Análise da aderência de recomendações explicadas de recursos educacionais para apoiar o ensino e a aprendizagem em um ambiente educacional online

Abstract

With the advancement of Information and Communication Technologies (ICTs) and Distance Education (EaD), online learning environments have acquired a prominent position as tools that enhance the learning process. And it can be seen that these tools have been gaining a great popularity due to the fact that they make the learning process more dynamic, making the content adaptable for each student, bringing an individualized teaching, among other factors. However, as these environments provide many resources for users, this can cause confusion in the construction of learning, because although these environments provide some recommendations to help users make good choices, the lack of explanation can cause low acceptance, due to the lack of transparency. And this has been a problem in AI, as predictions / recommendations are made, but are not accompanied by explanations. That is, even with the recommendations, many users are unable to progress, due to the fact that they do not understand what was suggested / recommended and do not understand the reason for such recommendation, thinking that the commendation does not apply to their profile or their need and simply reject the recommendation. Thus, this work aims to investigate the effect of the recommendations explained in relation to adherence (acceptance) of users in online learning environments, analyzing whether or not there is a significant difference in these environments (with explanation and without explanation). In order to validate the proposal, a learning system was designed and a descriptive statistics course was made available, where, in view of the students’ interactions, the student model was created. Thus, we removed a case randomly from the student’s model, where we elaborated a scenario and created a survey, which provided 4 recommendations (two with explanations and two without explanations) and some statements that were displayed gradually, for each exposed scenario, for that participants could respond to. And for each recommendation, the participant was asked to answer some statements that were proposed, using a likert scale. Therefore, data collection and analysis was performed, where we used a factor analysis, through a paired test (Wilcoxon), to compare the results of each scenario, in order to evaluate the participants’ opinions. Upon carrying out the experiment, it was seen that, as in other environments, the explanation also brought good results in the recommendations of educational resources, where there was a greater acceptance by the user, when he started to understand the reason for such recommendation, judge whether that is really relevant or not.