Skip to main content

Free IOA (Interobserver Agreement) Calculator

All seven common IOA formulas with instant results — total count, count-per-interval, trial-by-trial, scored and unscored interval, and duration/latency. No signup, and your data never leaves your browser.

IOA Calculator

Pick the type of data you collected, enter both observers' records, and the agreement percentages update instantly.

What is interobserver agreement?

Interobserver agreement (IOA) is the degree to which two independent observers, watching the same session, record the same behavior the same way. It is the most common index of measurement quality in applied behavior analysis: if two trained observers cannot agree on what happened, the data cannot be trusted to guide treatment decisions.

IOA matters in practice, not just research. Behavior-analytic supervision, school FBA/BIP teams, and insurance or district audits all lean on the credibility of the underlying data. Collecting IOA on a portion of sessions — at least 20%, preferably 25–33%, per Cooper, Heron, and Heward — and reporting the agreement percentage alongside your graphs is what turns “we took data” into defensible measurement.

Which IOA formula should you use?

Match the formula to your measurement system, and prefer the more stringent index when the data will support high-stakes decisions.

Total Count IOA

(smaller total ÷ larger total) × 100

Quick check for event/frequency data when you only have session totals. Least stringent — two observers can record entirely different events and still "agree."

Mean Count-per-Interval IOA

mean of each interval’s (smaller ÷ larger) × 100

Event data recorded in intervals. More sensitive than total count because agreement is evaluated interval by interval.

Exact Count-per-Interval IOA

(intervals with identical counts ÷ total intervals) × 100

The most stringent index for event data — observers must match exactly within every interval.

Trial-by-Trial / Interval-by-Interval IOA

(agreements ÷ total trials or intervals) × 100

Discrete trial data and occurrence/nonoccurrence interval recording. Can be inflated by chance at very high or very low response rates.

Scored-Interval (Occurrence) IOA

agreements ÷ intervals where either observer scored an occurrence

Low-rate behavior. Ignores the easy nonoccurrence agreements that inflate interval-by-interval IOA.

Unscored-Interval (Nonoccurrence) IOA

agreements ÷ intervals where either observer scored a nonoccurrence

High-rate behavior. Ignores the easy occurrence agreements instead.

Mean Duration / Latency IOA

mean of each pair’s (shorter time ÷ longer time) × 100

Duration-per-occurrence or latency data timed independently by two observers.

IOA frequently asked questions

What is an acceptable IOA percentage?

Cooper, Heron, and Heward (2020) give the common guideline: at least 90% agreement for an established measure and at least 80% for a new one. For research, the What Works Clearinghouse single-case design standards (Handbook Version 5.0) set the minimum at 80% agreement, or a kappa of .60. What is truly acceptable depends on the complexity of the behavior definition and the measurement system — a stringent index like exact count-per-interval at 82% can mean better measurement than total count IOA at 95%.

How often should IOA be collected?

Cooper, Heron, and Heward (2020) recommend IOA for at least 20% of sessions, preferably 25% to 33%. The What Works Clearinghouse research standard is at least 20% of data points in each phase. Spread those sessions across conditions, observers, and times of day; some funders and journals set their own minimums.

Which IOA formula should I use?

Match the formula to your measurement system. Session totals: total count IOA. Event data in intervals: mean or exact count-per-interval. Interval or discrete-trial recording: interval-by-interval, plus scored-interval for low-rate behavior or unscored-interval for high-rate behavior. Timed data: mean duration- or latency-per-occurrence IOA.

What should I do if my IOA is low?

Low IOA usually points to an unclear operational definition or drift in how observers apply it. Tighten the definition with examples and non-examples, retrain and recalibrate observers together, then re-check agreement before trusting new data for decisions.

Is this IOA calculator free?

Yes — it is completely free, requires no signup, and all calculations run in your browser. Nothing you enter is stored or sent to a server.

Stop calculating IOA by hand

Classroom Pulse collects second-observer data alongside your primary data and calculates IOA automatically, with reports ready for supervision and team review.