Internship assessment that produces evidence

Why completion forms are not assessment

Most internship programs collect a supervisor evaluation, and most of those evaluations are a satisfaction form: was the student punctual, would you host again, any comments. It confirms the placement happened. It does not establish what the student gained, it cannot be aggregated across a cohort, and it will not survive an accreditor asking how the program knows its students are developing professionally.

Assessment means measuring the same defined competencies, the same way, across students and across terms, from someone who observed the work.

The method that produces evidence

1. Assess against a shared framework. The eight NACE competencies are the

framework employers and accreditors already recognise, which is why a rating in them travels outside your office.

1. Have the student rate themselves first, before they see anything from

their supervisor. This is quick and it is the baseline for everything below.

1. Have the site supervisor rate the same competencies. Frequency-based

wording ("how often did this student demonstrate…") gets more honest answers than quality-based wording, because it asks about observed behaviour rather than inviting a verdict on a young person.

1. Compare the two. The gap is the finding. A student who rates themselves

above their supervisor on communication has a specific, coachable issue. A student who rates themselves well below is often the strongest intern in the cohort and does not know it.

1. Aggregate by cohort, major and term so the program, not just the student,

learns something.

Getting supervisors to actually respond

Response rate is the whole game, and account creation is where it dies. A site supervisor is doing you a favour in the middle of their own job: if the request requires a password, a verification email and a portal, a large share of them stop, and the ones who stop are not random.

Rating should be a single-use link that opens straight into the questions. Ours does, it requires no account, and it takes a few minutes. Reminders should be limited and honest, and declining should be one click, because a supervisor who cannot decline simply ignores every future request too.

What to do about inflated ratings

Supervisors rate interns generously. This is well known, it is not dishonesty, and it does not make the data useless. Three things keep it usable.

• Frequency wording rather than quality wording, as above.
• Report the spread, not just the mean. Where several observers rated the

same student, the disagreement between them is often more informative than the average of them.

• Read gaps, not absolutes. Inflation is roughly consistent across a cohort,

so the difference between self-rating and observer rating still tells you something even when both numbers sit high.

A caveat worth stating plainly: with a small number of raters per student, no statistical treatment makes the individual result precise. Use it to start a conversation with a student, and use the cohort aggregate for program decisions.

Doing this with the Career Readiness Report

Create an assessment window for the term, import the students on placement, and each student invites the supervisor who actually watched their work. The report that comes back carries the self ratings, the observer ratings, the gap and a written development note, and the cohort view rolls it up for accreditation evidence. It is free for every institution, and concurrent placements are supported, so a student with two internships gets one report per placement rather than a blurred average.

Not ready to adopt anything? The rating rubric behind all of this is published in full, free to copy into whatever survey tool you already use.