|
Using Path Analysis to Evaluate Multidimensional Models: Developing a Program Success Theory Model for Microenterprise Development Programs
|
| Presenter(s):
|
| Michele Cranwell-Schmidt,
University of Vermont,
mschmidt@uvm.edu
|
| Jane Kolodinsky,
University of Vermont,
jkolodin@uvm.edu
|
| Abstract:
This presentation will review the methodology and outcome of an evaluation that focused on factors that lead to client success in microenterprise development (MED) programs. Using a path regression analysis of data from the Vermont Micro Business Development Program, this evaluation examined the relationships between client characteristics, program activities, interim outcomes, and longer term impacts to develop a model of MED program success theory. The presentation will highlight these results and include a group discussion on the study findings, implications on program practice and public policy, and advantages and limitations of the evaluation methodology used. Statistics demonstrated excellent model fit to the data. Overall, the model demonstrates that client success in MED is influenced by a variety of interdependent factors, with access to more financial resources being the most prominent factor. The results suggest implications for business training and access to capital as well as methods to evaluate complex models.
|
|
Utilization of Structural Equation Modeling Techniques in Real-World Program Evaluation
|
| Presenter(s):
|
| Greg Welch,
University of Kansas,
gww33@ku.edu
|
| Bruce B Frey,
University of Kansas,
bfrey@ku.edu
|
| Jessica Oeth,
University of Kansas,
jessoeth@ku.edu
|
| Chris Nileksela,
University of Kansas,
chrisn@ku.edu
|
| Abstract:
Program evaluation can be enhanced by a data collection process which lends itself to statistical analysis via Structural Equation Modeling (SEM) techniques. The attractiveness of this family of statistical methods lies in their ability to address a variety of correlational questions. Unfortunately, many program evaluations cannot take advantage of SEM techniques. This, in part, is because of the lack of information available to evaluators about the advantages of applying SEM techniques, but also because SEM has fairly stringent data requirements. This presentation describes the challenges, advantages and disadvantages of utilizing SEM in an evaluation setting. An ongoing evaluation of a Kansas early childhood school readiness system will be used to provide a framework for this presentation and to illustrate the difficulties one can encounter when either the design of a particular evaluation or unplanned obstacles to data makes the use of SEM easier said than done.
|
|
Understanding Consumer Self-Perceptions Regarding Mental Health Recovery: A Structural Equation Modeling Evaluation
|
| Presenter(s):
|
| Christopher McKinney,
Mental Health Center of Denver,
christopher.mckinney@mhcd.org
|
| Kathryn DeRoche,
Mental Health Center of Denver,
kathyrn.deroche@mhcd.org
|
| Antonio Olmos,
Mental Health Center of Denver,
antonio.olmos@mhcd.org
|
| Abstract:
A consumer’s perception of his/her own mental health recovery is considered important in treatment selection, implementation, engagement, and other items related to mental health treatment. Utilizing structural equation modeling (SEM), the current study explores the relationships among five areas regarding consumers’ perceptions of mental health recovery; including hope, symptom interference, social networks, active growth/orientation, and perceived safety. A model of consumers’ perceptions of mental health recovery is presented, along with discussion of the implications regarding mental health treatment. Furthermore, the advantages of using SEM in exploring relationships among complex behaviors are discussed.
|
|
Structural Equation Model Evaluating Students’ Self-Perception Regarding Graduate Level Statistics Coursework: Demonstration of a Jackknife Approach for the Purpose of Item Reduction
|
| Presenter(s):
|
| Karen Larwin,
University of Akron,
drklarwin@yahoo.com
|
| Abstract:
The intention of the present investigation was to demonstrate the causal link between student’s prior mathematics preparation and Statistics Self-Perception, using structural equation modeling techniques. It is theorized that many graduate students continue to struggle with required statistics coursework due to poor preparation/minimal prerequisites in mathematics, even after many statistics instructors have shifted to more conceptual presentations of statistical concepts. A third-order factor structure of Statistics Self-Perception demonstrates the link between prior mathematics experience and students statistics-related attitudes/anxieties/self-efficacy. The number of observed items (k = 65) in this original third-order factor model was reduced (k = 40) arriving at a model that is parsimonious, while measurement and structurally invariant relative to the original model. Item reduction has been attempted in previous research using SEM, however these studies do not present clear evidence that model changes resulted in models that were measurement invariant while maintaining the integrity of the structural model.
|
| | | |