Learning analytics tools aim to boost student retention, outcomes


Learning analytics software helps professors evaluate their course structures and ensure better student outcomes.

As policy makers and campus leaders focus on boosting college completion rates, learning analytics is a field that has exploded in importance. A number of programs now exist to help instructors and campus leaders track student progress more closely, leading to better student outcomes.

Some of these programs are standalone software packages, while others are modules or features included in leading student information or learning management systems. Here’s an overview of some of the many products that can improve communication between students and professors, allowing everyone to gain a clearer perspective on students’ needs.

Canvas Analytics

Instructure’s Canvas Analytics, which launched this past June, presents dashboards for students and professors so they can make better data-driven decisions and achieve optimal learning outcomes.

Students can compare their individual progress with their peers’ performances, while professors can monitor the class as a whole and quickly detect at-risk students. Canvas Analytics also allows administrators to compare multiple courses and assess different departments and programs.

(Next page: How Canvas Analytics lets administrators use data)

“It has always been focused specifically on the learning side, making sure that we can help teachers and administrators, and students see at a glance how students are doing,” said Brian Whitmer, co-founder and chief product officer at Canvas.

[field name= iframe6]

 

Product developers used the same process to create Canvas Analytics that they used to create the larger open-source Canvas LMS. Designed with input from a dozen institutions, Canvas Analytics aims to address the major challenges experienced by students and faculty.

“I’m really proud of the fact that our analytics suite is as open as the rest of Canvas,” said Whitmer. “We think of Canvas as a learning platform. It’s the hub that brings the educational experiences together.”

Whitmer said the openness of Canvas Analytics allows administrators to harvest and control their own data. When asked about the outlook for learning analytics, he said he thinks well-organized data will lead to better decision making in higher education.

“I think there’s a great opportunity in the next couple of years to see this data-driven decision making come to the forefront,” he said. “Some of our heavier users of analytics have been students [who] are looking at their own data.”

 

Ellucian’s Course Signals

Course Signals, a product originally developed by Purdue University, uses online signals to alert students of their progress in a given course. Purdue officials piloted the software for three years before joining efforts with the former SunGard Higher Education to market the software to other schools.

Course Signals sends students graphical representations of their progress, depicted as red, yellow, or green circles resembling traffic signals. A green circle indicates good progress, a yellow circle signifies there is room for improvement, and a red circle shows that a student needs immediate attention. The colored graphics are sent to students via eMail and are based on the recorded grade data entered by professors.

Though SunGard Higher Education merged with Datatel in 2011, the newly formed company that resulted, called Ellucian, has continued to promote Course Signals. Ellucian Course Signals offers students the same traffic light alert system, indicates when professors need to intervene with struggling students, and suggests ways that students can improve their grades.

A major benefit of Course Signals is its ability to identify at-risk students as early as the second week of the semester, its makers say. Both students and professors benefit from the continuous feedback.

“The predictive model in Course Signals gives students a good indication very early in the course of how they are performing and whether they are starting to lag behind others in the class,” said John Campbell, Purdue’s associate vice president of the Rosen Center for Advanced Computing. “This very early alert to the students is extremely valuable, even in populations where you might not think it is necessary. Signals is helping Purdue improve [student] retention rates by identifying underperforming students early on and providing them with course-specific advice on how to change their trajectory.”

[field name= iframe2]

 

As student retention rates continue to pose a problem for many two- and four-year colleges, users of Course Signals say the product can help.

“We found in our research that this can improve student performance an average of one letter grade for many students,” said Gerry McCartney, Purdue’s chief information officer and vice president for information technology. “Course Signals is an important step forward for higher education that can be implemented successfully at many universities and community colleges across the nation to improve student retention and success.”

Jenzabar Retention

Through its custom-built, patent-pending predictive modeling techniques, Jenzabar Retention, (formerly known as FinishLine) helps to pinpoint at-risk students quickly and offer them the necessary resources to attain success.

As with Course Signals, the product includes early alerts based on a predictive model. However, this model is customizable for each institution, based on its own experiences.

With Jenzabar Retention, when a risk factor appears on a student’s profile, relevant administrators are notified immediately. Alerts are prompted not only through poor attendance or failing grades, but also by numerous other academic, social, and economic factors from the past five years.

“It’s not an absolute indicator, but it’s a predisposition,” said Burt Rubenstein, vice president of student success solutions at Jenzabar. If the data suggest that students are at high risk, professors can “proactively give them a better advisor or put them into a learning community.”

