Uso de la comprensión lectora para la construcción de un modelo predictivo del éxito de estudiantes de 4º de Primaria cuando resuelven problemas verbales en un sistema inteligente

This work presents a dynamic mathematical model designed to predict fourth grade primary students’ (9-10 years-old) achievement in the arithmetic solving of word problems. The work is organized into two parts. In the first part, we analyzed the variables that will be eventually included into the mod...

ver descrição completa

Detalhes bibliográficos
Autores: Sanz, María Teresa, González-Calero Somoza, José Antonio, Arnau Vera, David, Arevalillo Herráez, Miguel
Tipo de documento: artigo
Data de publicação:2019
País:España
Recursos:Universidad de Castilla-La Mancha
Repositório:RUIdeRA. Repositorio Institucional de la UCLM
OAI Identifier:oai:ruidera.uclm.es:10578/39154
Acesso em linha:https://hdl.handle.net/10578/39154
Access Level:Acceso aberto
Palavra-chave:Comprensión lectora
Educación Primaria
Intelligent tutoring systems
K-BIT
PIRLS
Primary Education
Problem solving
Problemas verbales
Reading comprehension
Resolución de problemas
Sistemas tutoriales inteligentes
Word problems
Descrição
Resumo:This work presents a dynamic mathematical model designed to predict fourth grade primary students’ (9-10 years-old) achievement in the arithmetic solving of word problems. The work is organized into two parts. In the first part, we analyzed the variables that will be eventually included into the model, the relationships between them and the effect on the predicted outcome variable. The model takes into consideration variables related to subjects’ features: prior proficiency in solving word problems, level of reading comprehension, and the ability to solve reasoning problems through figurative and abstract visual stimuli (also called fluid intelligence). Additionally, the model employs a variable that describes the task difficulty using the semantic category of the statements as a criterion of complexity. Sixty-four fourth-grade students took part in the experiment. For each student data were gathered by using three different instruments: a PIRLS reading comprehension test, the matrix subtest from the Kaufman test, and an intelligent tutorial system, which was used by participants to solve 26 word problems. All variables showed a strong statistical correlation with the dependent variable (the students’ achievement in word problems), except for fluid intelligence. The presented model reveals that reading comprehension has a significant relevance when it comes to predicting primary students’ performance in word problems. In addition, this model offers better behaviour in comparison to similar alternatives presented in the literature. Keywords: reading comprehension, PIRLS, problem solving, intelligent tutoring systems, K-BIT, word problems, Primary Education.