USING IB RESULTS DATA WITHOUT MISLEADING TEACHERS OR STUDENTS
Results data can reveal patterns worth investigating, but it rarely explains them alone. A cohort average is not a diagnosis, a global pass rate is not a school target, and a historical boundary is not a forecast. Use multiple evidence layers and state uncertainty plainly.
Keep evidence levels separate
Global bulletins describe large assessment sessions. School results describe a local cohort with its own subject mix and history. Student work provides the detail needed for intervention. Moving directly from a global pass-rate change to a classroom prescription skips the evidence that connects those levels.
Build a data dictionary before a dashboard. Define whether ‘students’ means all students in a session overview, diploma candidates, course candidates or another group. Define whether ‘pass rate’ applies to DP candidates and whether mean points uses the same population across the comparison. If the source changes definition, break the time series or annotate it.
| Level | Can show | Cannot show alone | Next evidence |
|---|---|---|---|
| Global bulletin | Session scale and aggregate trends | Why one school or student changed | School and subject context |
| School cohort | Local outcomes and recurring patterns | Whether teaching caused the pattern | Task, attendance and prior-attainment evidence |
| Subject/component | Where outcomes are concentrated | The mechanism behind lost marks | Marked responses and moderation notes |
| Student work | Specific knowledge and execution evidence | Whether a pattern generalizes without repeated tasks | New work under varied conditions |
Use a question-first analysis cycle
Begin with a decision the department could actually change. ‘Why did results fall?’ is too broad. ‘Did students lose more marks on evaluation tasks than on recall tasks across three common assessments?’ is testable. Decide the evidence and comparison before looking for a dramatic chart.
When a pattern appears, generate multiple explanations. A fall may relate to cohort differences, task design, attendance, curriculum sequencing, examination execution or random variation. Seek evidence that could disconfirm the preferred explanation. This protects the meeting from turning one plausible story into a fact.
1. Define
Write the metric, population, session and decision it will inform.
2. Validate
Check missing data, changed definitions, small groups and transcription errors.
3. Compare
Use an appropriate baseline and display denominators beside percentages.
4. Explain cautiously
List competing hypotheses and the evidence each predicts.
5. Intervene and test
Make one bounded change and collect new work that could show an effect.
A safe department dashboard specification
Data principle. The closer a decision is to one learner, the more the evidence must include that learner’s actual work and context—not only group averages.
- Show counts beside rates and suppress or aggregate very small groups where privacy is at risk.
- Keep May and November sessions distinct unless the analytical question justifies combination.
- Annotate syllabus, assessment and data-definition changes on trend charts.
- Separate subject-grade distributions from diploma-award outcomes.
- Do not rank teachers using unadjusted class outcomes.
- Link every global figure to the exact official bulletin and session.
- Give viewers a plain-language caveat and the date data was refreshed.
- Record the decision taken so the dashboard is evaluated by usefulness, not decoration.
Continue reading
References
Editorial review
Reviewed against the cited official IB and admissions sources. IBretake does not claim examiner, IB coordinator or alumni credentials unless a named credential is shown.
See something outdated? Email support@ibretake.com so the editorial team can check and correct it.
Ready to check real school availability?
Ready to check real school availability?
Start a retake search