An empirical examination of racial discrimination cases, The

Supreme Court and the protection of minority rights: An empirical examination of racial discrimination cases, The

Romero, Francine Sanders

This inquiry provides a basic assessment of the impact of three potential determinants of racial discrimination cases in the U.S. Supreme Court since 1954. The research design provides two improved methods of explicating this issue. First, the model allows for a comparison of basic Hamiltonian institutionalism (i.e., the bulwark thesis), majoritarianism, and attitudinalism in a single test, as opposed to previous studies that tended to examine only two theoretical approaches at a time. Second, the majoritarian approach is given more careful consideration through the use of theoretical and empirical evidence, which allows the subtleties of public opinion in this area to be assessed. The findings show some support for the basic bulwark prediction over majoritarianismdecisions fail to reflect majority opinion trends. The bulwark thesis fails to receive full support, however, since the ideologies of the Justices also display a significant influence on outcomes.


The assessment of the determinants of U.S. Supreme Court decisionmaking remains an intensely controversial aspect of judicial studies. Advocates of several broad approaches continue to debate which is the principal impetus of outcomes both in general and in specific legal fields.1 My inquiry offers one perspective to help untangle this controversy in the domain of racial discrimination cases in the post-Brown v. Board of Education (1954) period. Although the findings will not end the long-standing debate over what determines decisional outcomes, they do provide a clarified picture of the racial discrimination subfield and a suggestion for studying other specialized areas.

My basic approach and specific research design are premised on the assertion that a clear understanding of decisionmaking is obscured by previous studies (both general and particularized) that tend to inflate the influence of majority preferences (thus discrediting institutionalism to an unwarranted extent) and also fail to provide a full account by focusing only on two competing explanations at a time. The strategy for systematically interpreting outcomes in this area rests on two novel tactics. The first tactic is to expand consideration of the potential role played by majority opinion. (I furnish a complete description and justification of this approach later.) This tactic offers an improved test of the majoritarian thesis. The second approach of this inquiry allows for the explication of three potential determinants of decisionsthe rules and structure of the institution itself, majority public preference, and the ideological predilections of the Justices. It thus provides a core comparison of these broad categories rather bluntly defined, as opposed to an exhaustive assessment of all potentially meaningful determinants. Since such a basic measure of outcomes in this field has yet to be undertaken, however, this is a necessary first step.

The results of this investigation of constitutional challenges to racial discrimination suggest that even though the Supreme Court is insulated from majority preferences, its decisions are influenced by Justices’ ideological leanings. More specifically, although white Americans (who in this area represent the majority, as opposed to the African American minority) are much more amenable to government action designed to end blatantly discriminatory laws and practices (de jure discrimination) than to the eradication of entrenched patterns of inequity (de facto discrimination), this distinction is not reflected in the decision record. Whether a case represents a challenge to de jure or de facto discrimination does not significantly influence its outcome. Furthermore, fluctuations in the general ideological temper of the nation also fail to affect rulings. However, although distancing itself from majority influences, the Court is not consistently protective of minority rights. The ideological composition of the Court displays a significant influence on verdicts; thus, when the Court is relatively more conservative, cases are less likely to be decided in the minority interest.

Literature Review and Critique: The Bulwark and Alternative Theories

Below, I outline the very basic tenets of three standard explanations of judicial decisionmaking. This review of the literature does not delve into the many variations of each of these approaches, nor is this the only way to sort these diverse schools of thought. For example, one alternative means of grouping theories is through the legal and extralegal dichotomy, in which the former represents assertions that the Constitution and precedent determine decisions, while the latter includes all other explanations, including majoritarianism and attitudinalism (George & Epstein 1992). The approach I use here is simply more relevant to the framework of the research question and design.

Very broadly, the institutional school holds that rules and structure direct decisionmaking, the majoritarian approach asserts that decisions reflect the preferences of society at large, and attitudinalists argue that decisions represent an aggregation of Justices’ personal preferences. Naturally, the three viewpoints have some overlap. For example, a conservative turn in public opinion that leads to the election of several conservative presidents, the subsequent appointment of conservative Supreme Court Justices, and, eventually, conservative verdicts could conceivably be portrayed as illustrative of either the attitudinal or majoritarian theses.2 Or (as noted in the emerging field of strategic decisionmaking, which is addressed below), Justices’ personal preferences may be inextricably bound to the necessities of heeding institutional norms (Epstein & Knight 1998; Maltzman et al. 1999). But, at least for the present purposes, the literature is reasonably conceived in terms of these three streams, each suggesting the dominance of a single explanation (or class of explanation) of outcomes.

The Institutional Approach

A key question in the study of decisionmaking in any institution is whether structure and rules matter. In the judicial realm, proponents of a bare-bones institutionalism argue that the appointment/life tenure components of the office of Supreme Court Justice influence outcomes.3 The Founders in particular suggested that these institutional characteristics would result in the federal courts displaying a much greater tendency to protect minority rights than would the election-minded legislature or executive. As Hamilton predicted in Federalist 78 (through his socalled bulwark thesis), appointment and life tenure would afford Justices the necessary independence to become the bulwark defenders of minority interests against serious oppression by the majority. Judicial review is the tool with which the Court can actually enforce these protections by striking down unconstitutional legislation or actions.4

The scenario Hamilton sketches out in Federalist 78 is that the Constitution is designed to protect minorities (an observation even more true today, given the addition of the Bill of Rights and the Fourteenth Amendment), and that the Court will stand by the Constitution.5 Although he acknowledged that “it is easy to see that it would require an uncommon portion of fortitude in the judges to do their duty as faithful guardians of the Constitution, where legislative invasions of it had been instigated by the major voice of the community” (Hamilton et al. 1961:470), he at least suggests that life tenure would generate such fortitude, stating:

This independence of the judges is equally requisite to guard the Constitution and the rights of individuals from the effects of those ill humors which the arts of designing men, or the influence of particular conjunctures, sometimes disseminate among the people themselves, and which . .. have a tendency … to occasion dangerous innovations in the government and serious oppressions of the minor party in the community. (469)

Similarly, he refers to the independence of the judges as the “essential safeguard against the effects of occasional ill humours in the society” (470, emphasis added).

