| This dissertation attempts to synthesize and advance the ways in which the concept of social class, as an important component of social structure, is applied by social scientists in empirical research. A historical review of empirically grounded studies yields two ongoing, interrelated debates about the nature and existence of social classes. Are they continuous or discrete in nature, and are they subjective or objective in basis? Two divergent empirical approaches---attribute and structural approaches, are also implicated across three general bodies of research. These approaches are critically evaluated, with reference to their empirical applications and resulting contributions to our debates over social class.; A research method incorporating a unique combination of these approaches, that could significantly advance our open debates, was proposed in 1938. A critical obstacle to the implementation of this longstanding research method has been our inability to identify or collect data sets that are of sufficient scale and scope to systematically test for the existence of relational bases for social classes. Additionally, technological advances in computing power required for the development and implementation of the analytical tools that could enable identification of relationally-based patterns in these large scale data sets have only relatively recently been realized.; Using a consensus in the literature, kinship and marriage relations and occupation are selected as the relation and attribute for this study. The implications of longitudinal marriage choices within a given population for social class formation, and a method for detecting emergent patterns in these marriages choices that implicate group boundaries are discussed.; Application of the proposed research method to a data set from 19 th century London produces social groups that are not only substantively consistent with a social class-based stratification system, but further, that can be mapped to commonly used social class categories. Our overall findings support a discrete and an objective view of social classes. Limitations of this study and units of analysis are discussed. Identification of additional data sets containing both attributes and relations is necessary for further research, but sampling issues require careful consideration. |
