1. What is a survey?

If you’ve ever responded to the U.S. Census, then you’ve taken a survey, which is simply a questionnaire that asks respondents to answer a set of questions. Surveys are a common way of collecting data because they efficiently reach a large number of people, are anonymous, and tend to be less expensive and time-intensive than other data collection methods. The purpose of surveys is to collect primarily quantitative data. Surveys can be administered online, by phone, by text, or in print. 

2. Should I use a survey to collect data? 

In our last post we talked about how to decide which data collection method fits your evaluation. The first step is figuring out your evaluation question and determining if a survey can answer it. Surveys might be the right option if you want to collect information from a large number of people about their needs, opinions, or behaviors. For instance, they can help you determine what patrons learned from a program, the different ways people use resources at your library, or even what services non-users might be interested in, among other things. 

Surveys might not be the right method if: 

  You’re primarily trying to answer questions of why or how, as these work best as open-ended questions and are better suited for interviews or focus groups. Surveys can contain open-ended questions, but they are typically supplemental to the closed questions that make up the majority of the survey.

  Participant self-reported behavior is likely to be inaccurate. For instance, surveying children on how engaging a program was might not be the best approach.

In addition to these criteria, you should also consider time and costs associated with a survey and whether these line up with the resources you have available. A more thorough breakdown of the costs associated with a survey can be found here

3. How many of these question types have you used? (Mark all that apply)

Although survey questions can be written in a multitude of ways, ultimately every question is either closed, open-ended, or a combination of both. Open-ended questions ask the survey respondent to provide an answer in their own words, like in the example below.

Why did you decide to read this blog post? 

Open-ended questions allow the evaluator to collect robust data by not limiting the respondent to a list of possible answers. For instance, maybe you’re reading this blog right now because your cat walked across your keyboard and accidentally clicked on the link. The survey is unlikely to include that answer option on a closed question, but an open-ended question can capture that sort of qualitative data. 

Although there are many pros to using open-ended questions, there are also some downsides. Most open-ended questions take a long time and a skilled evaluator to analyze the qualitative data they produce. That’s why closed questions are more commonly used on surveys.

Unlike open-ended questions, closed questions provide a set of answer choices and produce quantitative data. Let’s explore some different types of closed questions.

Multiple choice questions allow respondents to select one or more options from a set of answers that you define. A common drawback of multiple choice questions is that they limit answers to a predetermined list like below, which may not reflect everyone’s responses. Often the problem is solved by adding an “other” option where a respondent can write in their answer if it isn’t part of the list. 

How do you feel today?

  Happy

  Sad

  Other, please specify: ___________

Adding an “other” option makes part of this question open-ended. When you analyze the data for this question, pay close attention to the percentage of respondents who chose “other.” If it’s a large portion (usually more than 10 percent), you will need to do some qualitative analysis of these answers.

Likert scale questions give respondents a range of options (usually five or seven choices). They’re often used to gauge someone’s feelings or opinions and can be written as statements instead of questions (see below). Writing a likert scale can be tricky because you need to make sure your response options are balanced. We’ll talk about that more in depth in our next post. Here’s an example of a likert scale question.

I am learning something from this post on surveys.

  Strongly agree

  Agree

  Neither agree nor disagree

  Disagree

  Strongly disagree 

Demographic questions ask respondents about characteristics that are descriptive, such as age, gender, race, income level etc. Demographic questions allow you to gain a deeper insight into your data. For instance, I could use a question that asks a respondent’s age to analyze whether younger respondents were more likely to say they “disagree” or “strongly disagree” on the question above. 

These are the most common question types you’ll find on a survey, but for a deeper dive on different question formats, such as matrix, dropdown, and ranking, check out this article from SurveyMonkey. 

4. Stay tuned for surveys pt. 2?       Yes       Definitely     I wouldn’t miss it for the world

We’ve all probably taken a survey, but there’s a lot that goes into making them balanced, understandable, and unbiased. In our next post we’ll cover why the question above should never be on a survey and other common mistakes people make when writing survey questions.