Though defenders of the bulwark theory have been criticized for inferring an empirical reality that is premised on nothing more than the theoretical obligation to protect minority rights (Railton 1983), the approach is not necessarily normative (and is not presented as such here). Hamilton, in this essay, was arguing not how Judges ought to behave, but how they would behave as a result of the influence of institutional structure, although he may have privately harbored some doubts regarding these predictions (Rakove 1996). Thus, Federalist 78 presents reasonable and testable hypotheses, but attempts to verify them have been limited, especially in this area.6

Rosenberg’s (1991) history of the role played by the federal courts in the facilitation of social change highlights the potential influence of the Court’s structure in protecting racial minorities from majority excesses, but does not provide a comprehensive analysis of the record. Other efforts have focused on a state-bystate comparison and have found the life tenure component in state supreme courts to be a significant predictor of outcomes (Brace & Hall 1995), although no studies have attempted such an analysis of racial discrimination claims. In fact, the potential influence of institutional structure on U.S. Supreme Court decisions is most often assessed indirectly and negatively, as the fallout of studies suggesting a strong influence for either majoritarian or attitudinal variables.

The Majoritarian Approach

Majoritarianism, the most distinct alternative to the institutional explanation, generally envisions a Durkheimian perspective in which a society is a single living organism and courts can merely be expected to mirror their host community and its majority preferences, regardless of institutional features (McIntosh 1991). A great deal of evidence has been gathered in support of the majoritarian thesis. Although the suggestion appears earlier (e.g., Commager 1943), Dahl (1957) was the first to provide empirical justification. This influential contribution showed how rarely the Supreme Court struck down federal policies (defined by Dahl as a surrogate for the majority will), particularly those that could be interpreted as harmful to minorities. Although Casper (1976) found that the Supreme Court was much more likely to overturn state and local policies, thus suggesting that Dahl’s conclusions were overstated, attempts to estimate majoritarian tendencies have flourished. Generally, support has been found for a positive relationship between public preferences and judicial outcomes both in terms of court decisions (Barnum 1985; Marshall 1989; Mishler & Sheehan 1993; Stimson et al. 1995) and the votes of individual justices (Flemming & Wood 1997). However, there is some question as to whether this influence is direct or indirect (Norpoth & Segal 1994) and questions about the causal direction of the relationship (Caldeira 1991). The many methodological problems with these studies led Segal and Spaeth (1993) to conclude that there is no convincing evidence of a majoritarian link.

In terms of the more specific question of the Court’s record on the protection of racial minorities against discrimination, there is little substantial evidence of majoritarianism. Spann (1993) argues that the Supreme Court has heeded popular will at the expense of minority interests in discrimination cases, but provides little systematic evidence. Link (1995) provides a rare empirical account of the question and finds that the Supreme Court’s likelihood of reaching a pro-minority decision is significandy and positively affected by concurrent levels of public liberalism on racial issues.

One problematic aspect of this general body of literature, however, is the overly-broad conclusions that have been extrapolated from its findings. Although the evidence can reasonably be accepted at face value, the extent to which it fatally cripples institutionalism may be exaggerated. Overstated interpretations of the bulwark thesis have led to a misguided tendency to declare majoritarianism the “winner” when in fact the rather narrow predictions Hamilton made have not effectively been discounted. Most existing tests of Court majoritarianism fail to truly account for the precepts of this contrary view. The many studies that support the idea that decisions in general are correlated with public opinion in general do not necessarily deny the actual bulwark thesis; nor do investigations (e.g., Link 1995) that include cases lacking even the assertion of a constitutional rights violation. The proper test would involve a comparison of the Court’s record with public sentiment in regard to minority rights in a set of cases involving constitutional claims. Unfortunately studies so designed are lacking.

Before moving on to the attitudinal approach, the concept of strategic judicial behavior should be mentioned, since it offers explanations of outcomes that overlap both institutionalism and majoritarianism, and because some of its most high-profile investigations have been in the racial discrimination field. Basically, this judicial version of positive political theory suggests that Supreme Court decisionmaking may be influenced by the balanceof-power dynamics between the Court and other federal actors. It thus displays an institutional component in that the basic structure of government is a key factor but might also be seen as majoritarian in nature because of the focus on the Court potentially bending its will to that of the popularly elected branches. Evidence in favor of strategic behavior is building (Epstein & Knight 1998), but there are some complications involved in drawing clear conclusions (Baum 1997), and debate continues over the extent to which strategic preferences outweigh sincere preferences (Segal 1997). However, the approach and the potential variables it suggests are, at this stage at least, extraneous to this inquiry. First, strategic theories have focused on the relationship between the Court and the Congress, the President, and/or the lower federal courts. The majority of the cases here, however, involve challenges to state and local policies or actions, so no clear expectation of strategic decisionmaking can be gleaned from this literature. Additionally, the strategic approach to racial discrimination has involved statutory interpretation issues (Eskridge 1991; Marks 1988), but the cases here solely involve challenges to constitutionality. Strategic behavior may indeed occur in this set of cases, but a consideration and investigation of its role is beyond the scope of this basic inquiry.

