Data Collection Methods in ABA

Summary: This page covers the various data collection methods used in ABA, including continuous methods (event recording, duration recording, latency recording, IRT) and discontinuous methods (momentary time sampling, partial interval recording, whole interval recording). You’ll learn when to use each method, their advantages and limitations, and how to select the appropriate measurement system for different behaviors.

Data collection is a cornerstone of Applied Behavior Analysis (ABA). Accurate and consistent data collection allows behavior analysts to make informed decisions about interventions, track progress, and demonstrate outcomes. This guide covers the various data collection methods used in ABA practice.

Continuous Measurement Methods

Continuous measurement involves recording every instance of a behavior during an observation period.

Event Recording

Event recording involves counting each occurrence of a behavior during an observation period.

Key Characteristics:

• Definition: Tallying each instance of a behavior
• Best for: Behaviors with clear beginning and end that occur at moderate rates
• Data reported as: Total count or rate (count ÷ time)
• Example behaviors: Hand raising, aggression, manding (requesting)

Implementation Steps:

1. Define the target behavior with clear start/end criteria
2. Select a recording method (counter, tally sheet, app)
3. Record each occurrence during the observation period
4. Calculate total count and/or rate

Advantages and Limitations:

• Advantages: Simple to implement, provides precise count
• Limitations: Difficult for high-rate behaviors, doesn’t capture duration or intensity

Duration Recording

Duration recording measures how long a behavior lasts from start to finish.

Key Characteristics:

• Definition: Timing the length of a behavior
• Best for: Behaviors where length is clinically significant
• Data reported as: Total duration, average duration, or percentage of time
• Example behaviors: On-task behavior, tantrums, engagement

Implementation Steps:

1. Define the target behavior with clear start/end criteria
2. Start timing when the behavior begins
3. Stop timing when the behavior ends
4. Record the duration in consistent units (seconds, minutes)

Advantages and Limitations:

• Advantages: Captures persistence of behavior, useful for endurance goals
• Limitations: Requires continuous observation, challenging for multiple behaviors

Latency Recording

Latency recording measures the time between a stimulus (e.g., instruction) and the initiation of a response.

Key Characteristics:

• Definition: Timing from stimulus to response start
• Best for: Measuring response speed or compliance
• Data reported as: Time (seconds, minutes)
• Example behaviors: Following instructions, transitioning between activities

Implementation Steps:

1. Define the stimulus and response with clear criteria
2. Start timing immediately after delivering the stimulus
3. Stop timing when the response begins
4. Record the latency in consistent units (typically seconds)

Advantages and Limitations:

• Advantages: Useful for compliance goals, captures hesitation
• Limitations: Only measures time to start, not completion

Interresponse Time (IRT)

IRT recording measures the time between consecutive instances of a behavior.

Key Characteristics:

• Definition: Timing between the end of one response and start of the next
• Best for: Analyzing response patterns and behavioral rhythms
• Data reported as: Time (seconds, minutes)
• Example behaviors: Self-injury, stereotypy, requesting

Implementation Steps:

1. Define the target behavior with clear start/end criteria
2. Start timing when one instance ends
3. Stop timing when the next instance begins
4. Record each IRT in consistent units

Advantages and Limitations:

• Advantages: Reveals temporal patterns, useful for understanding behavior chains
• Limitations: Requires continuous observation, complex to implement

Discontinuous Measurement Methods

Discontinuous measurement involves sampling behavior during portions of an observation period rather than continuously.

Momentary Time Sampling (MTS)

MTS involves recording whether a behavior is occurring at the exact moment of predetermined time intervals.

Key Characteristics:

• Definition: Observing behavior at precise moments
• Best for: Behaviors that last for extended periods
• Data reported as: Percentage of intervals with behavior
• Example behaviors: On-task behavior, engagement, posture

Implementation Steps:

1. Define the target behavior
2. Determine interval length (e.g., 1 minute)
3. At each interval point, record if behavior is occurring (yes/no)
4. Calculate percentage of intervals with behavior

Advantages and Limitations:

• Advantages: Less labor-intensive, good estimate for high-duration behaviors
• Limitations: May miss behaviors between observation points

Partial Interval Recording (PIR)

PIR involves recording whether a behavior occurred at any point during each interval.

Key Characteristics:

• Definition: Noting if behavior occurred during interval
• Best for: Behaviors that occur at low to moderate rates
• Data reported as: Percentage of intervals with behavior
• Example behaviors: Disruptive behavior, social interactions

Implementation Steps:

1. Define the target behavior
2. Determine interval length (e.g., 30 seconds)
3. Record if behavior occurred at any point during each interval
4. Calculate percentage of intervals with behavior

Advantages and Limitations:

• Advantages: Captures low-rate behaviors, less intensive than continuous
• Limitations: Tends to overestimate actual occurrence, doesn’t capture frequency within intervals

Whole Interval Recording (WIR)

WIR involves recording whether a behavior occurred throughout the entire interval.

