- Applied Behavior Analysis Measurement and Data Collection Questions and Answers Flashcards

A behavior analyst is measuring vocal stereotypy using a 10-second partial interval recording system across a 10-minute session. The behavior occurs continuously for the first 3 minutes, then stops entirely. What does the resulting data most likely OVERESTIMATE?

Answer: The total duration of the behavior

Partial interval recording scores an interval as 'occurrence' if the behavior happens at any point during that interval, regardless of how long. Because the behavior occurred continuously for the first 3 minutes (18 intervals all scored +), the system captures those accurately, but it cannot reflect that the behavior stopped — yielding a percentage that overestimates true duration relative to the session. Partial interval recording is known to systematically overestimate duration, especially for high-rate or continuous behaviors.

A practitioner uses momentary time sampling (MTS) every 60 seconds to measure on-task behavior. A client is on-task for 55 seconds of every 60-second interval but off-task at each observation moment. What will the MTS data show, and what is the measurement error called?

Answer: 0% on-task; systematic underestimation

Momentary time sampling only records behavior state at the precise observation moment. If the client is consistently off-task at each 60-second mark — despite being on-task for 55 of every 60 seconds — MTS will record 0% on-task, a severe and systematic underestimation of a high-frequency behavior. This is a known limitation of MTS: it underestimates behaviors that tend NOT to occur at observation moments.

An observer collects interobserver agreement (IOA) data using the scored-interval IOA method versus the total count IOA method on the same session. The behavior is high-rate and the two observers disagree on which specific intervals contained the behavior. Which statement BEST describes the expected relationship between the two IOA scores?

Answer: Scored-interval IOA will be lower than total count IOA because it accounts for interval-by-interval agreement

Total count IOA divides the smaller total count by the larger (e.g., 48/50 = 96%), which can appear high even when observers disagree on which specific intervals the behavior occurred. Scored-interval IOA compares agreement interval-by-interval only for intervals where at least one observer scored an occurrence, making it more sensitive to disagreements about when the behavior happened. For high-rate behaviors with distributional disagreements, scored-interval IOA will be lower and more informative than total count IOA.

During baseline data collection, a behavior analyst notices the data path shows a therapeutic trend (decreasing problem behavior). According to data decision rules in single-case research design, what is the MOST appropriate next step?

Answer: Continue baseline until the trend stabilizes or reverses, or document the limitation and proceed with caution

A therapeutic trend in baseline is a threat to internal validity because any subsequent improvement during intervention cannot be attributed solely to the independent variable. The standard practice is to extend baseline until the trend stabilizes, reverses, or reaches a level inconsistent with the therapeutic direction. If practical constraints prevent extension, the analyst must document this as a methodological limitation. Discarding data or immediately proceeding without acknowledgment would violate scientific and ethical standards.

A BCBA is training a new RBT to collect frequency data on aggression. After training, both observe the same 20-minute session independently. The BCBA records 12 instances; the RBT records 9. Using the smaller-divided-by-larger total count IOA formula, the agreement is 75%. The BCBA concludes IOA is unacceptable. Which additional analysis would BEST clarify whether the disagreement is due to observer drift, reactivity, or a measurement definition problem?

Answer: Analyze the session video to identify which specific instances each observer scored and whether missed instances share topographical features

Reviewing the video to identify which specific instances were scored/missed allows the BCBA to determine if the RBT consistently missed a particular topography (e.g., lower-intensity hits), suggesting a definition problem or insufficient calibration. If misses are random, drift is more likely. If disagreements cluster near the observer, reactivity may be the cause. Simply recalculating with a stricter formula, repeating BST without analysis, or collecting more IOA data does not diagnose the source of the problem.

A behavior analyst is comparing two measurement procedures for elopement: (1) duration recording starting from the moment the client leaves the designated area and (2) latency recording from the therapist's verbal instruction to return. A client who elopes immediately upon instruction and returns quickly would produce which pattern of data?

Answer: Short latency and short duration

Latency measures the time from the discriminative stimulus (instruction to return) to the onset of the response (returning). If the client elopes immediately upon instruction, the latency to RETURN would be long — but the question specifies elopement happens upon instruction, meaning the latency TO ELOPE from any prior cue might be short. Re-reading: latency is measured from the instruction to return, and the client returns quickly — so latency to return is SHORT. Duration of elopement is also short because the client returns quickly. Both measures are short for a client who elopes briefly and returns promptly.