The Attitiudinal Approach

Majoritarianism provides a clear alternative to Hamilton’s predictions of Court insularity, but the attitudinal approach, which stresses the role played by the ideological preferences of Justices, is relevant here as well. There is now a lengthy history of empirical substantiation for this theory (two classics are Schubert 1965 and Segal & Spaeth 1993), but applications to racial discrimination cases have been limited to impressionistic accounts of an increasingly conservative Court, leading to fewer pro-minority rulings (Davis & Graham 1995; Edsall & Edsall 1992). Yet, there is enough evidence for the general influence of Justices’ personal ideologies that this factor needs to be systematically considered in this issue area. The relevant ideological dimension here is variation on the basic liberal-to-conservative scale, which could potentially push Justices toward or away from minority protection.

Ironically, this approach is often linked to the very institutional factors that Hamilton asserted would lead to quite different outcomes. The potential for life tenure to allow not the Constitution but the Justices’ personal preferences to guide decisionmaking was recognized even in Hamilton’s times. In refutation of the bulwark argument, the pseudonymous Brutus (probably Judge Robert Yates of New York) wrote in The AntiFederalist Papers that judges under a system of life tenure “will not confine themselves to any fixed or established rules, but will determine, according to what appears to them, the reason and spirit of the constitution” (Ketcham 1986:295). As Segal (1997:29) notes, “For a variety of reasons, including the institutional protections granted the Court… attitudinalists argue that the justices of the Supreme Court are free to vote their sincere preferences” (emphasis added). But this is not the scenario Hamilton predicted (and that is tested here). Evidence in favor of attitudinal variables would prove problematic for the bulwark theory.7

Research Design

In summary of the basic literature outlined above, valid evidence supports the general majoritarian thesis, but a tendency to overstate Hamilton’s predictions of an insulated Court may have led to the underestimation of insularity in particular scenarios. A more precisely designed investigation of the determinants of outcomes in racial discrimination cases, which tests the essence of the institutional argument against the precepts of majoritarianism and also controls for the potential influence of the Justices’ ideologies, is outlined later.

Data Set

This study encompasses all Supreme Court racial discrimination cases from 1954 to 1996 published in the Supreme Court Reporter and indexed in West’s Decennial Digest under the “equal protection of laws, equal rights” and “personal, civil, and political rights” subheadings of the constitutional law category.8 Use of the West’s classification system allows for the construction of an objective and replicable data set in which only constitutionally based challenges are included, thus allowing for the more precise test of the Hamiltonian prediction.9 Since that basic thesis makes no assertions regarding outcomes in claims involving the mere interpretation or implementation of a statute, such cases are not triggered by the selection mechanism and are necessarily excluded. Only if the original minority plaintiff makes a constitutional claim is it so indexed in the Supreme Court Reporter.10

After selecting the initial data set, I eliminated three categories of opinions: First, I excluded a small number of per curium decisions because key facts could not be gleaned from these terse decisions. Second, I omitted challenges to private acts of discrimination as well. Although some constitutional claims were raised, thus triggering their initial appearance in the data set, they pertain to the constitutional right of the federal government to regulate these private actions, as opposed to the constitutional civil rights of minorities. Thus, they exist on a different legal plane than the other cases. Finally, I also eliminated criminal appeals in which an equal protection violation was claimed due to conviction by a racially disproportionate jury. These claims were too entwined with other constitutional provisions (specifically those dealing with the rights of the accused and convicted) for me to draw reasonable conclusions on the discrimination dimension. This method of selection resulted in a final data set of 48 cases (listed in the Appendix), the majority of which involve discrimination in public schools, public accommodations, voting rights, or employment.

Independent Variables

Classifying Cases To Capture Public Opinion Variations

Although the construction of the data set corrects one of the asserted problems of existing tests of majoritarianism (the inclusion of non-constitutional cases), a great deal of consideration must be given to the second problem-estimating the link between public opinion and decisions without a precise understanding of what that opinion is and how it may vary. A model that captures this subtle variation and its influence can provide a better test of the precepts of majoritarianism (and by extension the bulwark thesis) than investigations that utilize blunt measures of public liberalism or a composite of all racial issues. There are potentially myriad ways to construct such a model, but I offer a first step based on existing evidence.

My approach to gaining a greater grasp of majority opinion on this issue is based on the well-accepted proposition that the public is far more tolerant of de facto discrimination than it is of de jure discrimination, a tendency that remained stable for the years in this data set. This proposition is supported below, and this basic scenario also suggests an appropriate research design. The key independent variable is the cases themselves, categorized by the type of discrimination being challenged, with case decisions being the dependent variable. If, on one hand, the Court is significantly less tolerant of de jure than de facto discrimination (i.e., more likely to strike it down), a majoritarian link will at least be suggested (although a correlation between public opinion trends and Court decisions would still leave open the question of the causal direction of that relationship). If, on the other hand, the type of case has no impact on verdicts, my assertion will be supported (i.e., that a more nuanced investigation will report greater evidence for Hamilton’s bulwark thesis, and less evidence for the impact of majority preferences).11

Confirming this characterization of pubic opinion requires an overview of the available evidence that, as a whole, strongly supports the proposition. First, although the voluminous racial attitudes literature emanating from the fields of sociology, psy chology, and political science has moved beyond the de jure versus de facto question, general acknowledgment of the notion that white Americans are, and always have been, far more tolerant of non-governmentally imposed discrimination than that mandated by law is common (Campbell 1971; Edsall & Edsall 1992; Kinder & Sanders 1996; Kuran 1996).

Additionally, the complex debate over the true source of continuing resistance to policies designed to promote civil rights provides deeper theoretical confirmation. Three explanations have emerged within this literature, which I shall very briefly summarize. Whites who object to certain policies may be motivated by fear of the physical proximity of blacks (West 1993; Woodward 1974); by a commitment to the idea that, in the interest of fairness, laws should be truly “color-blind” and grant no difference in status, good or bad, on the basis of race (Kinder & Sears 1981; Sears et al. 1979; Sears et al. 1980); or by concern that the adoption of such policies would further empower a federal government that is already too large and intrusive (Bennett & Bennett 1990; Sniderman & Piazza 1993).

