Intelligent Questionnaire Design for Effective Participant Evaluation by Lisa Elias
Step 1: Before designing a survey it is critical that the objectives are identified – what is to be achieved and why is the survey necessary? This means considering the nature of the people who will be surveyed, those who will gain information and the purpose of the task itself.
Step 2. Then, write the questions in a clear, well thought out manner based on the objectives outlined in step 1. In this way, the data collected will be high quality and applicable to the needs of all concerned (Elias, 2015, p. 8).
Ensure that questions are:
Clear and unambiguous
Avoid embarrassment – omit or minimise sensitive topics
Ensure respondents’ privacy
Select the question formats with the objectives clearly in mind. A mix of question types will elicit the best data.
Question types to consider:
Yes/no – quick response enabling simple comparisons
Multiple choice – only one selection or multiple selection?
Likert scales – demonstrate a rating per respondent on a common scale
Open-ended responses – time-consuming to analyse but rich qualitative data. Use sparingly.
Alternative responses – allow respondents to opt out or provide their own answer
Ordinal/ranking – a series of items that respondents are asked to rank (for example from 1 to 5 where 1 is most important and 5 least important; each number can only be used once)
Format the survey by considering the most logical layout to achieve your aims. This avoids confusing the respondents and makes analysis easier.
Introduction – explain why the survey has been established and convince people that participation is valuable and worth their time and effort.
Order and group the questions according to the format you have deemed most logical.
Initial questions should be impersonal and easy to answer so that respondents continue.
Short is best (Elias, 2015, p. 9)…
But ensure the information will be adequate for the purpose.
Use contingency questions if applicable so that people do not have to answer questions irrelevant to them. A preliminary question should ascertain how many, if any, questions of the following set need to be tackled.
Use a progress indicator for online surveys – it shows respondents how far they have to go.
Thank participants and provide your contact details.
Likert Scales should be:
labelled e.g. Poor (1) ranging to excellent (5)
Consist of an odd numbered scale so there is a mid-point – 5 or 7 options have proven best
Follow the same value pattern – either left to right or right to left
Make sure the words applied to the scale allow for the full range of responses
Elias provides a very helpful checklist to use when construction questionnaires:
- Has the survey been test-driven?
- Do others find the layout clear?
- Is the purpose explicitly explained?
- Have respondents been thanked?
- Is anonymity and confidentiality of data been guaranteed?
- Are instructions clear and precise?
- No duplication?
- Are questions plain and unequivocal?
- Are all questions essential?
- Are questions correctly ordered?
- Will closed questions result in the expected numerical data required?
- Are open text options sparingly used?
- Is there sufficient time for completion(Elias, 2015, p. 10)?
Elias, L. (2015, February). Intelligent Questionnaire Design for Effective Participant Evaluations. Training and Development, 8-10.