Classroom Analytics 101: How Teachers Propel Learning With the Right Data

Teachers have always been great at data collection and analytics. This was true long before they had the tech tools to facilitate those activities.

Instead, they relied on gradebooks, roll calls and well-honed intuition to get a feel for how their students were doing.

Today, of course, the technology available to teachers can provide a much more robust view of their classrooms. They can pull attendance statistics, qualitative progress reports for each student, and so much more.

But at some point, classroom tech fatigue sets in. As we will see in a moment, some teachers have begun to feel more like data input specialists than actual educators.

The challenge for edtech moving forward is efficiency. Data certainly has a role to play in modern classrooms. It’s just a matter of understanding which data teachers should prioritize, and then facilitating the collection and analysis of that data. That way, teachers can spend less time worrying about data management and more time focusing on what’s important — teaching.

Below is a snapshot of where classroom analytics today can empower teachers to be the best educators they can.

What Teachers Have Learned From Years of Data Collection

Benjamin Herold at Education Week gives an excellent overview of how far edtech has come in just the last few years.

School districts now maintain huge databases on how students are performing, individually, and what attendance patterns are emerging at a collective level, he writes. At the classroom level, “learning software and digital games that generate extensive data that can be mined for evidence of student learning.”

The sophistication we see in edtech software today didn’t evolve overnight, however. As Blackboard’s Mike Sharkey argues, edtech’s hype cycle reached the trough of disillusionment sometime during the middle of the decade. Since then, “vendors and institutions alike are coming to understand that simply providing access to data doesn’t solve problems,” he writes. “Those who have done work with data know that it takes time, commitment, and some wrangling to get the brass ring.”

Today, educators understand that it isn’t enough to simply ingest tons of classroom data. Teachers need to be smart about what data they collect because they don’t have the time to sift through noise.

That means good classroom data analytics start with teacher decision making. More than administration-side or even vendor-side pressures, it’s up to individual teachers to decide what data is appropriate for their classrooms, writes Greg Watson, chief executive at GL Assessment.

“Ultimately, smart data leads to something — it isn’t an end in itself but it provides enough information for a teacher to make a decision about a student. It liberates, it does not control. The right data can make a real difference and simple data can make very smart schools.”

How Technology is Empowering Educators Today

So, what are teachers doing with data to create better learning outcomes and distinguishing very smart schools?

Georgia Tech researcher Jon Bidwell has spent a lot of time getting to the bottom of this question. A few years ago, he and a team of colleagues created a pilot program to measure and track student engagement in the classroom.

The premise was simple enough. They wanted to know how much attention individual students paid to their teachers. So, the researchers went to a third-grade classroom in Durham, North Carolina, to set up video equipment that captured where students focused their gaze during a lesson. On the backend, the software they developed could measure and model each student’s head position, and a trained researcher could then help classify the student’s level of engagement over time.

With that data, a teacher could know which students were attentive and for how long. Perhaps more importantly, the data could reveal which students were only casually observing but not actually devoting appropriate levels of attention to a lesson.

At the same time on the other side of the country in Washington, the Tacoma Public Schools district adopted a data-driven program to boost its graduation rate, which hovered just above 50 percent a decade ago.

As the Microsoft Education team notes, teachers in that school district began to collect data about student performance, attendance, discipline and behavior in general. All of these factored into a predictive model of student progress. With a more robust, classroom-level picture of what student success looks like, administrators were able to take action. The results were quick: By 2016, graduation rates had reached 85 percent, which is higher than the national average.

What Data-Savvy Teachers Are Measuring

A good model for classroom data collection takes into account what teachers are already great at — tracking grades, for example — and integrates some capabilities that technology makes possible.

Monica Fuglei, an adjunct faculty member of Arapahoe Community College in Colorado, offers one such model that can be molded to fit any classroom. Fuglei suggests there are three broad categories of metrics teachers should focus on. These include:

• Summative metrics. Think grades and performance evaluations.
• Formative metrics. This would include things like daily quiz scores or any information gleaned from “informal, low-stakes assessments.” This data will help teachers spot learning gaps so they can adjust lesson plans accordingly.
• Individual metrics. These are the harder data points to collect and analyze. In fact, this part of the model needs its own section to unpack.

How Teachers Can Track Each Student’s Needs and Achievements

The Go Pollock team argues that a data-backed lesson plan that speaks to each student on an individual level should address three components. This will give us a sufficient framework for understanding where individual metrics come into play.

First is the student’s skill level. Summative and formative metrics will be useful in assessing how well each student understands and is prepared to engage with a subject.

Next is the student’s interest level. To assess this, teachers need qualitative, open-ended feedback from students to get an idea of how much any given subject aligns with a student’s interests.

Here’s where the model dovetails with important, ongoing conversations about assessing student mindsets. There is now plenty of data to suggest that a student’s mindset is, above almost everything else, the factor that most determines their educational success.

