Page tree
Skip to end of metadata
Go to start of metadata


Creating Effective Surveys 
Brian Flaherty | Reading time: about 8 min

From working out what you want to achieve to providing incentives for respondents, survey design can take time. But when you don’t have hours to devote to becoming a survey-creation guru, a quick guide to the essentials is a great way to get started.

In this article, we’re going to reveal how to create a survey that’s easy for survey respondents to complete, hits the research questions you’re interested in, and produces data that’s easy to work with at the analysis stage.

1. Define the purpose of the survey

Before you even think about your survey questions, you need to define their purpose.

The survey’s purpose should be a clear, attainable, and relevant goal. For example, you might want to understand why customer engagement is dropping off during the launch of a new online product. Your goal could then be something like: “I want to understand the key factors that caused engagement to dip upon launch of the new product, including both internal and external elements.” Or maybe you want to understand customer satisfaction with a product or service. If so, the goal of your survey could be: “I want to understand how customer satisfaction is influenced by customer service and support, including online and offline channels.”

The idea is to come up with a specific, measurable, and relevant goal for your survey. This way you ensure that your questions are tailored to what you want to achieve and that the data captured can be compared against your goal.

2. Make every question count

You’re building your survey questionnaire to obtain important insights, so every question should play a direct role in hitting that target. Make sure each question adds value and drives survey responses that relate directly to your research goals. For example, if your participant’s precise age or home state is relevant to your results, go ahead and ask. If not, save yourself and the respondents some time and skip it.

It’s best to plan your survey by first identifying the data you need to collect and then writing your questions. You can also incorporate multiple-choice questions to get a range of responses that provide more detail than a solid yes or no. It’s not always black and white.

3. Keep it short and simple

Although you may be deeply committed to your survey, the chances are that the respondents, well... aren’t. As the survey designer, a big part of your job is keeping their attention and making sure they stay focused until the end of the survey. Respondents are less likely to complete long surveys or surveys that bounce around haphazardly from topic to topic. Make sure your survey follows a logical order and takes a reasonable amount of time to complete. Although they don’t need to know everything about your research project, it can help to let respondents know why you’re asking about a certain topic. Knowing the basics about who you are and what you are researching means they’re more likely to keep their responses focused and in scope.

4. Ask direct questions

Vaguely worded survey questions confuse respondents and make your resulting data less useful. Be as specific as possible, and strive for clear and precise language that will make your survey questions easy to answer.

It can be helpful to mention a specific situation or behavior rather than a general tendency. That way you focus the respondent on the facts of their life rather than asking them to consider abstract beliefs or ideas. Different question types will also allow for a variety of clear answers that help to uncover deeper insights. Good survey design isn’t just about getting the information you need, but also encouraging respondents to think in different ways.

5. Ask one question at a time

Although it’s important to keep your survey as short and sweet as possible, that doesn’t mean doubling up on questions. Trying to pack too much into a single question can lead to confusion and inaccuracies in the responses.

Take a closer look at questions in your survey that contain the word “and” – it can be a red flag that your question has two parts. For example: “Which of these insurance providers has the best customer support and reliability?” This is problematic because a respondent may feel that one service is more reliable, but another has better customer support.

6. Avoid leading and biased questions

Although you don’t intend them to, certain words and phrases can introduce bias into your questions or point the respondent in the direction of a particular answer.

As a rule of thumb, when you conduct a survey it’s best to provide only as much wording as a respondent needs to give an informed answer. Keep your question wording focused on the respondent and their opinions, rather than introducing anything that could be construed as a point of view of your own. In particular, scrutinize adjectives and adverbs in your questions. If they’re not needed, take them out.

7. Speak your respondent's language

This tip goes hand-in-hand with many others in this article – it’s about making language only as complex or as detailed as it needs to be when conducting great surveys.

Create surveys that use language and terminology that your respondents will understand. Keep the language as plain as possible, avoid technical jargon and keep sentences short. However, beware of oversimplifying a question to the point that its meaning changes.

8. Use response scales whenever possible

Response scales capture the direction and intensity of attitudes, providing rich data. In contrast, categorical or binary response options, such as true/false or yes/no response options, generally produce less informative data.

If you’re in the position of choosing between the two, the response scale is likely to be the better option.

Avoid using scales that ask your target audience to agree or disagree with statements, however. Some people are biased toward agreeing with statements, and this can result in invalid and unreliable data.

9. Rephrase yes/no questions if possible

As we’ve described, yes/no questions provide less detailed data than a response scale or multiple-choice, since they only yield one of two possible answers.

Many yes/no questions can be reworked by including phrases such as “How much,” “How often,” or “How likely.” Make this change whenever possible and include a response scale for richer data.

By rephrasing your questions in this way, your survey results will be far more comprehensive and representative of how your respondents feel.

10. Start with the straightforward stuff

Ease your respondent into the survey by asking easy questions at the start of your questionnaire, then moving on to more complex or thought-provoking elements once they’re engaged in the process.

This is especially valuable if you need to cover any potentially sensitive topics in your survey. Never put sensitive questions at the start of the questionnaire where they’re more likely to feel off-putting.

Your respondent will probably become more prone to fatigue and distraction towards the end of the survey, so keep your most complex or contentious questions in the middle of the survey flow rather than saving them until last.

11. Take your survey for a test drive

Want to know how to make a survey a potential disaster? Send it out before you pre-test.

However short or straightforward your questionnaire is, it’s always a good idea to pre-test your survey before you roll it out fully so that you can catch any possible errors before they have a chance to mess up your survey results.

Share your survey with at least five people, so that they can test your survey to help you catch and correct problems before you distribute it.


Survey Design Best Practices: Learn about a survey life cycle, general guidelines and best practices in survey design, and how to avoid common problems while gaining insight with this popular research method. More information and registration coming soon!



Brian Flaherty
Brian is currently a Senior Design Strategist with the Human-Centered Design Center of Excellence (HCD CoE). The HCD CoE is an organization that impacts the way the CCSQ delivers policy, products and services to its customers. Brian has been a graphic designer for more than 25 years, and has been practicing human-centered design for 11. Prior to joining Tantus as an HCD Strategist, Brian spent 12 years as a Creative Director, Communications Supervisor, and HCD Practitioner at the Johns Hopkins University Applied Physics Laboratory supporting classified and unclassified communications, primarily for the Department of Defense. Brian holds a BA degree from the University of Pittsburgh where he majored in Creative Writing and Public Relations. Brian is married, has a daughter just about ready to begin college, and considers two cats, two dogs, 26 chickens, three ducks, a crested gecko, and an Alpaca named Skinny Pete as his step children.


     






  • No labels