How to assess career readiness at an institutional level

A career-readiness assessment measures student proficiency against a competency framework. That sentence contains the two decisions that determine whether the result is usable: which framework, and which measurement method.

Four methods are in general use. They disagree with each other in a predictable direction, and understanding why is most of what institutional career-readiness assessment consists of.

The four methods

Self-report. Ask students to rate their own proficiency against the framework’s competencies. Cheap, scalable, easy to administer at census points, and it runs systematically high, because it measures confidence, which is a real and different variable.

Artifact scoring. Score what a student produced (a portfolio, capstone, project record, reflection, presentation) against a rubric tied to the framework. The most defensible method available to an institution and the most labour-intensive. It also produces evidence a student can carry into an interview, which none of the other three do.

Third-party ratings. Ask an internship supervisor, employer partner, clinical preceptor or community-project sponsor to rate proficiency. The only method that measures what employers are actually judging, and the one whose coverage is capped by placement volume.

Behavioural records. Participation, milestone completion, co-curricular transcripts. These measure whether a student did the developmental work, not how well. Weak as proficiency evidence, strong as reach evidence.

Why they disagree, and what to do about it

The direction of disagreement is consistent: self-report high, supervisor ratings low, artifact scoring somewhere between and dependent on rubric severity.

Our State of Higher Ed 2026 research shows the spread at national scale. 83.8% of undergraduates said their education prepared them for the workforce; among recruiters, 37% agreed. That 47-point difference is partly a skills gap and partly the arithmetic of asking two populations two different questions.

The wrong response is to blend the methods into a single institutional figure. Doing that hides the only interesting finding in the data (the distance between how ready students feel and what employers observe) and produces a number nobody can act on.

The right response is to report methods separately, with the method named on every figure. “76% of seniors rate themselves proficient; 41% of internship supervisors rate them proficient” is two useful sentences. Their average is zero useful sentences.

Coverage is the constraint nobody plans for

Third-party ratings are the highest-validity method and the least available. Only 19.4% of undergraduates have completed an internship, so for most institutions supervisor ratings can describe a minority of the graduating class at best, and a self-selected minority at that.

This is worth confronting before you design around it, because a plan that depends on supervisor ratings for the whole class quietly becomes a plan that describes your most advantaged students. Two workarounds are practical:

• Expand who counts as a third-party rater. Community-project sponsors, service-learning partners, clinical supervisors, employer judges on a capstone panel, and student-organisation advisors can all score against a rubric.
• Use artifacts to cover the rest. Artifact scoring has no placement dependency, which is why it is the backbone of any assessment plan meant to describe an entire cohort.

Two structural rules

Assess at two points, not one. A single senior-year measurement cannot separate what the institution added from who enrolled. An entry baseline, even a short self-report inventory built into onboarding, turns a snapshot into a gain, and gain is the claim leadership actually wants.

Disaggregate before you present. First-generation, commuter, part-time, transfer, adult and online students experience career preparation very differently. An overall figure carried by residential full-time students looks like an institutional result and is not one. Career-services engagement makes the point: 84.3% of undergraduates know career services exist, only 46% of those have used them, and the students least likely to convert awareness into a visit are the ones furthest from a first job.

Making it survive program review

Career-readiness assessment usually has two audiences: an internal one that wants to know whether the programming works, and a review body that wants to know whether it is being assessed at all. The second audience has a framework.

The CAS Learning and Development Outcomes, from the Council for the Advancement of Standards in Higher Education, organise student learning into six domains including cognitive complexity, intrapersonal development, interpersonal competence and practical competence. A program complying with CAS standards must identify relevant learning from those domains, assess it, and explicitly account for the domains it does not assess.

That last clause is the useful one. CAS does not require you to assess everything. It requires you to say what you assessed and what you did not, which is a far more achievable standard than most departments assume. A written statement of scope is what turns a partial assessment from a weakness into a documented design decision.

If your employer-facing reporting runs on NACE competencies and your program review runs on CAS domains, build the crosswalk once, write it down, and stop rebuilding it every cycle. The career readiness reference sets out how the frameworks relate.

What to write down before you collect anything

Five decisions, recorded before the first data point:

• The framework and its edition. Frameworks get revised. An unlabelled year-over-year comparison across two editions is not a trend.
• The method attached to each figure. Named on the figure, not in a footnote.
• The population and the threshold. Which students, and what counts as proficient, decided before you see the distribution.
• The disaggregations you will publish. Committing in advance prevents the quiet decision not to publish an inconvenient cut.
• The owner. One office accountable for the number, with the authority to report it upward.

Where the data comes from

The cheapest assessment plan is one where the evidence already exists because the programming generates it. A program built around real work with required reflection produces artifacts continuously; a program built around events produces attendance.

That is the practical case for hosting career-readiness assessment inside a structured co-curricular program rather than bolting a survey onto the senior year. The NSLS curriculum is designed this way (the Advanced credit-recommended course centres on a community team project with a DISC assessment, work that a rubric can score and a student can describe in an interview), and chapters receive annual program assessments as part of what’s included.

Decide the framework, name the method, then see how to design the program that produces the evidence.