Stronger than multiple choice, yet not quite as revealing (or time consuming to grade) as the essay question, the short answer question offers a great middle ground – the chance to measure a student’s brief composition of facts, concepts, and attitudes in a paragraph or less.
The University of Wisconsin Teaching Academy calls short answer questions “constructed response”, or “open-ended questions that require students to create an answer.” The Center for Teaching Excellence at the University of Illinois at Urbana-Champaign says that short answer questions allow students to present an original answer.
Like all assessment items, a short answer question should clearly assess a specific learning objective. It should ask students to select relevant facts and concepts and integrate them into a coherent written response. Question 1, below, is a typical example of a short answer question requiring such a constructed response.
 |
This question sets up a scenario with an expert role, a community history, and an environmental problem and asks the students to use a specific problem-solving strategy — the 4 A’s — to frame a response, which can most likely be completed in one paragraph.
Question 2 is slightly more problematic because of a very common error in constructing short answer questions.
 |
This question, while well-intended, actually asks two questions. This likely will leave the student confused as to which question is more important. Additionally, the student will have to write a longer response to answer both questions, leading this particular test question more toward an essay response than a short answer. Short answer questions should always ask one clear question, rather than confusing the issue with multiple queries.
Finally, one strategy professors use is to post a rubric in the test so that students will know how points will be distributed. Question 3, below, both shows such a rubric and demonstrates another common problem in short answer question development.
 |
Note in this question, a scoring distribution is provided to the students — not only do they know the question is worth six points, but they also know immediately that three points will be awarded for fully answering the question and two points for legibility, with the final point for spelling and grammar. Question 3 also demonstrates another common error — writing questions that close off a student’s potential answer. A better question would ask “How might two accidents be an acceptable level of risk…”, in order to promote a more meaningful answer.
Short answer questions are a great middle ground for professors. They are easier to develop than multiple choice and generate a more in-depth answer. Because of their brevity, they are easier to grade and they encourage students to integrate information into a coherent written answer. They can measure many types of knowledge when phrased correctly — even divergent thinking and subjective and imaginative thought. Best of all, they can provide professors with an open window into student learning — the real purpose of assessment.
Susan Codone, PhD is an associate professor in the Department of Technical Communication, School of Engineering at Mercer University.
This Post Has One Comment
While I agree in this particular case that the second example is flawed, I disagree with the premise. Part of exercising judgment is identifying which problem is the most important. Judgment can be demonstrated in short answers as well as longer, essay questions.
One control on length is to be clear on expectations. If the students knew that the professor would award two marks for each well-formed thought – typically one sentence – then the students would know they were expected to answer the question in roughly three sentences. This is similar to your example of the express rubric, but it is clearer since "complete answer" is undefined and may not allow for "overmarking" which is having multiple ways of getting the six marks.
A second control is to emphasize the marks per minute model. If the exam is one hundred marks and 150 minutes (2 1/2 hours) then there are one and a half minutes per mark. Assigning half the time (say) to reading and rough analysis, a six mark question would allow for four to six sentences.
I prefer considering short answer questions as primarily designed for students to apply course knowledge with a small degree of analysis. Accordingly, I agree with you that application questions are generally limited to a single area of knowledge: one issue per short answer question is generally enough.
In contrast, multiple choice questions tend to evaluate knowledge recall and comprehension. Obviously, it is possible to use multiple choice even to test evaluative skills, but publishers rarely provide those types of questions since they take longer to construct. Short answer questions provide greater scope for my B and A students to separate themselves from the C students tend to be rote learners and prefer the limited scope of multiple choice.
Thank you for your article.