Types of Learning Analytics Tools
The Canvas Learning Analytics Dashboard
The functions offered by LA tools typically include:
- Dashboards and reports. Some LA tools offer dashboards that provide instructors and advisors with at-a-glance information on students’ behaviors and success to date in a course, potentially enabling instructors and/or advisors to better guide individual students and better adjust instruction to an entire class or to subsets of students. Many LA tools allow for customization of reports, allowing faculty and advisors to focus on specific students or groups of students or on particular behaviors or performance metrics. Some LA tools also offer dashboards that students can use to view their progress in a course, examine their own behaviors, set goals, and, in some cases, obtain advice about study behaviors that might increase their likelihood of completing the course successfully.
- Messaging. A number of LA tools can provide students with instructor-, advisor-, or auto-generated-communication to prompt stronger study behaviors. They can also provide instructors and advisors with information about student behaviors. Automatic messages sent to students are often referred to as nudges, while those which are also sent to instructors and advisors are typically referred to as alerts. Depending on the capabilities of the system that generates them, nudges can be customized in advance or modified on the fly by instructors. When certain conditions are met, such as a set period of time following the last login to the system, failure to complete a quiz by a deadline, or failure to submit homework by a due date, nudges are sent to students. Depending on the settings in the system, alerts might also be sent to instructors and/or advisors. Messages may include updates on performance, suggestions about effective learning behaviors, information on associations between students’ uses of such behaviors and academic achievement, or other content. In some cases, alerts require a student to meet with the instructor or advisor in order to “clear” the alert.
A key issue associated with messaging involves how messages are constructed. On one level, it is a rhetorical issue, where the focus is on how to design messages that influence students in desirable ways. For example, designers of nudges and alerts seek to understand which rhetorical strategies are most likely to lead students to engage in behaviors that enhance their success in a course. On another level, it is a teaching and learning issue, where the focus is on which behaviors to address and which to encourage. For example, designers of nudges and alerts would benefit from understanding which behaviors are likely to lead to improved student learning and at what points it would be optimal to send messages. In the long run, the two levels intersect, in that the more knowledge is gained about which behaviors support learning, the more this information can be leveraged to increase the rhetorical effectiveness of nudges and alerts.
- Predictions of Student Success. Some LA tools offer dashboards and reports that provide instructors and advisors with predictions of student success at the level of individual students or entire classes. For ease of interpretation, these predictions are often communicated as scores on a specified scale. These scores are based on calculated probabilities of student success (e.g., earning a grade of A, B, C) or failure (e.g., earning a grade of D, F, W) in a course or program. Prior to the start of a course, such scores are based solely on academic history and demographic information. As the course unfolds, these scores are updated to reflect student behaviors and performance in the course. Typically referred to as “predictive analytics” or “predictive learning analytics,” these scores are derived through algorithms that operate on available data. On an individual level, these scores are subject to statistical variation, and are thus certain not to be completely accurate. However, improved accuracy results when these predictions are used in combination with what instructors learn from students through personal interaction and observation. Viewed in the aggregate, for example over all students in a course section, predictive analytics scores can provide an indication of the overall progress of students enrolled in a course.
In addition, a related set of tools—such as EAB’s Navigate—are being used to help institutions identify courses in which students struggle and, perhaps more importantly, to reveal course combinations within an academic term or course sequences across academic terms that appear to be correlated with lack of success.