RESEARCH METHODS

点击翻译按钮获取中文摘要

Methods Core activities will address specific challenges and opportunities associated with: (a) multilevel statistical modeling, (b) measurement methods, (c) classroom observational assessment systems, and (d) organizational construct measurement and analysis. In terms of the multi-level statistical modeling, and with the leadership of Dr. Gibbons, we will explore the development of methods for computing power that consider three levels of data. This will allow greater specificity with which to design future studies that involve multiple time points and clusters of individuals within multiple settings. In addition, we have the unique opportunity to test the new SuperMix program developed by Hedeker and Gibbons, which is the focus of SBIR award N44MH32056 through Scientific Software International (SSI), the company that distributes LISREL and HIM among other advanced statistical programs. SuperMix is a general program for performing 2- and 3-level mixed-effects linear and nonlinear regression models for analysis of continuous, binary, ordinal, nominal, count, and time-to-event response data. As a part of the Methods Core, Drs. Hedeker and Gibbons have invited us to be a beta-test site for the program and to use the SuperMix program for the proposed 3-level analyses for both the parent grant and the pilot studies in this Developing Center. Dr. Schoenwald also will coordinate the development of innovative mixed method studies toward understanding in greater depth and breadth the psychosocial and instructional environment of urban classrooms. This integration of qualitative and quantitative approaches begins with knowledge gained in the PRC pilots and Methods Core activities in which multiple quantitative methods (multi-level statistical modeling, classroom observational assessment systems, and organizational surveys) are implemented and adapted. Dr. Hopper will provide consultation to build understanding of and skill in qualitative inquiry among our research team, as well as to lead discussions around the use and adaptation of qualitative inquiry in urban school settings. In addition, two lead researchers in the proposed Developing Center have current or pending research training grants with consultation in qualitative methodologies (Dr. Birman with Dr. Tanya Luhrman, and Dr. Frazier with Dr. Mary Pattillo). Given the rich array of quantitative and qualitative skills in our network of researchers, and the complex and dynamic context of urban poor schools, we will have the opportunity to work in mixed method teams to develop writing and research projects that link qualitative inquiry to primarily quantitative studies - including, but not limited to, the Principle Research Core pilots - toward a more nuanced and inclusive understanding of the nature of, and factors that impact on, the climate and interactions within the urban school classroom (Shepard, Orsi, Mahon, & Carroll, 2002). With respect to the observational coding of classroom processes there are two primary challenges. These stem from the fact that, although there is strong evidence that the processes measured by the CLASS and the ICOS are valid indicators of classroom processes that are important in predicting students' social and academic development, and teacher effectiveness, across a broad spectrum of socioeconomic and ethnic groups (Hamre & Pianta, 2005; NICHD ECCRN, 2002; 2003; Schafferet al., 1994), there has not been a distinct focus on urban, high poverty settings. Thus, one challenge is to examine the degree to which there may be untapped dimensions of classroom process in these settings that are incrementally important predictors of student functioning. Second, although classroom ratings tend to be highly stable across activities and time in more global samples (NICHD ECCRN, 2002), classroom processes may be more contextdependent in urban poor schools in which there may be greater fluctuation in class composition, attendance, and stress. As a part of this Methods Core, we will systematically assess the degree to which the assumption of underlying stability in processes is met in our sample of high poverty, urban classrooms. Third, with the leadership of Drs. Pianta and Schaeffer, we will compare data collected across the two observational systems, one focusing on relational constructs (CLASS) and the other on instructional constructs (ICOS), in order to assess the extent to which these constructs are independent, interdependent, and overlapping toward the prediction of student outcomes. The adaptations in measurement dimensions and observational methodologies that result from addressing these challenges will in turn inform our conceptualization and measurement of classroom practices that data suggest are critical to student outcomes. Finally, there are particular challenges and opportunities afforded by the application of the organizational constructs and measures to the unique context of urban schools and classrooms. As mentioned above, two independent literatures - one from industry, business, and human services (e.g., Glisson, 2002), and the other from education and psychology (e.g., Comer etal., 1996; Rutter, 1983; Trickett & Moos, 1973)- have developed distinct definitions and methodologies to understand classroom climate. The systematic application of the first approach to the urban school-community context provides the opportunity to consider the unique nature of the school organization and classroom work units in which both the teachers and the students can be seen as "workers" with distinct and important contributions to the classroom and school context. The challenge will be to coordinate expertise in child development, urban schools, and organizational theory toward the creation of meaningful measures and understandings of the urban classroom climate and culture.