Questionnaire Design

Questionnaire Construction: Open and Closed Questions

A well-constructed questionnaire is essential for collecting valid and reliable self-report data. Poor questionnaire design introduces systematic biases that distort findings regardless of sample size or analytical sophistication. Key decisions in questionnaire construction include the choice of question type (open vs closed), the wording of individual items, the structure and order of the questionnaire, and the format of response options.

Open Questions

Open questions invite participants to respond in their own words, without constraining the response to pre-specified options. They produce qualitative data — rich, detailed, and potentially unexpected content. Open questions are most appropriate when:

• The researcher does not know in advance all the possible responses.
• The research aims to understand participants' perspectives in depth.
• Nuance, context, and individual variation are important.

Strengths: capture unexpected themes; allow participants to express complex views; avoid imposing the researcher's categories. Weaknesses: time-consuming to analyse (typically require thematic analysis or content analysis); difficult to compare across participants; responses vary enormously in length and detail.

Closed Questions

Closed questions offer pre-specified response options. They produce quantitative data that can be statistically analysed and directly compared across participants. Common formats include:

• Likert scales: participants rate their level of agreement on an ordered scale (e.g. 1 = Strongly disagree to 5 = Strongly agree), producing ordinal data.
• Rating scales: participants rate the intensity or frequency of an experience on a numerical scale.
• Dichotomous (yes/no): simple binary choices; quick but lose nuance.
• Multiple choice: participants select from a list of options; captures categorical data.

Strengths: quick to complete; easy to analyse statistically; allows direct comparison; standardised. Weaknesses: constrain responses to the researcher's predetermined options; may not capture the participant's genuine view if no option closely matches it.

Avoiding Bias in Question Wording

Several common pitfalls in question wording can introduce systematic bias:

• Leading questions: wording that implies an expected or preferred answer — e.g. 'Don't you agree that exercise is important for health?' Should be replaced with: 'How important do you think exercise is for health?'
• Double-barrelled questions: asking about two things in a single question — e.g. 'How satisfied are you with the quality and quantity of support you receive?' These should be split into two separate items.
• Loaded language: using emotionally charged words that predispose respondents towards a particular answer.
• Jargon or ambiguity: using technical terms or ambiguous words that participants may interpret differently.

Order and Structure

The order of questions can affect responses. Sensitive or personally revealing questions should be positioned after the questionnaire has been introduced and rapport established — not at the beginning. Demographic questions (age, sex) are typically placed at the end to avoid setting up categories that prime differential responding. Including reverse-scored items — statements worded in the opposite direction to the main construct — helps to identify acquiescence bias and produces more reliable scale scores.