Human-Centered Design (HCD) is an approach to problem solving that focuses on what is desirable for people. We focus on what people need and help develop solutions, technologies, and products that people will use while keeping viability and feasibility in mind. Our toolbox includes a wide range of methods to identify people’s needs and solve problems. This article on creating effective surveys is the first in our HCD Toolbox series.
We have found that conducting user research in the beginning stages of a business- or technology-related client project is a fundamental tool to uncover people’s needs, motivations, and goals. There are many different methodologies that we use to gather this information. For example, observations allow us to describe what people do and not what they say they do. Interviews can be used to explore people’s thoughts, motivations, and beliefs more deeply. Ethnographic Interviewing combines observation and interviews for immersive exploration of people’s behaviors in their natural environment. These techniques can be invaluable when you need to dive deeply into their why and have access to a smaller number of people.
The limitation to observation and interviewing, however, is that these techniques can only be conducted with an individual or small group. They require more time and cost, and the process of coding and synthesizing the qualitative information retrieved can be time-consuming.
Conducting surveys is an excellent way to validate information gathered from observations and interviews with a more significant number of individuals. Additionally, surveys allow us to ask more specific and targeted questions.
However, surveys can be easy for users to ignore or rush through without much thought, so your surveys need to be engaging, easy, and enjoyable.
Here, we cover strategies to ensure that your survey is designed to gain maximum attention and receive valuable, valid, and reliable input.
Why should you conduct a survey?
- Surveys are cost-effective
- They allow you to gather information from a larger number of people in a relatively short amount of time
- They help you to quantify users’ thoughts, feelings, motivations, etc.
- The results are less likely to be biased as long as questions are properly worded
- They are easy to distribute — online or in-person
When should you conduct a survey?
- Surveys are most effective if they are conducted after user observations and interviews
- Also, they can be an excellent tool to gather feedback during usability testing
What are your survey goals?
- Before you create a survey, it’s essential to be clear about why you are conducting the survey. What do you hope to learn?
- Then, ensure that your questions focus on those key objectives. This focus will help you eliminate questions that might be interesting but peripheral to your central question.
- Discuss the goals of the survey with your team so that everyone agrees on what you are hoping to learn.
- Decide how you plan to share the results of your survey. Who will be the audience for the survey results/ findings?
What’s the best approach to creating a survey?
Creating an effective survey that yields the desired, high-value feedback can be challenging and time-consuming. Using tried-and-true samples can be an excellent way to get started.
If you decide to create a survey from scratch, you want to ensure that you are asking the right questions. We believe that the most effective survey creation process is iterative and collaborative. Be prepared to develop various versions of the survey and seek out feedback from stakeholders who could provide balanced and useful feedback.
If you are sending surveys online (e.g., via email) you may want to include a deadline to complete the survey and close the survey after a specified time (e.g., 2 weeks)
How can you create the most effective survey questions?
- Ensure your questions are clear. The way a question is phrased affects how (or if) people respond. Poorly worded, unclear, or overly long questions force people to interpret your questions. Different interpretations can skew answers, making your survey results unreliable.
- Use simple and clear language. Avoid the use of double negatives, jargon, acronyms, unnecessary words.
- Avoid leading questions. These are questions that favor one way of responding. Instead, focus on creating neutrally worded questions.
- For example: “Would you rather use the old version of this improved version of the website?”; “Do you find this feature frustrating to use?”
- Avoid double-barred questions. These questions ask about two conceptually separate issues but allow only one response.
- For example, “Please rate the extent to which you have been feeling anxious and depressed.” This item should probably be split into two separate items—one about anxiety and one about depression.
Which types of questions should you use?
There are two main types of questions – closed- and open-ended questions.
These questions provide several response options, giving people greater choice. Two types of closed-ended questions that we find most compelling are multiple choice and dichotomous questions.
Multiple Choice Questions:
- Use this type of question to ask people about their preferences using a scale (e.g. 1- strongly disagree to 7-strongly agree) on a given topic (e.g., I believe that self-driving cars are a danger to public safety).
- Or, use multiple choice to ask people to choose between different alternatives (e.g., which of the following brands do you prefer?).
- Considering including an ‘Other’ or ‘none of the above’ response if you don’t want your respondents to feel that they are forced to make a choice.
- Multiple choice questions can ask respondents to choose one answer or can allow for multiple responses (e.g. Which of the following social media sites do you use? (select all that apply)).
These types of questions offer two answer choices only. Because this type of question provides less information than multiple choice questions, we recommend using them sparingly.
These give participants ultimate control over their answers.
- Open-ended questions can provide spontaneous and vibrant information; however, we recommend focusing, primarily, on using closed-ended questions.
- If you want to invite participants to elaborate on their thoughts and opinions, consider conducting select, follow-up interviews after the survey.
- These types of questions yield qualitative data that will require a different kind of analysis.
Many things can affect how people respond to survey questions. Be sure to avoid response biases that can affect the quality of response you obtain.
How can you ensure a successful survey?
- Surveys are most useful when you have access to a large sample. If you have access to fewer people, consider doing interviews instead.
- Choose participants that can be unbiased. If your sample is biased, you will likely get biased results. For example, if your sample were not representative of your target audience, your results might favor a particular point of view and the results would need to be interpreted with caution.
- Keep it brief. People are less likely to complete long surveys. They’ll get bored and lose interest. You also run the risk of having people rush through their responses.
- Memory is malleable. Avoid asking people about behaviors, thoughts, and experiences that occurred long ago as those answers are less reliable.
- Use an online survey tool like Qualtrics or Survey Monkey. These services are relatively easy to use and make the process of creating, delivering, and collecting data from surveys much faster.
- When possible, randomize questions and responses (tools like Survey Monkey and Qualtrics have these features).
- We highly recommend piloting your survey before you send it out. This will allow you to check if there are any problems with your survey (e.g., confusing questions, unclear answer choices, irrelevant questions).
- Consider whether ensuring anonymity is essential for your participants. For example, people may be reluctant to give honest responses if they think their boss will see their responses.
- Provide motivation. If people are not intrinsically motivated to complete the survey, provide incentives, like gift cards.
How should you present your survey findings?
- The benefit of a survey is that the results will be quantitative, providing numerical information that reflects your participants’ thoughts and attitudes. Using statistics will help support the concrete findings.
- For example, you will be able to make statements about frequency (e.g., how often something occurs), agreement/disagreement (e.g., 60% of employees agree that the lack of maternity leave is a reason they would consider leaving the company), etc.
- If you have a low number of participants, refrain from using percentages, which can be misleading. Discuss your findings in the context of raw numbers (e.g., 8 out of the 10 employees surveyed agreed that the lack of maternity leave is a reason they would consider leaving the company). Illustrating survey results using bar graphs can make it easier to explain.
- Keep your graphs simple and uncluttered as they will be easier for your audience to understand.
- Don’t forget to include legends and numeric values on the axis. This will help ensure clarity and ease of interpretation.
- Include a short section that summarizes the insights and novel patterns gathered. Be clear about how the data connects to outcomes and make it easy for your survey stakeholders to understand how the results support your goals or affect your survey hypothesis.
Surveys can be a valuable tool to complement other HCD techniques and to gather intelligence from a larger group. Creating effective surveys, however, takes a thoughtful approach stemming from your goals, your audience, and the type of input you require.
If you are embarking on a business or technology project and want to ensure, you build your project upon stakeholder and user intelligence, contact us.
Our human-centered design experts can ensure you choose the best technique for your needs and gather high-value user-centered feedback.