AGENTE DE SOFTWARE Q Y EL APRENDIZAJE DE MODELOS CONCEPTUALES

Resumen

We submit results of a research project focused on the relationship between a card
arrangement exercise and the learning of conceptual models, built based on hierarchical
ontologies. This approach is more frequent in measurement of attitudes - hot cognition, as it is
expressed by Brown (1987) - or in opinion or subjectivity studies (Brown, 1980), and it is
known as Q Methodology. A software agent - Q Agent - was programmed to be used either in
a hypertext environment or as a companion to study press document and it’s effects on
learning are explored. Two main comparisons are conducted: between use and non use of Q
Agent, and between Q Agent with hypertext and Q Agent with pressed documents. The
predictive value of student scores with the Q Agent in relation with student scores on a
cognitive competence test is assessed as well. Three trials were carried on with a ninth grade
student sample. Data are analyzed by means of ANOVA and multiple regression models.
Significant effects of using the Q Agent were found, and scores on the Q Agent Game are
significant predictors of scores on the cognitive competence test. Learning performance using
Q Agent in hypertext environments drift lightly to be superior in relation to Q Agent with
pressed documents. Results are discussed in the framework of information controlled
autonomous learning systems, and the meaningful learning.