Create a survey with a question you want to answer in mind. You should have hypothesis that you want to answer, but don’t make it obvious based off your questions.
A good survey is typically between 15 to 30 items. Any shorter and you may not get enough information, and any longer can decrease student interest.
If you have to cut questions, break them into ones that are essential, potentially useful, or lackluster/fringe.
Using a Likert scale can make it easier to use the data. However, it can require some additional planning.
Go with an odd number of responses. This allows for a truly neutral response.
Five is typically enough options. When going to a higher number, you can get more specific responses (but will increase the time it takes to complete the survey).
Make sure all best-to-worst responses all run down the same direction.
Place survey items strategically. Avoid starting with questions that are controversial or can impact later items. Try to place neutral, interesting questions at the start to draw students in.
Save demographic questions for the end of a survey, if applicable. Students will feel more comfortable answering them, and they require less thinking than other survey items.[1]
Executing the Survey
The higher the survey completion rate, the more valuable the data. If a survey is optional, it’s likely that only students with strong opinions (both positive and negative) will take the time to complete it. Encourage all students to take the survey.
Maintain student confidentiality whenever possible. If students do not feel like they can give anonymous answers, they’re less likely to give truthful responses. Make sure that students know that there are no right or wrong answers to the questions.
However, make sure students are aware that confidentiality can be broken in extreme circumstances, such as threats or vulgar comments. Making students aware of this can deter many potential unprofessional responses.
Consider tying a punishment or reward to a survey. Going with a carrot (or stick) can help ensure that all students take the survey seriously and complete it in a timely manner.[1]
Common Question Pitfalls
Questions that are inherently biased. These can be questions that frame a specific response in a negative light, or lead someone taking the survey to answer a particular way.[2]
Don’t you think that the professor was extremely polite and cordial?
Questions that contain multiple elements. These are often looking for either agreement or disagreement when students could possibly agree with only one of the elements.[2]
Do you agree that this class is the perfect length and covers only necessary content?
Questions those are too long, awkwardly phrased or contain unnecessary details. Students can become confused by them, or find themselves losing interest. Both of those factors can impact reliability.[2]
With how many students actively miss class even though the syllabus openly states that attendance is taken every day for points and is mandatory, do you think that information should not be placed on the Canvas site for students who perpetually skip to access at their own leisure?
Questions that are unrelated from the survey’s goals. These questions are ultimately unnecessary, and can distract students as to the actual purpose of the survey.[2]
Do you think that United States history is better than world history?
Creating a Survey, Florida Center for Instructional Technology, University of South Florida. Guidelines for survey creation, and the different types of surveys.