Taking a closer look: time sampling and measurement error

Area of Science:

• Behavioral analysis
• Applied behavior analysis
• Research methodology

Background:

• Accurate measurement of behavior is crucial in applied behavior analysis.
• Traditional interval time sampling methods may introduce measurement error.
• Understanding the limitations of different direct observation methods is essential.

Purpose of the Study:

• To compare the accuracy of whole interval, partial interval, and momentary time-sample measures against a continuous measure of in-seat behavior.
• To analyze the sources and characteristics of measurement error in interval time sampling.
• To determine the superiority of momentary time sampling in estimating behavior duration.

Main Methods:

• A participant engaged in in-seat behavior across sessions with varying occurrence rates (20%, 50%, 80%).
• Continuous measurement (percentage of time) was used as the standard.
• Whole interval, partial interval, and momentary time-sample data were collected and compared to the continuous measure.

Main Results:

• Interval time sampling (whole and partial) exhibited extensive, unidirectional, and inconsistent measurement errors.
• Errors in interval time sampling were found to be a function of the time per response.
• A procedural analysis revealed that frequency and duration are confounded in interval time sampling.
• Momentary time sampling demonstrated greater accuracy in estimating behavior duration compared to interval methods.

Conclusions:

• Interval time sampling methods are prone to significant measurement error, potentially compromising data integrity.
• The confounding of behavior frequency and duration in interval time sampling limits its utility for duration measurement.
• Momentary time sampling offers a more reliable and accurate method for estimating the duration of observed behaviors.