behavior-measurement

name	behavior-measurement
description	Use when defining target behaviors operationally, selecting appropriate measurement procedures, and designing data collection systems for behavior analytic assessment and intervention.

Behavior Measurement

Accurate measurement is the foundation of all behavior analytic practice. Without reliable data, clinical decisions are guesses. This skill covers operational definitions, dimensions of behavior, measurement systems, and observer training.

Operational Definitions

Every target behavior must be defined so that two independent observers would agree on whether the behavior occurred.

Requirements

• Observable: Describes what the behavior looks like, not inferred internal states.
• Measurable: Can be counted, timed, or otherwise quantified.
• Clear boundaries: Specifies when an instance begins and ends (onset and offset).
• Examples and non-examples: Include 2–3 examples of what counts and 2–3 examples of what does not.

Writing Process

1. Observe the behavior across multiple occasions and contexts.
2. Draft the definition based on the topography (what it looks like).
3. Test the definition: have two people independently score the behavior using only the written definition.
4. Refine until both scorers agree on at least 80% of instances.

Example

Target behavior: Aggression

• Definition: Any instance of forceful contact with another person's body using hands (hitting, slapping, grabbing), feet (kicking), head (headbutting), or teeth (biting) with sufficient force to produce an audible sound, visible mark, or pain response.
• Examples: Punching a peer's arm, kicking a staff member's shin, biting a teacher's hand.
• Non-examples: Accidentally bumping into someone while walking, giving a high-five, tapping someone to get their attention.

Dimensions of Behavior

Select the dimension that best captures the clinically relevant aspect of the behavior.

Dimension	Definition	When to Use	Unit
Frequency/Count	Number of times a behavior occurs	Discrete behaviors with clear onset/offset; consistent observation periods	Count
Rate	Count per unit time	When observation periods vary in length	Count/minute or count/hour
Duration	Total time engaged in the behavior	Behaviors where how long matters (tantrums, on-task)	Seconds, minutes
Latency	Time from a stimulus to behavior onset	Compliance, response initiation	Seconds
Inter-response time (IRT)	Time between consecutive responses	Spacing of behavior instances	Seconds, minutes
Magnitude/Intensity	Force or severity of the behavior	When topography is the same but intensity varies	Scale (1–5) or force measure
Topography	Physical form of the behavior	When you need to describe what the behavior looks like	Descriptive
Locus	Where on the body or in the environment	SIB targeting specific body parts	Location descriptor

Selecting Measurement Based on Behavior Characteristics

Decision Guide

1. Is the behavior discrete (clear start and stop)? → Frequency or rate.
2. Does the behavior vary primarily in how long it lasts? → Duration.
3. Is the concern about how quickly the client responds? → Latency.
4. Does observation time vary across sessions? → Convert to rate rather than raw frequency.
5. Is the behavior happening at high rates (hard to count individual instances)? → Use time sampling or duration.
6. Does the behavior leave a lasting product? → Permanent product recording.

Continuous vs. Discontinuous Measurement

Continuous Measurement

Observer records every instance of the behavior throughout the entire observation period.

• Event recording: Count each instance. Best for low-to-moderate rate discrete behaviors.
• Duration recording: Time each episode from onset to offset. Best for extended behaviors.
• Latency recording: Time from discriminative stimulus to response onset.

Discontinuous Measurement (Time Sampling)

Observer samples behavior at specified intervals rather than recording every instance.

• Whole-interval recording: Score "yes" only if the behavior occurs throughout the entire interval. Underestimates behavior. Best for behaviors you want to increase (on-task, engagement).
• Partial-interval recording: Score "yes" if the behavior occurs at any point during the interval. Overestimates behavior. Best for behaviors you want to decrease.
• Momentary time sampling (MTS): Score "yes" if the behavior is occurring at the exact moment the interval ends. Most accurate estimate of duration-based behaviors. Practical for classroom or group settings.

Interval Length Selection

• Shorter intervals (e.g., 10s) yield more accurate estimates but are harder to implement.
• Longer intervals (e.g., 1min) are more practical but less precise.
• Match interval length to the typical duration of behavior episodes.

Permanent Product Recording

Measure the outcome or product left by the behavior rather than observing the behavior itself.

• Completed worksheets (academic responses).
• Broken items (property destruction).
• Marks on skin (SIB).
• Written work samples (handwriting legibility).

Advantages: Does not require continuous observation; can be measured after the fact. Limitation: Does not capture the process or context of the behavior.

Data Collection Systems

Paper-Based

• Frequency tally sheets with defined observation periods.
• ABC (Antecedent-Behavior-Consequence) data sheets for descriptive assessment.
• Interval recording sheets with pre-printed interval grids.
• Scatterplot forms for identifying temporal patterns.

Electronic

• Tablet-based data collection apps (e.g., Catalyst, CentralReach, Hi Rasmus).
• Advantages: automatic graphing, timestamped data, reduced transcription errors.
• Ensure data security and HIPAA compliance for electronic systems.
• Train all users on the specific platform before data collection begins.

Training Observers

Steps

1. Provide the operational definition with examples and non-examples.
2. Review the data collection system and practice with the materials.
3. Score practice videos or role-plays together, discussing disagreements.
4. Score practice videos independently and calculate IOA.
5. Achieve criterion (typically ≥80% IOA across 3 consecutive practice sessions) before collecting data for clinical or research purposes.
6. Conduct periodic reliability checks (booster sessions) throughout the study or treatment.

Common Observer Errors

• Observer drift: Gradual change in how the observer applies the definition over time. Prevent with periodic retraining.
• Observer reactivity: Changing scoring behavior when aware of being assessed. Minimize by making reliability checks unpredictable.
• Expectancy bias: Scoring influenced by knowledge of the condition or expected outcome. Use blinded observers when possible.

Key References

• Cooper, J. O., Heron, T. E., & Heward, W. L. (2020). Applied Behavior Analysis (3rd ed.). Pearson.
• Johnston, J. M., & Pennypacker, H. S. (2009). Strategies and Tactics of Behavioral Research (3rd ed.). Routledge.
• Kazdin, A. E. (2010). Single-Case Research Designs (2nd ed.). Oxford University Press.