Interobserver Agreement
Also called: IOA, reliability, interobserver reliability
Definition
The degree to which two independent observers recording the same events report the same result, usually expressed as a percentage. It is evidence that a measure reflects the behavior rather than the observer, and it is collected on a planned sample of sessions rather than all of them.
In plain language
Two people watch the same session and record it separately, then compare. If they mostly agree, the data is probably about the learner and not about who was watching.
Examples
Two staff independently counted requests during the same session and agreed on nineteen of twenty intervals.
A supervisor records alongside a technician once a month to check agreement.
Not interobserver agreement
Two staff filling in one shared sheet together. They are not independent, so this shows nothing about agreement.
One person recording the same session twice from memory, which measures that person's consistency rather than agreement.
Commonly confused with
See also
Written from
- baer-wolf-risley-1968 — Analytic and technological dimensions
Written from the measurement standards described in the founding dimensions paper. No textbook glossary consulted.
This entry has not yet been through clinical review. Treat it as a draft.