Data Collection Methods in ABA

Data Collection Systems

Operational Definitions & Data Collection

• Operational definitions should already be in place.
• Focus is now on different data collection systems and how RBTs take data daily.

Types of Data Collection

• Count/Frequency: Tallying the number of occurrences of a behavior.
• Rate: Count/frequency of a behavior over a specified time period.
• Partial Interval Recording: Discontinuous measurement system where you record if the behavior occurs at any point during the interval.
• Momentary Time Sampling: Discontinuous measurement system where you record if the behavior is occurring at the exact moment the interval ends.
• Whole Interval Recording: Discontinuous measurement system where you record if the behavior occurs for the entire duration of the interval.
• Trial by Trial: Data taken after every trial, commonly used in discrete trial teaching (DTT).
• First Trial Data: Data only taken on the first response during DTT.
• Task Analysis Data: Data taken on each step of a task analysis.
• Latency Data: Measurement of time between the onset of a stimulus and a response.
• Duration: How long a behavior occurs.
• Permanent Product: Measuring the outcome/effect of a behavior on the environment, rather than directly observing the behavior itself.
• Planned Activity Check: Similar to momentary time sampling, but typically used in group settings.
• Estimation Data: Subjective data collection used within the Autism Partnership method, often in conjunction with other data collection systems.

Continuous Measurement: Rate, Count/Frequency, and Duration

Count and Frequency

• Simple tally of the number of times a behavior occurs.
  • Pick the behavior requested by BCBA and tally each occurrence.
    *Advantages:
  • Accurate for infrequent behaviors.
  • Provides a precise number of occurrences.
    *Disadvantages:
  • Not easy to take for high rates of behavior.
  • Can be time-consuming.
  • Does not specify a time period, making comparisons difficult.

Rate

• Combines count/frequency with observation time.
  • Count occurrences of the behavior.
  • Specify the observation period.
    *When to Use:
  • Behaviors with a clear start and finish (e.g., hand raising, aggression).
  • Best for low to mid rates of behavior.
    *When Not to Use:
  • Continuous behaviors (e.g., tantrum, crying).
  • High-frequency behaviors.
    *Advantages:
  • Provides an exact count over time.
  • Shows clear behavior change when compared to baseline, intervention, and maintenance phases.
    *Disadvantages:
  • Time-consuming.
  • Hard to get reliability if operational definitions are unclear.
  • May interfere with other RBT responsibilities, such as running programs or implementing behavior intervention plans.

Example: Jamie Throwing Toys

• Operational Definition: Jamie grabbing a toy, bringing his arm down, bringing it up, and releasing.
• Data Collection:
  • Frequency Count: 29
  • Observation time: 3 minutes 30 seconds (3.5 minutes)
  • Rate Calculation: \frac{29}{3.5} = 8.3 throws per minute

Duration

• Amount of time a behavior occurs.
  • From the onset to the end of a response.
    *When to Use:
  • Behaviors that occur for an extended period or at very high rates.
  • Stereotypy (physical or vocal).
  • Best with clear onset and offset.
    *Recording Methods:
  • Duration per occurrence: Time each instance of the behavior separately.
  • Total duration: Total time the behavior occurs throughout the session.
    *Advantages:
  • Provides total time per session.
  • Provides total time per occurrence.
    *Disadvantages:
  • Time-consuming to juggle data collection with other RBT tasks.
  • Not always accurate due to divided attention.

Latency

• Measure of time between the onset of a stimulus and a response.
  • Example: Time between a traffic light turning green and a driver stepping on the gas.
    *Advantages:
  • Useful to measure time elapsed between an instruction and a child’s response.
    *Disadvantages:
  • Does not measure the behavior itself.
  • Can be time-consuming.

Discontinuous Measurement Systems

• Used to make data collection easier for RBTs.
  • Do not require constant monitoring.

Partial Interval Recording

• Divide an observation period into smaller intervals (e.g., 5-30 seconds).
• Score if the behavior occurs at any time during the interval. *Advantages:
  • Easier to implement than frequency count.
    *Disadvantages:
  • Tends to overestimate behavior.
  • Underestimates high-frequency behavior.
  • Requires attention to the time passing and the intervals.

Momentary Time Sampling

• Divide an observation period into smaller intervals (e.g., 5-30 seconds).
• Score the target behavior if it occurs at the moment the interval ends. *Advantages:
  • Easier to implement than continuous data collection.
  • Interferes less with teaching.
    *Disadvantages:
  • Can both overestimate and underestimate behavior, but generally considered the most accurate discontinuous measurement system.
  • May miss the big picture of a behavior.
  • Requires attention to the time passing in the intervals.

Whole Interval Recording

• Divide an observation period into smaller intervals (e.g., 5-30 seconds).
• Score the interval only if the behavior occurs throughout the entire interval. *When to Use:
  • Best for continuous and high rates of behavior.
    *Advantages:
  • Discontinuous.
    *Disadvantages:
  • Underestimates behavior.
  • Time-consuming – requires watching the individual the entire time.
  • Interferes with teaching.

Task Analysis Data

• Best for behaviors broken down into smaller steps.
  • Social behaviors, daily living skills.
    *Requires a task analysis – skill breakdown into smaller parts.
    *Data Collection:
  • Data collected on each step (yes/no, prompted/independent).
  • Provides information on individual steps.
    *Advantages:
  • Data can be taken on each step.
  • Helps focus teaching on specific areas.
  • Helps teach the operational definition of the behavior.
    *Disadvantages:
  • Measures only one aspect of behavior at a time.
  • Relies on an accurate task analysis.
  • Can be time-consuming, especially when physically prompting through the steps.

Task Analysis Example: Conversation

*Joining into conversation.
* Data-sheet Includes the day and step-by-step skill breakdown to track what the learner is missing and how the skills are progressing.

Permanent Product Recording

• Measuring the outcome of behavior after it has occurred.
  • Picking up trash, doing homework.
    *Observe effects of the behavior produced in the environment.
  • Homework, cleaning.
    *Advantages:
  • Not time-consuming.
  • No need to measure in real time.
  • Good for interobserver agreement (IOA) and treatment fidelity.
  • Good for complex or multiple behaviors.
    *Disadvantages:
  • No data in the moment.
  • Unclear about who engaged in the behavior.
  • Unclear about how much assistance was required.

Trial by Trial Data

• Used most often during discrete trial teaching (DTT).

  • Learning opportunity → learner response → consequence.
    *Score behavior after every trial.
    *Performance Classifications:
  • Correct (+).
  • Incorrect (-).
  • Prompted Correct.
  • Prompted Incorrect.
  • No Response.
    *Data sheets Includes:
  • Dates.
  • Name of the target.
  • A check list of correct, incorrect, and prompt.

• Typical data are converted into a percentage.

  • \frac{Independent Correct}{All Trials}

• Helps determine if targets are mastered.
  *Advantages:

  • Clear data for responses on every trial.
  • Evaluate flow of teaching session.
    *Disadvantages:
  • Difficult with overlapping instructions or running programs together.
  • Can interfere with teaching.
  • Time-consuming.

First Trial Data

• Only take data on the very first trial for each target.

*Advantages:
* Less data collection.
*Disadvantages:
* Less representative.

Estimation Data

• Mark of the Autism Partnership Method.
  • Can be used across a variety of behaviors.
    *Subjective Data:
  • Uses a scale to estimate the percentage or number of correct responses, frequency, or duration.
  • Used in conjunction with hard data.
    *Advantages:
  • Increases teaching opportunities.
  • Easy to implement.
    *Disadvantages:
  • Less accurate.
  • Requires extensive training to become accurate.