We’re excited to announce the addition of sentiment analysis as part of Watershed’s latest feature releases—which means you can now analyze text based on how positive, negative, or neutral it is as well as categorize overall sentiment. This is a powerful new tool in your analytics toolkit, but what does it mean in practice? Let’s explore four ways you can use sentiment analysis to enhance your learning analytics and evaluation.
So, What Is Sentiment Analysis and Why Is It Useful?
Sentiment analysis means analyzing pieces of text to determine whether they express a positive, negative, or neutral sentiment. This process uses AI machine learning to understand text sentiment based on word choice and sentence structure, as shown in the following table.
| Positive | Neutral | Negative |
|---|---|---|
| I enjoyed attending the course. | I attended the course. | Unfortunately, I attended the course. |
| I like raspberry ice cream | Ice cream is made from milk. | Ice cream is too cold. |
| I have been successful at applying the learning in my work. | I work as a technician. | This learning was not relevant to my job. |
Sentiment analysis is useful because it aggregates qualitative data, allowing you to efficiently get an understanding of sentiment without having to review individual pieces of text—which could number in the tens or hundreds of thousands depending on how many learners are in your organization. And that’s great if you have an LMS that caters to thousands of learners!
And specific to the world of learning and development, sentiment analysis can help you:
- improve your learner feedback surveys,
- review forum responses and reflections,
- identify the most positive or most negative feedback, and
- highlight bland or unhelpful course descriptions.
Improve Your Learner Feedback and Enhance Smile Sheets.
Whether you’re asking learners for their views on a training course or about how they have applied learning in their work, qualitative textual data can provide additional insights beyond what you can collect through multiple-choice questions. Textual analysis—such as sentiment analysis—provides a way to analyze data efficiently and at scale.
Using freeform text fields, or textual data, alongside multiple-choice questions has the following benefits:
- Free text fields enable you to capture more-detailed reflections than are possible with multiple-choice or likert-style questions.
- Learners are forced to slow down and think about their answers, unlike multiple-choice questions where there’s a risk that learners will just quickly click any response to get through the survey. This means free text responses have the potential to be more accurate.
- Using free text questions alongside multiple-choice questions means you can use the textual data to evaluate the accuracy of multiple-choice responses. For example, if learners consistently give courses high ratings, but in their text responses they complain about the course or say they have not applied the learning, you know the course ratings are not a useful metric.
Automatically Assess the Sentiment of Learner Forums and Reflections.
Everyone says they want feedback about their learning programs, but ranking on a traditional smile sheet is devoid of detail and an easy way for learners to slip into a “checkbox” mentality.
Discussion forums and reflections are popular learning tools within academic learning and can be extremely helpful in a workplace learning context. When people write about what they have learned, they’re able to better process and remember their learning. Furthermore, the right discussion or reflection questions can help learners consider how they can apply that learning in their work.
Both forums and reflections generate a lot of textual data, so it isn’t normally feasible to manually read and assess it all. But sentiment analysis means you can automate part of this process by analyzing whether forum posts and learner reflections are generally positive, negative, or neutral.
This information can be helpful for mentors and managers when talking with learners about their experiences and for the learners themselves to reflect on their reflections and comments. For example, if reflections on a particular topic generally have a negative sentiment, it could indicate a topic where the learner needs more support.
Rank Comments and Feedback in Real-Time.
Sentiment analysis not only categorizes text as positive, negative, or neutral, but it also produces a numerical value based on how positive or negative a piece of text is. As a result, you can rank and sort comments or feedback based on their sentiment from most positive to most negative (or vice versa).
For instance, you might use this approach as part of the Brinkerhoff Success Case Method (SCM), which involves identifying a learning program’s most and least successful cases. While multiple-choice response data enables you to determine all cases where the top and bottom ratings were selected, this data does not allow you to differentiate respondents more granularly.
By contrast, because text responses are more likely to be unique, you can identify one or more that represent the most positive and the most negative responses. Following the Brinkerhoff SCM, you can then investigate each of these cases to better understand why certain responses are especially positive or negative with the textual feedback providing a good first step for that investigation.
Identify Downbeat Course Catalog Descriptions.
One interesting theme at this year’s Learning Technologies conference focused on using marketing practices to promote courses and increase engagement. For example, do the catalog course descriptions help encourage learners engage with a course? Do they include positive words and phrases to paint the course in the best possible light?
In this instance, a potential sentiment analysis use case might be to analyze the sentiment of your course descriptions to identify the ones with the least positive (or most negative) descriptions. You can then update them with something more upbeat and track usage to see if those changes help drive engagement.
Get Started with Sentiment Analysis in Watershed.
Understanding the sentiment behind learner feedback is essential for ensuring your training is effective, relevant, and impactful. That’s because sentiment analysis can help you improve your learner feedback surveys, review forum responses and reflections, identify the most positive or most negative feedback, and highlight bad-vibe course descriptions. To find out more and get started, check out our product update on sentiment analysis.
About the author
As a co-author of xAPI, Andrew has been instrumental in revolutionizing the way we approach data-driven learning design. With his extensive background in instructional design and development, he’s an expert in crafting engaging learning experiences and a master at building robust learning platforms in both corporate and academic environments. Andrew’s journey began with a simple belief: learning should be meaningful, measurable, and, most importantly, enjoyable. This belief has led him to work with some of the industry’s most innovative organizations and thought leaders, helping them unlock the true potential of their learning strategies. Andrew has also shared his insights at conferences and workshops across the globe, empowering others to harness the power of data in their own learning initiatives.
Subscribe to our blog