Ethnic and National Attachment in the Rainbow Nation: The Case of the Republic of South Africa

Abstract

Previous work concerning the interface between racial and national identification within multiracial states has suggested that dominant racial groups tend to express a firmer grip on ownership of and identification with the nation than is the case for racial minorities. This can occur despite inclusionary political rhetoric to the contrary and within nations regarded as civic rather than ethnic states. In this article, we explored the degree to which there were asymmetries in the interface between racial and national identities within the nation of South Africa, a state whose current political dispensation was founded on the principles of racial pluralism. We examined a large sample of South African citizens from the four officially recognized racial categories: Africans, Whites, Coloreds, and Indian/Asians. The results showed mixed support for the idea of South Africa as a “Rainbow Nation.”

Keywords

racial identity national identity social dominance orientation multiculturalism

The “Rainbow Nation” is a term originally proposed by Archbishop Desmond Tutu to depict the ideal of a postapartheid and democratic South Africa. This envisioned postapartheid nation is not only multiracial, a characteristic that describes many nations, but is a multiracial society in which all racial groups are equally valorized participants in the national project of creating an egalitarian and democratic society. President Nelson Mandela elaborated on this idea in his inauguration speech on taking office as the first democratically elected President of South Africa in 1994. He declared, “Each of us is as intimately attached to the soil of this beautiful country as are the famous jacaranda trees of Pretoria and the mimosa trees of the bushveld—a rainbow nation at peace with itself and the world” (Mandela, 1994). Furthermore, South Africa is probably one of the very few multiracial states in which the principles of equality, nonracialism, and nonsexism are etched into the constitution of the nation and, thus, function as the founding ideology. This constitutionally inscribed nonracialism is especially impressive, given the thoroughly implemented system of apartheid and oppression imposed on the non-White population of South Africa in 1948, until apartheid’s overthrow in 1994.

This is a fundamentally different conception of what it means to be a citizen compared with most other nations, where citizenship and belongingness to the nation are defined in ethnic/racial terms rather than in civic terms, where citizenship and national belongingness are determined by one’s devotion to race-neutral doctrines of civic life (e.g., support of democracy, pluralism, and rule of law). Even the United States, which considers itself a civic rather than an ethnic state, has functioned and continues to function as an ethnic state, where to be a “real American,” one must be White (see, for example, Devos & Banaji, 2005).

The civic and pluralist vision of the multiracial society is one in which all the society’s constituent ethnic/racial communities have unfettered and equal space to participate in the national project. Unfortunately, with very few exceptions, this is an ideal, which is seldom realized within multiracial/multiethnic societies. Rather, consistent with the logic of social dominance theory (Sidanius & Pratto, 2001), most multiethnic societies can be more accurately described as “Herrenvolk Democracies,” in which dominant ethnic/racial groups exercise disproportionate control of the state, its symbols, and its institutions, and which function for the disproportionate benefit of their own dominant constituents (see, for example, Peña & Sidanius, 2002; Sidanius, Feshbach, Levin, & Pratto, 1997; Sidanius & Petrocik, 2001; Sinclair, Sidanius, & Levin, 1998; Staerklé, Sidanius, Green, & Molina, 2010; Vickery, 1974).

Using the interpretive framework offered by social dominance theory (see also Pratto, Sidanius, & Levin, 2006), Sidanius et al. (1997) reasoned that one of the characteristics of Herrenvolk Democracies is an asymmetrical interface between ethnic and national identities across the dominance hierarchy. This asymmetry manifests by dominants experiencing a greater sense of ownership of, and identification with, the nation and its sacred symbols than is the case for subordinates. Because the state functions for the disproportionate benefit of dominant ethnic groups, one should expect that among dominants, there should be a more positive relationship between psychological identification with one’s ethnicity and identification with the nation. In contrast, because the state allocates less positive social value to subordinates, among them, one should observe a less positive or even a negative relationship between ethnic and national identities (see also Subhasish, 2018). This is an example of what social dominance theorists have labeled as “ideological asymmetry” and which has been found across a range of studies (e.g., Sidanius et al., 1997; Sidanius & Petrocik, 2001; Sinclair et al., 1998; Staerklé et al., 2010). Thus, most empirical research exploring the nature of the relationship between ethnic and national identities to date has been more consistent with a Herrenvolk hierarchy rather than a “Rainbow Nation” narrative. For example, Sidanius et al., 1997 found that (a) dominants (e.g., European Americans) were substantially more patriotically attached to, and identified with, their country (i.e., United States of America) than were subordinates (i.e., African Americans); (b) the correlation between patriotic attachment to the nation and attachment to one’s ethnic subgroup, whereas being positive among dominants, was negative among subordinates; and (c) attachment to the nation was positively associated with endorsement of ideologies of exclusion, classical racism, and social dominance orientation (SDO) among dominants (i.e., Whites), whereas being more negatively associated with these exclusionary ideologies among subordinates (e.g., Asian Americans, Latino Americans, and Black Americans; see also Peña & Sidanius, 2002). More recently, Staerklé et al. (2010), found similar results using multilevel analyses among respondents of some 33 nations around the globe.

