Combining dynamic and static analysis for automated grading SQL statements

Abstract

Learning and teaching Structured Query Language (SQL) is a challenge that is widely recognized within the computer science education community. Programming ac-tivities are considered to be the most important aspect of the learning process. Manually grading of SQL statements, however, is often regarded as a tedious and time-consuming ordeal. In this paper we propose an automated grading approach for SQL statements, which combines dynamic and static analysis. The proposed approach first identified the correct statements from the student SQL statements, and then grades the incorrect statements based on the similarity between the statements and the correct statements. Empir-ical evaluation of the proposal was carried out on a dataset consisting of 255 manually graded pairs of statements. Experimental results show that our approach is superior to three state-of-the-art grading approaches. Our study found that each grading approach has its particular advantages and disadvantages, and the combination of different grading approaches can help to improve the quality of grading results. Show more