How to write better multiple choice quizzes: An evidence-informed approach
(From Teaching with Moodle)
A recent paper examined the effects of multiple choice quizzes (MCQs) on students' learning* and best practices for optimising learning from MCQs (Butler, 2018). Simply put, Moodle's quiz module can be leveraged to produce substantial learning gains, especially with frequent low/no-stakes quizzes to review previously studied material. I've summarised and elaborated on the paper's findings as follows:
1- Don’t use complex (multi-stage) MCQs. Too complex, too easy to get it wrong, less reliable than simple, don’t add anything to learning/test validity.
2- Target specific (and hopefully authentic) cognitive processes with MCQs, e.g. individual definitions (item specific) and compare & contrast (relational). What decisions/distinctions do learners have to make in practice?
3- Don’t use “None of the above” or “All of the above” (see #1 = complex MCQs).
4- Only 3 options = 1 correct + 2 high quality (plausible) distractors. More aren’t necessary and can have negative effects.
5- MCQs should be challenging to learners but not too difficult (or too easy), i.e. should test ILOs at the level taught.
6- Extra: Provide appropriate feedback on wrong answers. Feedback can be:
corrective (i.e. give the correct answer),
directive (i.e. tell the learner how to get to the correct answer; the process),
or epistemological (i.e. use further questions to elicit the learners’ relevant knowledge about the
depending on learners’ proficiency/knowledge in the subject matter.
*Also see the testing effect (e.g. Roediger & Karpicke, 2006), also known as retrieval practice (e.g. Agarwal, Roediger, McDaniel, & McDermott, 2018)