It sometimes seems as if school segregation is one of those topics that is always “in fashion” among education policy commenters and journalists. This is a good thing, as educational segregation, and the residential segregation underlying it, are among the most important symptoms and causes of unequal opportunity in the U.S.
Yet the discussion and coverage of school segregation, while generally quite good, sometimes suffers from a failure to make clear a few very important distinctions or details, and it may be worthwhile laying these out in one place. None of the three discussed below are novel or technical, nor do they represent a comprehensive list of all the methodological and theoretical issues surrounding segregation (of any kind).
They are, rather, just details that should, I would argue, be spelled out clearly in any discussion of this important issue.
The first thing that should always be specified is the “type” of segregation – that is, which groups are being analyzed. This is a very obvious point, and probably goes without saying, but it bears mentioning anyway. The most common groups in the education context are those defined by income (usually using subsidized lunch eligibility as a rough proxy) and race and ethnicity. Schools, on the whole, are segregated by both race and ethnicity and income, and the two are interrelated, but the levels and trends can be different (e.g., Owens et al. 2014).
Moreover, particularly when it comes to segregation by race and ethnicity, the most common measures can be applied to different combinations of groups. For example, one can measure segregation between individual groups (e.g., Black from white students), or combinations of groups (e.g., minorities from white students). It is important to make clear such specifications and, perhaps, to note that different combinations/groups can yield different results (e.g., Reardon et al. 2000). This is particularly salient given the increasingly multiracial composition of U.S. public school students.
A second, and related distinction that should be highlighted in any discussion of school segregation is the type of segregation measure used. There are many different ways to measure segregation, school and otherwise. The two most common approaches are:
- Exposure: This is a measure essentially of contact or interaction between groups. For example, one might calculate the percentage of the typical higher income student’s peers who are lower income. These indicators are sensitive to compositional change (e.g., a change in the number of lower income students);
- Evenness: This type of measure focuses not on how many members of a given group there are, but rather how evenly they are distributed (e.g., between schools or districts). If, for instance, every school in a given district has roughly the same proportion of lower income students common evenness measures would assess that district as highly integrated. Note that, unlike exposure measures, evenness is not sensitive to composition – even a district with few lower income students can be interpreted as integrated if those students are evenly distributed among schools.
These and other types of measures in many respects represent different conceptualizations of segregation (see Massey and Denton 1988). For instance, exposure measures concentrate on the peer effects of segregation, whereas evenness is more about whether segregation causes some students to experience different situations than others (Reardon and Owens 2014).
And these different approaches can also yield very different results. For example, depending on whether one uses evenness or exposure indicators, one might conclude that school segregation by race and ethnicity is either increasing or stable over time (see, for example, Orfield et al. 2014 and Reardon and Owens 2014).
There is, as a result, a longstanding debate between researchers about which measures are appropriate in which contexts, but for our purposes here, suffice it to say that any discussion of a segregation analysis must pay particular attention to which measures are used, and, as a rule, any analysis that relies on just one measure may lead to incomplete conclusions (for more discussion, see Wysienska-Di Carlo et al. ).
The third important distinction is also the most frequently obscured, and that is the level or unit of segregation. Most often, this distinction refers to within- versus between district segregation. Put simply, districts are (or are not) segregated “internally” (i.e., between schools), and their students are (or are not) segregated from their peers in other districts (within, say, a metropolitan area).
For example, District X might have a 50/50 split between lower- and higher income students in all of its schools (i.e., perfect integration), but this district might be the only one in the metropolitan area that serves any lower income students, which means that all the lower income students in the area are completely segregated into a single district. In this sense, within- and between district segregation both "contribute" to total segregation, and, once again, this relationship can vary over time (Reardon et al. 2000).
Segregation within districts is of particular note in that most legal and policy efforts to desegregate schools concentrate within districts. Yet, to the degree segregation reflects and perpetuates inequality of opportunity, and insofar as such inequality can be just as, if not more, pronounced between as it is within districts, integration within a given district, while positive and important, may be only part of the story.