How final scores are calculated
When you mark in indicators mode, you give an overall mark and record indicator values. The app then derives per-criterion scores that sum to your overall mark, weighted by the indicators.
This page explains how that works.
The problem
Section titled “The problem”You’re in a live conversation. You form a holistic impression of the student’s performance. You notice specific things — strong evidence use, weak articulation, reading from a script.
But the student (or your institution) wants per-criterion scores: “Analysis: 18/25, Communication: 12/15” and so on.
IOA Marker bridges this gap. You capture what your brain actually produces (overall impression + observations); the app computes the breakdown.
Influences
Section titled “Influences”Each criterion declares which indicators affect it and how strongly. This is called an influence.
An influence has two parts:
- Indicator — which indicator affects this criterion
- Weight — how strongly (1 to 5)
For example, a “Communication” criterion might be influenced by:
- “Speaks clearly” (weight 3)
- “Makes eye contact” (weight 2)
- “Confident delivery” (weight 1)
You configure influences in the setup page using the matrix view. Click a cell to set the weight; click again to clear it.
Indicator values
Section titled “Indicator values”All indicators use a 5-point scale:
| Label | Meaning |
|---|---|
| + + | Well above expectations |
| + | Above expectations |
| ○ | Meets expectations |
| − | Below expectations |
| − − | Well below expectations |
Indicators start blank during marking. Blank values are treated as neutral for live calculations.
The derive algorithm
Section titled “The derive algorithm”Here’s what the app does when deriving per-criterion scores:
-
Calculate achievement: For each criterion, measure where the indicator values fall within the full possible range (0 to 1, where 0.5 is neutral)
-
Apply contribution: Scale the achievement by the indicator contribution setting — this controls how much indicators affect the distribution
-
Calculate adjusted scores: Multiply each criterion’s max points by its scaled achievement
-
Normalize: Scale all scores so they sum to the overall mark you gave
-
Cap and redistribute: Ensure no criterion exceeds its maximum; redistribute any excess
-
Round: Final scores round to the nearest 0.5 points
The result: per-criterion scores that reflect your indicator observations, constrained to sum exactly to your overall mark.
Indicator contribution
Section titled “Indicator contribution”The indicator contribution setting (0–100%) controls how much your indicator observations affect the distribution of scores across criteria.
| Contribution | What it means |
|---|---|
| 0% | Indicators have no effect — scores are distributed purely by criterion weights |
| 50% (default) | Balanced influence — indicators can shift scores moderately |
| 100% | Maximum effect — indicators strongly differentiate between criteria |
At 50% contribution with all indicators neutral, each criterion gets 75% of its maximum points before normalization. At 100% contribution, neutral indicators give 50% of max.
You can tune this in the setup page under the Marking tab.
Example
Section titled “Example”Setup:
- Overall mark: 75
- Criterion A: max 60 points
- Criterion B: max 40 points
- Indicator contribution: 50%
- Indicator 1 influences A (weight 2) — set to ”+”
- Indicator 2 influences B (weight 1) — set to ”−”
The result: because Indicator 1 is positive and influences A, criterion A gets a larger share of the 75 points. Because Indicator 2 is negative and influences B, criterion B gets a smaller share.
If both indicators were neutral (○), the 75 points would be split proportionally: A would get 45 (60% of 75) and B would get 30 (40% of 75).
With the indicators as set, the final scores might be something like A: 51/60, B: 24/40 — the exact values depend on the full calculation, but the principle is: positive indicators boost their linked criteria, negative indicators reduce them, all while maintaining the overall total.
Criteria score type
Section titled “Criteria score type”You can display criterion scores as:
- Percent — max points must sum to 100, displayed with %
- Points — max points can sum to anything, displayed as raw numbers
This is a display choice; the underlying math is the same.
Invariants
Section titled “Invariants”The algorithm guarantees:
- Scores always sum to the overall mark
- No criterion exceeds its maximum
- Same inputs always produce same outputs (deterministic)
- Neutral indicators give proportional distribution
This means every mark is reproducible and explainable — important for appeals and moderation.