Key Characteristics:

• Definition: Noting if behavior occurred for the entire interval
• Best for: Behaviors where duration is important
• Data reported as: Percentage of intervals with behavior
• Example behaviors: Sustained attention, on-task behavior

Implementation Steps:

1. Define the target behavior
2. Determine interval length (e.g., 30 seconds)
3. Record if behavior occurred continuously throughout each interval
4. Calculate percentage of intervals with behavior

Advantages and Limitations:

• Advantages: Good for measuring sustained behaviors
• Limitations: Tends to underestimate occurrence, misses brief instances

Interval Recording Comparison

Aspect	Momentary Time Sampling	Partial Interval	Whole Interval
When to record	At exact moment of interval	Any occurrence during interval	Behavior throughout entire interval
Tends to	Accurately estimate duration	Overestimate occurrence	Underestimate occurrence
Best for	High-duration behaviors	Low-rate behaviors	Sustained behaviors
Example	On-task at end of each minute	Any disruption during interval	Engaged for entire interval

Permanent Product Recording

Permanent product recording involves measuring the physical results or outcomes of behavior rather than the behavior itself.

Key Characteristics:

• Definition: Measuring tangible evidence left by behavior
• Best for: Behaviors that produce measurable outcomes
• Data reported as: Count, percentage, or quality measure
• Example behaviors: Academic work, cleaning, construction tasks

Types of Permanent Products:

1. Completed work: Worksheets, assignments, projects
2. Created items: Art, crafts, written work
3. Environmental changes: Cleaned areas, organized materials
4. Digital records: Computer files, online activities
5. Physical evidence: Property damage, self-injury marks

Implementation Steps:

1. Define the target behavior and its expected product
2. Establish measurement criteria (e.g., accuracy, quantity)
3. Collect and evaluate the permanent product
4. Record data according to predetermined metrics

Advantages and Limitations:

• Advantages: Can be measured after the fact, doesn’t require direct observation
• Limitations: Only applicable to behaviors that leave evidence, may miss process information

ABC Data Collection

ABC (Antecedent-Behavior-Consequence) data collection involves recording the events that occur before and after a behavior to identify functional relationships.

Key Characteristics:

• Definition: Documenting what happens before, during, and after behavior
• Best for: Functional behavior assessment, understanding behavior triggers
• Data reported as: Narrative or structured format
• Example use: Identifying patterns in challenging behavior

Implementation Steps:

1. Create a recording system with columns for A, B, and C
2. Record antecedents (what happened before the behavior)
3. Describe the behavior in objective terms
4. Document consequences (what happened after the behavior)
5. Note date, time, setting, and other relevant factors

Advantages and Limitations:

• Advantages: Provides contextual information, helps identify function
• Limitations: Labor-intensive, requires narrative recording, subjective elements

Scatterplot Analysis

Scatterplot analysis involves recording behavior occurrence across time periods to identify temporal patterns.

Key Characteristics:

• Definition: Mapping behavior occurrence across time
• Best for: Identifying temporal patterns in behavior
• Data reported as: Visual display of occurrence by time
• Example use: Determining if behavior is more likely at certain times

Implementation Steps:

1. Create a grid with time periods on one axis and days on the other
2. Define a coding system (e.g., no occurrence, mild, severe)
3. Record behavior occurrence in each time block
4. Analyze the completed scatterplot for patterns

Advantages and Limitations:

• Advantages: Reveals temporal patterns, helps with intervention timing
• Limitations: Doesn’t capture specific antecedents or consequences

Task Analysis Data Collection

Task analysis data collection involves breaking a complex skill into steps and recording performance on each step.

Key Characteristics:

• Definition: Measuring completion of sequential steps
• Best for: Complex skills with multiple components
• Data reported as: Percentage of steps completed correctly
• Example behaviors: Self-help skills, academic procedures, social routines

Implementation Steps:

1. Break the skill into discrete, sequential steps
2. Create a data sheet listing all steps
3. For each step, record whether it was performed correctly
4. Calculate percentage of steps completed correctly

Types of Task Analysis Data:

1. Total task presentation: Presenting all steps in each session
2. Forward chaining: Adding steps from beginning to end
3. Backward chaining: Adding steps from end to beginning
4. Most-to-least prompting: Fading assistance across trials
5. Least-to-most prompting: Increasing assistance as needed

Advantages and Limitations:

• Advantages: Detailed skill assessment, clear progress monitoring
• Limitations: Time-consuming to develop and implement

Selecting Appropriate Measurement Systems

Choosing the right measurement system depends on several factors:

Behavior Characteristics

• Frequency: How often does the behavior occur?
• Duration: How long does the behavior last?
• Topography: What does the behavior look like?
• Intensity: How forceful or severe is the behavior?
• Complexity: Is it a simple or multi-step behavior?

