Survey Design

How Survey Question Wording Can Skew Responses [Video]

How much can a single word in a survey question shift the answer? Substantially more than most survey designers think. This short video walks through a real example — "easy" vs "difficult" in the same question — and shows why word choice quietly biases your data before the respondent even gets to the scale.

By Adam Ramshaw 2 min read
How Survey Question Wording Can Skew Responses [Video]
How Survey Question Wording Can Skew Responses

A simple change in words can influence the response. Choose your words wisely as it may skew your results!

Transcription of this Video

Hi, my name’s Adam Ramshaw. I’m from Genroe.

Today’s customer feedback question is

“How much can one word in the question influence the answer?”

When we write a customer survey, we’re really trying to get a good understanding of what our customers think. So we really want to influence them as little as possible in changing what the score would be, or changing how they would respond.

Let me give you an example of how you might do this.

You might write a question that says,

“Please write how easy it was to contact our service desk, from very easy to very difficult?”

Now you see we have the word “easy” twice in that question, once in the question and once in the answer.

Just by human nature people are more likely to answer closer to the very easy end of the scale than the very difficult.

Skewing the results like this is something that you really don’t want to do. You want to get a clear understanding of what the customer is thinking. You really don’t want to influence them.

A much better way to ask the same question would be to simply say:

“How did you find contacting our service desk, from very easy to very difficult?”

Same idea, but now you’re not influencing people and influencing the response they’re going to give you. So you get much more accurate information.

One final thought to think about when you write questions is to try to be as neutral in the wording and in the lead-in to the question as you possibly can.

Try not to do things like saying:

“We try to provide excellent service. How was our service today?”

By adding that first sentence in you’re going to change the way the customers respond.

You really don’t want to predispose them to answer one way or the other. You want to get that neutral response from the customer.

Back to Blog

Related Posts

View All Posts »
The Only 6 Training Survey Questions You Will Ever Need

The Only 6 Training Survey Questions You Will Ever Need

Most training feedback forms ask 15 questions and tell you nothing actionable about the course. After analysing tens of thousands of responses across hundreds of training surveys, six questions consistently emerge as the only ones that matter. Here's the list, why each one earns its place, and the popular questions you can confidently drop.

The 11 Worst Customer Feedback Question Mistakes You Can Make

The 11 Worst Customer Feedback Question Mistakes You Can Make

Customer surveys vary a lot; the mistakes inside them really don't. The same eleven question errors keep showing up across industries — mandatory questions, missing N/A options, asking what you already know, the 30th question. Here's the list, with how to identify each one and the fix that takes minutes, not weeks.

Club Med’s Customer Survey Fail and How You Can Do Better 

Club Med’s Customer Survey Fail and How You Can Do Better 

Club Med's post-stay survey arrives personalised, well-timed, and immediately falls apart — images-as-content, question overload, and a hidden lead-gen form at the end. Here's a teardown of where a well-resourced brand's customer survey went wrong, and the specific mistakes that are worth not copying in yours.

How Long Should My Customer Survey Be? [Video]

How Long Should My Customer Survey Be? [Video]

How long should your customer survey be? "As long as it needs to be, but no longer" is a fine principle and a useless answer. Here's a sharper rule of thumb, split by B2C and B2B, with concrete question counts — drawn from the analysis of thousands of real survey response patterns.