¿Es el entremezclado amigo de la inducción de categorías? Una replicación y extensión de Kornell y Bjork (2008)
Loading...
Files
Identifiers
Publication date
Reading date
Authors
Collaborators
Advisors
Tutors
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Share
Department/Institute
Abstract
El aprendizaje inductivo es un proceso cognitivo mediante el cual podemos inferir reglas y
patrones generales a partir de ejemplares particulares. Trabajos previos han investigado la
influencia de la presentación entremezclada de ejemplares para mejorar el aprendizaje
inductivo. En una condición de espaciado-entremezclado (SI), estímulos de diferentes
categorías aparecen mezclados y separados en diferentes bloques. Por otra parte, en una
condición de masivo-bloqueado (MB) ejemplares de una misma categoría aparecen de forma
secuencial y cercana en el tiempo en un mismo bloque. En 2008, Kornell y Bjork realizaron
dos experimentos utilizando pinturas de diferentes artistas. En el primero encontraron que SI
mejoraba la categorización frente a MB en una tarea a corto plazo con retroalimentación. En
el segundo se mostró que SI era mejor que MB en una tarea de reconocimiento en el que los
participantes debían reconocer elementos nuevos y viejos de diferentes categorías. El
presente trabajo replica y extiende el trabajo de Kornell y Bjork (2008) con cuatro
experimentos. En el Experimento 1 y 2 se replicaron los experimentos originales, añadiendo
el uso de análisis derivados de la Teoría de Detección de Señales. En el Experimento 3 se
realizó una tarea de categorización eliminando la retroalimentación para evitar el
reaprendizaje durante la prueba. En el Experimento 4 se probó el mismo fenómeno, pero a
largo plazo, introduciendo una semana de demora entre el entrenamiento y prueba.
Consistentes con resultados previos, las categorías presentadas en SI obtuvieron mayor
proporción de aciertos comparado con MB en todos los experimentos.
Inductive learning is a cognitive process by which we can infer general rules and patterns from particular exemplars. Previous works has investigated the influence of interleaved exemplar presentation to improve inductive learning. In a spaced-interleaved (SI) condition, stimuli from different categories appear mixed and spaced apart across different blocks. On the other hand, in a massed-blocked (MB) condition, exemplar from the same category appear sequentially and closely spaced within a single block. In 2008, Kornell and Bjork conducted two experiments using paintings by different artists. In the first one, they found that SI improved categorization compared to MB in a short-term feedback task. In the second one, they found that SI was better than MB in a recognition task in which participants had to recognize new and old exemplars from different categories. The present work replicates and extends the work of Kornell y Bjork (2008) based on four experiments. Experiments 1 and 2 replicated the original experiments including analyses derived from Signal Detection Theory. In Experiment 3, a categorization task was performed without feedback to prevent relearning during the test. Experiment 4 examined whether the same effect could take place including a one-week delay between training and testing. Consistent with previous results, categories presented in SI obtained a higher proportion of correct answers compared to MB in all experiments.
Inductive learning is a cognitive process by which we can infer general rules and patterns from particular exemplars. Previous works has investigated the influence of interleaved exemplar presentation to improve inductive learning. In a spaced-interleaved (SI) condition, stimuli from different categories appear mixed and spaced apart across different blocks. On the other hand, in a massed-blocked (MB) condition, exemplar from the same category appear sequentially and closely spaced within a single block. In 2008, Kornell and Bjork conducted two experiments using paintings by different artists. In the first one, they found that SI improved categorization compared to MB in a short-term feedback task. In the second one, they found that SI was better than MB in a recognition task in which participants had to recognize new and old exemplars from different categories. The present work replicates and extends the work of Kornell y Bjork (2008) based on four experiments. Experiments 1 and 2 replicated the original experiments including analyses derived from Signal Detection Theory. In Experiment 3, a categorization task was performed without feedback to prevent relearning during the test. Experiment 4 examined whether the same effect could take place including a one-week delay between training and testing. Consistent with previous results, categories presented in SI obtained a higher proportion of correct answers compared to MB in all experiments.
Description
Bibliographic citation
Collections
Endorsement
Review
Supplemented By
Referenced by
Creative Commons license
Except where otherwised noted, this item's license is described as Attribution 4.0 International







