Toggle open
Learning Library Blog Measuring what matters
Expand breadcrumbs

Measuring what matters

By Chris Frisella
September 25, 2019
Measure20495 Version Id C27 LV031kvz j L Fe Lpy M vs F Pq B p Xv

As educators pursue more personalized, student-driven and inquiry-based approaches to learning, one obstacle that continues to rear its head is the inability of widely used assessment models to successfully measure student learning and growth.

Critics say traditional models simply can’t assess the type of learning that schools need to promote. They provide incomplete, inadequate and ill-timed feedback. They advantage some students and disadvantage others. And at their worst, they burden teachers, disrupt classrooms and stress out students.

Like many adults, Kristen Brooks remembers the stress that tests caused her as a child. “I wasn’t a great test taker. Even when I knew that I knew the information, I would still get test anxiety,” Brooks says. “I remember thinking, ‘I wish I could do it a different way.’”

As an instructional technology specialist for the Cherokee County School District in Georgia, Brooks works to make that wish come true as one of the many educators and researchers championing and working to develop new ways of assessing learning.

Take educator Cindy Herren, K-5 technology teacher for Waukee Community School District in Iowa, for example. Like Brooks, she advocates for using widely available creation applications – such as Adobe Spark, FlipGrid, Book Creator and Seesaw – to assess student mastery. Using these tools, students tap academic, technical and problem-solving skills, and ultimately “feel empowered” rather than stressed out, says Herren.  One example Herren cites involved helping a first-year German teacher transform one of her school’s long-standing summative writing assessments into a more dynamic creative project. The traditional assignment asked students to keep a three-day diary in which students typically used simple German phrases to recount mundane activities, such as “I got up,” “I ate breakfast,” “I rode the bus to school.” The teacher then would need to review all of those phrases – for each of 100 students.

Herren suggested letting students instead make videos using the Clips app to capture images of them doing an activity, audio of them describing it in German and edited captions demonstrating their writing skills.

“These videos were amazing, and the kids had so much fun,” Herren says. What’s more, when colleagues saw how successful the new teacher’s approach had been, they adopted it in their classrooms, too.

Bringing the learning sciences to assessment

Such project-based assessments are one of the alternatives to traditional testing that combine the learning sciences and technology, a potent blend that offers many approaches to assess and help students retain and use knowledge. The learning sciences point to three stages of learning: encoding, storage and retrieval, says Patrice Bain, author, educator and consultant. Studies show that last stage – pulling information back out of storage – improves learning and test scores.

“What the learning sciences say is that the best way we can help our students learn is through frequent retrieval, and it is best done as no stakes or low stakes,” Bain says. “Studies also have shown learning and teaching this way directly reduce anxiety in students versus those high-stakes tests.”
Along with retrieval, students need feedback that clarifies what they know and what they don’t know, and increases metacognition, or their understanding of how they learn, Bain says. “When we are able to do that – to teach them how to discriminate and to focus their time on what they don’t know – that really aids learning.”

Two other principles from the learning sciences that technology and new assessment tools can harness are spaced practice and interleaving. Spacing is spreading lessons and retrieval out over time. Interleaving involves mixing skills or topics so that students have to think about the process or concepts involved, Bain explains.

Clearly, human learning is a complex process.

Defining 2 key principles

Spacing and interleaving are powerful strategies that boost learning, according to cognitive scientist Pooja Agarwal, a learning science website founder and co-author with Patrice Bain of the book Powerful Teaching: Unleash the Science of Learning.

Spacing involves taking a given amount of learning time – whether instructional time or study time – and breaking it into multiple sessions spread out over time, Agarwal’s website explains. By spreading out lessons and retrieval opportunities, students’ knowledge has time to rest and be refreshed.

For example, instead of cramming the night before a Spanish vocabulary exam by trying repeatedly – say, three times – to retrieve the English translation for each word, a student might practice retrieving the translations for the vocabulary words one time on three separate occasions – say, a week before the exam, a few days later and again a few days after that.

The two approaches involve the same amount of learning time but allocate it differently, explains a resource guide on the website.

Interleaving is the practice of switching between ideas during a learning session and not focusing on any one idea for too long. It also strengthens understanding and helps students connect different ideas.
According to Agarwal’s, educators can incorporate these same practices to make cumulative assessments more powerful.

