Establishing Equivalence At Different Stages of the Research Process

BACKGROUND:  Diversity exists in many forms among those participating in social work research.  As social work researchers, we often finds ourselves concerned with how observable forms of diversity, for example, race, ethnicity, sexual orientation and gender, shape such things as our understanding of the social phenomena of child maltreatment  or the effectiveness of an social intervention.  Understanding the impact of human diversity on research findings requires that we use methods that allow for the appropriate comparison of findings across groups.   Appropriate comparisons begin with using methods that allow for the determination of the extent to which our research design functions in the same matter across the diversity of study participants.   

PURPOSE:  The purpose of this workshop is to illustrate the use of methods during each phase of the research process that can ensure that studies involving diverse groups reflect true group differences and not flaws in the research design. 

LEARNING OBJECTIVES: Workshop participants will learn about the ways in which a study can be designed to establish research-design equivalence among diverse groups.  The establishment of equivalence during the problem formulation, research design, sampling, measurement selection, data collection, and data analysis phases will be addressed.   Workshop participants will learn about the major threats to equivalency that occurs during each phase of the research process.  For example, during the problem formulation phase methods to ensuring construct equivalency will be discussed.   The participants will learn that there are three levels of construct equivalency (conceptual, functional, and scalar) that must be addressed during the problem formulation phase of the study’s design.  During the research design phase participants will learn about how to address threats to causal inferences in comparative research that are important to consider when conducting research with diverse groups.  These threats are statistical-conclusion validity, internal validity, and construct validity. 

INSTRUCTIONAL STRATEGIES:  Illustrations from published and hypothetical studies will be presented.  Results from the analysis of data from national longitudinal studies (Add Health and National Educational Longitudinal Study) will be used to illustrate the use of multiple group confirmatory factor analysis to establish measurement equivalency.  The workshop will demonstrate the use of Latent Class Analysis to identify latent diversity of groups, which can be used to identify latent within group diversity.