OSCE Standard-Setting Calculator
Paste each candidate's station totals and examiner global grades, and get defensible pass marks by all four standard methods — custom, Angoff, borderline group and borderline regression — with the regression fit shown, not hidden. Multi-station exams, must-pass rules and pass-rate impact included. Everything is computed in your browser.
Load candidate scores
First row: headers. First column: candidate. Then two columns per station: checklist total, global grade (the numeric value of the rating scale).
Scores are computed locally — open your browser's network tab to verify nothing is uploaded.
Choose the standard-setting method
Which method to use is an institutional policy decision — this calculator computes, it doesn't choose. AMEE Guide 85 (McKinley & Norcini) covers how to pick.
Exam rules
Stations and scores
| Candidate name or ID | total | grade | total | grade | |
|---|---|---|---|---|---|
Borderline methods need each station's borderline grade — set it in the station header above.
Results — pass marks and their impact
OSCE standard-setting report
Add stations and candidate scores above to compute pass marks.
| Station | Pass mark | n | Mean | SD | Min–max | Passed | Failed | R² | Note |
|---|---|---|---|---|---|---|---|---|---|
| Station 1 | — | 0 | — | — | — | — | — | — | Borderline grade value not configured for this station |
| Station 2 | — | 0 | — | — | — | — | — | — | Borderline grade value not configured for this station |
Export the report
The full report is free too — leave an email and it unlocks. Only your email and the tool's name are sent, never any scores.
- Print-ready PDF: pass marks, statistics and the regression charts — the page for the exam board
- CSV of the computed station results (opens in Excel)
- Every method's guards and rounding documented, so the marks are defensible
This is the after-the-exam spreadsheet ritual. StudyDrome runs all four standard-setting methods on your real OSCEs automatically — examiner marking on tablets feeds straight into borderline regression, per-station quality metrics and the pass/fail list, no pasting.
See StudyDrome's OSCE engineExactly how each mark is computed
Published so you can check (or disagree with) every number. All rounding is half-even (banker's), matching the product engine; statistics use the sample standard deviation.
- Custom — each station's mark is the one you enter; no mark means the station contributes nothing until set.
- Angoff — station marks are the entered expert judgments; the exam mark is their sum, or their average rounded to 2 decimals (McKinley & Norcini, AMEE 85).
- Borderline group — the mark is the mean (or median) checklist total of candidates whose global grade equals the station's borderline grade. Even-sized groups take the average of the two middle values. No borderline candidates → no mark (Wood 2006).
- Borderline regression — ordinary least squares of checklist total (Y) on global grade (X) over candidates with both values. It needs at least 3 such candidates and variance in the grades. Slope, intercept and R² are rounded to 4 decimals first; the mark is intercept + slope × borderline grade, clamped between 0 and the station's max points, rounded to 2 decimals (Wood 2006; Pell 2010 for R²).
- A locked station uses its preset mark instead of computing one; a withdrawn station keeps its statistics but contributes no mark.
- Exam pass mark = the sum of the non-withdrawn station marks (Angoff optionally averages). You can override it manually — the report says so when you do.
- Pass/fail — a candidate's exam total is the sum of ALL their entered station totals (withdrawn stations still count toward the total, mirroring the product), compared against the exam mark; then the minimum-stations rule and must-pass stations are applied in that order.
- Candidates with no checklist total on a station are excluded from that station's statistics, group and regression; pass rate = passed ÷ all candidates, rounded to 2 decimals.
This is a calculator, not a policy: which method (and which borderline grade) your institution uses is a governance decision documented in AMEE Guide 85. The maths here mirrors the StudyDrome product engine and is pinned by the product's own test fixtures — but a defensible mark still needs enough candidates, sane grading and expert oversight.
Sources
Frequently asked questions
Are candidate scores uploaded anywhere?
No. Every mark is computed in your browser. The only network calls this page makes are anonymous aggregate counters (which tool, which action) and — only if you use it — the export form, which sends your email address and nothing else. Verify it in your browser's network tab.
Which standard-setting method should I use?
That's an institutional policy decision, not a calculator's. AMEE Guide 85 (McKinley & Norcini 2014) is the standard reference; Wood 2006 compares the two borderline methods directly. Borderline regression uses every candidate's data rather than only the borderline group's, which is why many OSCE units prefer it — when the grade–score relationship is strong. This tool computes all four so you can compare them on your own data.
Why is a station's pass mark blank?
The station note says exactly why: no borderline grade configured, fewer than 3 candidates with both a total and a grade (regression), no variance in the grades, no candidates at the borderline grade (group), or no preset mark entered (custom/Angoff). Blank means "couldn't be computed defensibly", never zero.
What does the R² number mean?
R² measures how much of the checklist-score variation the global grades explain. Pell 2010 (AMEE Guide 49) treats a correlation above 0.5 as a reasonable relationship, so below that the grades and the scores agree poorly and a regression-derived mark rests on a weak link. The chart shows the fit rather than hiding it — that's the point.
Is this really free?
Yes. Computing pass marks needs nothing at all; the print/CSV export asks only for an email address. The tool exists so you can see the standard-setting engine StudyDrome runs on real OSCEs automatically — same methods, same maths, pinned by the same test fixtures.