ABAT Data Collection and Analysis 2 — Questions and Answers

Question 1: A BT records whether a learner is on-task every 5 minutes by checking at exactly the moment the timer sounds. This is:

Whole interval recording
Partial interval recording
Momentary time sampling (Correct answer)
Continuous event recording

Correct answer: Momentary time sampling

Checking behavior only at the moment of a timed signal — not throughout the interval — is momentary time sampling.

Momentary time sampling requires the observer to record whether the behavior is occurring at a single designated moment (e.g., when the timer sounds). Unlike interval recording procedures that require monitoring throughout the interval, momentary time sampling minimizes observer effort and is suitable for naturally occurring settings where continuous monitoring is difficult. It provides an estimate of the proportion of time the behavior occurs and is less prone to overestimation than partial interval recording.

Question 2: A behavior analyst graphs data and observes a flat trend with high variability in the baseline phase. This means:

The behavior is stable and ready for intervention
The behavior is unstable — something may be controlling the behavior that is not yet identified (Correct answer)
The intervention is ineffective
The data collector made errors

Correct answer: The behavior is unstable — something may be controlling the behavior that is not yet identified

High variability in baseline means the behavior is inconsistent, suggesting uncontrolled variables are influencing it — intervention should typically wait until a stable baseline is obtained.

Visual analysis of baseline data considers level (average performance), trend (directional change), and variability (consistency of data points). High variability indicates the behavior is fluctuating due to unidentified variables. Intervening on a variable baseline makes it difficult to determine whether observed changes are due to the intervention or to the same uncontrolled variables causing variability. Behavior analysts should try to identify and control sources of variability before introducing an intervention, or at minimum, continue baseline until stability improves.

Question 3: For which type of behavior is frequency recording LEAST appropriate?

A behavior that occurs discretely and has a clear beginning and end
Behaviors with very consistent, brief durations
A behavior that occurs continuously for extended periods (e.g., on-task behavior that lasts 20 minutes) (Correct answer)
Behaviors that occur 3–5 times per session

Correct answer: A behavior that occurs continuously for extended periods (e.g., on-task behavior that lasts 20 minutes)

Frequency recording is poorly suited for behaviors with highly variable or long durations because the count does not capture how long the behavior lasted.

Frequency recording counts the number of occurrences of a discrete behavior. It is poorly suited for behaviors that do not have clear, consistent episode lengths, particularly behaviors that occur continuously or for long durations (e.g., on-task, stereotypy, self-injurious behavior that may last seconds or minutes per episode). For these behaviors, duration recording or time sampling procedures are more informative. Frequency recording is most valid when episodes are brief, discrete, and relatively uniform in duration.

Question 4: Cumulative graphs in ABA display data so that:

Each data point shows the behavior count for that session only
Each data point adds to all previous data points, so the line never decreases (Correct answer)
Each session is displayed on a separate graph panel
Behavior rates are shown as percentages of intervals

Correct answer: Each data point adds to all previous data points, so the line never decreases

In a cumulative graph, each new data point is added to the cumulative total — the graph line can only move upward or remain flat, never decrease.

Cumulative records plot the running total of responses across sessions. Because each session's count is added to all previous counts, the graph line can only ascend or remain flat — it never decreases (assuming the behavior is a countable event). The slope of a cumulative record represents response rate: a steep slope indicates a high rate, a flat segment indicates a period of no responding. B.F. Skinner made extensive use of cumulative records. They are useful for detecting changes in response rate but make it harder to see session-to-session variation directly.

Question 5: Which calculation gives the rate of behavior per minute during an observation session?

Total occurrences × observation time in minutes
Total occurrences ÷ observation time in minutes (Correct answer)
Total occurrences ÷ number of sessions
Total occurrences × number of intervals

Correct answer: Total occurrences ÷ observation time in minutes

Rate (responses per minute) = total count divided by the total observation time in minutes, allowing comparison across sessions of different lengths.

Rate of behavior normalizes frequency data by the observation time, making it possible to compare data across sessions or observers that differ in observation duration. Rate = total occurrences ÷ observation time in minutes (or hours). For example, if a learner asks 12 questions during a 30-minute session, the rate is 0.4 questions per minute. Rate data are especially useful when session lengths are not constant and when comparing behavior across different time periods.

Question 6: Operational definitions of target behaviors in data collection should be:

Broad enough to capture all related behaviors without specifics
Clear, objective, and complete — observable, measurable, and unambiguous (Correct answer)
Written in general clinical terminology only
Changed frequently to reflect new observations

Correct answer: Clear, objective, and complete — observable, measurable, and unambiguous

Operational definitions must be observable, measurable, and specific enough that multiple observers can reliably identify the behavior consistently.

An operational definition describes a behavior in observable and measurable terms, specifying exactly what the behavior looks like (topography), when it starts and stops, and what distinguishes it from similar or related behaviors. Well-written operational definitions allow any trained observer to identify the behavior consistently, which is essential for reliable data collection and high IOA. Poor definitions (e.g., 'aggression' without specifying what counts) lead to inconsistent data. Operational definitions should be tested before implementation by having multiple observers practice and comparing their data.

ABAT Data Collection and Analysis 2

A BT records whether a learner is on-task every 5 minutes by checking at exactly the moment the timer sounds.
This is: