---
title: "Free OSCE calculators: standard setting, item analysis and checklist building"
description: "Three free OSCE calculators that run in your browser: what each one answers, the data it needs, and the questions it cannot settle for you."
canonical: https://studydrome.com/docs/osce/toolkit/free-calculators/
updated: 2026-10-10
---

# Free OSCE calculators: standard setting, item analysis and checklist building

Three free OSCE calculators that run in your browser: what each one answers, the data it needs, and the questions it cannot settle for you.

Three calculators cover the arithmetic a small OSCE needs: building a station checklist, computing the pass mark, and reading the station statistics afterward. All three run in your browser, need no account, and send nothing anywhere. They are free because the arithmetic is not the hard part of an OSCE.

This page says what each one answers and where each one stops. The stopping points matter more than the features. A calculator computes. Choosing the method, judging a flagged item and defending a result stay with your committee.

## The three calculators

| Calculator | What it answers | The data it needs | Where it stops |
| --- | --- | --- | --- |
| [Checklist builder](/tools/osce-checklist-builder/) | Is this station checklist well formed, and can I export it? | A station: sections, items, a scale, any critical item | Structural checks only. It cannot judge whether the station measures the right clinical skill. |
| [Standard-setting calculator](/tools/osce-standard-setting/) | What is the pass mark under each method, and how do they compare? | Station totals per candidate, plus global grades for the borderline methods | It computes, it does not choose. The method is an institutional decision. |
| [Item-analysis calculator](/tools/osce-item-analysis/) | Which items and stations behaved strangely at this sitting? | The examiner-scored item matrix for one station | Screening statistics only. They find odd behavior, not the wrong construct. |

## The checklist builder

It walks a station in four steps: how you want to start, the station and its scale, the checklist itself, then a report and an export. You can start from scratch, load a sample station, paste rows from a spreadsheet, or upload a file you already have. The validator runs two tiers of rules: format rules, which would stop a file importing, and quality screening drawn from the OSCE literature, each with its source named.

The export is a spreadsheet in the same import format as the [OSCE station template](/resources/templates/osce-station/). The two are siblings: one interactive, one a static file to edit in Excel.

Two rules inside it are worth knowing before you build. At most one item can be flagged as the global rating, because that is the item borderline standard setting reads. And the station total is the sum of each item's top score, excluding the global rating, which feeds the standard rather than the score.

## The standard-setting calculator

It runs four methods on the same data: custom pass marks, Angoff, borderline group and borderline regression. Seeing four results side by side is the point. A school that has never compared them usually discovers the gap is smaller than feared, which makes the choice easier to defend.

Borderline regression needs at least three candidates carrying both a total and a global grade, and it needs variation in those grades. If every candidate is graded the same, there is no line to fit. You can also set an exam-level pass mark by hand and require a minimum number of stations passed, which is the conjunctive rule described in [conjunctive rules](/docs/osce/scoring/conjunctive-rules/).

The tool prints the policy note itself, and it is the right note: which method to use is an institutional decision. [Choosing and defending a pass standard](/docs/osce/scoring/choosing-and-defending-a-pass-standard/) is the page that helps you make it.

## The item-analysis calculator

It takes the examiner-scored matrix for one station and returns the polytomous item analysis: facility, discrimination, item-rest correlation, Cronbach's alpha and the standard error of measurement, with every flag explained and cited.

Three input rules decide what you get. A blank cell means not assessed and leaves that item's count. A zero is a real score and stays in. And item maximums come from a Max row when you supply one. Without it, the tool assumes each item's highest observed score and tells you it has done so, which matters because facility is computed against the maximum.

Its own caution is the one to repeat to a committee. Small cohorts make every statistic noisy, so a flag from a single sitting is a prompt to look, not a verdict. [Reading station metrics](/docs/osce/quality/reading-station-metrics/) covers how to act on one.

## When a calculator stops being enough

Three signs. You are retyping marks from paper into a spreadsheet for more than an hour. You need the statistics for twelve stations rather than one. Or you need a record of who changed a score, because a result was challenged. At that point the work is data handling, not arithmetic, and the [software requirements checklist](/docs/osce/toolkit/osce-software-requirements-checklist/) is the next page to read.

## Frequently asked questions

### Is our candidate data uploaded anywhere?

No. All three calculators compute in your browser. The tools say so on the page and invite you to check the network tab yourself. Nothing is stored and nothing is sent, which is also why there is no account and no saved history.

### Can the item-analysis calculator handle a whole exam?

It is built for one station's matrix at a time. For a twelve-station exam you would run it twelve times and combine the results by hand. That is the point where a platform that computes station and exam reliability together starts to pay for itself.

### Which standard-setting method should we pick?

The calculator will not answer that and neither will this page in one line. Borderline regression is the most widely used for OSCEs where global grades are recorded. The decision belongs in your regulations before the exam runs. See [choosing and defending a pass standard](/docs/osce/scoring/choosing-and-defending-a-pass-standard/).

### Do the checklist builder's warnings mean our station is wrong?

No. They are screening heuristics from the literature plus the format rules of the import file. A station can pass every check and still test the wrong thing, and it can trip a warning for a good reason. Subject experts settle it.

### Can we use these for a written exam?

These three are built for OSCE data. The item-analysis tool expects partial-credit scores rather than right and wrong answers. There are separate calculators on the [tools page](/tools/) for written-exam item analysis.

> [!NOTE]
> **Book a pilot**
> If you have outgrown the calculators, the same arithmetic runs inside the platform on live exam data, with the exports your board reads. Start at /register/.