ABA percent change from baseline equals current value minus baseline value, divided by the baseline value, multiplied by 100. Define how each value was summarized and state the therapeutic direction. Report the absolute change beside the relative change. A zero baseline makes the ratio undefined, and a near-zero baseline can create an unstable, dramatic percentage.

Define baseline and current values first

Percent change compares two declared summaries, not two unlabeled points. Define the target, measurement dimension, unit, baseline phase, current phase, observation dates, exposure, inclusion rules, and how each phase value was calculated. State whether the summary is a mean, median, rate, percentage, duration, or another reviewed measure. Use the same definition and compatible denominator in both phases.

Yara's worksheet compares rates per hour. The baseline summary is 4 events per hour and the current summary is 3 events per hour. The worksheet links each phase value to its underlying observations and keeps the sample size, missing records, exclusions, and phase dates visible. A missing observation does not become zero, and a partial observation needs its reviewed exposure before it contributes to a rate.

Verify source data and denominator compatibility

Reconcile raw counts, observation time, opportunities, phase labels, duplicates, corrections, and changed definitions before calculating. A baseline count from two hours cannot be compared directly with a current count from 30 minutes. Convert both to an appropriate common rate only when the measurement system supports that calculation.

Look for shortened sessions, unavailable AAC, different observers, new settings, altered opportunity definitions, and implementation changes. These factors may affect interpretation even when the arithmetic remains correct. Preserve the original phase summaries and the rule that produced them.

Calculate Yara's percent change

Use the signed formula:

Percent change = ((current value minus baseline value) / baseline value) × 100

For Yara:

  1. Current minus baseline = 3 minus 4 = negative 1 event per hour.
  2. Divide by baseline = negative 1 / 4 = negative 0.25.
  3. Multiply by 100 = negative 25%.

The signed result is a 25% decrease from baseline. Because the stated therapeutic direction is reduction, the report can say that the observed rate decreased by 25%. It should also report the absolute change: a decrease of 1 event per hour, from 4 to 3 per hour. Keep the negative sign in the calculation even if the narrative uses the word decrease.

Handle zero, near-zero, and negative baselines

When baseline equals zero, division by baseline is undefined. Do not report an infinite or 100% increase. Report the current value, absolute change, and a different predeclared comparison if clinically appropriate. A baseline near zero can produce a very large percentage from a small original-unit change, so show both magnitudes and interpret cautiously.

Negative baseline values can occur on some centered or standardized scales, but ordinary count, rate, duration, and percentage measures are usually nonnegative. Relative change around a negative value is often hard to interpret and may not answer the clinical question. Seek quantitative review and use the original-unit difference when the ratio lacks a clear meaning.

Address direction, rounding, and missingness

A negative signed result indicates that the current numerical value is lower. Its therapeutic meaning depends on the target. Lower latency or a lower rate of a harmful event may fit the goal, while lower independent communication may be concerning. State the intended direction before interpreting the sign.

Calculate with unrounded phase summaries and round the final percentage once under a declared rule. Report expected and observed records for both phases. If missingness, unequal exposure, or a source correction materially changes either summary, recalculate and preserve the earlier result with its reason.

Show the change on a graph

Graph the raw time-ordered data with phase lines, observation units, relevant condition changes, and ordinary supports. Add baseline and current summaries as clearly labeled annotations or a companion table. Avoid replacing the raw series with a single arrow labeled 25%, since that removes variability, trend, overlap, and sample size.

Lobo and colleagues and Wolfe, Barton, and Meadan provide single-case and visual-analysis context. NIST's moving-average guidance describes smoothing, which can change the appearance of recent shifts and should be labeled when used.

Use report language that preserves both magnitudes

Yara's observed rate decreased from 4 events per hour in the declared baseline summary to 3 per hour in the current summary. The absolute change was negative 1 event per hour, and the relative change from baseline was negative 25%, reported as a 25% decrease for this reduction target.

Add phase dates, sample sizes, missingness, calculation method, rounding, setting, exposure, treatment-integrity evidence, and client feedback. Avoid saying that Yara "improved by 25%" because the calculation describes one measure and does not establish overall improvement or cause.

Review meaning and access with the client

Discuss the measure with Yara through accessible communication and with the responsible clinician. Maintain AAC, interpreters, mobility supports, food, water, bathroom access, prescribed care, rest, relationships, and emergency help. ASHA supports continuous AAC access.

Ask whether the target and direction matter to Yara, whether context or burden changed, and whether the numerical difference is useful in everyday life. A qualified clinician integrates this summary with the graph, measurement quality, implementation evidence, risks, benefits, and the person's perspective.

Keep the source set in scope

The BACB ethics hub and CASP public summary provide professional and clinical context. The BCBA Test Content Outline lists measurement and data-based decisions as examination content. NIST's distribution page, outlier guidance, and box-plot page support broader descriptive review. None sets a universal ABA percent-change threshold.

Clinician review checklist

For an ABA percent change from baseline review, confirm the question, target definition, therapeutic direction, phase dates, compatible units, exposure, underlying records, missingness, baseline summary, current summary, absolute difference, nonzero baseline, signed formula, arithmetic, rounding, graph, context, access, client perspective, implementation evidence, qualified decision owner, and review date. Recalculate when a phase value or definition changes.

Limits of percent change

Percent change does not establish measurement validity, clinical importance, treatment integrity, client benefit, statistical uncertainty, generalization, future performance, or cause. It is undefined at zero and unstable near zero. The same relative change can represent very different original-unit differences, and the same absolute change can yield different percentages from different baselines. Report the raw values and absolute change beside the percentage, and never use the ratio alone as a treatment rule.

Related resources

Sources