Writing better questions by focusing on the participant

Writing better questions by focusing on the participant

By Ray Poynter, 24 August 2020

At Potentiate, we are committed to being human-centric in everything we do. Focusing on Human Experience (HX) has major implications for how online questionnaires are created and, in particular, the construction of questions.

The four basic rules
There are four basic rules that determine whether a question is suitable:
I. People have to understand what is being asked. I remember one survey that asked people “How often do you encounter a 404 error?” We need to think about (and research) which terms are understood by the people we are talking to.
II. People should be able to answer the question. For many people, questions such as “How many units of alcohol do you consume in a typical month?” are understood, but people’s memories are not reliable, they may not have a ‘typical month’, and they may not think of their drinks in terms of ‘units’.
III. People need to be willing to answer the question. Questions such as “Do you cheat on your expenses?” and “Do you take illegal drugs?” may not attract truthful answers.
IV. The researcher must be able to interpret the answers. If we ask a double-barrelled question, such as “Was the train clean and on time?”, then the answer “No” is hard to interpret. Was the train clean but late, dirty but on time, or dirty and late?


10 Human-centred tips for writing better questionnaires and questions

Here are ten key tips for making sure your questions and questionnaires focus on being human-centric:

1. The questionnaire should work well on a phone. We need to respect the choices that our customers and participants make, and for many that means completing questionnaires on their smartphone.

2. In the answer list for every question, there should be a suitable option for every person taking the survey. This means thinking about whether the question needs a Don’t Know, None of these, Not Applicable or Other option adding. For example, “Which of these did you eat in the last month? Beef, Chicken, Pork” needs a None of these. The question “What is your weight in kilograms?”, needs a Don’t Know option.

3. If the question is a ‘choose one’ question, there should only be one option suitable for each person. For example, the question “Which type of wine do you drink (select one) White, Red, Sparkling, None of these” could be altered to “Which of these is your favourite type of wine? (select one) White, Red, Sparkling, None of these” or it could be changed to a multi “Which of these wines do you drink? (select all that apply) White, Red, Sparkling, None of these”.

4. Avoid leading questions. “Should the police do more to tackle the growing problem of anti-social behaviour?” is obviously a leading question. But so is “Is French wine better than Australian wine?” – acquisition bias means people are more likely to agree – so this question would better written as “Which of these is best? Australian wine, French wine, Neither, Both” – with the order of Australian and French wine being randomized between participants.

5. Intuitive. If the question is long (more than one line), many people won’t read it. The longer the question gets, the more people will feel TLDR (too long didn’t read).

6. Only collect sensitive data if it is needed. If you do need to collect it, explain to the participant why you need it and ensure you have the right policies in place to keep it secure.

7. Try to use questions that have been used/tested before. You will reduce the risk of confusing and offending people if you use questions that have been tested and honed over time. If you do create a new question, test it, check it works for the participant and that it delivers what the end client needs.

8. Only force questions when necessary. We should recognise that even the people who are happy to share their views with us do not want to be ‘forced’ to answer every question. This is especially true of open-ended questions – there are lots of reasons and situations where forcing an open-ended question is too onerous.

9. Keep the survey as short as possible. How short? Less than five minutes is good, less than ten minutes is often OK, more than 15 minutes is usually too long.

10. Ask a closing open-ended comment question. People often want to tell us something about the topic we are researching, or about the questionnaire we have used – we should respect that and finish with something along the lines of “Is there anything else you’d like to tell us about this topic or this questionnaire?”

Balancing Research & Engagement
A questionnaire has two key constraints, and these sometimes pull in different directions. We want a research instrument that allows people to provide us with the information we need. However, we also want the questionnaire to be engaging, so that people feel valued and want to take part. This requires balance, neither can be the key driver if we want to adhere to the principles of HX, human experience.


To find out more about how we can help you deliver exceptional human experiences,
contact us today.

Valid name is required.
Valid company name is required.
Please enter a valid email address.
Please enter your phone number.
Please select a valid country.