Bloom's Taxonomy-Based Approach for Assisting Formulation and Automatic Short Answer Grading

Abstract

This paper presents an approach to enhance automatic short answer grading accuracy by using Bloom’s Taxonomy as a reference for questions formulation. We sought to address the semantic aspects related to the answer by using WordNet and Latent Semantic Analysis models, which supported automatic short answer grading with size ranging from a single sentence to a short paragraph. The responses for three questions answered by high school students were graded automatically resulting in a high correlation with teacher grading (0.82, 0.91, 0.80). Another discovery is that automatic correction might vary according to the type of question, the application context and that the representativeness and concision of the expected response.