The questionnaire creation is easy at first. It is, as long as some basic criteria are considered.
The criteria can be answered well with W-questions.
- What should be charged?
- How much should be levied?
- When should it be levied?
- Who should participate in my survey?
In summary: Content, coverage, procedure, target group.
The content follows the aim or purpose of the questionnaire. For example, the aim can be to compare the attractiveness of one product with another. A very well known questionnaire is the AttrakDiff or its successor, AttrakDiff2.
That sounds so simple. In this case, it is, too. A concrete use case for which a questionnaire already exists. Adapting a questionnaire for something specific is just exchanging a word in the questions. The questions are already validated, and therefore evaluation as a survey instrument is no longer necessary.
However, you should be warned against blindly adopting questions. Each question should serve a purpose. Meaningless questions confuse the person giving the answer. In the worst case, the respondent aborts the questionnaire or blindly ticked off so that the answers are useless.
However, there are two other ways to get to the questionnaire. Derive from the specialist literature - usually combined with literature research and an expert discussion.
Derivation From The Literature
Open topics are searched for in the literature to find the same or related topics. Theoretical connections can be found as well or theoretical contributions, which contradict the use in practice.
The expert discussion needs more effort. First, experts have to be identified. This can be the supervisor at work. However, an expert is very little to be able to make valid statements. The aim is to define a topic area that is meaningful for the purpose of the survey questionnaire in a framework with several experts. Thus, some topics and statements have to be formulated in advance in order to be prepared in the discussion and to be able to guide the discussion.
Basically, only those questions should be asked, which are necessary to achieve the goal. Necessary in the sense that the longer a questionnaire is used, the greater the probability that the participant will abandon the survey. An old Chinese proverb says: "Everything important fits on a thimble."
The scope depends on the objective of the survey questionnaire. I do not recommend that the questionnaire be longer than 15 minutes. In my experience, the willingness to participate decreases significantly after 30 minutes. It, therefore, requires incentives to bind participants to the question-answer for a longer time.
Note: Better shorter than longer. Brevity is the spice.
The process is significantly influenced by time. However, time should not only be based on the duration of the survey, but also on the time when the test group is available. Finding students during the semester break could become problematic. The same applies to teachers during the holidays. Understandably, these groups of people are mental with their children or in the learning phase.
Furthermore, the manner in which this is done should be taken into account. It may go online, particularly quickly. However, it is not always clear whether the person who has also been contacted answers. For example, when calling up the website via Facebook. In the case of an alternative survey by post, the times of postal delivery must be taken into account.
The target group/group of participants should consequently consist of persons who can provide information about the content. Asking students about the content of strategic business decisions could certainly be interesting. However, it seems to be more suitable to ask people who are active in the field to contribute their experiences. CIOs are a good option here.
It should be considered: The higher the level of people in the company, the more difficult it is to recruit them for the survey.
Creation of Questions
Once the most important criteria have been determined, the questions are created. There are two basic types of questions: closed and open questions.
In the case of open questions, the participant can answer freely, i.e., the answer options are not predetermined. An evaluation is carried out using a code to highlight similarities or differences.
For closed questions, the possible answers are given. The possible answers already indicate the later significance. The significance depends on the scale level. Four types are distinguished: nominal, ordinal, interval, and rational.
Nominal answers enable statements to be made about equality or difference, as with gender (man != woman, woman = woman, man = man).
Ordinal allows us to make additional statements about smaller or larger. For example, assessment using a Likert scale from very poor = 1 to very good = 5 (→ 1<2<3<4<5). Note: Very good is not five times better than very bad. There are no distances between the values themselves.
The interval scale allows the comparison of differences in addition to the ordinal scale. Example temperatures: Difference between 20 °C and 15 °C is equal to the difference between 10 °C and 5 °C Note: 20 °C is not twice as warm as 10 °C. A comparison of ratios is therefore, not possible.
Rational answers are the only way to compare conditions. Example income: 4000 € is half as much as 8000 €. Such a scale has a natural zero point.
The scales to be queried should always be balanced. Example: Instead of neutral to agree to better in both directions from disagree to agree.
Notes on the Formulation of Questions
How do I formulate a question?
- The questions should be kept simple, i.e., switch phrases should be avoided and confuse the participant. The statement behind a question should, therefore, be clear and unambiguous.
- Unambiguous means: one question is one answer. The questions, therefore, contain no AND and no OR. For example: How high is your gross income, and are you satisfied with it? Better create two questions here.
- Use the language of the target group. In case of doubt, include a comprehension question like: Did you have any language difficulties? If yes, which?
- The question asked will determine the chosen method (How - process, What - concrete effect).
A thesis and antithesis are formulated for each question. Each question has a concrete goal. Each question is purpose fulfilling and can be justified in case of doubt about its necessity. Here, data protection must be taken into account. Basically, the more personal the questions are, the less willing participants are to answer them, and the more critical data protection becomes.
Before the questionnaire is released, a pretest should be performed. The pretest serves several purposes. It shows whether questions are understandable, how long the questionnaire takes, whether the data in the tool used are the ones to be evaluated, and whether the participants are even willing to fill out the questionnaire. If the response rate is already very low during the pretest, the questionnaire should be improved, or the incentives to participate should be increased.