The best approach to building a survey is to work backwards. Identify what you want to learn and how you want to measure the results before you begin designing. Much of effective survey design is about not just what to collect, but also how it's collected. Thinking about the end goal keeps you focused on results first, and process second. Starting at the end lets you create a process that matches your desired end state. Once you've identified what "success" looks like, you can analyze the details of what data you need, in what order, and more.
- What do I want to know?
- What metrics should I use?
- What data should I collect?
- How should I lay out my survey?
- Handling changes
For your data to answer all the right questions, the first step is to identify the information you need to have in the end for success. By starting at the end, you can visualize how you want to present your information. Create a hypothesis that is testable and measurable, and from that determine what data you need to collect.
A project trying to assess "fire risk" may want to map out quantitative information on hydrants and fire prevention equipment within an area. A political poll trying to map out supporters might want to gather qualitative data on opinions and party affiliations without biasing the responses.
Understanding what you want to learn is crucial to determine what you need to collect. It allows you to properly frame up your survey and collect data that is relevant for examining your hypothesis.
The next step is to identify metrics that will guide your decision making. Identifying these before conducting your survey will ensure objectivity in your survey by avoiding bias through after-the-fact analysis. The point in conducting a survey is to gain real-world knowledge, not to prove a specific point. Your data should be factual information that you can then use to derive answers. Identifying these key indicators will feed into the specific information to capture with your survey.
If your data supports your hypothesis, great! If it doesn't, it's never advisable or ethical to bend the data in a way to support your position. Also, avoid framing questions in a way that leads people to a desired conclusion, in order to avoid bias in your data.
Next you should break down your desired result into its individual components. Perhaps to answer your questions, your survey needs to capture 25 different pieces of information. You then need to organize them in a fashion that makes it efficient for a field collector to fill out the survey. People often make 2 major mistakes in this process: They either try to collect everything they can, or as little as they can, and both of these approaches can cause problems.
Collecting too little data can make your survey lack enough information for meaningful answers. Collecting too much data can lead to people terminating a survey midway through. Both have their advantages, of course. A short and sweet survey means rapid data capture, but light on the depth. An incredibly detailed survey maximizes the time in the field with a lot of content in the end, but takes longer to perform. Get a feel for the time it takes to complete the survey by conducting some mock collection. The goal should be to strike the right balance of data depth and efficiency.
Try to isolate the fewest number of key data needed for your project, and add additional fields or a comment section as a catchall for general use. It's a good idea to have general "comments" fields for collectors to leave notes or other insightful info from the field.
The way you format and layout your survey can help save time and reduce the work needed to collect data in the field. We'll focus on 4 major ways to improve the collection speed of your survey: Chunking, Labeling, Skip Logic, and Calculation Fields.
Chunking is the grouping of similar questions together, to help organize your survey and allow it to flow easier. Grouping related questions together into sections helps collectors follow the process. If you have 5 questions on history, and 5 questions on politics in your survey, grouping them into 2 clusters make logical sense. It simplifies the flow for your field collection teams.
Labeling is the method of adding a label to chunks of questions you've already grouped together. This helps the survey administrator understand the purpose behind the various of blocks of questions they are collecting, and gives them an idea as to how far along they are in the survey. When a survey is well labeled, a user can parse a long question set easily, and find a specific question without having to read each question and entry to find a piece of information later on. Learn how to add a label in your Fulcrum app by visiting the help page on Label Fields.
Skip Logic refers to the use of logical conditions to determine how the survey flows. Skip logic is key for keeping only the relevant questions in front of a collector. If a user answers a question that makes later questions irrelevant, skip logic allows you to bypass those questions entirely rather than wasting time having to scroll past them. In an ideal setup, a survey should only ever present relevant questions to the individual. Learn how to set this up in Fulcrum by visiting the help page on Conditional Logic.
Calculation Fields allow you to conduct simple or complex calculations instantly by using other questions as inputs. For example, rather than having to tally up responses by hand, or use a calculator to determine the volume or area of a space, a calculation field can be used to take your initial measurements and generate an answer instantly. This is useful for individuals in engineering and construction who have to take specific measurements to create an estimate, right on site. Learn how to add this feature to your Fulcrum surveys by visiting the help page on Calculation Fields.
After you initially deploy a survey to a collection team, there will usually be modifications to make after it's in the field. In the days of pen and paper surveys, this was naturally a hassle to deal with, requiring editing word documents or PDF templates and re-printing new batches of forms to distribute. As a result, organizations using a method like this are tempted to overload their survey with every possible thing (see the chapter on best practices). Because of the inefficiency of making changes to paper surveys, many positive survey structure changes would never be made.
With digital tools like Fulcrum, updating survey template structures is more realistic. Given that the survey on a collector's device is digital, syncing with the server allows the user to receive the latest version automatically, without any need for memos or meetings to bring a team up to speed.