Although some degree of Herrenvolk hierarchy seems to describe the relationship between ethnic and national identities in many multiracial/multiethnic societies, Sawyer, Peña, and Sidanius (2004) found this not to be the case within the three Caribbean and multiracial nations of Puerto Rico, Cuba, and the Dominican Republic. Within these three nations, there were no statistically significant differences in patriotic attachment to the nation as a function of “race,” despite the consistent finding that “Blancos” (i.e., Whites) were regarded as enjoying significantly higher social standing than either Mulatos (i.e., mixed race) and especially “Negros” (i.e., Blacks). In the case of Cuba, patriotic attachment to the nation was found to be an inverse function of racial status, such that the greater one’s African (i.e., low status) heritage, the greater one’s patriotic attachment to the nation. This finding was interpreted as caused by the antiracialist discourse and the domestic and foreign policies pursued by the Cuban Government. Moving beyond mere rhetoric, these antiracialist policies have expressed themselves in Cuba’s extensive military, financial, and medical aid to the anticolonial and antiracialist insurgencies in Africa (see, for example, de la Fuente, 1998; see especially Falk, 1987). In other words, the data seem to suggest that there are sociopolitical contexts within which the standard asymmetrical relationship between ethnic and national identities, specifically where dominant social groups exhibit higher levels of identification with the superordinate construct of the nation, might not hold.

Given the 1994 transition from a state in which political power rested exclusively in the hands of South African Whites, into a political regime based on majority rule, and, thus, largely in the hands of South African Blacks, and using the Cuban case as a model, there is at least some reason to expect reverse asymmetry in the interface between ethnic and national identities in South Africa. This is to say that one might expect a stronger relationship between one’s ethnic/racial and national identification among Blacks than among Whites. However, if the egalitarian discourse of the “Rainbow Nation” is determinative, one should expect a symmetrical interface between racial and national identities across the South African racial hierarchy such that the connection between one’s ethnic/racial and national identities are equally strong across all major ethnic/racial groups.

Exploring the “Rainbow Nation” status of South Africa would be particularly interesting given its four major racial categories. The official and major racial groups in South Africa are “Black Africans” (80.8%), Coloreds (a mixed racial group including people of British, Irish, German, Mauritian, South Asians, Zulu, and Xhosa; 8.6%), Whites (8.0%), and South Asians (2.6%; see Statistics South Africa, 2016). Blacks not only are the racial majority in the country but also hold the lion’s share of political power and have the status of being the aboriginal people of South Africa. However, although Whites are a racial minority, they possess the greatest wealth per household (Orthofer, 2016), and they have played a prolonged and exclusive role in governing the country up until 1994 when the country became a full democracy.

Despite South Africa’s relatively new status as a democracy, it is widely recognized that a vastly disproportionate share of the nation’s economic wealth remains in the hands of Whites, which has contributed to a high and continuing level of racial tension within the nation (Colin, 2016; but see also Subhasish, 2018). However, consistent with the expectations of the common ingroup identity model (Gaertner, Dovidio, Anastasio, Bachman, & Rust, 1993), positive identification with both an inclusive superordinate social category, such as a Rainbow Nation, and one’s ethnoracial particularism is still possible, despite great inequalities in wealth and general social standing. One might even credibly argue that a positive connection between these two types of social identities is a prerequisite for a stable and well-functioning democracy (however, see Subhasish, 2018).

