Background
Personal control [
1‐
5], sometimes referred to as locus of control [
6], sense of control [
7,
8], sense of coherence [
9], mastery [
10], or self-efficacy [
11] has played an important role in the study of health and health behaviour [
12‐
17], because:
“…compared to people who feel powerless to control their lives, people with a sense of control know more about health, they are more likely to initiate preventive behaviours like quitting smoking, exercising, or maintaining normal weight, and in consequence, they have better self-rated health, fewer illnesses, and lower rates of mortality” [15; p. 66].
As a result, personal control among older adults is viewed as a marker of successful aging, and further study of this construct is warranted [
14,
18].
A frequently observed association with personal control involves race [
15]. African Americans generally report lower levels of personal control than non-Hispanic Whites or other minorities [
15,
19,
20]. In most studies, however, race is simply used as a covariate, with primary attention focused on the effects of age, education, and health. Indeed, only two studies [
20,
21] have modelled personal control separately for older non-Hispanic Whites and African Americans. Thus, much remains to be known racial differences in personal control, especially since the dominant view is that personal control becomes relatively stable by early adulthood, with declines commencing after age 50 due to the progressively less favourable ratio of gains to losses in later life [
9,
13,
14,
22,
23]. Therefore, any social processes that adversely affected school-aged African American children may have played an important role in the development of their personal control.
Along those lines, Buck [
24] recently argued that desegregation left an ironic legacy for African Americans. He claimed that desegregation:
“…undermined one of the traditional centres of the black community: the school. In the segregated schools, black children had consistently seen other blacks succeeding in the academic world. The authority figures and role models—that is, the teachers and principals—were virtually always black. And the best students in black schools were black as well. That ended with desegregation” [24; pp. 3–4].
Based on Buck’s argument, we hypothesized that the sense of control among African Americans adults who attended segregated schools as children would be higher than that among African American adults who attended desegregated schools. Moreover, given the extensive literature linking the sense of control with better health outcomes, we hypothesized that the higher sense of control among African Americans adults who attended segregated schools as children would be independently associated with better physical performance in later life. Furthermore, we hypothesized that these effects were independent of potential confounders such as those risk factors and covariates traditionally found to be associated with the sense of control (e.g., demographic factors, socioeconomic status, and self-rated health), as well as regional differences in where the childhood schooling occurred, and current racial attitudes and beliefs, and resilience. Finally, we hypothesized that the association between childhood school segregation and physical performance in later life would be partially or fully mediated by the sense of control. Note that while our hypotheses and our evaluation of them flow directly from Buck’s argument, they do not address the etiologic pathways that he postulated.
Essentially, the etiological mechanism that Buck [
24] proposed was that desegregation-induced erosions of traditional “black community” schools may have reduced or eliminated three protective effects for the children who attended them—the crystallization of racial solidarity, the advantage of having same-race students as role models and same-race teachers and principals as authority role models, and not having been discriminated against or antagonized at school on the basis of race [
25,
26]. Racial solidarity would have arisen from the collective identification associated with segregated schooling where all of one’s peer models, including the best students, star athletes, and most popular individuals were African American. African American students in segregated schools would also have had more exposure to role models of successful achievement reflected by their African American teachers and principals. And by attending segregated schools, African American students would not have been discriminated against or antagonized at school due to their race. Moreover, they would have been less likely to be discouraged due to any academic failures, and they would be more likely to have received greater levels of support encouraging their success. All of these are well-documented benefits for African Americans attending Historically Black Colleges and Universities (HBCUs) [
27‐
29].
Buck’s [
24] argument also resonates with Allport’s contact hypothesis [
30] for optimizing intergroup relations, which provides a conceptual backdrop for understanding why U.S. school desegregation did not equate with integration [
31,
32]. That is, subsequent to the 1954
Brown v Board of Education U.S. Supreme Court ruling, a cascade of federal policies were implemented that were designed to better integrate African Americans into American society. Essentially, those policies assumed that desegregation would provide equal access to quality education, which had long been considered the springboard for achieving the American dream. As Pettigrew [
33] noted, however, for desegregated schooling to achieve these goals, the four conditions necessary for the contact hypothesis—“equal group status within the situation, common goals, intergroup cooperation, and authority support”—were not met. Thus, while schools were legally desegregated, African American students were not functionally integrated into their new schools, and not all of the benefits that those schools should have provided African American students were received.
