Why do small cohorts mislead?

If you're trying to compare TCM schools by CALE (California Acupuncture licensing exam) pass rate, it's a fair worry — a single percentage can look like a clean, objective number, but it can be almost meaningless if you don't also know how many people it's built from.

Why do small cohorts mislead?

If you're trying to compare TCM schools by CALE (California Acupuncture licensing exam) pass rate, it's a fair worry — a single percentage can look like a clean, objective number, but it can be almost meaningless if you don't also know how many people it's built from.

Here's the arithmetic underneath it: when a class has only 2–3 test-takers, one person retaking the exam swings the percentage by 20–50 points. When a class has 30–40 test-takers, that same one retake moves the number by only a couple of points. Same event, wildly different effect on the headline figure — purely because of sample size, not program quality.

Here's what that looks like with real published numbers, per the school's own multi-year CALE analysis (2021–2025, California Acupuncture Board data, as checked 08/2026):

SchoolCumulative N (2021–2025)Cumulative pass rateWhat the swings look like
Five Branches University16985.8%Every year 26–42 takers — never single-digit
University of East-West Medicine7485.1%16–23/year, but two 100% years next to two 68–78% years
Academy of Chinese Culture & Health Sciences8883%Shrinking 26→10, swinging 69%–94%
California University–Silicon Valley1573.3%~2/year average, one 7-person year swung it to 42.9%

The lesson isn't "small school = bad school" — it's that a rate from a handful of people isn't yet a reliable signal of program quality, while a rate built on dozens of test-takers every single year is much harder to move by chance. That's why leading with the N (test-taker count), not just the percentage, is the honest way to read these numbers.

Are you comparing a couple of specific schools, or trying to get the general lay of the land first?

Sources

Last reviewed:

Continue this conversation

Omni picks up right where this answer left off — edit the text to ask something else.