A questionnaire is a tool to make reality as reliable and valid as possible to map. A question is valid if the answer to the question matches with the true value of the measured construct. A question is reliable as a respondent who answers the question twice without changing the circumstances have changed, gives the same answer. Also, two respondents of which the circumstances are the same, give the same answer.
Our goal is to make questionnaires as reliable and valid as possible. Important it is to realize that we can never make 100% reliable and valid questions. We can only strive to make the measurement error as small as possible.
Four rules for reliable and valid questions
If we further elaborate the definition of reliable and valid, there are four rules that a question must meet in order to be valid and reliable:
- A respondent must understand the question;
- A respondent must have the necessary information to answer the question;
- A respondent must be able to translate the relevant information into a form that is is sufficient to answer the question;
- A respondent must be able to give the answer.
Understanding the question
It seems very logical; the respondent must understand the question to be correct can answer. Yet this is more difficult than it seems. Include your target audience thoughts when drafting questions. The goal is that all respondents understand your questions.
Formulate questions as briefly and concisely as possible. Respondents more often no longer read than two lines. If you include long questions or detailed explanations, the chances are that respondents will not read them at all. The result is that respondents also not being able to understand your question as you intend it.
Have the necessary information
The respondent cannot answer questions to which he does not have the answer. We can only ask questions for which the respondent has the necessary information has to give the answer. Consider carefully whether a respondent is able to answer your questions. If you have any doubts about this, ask a question that respondents can answer and translate the answers into the information you are looking for.
The human brain
The human brain is not a computer. Many things are forgotten or wrong estimated. Some well-known memory effects to consider, to be:
- Recency: The longer ago the event occurred, the more it is more difficult to remember them;
- Salient: the less important the event, the more difficult it is to retrieve of it becomes;
- Frequency: The more times an event has happened, the more difficult it becomes to recall a concrete example.
Because of these kinds of memory effects, Crowdtech's advice is always close possible to measure the experience.
Being able to provide the answer
Finally, it must be possible for the respondent to provide the desired answer to give. If this is not possible, you cannot possibly expect your results to be entirely accurate.
Common mistakes in this are:
• Asking double questions (eg do you want to be rich and famous?)
• Not fitting the question and the answer option (multi- or single response).