How to conduct a survey
Conducting a survey is a tricky business. There are a number of considerations to think about in order to achieve the desired response rate. The points below indicate some of the key issues that can affect the responses that you receive and how to make your survey easier to process.
Questions
Questions should be ordered in a way that does not influence the responses to other questions.
Questions should not be phrased as to lead the respondent to a particular answer.
Begin the questionnaire with the questions which will be of most interest to the recipient.
Use closed questions instead of open ones. Open questions are those which allow comments. Closed questions are those which comprise a choice of answers. For example, the question “How do you rate the IT department?” is much easier to process if the recipient is given options such as good/bad/indifferent.
Refer to existing survey questions. These can be found in the SNAP .mdf file under the Downloads section.
Try to avoid questions that are likely to generate a “not applicable” response. For example, “How often have you been to the moon?” can only be answered by a very small number of people.
Try to keep your questions varied throughout the survey.
Survey design
Keep the questionnaire short (or as short as possible).
Use the University of Aberdeen SNAP templates. The layout of these has been optimised for readability and conforms to the Disability Discrimination Act (DDA).
Make sure you include all relevant answer options. A simple agree/disagree question may return a completely different response to one that has agree / generally agree / indifferent / generally disagree / disagree.
Limit score or rating scale questions to five points. For example, 1 is the highest score and 5 is the lowest
Use a consistent page layout. Make sure your questions remain on the left and the answers on the right.
You can use images to enhance the front page but ensure you restrict the number of images you use and keep them small in file size.
Ensure you preserve anonymity. In your e-mail introduction, highlight to the recipients that their details and answers will be kept private but they should not forward their e-mail link to other people. See the section on Web surveys and preserving anonymity.
General
Recipients of surveys should not be briefed beforehand. A good introduction to explain the survey and what it is for will be beneficial. However, make sure your introduction does not bias the respondents in any way.
Use simple and direct language.
Give your survey a short and meaningful title.
Surveys are generally anonymous. You can add a space at the bottom to let the respondent enter their details should you wish to contact them. However, this should not be compulsory and it should not be expected that respondents will complete this field.
Avoid technical terms and acronyms.
Have your questionnaire proof read. Test it on some of your colleagues before exposing it to the wider audience.
Limit the number of questions you display on each page to avoid excessive scrolling.
Avoid horizontal scrolling.