Dubrow, Joshua Kjerulf. 2008. “How Can We Account for Intersectionality in Quantitative Analysis of Survey Data? Empirical Illustration of Central and Eastern Europe.” ASK: Society, Research, Methods 17: 85-102.
Approaches to Understanding Demographics’ Influence
In studying the influence of demographic variables on attitudes and behavior, there are two major approaches, with some important internal divisions. The first is unitary, or additive, approach, and the second can be called the multiplicative approach1. I present the essence of these approaches, referring the reader to the literature for the details (Hancock 2007; McCall 2005).
NOTE: As researchers use various names for these approaches, I chose names based on what I considered the most intellectually appealing. For unitary, I use Hancock’s 2007 term, and Weldon (2006: 240) suggests the term “monism” for this concept. For variants of intersectionality, I refer to Leslie McCall’s (2005) terms: anticategorical, intracategorical and intercategorical, the last reduced to “categorical approach,” which I refer to as the multiplicative approach, as multiplicative interaction terms are the key element. Anticategorical and intracategorical have nuances that separate them, but the differences are not sufficiently pronounced to suggest fundamentally different statistical approaches. Weldon (2006) refers to McCall’s anticategorical and intracategorical as the Intersectionality-Only model and the categorical approach as the Separable-But-Mutually-Reinforcing model.
Unitary, or Additive, Approach
Most researchers employ the unitary approach. Its main theoretical assumptions are that (a) demographic categories have social properties that are distinct from the other characteristics of individuals, and (b) a separate category could be a the best predictor of the dependent variable (Hancock 2007; Weldon 2006). In the unitary approach, it is assumed that demographic variables have additive effects. For example, the joint effect of being a woman, belonging to an ethnic minority, and representing disadvantaged social class is seen as a sum of the effects of these three demographic variables.
Multiplicative approach is also called categorical (McCall 2005). In intersectionality theory, the influence of demographics on a social outcome is conditional on the intersections of the demographic categories. Thus, interaction terms of categories identifying respondents are advocated for in this approach. In this more relaxed version of intersectionality theory, it is assumed that the constituent elements of intersections have valid social meaning. Moreover, this approach “begins with an analysis of the elements [of the intersections] first because each of these is a sizable project in its own right” [emphasis mine] (McCall 2005: 1787). The constituent elements of intersections are there to provide context for the intersections themselves (Weldon 2006).
Versions of intersectionality theory range from more relaxed to more strict. In the strictest version, it is assumed that “social life is considered too irreducibly complex… to make fixed categories anything but simplifying social fictions…” (McCall 2005: 1773). Gender, ethnicity, and class are inseparable as each has no valid social meaning on their own. This “anticategorical” approach assumes that each constituent category of the intersection has no autonomous effects (McCall 2005; Weldon 2006: 240). While I focus on the multiplicative approach, I briefly discuss methodological implications of the anticategorical approach.