Writing effective multiple choice questions

Multiple choice questions are useful tools for quickly assessing student learning, but ensuring that they actually achieve this goal is not as easy as it sounds. In fact, writing effective multiple choice questions is anything but quick. Here are three tips to help you build better MCQs.

1. Review your learning objectives

When you’re trying to come up with a nice round number of MCQs, it’s easy to start writing questions on details that stray from the things you really want your students to learn. Keep your learning objectives close, and make sure each question addresses these objectives. Since best practices call for transparency, hopefully your students have access to your learning objectives. If so, by sticking to your objectives when you build your questions, you can also ensure that you are testing your students on the material they studied, rather than less important details they may have missed.

2. Keep stems clear

So you’ve identified a learning objective you want to test, now it’s just a matter of making it into a question, right? The next thing you know you have a paragraph-long stem with three blanks for students to complete. Effective multiple choice questions should be as clear and concise as possible, so you will need to refine your question. That doesn’t mean your questions must always be short, but it does mean cutting the fat and keeping your wording as concise and simple as possible. Don’t weigh your questions down with unnecessary information and avoid long, complex sentences that might be confusing to students. Remember, you don’t want to test their comprehension of the test itself–you want to be testing their mastery of the learning objectives!

Consider the difference between these two questions:

In 1793, Eli Whitney revolutionized the cotton industry by creating what piece of technology that automated the separation of cottonseed from cotton fiber?

or,

What invention revolutionized the cotton industry by automating the seed separation process?

Both are asking for the same answer (the cotton gin, if your history is rusty) but the second is a much neater question. Whitney and the year are not important to this question, which is about the name of the invention. If you wanted to check that the students know the year of the invention or the name of the inventor, those should be different questions.

3. Choose believable distracters

Once you’ve built a good question, it’s equally important that you produce good distracters, or wrong answers. Distracters should be responses that a student might reasonably choose if they have not studied the material. In order to achieve this, two main principles apply. First, make the distracters match the question. Consider the cotton gin question above. The question clearly asks for an invention. In order for the distracters to be believable, they should reasonably sound like inventions. They should also be singular, since the question call for one invention. Second, make the distracters match each other. Answers should be about the same length and have similar wording. If the correct answer is a one-word noun, make all of the answers one-word nouns. If the correct answer calls for a complete sentence, make all of the answers complete sentences.

Consider these three answer sets that could go with the question above:

a. the spinning jenny
b. sewing machines
c. cotton gin
d. weaving loom
a. spinning jenny
b. sewing sally
c. cotton gin
d. cotton ale
a. cotton loom
b. cotton seeder
c. cotton gin
d. cotton sifter

The first answer set is all over the map. Answer (a) has an article, but none of the rest do. Since this isn’t necessary, remove it. The second answer is plural, so it doesn’t match the question. Also, students will most likely recognize that neither sewing machines nor weaving looms remove seeds from cotton, so those are implausible distractors. Any attentive student would most likely land on the correct answer, whether or not they read the material.

The second answer set is a little better because the answers are parallel, but students are likely to notice that two answers have “cotton” and the other two have different words, and throw out (a) and (b) without reflecting on the material. Answer (d) might throw them off, but again, the chances are high that a student will answer this question correctly even if they don’t know the material.

The third answer set is best. First, since all four answers have “cotton” in them, there’s no problem with cuing between the question and answers (cuing is when the student identifies the correct answer because it has the same words as the question). Second, loom, seeder, and sifter all seem like reasonable names for an invention that does the task described in the question. Third, by providing two answers that end with “er” and two that do not, none of these stands out obviously from the rest. With these distracters, you’re really testing the students knowledge of the material, rather than their mastery of test-taking techniques.

Summary

Multiple choice questions can be useful for assessing student achievement and they are quick to grade. But, it’s important to avoid over-using MCQs as an assessment method. To assess higher-level learning, oral and written exams are preferable whenever course circumstances make them a possibility. Our assessment item-writers provide MCQs with each course, but we also include numerous writing prompts to further assess student achievement. We are committed to promoting reading and writing across the curriculum, and suggest that multiple choice questions are only a small part of your online course assessment item bank.

For further reading, take a look at Mike Dickinson’s excellent article on writing multiple choice questions for higher-level thinking.