An Empirical Study of Measurement Error Using Partial Interval, Whole Interval, and Momentary Time-Sample Recording Methods
Steffen, William F.
MetadataShow full item record
The problem. An examination was made of the effect varying Session Response Duration, Mean Individual Response Duration, and length of recording interval on the accuracy of data using whole interval, partial interval, and momentary time-sample methods. The Session Response Duration was defined as the total cumulative in-seat response duration, and the Mean Individual Response Duration as the average in-seat response duration. Procedure. Eight 30-minute sessions having a 20% and 50% Session Response Duration with four Mean Individual Response Duration seach were videotaped. Observers viewed the videotapes using the three recording methods. The data from the observers were compared to the Session Response Duration and predicted scores derived from models by Milar and Hawkins, and by Ary. Findings. The momentary time-sample method always gave a more accurate estimate of the Session Response Duration. Increased Mean Individual Response Duration increased the accuracy of the whole and partial interval estimates. Predictions of the momentary time-sample data were based on valid assumptions, whereas assumptions of the models were not met with the whole and partial interval methods. Conclusion. The momentary time-sample method is more accurate than the whole and the partial interval methods when measuring the Session Response Duration. Recommendation. More studies of the three recording methods in the applied settings are needed to extend generality.
v, 50 leaves. Advisor: Kenneth E. Lloyd