Abstract:

The objective of this article is to analyze the didactic functionality of a chatbot to improve the students’ communicative competence in Primary Education. Good chatbot design can make learning a more fluid, automatic process, and can integrate deep-learning didactic proposals. The use of chatbots is still an incipient technology in pre-university education. The influence of chatbots and virtual agents have generated scientific debates about their didactic utility (Crown, Fuentes, Jones, Nambiar, & Crown, 2010; Heidig & Clarebout, 2011; Schroeder, Romine, & Graig, 2017). Likewise, there are studies that show that the use of these agents or chatbots increases motivation among students and their academic performance in virtual learning environments (Liew, Mat-Zin & Sahari, 2017). Adverse aspects in the use of these resources in the educational scenarios have also been identified (Liew, Mat Zin, Sahari & Tan, 2016). The study by Heidig and Clarebout (2011) of scientific review of chatbots and virtual agents showed that there is evidence of their didactic functionality, although it also showed that they do not increase motivation and interest in academic activities. For this reason, in this research we created a gamified chatbot and analyzed the perception of students, teachers and parents on the possible educational benefits. This chatbot was implemented as an extracurricular activity during one month in four Spanish schools. The total sample consists of 120 students (4th grade), 40 parents, and 4 primary teachers. The research was approached from a quantitative and qualitative methodological approach. Firstly, an “ad hoc” questionnaire based on likert scale items was implemented and secondly an open digital forum was created to discuss the positive and negative aspects of this didactic experience. This forum was anlyzed through text mining with tests of Latent Dirichlet Allocation (LDA), pairwise distance matrix and bigrams. We have grouped the parents, students and teachers opinions in three subsets: DP1 Parents, DP2 Students and DP3 Teachers; within each group we have grouped the text-subsets DP123n. The comparison criterion depends on the topics discovered using latent Dirichlet allocation or LDA (Blei et al., 2003) and a pairwise distance matrix. The computation of this matrix is done only for the lower triangular values and then reconstructed to form the full matrix. The results show that the propagation dynamics in forum is appropriate with an alpha exponent. Main results show that parents (79%) and teachers (80%) consider the chatbot as a useful resource to favor the educational use of digital devices outside the classroom. In this sense, teachers value the feedback they receive from the results on the chatbot platform, which allows them to improve the assessment and attention to the student in the class. Finally, the students (91%) find the activity fun and creative, allowing them to practice several times until they understand complicated concepts on spelling. As negative aspects, it should be noted that the use of chatbot in smartphones makes writing difficult, being more advisable the use of a tablet or laptop.

Keywords:

Chatbot, gamification, communicative competence, primary education.