Even given the lack of consensus within this subfield as to which is the most important source of opposition, the conclusions still provide a great deal of useful information for my purpose of justifying the de jure/de facto support variations. For whether the public (or that portion who in fact objects to any, or all, policies) is motivated by the fear of physical proximity, unfair policies, or government intervention, attempts to end de jure discrimination generate far less opposition than attempts to end de facto discrimination. I examine each of these explanations in regard to both types of discrimination.

First, decisions limiting legally mandated discrimination would end some separation between the races but would not offect the self-imposed separations (common outside of the South) that could substantially be altered by successful challenges to de facto patterns. Thus, limits on de jure actions would not trigger racial proximity fears to the degree that an end to de facto actions would. Second, for those concerned with the preservation of perceived principles of fairness, de jure discrimination is troubling and should be struck down. But de facto discrimination does not elicit the same reaction. In fact, one common response to attempts to limit de facto patterns is that such a course of action itself violates the color-blind principle. Third, the fear of overly intrusive government not only leads to support for the removal of de jure discrimination (which in fact represents a great deal of government power) but also leads to opposition to the increased government presence, at least perceived as necessary, to end de facto situations.

Finally, empirical support for this de jure/de facto dichotomy is provided by polling data.12 Surveys consistently show greater than 50% support among white Americans for striking down various types of de jure discrimination, such as limits on where children can go to school, segregated public accommodations, discriminatory voting practices, and miscegenation laws. However, support for ending de facto patterns of separation in areas such as housing, employment, and schools through practices designed to achieve equal outcomes is regularly below 50%.13

The previous evidence confirms the assertion that public support declines when policies move from the striking down of discriminatory laws to the implementation of policies attempting to actively redress discrimination/segregation occurring through patterns and practices linked only distantly, if at all, with state action. Additionally, this trend is consistent over time. Thus, the next logical step is to divide all cases in this data set into de jure and de facto categories. There are complications, however, with this dichotomization.

First of all, the de jure designation is not as clear-cut as it may seem. Certainly, cases involving challenges to discriminatory laws-such as the miscegenation statute in Loving v. Virginia (1967) or the city segregation ordinance in Peterson v. Greenville (1963)-clearly belong in this category. And even instances in which there is no codified discrimination but the state behaves as if there is easily fit the de jure designation (e.g., Lombard v. Louisiana [1963] in which a city truthfully claimed to have no segregation ordinances, yet arrested blacks for attempting to eat in certain restaurants).

But what of challenges to discrimination that, by plaintiffs’ own admissions, is only indirectly tied to state action? Examples include Burton v. Wilmington Parking Authority (1961), in which a coffee shop leasing space from a city refused to serve blacks; or more complicated cases such as U.S. v. Fordice (1992), in which the claim was not that a state university was continuing to officially discriminate against blacks, but that it had not done enough to rectify discriminatory behavior patterns that were institutionalized in prior years, when de jure segregation was in effect. It seems unjustified that, for the purpose of capturing public reaction, these actions should also be considered as instances of de jure discrimination.

Based on the public opinion trend outlined previously, these types of cases are likely to elicit a different reaction from the majority. Because the evidence of state involvement is much more tenuous, the strong public disapproval that would be evident in the clear de jure cases would likely be eroded here. Therefore, these two different types of “de jure” discrimination are captured by two separate categories. Category One cases encompass challenges to discrimination by direct government action, and Category Two comprises challenges to discrimination linked indirectly to government action.

There is also some complexity to the de facto category. In fact, none of the cases in this set involved a constitutionally based challenge to discrimination or segregation that was claimed to be purely de facto.14 The ironic manner in which the Court is asked to weigh in on the issue of de facto discrimination is through cases in which white plaintiffs challenged state or federal policies designed to rectify de facto patterns as violative of their constitutional right to equal protection of the laws. Category Three comprises these challenges to de facto remedies.

I made coding decisions in regard to type of discrimination on the basis of the claim of the original plaintiff (confirmed by a careful reading of the facts of the case), for, as noted by George and Epstein (1992), ascertaining case facts from the Court’s own conclusions can skew findings. Cases in which the original plaintiff asserts that discrimination is mandated by a particular governmental policy or practice fit Category One. Cases in which the state is claimed to be only an accessory to another party’s discrimination, or where direct state action is claimed to be only part of the problem, compose Category Two. When the challenge is to policies designed to rectify de facto patterns, I coded the case as Category Three. If these opinion trends are reflected in decisions, the likelihood of a pro-minority decision should decrease as case type increases.

B. Controlling for an Alternative Scenario of Majoritarianism

Although evidence supports the previous characterization of public opinion, the characterization still represents only one conception of this variable. Even if the Court fails to respond to the opinion variances regarding different types of discrimination, it is conceivable that it is still responsive to the public in some other way. I included Stimson’s (1999) measure of public mood to test the most likely possibility-that my test of majoritarianism is simply too nuanced and that it is the degree of general public liberalism that influences outcomes in these cases.15 For this variable (Public Liberalism), higher numbers represent greater levels of public liberalism. If the Court is bluntly majoritarian, then an increase in liberalism should have a positive effect on the handing down of pro-minority decisions.

C. Measuring the Attitudinal Variable

In order to provide the essential control for the attitudinal model, I assigned each case a score based on the mean ideology of the Justices hearing that case. I calculated the means from an ideological index gleaned from newspaper editorials (Segal & Cover 1989; Segal et al. 1995).16 For this variable (Court Ideology), higher numbers represent a more liberal court. If the variable is positively linked to pro-minority decisions, attitudinal predictions will be supported.

Dependent Variable

The outcome to be predicted is simply the decision in each case. I coded these decisions as to whether or not the minority interest was upheld. When a challenged act of direct or indirect state discrimination (Case Type 1 or 2) was struck down, I coded the decision as pro-minority; if not, as anti-minority. In Category Three cases, I coded decisions as pro-minority when the Court failed to strike down the challenged de facto remedy. When such remedies were stuck down, the decision is anti-minority.


