Colorado universities send about 20,000 people a year to community colleges because they aren't ready for college-level math and English. But they're losing many of these students in the process.
"Developmental education has the potential to be a black hole for students," said Casey Sacks, project manager at the Colorado Community College System.
Nationally, nearly 1.9 million students started college in 2006, but six years later, only 54.1 percent of them had graduated, according to a National Student Clearinghouse study published last year.
Yet by 2020, President Barack Obama expects the United States to have the highest proportion of college graduates in the world. In order to reach this goal, community colleges, four-year colleges and universities would need to graduate 10 million more students than expected, according to the U.S. Education Department.
Data mining and analytics technology can help colleges fill the black hole and reach for Obama's college completion goal.
"We have at our fingertips the capabilities to have more students succeed than ever before by leveraging the technology tools we have at our disposal," said Mark Milliron, co-founder and chief learning officer of Civitas Learning who gave a keynote address at the Blended Learning Conference and Workshop on July 8 in Denver.
The problem is actually getting student learning data to the front lines where faculty can use it to test innovations, create interventions and predict actions such as likelihood of course completion and graduation. Too often, colleges focus so much on accountability analytics that they hamstring their ability to get data to faculty.
Instead, they need to have the right infrastructure to get the right data to the right people in the right way, Milliron said. If faculty, advisers and students have access to learning data, they can make more informed decisions.
Colleges turn to apps for student success
Practically speaking, a number of colleges are working on different ways to help students stay on track to finish their degree. Austin Community College made a Degree Map application available, and while it wasn't mandatory, 90 percent of advisers encouraged students to use it. The app shows students where they stand in their degree program, helps them plan their next level of courses and shows the monetary and time cost of switching degree programs. It also predicts when students choose toxic course combinations that they're unlikely to do well in.
At El Paso Community College, an app allows students to choose an academic goal, such as getting straight A's. Throughout the year, the app shows them whether they're on track to meet that goal.
University of Texas at Austin students created an app called hoot.me that crowdsources questions and answers in large lecture classes. Students in larger classes don't have the same level of individual help from faculty that their peers in smaller classes do, so this app empowers students to take charge of their own learning.
Five or six high-achieving students serve as "owls," which means they answer questions that other students ask them in the Facebook app. Before faculty members teach the next class, they can log into their learning management system that's connected to the Facebook app and look at the top three themes that students aren't getting. That allows them to address thorny topics in class right away.
Developmental education gets a makeover
Along with apps, colleges are redesigning and experimenting with developmental education so they can move more students through post-secondary education in a timely manner. After looking at course completion and graduation data for developmental education students, Colorado Community College cut down the number of developmental classes students need, which included the integration of college composition and reading courses. They also provided extra academic support for students who scored the lowest on placement tests.
"It's really about putting students in the environment that they need to get the education support that they're looking for," Sacks explained in a session at the Blended Learning Conference and Workshop.
The University of Hawaii also went through a redesign of developmental education after seeing average passing rates of 44 percent. If students took one or two online courses every year, they made progress toward graduation, but developmental math online did not work, said Hae Okimoto, director of academic technologies for the University of Hawaii System.
So the university moved to a blended format that allowed students to work at their own level and pace. Instead of requiring students to sit through an entire course's material, they only had to go through the topics that they struggled with. By using ALEKS software, they could take individualized assessments and learn through various modules.
Students were together in class at the same time on Monday, could go to the lab with a faculty member at any time on Wednesday and could do their work from anywhere on Friday if they were doing well. This structured face-to-face time coupled with online learning modules and social interactions worked out better than lectures for these students.
In addition to implementing blended courses, the university cut the number of developmental math courses from four to three and asked college algebra faculty what they wanted students to know by the time they came to their college-level classes. This allowed them to work backward to plan the streamlined developmental math courses.
The course redesign bumped passing rates up to 73 percent — an improvement of 29 percent — and allowed students to go through an entire sequence of developmental education courses in one semester if they worked fast enough.
University leadership buys into analytics
At the executive level, the University of Maryland University College committed to understand and use data to make changes, said Karen Vignare, associate provost of the university and leader of the Center for Innovation in Learning. With executive support, the university added an Office of Analytics, created a cross-functional interventions team and partnered with the Predictive Analytics Reporting Framework, Civitas Learning and Kresge Transfer Data Repository.
In the Predictive Analytics Reporting project, data on retention and progression from 20 institutions is aggregated into a huge database. By looking at this student data, the University of Maryland discovered that it did well in freshman retention, but had a bigger retention problem in year two and three of students' education than the other universities in the project did.
That discovery has sparked experimentation with different interventions in students' second and third years of college, particularly in gateway courses that present serious roadblocks. Some of the tactics they've tried include online tutoring, sending messages to students and starting a required class to help them with career planning, among other things.
"If we don't get the results the hypothesis was telling us, then we're going to have to move forward and figure out the next intervention," Vignare said.
One of the most important things to keep in mind when dealing with analytics and interventions is to balance out the hype and the skepticism. True analytics believers promise too much and don't deliver enough, while skeptics fight back against innovation and change. As a result, true believers and caustic cynics hijack important conversations, which are complicated and require tough mindedness, Milliron said.
Universities need to calm the caustic cynics, temper the true believers and create a space for learning analytics to move forward to the next level, he said.