Discontinuous Measurement Distortion & Bias
When collecting behavioral data, continuous measurement systems (like frequency and duration) capture every instance of a target behavior. However, when clinical or resource constraints force you to select discontinuous measurement procedures, you systematically introduce measurement artifact.
On the 6th Edition BCBA exam, you must go beyond knowing how to run interval recording systems; you must master the exact mathematical reasons why these systems introduce predictable, structural bias. Choosing between Partial Interval Recording (PIR) and Whole Interval Recording (WIR) dictates whether your graph will suffer from severe overestimation or underestimation of the true behavioral rate.
The Mathematical Flaw of Partial Interval Recording (PIR)
Partial Interval Recording requires the observer to score an occurrence if the target behavior occurs at any point during the interval, regardless of how brief that instance is.
⚠️ The PIR Bias: Partial Interval Recording tends to consistently overestimate high-rate or continuous behavioral metrics.
Why PIR Overestimates
Imagine a 10-second interval tracking stereotypical vocalizations. The client emits a brief, 1-second burst of the behavior at the very beginning of the interval and then stops entirely for the remaining 9 seconds. Because the behavior occurred for a fraction of a second, the observer checks the box as an occurrence ($100\%$ of the interval block is marked).
On a data graph, this interval looks identical to a session where a client engaged in vocalizations for the entire 10 seconds. Thus, PIR artificially inflates your metrics because it converts low-duration, high-frequency behavior into an illusion of prolonged, continuous duration.
The Mathematical Flaw of Whole Interval Recording (WIR)
Whole Interval Recording requires the target behavior to persist uninterrupted for the entire duration of the interval to be scored as an occurrence.
📉 The WIR Bias: Whole Interval Recording inherently underestimates duration-based behaviors.
Why WIR Underestimates
Consider a 10-second interval tracking on-task behavior. A learner is perfectly focused, working on their assignment from second 0 through second 9. At second 9.5, they glance up at the clock for half a second. Because the behavior did not persist through the entire block uninterrupted, the observer must mark that interval as a non-occurrence ($0\%$ of the interval block is marked).
Even though the learner was engaged for $90\%$ of the interval, the data log shows a flat zero. This is why WIR creates an underestimation artifact, systematically hiding true behavioral progress from your visual data displays.
Structural Summary for Clinical Decision Making
To ensure you choose the correct system on applied exam questions, evaluate the clinical goal of your tracking protocol:
| Measurement System | Scoring Requirement | Systematic Bias Profile | Best Clinical Matching |
| Partial Interval (PIR) | Behavior occurs at any point during the interval. |
Overestimates high-rate metrics and overall duration. |
Use when the clinical goal is to decrease behaviors (e.g., aggression, scripting), ensuring you don’t miss instances. |
| Whole Interval (WIR) | Behavior occurs for the entirety of the interval. |
Underestimates true duration and behavioral occurrence. |
Use when the clinical goal is to increase behaviors (e.g., on-task attention, cooperative play), ensuring a strict progress baseline. |