|
|
Using Multielement Designs in Applied and Practical Settings |
Wednesday, November 15, 2017 |
11:00 AM–11:20 AM |
Scene DEF, Niveau 0 |
Area: PRA |
Instruction Level: Intermediate |
Chair: Marc J. Lanovaz (Université de Montréal) |
|
Using Multielement Designs in Applied and Practical Settings: An Examination of False Outcomes |
Domain: Service Delivery |
MARC J. LANOVAZ (Université de Montréal), Mary Francis (Université de Montréal) |
|
Abstract: Researchers and practitioners alike often use multielement designs in applied settings to examine the effects of variables or interventions on behavior. Although visual inspection is a hallmark of behavior analysis, our knowledge of the probability of reaching incorrect conclusions from data collected within multielement designs is limited. This issue is often a concern when sharing the results of multielement studies with practitioners and researchers from other professions who have mainly adopted inferential statistics to test their hypotheses. Thus, the purpose of the presentation is to address this issue by examining the probability of reaching false outcomes when adopting multielement designs. Our analyses are currently ongoing, but our preliminary findings indicate that having at least five data points within each condition provides adequate power (.80) and sufficient control over type I error rates (.05). The potential implications of these findings will be discussed in terms of applying structured criteria to render the analysis of the results of multielement designs more reliable. |
|
|
|