|
Implementing a Randomized, Controlled Study Design in a Community-Based Program Setting: Learning to Balance Service Provision and Robust Science
|
| Presenter(s):
|
| Melanie Martin-Peele,
University of Connecticut Health Center,
peele@uchc.edu
|
| Daria Keyes,
The Village for Families and Children Inc,
dkeyes@villageforchildren.org
|
| Patricia Schmidt,
The Village for Families and Children Inc,
pschmidt@villageforchildren.org
|
| Cheryl Smith,
University of Connecticut Health Center,
casmith@uchc.edu
|
| Lisa Daley,
The Village for Families and Children Inc,
ldaley@villageforchildren.org
|
| Toral Sanghavi,
The Village for Families and Children Inc,
tsanghavi@villageforchildren.org
|
| Abstract:
Despite an abundance of published community-based studies, and existing guidelines on community-based research and methods, implementing robust study designs in service-orientated community programs can be challenging. This paper describes the authors’ process of changing a federally-funded program and evaluation study from a simple comparison design (of two similar programs in different cities) into a randomized, controlled trial that includes safety exclusions and an opt-in/opt-out choice for participants. Providing detailed description of the more than six month decision-making process, this paper presents the barriers each competing agenda generated, including the program agency’s administration and staff, the funder, and the evaluator. Solutions to these barriers and the reasoning behind each choice are discussed, including mid-course corrections as the new study design was implemented. Recruitment data from the first year of recruitment and data collection are described and analyzed for necessary problem-solving. Finally, suggestions for future evaluation planning and methods are offered.
|
|
A Mixed Methods Approach to Evaluating Individualized Services Planning in the Human Services
|
| Presenter(s):
|
| Michel Lahti,
University of Southern Maine,
mlahti@usm.maine.edu
|
| Abstract:
This paper will present the findings of a four year study of an individualized services planning approach, Family Team Meetings, in child welfare – human services. This paper will explore the extent to which quantitative results can help to describe the implementation and (perhaps) outcomes of Family Team Meetings and will present initial findings on the study of interaction in a team meeting setting. An argument will be made for mixed-methods designs and how evaluations of individualized service planning approaches need to include observation of interaction. Research approaches such as Conversation Analysis focus on the study of interaction and this paper will present the authors’ experiences at learning and applying this technique.
|
|
Impact Evaluation of a National Citizenship Rights Program In Brazil
|
| Presenter(s):
|
| Miguel Fontes,
Johns Hopkins University,
m.fontes@johnsnow.com.br
|
| Fabrízio Pereira,
Brazilian Industrial's Social Services,
fpereira@sesi.org.br
|
| Lorena Vilarins,
Brazilian Industrial's Social Services,
lorena.vilarins@sesi.org.br
|
| Milton Mattos de Souza,
Brazilian Industrial's Social Services,
milton.souza@sesi.org.br
|
| Rodrigo Laro,
John Snow do Brasil,
r.laro@johnsnow.com.br
|
| Abstract:
Objectives: The Brazilian Industrial’s Social Services (SESI) implemented in 2007 a Citizenship Rights Event in 34 municipalities offering low-income populations access to odontological/medical exams, cultural/sports activities, social security/labor identification cards, and professional workshops. The Event reached 1 million individuals in 2007. The objective is to demonstrate the impact of the program.
Methods: An ex-ante/ex-post survey was carried out in November 2007. The sampling error for national representation reached 2.4% (n=1,570). A Healthy Citizenship Scale, ranging from -65 to 65 points, was generated by combining results from and attributing specific weight values to 15 types of services.
Results: Scale reliability reached 0.60 (Cronbach). Differences between ex-ante and ex-post results were found for all scale’s 15 items (p-value<0.05). Ex-ante average of -2.0 increased to 9.9 in ex-post. Increases were observed in all five regions of the country, by gender, age of respondent, and income.
Conclusions: The scale is a reliable evaluation tool.
|
| | |