Typical cumulative exams assess learning of all content covered throughout the semester or school year, the website notes. While students may re-study material learned months earlier, they may simply cram, which fails to reap spacing’s full benefit to learning. Similarly, cumulative exams may fail to take advantage of interleaving. That’s because, while course topics on a cumulative exam are likely to be related (e.g., covering key concepts from earth science), they may not require discrimination (e.g., the atmosphere vs. plate tectonics), the site explains.

Agarwal’s website recommends encouraging even more spacing and discouraging cramming with cumulative mini-quizzes throughout the semester, not just as an end-of- semester exam, and ensuring that cumulative mini-quizzes, activities and exams include similar concepts that require careful discrimination from students, not simply related topics.

Capturing students’ thought process

Every learning activity draws on and has the potential to affect identity, social and emotional capacities, cognitive capacities including executive function and metacognition, and even physical and mental wellness, explains Bror Saxberg, vice president of learning science at the Chan Zuckerberg Initiative (CZI), citing work by CZI colleague Brooke Stafford-Brizard on the comprehensive student development framework.

“There’s just no way to separate these things,” says Saxberg. “Inside a real brain, you can’t just pull out the math processor and tinker with it on a bench and upgrade it and then reinstall it.”
One of the best ways technology can improve assessment and evidence-gathering is to capture more of the student’s thought process toward a solution, Saxberg says. “We need better formative assessments – and not just assessments of did you get the practice ex- ample right or wrong, but actual evaluations and evidence-gathering around the students’ process.”

Instead of simply handing in a finished essay, a student could work in a tool, for example PowerNotes, that lets them start with an outline, access references and resources from within it, then re- vise their outline, craft introductions and conclusions, and finalize their essay.

“The technology can start to capture that and maybe even scaffold that,” Saxberg says. “It can see that process and compare it with others” – possibly using machine learning.

Andreas Oranje, general manager for research at Educational Testing Service (ETS), agrees. He says technology-based curriculum activities can produce and analyze a wealth of process data and other information to help the student and teacher make decisions about what to learn next. That data can help ETS and others figure out the cognitive processes underlying a student’s learning and thinking, identify possible student mistakes and misconceptions, then anticipate and prevent them, or determine how to intervene in a targeted way if they happen.

Process data also could inform a fuller picture of student performance that augments end-of-year standardized testing, Oranje says.

Artificial intelligence (AI) similarly could mine process data to match peers and near peers to enhance collaborative learning, says ISTE Chief Learning Officer Joseph South. “Imagine a system that understands that Student A has a particular misconception that Student B just mastered a couple of days ago,” he says. “We can use that system to match Student A to Student B and have Student B tutor Student A and help them find the same mastery they just achieved.”

“Just knowing where you are can make a big difference in the motivation to improve.”

Another important way researchers believe technology can better gauge and support learning is by providing real-time feedback. “If we can shorten the gap between when a student performs and when a student gets feedback on that performance, then we can be making course corrections sooner, which can allow the student to improve that much quicker,” South says. “We don’t want to under- estimate the visibility these tools can provide to the students and the parents.

“Just knowing where you are can make a big difference in the motivation to improve,” he says. “That’s why we carry around FitBits. When we know how many steps we’ve taken, suddenly we’re motivated to take the extra 2,000 to get us to our 10,000-a-day goal. Kids are the same way.”

Improving data steps

Technology can also improve how assessment data are gathered, analyzed and presented.

By rolling assessment data up into an actionable dashboard, it can give an educator “a 360-degree view” of a student’s schoolwide performance and how to help them, South says. Similarly, some new assessment tools not only tell parents sooner about their student’s areas of need, but provide recommendations about how to help the student at home.

“Through technology, they could build bridges out of simulated steel and buildings out of simulated materials.

Then you could simulate an earthquake and see if their structures can withstand those forces. It’s a whole other level of learning and understanding and a whole other level of fidelity to the real world.”
At Northwest Evaluation Association’s (NWEA) Product Innovation Center, Senior Director Mike Nesterak and his team are looking for better feedback tools for teachers, students and parents. “We want to make sure that the language and method of providing information is appropriate for each stakeholder,” Nesterak says.

