A Comparison of Observed Functional Relations across Measurement Methods Used to Assess Vocal Stereotypy in Two Adolescents with Autism
Restricted (Penn State Only)
- Author:
- Mantzoros, Theoni
- Graduate Program:
- Special Education
- Degree:
- Doctor of Philosophy
- Document Type:
- Dissertation
- Date of Defense:
- April 15, 2022
- Committee Members:
- Jonte Taylor, Major Field Member
Pamela Wolfe, Major Field Member
David Lee, Chair & Dissertation Advisor
Charles Geier, Outside Unit & Field Member
David Lee, Professor in Charge/Director of Graduate Studies - Keywords:
- autism spectrum disorder
vocal stereotypy
measurement
continuous duration recording
partial-interval recording
momentary time-sampling - Abstract:
- Measurement across single-case research and application varies from continuous to discontinuous methods. Although there is agreement across fields that continuous methods are ideal and should be used whenever possible, there is also an understanding that dependent on resources such as time, finances, and personnel, it may not always be a viable option. Discontinuous measures, such as MTS and PIR, have been used for more than 50 years in measuring behaviors of interest. The literature on the comparison between discontinuous and continuous methods has primarily focused on the comparison of accuracy across measures. A small subset of this literature has also assessed functional relations across measures, with two articles implementing visual analysis based on the assessment of experts. This study expands the literature by assessing MTS and PIR measures of 10-s, 20-s, 30-s, and 60-s intervals against CDR for accuracy, as well as the determination of functional relations and effect size as compared to Tau-U. Outcomes of this study are consistent in that PIR overestimates behavior, however MTS results indicate that although MTS over- and underestimates behaviors, a majority of the sessions overestimate. Additionally, as measures increase in interval length, PIR systematically increases in error, while MTS mean error remains stable, however variability in individual sessions increases. For functional relations, up to 60-s MTS measures present consistently with CDR, while only 10-s and 20-s PIR measures present with similar reliability. Effect size analysis varies across methods. Limitations and future research are discussed.