SQL tester: an online SQL assessment tool and its impact

Abstract

Learning SQL can be surprisingly difficult, given the relative simplicity of its syntax. Automated tools for teaching and assessing SQL have existed for over two decades. Early tools were only designed for teaching and offered increased feedback and personalised learning, but not summative assessment. More recently, however, the trend has turned towards automated assessment, with learning as a side-effect. These tools offer more limited feedback and are not personalised. In this paper, we present SQL Tester, an online assessment tool and an assessment of its impact. We show that students engaged with SQL Tester as a learning tool, taking an average of 10 practice tests each and spending over 4 hours actively engaged in those tests. A student survey also found that over 90% of students agreed that they wanted to keep trying practice tests until they got a "good" mark. Finally, we present some evidence that taking practice tests increased student achievement, with a strong correlation between the number of practice tests a student took and their score on the assessed test.

Exploring the Prevalence of SQL Misconceptions: a Study Design

“…explained by this correlation) [5]. Kleerekoper and Schofield find that formative assessment in the shape of practice tests can improve students' score on the test [4], providing motivation for the students to participate in our study.…”

“…Brass and Goldberg were the first to extend the work on errors of Reisner, Welty and Stemple by introducing logical errors: semantic errors that produce a -by definition incorrect-result, such as an empty result table, because of a contradictory WHERE clause [3]. Other recent research calculates query formulation success rates [4,9], or creates more specific lists of errors of a certain type: syntactic [1], semantic [2] or both [11,14]. Another newly introduced category is that of complications: queries that give the correct result but contain unnecessary elements [16].…”

The False-Positive Rate of Automated Plagiarism Detection for SQL Assessments

“…The first dataset, which we refer to as exam, is taken from an automated assessed test taken under exam conditions using the tool described by Kleerekoper and Schofield [7]. Students were not able to see the questions in advance and could not communicate with each other during the test.…”

“…At first this result is surprising because the SQL assessment tool used for the exam dataset requires specific column names for an answer to be correct (see [7]). We would therefore expect that removing alias names would make no difference.…”

“…Despite the increasing use of automated and online assessment tools for SQL [1,7,8,12], there has been relatively little research into detecting plagiarism in SQL queries [14,15]. Indeed, the systematic review of automated detection tools from Novak et al identifies 29 languages used in 131 articles and none of them are SQL [10].…”