Finally, past work in the United States has shown there to be an asymmetrical relationship between identification with the nation and group-domination ideologies, such as classical racism and SDO, in theoretically expected ways. Consistent with the logic of social dominance theory, these relationships tend to be positive among dominants (e.g., White Americans) and negative among subordinates (e.g., Latinos; see; Peña & Sidanius, 2002; Sidanius et al., 1997; Sinclair et al., 1998). Social dominance theory argues that because most multiethnic societies are structured as group-based hierarchies, asymmetry in the relationship between ethnic and national identities should characterize most nations. If, however, a multiethnic/multiracial society were to function as advertised, that is, in a genuinely egalitarian fashion, then one would expect symmetry rather than asymmetry in the relationship between ethnic and national identities. Thus, the broad question is, and one not asked in the previously cited literature, does the psychological interface between ethnic and national identities follow the “Rainbow Nation” narrative, within the South African context?

To explore this broad question, one can derive four hypotheses regarding the expected relationships between ethnic/racial identity and national identity, and the relationship between national identity and the endorsement of group-based hierarchy.

Hypothesis 1: On average, people across all major ethnic/racial categories feel an equal degree of identification with the nation.

Hypothesis 2: The degree to which one identifies with the superordinate construct of the “nation,” will be greater than the degree to which one identifies with one’s ethnic/racial subgroup within all major ethnic/racial categories.

Hypothesis 3: There will be an equally positive correlation between one’s ethnic/racial and national identities within all major ethnic/racial groups.

Hypothesis 4: There will tend to be a negative relationship between national identity and ideologies of group-based hierarchy and dominance across all major ethnic/racial groups.

Method

To explore these hypotheses, we contacted 2,390 students enrolled at the University of South Africa in the spring of 2017. Only those students with South African citizenship and who belonged to one of the four major racial categories (i.e., Blacks, Whites, Coloreds, South Asians) and who completed the survey, were included in the analyses. This left a total analyzable sample of 1,577 respondents, 66% of whom were women. The ethnic/racial breakdown of the sample was Blacks (70%), Whites (18%), Coloreds (8.0%), and South Asians (5%). These proportions were not terribly unlike the ethnic breakdown of the South African population (Central Intelligence Agency, 2018; i.e., Blacks = 80.2%, Whites = 8.4%, Coloreds = 8.8%, and South Asians = 2.5%). Thus, compared with the population of the country, in our sample of university students, Blacks were underrepresented, whereas Whites and South Asians were overrepresented. These disproportions are not surprising, given the history of apartheid in South Africa.

The respondents were given a survey questionnaire asking them, among other things, to indicate the degree to which they identified with their race and with the nation of South Africa.

Operationalization of Variables

Racial identification was defined by use of three, 5-point Likert-type scales. The items and the response alternatives were as follows: (a) “How strongly do you identify with your race?” (1 = very weakly to 5 = very strongly). (b) “How often do you think of yourself as a member of your race?” (1 = very seldom to 5 = very often). (c) “How important is your race to your sense of who you are?” (1 = very unimportant to 5 = very important; α = .77).

National identification was defined by use of three questions using 5-point Likert-type response scales. The questions read as follows: (a) “How strongly do you identify with the country of South Africa?” (1 = very weakly to 5 = very strongly). (b) “How often do you think of yourself as a South African?” (1 = very seldom to 5 = very often). (c) “How important is being a South African to your sense of who you are?” (1 = very unimportant to 5 = very important; α = .81).

SDO is a self-report measure, which assesses the degree to which people endorse and support antiegalitarian and hierarchically structured intergroup relations. We employed the SDO₇ version of the scale (see, for example, Ho et al., 2015).

Results

The means, standard deviations, and intercorrelations among the three focal variables (i.e., racial identity, national identity, and SDO) are found in Table 1. As can be seen in this table, there was a substantial correlation between national and racial identities (r = .40, p < .01), whereas there were modest but significant negative correlations between SDO and national identity (r = −.06, p < .05); there was a slight, but small positive correlation between SDO and racial identity (r = .08, p < .01).

