What triggers survey dropout, and how do you stop it?

Learn why people might not be finishing your surveys and what you can do to see better engagement and overall data quality.
14 September 2020
online survey engagment
May Ling Tham
May
Ling Tham

Director, Modern Survey Design, Profiles Division

Contact us

When people get bored taking a survey, expect to see an impact on your data in addition to early exits.

This is an important fact to be aware of, but what’s equally vital is understanding what you can do in the way of survey design to prevent dropout due to respondent boredom. Here are three examples of the impact of respondent fatigue, and five simple steps to reduce it and improve overall data quality.

1. Inconsistent answers

As respondents get bored completing surveys, their overall engagement starts to drop. They start to click less and it can worsen throughout the survey duration. This results in a lack of attention and leads to different answers at the beginning versus the end of the survey.

2. Neutralised responses

Answers in Likert scales and sliders become more neutralised as people get bored. Kantar ran an experiment by moving the same questions from the start of a survey to the end and compared the answers. The neutral answer or “I don’t know” responses rose by 18%.

We also saw very positive and very negative responses falling significantly, by an average of 25%. A reduction in the variation of answers also dropped as respondents got bored. This is reflected in reductions in the levels of Standard Deviation, delivering less clarity in results as less consideration is given by the respondent in their answers on a scale.

3. High dropout levels

Repetitive questioning can trigger significant levels of dropout. A list of 14 statements compared to six sees a difference of 10% in dropout rate (12% to two percent respectively). Furthermore, the survey enjoyment score decreases by nine percent, suggesting that fewer repetitive lists is more engaging for respondents to complete.

The longer the survey the higher the dropout rate as well. Kantar has found that a survey over 25 minutes loses more than three times as many respondents as one under five minutes. We see that more people complete shorter surveys because it’s easier to keep them engaged.

What survey design techniques can you apply to reduce survey length, improve engagement and maximise the quality of your overall research findings?

1. Think like a respondent

You may be reading the examples above and thinking they are common sense. Well in essence, they are. Write your questionnaire from the perspective of person taking it. Try completing it yourself or ask your colleagues to. Taking the survey yourself will give you the understanding of how a respondent will navigate the survey and what needs to be addressed for an easier, clearer and more engaging experience.

2. Hunt out the redundant questions

One of the first things to look for in a survey that needs trimming is redundant questions: questions that give roughly the same or similar answers to other questions in the survey. These redundancies are commonly found in banks of attitudinal statements and word-association lists. There are some sophisticated statistical ways to analyse this, but the simplest test is to run is a basic Pearson correlation. A Pearson correlation is a standard Excel function that compares the answers of every question with every other in the survey and eliminates those with high levels of correlation.

3. Make use of iconography

Generally, respondents don’t like processing long lists because it requires either memory or reading the list several times. Using iconography can alleviate those issues. A good icon, once read, acts as a quick visual prompt that can be processed in a fraction of a second. The icons reduce the memory load of choices for the respondent, making the question much more straightforward to answer.

4. Switch to shorter, more effective question techniques

You can reduce the amount of work involved in answering questions by simply changing the way you ask the question. For example,

  • Ask pre-shortening or filter questions to create less work for the respondent and to more efficiently identify the key issues. Preliminary filter question will also illuminate the most important factors.
  • Use selective ranking by asking respondents to “rank the top 5 factors”. This saves 15 seconds on average and reduces the minimum sample requirements to roughly 300 to gain statistical clarity.
  • Use choice based max differentiation by presenting small random sub-sets of say five choices and asking them to pick the most and least important factor.

5. Ask more rewarding questions

Shorter surveys will yield more engaging experiences, but so will smarter, rewarding questions. Here are a few examples;

  • Switch a rating scale to a binary question. They can deliver more reliable, better differentiated answers.
  • Ask respondents to participate in a gamified scenario to achieve higher concentration and thought out answers
  • Use predictive techniques that ask respondents to solve a problem, requiring more brainpower that can reach far more considered and accurate answers.

Learn more

Speak to our expert team using the 'contact us' button, or use the form below to download 12 tips for writing stronger survey questions.

Keep an eye out for the October launch of Kantar’s Survey Design Training modules to build on your skills.

Contact us
Related Solutions
Our Modern Survey Design techniques help you understand people better, by asking the right questions in the right way.
Survey programming and fieldwork services, from fully customised research designs to sample-only fieldwork.
Flexible quick-turn survey solutions designed to get you answers when you need them.