based on motivational strategies from Keller's ARCS model, in the area of game user experience (Game User Experience), as well as Bloom's taxonomy of educational objectives. The results were analyzed using descriptive statistics. As a choice for this research, we used the calculations of the mean and standard deviation in relation to the domains: Motivation and User Experience. Statistical inference tests were not used, but the Coefficient of Variation (CV) of each item that made up each of the domains was considered. This variability indicated agreement between students regarding the domain (Latency). In other words, to what extent would there be agreement on the aspects investigated in this research? An agreement equal to or less than 30% would be very similar among the participants, but values greater than 30% would represent high variability, which would imply low agreement between the analyzed items. The main findings were: With regard to the domain of motivation, most indicators showed a high level of agreement among users, since attention (11%), relevance (9.9%) and satisfaction (10.5%) did not have a value greater than 30% with regard to the coefficient of variation (CV). That is, this indicates that the students reacted favorably to the game in the learning situation about the proposed content. However, the confidence indicator had a high value of disagreement, that is, it was possible to observe a variability coefficient of approximately 63.2%. Regarding user experience, it can be seen that all indicators showed a high level of agreement, as immersion (7.7%), challenge (11.4%), competence (12.5%), social interaction (11.9%) and fun (14.5%) were below 30% of the coefficient of variation. As final considerations, the study allowed observing the possibility of some improvements (redesign) in the face of some difficulties that were observed throughout the application of the game. The choice of cards within a board game as the main component of interaction with the game system, despite making the experience more practical and easier, compromised the participant's experience. Faced with the reality that today's students are digital natives, the hypothesis that the immersion attribute would be better used with the use of electronic media, since it would be closer to the reality of the participants, as it would allow the insertion of more technological resources attractive and elaborate. In this way, a closer exploration of the students' reality with the use of digital tools could bring greater coverage to this study. As an idealized product, built and tested in this research, the SaveTheMoney educational game could be applied in other educational contexts in basic education, specifically in elementary school.
Files
Steps to reproduce
For the construction of this work, the Design Science Research (DSR) research method was used, which has as its main premise the creation of an artifact to solve the problem, in our case, the development of a prototype of a board game, which we call by SaveTheMoney. The research participants (n=13) were elementary school students from a school in the city of São Bernardo do Campo, located in ABC Paulista. Students were chosen at random and were not identified to preserve the confidentiality of their identities. After the pedagogical intervention, a quantitative assessment was carried out, with the intention of verifying cognitive latency (Likert-type scale). The instrument used was the questionnaire for evaluating educational games based on motivational strategies from Keller's ARCS model, in the area of game user experience (Game User Experience), as well as Bloom's taxonomy of educational objectives. he results were analyzed using descriptive statistics. As a choice for this research, we used the calculations of the mean and standard deviation in relation to the domains: Motivation and User Experience. Statistical inference tests were not used, but the Coefficient of Variation (CV) of each item that made up each of the domains was considered. This variability indicated agreement between students regarding the domain (Latency). In other words, to what extent would there be agreement on the aspects investigated in this research? An agreement equal to or less than 30% would be very similar among the participants, but values greater than 30% would represent high variability, which would imply low agreement between the analyzed items.