Whole-interval measurement bias arises because a brief disqualifying interruption makes the interval negative even when responding occupied most of it. Longer intervals create more opportunity for interruption. State the full-interval and gap rule, interval length, and coverage. Compare with continuous duration when the decision needs the actual amount of time engaged.

Match the method to a sustained-response question

Whole-interval recording asks whether the response met its definition for the entire interval. It is suitable when uninterrupted occurrence is the feature the clinician truly needs to sample. It is a poor substitute for total duration when the decision concerns how much of the observation was occupied by the response. State the question first, then specify the response, start and stop rules, interval length, observation period, and what disqualifies an interval.

Kai's worksheet defines five consecutive one-minute intervals within a five-minute observation. Any five-second gap makes that interval negative under the chosen full-interval rule. The method was declared before collection; it was not selected after seeing the data.

Build the raw interval record

Keep one row per planned interval with start and end times, positive or negative result, valid or invalid state, reason, observed response duration when available, observer, setting, and procedure version. Separate planned, observed, valid, positive, negative, missed, invalid, and outside-exposure intervals. A technical failure or safety pause does not become a negative client score.

Preserve the continuous timing record when it was collected. Whole-interval scores remove information about where and how long a gap occurred. That lost detail matters when clinicians later review measurement fit or compare the sampled percentage with total duration.

Calculate both summaries for Kai

Kai meets the response definition for 55 seconds in each one-minute interval. Each interval also contains a five-second interruption. All five intervals are valid observations.

For whole-interval recording:

  1. Positive intervals = 0, because no interval contains the response for all 60 seconds.
  2. Valid intervals = 5.
  3. Whole-interval percentage = 0 / 5 × 100 = 0%.

For continuous duration:

  1. Total response duration = 55 seconds × 5 intervals = 275 seconds.
  2. Observed duration = 60 seconds × 5 intervals = 300 seconds.
  3. Duration percentage = 275 / 300 × 100 = 91.7%, rounded to one decimal place.

Both results are arithmetically correct. They answer different questions.

Explain the apparent contradiction

The 0% whole-interval result means that none of the five intervals met the uninterrupted full-interval rule. It does not mean Kai was unengaged for the whole observation. The duration record shows engagement during 275 of 300 observed seconds. A brief gap carries the same negative whole-interval score as a gap lasting most of the interval.

Interval length changes the opportunity for a disqualifying interruption, and the response pattern changes how often gaps cross interval boundaries. Pritchard and colleagues and Wirth and colleagues examine accuracy and decision implications of discontinuous procedures. There is no universal correction factor that turns Kai's 0% into exact duration.

Check missingness and edge cases

Report valid coverage as valid intervals divided by planned intervals. If one interval had been unobservable, the whole-interval denominator would follow the prewritten missing-state rule, while coverage would show four valid of five planned. Never place an invalid interval in the negative numerator by default.

If there are no valid intervals, the sampled percentage is undefined. If every valid interval is positive, the result is 100% under the rule, yet short gaps at interval boundaries or unobserved periods may still matter. Repeated interruptions tied to one activity, observer, access problem, or health need can make the sample unrepresentative and should prompt review rather than blame.

Graph and report the method honestly

Graph interval scores in time order when the pattern of positives and negatives matters, and show continuous duration alongside them when available. Label interval length, rule, planned exposure, valid coverage, and missing states. Avoid presenting a whole-interval percentage as exact percent of time.

Across five valid one-minute intervals, Kai met the uninterrupted whole-interval rule in 0 of 5 intervals, or 0%. Continuous timing showed 275 of 300 observed seconds, or 91.7%, meeting the response definition. The difference reflects the full-interval rule and the five-second gap in every interval.

Lobo and colleagues provide broader single-case measurement and analysis context. Retain the raw time series, access conditions, implementation evidence, and client feedback with the summary.

Review meaning and access with Kai

Use accessible communication to ask Kai whether the response and uninterrupted-duration question are meaningful. Maintain AAC, interpreters, mobility supports, food, water, bathroom use, prescribed care, rest, relationships, and emergency help throughout observation and review. ASHA supports ongoing access to AAC tools or devices.

Do not treat use of AAC, a needed movement, a health action, or a request for help as an interruption to suppress merely to improve the score. The qualified clinician should examine whether the measurement rule fits the person's goal, access, burden, risk, and everyday context.

Keep professional sources in scope

The BACB ethics hub and CASP public summary provide professional and clinical context. The BCBA Test Content Outline covers continuous and discontinuous measurement as examination content. These materials support qualified method selection; they do not prescribe a universal interval length or clinical threshold.

Clinician review checklist

Confirm the decision question, response definition, onset and offset, interval length, full-interval rule, planned exposure, valid coverage, positive and negative counts, invalid and missed states, raw timing, arithmetic, rounding, graph label, access conditions, client perspective, implementation evidence, qualified decision owner, and review date. Reopen the method when the response pattern, interval, setting, access, observation system, or decision changes.

Limits of whole-interval measurement

Whole-interval percentage is a procedure-specific sampled result. It does not equal exact duration, establish measurement validity, prove treatment integrity, show client benefit, predict future performance, or identify cause. Brief interruptions can sharply lower the score, especially with longer intervals. Use continuous duration when exact occupied time matters, retain coverage and raw evidence, and avoid using the percentage alone to judge treatment or the person.

Related resources

Sources