Although a growing number of studies have considered the effects of school segregation on health outcomes for children and young adults [
34‐
38], no studies have focused on this association among older adults. Moreover, in a review of 39 studies on the association between segregation and health outcomes, 38 were found to have relied solely on measures of residential (rather than school) segregation [
39], with the only study using school segregation having combined it into a composite with neighbourhood and church segregation [
40]. Nonetheless, several studies have reported protective effects of living in clustered African American neighbourhoods on health outcomes [
41]. Thus, the hypotheses for this study were consistent with prior reports, although our hypotheses focused on the independent relationship between school segregation in childhood (as a marker of exposure to the rich benefits of the “black community” school) and the sense of control in later life, and the sense of control in later life and physical performance with these constructs being measured at the individual rather than the community level. Measuring school segregation at the individual level enhances the validity of correct exposure classification by focusing on whether each study participant actually attended segregated schools rather than merely lived in an area where segregated schools were available within the community. Moreover, school segregation and residential segregation are rather different phenomena that would not necessarily have the same effects [
39‐
41].
Discussion
This study was driven by four hypotheses, all of which were logical extensions of Buck’s [
24] work on the unintended consequences of desegregation leading to the erosion of the traditional “black community” school. The first hypothesis was that African Americans who attended segregated schools for their 1
st-to-12
th grade education would have higher sense of control scores [
13,
54‐
57] (i.e., more claiming and less denying responsibility for both the good and bad things) in later life. The second hypothesis was that higher sense of control scores in later life would be related to better physical performance. The third hypothesis was that the effects of childhood school segregation on the sense of control in later life, and the effects of the sense of control on physical performance would hold even after adjustment for all of the traditional covariates and potential confounders, racial attitudes and beliefs, and resilience. The fourth hypothesis was that the beneficial association between childhood school segregation and physical performance in later life would be partially or fully mediated by the sense of control. Sense of control [
13,
54‐
57], school segregation, the traditional covariates and potential confounders [
13,
19,
58], racial attitudes and beliefs [
59,
60], resilience [
62], and physical performance [
42,
63‐
69] were all measured at the 2010 in-home follow-up for 582 of the 998 original AAH participants. Propensity score re-weighting methods [
49‐
52] were used to adjust for potential attrition bias.
In terms of the first and third hypotheses, we found that the unadjusted effect of having attended segregated schools for at least half of one’s 1
st-to-12
th grade education led to scores on the sense of control that were .990 points higher (
p < .01, standardized effect size = .25) than those who did not attend segregated schools, indicating more claiming rather than denying control over both the good and bad outcomes in life. This significant association held even after adjusting for all of the traditional covariates and potential confounders [
13,
19,
58], racial attitudes and beliefs [
59,
60], and resilience [
62]. In the final model, African Americans who had received half or more of their 1
st-to-12
th grade education in segregated schools had sense of control scores that were .886 points higher (
p ≤ .01) than those who did not attend segregated schools, representing a standardized effect size of .22.
Sensitivity analyses that looked at the effects of childhood segregated schooling on each of the four-item component scores [
13,
55‐
57] of the sense of control—one reflecting claiming control and the other reflecting denying control—led to a better characterization of the observed association. Those results showed that the effect of childhood school segregation on the sense of control in later life was almost entirely derived from increased disagreement with the four items reflecting denying control (and therefore reflective of having a greater sense of control). Indeed, the unstandardized regression coefficient (
b) for having attended segregated schools for at least half of one’s 1
st-to-12
th grade education obtained from the fifth (final) model was .867 (
p = .003) for the denying control component, which was almost as large as the effect on the overall sense control scale (where
b = .886,
p < .01). In contrast, there was no effect of segregated schooling on the claiming control component (
b = .020,
p = .932).
