The big idea
Every measure answers the same question: how close is the thing being recorded to the behavior you care about?
- Direct measurement records the behavior of interest itself, usually as it happens.
- Indirect measurement records something else, such as an interview answer or a rating, and you infer the behavior from it.
- Permanent-product measurement records the behavior after it has occurred, through a lasting change it left in the environment.
Behavior analysts prefer direct measures because they require the fewest inferences. Indirect measures are useful early in assessment, but when they stand in for direct measures they threaten the validity of the data (C.8). Permanent products can be excellent data, provided the product truly reflects the behavior.
Breaking it down
Direct measurement
An observer watches or listens and records the target behavior: counting greetings, timing a tantrum, scoring each trial. What is measured and what you want to know about are the same thing.
Direct measurement can capture any dimension (count, duration, latency) and supports the strongest conclusions. Its costs are practical. Someone has to be there when the behavior happens, which is hard for behavior that is rare, private or spread across the day. Being watched can also change the behavior, an effect called reactivity. Unobtrusive observation, and a few sessions that let people get used to the observer, reduce it.
Indirect measurement
Indirect measures include interviews, questionnaires, rating scales, checklists and most self-reports. The data are someone’s verbal report about the behavior, filtered through memory, attention, bias and that person’s own idea of what the behavior is. A parent’s rating of how often a child melts down depends on what that parent counts as a meltdown and how the week felt.
Indirect measures are quick and can reach behavior no observer could see. That is why they are common early in assessment (F.5). When possible, check them against a sample of direct observation before basing decisions on them.
Permanent-product measurement
A permanent product is a lasting outcome of behavior: completed math problems, assembled parts, a made bed, litter picked up, words in a written story. Video and audio recordings are contrived products that preserve the behavior itself so it can be scored later.
Products have real advantages. The practitioner need not be present when the behavior occurs, products can be scored carefully and rescored by a second person for agreement checks, and they work well for behavior that happens at inconvenient times.
Two rules decide whether a product is a valid measure:
- Every occurrence leaves the product. If a student sometimes solves problems in their head without writing anything, the worksheet undercounts.
- Only the target behavior can produce the product. If a sibling sometimes does the homework, or several behaviors affect the outcome, the product overstates or blurs the behavior.
A product also hides how the behavior happened. A finished worksheet does not show whether the learner was prompted, copied, or took an hour. If those details matter, add occasional direct observation.
Seeing it in practice
| Situation | Type of measure | Note |
|---|---|---|
| A BCBA watches recess and records each time Chloe asks a peer to play. | Direct | Records the behavior itself as it happens. |
| A job coach counts the boxes Mrs. Oyelaran folded by counting the stack at the end of the shift. | Permanent product | Valid if only Mrs. Oyelaran folds boxes at that station. |
| Parents fill in a weekly scale rating how often Jun “melts down.” | Indirect | A report about the behavior, shaped by memory and judgment. |
| A teacher records reading group on video and later counts Daniela’s answers. | Permanent product (contrived) | The recording preserves the behavior for direct scoring later. |
| A district tracks fighting by counting office discipline referrals. | Indirect | Each referral is a product of a staff member’s behavior, not of the fight. |
The last row is a trap. A referral is a written record, but it exists only when an adult notices a fight and decides to write it up. It fails the second rule, because staff behavior produces it, so it is better treated as an indirect measure of fighting.
Common mix-ups
- “Indirect” does not mean “not live.” Scoring a product later is not indirect if the product reliably reflects the target behavior. Indirect means the thing scored is a report or a stand-in.
- Not every record is a product of the target behavior. Incident reports, referral slips and notes written from memory are products of the writer’s behavior.
- Outcomes with many causes are weak products. Body weight changes with eating, exercise, illness and medication, so it is a poor measure of any one of them.
- Rating scales do not yield counts. “Often” and “sometimes” mean different things to different raters.
- Direct measurement has a cost too. Reactivity can make observed behavior unrepresentative.
Exam tip: Ask two questions. Is the thing being scored the behavior itself, or someone’s account of it? If it is the behavior, was it scored as it happened (direct) or afterward from what it left behind (permanent product)?
Quick recap
- Direct measures record the behavior of interest itself; indirect measures record a report or stand-in and require inference.
- Permanent products measure behavior after the fact through its lasting effects, and video recordings count as contrived products.
- A product is valid only if every occurrence leaves it and nothing else produces it.
- Indirect measures are useful early in assessment but threaten validity when they replace direct observation.
- Reactivity is the main cost of direct observation.
Further reading
- Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied behavior analysis (3rd ed.), chapters 4 and 5.
Key terms
- Direct measurement
- Measurement in which the behavior recorded is the same behavior the practitioner wants to know about, usually observed and recorded as it occurs. It requires the fewest inferences and is preferred in behavior analysis.
- Indirect measurement
- Measurement in which what is recorded differs from the behavior of interest, such as an interview answer, rating scale or questionnaire about the behavior, so conclusions about the behavior require an inference. It is generally less valid than direct measurement.
- Permanent product (Permanent-product measurement, Outcome recording)
- A lasting change in the environment produced by a behavior, which can be measured after the behavior has occurred. A product is a valid measure only when every occurrence of the behavior leaves it and nothing other than the target behavior can produce it. Video and audio recordings are contrived permanent products.
- Reactivity
- A change in behavior caused by the act of observing or measuring it. It applies to clients who behave differently while watched, and to observers whose recording changes when they know their data will be checked (observer reactivity).
Check yourself
4 questions on C.2. Choose an answer to see whether you're right and why.
Show the answer
Answer: D. The worksheet is a lasting outcome of Femi's writing, scored after the behavior occurred, so this is permanent-product measurement. It is not indirect, because the thing scored is what Femi produced, not someone's report about it. It is not direct observation, because nobody watched him write.
Show the answer
Answer: D. A product is a valid measure only if the target behavior alone produces it. Body weight also changes with exercise, meals, illness and medication, so a change in weight cannot be credited to snacking. Weight is easy to graph, and it is not a direct measure of snacking.
Show the answer
Answer: D. The data are the parent's after-the-fact judgment about sleep, not a record of the sleep itself, so the measure is indirect. Writing a rating down does not make it a permanent product of the target behavior; it is a product of the parent's rating behavior. How often a measure is collected does not decide whether it is direct.
Show the answer
Answer: A. A referral is produced by a staff member's behavior, so fights that go unseen or unreported never appear, and reporting habits can change while fighting does not. That makes referrals an indirect measure of fighting. True permanent products of the target behavior are often very good data, not worse than ratings.
Adds a step to your study record. Saved in this browser only.