You might be able to get more out of your open-ended questions in an online survey than you might think. With the power of text analytics and automation, many survey tools now have the ability to calculate the sentiment of a respondent’s answer.
A sentiment analysis takes a look at the text of an open-ended response and interprets whether it is positive, neutral, or negative in nature. Another popular application of this technique is the analysis of social media content. In the context of an online survey, this insight can be used in a number of ways for the remainder of the survey or during reporting.
Sentiment Analysis in a survey tool may make text analytics more accessible than ever before to clients. Get insightful, yet simple answers from your open-ended questions.
How Does Sentiment Analysis Work in a Survey?
Drive Research uses an advanced survey tool that offers the ability to ask almost any question you can think of. Among the text box and essay style of questions, the tool has recently developed a sentiment analysis feature.
Imagine a satisfaction question concerning a company’s product. This question may be followed-up with an open-ended probe into why the respondent selected a particular rating. A typical use of this data is a qualitative review for themes or coding the responses into buckets if the sample size is large enough.
With the survey tool, an alternative option is to run a sentiment analysis on the responses. The words and phrases within a response are tested against a dictionary that ultimately provides a quantifiable score for sentiment.
After a score is calculated, the remaining survey path can include logic based on a respondent’s answer to that question being positive, neutral, or negative. For more precision, we can set a threshold for the scores that filters respondents down a certain path.
How is a Sentiment Score Calculated?
So, what’s the magic behind assigning a score to an open-ended response? The survey tool leverages the Valance Aware Dictionary and sEntiment Reasoner (VADER) Sentiment Analysis Package, which is part of the Natural Language Toolkit (NLTK).
Text is fed through a dictionary that has predetermined sentiment values for words, acronyms, slang terms, and even emoticons. Everything has a value from negative 4 to positive 4, with 0 being neutral. The values from every piece of the response are then added together, and the score is normalized to fall between -1 and +1. If a respondent’s answer is multiple sentences, the sentences' scores are averaged to determine the overall sentiment score.
However, everyone knows that the English language is complicated. The number of exclamation points or question marks used plays a role in the score. Capital letters are recognized as a sign of emphasizing a word and its value.
Booster words, such as “very” or “hardly” may also increase or decrease the value of a word. Negation is factored in, as well, to reverse the sentiment of the following word or words.
What are the Benefits?
Once you have the score for a respondent’s answer, there are several ways the new data can provide value. Since the calculation is performed in real-time, the online survey can immediately use the scores for the remaining programmed logic.
One useful application here is interpreting the sentiment of an open-ended question about a service experience. The sentiment analysis will determine if a response is generally positive, which can trigger a follow-up question asking if the respondent would be willing to provide a testimonial.
On the other side, respondents who received negative sentiment scores can be flagged to understand what made their experience less than satisfactory. In either case, sentiment analysis may be a useful component of a voice of customer (VoC) survey to understand the feedback and thoughts of customers.
Sentiment analysis also provides data to generate standalone cross tabulations or charts that make an open-ended question easier to digest without the effort of coding. Through the survey tool, this data is automatically calculated and ready for use.
Contact Our Online Survey Company
Have a question about including Sentiment Analysis in your next online survey? Call, email, text, or message our team at Drive Research to learn how you and your business can best utilize it. Drive Research drafts, manages, and reports on surveys across a variety of clients and industries every day.
Message us on our website
Email us at [email protected]
Call us at 888-725-DATA
Text us at 315-303-2040