I first present the results of a simple logistic regression model, using only type of case as the independent variable. As the data in Table I show, the categorical movements appear to have a limited effect on the likelihood of an anti-minority decision. Cases involving challenges to indirect state discrimination (Case Type 2) or de facto remedies (Case Type 3) are both significantly less likely to be decided in a pro-minority manner than are challenges to direct state discrimination (Case Type 1).17

With this model, the majoritarian argument appears to receive a degree of support, with the Court’s sympathy to minority claims reflecting, to a limited extent, the opinion patterns of white Americans. But further analysis shows that this simple approach is misleading to the extent to which it attributes even modest significance to the case categories. The catch involves the time bound nature of these cases. As might be expected, the “easier types of discrimination, (i.e., Case Type 1, the direct de jure cases) were challenged first, with the more difficult indirect cases coming later. And because remedies to de facto discrimination (Case Type 3) did not occur until well into the years of this data set, cases involving challenges to these remedies naturally occurred last. This chronology is important because it means that different types of cases were heard by different Courts. Table 2 illustrates this effect.

The obvious trend is the varying racial discrimination docket of each Court era, clearly highlighted by the fact that the Rehnquist Court heard no pure de jure cases, while the Warren Court heard no challenges to a de facto remedy. It is thus difficult to conclude if the Warren Court’s more liberal record was a result of the “easy” cases on its docket or the result of the liberal tilt of that body.18 Similarly, there is no simple way to determine if the largely anti-minority Rehnquist Court record was an artifact of tougher cases or conservative Jurists. This distributional skew clearly illustrates the need for a full model in which the attitudinal variable is included as well. Findings based on this complete model, in which the effects of type of case, ideological composition of the Court, and public mood are estimated on the case outcome, are presented in Table 3.

In this more nuanced test of the majoritarian thesis, the public preference variables fail to achieve the statistical criteria necessary to reject the null hypothesis. The category of discrimination being challenged exhibits no significant effect. Thus, public opinion variations as to the necessity of ending different types of discrimination are not reflected in these decisions. Furthermore, national public liberalism also fails to influence outcomes.19 In general, the Court is predominately supportive of minority claims, ruling in the minority interest in 65% of the cases. That finding provides a degree of support for Hamilton’s predictions. The attitudinal variable clearly exerts its own significant effect, however. Pro-minority decisions, though the norm, are attenuated by relatively conservative Courts.20

Some scholars would consider the significant effect of the Court’s ideological composition evidence of indirect public influence, and thus supportive of a majoritarian link. This issue cannot adequately be addressed here, but such a conclusion would rely on a scenario in which the public elects a President in keeping with its current ideological temper, and this President then selects sympathetic Justices who in turn push the Court in the particular ideological direction. Given the insignificance of the general public liberalism measure, this scenario does not seem very likely, although perhaps the three-year lag time is not long enough for the process to reach fruition.

Despite the potential controversy over the extent to which attitudinal variables actually reflect majoritarianism (a debate I will respectfully bow out of), the more troublesome aspect of the attitudinal variable for the Hamiltonian argument is the extent to which it undercuts the expectation that the precepts of the Constitution will always reign supreme in the disposition of minority claims of rights violations. The only chance of salvation would be to argue that the strength of the attitudinal variable does not necessarily mean that Justices ignore the Constitution in heeding their own ideological leanings, and that the Justices’ preferences could simply be in serendipitous harmony with the “correct” constitutional interpretation, whether liberal or conservative. But that argument is clearly forced. Although this is not the forum for a consideration of what the proper ruling on these claims should be, the finding that Justices’ attitudes shape decisions renders problematic an unconditional acceptance of Hamilton’s claim that Justices will protect minorities when, and only when, the Constitution mandates.


Based on its structure, the Supreme Court was expected by the Founders to function as an anti-majoritarian defender of the constitutional rights of minorities, but the many studies supportive of a link between public opinion and decisions have weakened the perceived strength of that hypothesis. However, the results of this investigation support, to an extent, the core predictions of the bulwark thesis, suggesting that the white majority’s vicissitudes on the necessity of ending various types of discrimination were not reflected in the Court’s decisions. And, on a broader level, general public mood also failed to influence outcomes.

Naturally, this is not to say that majoritarianism in general is debilitated by these findings. The purposefully limited data set definitively precludes such an overblown conclusion. Additionally, the particular measures used to capture public opinion (the de jure/de facto distinction and the general liberalism index) may not have represented the proper operationalization of this variable. But the initial suspicion has been substantiated. In this more suitable test, in which only those cases relevant to Hamilton’s bulwark thesis were selected and where the nuances of public opinion were better captured, evidence of majoritarianism failed to materialize.

Yet, even though the Court does not appear responsive to public opinion on any level here, neither can it be described as the consistent defender of racial minorities against violations of their constitutional rights. A healthy majority of cases in the data set were decided in favor of minority interests, but approximately 35% were not. The data suggest that, in a perversion of the bulwark thesis, the Court can be expected to defend minorities only when its membership has a relatively liberal tilt.

There are clear limitations and boundaries to this fundamental and preliminary study. Not all potential determinants of decisionmaking are controlled for. In particular, I did not assess more modern takes on the role of institutional structure, largely because they were not relevant to this very limited data set. The strict criteria for case inclusion makes this a test of the basic Hamiltonian anti-majoritarian thesis in racial discrimination cases as opposed to a sweeping analysis of the Court’s complete record in this area.

