Concept Inventories (CIs) have been developed to assess students' understanding of the main concepts in a study area. They typically consist of a set of Multiple Choice Questions (MCQs). What is unique about these CIs that, in addition to the correct answer, they are designed, through a very thorough and collaborative process, to include a number of distractors that represent commonly observed conceptual misunderstandings.
The ease of marking MCQs has made them an attractive method for use in concept inventories. However, this may pose a limitation on how the test scores can be used to provide better insights into students' misconceptions, and this hinders efforts to attempt to correct them.
In order to address this problem, our recent research proposed extending multiple choice questions in concept inventories to include a text field for students to provide short answers, allowing students to express the reasoning behind their choice of a specific answer.
- Conducting literature reviews
- Developing and implementing algorithms
- Collecting data and interpreting and analysing results
This project will build on our previous work on analysing students’ textual responses, to investigate the use and utilisation of deep learning, to assess their conceptual understanding.
The variety of forms in which students could present responses is an important point to consider. In addition, the effect of language fluency on students’ ability to appropriately express themselves, and subsequently on the interpretation and analysis of their answers.
Skills and experience
- Fluent in Matlab and Python programming, excellent math and calculus skills.
- Machine learning, deep learning, text analysis, programming, engineering
You may be able to apply for a research scholarship in our annual scholarship round.
Contact the supervisor for more information.