Discontinuous Measurement in ABA
In applied behavior analysis, continuous measurement (recording every instance of a behavior) is the gold standard. However, continuous measurement is not always practical. When a behavior occurs at high rates, when an observer is monitoring multiple clients, or when resources are limited, discontinuous measurement procedures provide a practical alternative.
The three main discontinuous measurement methods are partial interval recording, whole interval recording, and momentary time sampling. Each divides the observation period into intervals and uses different rules for scoring behavior within those intervals. Understanding the differences, including how each method distorts the data, is essential for the BCBA exam.
Partial Interval Recording
How It Works
In partial interval recording, the observer divides the observation period into equal intervals (for example, 10-second intervals within a 5-minute observation). The observer scores an interval as "occurrence" if the target behavior occurs at any point during that interval, even briefly. Once the behavior is observed, the interval is scored, and the observer waits for the next interval to begin.
Overestimation of Behavior
Partial interval recording tends to overestimate the actual occurrence of behavior. This is because a behavior that occurs for just one second within a 10-second interval results in the entire interval being scored as occurrence. The data might suggest the behavior occurred during 80% of intervals when the actual duration was much less.
The shorter the actual duration of the behavior relative to the interval length, the greater the overestimation. This is an important point for exam questions.
When to Use It
Partial interval recording is appropriate when you want to ensure you capture the occurrence of a behavior that might be brief or easy to miss. It works well for behaviors that are difficult to count individually because they occur at very high rates or because individual instances are hard to distinguish.
It is commonly used for behaviors such as stereotypy (repetitive motor movements), off-task behavior, or vocalizations that occur in rapid bursts.
Whole Interval Recording
How It Works
In whole interval recording, the observer scores an interval as "occurrence" only if the target behavior occurs throughout the entire interval. If the behavior stops at any point during the interval, even briefly, the interval is scored as nonoccurrence.
Underestimation of Behavior
Whole interval recording tends to underestimate the actual occurrence of behavior. A behavior that occurs for 8 out of 10 seconds in an interval would still be scored as nonoccurrence because it did not persist for the entire interval. This means the data will show less behavior than actually occurred.
The longer the interval relative to the typical duration of the behavior, the greater the underestimation. This point is frequently tested on the exam.
When to Use It
Whole interval recording is appropriate when you want a conservative estimate of behavior, particularly for behaviors you want to increase. Because it underestimates, if whole interval data show high levels of a desirable behavior, you can be confident the behavior is truly occurring at high levels.
It is commonly used for behaviors such as on-task behavior, engagement, or sustained attention, where the goal is to demonstrate that the individual is maintaining the behavior for extended periods.
Momentary Time Sampling
How It Works
In momentary time sampling (MTS), the observer looks at the individual only at the exact moment each interval ends (or begins, depending on the protocol). If the behavior is occurring at that precise moment, the interval is scored as occurrence. If not, it is scored as nonoccurrence. The observer does not need to watch continuously between sampling moments.
Most Accurate Estimate
Research has shown that momentary time sampling provides the most accurate estimate of the actual percentage of time a behavior occurs when compared to partial and whole interval recording. While it can both overestimate and underestimate depending on the specific pattern of behavior, on average it provides a closer approximation to continuous measurement than either of the other methods.
When to Use It
Momentary time sampling is particularly useful when an observer needs to monitor multiple individuals simultaneously or when continuous observation is not feasible. Because the observer only needs to look at the individual at specified moments, it reduces observer fatigue and allows monitoring of several clients at once.
It is commonly used in classroom settings where a teacher or aide monitors multiple students, in residential settings with multiple residents, and in any situation where continuous observation of a single individual is impractical.
Choosing the Right Method
When selecting a discontinuous measurement procedure, consider these factors:
- Purpose of measurement: If you need a conservative estimate (to ensure you are not overreporting a behavior you want to decrease), whole interval recording may be appropriate. If you want to make sure you capture occurrences of a low-duration behavior, partial interval recording may be better.
- Accuracy needs: If accuracy is the priority, momentary time sampling generally provides the best estimate of actual behavior levels.
- Practical constraints: If the observer cannot watch continuously, momentary time sampling is the most feasible option.
- Behavior characteristics: Brief, discrete behaviors may be better captured with partial interval recording. Sustained behaviors are better assessed with whole interval recording or momentary time sampling.
Interval Length Considerations
The length of the interval affects the accuracy of all three methods. Shorter intervals generally produce more accurate estimates because the sampling is more frequent. However, shorter intervals also increase the demands on the observer. A common interval length is 10 to 15 seconds for partial and whole interval recording, and 1 to 5 minutes for momentary time sampling in applied settings.
Key Exam Takeaways
For the BCBA exam, remember these critical points: partial interval recording overestimates, whole interval recording underestimates, and momentary time sampling provides the most accurate estimate of duration. Know which method is best for different types of behaviors and measurement situations. Be prepared for scenario questions that ask you to identify the most appropriate method or to predict how a measurement method would distort the data for a given behavior.