What this inquiry does contribute is a broad empirical comparison of Federalist conjectures with current theories of majoritarian and attitudinal influence, which has been lacking to this point. This comparison provides a baseline from which to move on to more sophisticated and encompassing estimations of these approaches. Additionally, the case categorization framework I utilized here to capture public opinion offers an initial step toward the application of more nuanced and accurate assessments of this variable when it is offered as a potential determinant of decisions in any area.

One logical next step in this issue area might be to compare the outcomes in these constitutionally based challenges with other cases in which the discrimination claim was based only on a disputed statutory interpretation, to further underscore the useful but limited nature of Hamilton’s bulwark predictions. Alternatively (although, as noted, there is no common precedent to this set of opinions), an in-depth study of the cases that were potentially linked to a controlling decision may uncover information on one particular aspect of the institutional approach that I did not analyze here. Finally, one might explore the influence of the U.S. Congress (as suggested by Wasby et al. 1977, as well as by strategic theorists) or even state legislatures on the relevant outcomes in this area. For now, the picture that has emerged from this basic inquiry is of a Supreme Court that adjudicated minority rights claims independently of public preference. In the relatively few instances where the Court did not sustain minority claims in this set of cases, its decisions were traceable to ideological factors-conservative Judges, not public opinion, pushed these decisions in a conservative direction.

I refer solely to the debate over what does influence outcomes. The separate controversy over what should determine decisionmaking is well beyond the realm of this inquiry. Although recognizing that even objective studies may potentially be normative discourse in disguise, the various analyses cited herein are accepted as impartial evaluations of observed trends.

This debate is illustrated through the findings of Mishler and Sheehan (1993), as critiqued by Norpoth and Segal (1994). I briefly return to a consideration of this point in the findings section.

Naturally, there are a great many more sophisticated institutional approaches as collected, for example, in Clayton and Gillman (1999).

Broader models assert that institutional variables result in Justices heeding not just the Constitution but case precedent as well. But precedent is not a major component of Hamilton’s thesis, which more narrowly focuses on the Constitution’s dominance. Furthermore, the notion of precedent lacks relevance to the set of cases examined here, which are linked only by constitutional challenges to some instance of discrimination. There is no common precedential thread to test the effects of here, and an attempt to code each case individually in regard to the existence of controlling precedent would lead to violations of objectivity and comparability.

As many scholars have observed, Hamilton’s personal vision of a minority was quite different from current definitions. But the basic concept, in which a minority is identified by particular characteristics that might result in majority tyranny, is consistent.

There are some disagreements as to how Federalist 78 should be interpreted, and a reasonable argument could be made that Hamilton believed life tenure to be a necessary but not sufficient factor in the bulwark thesis. If so, this would still warrant empirical investigation.

Hamilton does suggest that structural features will lead to “a spirit of independence” among Justices, but the implication is that this independence is from elected bodies and the public, not the Constitution. The logic behind constitutional fealty overriding personal preference is more clearly suggested in Rostow’s (1952) essay on judicial review. Rostow asserts that the potentially frightening (to the public) scenario, in which Justices appointed for life have the ability to strike down legislation, is mitigated by the assurance that the Constitution reigns supreme over any legislative body and that Justices merely interpret its directives. Thus, Justices realize that a pattern of disregard for the Constitution in favor of their own preferences could result in widespread rejection of decisions. In order to preserve the legitimacy of their own branch, they must adhere to the Constitution, a purposeful arrangement also suggested by Madison in Federalist 51.

As noted, I do not expect the Brown precedent to be directly relevant to most of these cases. The choice to begin the study in 1954 is simply premised on the general acceptance of the Brown decision as ushering in the modern era of racial discrimination doctrine. The specific key numbers used are 215 through 220.5 in the equal protection of laws, equal rights subheading; and 82 through 91 (excluding 84 and 90, which respectively represent religious liberty and free expression claims) in the personal, civil, and political rights subheading.

The Spaeth Supreme Court database, which relies on the numbered holdings in U.S. Reports’summaries to code cases, offers a substantially similar method of constructing a data set. Using constitutional issue and racial discrimination topic codes from that database, I constructed an alternative data set that was almost identical to my own. A few cases appeared in the Spaeth-generated set, but not in the original, suggesting that the U.S. Reports summary attributed a constitutional claim, but the Supreme Court Reporter did not. In checking this, however, I found that, in these cases, there was in fact no reference to a constitutional claim in the U.S. Reports summary, thus suggesting some potential replicability difficulties with the Spaeth data.

10 A typical racial discrimination case that lacks a constitutional claim involves school integration disputes. In many such cases the question was not whether segregation itself was a constitutional violation of minority rights but-in a much narrower and technical inquiry-whether the school district was properly implementing a previously approved integration plan. Another example is Griggs v. Duke Power Co. (1971), in which the plaintiff challenged the use of alleged racially biased intelligence tests for hiring purposes. The complaint was based not on constitutional grounds but on a claimed violation of Title VII of the 1964 Civil Rights Art.

11 It is somewhat unusual to test two alternate theories with the same variable (i.e., if the variable is significant, the majoritarian explanation is supported, and if not, the Hamiltonian explanation is supported). But the tactic is reasonable, given that the two approaches are virtual mirror images of each other, and there is no independent way to show that a single Court (which remains constant across all observations) is insulated from public preferences.

12 Many of these surveys are collected and presented in Schuman et al. (1985). Because these authors are concerned with the “principle-implementation” phenomenon, in which whites support general principles of equality but not specific governmental actions necessary to carry them out (also see Tuch & Hughes 1996), they are careful to include polling questions that target attitudes on the latter. This is useful for my purposes since the cases themselves revolve around policy issues and not general principles of right and wrong.

