Measurement & recording Glossary

Define the event and opportunity

Count is the number of response occurrences. Event recording records each defined occurrence during an observation period. Count becomes rate only when divided by an appropriate time denominator.

Define when a response begins and ends, how repeated responses separate, what counts as an eligible opportunity, and how interruptions or missing observation are handled. A high count can reflect a longer observation period, more opportunities, a setting change, or observer differences.

The BACB BCBA Test Content Outline includes measurement, time sampling, data display, experimental design, assessment, and evaluation as examination content. ABAI's basic-principles page lists measurement and behavioral concepts for higher-education training. Neither source supplies a case-specific method.

Measure time and intensity

Latency is elapsed time from a defined antecedent or opportunity to response onset. Interresponse time is elapsed time between defined responses. State the start event, end event, unit, timeout rule, and treatment of responses that never occur.

Response magnitude measures a defined dimension of response intensity, such as force, volume, distance, or another calibrated quantity. Avoid subjective labels like “severe” unless they map to observable criteria and meaningful safety or clinical decisions.

Trials to criterion counts learning opportunities required to meet a predeclared performance rule. Define the trial, criterion, maintenance expectation, prompts, errors, resets, and maximum exposure. Fewer trials can indicate efficiency, while comfort, generalization, durability, and client preference still matter.

Choose continuous or sampled measurement

Continuous measurement attempts to record every occurrence of the selected response dimension during observation. Discontinuous measurement samples behavior across intervals or moments.

Partial-interval recording marks an interval if the response occurs at any point. Whole-interval recording marks an interval only if the response occurs throughout. Momentary time sampling observes at a defined instant, usually at interval end. A planned activity check samples the behavior or engagement of several people at scheduled moments.

Peer-reviewed comparisons of time-sampling methods and measurement accuracy show that interval length, response pattern, observation schedule, and method can bias estimates. Use the sampled unit directly when possible rather than relabeling it as exact duration.

Use products when they preserve the event

Permanent product recording measures a lasting, observable result of behavior. Examples can include a completed form, prepared item, saved message, or cleaned workspace. Confirm that the product can be attributed to the target response and that quality criteria are defined.

Product measures can reduce live-observation burden. They may hide assistance, retries, time, distress, or who performed the work. Combine them with direct observation or contextual fields when those facts matter.

Report percentages with their counts

A percentage is a ratio multiplied by 100. State the numerator, denominator, eligibility rule, time window, and exclusions. “80% independent” means little without the number and type of opportunities, prompt definition, ordinary supports, and who observed.

Research on measurement and observer effects reinforces the need to evaluate how recording procedures affect data. Train observers on the same examples, sample agreement across relevant conditions, and report observer agreement separately from performance.

Find and explain artifacts

A measurement artifact is a pattern created or distorted by the measurement system rather than the underlying behavior. Device latency, app defaults, rounded timestamps, overlapping intervals, missed opportunities, observer drift, duplicate taps, and phase-boundary changes can create artifacts.

Audit raw records before interpreting a surprising trend. Preserve the original data, correction history, device or form version, observer, schedule, and known outage. A clean graph can still display a flawed input.

Write a measurement plan before collection begins. Include the purpose, operational definition, inclusion and exclusion rules, observation schedule, response unit, opportunity unit, timer or device, observer role, training examples, agreement sample, privacy limits, and decision rule. Pilot the method in the real setting. A measure that consumes the session, interrupts communication, or changes partner behavior may be technically precise while poorly fitted to routine care.

Review feasibility with the person being observed. Explain what will be recorded, who will see it, and how they can ask questions or pause when applicable. When observation itself changes comfort or participation, record that context rather than treating the resulting data as ordinary performance.

Retire measures that no longer answer a useful question.

A recording example

Lena is a fictional twelve-year-old who wants adults to answer her AAC break message reliably. Across nine eligible activity periods, Lena sends seven independent messages. Adults respond within ten seconds to five of seven messages.

Report Lena's message use as 7/9, 77.8% and partner response as 5/7, 71.4%. If the device is unavailable during one period, retain that system-access failure rather than scoring it as a child skill opportunity. The two percentages describe different responsibilities and establish no causal effect.

Explore clinical roles at Finni practices. Confirm the practice's measurement, calibration, data-integrity, client-access, supervision, and decision-review systems during diligence.