McKinsey researchers Mona Mourshed, Marc Krawitz and Emma Dorn even demonstrate how a student’s mindset — their attitudes, their beliefs, their very desire to learn — is a better predictor of their success in school than even their socioeconomic background. Mindset carries twice the weight of background.

Finally, there is the student’s learning profile. This data speaks to the variety of ways students grasp and comprehend the material that teachers present. Some students do better in groups, for example, whereas quiet students might prefer to study on their own. Teachers have long been able to intuit this info. Tracking it on a dashboard, however, can paint a better picture of the entire class’s collective learning profile.

Deploying These Analytics: Three Major Challenges Remain

At both the classroom and the administrative levels, educators face a few key hurdles to using student data in a robust way.

1. Student Data Privacy

The first challenge is pretty straightforward, so let’s start there. What does it mean for teachers and administrators to handle student data responsibly? How do they safeguard that data?

As The Atlantic’s Andrew Giambrone reported, one very well-funded nonprofit folded in 2014 because it couldn’t overcome this hurdle. “Parents worried that their children’s personal information — stored by inBloom to help improve academic performance in public schools across the country — could be misused, sold, or breached,” Giambrone wrote in 2015.

That data included family composition profiles, whether students were eligible for free lunch and even some medical conditions. Understandably, parents and teachers don’t want that information to leak.

Mike Patterson, a former education strategist at CDW•G, offers a handful of best practices educators should be aware of when handling sensitive student data:

• Create a privacy pledge that outlines how relevant data will be collected.
• Facilitate transparent conversations between parents and educators about how data are collected and used.
• Loop students in, too. They need to know how their data are being used so they can ultimately learn to be responsible digital citizens.

2. Learning How to Measure What Really Matters

Perhaps more than any other organization, the Clayton Christensen Institute in California has been a driving force in helping educators understand what data they should be tracking — and what data they can ignore.

For example, think back to what the McKinsey researchers found about socioeconomic background. They don’t argue that it’s not important. They simply argue that it’s less predictive of success than student interest.

But background is still something worth tracking. It just requires a better framework so real meaning can be extracted from that data.

Julia Freeland Fisher, director of education at the Clayton Christensen Institute, offers a model in which educators could measure the opportunities available to students. “Students today are competing on an even more complex playing field, one that’s often masked by statistics on income and achievement gaps,” she writes. “A well-resourced childhood introduces a whole new set of inequities between rich and poor students, and those whose parents have or have not earned college degrees: social gaps.

“Gaps in students’ networks matter immensely in both immediate and longer-term measures. Research groups like the Search Institute have shown that developmental relationships drive everything from higher grades to persistence in school. Down the line, an estimated 50 percent of jobs come through personal relationships.”

Nevertheless, Thomas Arnett, a senior research fellow at same Institute, argues that no system should discount the wisdom and insights that teachers accrue over their careers. “Even in classrooms with the latest adaptive learning technology, an expert teachers’ professional intuition is still the best way to understand and address the myriad cognitive, non-cognitive, social, emotional, and academic factors that affect students’ achievement,” he writes.

Eric Sheninger, a senior fellow at the International Center for Leadership in Education, echoes this sentiment, and says that technology must always follow pedagogy. Each tool must demonstrate how it supports learning outcomes.

3. Making It More Efficient to Collect Qualitative Data

Finally, let’s return to Fuglei’s model of summative metrics, formative metrics and individual metrics.

Each component of that model requires more qualitative data than the component before it. Whereas summative data (i.e., grades) is fairly easy to measure, formative metrics are a little bit more subjective. Individual metrics can be even more subjective and open-ended.

In the past, that kind of data has been hard for teachers to collect, track and analyze. However, authors Ron Berger, Leah Rugen and Libby Woodfin write in their book “Leaders of Their Own Learning: Transforming Schools Through Student-Engaged Assessment” that this harder-to-quantify data is crucial for improving student learning outcomes.

Accessing this data is just a matter of knowing where to look, and knowing how to capture that data. “Rubrics, which are composed of qualitative descriptions of student work, are filled with this kind of data,” they write. “Many recording forms, such as journals, note catchers, and entrance and exit tickets can be powerful data sources to track the why and how of student thinking.”

We have built Jotform to be an indispensable tool for collecting this kind of data, too. It’s perfect for things like online quizzes, homework submission and even student surveys.

The key to Fuglei’s model — and to data-driven education in general — is ensuring that teachers have easy access to tools that allow them to collect hard-to-quantify information efficiently. They’re already masters at grade-keeping and using their own intuition to connect with students.

It’s the job of edtech companies to simply provide them with the right tools, then step out of the way so teachers can do what they do best: Teach.

images by: Rawpixel, Avel Chuklanov