Survey Tips & Best Practices
Before you begin building your survey, think about what your goals are. What do you hope to learn from the data you collect? Ensuring that your survey questions fit your objectives will help you design a more successful survey.
Surveys are a powerful and effective way to gather information. Unfortunately, not all surveys are created equal. A survey that is designed poorly can give misleading or even false results. Therefore, knowing the best way to design a survey can go a long way towards helping get the results you need.
Research Phase
Surveys can be designed for a variety of reasons. Businesses can use surveys to measure customer satisfaction, test the viability of a new product, or to assess the performance of advertising activities. Students can use surveys to assist with research projects, and non-profits can use them to better understand the mindset of donors. No matter what you are using your survey for, make sure to start with a clearly defined objective to help you decide which questions to include. Make sure to thoroughly research your target audience as well.
Example: I have an idea for a new online dating website. I know that there is a lot of competition, and I want to gage the interest in my plan.
Work Backwards
Once you’ve defined your objectives, start thinking about the information you would like to obtain from your survey. Then, work backwards to figure out how to word the questions so that they will obtain the most accurate results. Keep your questions focused and specific. You don’t want to ask questions that will collect unnecessary data.
Example: I would like to identify the demographic that most frequently uses dating websites. Therefore I will create a survey question asking my respondents to identify their demographic information, and a separate question asking how often they visit dating websites. I will not ask vague questions like, “Do you use dating websites a lot?”
Structure
Keep your survey as short and concise as possible. If your survey has to be long, try to make only the most relevant or important questions required. Divide the questions into pages to make the survey appear shorter to the respondent. You can also use Skip Logic to lead respondents through the most relevant path through the survey, and avoid asking questions like, “If you answered yes above…” Questions should flow in a logical order, and should not have too many response choices.
Check for Bias
Make sure that your questions are neutral and don’t lead respondents in the direction you want them to go. Avoid using statements or ambiguous questions; instead, start your questions with who, what, where, why, or how. Randomize your response choices wherever possible to make sure that survey fatigue doesn’t corrupt your data. Remember to include an ‘other’ or ‘not applicable’ response when possible so respondents don’t choose irrelevant answers.
Example: I won’t ask, “Do you agree that online dating is the best way to meet new people?” because that is a leading question. Instead, I will ask, “What do you feel is the best way to meet new people?”
Test Run
Before you send out your survey, make sure to test it with a handful of friends or coworkers. They can help you ensure that your survey is easy to understand, doesn’t contain any technical errors, and isn’t longer than most people have patience for. If necessary, revise your survey to reflect the feedback from your test run. Make sure to delete your test responses before distributing your survey so they don’t corrupt your final data. It is also helpful to include an open ended question at the end of your survey inviting your respondents to tell you what they thought of the survey itself; you can use this feedback for future surveys.
Respect Respondents Privacy
In the introduction to your survey, it is a good idea to explain to potential respondents who you are, why you are conducting this survey, and what will happen with the results. If you don't need to be able to identify respondents, make sure to turn off the setting in distribution that collects IP Address. You should also post your privacy policy to reassure respondents that their identities and responses are safe.
Analyze your Results
Make sure your sample size is large enough to achieve accurate results. When you have reached your desired number of responses, compare the data you received to the objectives you defined before creating your survey. Results to open ended questions are particularly helpful to gain valuable insights.
Make sure your sample size is large enough to achieve accurate results. When you have reached your desired number of responses, compare the data you received to the objectives you defined before creating your survey. Results to open ended questions are particularly helpful to gain valuable insights.