Successful data collection starts with a good survey. How you design your questions will have a tremendous impact on the way data is collected, processed, and analyzed. We’d like to share some of our best advice on building a proper field data collection survey.
Effective survey design is not just about what is collected, but how.
We find that the best approach is to work backward. Starting with an eye on your goal keeps you focused on results, not process. Once you know what “success” looks like, you can determine what information you need to collect, and in what order.
Questions to ask to build a survey
1. What am I trying to find out?
Identifying the information you need is the first step to ensuring your data asks the right questions. To create a successful survey, start with the end in mind so you can visualize how you want to present your information. This will help you determine what data you need to collect.
Remember: The point of field data collection is to gain real-world knowledge, not to prove what you already believe to be true. You want your data to be factual information from which you can derive answers.
2. What are the appropriate metrics to use?
Knowing what metrics you will use will ensure objectivity by helping you avoid creating bias in after-the-fact analysis. Data types fall into four categories:
- Categorial: Unordered labels, such as brand names or colors (also known as “nominal”)
- Ordinal: Scaled answers, such as from “strongly disagree” to “strongly agree” or “never” to “often”
- Interval: Ranges of numbers (e.g., 1-100)
- Ratio: Numbers (e.g., 99)
3. What data should I collect from the field?
In field data collection, people often make two major mistakes: either collecting as much information as they can, or as little information as they can. Both approaches come with advantages, but can also cause problems.
A short and sweet survey results in rapid data collection, but may not capture all the needed information. If you collect too little data, your survey will not generate meaningful answers.
A very detailed survey will create a lot of content, but takes longer to conduct. Trying to collect too much data can result in people ending a survey midway through and potentially missing the most important questions.
You want to strike the right balance of depth and efficiency. We recommend isolating the fewest number of key data for your questions, and add a comments section so collectors can leave notes or other insights from the field.
4. How should I format my survey?
The layout of your survey has a direct effect on the time and effort needed in the field. There are four main ways to improve the efficiency of your data collection:
Chunking: Organize your survey by grouping similar questions together into sections, which allows collectors to follow the process. If you have 5 questions about history and 5 questions about politics, it makes sense to separate them into two groups.
Labeling: Once you’ve grouped similar questions together, label them to help the collector understand the focus of the questions they are asking. (It also helps give them an idea of how far along they are in the survey.) This way, a surveyor can parse a long set of questions easily, finding a specific piece of information without having to read through them all, when they are well labeled.
Skip logic: Skip logic can determine the flow of a survey by keeping only the relevant questions in front of the collector. If an answer makes some of the questions that follow irrelevant, skip logic allows the surveyor to bypass those questions entirely rather than having to scroll past them. A well-designed survey should only ever present relevant questions to the collector.
Calculations fields: Using calculation fields allows you to conduct calculations (simple and complex) instantly by using other questions as inputs. A calculation field can be used to take initial measurements and generate an answer instantly, saving the collector from having to tally up responses by hand or, say, use a calculator to determine the total area of a space. This is especially useful for engineers or construction workers who take measurements to create an estimate on site.
Now that you you know what data you’re collecting, it’s time to write your survey. Here are some best practices for framing your questions.
Keep questions short.
The more direct the question, the less risk you run of a surveyor misreading it and answering incorrectly. If you break up your survey into several smaller questions rather than a few long ones, you’ll end up with richer, more useful data. Short questions also keep the process moving along efficiently.
Make questions relevant and specific.
Staying focused only on critical information prevents unnecessary delays and feelings and opinions from being interjected into your data. You want your information collected to be comprehensive, but optional data can detract from your primary purpose.
Use simple language.
Avoid flowery language and euphemisms and be clear what you’re asking for. Keep your audience in mind: Avoid technical jargon unless you are certain it will be understood by the person completing the survey. The best surveys can often be conducted with minimal training.
Ask one thing at a time
Avoid asking open-ended questions or questions with multiple answers. It’s best to break up a survey into more small questions rather than fewer large ones; this will provide the added benefit of being able to slice up your data in many different ways once your project is complete.
Avoid biased language and leading questions.
Make sure your questions are phrased in a way that is considerate, inclusive, and respectful. Don’t use unnecessary adjectives or adverbs that may add bias to a response. Be objective and steer clear of questions that may lead the subject to answer in a particular way.
Phrase questions in a positive form.
Asking questions in the negative can be confusing and may lead to biased responses. Don’t ask what is not, ask what is.
List questions in a logical order.
Your questions should follow the order in which the collector would typically conduct the survey. For example, if you’re doing a construction walk-through inspection, structure the questions in the order that the inspector would move through the site.
Offer a balanced set of responses.
Avoid errors of omission by giving a range of answers that are inclusive and thorough. Include multiple options that cover every possible response to the question. If all else fails, include an “other” option.
Add a comments section for additional data.
Sometimes, your collector will find important information in the field that should be relayed to the analysts. An optional comments or notes section at the end of the survey provides a catch-all for anything your survey model might not have accommodated for.
Now you’re ready to build your first survey!
Check out our Guide to Field Data Collection to learn more tips, tricks, and best practices for collecting field data!