Theory-Based Data Analysis for the Social Sciences Paperback - 2002

The advent of complex and powerful computer-generated statistical models has greatly eroded the former prominence of social theory in data analysis, replacing it with an emphasis on statistical technique. To correct this trend, Carol S. Aneshensel presents a method for bringing data analysis and statistical technique into line with theory. She approaches this task by first providing an overview that explains the connection between data analysis, statistical technique, and theory. This section includes a description of the elaboration model for analyzing the empirical association between two variables by adding a "third variable" to the analysis.

Aneshensel then introduces a new concept into this model, the focal relationship. This concept is the one cause-and-effect type of relationship of primary significance that is indispensable to the entire theory. Building upon the focal relationship as the cornerstone for all subsequent analysis, two analytic strategies are developed to establish its internal validity:

• An exclusionary strategy to eliminate alternative explanations for the focal relationship using control and other independent variables to rule out spuriousness and redundancy.
• An inclusive strategy to demonstrate that the focal relationship fits within an interconnected set of relationships predicted by theory using antecedent, intervening and consequent variables.

Using real examples of social research, the author demonstrates the use of this approach for two common forms of analysis, multiple linear regression and logistic regression.

Whether learning data analysis for the first time or adding new techniques to your repertoire, this book provides an excellent basis for theory-based data analysis.