Table 1.

Descriptive Statistics of Focal Variables.

Variables	M	SD	1	2
1. National identification	4.39	.84
2. Racial identification	4.10	.98	.40**
3. Social dominance orientation	2.38	.99	−.06*	.08**

p < .05. **p < .01.

It is well known that when conducting cross-cultural research, it is important to be assured that the meaning of the survey responses be relatively uniform across group boundaries. This issue is referred to as the problem of measurement invariance. There are approximately three major types of measurement invariance, the simplest of which is configural invariance. This type of invariance is one in which the number of underlying latent continua and the pattern of factor–indicator relationships are identical across groups (e.g., Ariely & Davidov, 2012; Steenkamp & Baumgartner, 1998).

We assessed the configural invariance of the two social identity measures (i.e., racial and national) by use of principal components analysis. If there was general agreement across racial category that these two scales were assessing the same latent dimensions, then we expected to find a clear two factor solution within each ethnic/racial group, in which one factor was defined by the three racial identity items and the second factor was defined by the national identity items. This is precisely what the data indicated. The principal components provided a clear simple structure for the entire sample (see Table 2), and this simple structure was replicated within each of the four racial categories.

Table 2.

Principal Components Analysis of Racial and National Identities for the Entire Sample.

National identity	Racial identity
1. How important is being a South African to your sense of who you are?	.81
2. How often do you think of yourself as a South African?	.87
3. If you were born in South Africa, how strongly do you identify with the country?	.86
4. How important is your race to your sense of who you are?	.85
5. How often do you think of yourself as a member of your race?	.82
6. How strongly do you identify with your race?	.81

Note. Blank loadings were ⩽ .25; correlation between components = .39.

Unlike the racial or national identification scales, the SDO₇ consists of a much larger number of items (i.e., 16). Furthermore, the sample sizes of two of the four ethnic/racial categories were relatively small. Therefore, because principal components is a large sample technique, use of principal components within each group was considered inappropriate. Instead, to ascertain the configural measurement invariance, we divided the sample into two broad ethnic/racial categories: Blacks (n = 1,096) and non-Blacks (n = 482). Inspection of the scree plots for each of the racial categories indicated only four substantive principal components among both Blacks and non-Blacks. In addition, the pattern of the factor loadings of the variables on the four principal components was the same for both Blacks and non-Blacks. These results indicated that a high degree of configural measurement invariance held across the Black and non-Black groups in South Africa. Furthermore, not only did configural measurement invariance hold across Black and non-Black South Africans, there was also a high degree of configural measurement equivalence when comparing the four-dimensional structure in South Africa with the four-dimensional structure found within scale validation studies in the United States (see Ho et al., 2015; see also Table 3).

Table 3.

Four-Dimensional Factor Structure of SDO₇ Scale for Total South African Sample.

Antiegalitarianism/contrait 1. We should do what we can to equalize conditions for different groups (.87). 2. No matter how much effort it takes, we ought to strive to ensure that all groups have the same chance in life (.86). 3. We should work to give all groups an equal chance to succeed (.83). 4. Group equality should be our ideal (.80).Dominance/protrait 5. An ideal society requires some groups to be on the top and other groups to be at the bottom (.81). 6. It is probably a good thing that certain are at the bottom (.80). 7. Some groups of people are simply inferior to other groups (.70). 8. Some groups of people must be kept in their place (.69).Dominance/contrait 9. Groups at the bottom should not have to stay in their place (.75). 10. Group dominance is a poor principle (.70). 11. No one group should dominate in society (.66). 12. Groups at the bottom are just as deserving as groups at the top (.63).Antiegalitarianism/protrait 13. We should not push for group equality (.75). 14. Group equality should not be our primary goal (.72). 15. It is unjust to try to make groups equal (.68). 16. We should not try to guarantee that every group has the quality of life (.66).(Entries) are factor loadings. Factors are rotated with oblimin = 0.

Note. SDO = social dominance orientation.

Table 4.

Two-Stage OLS Multiple Regression Analysis of South African Identification as a Function of Four Background Variables and Racial Category.