Recent work by Kraus et al. [
70] provides insight into why denying control may have been most affected by childhood school segregation. Building on classic social science theory about why social class plays such a pervasive role in everyday life, they present a social cognitive theory as the underlying process by which social class creates clear cultural differences between the rich and the poor. The core of Kraus et al.’s argument is that:
"“…social class contexts elicit reliable social cognitive patterns among lower-class individuals—characterized by a contextual, externally-oriented cognitive and relational orientation to the world—and upper-class individuals—characterized by a solipsistic, individualistic cognitive and relational orientation to the world.” (p. 546)"
Although the contrast between the poor and the rich is not directly relevant here, the notion of how that contrast leads to the development of differential life orientations and outlooks is. Kraus et al. argue that:
"“Diminished resources and lower rank create contexts that constrain social outcomes for lower-class individuals and enhance contextualist tendencies—that is, a focus on external, uncontrollable social forces and other individuals who influence one's life outcomes.” (p. 546)"
African American children attending desegregated schools would have faced diminished educational resources and discrimination based on their ascribed racial status, and this may have led to their greater focus on and acceptance of external, uncontrollable forces in their lives. Thus, Kraus et al.’s explanation is entirely consistent with and provides the context for our findings. Indeed, if one were to substitute “childhood desegregated schooling” for “lower social class” and “childhood segregated schooling” for “upper social class” in their second and fifth hypotheses, our results would be exactly what they predicted.
Furthermore, our results are also consistent with, but clearly do
not demonstrate Buck’s [
24] main argument concerning the major unintended consequence of desegregation—the erosion of the traditional “black community” school. Buck argued that the erosion of the traditional “black community” school may have led to a gradual but significant cultural shift for subsequent generations of African Americans attending desegregated schools. Building on his argument, we hypothesized that this may specifically have led African Americans attending desegregated schools to be more likely to deny having any control over either the good or bad things happening in their lives. Buck also argued that over time, such cultural shifts may also have led to the emergence in the “black community” of the sentiment that African Americans with a stronger sense of responsibility for both the good and bad things occurring in their lives were “acting white.”
In terms of the second and third hypotheses, we found significant unadjusted associations between sense of control [
13,
55‐
57] scores and seven of the eight physical performance tests [
38,
63‐
69]. Higher scores on the sense of control were associated with significantly lower systolic blood pressure, and with significantly better grip strength, peak expiratory flow, gait speed, chair stands, balance tests, and the SPPB (all
p < .001). Furthermore, adjustment for the two childhood school segregation measures, the traditional covariates and potential confounders [
13,
19,
58], racial attitudes and beliefs [
55,
56], and resilience [
62], did not appreciably alter the beneficial independent effects of the sense of control scores on systolic blood pressure, grip strength, peak expiratory flow, chair stands, balance tests, and the SPPB. While these effects are modest, they are important. Indeed, they indicate that a difference of just four points on the sense of control score would be associated with a one-point improvement on the SPPB, which would be enough to move the average AAH participant (mean = 7.9) above the cut-off for increased risk of future mobility impairment, poorer prognosis and recovery from newly diagnosed disease, and death [
64‐
69].
In terms of the fourth hypothesis, we found statistically significant beneficial crude associations between childhood school segregation and physical performance for peak expiratory flow, gait speed, the balance tests, and the SPPB. The associations with peak expiratory flow and the balance tests were partially mediated by adjusting for the sense of control, and the associations with gait speed and the SPPB were fully mediated by adjusting for the sense of control. Thus, it would appear that childhood school segregation affected the sense of control in adult life, which resulted in better physical performance, suggesting that the sense of control plays an intervening variable role in the underlying causal sequence.
