As the questionnaire represents your contact with the respondent, how you design the questions in it is very important. Keep the following in mind:
1. Designing your own questions means more freedom, both in wording and format. However, this raises doubts about objectivity, reliability (to what extent repeated measurement under the same conditions and on the same units will give the same results) and validity (whether we are measuring what we intended to measure in the first place). Many different measurement scales and questionnaires have been developed in order to improve these three components as much as possible. So check if there are already any that look for the information you want (e.g. in a survey library).
2. Do not force respondents to answer (e.g. mandatory questions), but rather check that respondents have provided the information you are looking for. E.g. if you want to know the year of birth, limit the answer field so that only four-digit numbers can be entered. The alert should therefore only remind you that the information provided was incorrect and ask for a correction, but should not prevent you from continuing. The latter may lead respondents to make up answers or even to stop answering the survey.
3. The content of the questions should be simple and easy to understand. This means avoiding jargon, slang and overly technical words. Avoid questions that have more than one dimension.
4. The order of the questions is very important:
- The order should be logical. E.g. questions about events should follow the order of events, questions about the evaluation of an experience should follow questions about the description of that experience, etc.
- Questions relating to the same topic should follow each other. Constantly changing topics reduces the likelihood that answers will be well thought out. If the questions are organised by topic, the respondent can focus on only one topic at a time.
- The topics (and questions) should be ordered from those that are most important to the respondent to those that are least important to the respondent. Research shows that surveys that address topics that are more important to respondents have higher response rates.
- Sensitive (difficult) questions should be more towards the end, as you do not want the respondent to stop completing the questionnaire at the beginning because of these questions.
- Demographic questions are usually at the end of the questionnaire.
- Questions with similar parts should be together - e.g. questions with the same scale should be placed together. In this way, the cognitive burden on the respondent is reduced.
5. The first question should be suitable for everyone (the only exception is when the first question is to determine the "suitability" of the respondent), namely:
- Keep it simple so that respondents can read it, understand it and answer it in a few seconds. This is not the place for long questions with many possible answers or for open questions.
- Make it interesting, as this increases the likelihood that the respondent will participate in the survey.
6. The measurement scales should be consistent throughout the survey. If you have used a scale from 1 to 5, this should be the case throughout the survey. The naming of the highest and lowest values should also be consistent (e.g. 1 should be the lowest and 5 the highest everywhere, or vice versa).
7. Care should be taken when drafting the first and last pages of the questionnaire. The first page should primarily motivate the respondents, so do not put instructions on it. The last page should be simple; it usually includes a thank you.
8. Determine which question type is best for each question.
9. If there are many questions in the survey, it should also have several pages, as too much scrolling is not desirable. It is also not advisable to have only one question per page, as this increases the time it takes to complete the survey and also increases the chance that respondents will stop answering.
10. Use conditions to branch the survey when necessary. Never include text in the survey to the effect of "If you answered 'yes' to Q1, then answer Q2, otherwise go directly to Q3". Conditions are only available for users of the 2KA and 3KA packages for individuals, and for users of the business packages for groups.
11. Check that you have included all possible answers for the closed question types. If you are not sure, use the "other" option.