Data Collection and Analysis Flashcards

A behavior analyst is measuring self-injurious behavior (SIB) using a partial interval recording system with 10-second intervals over a 30-minute session. The data show 80% interval occurrence. A colleague suggests switching to whole interval recording to get a 'more accurate' picture. Which statement BEST describes the measurement implications of this switch?

Answer: Whole interval recording will likely yield a lower percentage than partial interval, potentially underestimating the true frequency of SIB

Partial interval recording overestimates behavior occurrence because any instance within the interval is scored, while whole interval recording underestimates it because the behavior must persist for the entire interval to be scored. Switching from partial to whole interval when SIB is at 80% would likely produce a lower percentage, potentially masking the severity of the behavior.

During baseline, a client's disruptive behavior shows a decelerating trend across 7 sessions. The behavior analyst proceeds with a DRO intervention anyway. From a data-based decision-making standpoint, what is the PRIMARY concern with this approach?

Answer: The decelerating baseline trend confounds the ability to attribute any subsequent reduction to the intervention

When a baseline shows a decelerating trend in the target behavior (already moving in the desired direction), introducing an intervention makes it impossible to determine whether any continued reduction is due to the treatment or the natural continuation of the existing trend. This is a fundamental threat to internal validity. The intervention effect cannot be isolated from the pre-existing trend.

An analyst collects IOA data using the exact count per interval method across 20 intervals. Observer 1 records: 2, 0, 1, 3 (for the first four intervals). Observer 2 records: 1, 0, 2, 3 (for the same intervals). What is the IOA score for these four intervals using the exact count per interval method, and what does a score below 80% indicate?

Answer: IOA = 75%; the observers disagree on one of four intervals; a score below 80% suggests the operational definition or training requires improvement

Exact count per interval IOA scores each interval as an agreement only when both observers record identical counts. Intervals 1 (2 vs 1), 3 (1 vs 2) are disagreements; intervals 2 (0 vs 0) and 4 (3 vs 3) are agreements. That gives 2 agreements out of 4 intervals = 50% — making option A more technically accurate on the math, but option B correctly identifies the threshold implication. However, the correct calculation is 2/4 = 50%. A score below 80% signals the operational definition, observer training, or measurement conditions need refinement to ensure data reliability.

A BCBA is evaluating treatment effectiveness using a changing criterion design. The criterion is shifted upward after the client meets the previous criterion for 3 consecutive sessions. After the 4th criterion shift, the client's performance plateaus significantly below the new criterion for 8 sessions. Which action is MOST consistent with data-based decision-making within this design?

Answer: Recalibrate the criterion step size downward based on the magnitude of the performance plateau and the client's rate of acquisition

In a changing criterion design, when a client plateaus well below the new criterion, the data indicate the criterion step size was too large relative to the client's current rate of acquisition. The appropriate data-based response is to recalibrate — set a smaller, more achievable criterion increment. Returning to a previous criterion (reversal) is not part of a changing criterion design's logic, and abandoning the design is premature and clinically unjustified based on one plateau.

A behavior analyst is using momentary time sampling (MTS) with 5-minute intervals to measure on-task behavior during a 60-minute work session. The resulting data show 70% on-task. A direct observation using continuous duration recording during the same session yields 85% on-task. Which explanation BEST accounts for this discrepancy?

Answer: MTS systematically underestimates behavior when the target behavior tends to terminate just before observation moments, which is a known artifact of this method when behaviors cluster at interval beginnings

MTS accuracy is highly dependent on the temporal distribution of the behavior relative to observation moments. If on-task behavior tends to occur at the start of intervals and the client transitions off-task before the observation moment (end of interval), MTS will underestimate true duration. This is a recognized limitation: MTS neither systematically overestimates nor underestimates but is sensitive to behavioral patterns within intervals. When behavior terminates before sampling points, MTS will yield lower estimates than continuous recording.

When calculating the mean count per interval IOA for frequency data, an analyst finds that one observer consistently records higher counts than the other across all intervals, but the ratio between their counts is nearly constant (approximately 1.5:1). Which interpretation and recommendation is MOST appropriate?

Answer: The constant ratio suggests systematic observer bias rather than random measurement error; recalibrating both observers against video-recorded sessions and revising the operational definition to reduce interpretive ambiguity is indicated

A constant ratio discrepancy across all intervals signals systematic observer bias — one observer is consistently applying the operational definition differently than the other (e.g., using a broader vs. narrower topographic definition, or differing on what constitutes a discrete response). Random measurement error would produce variable ratios. Systematic bias is a more serious validity threat and requires identifying the source — typically through calibration sessions with video-recorded behavior and clarification of the operational definition to reduce interpretive latitude.