13 A sample of survey results from various polling organizations shows the range of support levels in a given years) among whites for ending various types of discrimination: de jure accommodations: 55-65% (1964-1974)-Institute for Social Research voting rights: 76% (1965)-Gallup

miscegenation: 60-70% (1973-1982)-National Opinion Research Center

schools (open to all races): 38-66% (1964-1990)-National Election Studies

de facto housing: 30-40% (1972-1984)-National Opinion Research Center employment: 40% (1964-1972)-Institute for Social Research schools (busing): 9-11% (1972-1984)-National Election Studies

14 This point is illustrated by Milliken v. Bradley (1974). In that case, involving attendance patterns in Detroit public schools, the plaintiffs claimed not that voluntary patterns of residential segregation led to segregated schools (which would have been a true de facto challenge), but that, despite integrated residential patterns, schools were still segregated. The school board was blamed for making decisions that ultimately led to this outcome.

15 This measure, a compendium of survey questions, is a barometer of the American public’s relative liberalism by year, in regard to “preferences for a larger, more active federal government as opposed to a smaller, more passive one across the sphere of all domestic policy controversies” (Stimson et al. 1995:548). Typically, the measure is lagged when estimating its impact on policy outputs. Stimson (1999) also offers a specialized racial index, but the general score better captures the broad public mood that I control for here (although the difference is largely moot, since the two measures are quite similar). Link (1995) finds a three-year lag results in the most significant effects of the racial index on Supreme Court outcomes. I utilize the three-year lag as well in order to give this variable the best chance to succeed.

16 Information from editorials is one of the few ideological measures that does not use a justice’s previous votes to predict future votes, thus avoiding the problems of circularity and non-falsifiability. Epstein and Mershon (1996) note this measure’s misuse by

scholars who apply it to studies where it does not fit, but their study confirms that the measure is valid in the prediction of civil rights and liberties cases, as the editorials that were used as the basis for this measure largely focused on Justices’ pre-Supreme Court record in these two visible areas.

17 Since type of case is an ordinal level variable, the three categories are run as a series of dummies with Category One withheld. The significance of the movement from Two to Three is reported at the bottom of the table.

18 Baum (1988) utilizes an issue dimension model to confront a similar quandary in his analysis of civil liberties cases, but that approach does not fit here. He states that the method is necessary only where case content cannot be controlled for, but that is not the case here, since the categorization framework provides this function. Additionally, as the method assesses issue and personnel change but not member change, it does not allow for a test of majoritarianism.

19 To test for the possibility that the three-way categorization was too subtle, I also ran the equation with Categories One and Two collapsed-thus resulting in one category that captured all de jure challenges (whether direct or indirect) and one category composed of challenges to de facto remedies. The results remained insignificant. Measuring the effects of categorical variation and general public liberalism separately also failed to yield significant results.

20 Because I use case as the unit of analysis, there are instances of multiple cases in a given year. This means that there are occasions where Court ideology and public liberalism are invariant, making the same prediction on the case outcomes in a given year (actually, it is possible that Court ideology is not invariant in a given year because one or more Justices may resign during the term or simply not participate in a certain case). But this prediction error is an artifact of the actual distribution of the cases over time, and thus unavoidable. This condition has two statistical results: coefficient attenuation, and

standard error inflation. In other words, this condition results in a slight bias toward the null hypothesis for these independent variables. For a more detailed discussion of this effect, see Cain, Ferejohn, and Fiorina’s discussion of “Overlooked Difficulties in CrossLevel Analysis” (1987:130).


Barnum, David G. (1985) “The Supreme Court and Public Opinion: Judicial Decision Making in the Post-New Deal Period,” 47 J of Politics 652-65.

Baum, Lawrence (1988) “Measuring Policy Change in the U.S. Supreme Court,” 82 American Political Science Rev. 905-12.

– (1997) The Puzzle of Judicial Behavior. Ann Arbor: Univ. of Michigan Press.

Bennett, Linda, & Stephen Earl Bennett (1990) Living with Leviathan. Lawrence: Univ. Press of Kansas.

Brace, Paul, & Melinda Gann Hall (1995) “Studying Courts Comparatively: The View from the American States,” 48 Political Research Q 5-30.

Cain, Bruce, John Ferejohn & Morris Fiorina (1987) The Personal Vote. Cambridge: Harvard Univ. Press.

Caldeira, Gregory A. (1991) “Courts and Public Opinion,” inj. B. Gates & C. A. Johnson, eds., The American Court. Washington, DC: CQ Press.

Campbell, Angus (1971) White Attitudes Toward Black People. Ann Arbor: Institute for Social Research.

Casper, Jonathan D. (1976) “The Supreme Court and National Policy Making,” 70 American Political Science Rev. 50-66.

Clayton, Cornell W., & Howard Gillman, eds. (1999) Supreme Court Decision Making. Chicago: Univ. of Chicago Press.

Commager, Henry Steele (1943) “Judicial Review and Democracy,” 19 Virginia Q. Rev. 417-28.

Dahl, Robert (1957) “Decisionmaking in a Democracy: The Supreme Court as a National Policymaker,” 6 J. of Public Law 279-95.

Davis, Abraham L., & Barbara Luck Graham (1995) The Supreme Court, Race, and Civil Rights. Thousand Oaks, CA: Sage.

Edsall, Thomas Byrne, with Mary D. Edsall (1992) Chain Reaction. New York: W.W. Norton & Co.

Epstein, Lee, & Carol Mershon (1996) “Measuring Political Preferences,” 40 American J of Political Science 261-94.

Epstein, Lee, & Jack Knight (1998) The Choices Justices Make. Washington, DC: CQ Press.

Eskridge, William N., Jr. (1991) “Reneging on History? Playing the Court/Congress/President Civil Rights Game,” 79 California Law Rev. 613-85. Flemming, Roy B., & B. Dan Wood (1997) “The Public and the Supreme Court:

Individual Justice Responsiveness to American Policy Moods,” 41 American J of Political Science 468-98.

George, Tracey E., & Lee Epstein (1992) “On the Nature of Supreme Court Decision Making,” 86 American Political Science Rev. 323-37.

