Everyone agrees durable skills matter. Measuring them is the harder question, and the one Skills for the Future (SFF) is working to address. SFF partners with states to build the infrastructure they need to measure, document and recognize the durable skills students rely on after high school. With Missouri recently joining SFF as a pilot state, we sat down with Danielle Eisenberg, Executive Director of Skills for the Future, to talk about why durable skills have been so hard to capture and what credible measurement looks like.

Eisenberg: One way to explain this gap is a fundamental mismatch between the nature of durable skills and the methods we have traditionally used to assess them.

Standardized assessments are optimized for constructs that are relatively context-independent and can be reliably elicited in a controlled setting. Durable skills do not fit that model. Collaboration looks different in a science lab than it does in a community service project.

That's not a flaw in the measurement; it's a feature of the construct. The field has often tried to solve the variation by stripping out context, standardizing the task and controlling the conditions, so that assessments are comparable across students. But for durable skills, doing that trades validity for reliability. You end up with a score you can compare across students that no longer measures the actual capacity you cared about. SFF's work includes context as a critical part of the signal.

There is also a subtler problem: durable skills are not just context-dependent, they are cumulative. A single strong artifact, like a compelling essay or one group project, cannot tell you on its own whether a student can apply what they learned somewhere else. Can they collaborate just as well on a science experiment as they did on a community service project? Can they communicate clearly in a written report and in a live presentation? That ability to carry a skill across different contexts is what we mean when we say a skill is durable.

Eisenberg: Skills for the Future is built on a specific premise: valid measurement of durable skills requires multiple, varied forms of evidence, interpreted in context. Authentic work artifacts, what students produce through coursework, extracurriculars, employment and community engagement, form the foundation. Those artifacts are assessed against shared Skills Progressions through a workflow that pairs AI with educator review. The AI scans each artifact, identifies passages that constitute evidence of specific skills and proposes ratings against the Progressions. Every proposed rating is tied to the specific evidence in the student's work that supports it. Educators then verify and adjust. The point of the AI is not to remove the educator from the loop, it's to do the initial extraction at scale and surface a structured starting point that the educator can confirm, refine or reject. Every rating ends up with an auditable evidence chain behind it, which matters both for educator trust and for the longer-term work of validating the system itself. Including work from outside school is intentional. If evidence is restricted to school-generated work, the system systematically undercounts what many students can do, particularly those whose strongest demonstrations show up outside the traditional classroom.

Direct assessments help fill in gaps for skills that are hard to see in a finished product. Collaboration, for example, does not really show up in a final essay or report, so we use interactive scenarios that let students demonstrate it in action. Direct assessments aren’t the primary measurement engine. They function instead as signal boosters in this model, deployed strategically to fill evidence gaps that authentic artifacts can't address on their own, like the interpersonal dimensions of collaboration.

Student reflections add another layer: the student's own account of how they built a skill over time, which adds an interpretive dimension and serves a metacognitive role for the student.

Aggregation is what changes the inference. A single assignment can only tell you so much. Across subjects and over time, patterns emerge that no individual artifact can show, and the questions the system can answer shift from "did this student demonstrate critical thinking in this essay" to something closer to what college admissions officers and employers actually want to know: does this student consistently bring this capacity to bear across varied situations?

Since ETS launched its state pilots last year, more than 8,500 students across 79 schools and 336 educators have participated. Early findings suggest the approach is viable, and the conditions for meaningful adoption are becoming clearer. Validation across contexts is ongoing.

Eisenberg: Skills Progressions are research-grounded definitions of what collaboration, communication and critical thinking look like at four developmental levels. These progressions were written by teams including ETS researchers and then went through rounds of review with national domain experts, educators and even stakeholders from higher ed and workforce.

These progressions are attempting to do something less common in this space: they distinguish what's construct-essential about a skill, the features that have to be present for, say, collaboration to be collaboration at all, from what's culturally and contextually variable about how it shows up. Joint intentionality, mutual responsiveness and distributed contribution: those are construct-essential to collaboration. Whether collaboration shows up as visible verbal assertion or as quieter facilitation, whether disagreement is direct or indirect, whether contribution is attributed to the group or the individual: those vary by community and context.

The Progressions are designed to recognize both. That's what lets the same framework score a student leading a science lab, a student facilitating a community organizing meeting and a student supporting a multilingual family decision without flattening any of them into a single template.

If you want to score a written argument, a group project and a reflection on an out-of-school experience against the same skill, you need a common reference for what growth in that skill looks like. We call them a shared language because they give educators, students and policymakers the same vocabulary. The person scoring an artifact, the student understanding where they are and the state interpreting results all work from the same definition.

Eisenberg: Missouri joining this effort, alongside Rhode Island, Nevada, North Carolina, Indiana and Wisconsin, reflects both the growing state-level appetite for this work and the maturation of the infrastructure needed to support it.

What we have learned from North Carolina, where sustained state sponsorship, dedicated educator professional development and clear connections to Portrait of a Graduate frameworks have produced our largest and most rigorous implementation, is now informing how we design for scale elsewhere.

Adoption accelerates when there is alignment at every level, from leadership and policy to technology, and above all when educators themselves are engaged and enthusiastic. That means bringing the right partners to the table early and building the infrastructure and educator supports alongside the tools themselves.

Eisenberg: Durable skills have been undermeasured. According to the Walton Family Foundation and Gallup's Voices of Gen Z study, only 35% of K-12 students feel their education is giving them skills relevant to their future careers. That figure does not mean students are leaving school without durable skills. It means schools have not had effective, credible ways to recognize and document them.

Measuring those skills well requires accepting complexity that simpler systems are designed to avoid: complexity about what counts as evidence, where it can come from, and what it means in context.

It also requires being explicit about what the measurement system can claim today, what is still being validated and what evidence sits behind each piece. We try to be as clear about the limits of our current claims as we are about their strength. That discipline is part of what distinguishes this work from earlier durable-skills efforts, and it's part of why state partners and funders have been willing to invest in the longer arc rather than the quick result.

When students contribute varied forms of evidence, when work from inside and outside school is evaluated through a consistent, research-grounded framework and when insights accumulate over time rather than collapsing into a single score, something shifts. Students are seen more fully. Educators teach with a more complete picture of who is in front of them. And the transcript becomes closer to what it should be: a true representation of a student's body of evidence. We are early in building toward it. The foundation is sound and the learning is real.

To learn more about the Skills for the Future initiative, visit https://www.ets.org/skills-for-future.html