Structural Effects in Education 119

Structural Effects in Education 119
reminiscent of Union Democracy's discussion of the relation between chapel size and political
participation). They found that the correlation between school status and college plans was
reduced with friends' status (a proxy for peer group relations) held constant, but that between
friends' status and plans was not reduced with school status held constant. This suggested that
interaction among friends was the mediating mechanism, constrained by opportunities for
interaction, rather than by school climate. Campbell and Alexander brought a familiar line of
reasoning into play: individuals' values and attitudes develop through sustained interaction in
social situations with others who are important to them. The mechanism is reciprocal interaction
among friends, rather than the direct normative influence of school climate on conduct
(Alexander & Campbell, 1964).
STRUCTURAL EFFECTS POST-EEO
The body of work just discussed was less about the nature of schools and schooling than about
contextual analysis based primarily on neighborhoods and on schools as climates and on interaction
among peers. EEO (Coleman et al., 1966) brought school characteristics into prominence
because its commission (by the Civil Rights Act of 1964) mandated the examination of
the degree to which White and Black students, as well as other minority group students, attended
schools of comparable quality. It thus entered the realm of public policy controversy
over civil rights, social equality, school desegregation, and educational excellence. This policy
preoccupation would turn out to have a long life and influence how structural effects arguments
colored school reform thinking in the 1980s and 1990s.
Publications appeared in the early 1970s reacting to EEO (Coleman et al., 1966), Mosteller
and Moynihan's On Equality of Educational Opportunity (1972) prominent among them. Several
of its constituent essays employed structural effects arguments; Armor's, for example,
(1972, p. 176) noted problems of interpretation attributable to the ecological fallacy. McDill
and Rigsby's Structure and Process in Secondary Schools (1973) questioned the overreliance
on aggregate socioeconomic status to define school climate, an indirect measure from which
normative influence was inferred (p. 20). Acknowledging that curriculum, teacher quality,
physical facilities, structural features, and socioeconomic composition were proper subjects
of investigation, they "concentrate[d] primarily on differences in what has been called educational
and social climates among schools" (p. 2), measured directly rather than inferred. (For
example, an emphasis [i.e., climate] on "Academic Emulation," based on questions tapping
the academic side of school life, referred to a "general academic and intellectual tone" [p. 38];
similar concepts applied to five other dimensions of climate.) This conception was based on
their belief that structural characteristics "are simply too gross" (p. 118) to explain academic
performance, and on their earlier findings showing that the effects of nonclimate (structural)
school characteristics were small (McDill, Rigsby, & Myers, 1969). They tested their conception
of climate against the indirect form (inferred from aggregate school characteristics) and
assessed the results of earlier studies of whether peer-group processes represented mechanisms
by which climate influenced achievement and college plans. McDill and Rigsby reinforced
the normative meaning of school climate.
Others (Brookover, Beady, Flood, Schweitzer, & Wisenbaker, 1979; Rutter, Maughan,
Mortimore, & Ouston, 1979) expressed disbelief that the effects of school differences on
achievement were as small as EEO (Coleman et al., 1966) reported. Brookover's methodology
(like McDill and Rigsby's) used teachers, students, and principals as informants, summarizing
their perceptions of norms and structure at the school level, and summed multiple indi





120 Robert Dreeben
ces of climate and structure into global naeasures. The result was to obscure how the components
of school organization fit together and which components of the summary singly or in
combination affected the outcomes (as in Michael, 1961). The mechanisms by which school
organization and climate influenced outcomes (themselves aggregated) were indeterminate
because there was no way to ascertain which students, whose achievement was expressed as a
school rate, were subject to climate effects.
A study of 12 London secondary schools (Rutter et al., 1979) argued that EEO (Coleman,
et al., 1966) was
quite unable to consider wliether children were influenced by differences in things such as the style
or quality of teaching, the types of teacher-child interaction in the classroom, the overall social
climate of the school, or its characteristics and qualities as a social organisation, (pp. 5, 31)
Rutter accordingly examined whether children's school experiences made a difference in
outcomes. Acknowledging the importance of classroom instruction, he considered not the
"details of curriculum," but "the broader curriculum of the social environment within which
lesson teaching takes place" (p. 54). He gathered observations of classrooms, of school entry
patterns, and of student placements in different kinds of courses and aggregated them at the
school level to derive an abstract, global notion of school structure. This meant conceptually
subordinating the direct influences of classroom organization, teaching practices, and social
interaction (pp. 62-64) on students' experience. He maintained that the importance of school
events and behavior was their contribution to "the establishment of an ethos which would
enable all those in the school to function well" (p. 56; my italics).
This analysis distinguished several kinds of school process (e.g., academic emphasis,
teacher action, rewards and punishments, etc.), each represented by a set of items and correlated
with four school-level measures of outcome: attendance, student misbehavior, academic
attainment, and delinquency. The numerous correlations differed widely in magnitude, which
led to speculation about whether "process" should be construed as a multitude of unstable,
substitutable effects or as a coherent phenomenon, identified by combining those items into a
single index, strongly related to the outcomes. Rutter (1979) concluded:
[T]he association between combined measure of overall school process and each of the measures of
outcome was much stronger than any of the associations with individual process variables" [i.e.,
taken one at a time] . . . The implication is that the individual actions or measures may combine to
create a particular ethos, or set of values, attitudes and behaviours which will become characteristic
of the school as a whole, (p. 179; Rutter et al.'s italics)
The correlation of the global scale with the outcomes was strong. However, difficulties
of interpretation arise, attributable to the combining of variables and to the familiar uncertainties
of ecological correlation. Despite its critical posture toward EEO, the study employed the
same form of analysis: it summed elements of school structure and operation and related them
to outcomes, thereby sacrificing the portrayal of internal school life and the mechanisms by
which ethos might affect outcomes.
The prevailing school effects formulation had by the late 1970s thoroughly absorbed
Wilson's version of Durkheim's perspective. Though Blau and Wilson employed the same
analysis, Wilson's conception, not Blau's, gained currency in its own right and also indirectly
through Coleman's. It is also interesting that Durkheim's, Merton's, and Blau's concerns with
deviance dropped from sight, as if conventionally salutary academic values could not have
unanticipated consequences of a negative sort (Stinchcombe, 1964). Sociologists of education
could have drawn from Blau as well as from other contributors to the field of organizations,
such as Etzioni (1975), Homans (1950, 1974), Perrow (1970, 1972), and Stinchcombe
(1959, 1965). The field might then have developed in quite different directions.