How to write good survey questions

Explain why you’re asking these questions

Respondents are more likely to answer your survey questions if they can see value in them. Explain the purpose of your survey right at the beginning – to improve customer service or make a product better, perhaps. And when you can make the value specific to the respondent, you’ll get a better response rate.

Speak your audience’s language

A good rule of thumb is to write for your least-informed respondent. Use clear language and terminology to make it easy to understand what you are asking. It’s important to avoid technical jargon and acronyms, as these unintentionally create survey bias – only some people in your audience know what you are talking about, which will alienate others and skew your data.

Keep the questions simple

Make your questions short, simple and direct. When you increase the length of questions and surveys, you decrease the chance of receiving a completed response. Respondents will get bored, frustrated, irritated – and drop out of the survey. Questions that are vague and fail to communicate your purpose can also limit the usefulness of your results.

Here’s an example of a vague question:

What do you like to do for fun?

Finding out that respondents like to play Scrabble isn’t what the researcher is looking for, but it may be the response received. It is unclear that the researcher is asking about movies vs. other forms of paid entertainment. A respondent could take this question in many directions.

A better, more specific question would be:

In your spare time do you watch movies online?

It costs money to find respondents, so it pays to keep them engaged to the end.

‘Funnel’ your questions

The funnel technique is a way of arranging your survey questions in the best sequence to engage respondents. Use the following structure:

1. Start with broad, general interest questions that are easy to answer, ‘warm up’ the respondent and get them involved in the survey
2. Place the harder, more specific questions that take time to think about and answer in the middle
3. Finish with easy-to-answer questions, such as demographics.

Survey question types

Did you know that we provide 23 question types you can use in your surveys? Some are popular and used frequently by a wide range of people from students to market researchers, while others are more specialist and used to explore complex topics.

To write a successful survey, you’ll need to understand the different survey question types and how they work. Each type needs a slightly different approach to question-writing. Here are some of the most common ones, and how to write them well.

Multiple choice

Familiar to many, multiple choice questions are the longest and most complex. They ask respondents to pick from a range of options. You can set up the question so that only one selection is possible, or allow more than one to be ticked.

When writing a multiple choice question:

• Be clear about how many a respondent should choose: one (Pick only one) or several (Select all that apply).
• Think carefully about the options you provide, since these will shape your results data.
• The phrase ‘of the following’ can be helpful for setting expectations:

For example, if you ask: What is your favourite meal? and give the options hamburger and chips, spaghetti and meatballs, there’s a good chance your respondent’s real favourite won’t be included. If you add of the following the question makes more sense.

• Make multiple choice response options mutually exclusive so that respondents can make clear choices. Don’t create ambiguity for respondents. Review your survey and identify ways respondents could get stuck with either too many or no single, correct answers to choose from.

For example, spot the problem here:

What is your age?

0–10
10–20
20–30
30–40
40+

What answer would you select if you were 10, 20, or 30? Ambiguous survey questions like this will frustrate respondents and invalidate your results. 

Rank order

You can use a rank structure to ask respondents to rank things in order, such as preference, frequency or perceived value. There’s a variety of tools, including drag-and-drop, radio buttons, tick boxes and more.

When writing a rank order question…

• Explain how the tool works and what the respondent should do to select their choice.
  • For example: Drag and drop the items in this list to show your order of preference.
• Be clear about which end of the scale is which.
  • For example: With the best at the top, rank these items from best to worst.
• Be as specific as you can about how the respondent should consider the options and how to rank them.
  • For example: Thinking about the last 3 months’ viewing, rank these TV streaming services in order of quality, starting with the best.

Slider

Sliders ask the respondent to move a pointer or button along a scale, usually a numerical one, to indicate their answers.

When writing a slider question…

• Consider whether the question format will be intuitive for your respondents, and whether you should add help text such as Click/tap and drag on the bar to select your answer.
• Our surveys include an open field option where your respondent can type their answer instead of using a slider. If you use this, make sure to reference it in the survey question so the respondent knows what to do.

Text entry

Also known as an open field question, this format allows respondents to type their own answer into the box.

When writing a text entry question…

• Use open-ended questions such as How do you feel about…, If you said x, why? or What makes a good x?
• Use text entry questions sparingly, as open-ended questions take more effort to answer.
• Be as clear and specific as possible when framing the question and give respondents context to help make answering easier.
  • For example, rather than, How is our customer service? write: Thinking about your interaction with us today, what could we do better?

Matrix table

Matrix structures allow you to address several topics using the same rating system, for example a Likert scale (Very satisfied – satisfied – neither satisfied nor dissatisfied – dissatisfied – very dissatisfied).

When writing a matrix table question…

• Make sure the topics are clearly differentiated from each other, so that respondents don’t get confused by similar questions placed side by side and answer the wrong one.
• Keep the text brief and focused. A matrix includes a lot of information already, so make it easier for your respondent by using plain language and short, clear phrases in your matrix text.
• Add detail to the introductory text if necessary to help keep the labels short.
  • For example, if your introductory text says, In the Birmingham store, how satisfied were you with the… you can make the topic labels brief, for example, staff friendliness, signage, price labelling

Ask single questions only

What do you think of our new store layout and our refurbished café? This is an example of a ‘double-barrelled’ question where the respondent is forced to answer two questions at once. It can cause problems for respondents and influence their responses.

They may either skip the question or refer to just one in their answer – and you won’t necessarily know which. Better to create two single questions:

• What do you think of our new store layout? and
• What do you think of our refurbished café?

Avoid survey bias

Biased survey questions are worded in such a way as to favour one particular answer over the others, and they give you misleading data. There are many types of biased questions, the most common being:

• The leading question: For example: How dissatisfied are you with our delivery service? assumes everyone’s unhappy with the service. What do you feel about our delivery service? is a more neutral question.
• The loaded question: For example: Where do you go shopping at the weekend? assumes everybody does this and excludes those who don’t. What do you do at the weekend? is easier for everyone to answer.
• The absolute question that requires only yes or no answers creates bias because it doesn’t reflect a bigger picture, for example: Do you eat bread every day? Many people don’t every single day of the week, so they simply answer no. How many times a week do you eat bread? gives you much more quantitative data.

Don’t force participants to answer

Respondents may not know the answers to all your questions, and there may be some questions they simply don’t feel comfortable answering. Privacy is an important issue for most people and sometimes respondents may not want to provide you with certain types of information.

Examples of sensitive questions include:

• What is your annual income?
• What is your occupation?
• How many children do you have?
• How did you vote in the last election?
• How often do you shower?

Only ask such questions when they are absolutely necessary to your survey, and justify why you need them – for benchmarking, or only within your organisation, for instance. Incentives and assurances of confidentiality may make it easier to obtain private information.

Always include a Prefer not to answer option for sensitive questions.  While you may not get an answer, respondents won’t feel forced into giving up sensitive information. And if respondents feel they’re being forced into answering, there’s a risk they may select a random answer or abandon the survey.

Include a Not applicable or Don’t know option, as some respondents can’t or won’t answer certain questions because they lack experience or aren’t sure how to respond.

If you find you are getting a disproportionate number of Prefer not to answer/Not applicable/Don’t know answers, take the opportunity to revisit your survey and check that it’s fit for your purpose and sufficiently engaging.