The five ARC/STSA outcomes thresholds, and the math that trips programs up

August 16, 2026 · 7 minute read · For surgical technology program directors

Five numbers decide whether a surgical technology program is in good standing: retention, exam participation, exam pass rate, graduate placement, and survey results.

Most programs know the thresholds. What causes trouble is the arithmetic underneath them, because the same roster can produce two different percentages depending on how you add it up, and only one of those is the number being read.

The thresholds

Participation is the one programs underweight. It is 100 percent, and it is scored separately from the pass rate. A graduate who never sits for the exam does not quietly drop out of the calculation. The nonparticipation is itself the problem.

The four places the math goes wrong

1. Averaging percentages across cohorts

This is the big one. When you roll two or more cohorts into an academic year summary, you sum the numerators and sum the denominators. You do not average the two cohort percentages.

Take a program with two cohorts in one academic year. One retains 5 of 6. The other retains 10 of 14. Averaged, that reads 83 percent and 71 percent, so 77 percent. Summed correctly, it is 15 of 20, so 75 percent.

Two points does not sound like much until the threshold sits between them. Averaging inflates small cohorts, and small cohorts are exactly where the variation lives. Sum the counts, then divide once, every time.

2. Counting a graduate more than once for placement

Placement counts each graduate one time. A graduate working two jobs, or employed and also continuing their education, is still one graduate.

Rosters built from employment records rather than from the graduate list tend to double count, and the error always runs in the flattering direction. That is what makes it easy to miss.

3. Treating survey satisfaction as an average score

Satisfaction is not the mean of the ratings. A returned survey counts as satisfactory when at least 7 of the 10 items are rated 3 or higher. It is a pass or fail per survey, then a percentage of the surveys that passed.

Averaging the ratings instead produces a number that looks close most of the time and diverges exactly when one respondent rates everything low. Grade each survey, then count the passes.

One live question worth settling for your own program: how a "does not apply" response is treated inside the 7 of 10. The conservative reading is that it does not count toward the seven. Ask ARC/STSA and get the answer in writing before it decides a reporting year for you.

4. Losing track of re-entries and delayed graduates

A student who withdraws and re-enters with a later cohort belongs to a specific denominator, and it is not always the one you would guess. Delayed completions, re-entries, and transfers are where a roster quietly stops matching the report.

The Outcomes Tracking Tool has a color language for exactly this reason. Use it, and record the decision at the moment you make it rather than reconstructing it a year later from memory.

Three consecutive years is the signal that matters

A single year below a threshold is a data point. Three consecutive years below the same threshold is what triggers review.

Which means the number to watch is not this year's percentage. It is the streak. A program sitting at 58 percent retention for the second year running is in a materially different position than one that dipped once, and that difference is invisible if you only look at the current reporting year.

Put your last three years side by side, per threshold, at least once a year. If you have two in a row anywhere, that is the thing to fix, and you have roughly one reporting cycle to do it in.

Do the reconciliation before you need it

The point of keeping outcomes current is not the report. It is that the report becomes a printout instead of a project.

A program that updates the tool as students move, term by term, spends an afternoon on the annual report. A program that reconstructs the year in June spends three weeks on it, finds the survey return rate too late to chase it, and answers site visit questions from memory.

Same math. Completely different job.