One tool NWEA is exploring is a website prototype where parents can enter one of their child’s assessment scores and then view an automatically generated video discussing how their child did relative to past scores and the scores of other students, and get predictions about growth.

Another innovation center project involves inquiry-based reporting, mining NWEA’s comprehensive reporting system to pro- duce reports answering principals’ top questions about assessment data.

Principals, teachers and administrators have particular information needs at any particular point in time, Nesterak says. The problem is NWEA has so much data in its system that it can be difficult to know what, when and how to extract relevant information. So, the innovation center surveyed middle school principals, asking for the 10 most important pieces of information they needed from NWEA’s assessment data. Researchers then used that feedback to shape a 20-page “insight report” answering those key questions for principals in text, graphs and tables. Now, researchers are prototyping a similar report for district-level administrators, and they plan to develop reports for teachers and parents, too.

Technology also makes through-year testing a more manageable proposition.

NWEA has been piloting a through-year model for Louisiana that would assess students over two days four times a year on materials related to a just-completed instructional unit. To make the tests more fair, the “cold reads” used in most assessments are replaced with a “hot read” that allows students to refer to the text during the test, a “warm read” involving material that’s new but related to the topic they’ve been studying, and a final section that asks them to analyze and synthesize information from both passages and draw conclusions.

The data from all four assessments will be aggregated into an English language arts score. “We assess knowledge, as opposed to just skills,” Nesterak says. “We’ve eliminated the end-of-year assessment, and we’ve embedded that assessment into the instructional assessment.”

Technology and formative assessment

The Louisiana pilot model underscores another broad benefit of technology. “If you use technology for the formative assessments right, then you can get the same or most of the same information you would gather in summative assessments along the way,” South says. “It doesn’t mean we don’t need some sort of summative assessment, but the summative assessments should be the cherry on top.”

Philadelphia educator Chris Lehmann also stresses the interplay between technology and formative assessment. “Because of its levels and speed of communication, technology allows us to make formative assessment far more often,” says Lehmann, founding principal at Science Leadership Academy, which emphasizes inquiry-driven, project-based learning.

For inquiry-based models, immersive technology holds great promise. It can make teaching and assessing science much more vivid, clear, efficient, accessible and safe, Oranje says. “There are certain things … you just cannot do in the classroom,” he says. “You cannot zoom in to the molecular level in a classroom. You cannot do experiments that require a lot of time – for example, if you want to explain something about fossils.”

“We’ve eliminated the end-of-year assessment, and we’ve embedded that assessment into the
instructional assessment.”

Virtual laboratories and dialogic systems let students work through an experiment or practice complex skills or behaviors, learning from their mistakes without real-world consequences.

“For a long time, we’ve had students build toothpick or Popsicle stick bridges to understand basic principles of engineering,” South says. “Through technology, they could build bridges out of simulated steel and buildings out of simulated materials. Then you could simulate an earthquake and see if their structures can withstand those forces. It’s a whole other level of learning and understanding and a whole other level of fidelity to the real world.”

There’s tremendous power in that ability to layer onto tasks additional complexity and texture, such as time constraints and distractions, when real-time data shows that the student is ready for it, CZI’s Saxberg says.

At NWEA’s Product Innovation Center, where researchers have built an augmented reality science lab prototype that allows students to conduct three experiments anywhere in a room, Nesterak sees the immersive technology as a way to level the playing field for schools that lack lab space, equipment and materials. It also boosts engagement. “Students love it,” he says, adding that researchers are still evaluating how it affects student performance.

Nesterak sees possibility for augmented reality in other areas, including English language arts, where it might allow a character to “come to life” while the student reads.

Another promising emerging technology is conversational computing, Nesterak says. At NWEA, researchers have built a prototype that would allow teachers to converse with Amazon’s Alexa to retrieve information about their students’ math assessment performance. To move past prototype, the tool would need to be FERPA compliant. NWEA researchers also are exploring whether they can develop an assessment that’s given through a conversational avatar or conversational entity like Alexa.

The device could follow a student’s approach to solving a problem by tracking the questions that the child poses to it. That data, in turn, could provide a fuller picture of student performance and help improve assessments, perhaps even tailoring future tests to an individual student’s problem-solving approach.