Every institution has a customized array of indicators that suggest low-, medium-, and high-risk students. Jenzabar works with institutions to understand how administrators define risk. Called “candidate factors,” these behaviors could include anything from poor attendance to skipping payments on bills. Jenzabar takes those candidate factors and runs them against the institution’s historical data to see if the two correlate; if they do, those factors become reputable indicators of low, medium, or high risk.

“We intentionally don’t have a generic model,” said Rubenstein, because every school has its own unique challenges.

[field name= iframe3]

 

Currently, Jenzabar Retention does not alert students of their personalized risk factor, though the company is researching whether such a feature would benefit users.

When asked if campus leaders are feeling pressure to invest in learning analytics in order to improve student outcomes, Rubenstein didn’t hesitate.

“I don’t think they have a choice,” he said. “There are both private and government-funded grants that are related to improved outcomes.” He noted that many states are moving to a funding model that bases support for public institutions on the number of students they graduate instead of how many are enrolled (see “Higher education funding models changing“).

Blackboard Analytics for Learn

Providers of learning management systems (LMS) software recently have begun to develop and integrate learning analytics tools into their product designs as well.

Launched in 2011, Blackboard Analytics for Learn uses learning analytics to track multiple metrics, such as gradebook scores and involvement in online discussions, to help students monitor their performance in a given course, and compare it to how their peers are performing.

“The data come from multiple sources, not just the LMS,” said Jim Hermens, general manager of Blackboard Analytics. “Right away, we’re co-mingling traditional student management data and LMS data and exposing that to the student.”

Professors can analyze and monitor student activity patterns and trends over time, identify at-risk students early on, and keep performance data current for students quite easily. The product offers several user-friendly tools for self-service, including pre-built dynamic report formats.

[field name= iframe4]

 

Hermens said Blackboard Analytics allows administrators to examine not only students’ performance, but also their progression to determine whether there might be problems in the design, structure, or delivery of a course, among other things.

He called learning analytics a necessary tool for colleges and universities, explaining: “Having the data is a critical component, [but] having a program [to help] you use [the data] for constant improvement” is important as well.

Though some colleges use Blackboard Analytics for Learn to improve course retention rates and monitor student progress, Hermens said the product also is used in many institutions to assess course structure. Some administrators would like to attract students to less popular majors, and Hermens believes the presentation and breadth of data collected by Blackboard Analytics for Learn can help them evaluate their courses with an eye toward this goal.

Desire2Learn Analytics

Desire2Learn Analytics aims to improve student outcomes through four primary objectives: designing curricula-aligned assessments, analyzing results and reporting on evidence, making informed improvements, and defining outcome standards.

The product compiles data using quizzes, rubrics, competencies, and learning outcomes to compare students’ results with learning objectives so that professors can better assess whether their assignments are effective. Students also are able to compare their performance to that of their peers.

Students “can understand how they’re doing relative to others in the class,” said Alfred Essa, director of innovation and analytic strategy at Desire2Learn. “It gives them more insight into how they’re performing, and it [could] motivate them to do better.”

Professors and students can view data in multiple formats, such as charts, heat maps, or decision trees, in order to observe trends. At-risk learners are identified quickly, and by the second week of classes, the product can predict what a student’s grade will be by the end of the course, Essa claims.

“Think of it as an early-warning radar system,” said John Baker, president and chief executive officer of Desire2Learn. “We want to have the capability to identify students at risk, not only how they’re doing now, but be able to forecast with incredible precision about where they’re going.”

[field name= iframe5]

 

Though projection is important, Desire2Learn believes personalization is the key to better learning outcomes. “The promise of learning analytics is to personalize learning,” said Essa. “Less data, more insight.”

Essa and Baker also highlighted the importance of comprehending the social dimension of learning. Desire2Learn recently used predictive modeling to create a sociogram, or a dynamic network graph of interactions among students, to determine which types of students were “isolated,” or more unwilling to work in groups in specific courses.

Essa and Baker believe this type of research can help not only when intervening with at-risk students, but also in advancing enrollment processes.

Sign up for our newsletter

Newsletter: Innovations in K12 Education
By submitting your information, you agree to our Terms & Conditions and Privacy Policy.

Oops! We could not locate your form.

Sign up for our newsletter

Newsletter: Innovations in K12 Education
By submitting your information, you agree to our Terms & Conditions and Privacy Policy.