Skip navigation

Top 10 Tips for Survey Writing

Posted 11th September 2014 | Category: MarketResearch

I was recently sent a postal survey which on first opening looked to be from the NHS. Through a combination of professional curiosity and trying to be a good respondent, I decided to complete it. As I began to read through the opening instructions and first few questions I realised that this was in fact a survey from the Conservative party. On the principle of misleading respondents I changed my mind about filling it out!

A few weeks later I noticed a tweet from the MRS with a link to an article about this same survey -  Jane Frost had written to the Conservative party about how the wording at the end of the survey broke several code rules about correctly informing respondents how their data will be used:

"..As such I am compelled to alert you that your " What matters most to you?" survey is in breach of a basic legal requirement as it fails to specify the purpose for which the data collected will be used. I am afraid to say that it is also unethical as it is being published under the guise of research but is in effect being used as means to build marketing or contact databases."

As well as the legal side of things, it's important to remember that the way we write surveys should serve the respondents as well as our clients. Properly informing participants of the purpose of the survey and the way in which their responses will be used not only complies with MRS regulations but can also make for better survey results - understanding the question helps respondents to make more informed response choices!

Top 10 Tips for Survey Writing

1. Only ask one question at a time, for example asking respondents both the likelihood of purchase of a product and a timescale of purchase in one question will probably not give you the answers required! 

Try breaking down questions so that one follows logically from the other

2. Consider the options you give in a question with a scale and ensure all opinions can be covered! If I was to ask you "How would you rate this blog?" and gave response options a) Good, b) Great and c) Excellent how would you respond?!

3. Speaking of scales...make sure they don't overlap or miss answer options:

You might want to find out the age bracket of respondents and give options such as: a) 18-24 b)24-35 c)35-45 - this would likely end up with some double counting as a respondent aged 24 could answer both a) & b)! 

Similarly, asking a respondent the turnover of their company and giving options such as a) £10,000 - £50,000, b) £51,000 - £100,000...what do you pick if the turnover is £50,500?!

4. Avoid unnecessary questions / overly long surveys - think about what you are going to use the responses from a particular question for. If the answer is "not a lot" then don't ask it! 

5. Avoid the use of jargon and abbreviations where possible. You want your respondents to understand what you are asking them! If you absolutely have to use abbreviations make sure these are explained somewhere in the question text.

6. Make sure that filters and question skipping works effectively before launching a survey. For example, asking, "How many diets have you tried in the last 3 years?" with an answer option of "None", then asking a follow up question of, "Which of these diets have you tried?" without filtering out those who answered "None" in the original question is very annoying for the respondent and means you will end up with some fairly useless data!

7. Speaking of those Don't Know/None questions ... ensure these are listed as options where applicable 

8. Ensure that the answer options you give match the question you ask! Seems obvious I know..

9. Avoid unnecessary repetition of questions

a. Avoid unnecessary repetition of questions...Haha just kidding!

10. And the golden rules:

a. Clearly explain the purpose of the survey

b. Explain who you are and reassure respondents that the research is genuine

Post a comment

Comment Form

There are currently no comments on this blog.