|
Evaluating Intervention Effects On Marijuana Use Among Youth by Comparing Alternative Structural Equation Modeling: Decision-Trees Based On Formal Model Comparisons
|
| Presenter(s):
|
| Marlene Berg, Institute for Community Research, mberg_84@yahoo.com
|
| Emil Coman, Institute for Community Research, comanus@netscape.net
|
| Abstract:
This is a systematic method of testing for true estimated post-treatment differences between intervention and control groups, by testing all assumptions regarding group differences in actual data from a quasi-experimental design through nested model SEM model comparisons and advancing from basic all-different baseline model to more parameters constrained to be equal nested models.
1. We will use AMOS to show how a small simple 2 group model can be specified, and alternative settings.
2. How a test of nested model comparison can be conducted, by adding a constraint to a baseline model.
3. We will illustrate the decision-tree paths and their meaning (significant difference testing).
4. We will interpret the meaning of various model comparisons and the impact on the evaluation final judgment.
The decision-tree SEM indicate that some test of post-intervention equality constraint test of means with different assumptions as baseline model show that the equality of means does not hold.
|
|
Finding Meaning in Numbers: How Well Do Our Measures Capture What We Evaluate?
|
| Presenter(s):
|
| Ann Doucette, George Washington University, doucette@gwu.edu
|
| Abstract:
The research literature supports therapeutic alliance and motivation to change, as key determinants of favorable treatment outcomes for individuals receiving care for mental health and substance abuse disorders. In response, models of therapeutic alliance and stages and willingness to make changes within the therapeutic setting have been generated to reflect how these constructs function and relate to mental health and substance abuse treatment outcomes. This paper focuses on the measurement of these constructs and investigates the implications of unexamined measurement properties in modeling behavioral health treatment effectiveness and consumer outcomes. Secondary analyses using existing data from Project Match, a multi-site clinical trial for alcohol treatment, the Depression Collaborative and a cocaine treatment study question whether the assumptions about measurement dimensionality, response scale adequacy, and invariance across diverse samples (diagnostic, age, gender, race/ethnicity, etc.) hold, and whether findings accurately reflect treatment effectiveness.
|
|
Barriers to Treatment of the Chronic Homeless with Co-occurring Disorders
|
| Presenter(s):
|
| M David Miller, University of Florida, dmiller@coe.ufl.edu
|
| Abstract:
The STAR (Seeking Treatment and Recovery) program provides treatment for the chronically homeless with co-occurring disorders (substance abuse problem and mental illness). The program provides a range of services through intensive case management. The program includes treatment of mental illness, substance abuse problems, and establishing housing and employment.
In our context evaluation, we have found that considering why chronic populations have not taken advantage of treatment in the past is an important part of understanding the population and being able to better serve them. Consequently, we developed an instrument to measure the barriers to treatment (BTI). This paper shows the results of that instrument development including reliability and validity, and use of the instrument (N=136). Clients tend to not receive treatment due to personal issues rather than program or societal issues.
|
| | |