Core academic subjects have long been the top priority for districts preparing students for life after high school. Now districts are elevating social and emotional learning, as well, drawing from a large body of research showing that skills like grit and self-regulation are critical to life success. If they’re so valuable, the logic goes, schools should find concrete ways to teach them. And with teaching comes testing — not only to ensure students’ learning can be measured, but so that educators can determine how well the new efforts are working.
“As more programs are being taken up in schools and districts, there becomes this greater demand to assess them, to see if they’re working, to see if students are, in fact, learning the skills that are being taught,” said Lindsay Read, manager of research at the Collaborative for Academic, Social and Emotional Learning, or CASEL.
While there are some assessments on the market, Read said there is a clear need for more direct assessment options, as distinct from student or teacher surveys of social and emotional skills. And when it comes to direct assessment, there’s a need for tests that can be administered quickly in classrooms and return results to teachers right away.
Enter the Measuring SEL design challenge. CASEL is spearheading a collaborative effort to help the field coalesce around practical and appropriate design principles that developers can use to make future assessments. The first of three annual design challenges produced seven top designs, all at varying stages of development. Some haven’t been used in schools at all, while others are already being piloted.
The first-place design, submitted by NWEA and Santa Ana Unified School District, makes use of metadata collected when students take other computer-based tests. It assesses self-regulation and self-management based on how long students spend on each test question, effectively combining a measure of social-emotional learning with existing standardized tests.
Raegen Miller, research director at the Georgetown University think tank FutureEd, expects clever technology approaches like this to become more common. He has a forthcoming report about the focus on social and emotional learning in a coalition of California urban districts that have been early leaders in SEL measurement, primarily with the use of surveys.
The second-place design, reflecting a very different assessment method, is called Social Detective, and aims to measure students’ ability to assess other people’s values, interests and perspectives based on video clips of them.
Sam Moulton, research director at Panorama Education, which created the tool, said social perspective-taking is a foundational human capacity that underlies a variety of everyday interactions. It also happens to be one that research has indicated is both malleable and teachable. The umbrella of social-emotional skills is broad and some skills seem to be more fixed based on individual personalities or not as easy to teach in a classroom setting. Social perspective-taking seems to avoid both pitfalls.
In Social Detective, students watch video clips of people answering questions like “What makes a good friend?” and then they describe the person in their own words, answering multiple-choice questions like “What is more important to this person, loyalty or honesty?” In the assessment itself, students can practice social perspective-taking. They get immediate feedback about how their answers match the answers of the people who appeared in the videos. A full assessment includes 15 minutes of watching a video, answering questions, getting feedback and repeating the exercise at least three times.
With additional design challenges planned for the spring of 2018 and 2019, the hope is that schools soon will have many good options for measuring social-emotional learning. But that will cross only one challenge off the list.
Miller said the next concern for the field of SEL measurement is how educators should interpret assessment results. What do they say about students and teachers?
“Validity is a question of what inferences are appropriate,” Miller said. “And that’s not a purely scientific thing.”