Hamilton, Alexander, James Madison & John Jay (1961) The Federalist Papers. New York: NAL Penguin, Inc.

Ketcham, Ralph, ed. (1986) The Anti-Federalist Papers and the Constitutional Convention Debates. New York: New American Library.

Kinder, Donald R., & David 0. Sears (1981) “Prejudice and Politics: Symbolic Racism Versus Racial Threats to the Good Life,” 40 J. of Personality & Social Psychology 414-31.

Kinder, Donald R., & Lynn M. Sanders (1996) Divided by Color. Chicago: Univ. of Chicago Press.

Knight, Jack, & Lee Epstein (1996) “The Norm of Stare Decisis,” 40 American J. of Political Science 1018-35.

Kur-an, Timur (1996) “Seeds of Racial Explosion,” in P. A. Winters, ed., Race Relations. San Diego: Greenhaven Press.

Link, Michael W. (1995) “Tracking Public Mood in the Supreme Court: CrossTime Analyses of Criminal Procedure and Civil Rights Cases,” 48 Political Research 0. 61-78.

Maltzman, Forrest, James F. Spriggs II & Paul J. Wahlbeck (1999) “Strategy and Judicial Choice: New Institutionalist Approaches to Supreme Court Decision-Making,” in C. W. Clayton & H. Gillman, eds., Supreme Court DecisionMaking. Chicago: Univ. of Chicago Press.

Marks, Brian A. (1988) A Model ofjudicial Influence on Congressional Policymaking: Grove City College v. Bell. Working Papers in Political Science, P-88-7, the Hoover Institution, Stanford Univ., Stanford, CA.

Marshall, Thomas (1989) Public Opinion and the Supreme Court. New York: Longman.

McIntosh, Wayne V. (1991) “Courts and Socioeconomic Change,” inJ. B. Gates & C. A. Johnson, eds., The American Courts: A Critical Assessment. Washington, DC: CQ Press.

Mishler, William, & Reginald S. Sheehan (1993) “The Supreme Court as a Countermajoritarian Institution? The Impact of Public Opinion on Supreme Court Decisions,” 87 American Political Science Rev. 87-101.

Norpoth, Helmut, & Jeffrey A. Segal (1994) “Popular Influence on Supreme Court Decisions, 88 American Political Science Rev. 711-24.

Railton, Peter (1983) “Judicial Review, Elites and Liberal Democracy,” in R. Pennock & J. Chapman, eds., Liberal Democracy. New York: New York Univ. Press.

Rakove, Jack N. (1996) Original Meanings. New York: Alfred A. Knopf. Rosenberg, Gerald (1991) The Hollow Hope. Chicago: Univ. of Chicago Press. Rostow, Eugene (1952) “The Democratic Character of Judicial Review,” 66 Harvard Law Rev. 193-224.

Schubert, Glendon (1965) The Judicial Mind: The Attitudes and Ideologies of Supreme Court Justices, 1946-1963. Evanston, IL: Northwestern Univ. Press. Schuman, Howard, Charlotte Steeh & Lawrence Bobo (1985) Racial Attitudes in

America. Cambridge: Harvard Univ. Press.

Sears, David O., Carl P. Hensler & Leslie K. Speer (1979) “Whites’ Opposition to Busing: Self-Interest or Symbolic Politics?” 73 American Political Science Rev. 369-84.

Sears, David O., Richard R. Lau, Tom R. Tyler & Harris M. Allen, Jr. (1980) “Self-Interest Versus Symbolic Politics in Policy Attitudes and Presidential Voting,” 74 American Political Science Rev. 670-84.

Segal, Jeffrey A. (1997) “Separation-of-Powers Games in the Positive Theory of Congress and Courts,” 91 American Political Science Rev. 28-44.

Segal, Jeffrey A., & Albert Cover (1989) “Ideological Values and Votes of Supreme Court Justices,” 83 American Political Science Rev. 557-65. Segal, Jeffrey A., & Harold J. Spaeth (1993) The Supreme Court and the Attitudinal Model. New York: Cambridge Univ. Press.

Segal, Jeffrey A., Lee Epstein, Charles M. Cameron & Harold J. Spaeth (1995) “Ideological Values and the Votes of Supreme Court Justices Revisited,” 57 J of Politics 812-23.

Sniderman, Paul M., & Thomas Piazza (1993) The Scar of Race. Cambridge: Belknap Press of Harvard Univ. Press.

Spann, Girardeau A. (1993) Race Against the Court. New York: New York Univ. Press.

Stimson, James A. (1999) Public Opinion in America. Boulder: Westview Press. Stimson,James A., Michael Makuen & Robert S. Erikson (1995) “Dynamic Representation,” 89 American Political Science Rev. 543-65.

Tuch, Steven A., & Michael Hughes (1996) “Whites’ Racial Policy Attitudes,” 77 Social Science Q 723-45.

Wasby, Stephen L., Anthony D’Amato & Rosemary Metrailer (1977) Desegregation from Brown to Alexander. Carbondale: Southern Illinois Univ. Press. West, Cornel (1993) Keng Faith. New York: Roudedge.

Woodward, C. Vann (1974) The Strange Career of Jim Crow. New York: Oxford Univ. Press.

Cases Cited

Brown v. Board of Education, 347 U.S. 483 (1954) Griggs v. Duke Power Co., 401 U.S. 424 (1971) (See also Appendix.)

Francine Sanders Romero is Assistant Professor in the Division of Social and Policy Sciences at the University of Texas at San Antonio.

She has studied the implication of institutional setting and public opinion on the formulation and implementation of civil rights policies, with previous efforts examining federal district courts and the U.S. Congress. She has also conducted research on institutional policymaking in other realms, with a current focus on land use.

Copyright Law and Society Association 2000

Provided by ProQuest Information and Learning Company. All rights Reserved