Nesterak also sees AI as a tremendous opportunity not only for scoring assessments, where it’s already built a good track record, but also for big data analysis. NWEA already is using AI to analyze its extensive database of individual assessment scores for patterns of performance and find a more empirical learning progression.

Finally, technology can help eliminate some of traditional assessment’s blind spots. “There are things that are important to a student’s success that are difficult to measure on a piece of paper,” South says, pointing to things like social and emotional learning and creative or collaborative problem-solving. “Technology can really be an important tool in helping us do that.… It opens up many more avenues for expression that can help us understand a more nuanced array of their abilities.”

Engaging educators

So, what needs to happen to realize technology’s potential to reshape learning and assessment?
First, solution providers need to engage educators throughout the development process to ensure they’re creating a tool that meets their needs and makes their lives easier, not harder, South says. Then, educators need to be aware of what’s out there and be willing to learn and use those tools. Finally, school leaders need to support competency-based approaches, which may require more than just flexibility in how and when testing takes place, he says.

It may take changes to classroom procedures and culture, teacher education, and student autonomy and responsibility.

“The best thing that teachers and policymakers and purchasers and school district people and principals can do is make sure they’re paying attention to learning first,” Saxberg says. “You’ve got to think about what is it that my learners really need to improve their performance and then ask how can technology help make that happen.”

The pursuit of equity

Alicia Johal agrees putting learning first is key, and she sees improving equity as part of that. Before becoming assistant director of the Center for Innovation and Entrepreneurial Thinking at the San Diego Jewish Academy, Johal taught middle school science, biotechnology and marine biology for seven years. Many of the children in the south San Diego public schools where she taught came from low-income households, and most were English language learners.

Johal soon realized that the traditional assessments she was giving her students failed to show how much they actually knew about science because the tests imposed barriers for them. Those barriers, or biases, can present in different ways. Does the assessment accurately measure what it’s meant to measure? Does its content give one student subgroup an advantage over another? Do test questions assume background knowledge or shared experiences that one subgroup is less likely to possess?

For example, one gender, cultural or ethnic group may be at a disadvantage answering a word problem about velocity and acceleration based on a particular competitive sport. The question may be a barrier for students who have never played the sport. “On top of trying to solve this velocity problem, they have no idea why a football is being thrown in a spiral toward the end zone. What if they don’t even know what the end zone is?” Johal says. “There’s so many ways to write this stuff that you’re not singling out students based on the experiences that they have.”

To improve equity and outcomes, Johal made a cultural shift, changing how she assessed, graded and taught her students.

In addition to writing more of her own questions to reduce barriers, Johal adopted two-part standards that separated skill from content.

Johal also began providing a rubric at the bottom of each assessment. Using student-friendly terms, the rubric spelled out the skill and content criteria required to earn a grade of “Mastery.” The only other grading option was “Work in Progress,” and Johal allowed students to fix mistakes and resubmit an assessment until they achieved Mastery. “It became part of the culture of the class,” Johal says.

Johal found another way to make some assessments more accessible and learning-supportive: student videos. For example, instead of requiring her students to write a traditional lab report, she first had them use the Explain Everything app to narrate a video lab report complemented by photos, data tables and text. “A lot of language learners are really good at speaking to you about science and telling you what they know, but when you ask them to write it, that’s more difficult,” Johal says. “It’s easier for them to write if they can speak about their knowledge first.”

Johal used the videos as a scaffold for written reports. “It was a way for me to use technology to support language learning, and it was a game-changer with every class,” she says.

Like Johal, NWEA’s Nesterak and others, Oranje at ETS emphasizes the pursuit of equity.

“It’s great to come up with new things, but we have to make sure it’s fair and accessible,” Oranje says. “It needs to be affordable. It has to be culturally unbiased or cross-biased. And it needs to work for students with a range of needs.

“It’s a moral imperative: That’s what we need to do.”

Reshan Richards, director of studies at New Canaan Country School in Connecticut, suggests teachers ask themselves three questions before assessing students. Find out what those three things are in the video below: 



Got learning science? Design lessons that work! Explore ISTE's book Design Ed.

Chris Frisella is a freelance writer who explores educational technology and its power to reshape learning and lives. He previously spent more than 30 years writing and editing for newspapers in Oregon and Alaska. He lives in Eugene, Oregon, where he's enjoyed volunteering in area schools.