Your survey answers are as important to collecting high-quality data as the question you write. So, we’ve provided some tips on writing answers options for survey questions, from multiple choice to Likert scales.
Top tips for writing survey answers
1. Answers should be clear and concise
Ensure that respondents understand each answer option clearly.
E.g.
- Always
- Very Often
- Sometimes
- Rarely
- Never
The difference in meaning between ‘Sometimes’ and ‘Rarely’ isn’t obvious. But if you need to include words closely similar in meaning, it’s a good idea to define each.
You can use a text box from Shouts’ Presentational Items for this purpose.
2. Don’t use ‘extreme absolutes’
This is a suggestion for both your question and answer text. These are words such as ‘All’, ‘Any’, ‘Best’, ‘Worst’, ‘Never’, ‘Always’ and tend to assume that respondents will fully agree or disagree with a statement.
In truth, respondents aren’t likely to fully agree or disagree with a statement and are more responsive to specifics.
E.g. ‘Once a Day – Once a Month’ rather than ‘Never – Always’.
3. Answers should be ‘collectively exhaustive’
All relevant survey answers should be made available to respondents.
For example, when collecting data on customer expectations:
Question: ‘What level of care are you expecting for the duration of your stay?’
Scale: Excellent | Very Good | Good | Fair
In this case, all expected answers have been provided. There’s no need to include the negative answer options, as no one would expect a poor level of care.
4. Limit your answer options per question
Offering too many or too few answer options is bad practice. Too few answer options may be restrictive, whilst too many can be confusing.
If you have more than four or five answer choices, your questions may be too broad. Ask yourself if they can be split into two or simplified in some way.
5. Provide ‘No Opinion’ answer options (when necessary)
By including a ‘neutral’ answer option, such as ‘N/A’, ‘Prefer not to Answer’ or ‘No Opinion’, you give respondents the choice to opt-out. If they aren’t given this choice, they have to select another answer. Which means that data isn’t accurate.
However, you should read more about the use of ‘no opinion’ answer options to determine what effects they may have on your research.
6. Good survey answers are ‘Mutually Exclusive’
No two answers should be so similar that respondents are unsure of which to pick.
Below is the most common example of this issue.
Poor Example:
Select Age Bracket: 18-25, 25-35, 35-45, 45-55, 55-65
Good Example:
Select age bracket: 18-25, 26 -35, 36-45, 46-55, 56-65
In the first example, respondents wouldn’t know which age bracket to pick if they were 25, 35, 45 or 55.
7. No more than 7 options for scaled questions
Try not to exceed more than 7 answer options for scaled questions, i.e. 3 positive answers, 1 neutral answer and 3 negative answers.
Too many choices can be taxing for respondents and will lower your completion rates.
This only really applies to Matrix question types (Likert scales). An Opinion Scale can contain up to 10 or 11 points (e.g. NPS ratings).
8. Use labels for rating scales
When creating a rating scale, always remember to label each point. Leaving these fields blank will only cause ambiguity.
For example, if you have a Net Promoter Score question you’d be asking: “How likely are you to recommend this product/service to a friend or colleague?”. Then you’d set a scale from 0-10.
You must label one end ‘Not at all Likely’ (0) and the other ‘Extremely Likely’ (10), or you may confuse participants. And this is the last thing you want to do when asking for a review of your product or service.
Learn more about how to measure customer loyalty and calculate NPS.
You should also ensure that the opposing labels are directly contrasting. I.e. if one end of the scale is ‘Very Good’, the other end should be ‘Very Poor’ and not ‘Poor’ or ‘Awful’.
9. Convert scales to contrasts
You don’t always have to do this, but if you’re using a lot of scaled questions it might be better to re-frame some of them as contrasting statements instead.
Rating scale question:
Question: ‘How much do you agree with this statement: There is a good work-life balance in my department’
Scale: Strongly Disagree | Disagree | No Opinion | Agree | Strongly Agree
Contrasting statement question:
Question: ‘Is there a good work-life balance in your department?’
A: ‘There is a positive work-life balance’
B: ‘There is a negative work-life balance’.
If someone constantly answering the same question type, there likely to suffer from survey fatigue. Filling out an online survey isn’t the most exciting task and so it’s easy for respondents to lose interest or motivation.
You want to do everything you can to keep them engaged.
This also occurs if answers are repeatedly in the same order for scaled (or any multiple choice) questions.
10. Switch the direction of scale options
Some respondents may speed through your survey, selecting survey answers that don’t reflect their opinion. With scaled questions, this usually manifests as straight lines in your survey data (i.e. always picking the same point of a scale for each question).
This is more likely to occur is you’re offering an incentive of some kind (prize/ promotion) but can be caused by survey fatigue or disengagement.
Unfortunately, there’s no real way of preventing respondents doing this (other than ensuring your survey is as engaging as possible). However, you can reduce the effect of this on your results by reversing the direction of scales for each question.
E.g.
Question 1: Strongly Disagree – Strongly Agree.
Question 2: Strongly Agree – Strongly Disagree.
This ensures that your data set isn’t weighted in favour of these respondents’ answers. Meaning, one end isn’t extremely disproportionate to the other.
If you’ve already written all your questions and answers out, Shout has a feature for reversing the points of your scale.
11. Always balance your scales
You should always have an equal number of positive and negative answer choices for scaled questions.
E.g. Very Poor | Poor | Good | Very Good | Excellent
In the example above, you can see there are more positive answer choices than negative ones.
As a result of this, response bias could be created. For instance, respondents could be more likely to pick a positive answer because there’s more of them or start picking negative answers because they feel they’re being led the other way.
By keeping survey answers balanced, you reduce the number of variables influencing respondent choices.
12. Provide a neutral scale option
Respondents may not have an opinion on a subject or be unable to decide how they feel. To prepare for these cases, you’ll want to provide a neutral option. These are usually found at the centre of a Likert scale.
E.g. Very Poor | Poor | No Opinion | Good | Very Good
These options can be very informative. Say you’re launching a customer satisfaction survey and receive a high percentage of respondents selecting ‘No Opinion’.
This could be an indicator that you’re putting too much emphasis on the wrong aspect of service or product, and your customers are looking to be engaged in some other way.
With close-ended questions, the answer choices are just as important as the question text when it comes to gathering data.
But with open-ended questions, there’s more pressure on how you frame your question if you want the same quality in survey responses. Luckily, we’ve written an article on how to write survey questions.
If you just want some help improving your project, take a look at these survey design best practices.
Other than that, let us know if you have any questions or tips for writing survey answers to be added to this list.
