One of the most important questions you might come across when building your survey is whether or not to include any open-ended questions, a type of question that will give you free-form text data. You think—is the value of open-ended questions worth the time and effort they take to analyze?
In this article, we explore the importance of including these types of questions in questionnaires, best practices for analyzing free-form text data from surveys, and how aytm can help.
What is free-form text data?
Free-form text data can come from any question where you allow respondents to provide answers or comments in their own words. This could be an open-ended question, such as “Why do you get popcorn at the movie theater?” or it could be an answer option in a closed-ended question that allows a respondent to elaborate on their answer choice. An example could be an “other” answer option that requires respondents to provide a comment.
In general, you’ll receive some insights answers from respondents to these questions that are important to capture in the survey results you share with stakeholders, but oftentimes, free-form text data is forgotten or pushed to the side in favor of saving time.
Benefits of free-form text data
Even though text data isn’t the easiest to make sense of, it is a key factor in determining how actionable your survey insights can be. For example, if you asked “What do you put in your coffee?” in a closed-ended question to determine if coffee drinkers could be an addressable market for your new dairy alternative product, you wouldn’t understand why they use what they use. By following up with the open-ended question “why do you put that in your coffee?” you can gain valuable information as to why coffee drinkers use certain products in their coffee.
Free-form text data can also be used to support and enrich quantitative data. Imagine you asked respondents why they only exercise a certain amount of time in a close-ended question. If you follow that up with what activities they believe prevent them from exercising more, you can add additional context to the results from your close-ended question.
Still, it may be difficult to weigh the benefits of including questions leading to free-form text data in your surveys given the time they take to analyze. Like all types of analyses, it is important to plan your strategy ahead of time, keeping in mind any time constraints and how much detail you’ll need to uncover.
How to analyze free-form text data
There are three common methods researchers use to analyze free-form text data, depending on their needs and the resources they have available:
- Word Clouds
Generating a word cloud is a quick and easy way to visualize free-form text data. All text from respondents is clustered and presented in different sizes. The larger the word appears in the visualization, the more times it was mentioned. In general, the larger words will be more important to developing and supporting your survey insights.
In some cases, you might even want to include a word cloud in your report. Many free word cloud generators available offer customization, like excluding certain words or changing colors and fonts so that you can fine tune your visualization.
Coding involves looking through all text data by hand and grouping responses by common themes, depending on your analysis goals. When using this method, it’s important to create criteria for categorization and rules. Two common categorization schemes are described below.
In sentiment analysis, responses are grouped by their underlying attitude. Oftentimes, researchers use the following categories: positive, negative, and neutral.
In topic analysis, responses are grouped by assigning categories or tags based on themes. For this method, it is important to at least skim through responses in determining the best categories to include. An example of this for a new consumer packaged good product could be the following categories: packaging, price, label, colors. name.
- Text Analytics
Text analytics is an evolving method and although it’s not as robust as manually combing through data, it is a valuable tool that is fast and less expensive than coding. This method involves algorithms that automatically categorize free-form text inputs. Sentiment analysis and topic analysis can be done hands-free using text analytics, but results will not be perfect.
A good example that illustrates why humans are still better than machines at categorizing free-form text is the use of sarcasm. Imagine a respondent writes “Well, what a surprise” about a new product concept they were presented with. A computer may have difficulty categorizing this comment as “negative” since the word “surprise” has a positive connotation.
How aytm can help you analyze free-form text data
aytm has several tools that can help you analyze text more efficiently and effectively. Word clouds will be automatically generated for all questions with free-form text in your survey. You can customize word clouds by changing fonts, font sizes, colors, excluding certain words, and more.
Besides word clouds, aytm offers robust coding tools to evaluate all comments in your surveys. Sentiments are automatically assigned to each comment, and you can adjust all assigned sentiments as needed. In addition to automatic sentiments, aytm offers the ability to create custom tags so that you can easily categorize comments as you read through them on the platform.