|
| Interval Methods for Detecting Changes in Frequency and Duration Events: What You Should Know from the Standpoint of Functional Control |
| Saturday, May 24, 2008 |
| 2:30 PM–3:50 PM |
| Stevens 1 |
| Area: DDA/AUT; Domain: Applied Research |
| Chair: John T. Rapp (St. Cloud State University) |
| CE Instructor: John T. Rapp, Ph.D. |
| Abstract: The utility of partial-interval recording (PIR) and momentary time-sampling (MTS) for detecting changes in simulated events was evaluated using single-subject experimental designs. Colby-Dirksen et al. evaluated the extent to which various interval sizes of PIR and MTS detected the same functional control that was demonstrated with continous duration recording (CDR). Michalski et al. evaluated whether various interval sizes of PIR and MTS detected changes in various event-rates. In this sense, both Colby-Dirksen et al. and Michalski et al. evaluated the probability that interval methods produce false negatives (i.e., failed to detect effects that were evident with CDR) when evaluating the effects of independent variables. To this end, Carroll et al. evaluated the possibility that interval methods depict false positives (i.e., depict functional control that is not evident with CDR). Finally, Devine et al. evaluated the extent to which the length of observation periods influenced the sensitivity of interval methods for detecting various changes in duration events. |
| |
| An Extensive Evaluation of Functional Control with Interval Methods I: Duration Events. |
| AMANDA M. COLBY (St. Cloud State University), John T. Rapp (St. Cloud State University), Ally Lindenberg (St. Cloud State University) |
| Abstract: The sensitivity of partial interval recording (PIR) and momentary time sampling (MTS) methods for detecting functional control was evaluated for events that occurred for 25%, 33%, 40%, 50%, 66%, and 75% of the time during 10-min sessions. Simulated data derived from continuous duration recording (CDR) were re-calculated using 10-s, 20-s, 30-s, 1-min, and 2-min intervals for PIR and MTS. Each data set was evaluated with reversal designs to determine the extent to which changes in varying durations of events were detected with each interval method. For MTS, the results showed that (a) interval sizes up to 30 s detected the small effects and (b) interval sizes up to 1 min typically detected the large effects. Conversely, for PIR only 10-s intervals detected changes in duration events and such changes that were detected only for the large effects. As a whole, the results show that the sensitivity of interval methods was influenced by both the ratio of the interresponse time to event-run within each session and the percentage of the change from A-phase to B-phase that was evaluated. |
| |
| An Extensive Evaluation of Functional Control with Interval Methods II: Frequency Events. |
| DARA N. MICHALSKI (Redwood Learning Center), John T. Rapp (St. Cloud State University), Amanda M. Colby (St. Cloud State University) |
| Abstract: This study evaluated the accuracy of partial-interval recording (PIR) and momentary time sampling (MTS) for measuring frequency events. Sessions with simulated data for continuous frequency recording (CFR) were generated for seven event-rates and were subsequently rescored using 10 s, 20 s, 30 s, 1 min, and 2 min PIR and MTS measures. The data that were produced with each interval method were depicted in line graphs and evaluated using ABAB reversal designs. Each line graph was compared to the respective CFR line graph to determine if the interval method produced the same conclusions about functional control. The results show that PIR with interval sizes up to 1 min detected the large effects; however, only 10 s PIR reliably detected the small, the moderate, and the large effects. Conversely, each interval size of MTS was insensitive to small effects, but 10 s MTS detected over two thirds of the moderate and the large effects. The results support prior conclusions regarding the utility of 10 s PIR for evaluating the effects of independent variables on frequency events. |
| |
| An Extensive Evaluation of Functional Control with Interval Methods III: False Positives for Duration Events. |
| REGINA CARROLL (St. Cloud State University), John T. Rapp (St. Cloud State University), Amanda M. Colby (St. Cloud State University), Ally Lindenberg (St. Cloud State University) |
| Abstract: The extent to which partial interval recording (PIR) and momentary time-sampling (MTS) methods generate false positives was evaluated for events that occurred for 25%, 33%, 40%, 50%, 66%, and 75% of the time during 10-min sessions. Simulated data that were produced in the Colby et al. (2007) study were re-evaluated in this experiment. For each targeted percentage, low, moderate, and high inter-response time to event-run ratios were compared with reversal designs to determine whether interval methods depicted functional control that was not evident with continuous duration recording. The results show that PIR with 10-s intervals generated a high percentage of false positives whereas MTS did not generated false positives with any interval size. Specifically, 10-s PIR generated false positives for each of the low to moderate, low to high, and moderate to high comparisons for events that occurred for 25%, 33%, and 40% of a session. Potential problems with using 10-s PIR to evaluate the effects of independent variables on duration events are briefly discussed. |
| |
| Evaluating the Accuracy of Interval Recording Methods in Estimating Duration Events: Assessing the Effect of Session Length. |
| SHERISE LORRAINE DEVINE (St. Cloud State University/St. Amant), John T. Rapp (St. Cloud State University) |
| Abstract: This study extends the body of research that exists in assessing the accuracy of partial-interval recording (PIR) and momentary time sampling (MTS) in estimating duration events. Simulated data were generated to produce various absolute durations of behavior (25%, 33%, 40%, 50%, 66% and 75%) for various session lengths (10 min, 30 min, and 60 min) using an ABAB reversal designs. The average inter-response time to event-run ration was simulated to be low, medium, or high for each percentage. The generated data were scored using continuous duration recording (CDR) and then rescored using PIR MTS with intervals sizes of 10 s, 20 s, 30 s, 1 min, and 2 min. The resulting data paths for PIR and MTS were visually inspected for similarity with CDR regarding functional control. In addition, within-session patterns of events produced by PIR and MTS methods were compared to the within-session patterns of behavior produced by the CDR measure for each data set. The results provide further support for the conclusion that a number of variables influence the sensitivity of intervals methods. |
|
| |