Calculation result is unexpected

Requires: Access to the source answers and calculation; write access to change the questionnaire.

Check:

1. You are inspecting the intended submission. Test calculation uses the latest accessible answer, not necessarily your own.
2. The questionnaire was saved before testing.
3. Each operator selects the intended numeric, text or answer-item references.
4. Answer positions match the values you intend to score. A visible choice label is not necessarily its numeric position.
5. The operator mode is correct: Per item differs from aggregation. In per-item SUBTRACT, the result is Value minus answer.
6. A later operator is not counting original inputs and derived results together. Most operators retain previous values.
7. Missing/nonnumeric values have not reduced the input set. An empty aggregate AVERAGE can return zero; zero is not proof that every question was answered.
8. Earlier pipelines precede dependent ones, and external source questionnaires have complete answers.
9. The headline result is the final intended entry. End a scoring pipeline with an explicit aggregation if earlier entries were overwritten.
10. Historical answers were recalculated only if intended. Saving a new configuration does not update them all automatically.

Avoid zero divisors and inspect intermediate results before changing several operators at once.

Still not working?

Prepare one synthetic submission, its expected result, the pipeline/operator order and actual intermediate outputs. Include which references were selected.

Operator reference · Examples.