Questionnaire Design

This workshop was conducted by Dr Mehmet Omen from the Department of Econometrics & Business Statistics, Monash University. The focus was on how to design questionnaires aligned with one’s research objectives, what are the pitfalls of poor questionnaire design, become familiar with different types of questions and answer formats and to understand how physical characteristics and deployment of the questionnaire affect response rates.

Date: Wednesday 18 May 2016
Time: 2 – 4pm
Location: Rm 340, Level 3, 9 Rainforest Walk, Clayton Campus

In most research projects, questionnaire are deployed to ascertain opinion or feedback from experts or the general public. Having a clear research objective and hypotheses make the questionnaire design easier. A researcher should also be mindful that ethics clearance is required for all Monash University survey research (See: www.monash.edu.au/researchoffice/ethics.php).

It is very important also to evaluate that a survey is the best approach. Sometimes options such as in-depth interviews (5-10 people) or focus groups work quite well. However, if survey is agreed to be the way forward, one must decide on :

  1. How will it be administered?
    • In person, by you.
    • by someone else.
    • Telephone
    • internet (most common these days)
  2. Will the respondents need assistance (techical questions)?
  3. How long will it take ?
    • research shows ideally 3 questions (cognitive load 🙂 )
  4. Will it require follow-up?
    • Setup as anonymous or not?

Some of the important tips to note are:

  • Ask one question at a time.
  • Determine the form of response for each question (to allow analysis).
  • Provide some hints or prompts.
  • Good to include selection choices (limited choice, multiple choice, checklist, partially closed)
  •  Dont reinvent things if existing surveys and results exist. Draw from literature.
  • Use simple , non-threatening words (use terminology of audience).
  • Try to keep questions short.
  • Avoid ambitious words, leading questions, making assumptions, persuasive tone (may produce bias) and generalisations.
  • Use funnel approach (start with broad questions and narrow down).
  • Place important questions as far forward as possible (leave demographic to the end).
  • Always see a statistician before you go out and start collecting (the real) data.

Finally, one should remember that ordering of questions is also crucial as it can influence responses.