|
Methods |
Saturday, May 23, 2015 |
2:00 PM–3:50 PM |
007A (CC) |
Area: EAB |
Chair: Abdulrazaq A. Imam (John Carroll University) |
|
Mathematics Behind the Stimulus Equivalence |
Domain: Theory |
CELSO SOCORRO OLIVEIRA (UNESP - Sao Paulo State University) |
|
Abstract: This paper presents a review of mathematical operations and terms that are commonly used in Matching-To-Sample (MTS) during Stimulus Equivalence experiments. Most articles uses various mathematical operations atributing them to Set Theory. Many of those operations and concepts are not pertinent to that Set of Knowledge. Some operations belongs to Geometry and others to Graph Theory. On the same way, various terms used on articles about Stimulus Equivalence belong to other Mathematical areas of knowledge. The correct use of operations and terms helps to understand the psychological phenomena and may include other important aspects that could be the object of future studies. Studies of complex networks, with many sets interconnected in different ways may use properties that typical of Analitical Geometry or Graph Theory, such as nodal distance, or the strategies of teaching as SaN or CaN. Although MTS might be considered an Addiction Operation in some Mathematical areas, other operations such as Multiplication or Composition are not found in the Stimulus Equivalence literature. |
|
Small-Sample Statistical Analysis Using Multiple Single-Case, Nonparametric and Monte Carlo Methods for Behavior-Analytic Research |
Domain: Theory |
ANDRE V. MAHARAJ (Florida International University), Jacob L. Gewirtz (Florida International University) |
|
Abstract: The results of single-case design studies in behavior-analytic research typically rely solely on visual analyses for reporting findings, often without any consideration of difference-testing or inferential-statistical processing. Although graphical methods are appropriate, there are instances in research using small samples where non-parametric analyses may be useful, or indeed warranted. Examples of such occurrences may be: 1) When agreement on the trend in data is not apparent using graphical methods, 2) To aid in the identification of possibly significant differences among participants or behaviors, and 3) To bolster the assertions of the experimenter with respect to predictions indicated by visual trends. Additionally, advances in computing have enabled resampling techniques based on Monte Carlo methods to be used effectively for small samples in both difference and prediction testing. Visual analysis alone may be insufficient to extricate the findings, and a combined approach using appropriate statistical techniques is suggested. Various methods and their respective benefits are discussed. |
|
Identifying and Analyzing the Components of the Topology for Improving Baseline Methods in Interventions and Experimental Studies |
Domain: Basic Research |
RAY BROGAN (Kaplan University) |
|
Abstract: Because the topographies of various applications in baseline studies are intrinsically different, replications of successful approaches become difficult. As an intervention, what worked in a seemingly similar situation does not prove to be similarly effective. As an experimental study, the external validity of any given study can only be established by the power of that study. Still, the effectiveness of an intervention or the power of an experimental study can be recognized by a dramatic change from baseline due to the treatment. However, for operational analysis, even the measures of moderate change have to be statistically available. The antecedents and consequences of a treatment are expected to be observed and recorded. However, too often unmeasured conditions influence the effect of the treatment in ways that are not observed. These unmeasured influences can be used to explain the dissimilarities in results of seemingly similar approaches. Using archival data collected from various sources, this presentation will demonstrate how the topography of baseline methods can be modified for greater effectiveness and experimental power. Recommendations will be appropriate for researchers and practitioners. |
|
Place of Behavior Analytic Research in the “New” Psychological Science |
Domain: Theory |
ABDULRAZAQ A. IMAM (John Carroll University) |
|
Abstract: The “new” psychological science seeks to promote a culture of replication in response to rampant publication bias and some controversial failures to replicate. Two of the solutions adopted are the emphases on the New Statistics by Psychological Science and the growing use of replication repositories. Where does behavior analytic research fit in this effort? Although experimental and applied research methods in behavior analysis naturally are replication focused and replication friendly, trending growth in group designs in areas of behavior analytic research suggests that we should be sharing the concerns in the broader psychological science community. What measures, if any, are behavior analytic journals taking or should be taking to address these concerns? Not all of the solutions being considered are amenable to behavior analytic research. How do we proceed? Recommendations are offered to begin the conversation |
|
|
|