My last two blogs looked at ways in which you can ensure your questionnaire is a vehicle which will gather great data, instead of looking like a jumbled mess of words! If you haven’t read Parts 1 and 2, the links are here and here.
This final blog in the series looks specifically at the structure of the questions you might ask and advises you on five things to avoid.
INCOMPLETE ANSWER OPTIONS
Time and time again I’ve abandoned a survey part-way through because a question I’m being asked doesn’t have an option that applies to me. Look at this example –
Q Which is your favourite boy band?
Westlife
Boyzone
JLS
The Wanted
This question forces people to choose from a list of four options.
Where do I tick to show my preference for Take That?
Where do I tick to show that I don’t have a favourite boy band, because I love thrash metal?
The results won’t be a true representation of opinions on favourite boy bands. This is because the question actually being asked is, “From this list of four, which is your favourite boy band?”
At the very least you need to allow an “Other” option. The ideal version would show an alphabetical list of 20 or so boy bands plus an “Other – please specify” option and an “I don’t like boy bands” option.
BIASED SCALES
I see this ALL.THE.TIME. The example below is about gathering satisfaction with customer service, but four of the five options indicate a positive response. The results will therefore be skewed towards a positive response. So, yes, you are likely to be able to say that most people are satisfied, but because of the poor question structure those results will be meaningless.
Q How satisfied are you with the level of customer service we provide?
Extremely satisfied
Very satisfied
Quite satisfied
Satisfied
Not satisfied
The example below has a balance scale, with an equal number of options available for positive and negative replies, along with a middle ground option for those who don’t have a strong opinion either way.
Q How satisfied are you with the level of customer service we provide?
Very satisfied
Quite satisfied
Neither satisfied nor unsatisfied
Quite unsatisfied
Very unsatisfied
VAGUE QUESTIONS
Be direct – if you want to know about customer service then ask about customer service. Don’t be vague.
Q Do you like our company?
A Well, I quite like the name, and I love Jane because she always answers the phone really quickly. But last time I ordered it took four weeks to arrive, and then my order was wrong.
Which bits of that answer are relevant for you?
LEADING QUESTIONS
Questions should be neutral. Don’t assume anything. Don’t ask a question which, by its nature, sways respondents to answer in a particular way. You will alienate people if you assume that they have certain behavioural traits. And, again, your data won’t be meaningful because you’ve led respondents to answer in a particular way. These examples show you what I mean –
Example 1
Q Please tell us what you love most about our fabulous new and exciting website?
You are leading people with the words “fabulous”, “new” and “exciting”. Much better to ask..
Q Please tell us what you like best about our website?
Example 2
Q Where do you enjoy drinking tea?
You are assuming people enjoy drinking tea. Much better to ask..
Q Do you drink tea? Yes/ No
Q If yes, where do you enjoy drinking tea?
ARE YOU CONDUCTING RESEARCH?
I hope that this series of blogs about writing a great questionnaire has been useful. Hopefully they’ve given you ideas and tips to help with your own questionnaire design process.
If you’re thinking of conducting research and would like some advice about questionnaire design, then please give me a call – I’ll be happy to talk through the options with you. Or, if you want the whole thing taken off your hands, then I can help you with that too.