Although no other studies have focused on the association of segregated schooling and the sense of control [
13,
55‐
57] among older adults, recent rigorous studies of the effects that school segregation has on health outcomes for children and younger adults provide context and support for the results reported here [
25‐
29]. In particular, data from the National Longitudinal Study of Adolescent Health (NLSAH) found that risky health behaviours such as smoking and drinking were lower for African American women in more segregated schools than for Non-Hispanic White women [
34]. The two explanations given for this finding were that (a) because Non-Hispanic Whites had higher smoking and drinking rates, African American women in segregated schools may have been protected by the absence of cross-race friendships, and that (b) African American women in segregated schools may have engaged in other resiliency strategies [
34]. Similarly protective functions associated with segregated schooling were found when the NLSAH was used to study depressive symptoms, which led the same authors to conclude that “attending predominantly-minority schools may buffer African American students from discrimination and increase their school attachment, which may reduce their risk of experiencing depressive and somatic symptoms” [35; p. 1873]. These explanations are consistent with the etiologic mechanism proposed by Buck [
24] that drove our hypotheses extending his argument. It is important to note, however, that the NLSAH and the AAH reflect rather distinct cohorts of African Americans that lived through quite different historical and cultural periods in the United States, and are at rather different stages in the life course. Nonetheless, building on life course theory [
71], both studies focus on the accumulative advantage [
72] of segregated schooling for African Americans, at least in terms of less risky health behaviours and fewer depressive symptoms (in the NLSAH), and stronger sense of control and better physical performance (in the AAH).
It is important to note that this study is not without limitations. Three of these are most salient. The first involves the geographic generalizability of these associations beyond St. Louis, Missouri, and the extent to which these associations may be time-bound to the AAH cohort (born in 1936–1950). Therefore, further research on desegregation consequences from other communities and regions, as well as from other birth cohorts at different life stages is needed and should be encouraged. The second limitation is that the sense of control [
13,
55‐
57], school segregation, traditional covariates and potential confounders [
13,
19,
58], racial attitudes and beliefs [
59,
60], resilience [
62], and physical performance [
42,
63‐
69] were all measured at the 2010 in-home follow-up assessment. That makes ours a cross-sectional study in which we can only demonstrate associative relationships rather than causal effects. The third limitation is that our focal measures were retrospective self-perceptions of segregated schooling and the perceived extent of that exposure. As such, they are not objective, independent assessments of school segregation, and are subject to traditional retrospective reporting biases.
Conclusions
In conclusion, we have shown that African Americans who attended segregated schools as children for at least half of their 1
st-to-12
th grade educations were significantly more likely to claim rather than deny control over both the good and bad outcomes in their later lives, when compared to their counterparts who did not attend segregated schools as children. Moreover, higher sense of control scores in later life were associated with significantly lower systolic blood pressure, and with significantly better grip strength, peak expiratory flow, gait speed, chair stands, balance tests, and the SPPB in later life. While the etiologic mechanism proposed by Buck [
24] could not be evaluated in these data, our results are consistent with the essence of his argument.
That said, under no circumstances should our results be misconstrued as either evidence for or an argument that desegregation was bad social policy. Rather, our results suggest that (a) while desegregation may be a necessary condition for integration, it is not a sufficient condition, and (b) desegregation may have had an unfortunate, unintended consequence that led to lower sense of control in later life, with that lower sense of control associated with poorer physical performance. Assuming that our results are replicated in analyses from other studies, they have important policy and public health implications. To overcome these unfortunate side effects of desegregated schooling, greater emphasis may need to be placed on helping today’s desegregated schools (a) develop more and better same-race peer and authority role models for African American students, (b) emphasize that successful achievement avenues are not only possible but readily available, (c) minimize racial discrimination, and (d) promote cultural solidarity. Although achieving these goals will be difficult, the vibrancy of our society and civilization makes finding ways to accomplish them essential.
Competing interests
The authors state that they have no competing interests.
Authors’ contributions
FDW conceived of and planned the analyses reported here, performed all data analysis and reanalysis, and wrote and revised the manuscript. DKM conceived the overall plan for the AAH study, and contributed to the review of the analytic approach and revision of the manuscript. TKM contributed to data preparation, review of the analytic approach, and revision of the manuscript. JPM, MS, and EMA contributed to the review of the analytic approach and revision of the manuscript. All authors read and approved the final manuscript.