Automated generation of vocabulary quizzes en masse for large-scale testing
Vocabulary testing is a common element of most language teaching curricula, but presents certain problems (in quiz consistency, difficulty, and security) when used in programs that have large student enrollment. This presentation describes and demonstrates the latest version of Word Quiz Constructor (WQC), an open-source free application the generates word quizzes from online and offline sources.
WQC creates multiple-choice cloze questions, synonym questions, and free response cloze questions in a customizable configuration (e.g., 10 questions of each type). Stem sentences are drawn from online sources such as Wikipedia or offline corpora such as the British Academic Written English (BAWE) corpus. The target words for a quiz can be drawn from the Academic Word List (AWL: Coxhead 2000) or other customized lists. Google NGrams are used to ensure that the target word appears in contexts that are high frequency contexts. Finally, the select stem sentences are filtered using readability metrics (e.g., Flesch-Kincaid, Linsear Write) to control the difficulty level.
Quizzes produced by WQC have been evaluated with various populations: native Japanese learners of English, native speakers of English, and experienced teachers of English as a foreign language. Results (published previously) show that WQC produces questions that are as well-formed as manually-produced questions. It has also been used productively for several years in a large-scale ESP program in Japan.
This presentation will describe the architecture of WQC, demonstrate its use, and explain how to download and make use of the application.