How Predictive Analytics Will Improve Learning

For years, educators have embraced metrics and performance standards as a way to improve everything from test scores to budgeting. But the road to progress has always been paved with more than a few potholes. Although databases and business intelligence tools offer valuable information and insights into overall performance, they provide little more than a snapshot of a district at any given moment.

Predictive analytics is poised to change things — and offer a motion-picture view of events. The software uses statistical techniques and algorithms to process current and historical data and make predictions about future events. It is gaining acceptance at the K–12 levels, because “a consensus has clearly emerged that educators need a greater understanding of how children learn in order to better shape their progress,” says Yeow Meng Thum, assistant professor of measurement and quantitative methods at Michigan State University.

When predictive analytics is used effectively, Thum says, it can help mold classroom instruction, direct resources and funding, and measure which strategies and approaches are likely to produce the best results. He says that it is possible to develop “a culture that fully embraces the use of information in planning for results.”

Districts in Alabama, California, Colo­rado, Illinois, Iowa, Louisiana, Minnesota, New York and Tennessee are turning to predictive analytics to blaze paths to success.

Bold Predictions

Although the use of analytics in the educational arena is common, predictive analytics is only now catching on at the K–12 levels. One reason is that software providers such as SPSS and SAS just began to offer products tailored for use in schools. Another factor is that the effective use of predictive analytics requires trained analysts who are able to manage large amounts of data and interpret results. Finally, district administrators and teachers must accept that changes will take place as predictive analytics takes hold.

One school district that has embraced predictive analytics is Naperville District 203 in the Chicago suburb of Naperville, Ill. With 18,500 students and 21 schools, there’s a growing need to understand academic trends and patterns — despite the district’s ranking as one of the highest-achieving districts nationally. “We want to use predictive analytics to keep the growth curve up and improve on the way we use data,” explains Superintendent Alan Leis.

The district is using predictive analytics to understand learning trends, catch problems and identify successful methods. It is also tracking enrollment patterns so that it can better allocate resources. “Rather than examining whether a student is merely getting better, we’re able to examine the growth curves and get an idea of how they are likely to perform in the future,” he says.

That might translate into the district devoting more time to fractions for a particular class or examining why a student is losing momentum in English. “It can be about curriculum, teachers, special program services, or emotional or learning problems in a kid’s life. This allows us to be more proactive,” Leis says.

Another district turning to predictive analytics is the Hamilton County Education Department in Chattanooga, Tenn. With performance pressures resulting from No Child Left Behind, the district, which has more than 80 schools and 40,000 students, turned to predictive analytics to estimate how children will do on standardized tests. The software displays the probability — expressed as a percentage — that a child will score at levels such as “proficient” or “advanced.”

It’s no small matter for the department. In 2000, Chattanooga held the dubious distinction of having nine of the worst elementary schools in the state, with only 18 percent of third-graders reading at or above grade level. Through an array of initiatives, including the use of analytics, the district has taken a more proactive stance. “The district uses the information to assess whether it needs an in-service in reading or math. We are able to predict that we’re going to have problems based on scoring patterns,” says Kirk Kelly, director of accountability and testing.

Today, 74 percent of students test “proficient” or “advanced” in reading, and elementary students are performing better than 90 percent of the students in the state. “Predictive analytics provides visibility that otherwise wouldn’t exist,” Kelly says.

Although the tool is particularly valuable for monitoring performance, it can also pay dividends in areas such as budgeting and finance. For instance, the Poway Unified School District in San Diego County, Calif., which has 34 schools and more than 32,000 students, is introducing an analytics tool to better understand budgeting and manage cash flow more effectively. The task is complicated by the fact that the district’s overall enrollment remains steady — though certain schools are affected by a decline in enrollment while others are growing quickly. “Allocating staff and resources is an ongoing challenge,” Poway CFO Randie Allen says.

Thum says that success with predictive analytics requires equal parts technical proficiency — including understanding how to use the technology effectively — and business acumen. It’s necessary to focus on achieving principal objectives and revisit goals and problem points. “Predictive analytics is not so much about responding to an educational crisis as it is about improving education,” says Thum.