Practical Considerations

• Observer resources: How much time and attention is available?
• Setting constraints: What’s feasible in the environment?
• Multiple behaviors: How many behaviors need tracking?
• Precision needed: How exact must the measurement be?
• Purpose of data: What decisions will be made with the data?

Decision-Making Guide

If the behavior…	Consider using…
Has clear beginning/end and moderate rate	Event recording
Varies in how long it lasts	Duration recording
Involves response to instruction	Latency recording
Shows patterns in timing between occurrences	IRT recording
Lasts for extended periods	Momentary time sampling
Occurs briefly but is important to catch	Partial interval recording
Should be sustained over time	Whole interval recording
Produces measurable outcomes	Permanent product recording
Has unclear function	ABC recording
Shows temporal patterns	Scatterplot analysis
Involves multiple steps	Task analysis recording

Data Collection Tools

Paper-Based Tools

• Data sheets: Customized forms for specific measurement systems
• Counters: Mechanical devices for tallying occurrences
• Timers: Devices for measuring duration, latency, or intervals
• Checklists: Lists of steps or behaviors to mark
• Grids: Visual formats for interval or scatterplot recording

Electronic Tools

• Mobile apps: Specialized behavior tracking applications
• Tablets: Devices that can run data collection software
• Wearable technology: Devices that can automatically record data
• Video recording: For later coding and analysis
• Electronic data systems: Comprehensive platforms for data collection and analysis

Ensuring Data Quality

Operational Definitions

• Clear: Unambiguous description of the behavior
• Objective: Based on observable characteristics
• Complete: Includes examples and non-examples
• Measurable: Can be consistently identified
• Agreed upon: Understood by all observers

Observer Training

• Initial instruction: Teaching measurement procedures
• Practice opportunities: Rehearsing with feedback
• Calibration: Comparing to expert measurement
• Ongoing support: Addressing questions and concerns
• Refresher training: Preventing observer drift

Interobserver Agreement (IOA)

• Definition: The degree to which independent observers agree
• Calculation methods:
  • Total agreement: (Smaller count ÷ Larger count) × 100
  • Exact agreement: (Agreements ÷ Total observations) × 100
  • Interval-by-interval: (Agreement intervals ÷ Total intervals) × 100
• Acceptable levels: Typically 80% or higher
• Frequency: Regular checks throughout data collection
• Addressing discrepancies: Retraining and clarifying definitions

Practice Examples

Example 1: Classroom Disruption

Scenario: A teacher wants to measure a student’s disruptive behavior during 30-minute lessons.

Measurement decision process:

• Behavior characteristics: Brief instances, moderate frequency, multiple topographies
• Practical considerations: Teacher can’t continuously observe while teaching
• Selected method: Partial interval recording with 1-minute intervals
• Implementation: Teacher sets timer to vibrate every minute and notes if disruption occurred
• Data reporting: “Disruption occurred during 40% of intervals, down from 65% baseline”

Example 2: Hand Washing Skill

Scenario: A therapist is teaching a child with autism to wash hands independently.

Measurement decision process:

• Behavior characteristics: Complex skill with multiple steps
• Practical considerations: Need to track progress on specific components
• Selected method: Task analysis with total task presentation
• Implementation: Therapist creates checklist with 12 steps, records performance on each
• Data reporting: “Client independently completed 75% of hand washing steps, up from 25% baseline”

Example 3: Tantrum Behavior

Scenario: A behavior analyst is assessing a child’s tantrums to develop an intervention.

Measurement decision process:

• Behavior characteristics: Variable duration, unclear triggers
• Practical considerations: Need to understand function and patterns
• Selected methods: Duration recording, ABC data, and scatterplot
• Implementation: Staff record tantrum length, antecedents/consequences, and time of day
• Data reporting: “Tantrums average 12 minutes, typically occur before transitions, and are most frequent between 10-11am”

Key Points to Remember

• Select measurement methods based on behavior characteristics and practical considerations
• Continuous measurement provides the most precise data but requires more resources
• Discontinuous measurement offers efficiency but may sacrifice some accuracy
• Clear operational definitions are essential for reliable measurement
• Regular IOA checks help ensure data quality
• Different measurement methods can be combined for comprehensive assessment
• The goal of measurement is to gather meaningful data for decision-making
• Technology can enhance efficiency and accuracy of data collection
• Consistency in measurement procedures is critical for valid comparison over time
• Data collection methods should be reviewed and adjusted as needed based on results