Exam confidence isn't a fixed trait. It cracks at exam time.

The finding

Exam confidence, a Confidence & Self-Efficacy measure, averages 6.7 out of 10 over the last 12 months across roughly 13,900 responses. That single number looks stable. The monthly pattern underneath it is not.

Confidence moves with the assessment calendar. It sits near 7.4 at term start in late summer and early autumn, eases to 7.2 in October, 6.8 in November, and 6.5 in December as winter exams arrive. It recovers to 7.0 in January and peaks near 7.7 in February. Then it falls to 5.4 in April, the spring exam window and the year’s sharpest dip, measured on roughly 2,400 responses that month. It recovers to 6.9 in May and 7.2 in June.

So confidence swings by roughly two points across the year, and the low points line up exactly with exam windows: December, and most sharply, April. This is not a stable personality trait. It is seasonal and predictable.

What separates the calm from the panicked

Reading the comments, two things divide the confident students from the anxious ones: preparation and rehearsal, and access to a teacher in the run-up. The lowest scores cluster around unfamiliar formats (a first big written project, an oral exam), exams bunched too close together, and the prospect of performing in front of teachers and peers. The highest scores credit having practised in class and being able to reach a teacher for guidance.

Lowest-scoring students 2.5

"I just don’t like presenting, because what if I haven’t done it well enough. Before every exam or test so far, I’ve struggled to breathe and find calm."

"I’m very nervous about the final project. I’ve never written a big assignment like this before, so I could really use guidance on what it should contain and how to build it."

"Because the exams come so close together, I prepared well for the first two but naturally have less time for the last two, which are also the hardest."

"I already feel stressed because there are so few days left, and it’s in front of the teacher and half the class."

Highest-scoring students 9.3

"What’s given me calm about the coming exam is that we’ve practised a lot in class and had good support from the teachers. I feel better prepared through the experience I’ve gained."

"I have a clear problem statement and a clear plan, and the teachers helped me get there. They’re always ready to guide when there’s a problem."

"There’s good flexibility to revisit tests and assignments; it’s easy to arrange with the teachers if you feel you’re falling behind."

Representative, de-identified comments from each segment. Translated from the original where needed.

Why the predictability matters

The single most useful fact here is that the dip is foreseeable. Confidence does not erode at random. It collapses on the exam calendar and recovers afterwards. December and April are not surprises; they are scheduled. That changes what an institution can do about it. A trait you cannot move, you can only screen for. A seasonal pattern you can plan against. Because the low points are known months in advance, support can be timed to land before them, not after.

What it means for institutions

1. Treat exam confidence as seasonal and time support to the calendar. The April collapse is foreseeable, so put guidance, rehearsal and check-ins before the spring window, not after it.
2. Demystify unfamiliar formats. The sharpest fear is the first big written or oral assessment, so show its structure and expectations explicitly and early.
3. Keep teachers reachable in the run-up. The calm students credit rehearsal and access, not innate confidence, and both are things the institution controls.

How we measure it

Exam confidence is a Confidence & Self-Efficacy measure, scored 0 to 10 and paired with an optional free-text comment. The figures here cover roughly 13,900 responses over the last 12 months, with monthly means computed across all participating institutions. The April reading rests on roughly 2,400 responses. Quotes are de-identified and, where the original was not in English, translated. They are chosen to represent the lowest- and highest-scoring segments, not cherry-picked outliers.