	B	95% CI [lower bound, upper bound]
Stage 1
Gender	0.04	[–0.05, 0.14]
Age	0.01**	[0.01, 0.02]
Social class	−0.01	[–0.05, 0.02]
Academic year	−0.02	[–0.04, 0.01]
R ² _adj	.01**
Stage 2
Gender	0.00	[–0.09, 0.10]
Age	0.01**	[0.01, 0.02]
Social class	0.01	[–0.02, 0.05]
Academic year	−0.01	[–0.04, 0.02]
Whites vs. Blacks	−0.47**	[–0.59, –0.35]
Coloreds vs. Blacks	−0.29**	[–0.46, –0.12]
South Asians vs. Blacks	−0.02	[–0.24, 0.20]
R ² _adj	.06**

Note. OLS = ordinary least squares; CI = confidence interval.

p < .01.

Major Hypotheses Testing

After addressing the issue of configural measurement invariance, we went on to test our four substantive hypotheses. The first of which was that all four major ethnic/racial categories would be equally identified with the nation. This was tested by use of a two-stage hierarchical ordinary least squares (OLS) multiple regression analysis using identification with the nation as the dependent variable and the four background variables (i.e., gender, age, social class, and academic year) as controls variables at Stage 1. Next, we added three dummy vectors capturing the variance in national identity associated with ethnicity/race at Stage 2. The results from Stage 1 showed that the four background variables accounted for a statistically significant, but very slight, amount of the variance in national identification, that is, R²_adj = .01, F_{(4, 1,259)} = 4.56, p < .01. The addition of the three dummy vectors representing the four racial categories (where Blacks served as the reference category) were entered into the regression equation at Stage 2. Next to the influence of the four control variables, the addition of the ethnic/racial category information resulted in a significant increase in the amount of variance in national identity accounted for, that is, R²_adj = .05, F_{(4, 1,256)} = 22.32, p < .01; See Table 4.

Two of the three unstandardized multiple regression coefficients representing the ethnic/racial categories were statistically significant. Blacks were significantly more South Africa identified than Whites (B = −0.47, t = −7.85, p < .001, confidence interval [CI] = [−0.59, −0.35]) and Blacks were more South Africa identified than Coloreds (B = −0.30, −0.12, t = −3.38, p < .001, CI = [−0.59, −0.35]). However, Blacks were not significantly more South Africa identified than South Asians (B = −0.02, t = −0.19, p < .85, CI = [−0.24, −0.20]).

A second approach exploring the validity of the “Rainbow Nation” narrative was executed by comparing the strength of respondents’ racial identity versus the strength of their national identity. Following the logic of the “Rainbow Nation,” one would expect one’s national identity to be more compelling than one’s racial particularism (Hypothesis 2). Inspection of the relative strengths of national and racial identity ratings seemed to confirm this Rainbow Nation expectation (see Table 5).

Table 5.

Comparisons of Racial and National Identities Within Each Racial Group.

Racial group	Racial identity	National identity	F	p <	_partialη²
Blacks	4.37	4.48	21.85**	.00	.02
Whites	3.42	4.09	84.32**	.00	.23
Coloreds	3.57	4.24	42.37**	.00	.26
South Asians	3.79	4.46	25.68**	.00	.27

p < .01.

To take a closer look at the question raised above, we deployed a two-way mixed effects ANOVA, with identification target (i.e., racial vs. national identification) as the within-subjects factor, and race of the respondent as the between-subjects factor. Although there were racial differences in degree of identification with both race and nation, F_{(3, 1,569)} = 78.97, p < .01, _partialη² = .13, more pertinently, national identity appeared to be the more compelling social identity than “race” across all four racial categories, F_{(1, 1,569)} = 183.92, p < .01, _partialη² = .11.

However, the degree to which nation was a more powerful social identity than race was not uniform across the four ethnic/racial categories. This is witnessed by the fact that the interaction between race of subject × type of identification (i.e., racial vs. national) was statistically significant, F_{(3, 1,569)} = 36.52, p < .01, _partialη² = .07. As seen in Figure 1, the gap between the degree of racial identification versus degree of national identification was much smaller among Blacks than among the other three ethnic/racial categories.

Figure 1.

Degree of social identification by target and race.

The third manner of assessing the validity of the “Rainbow Nation” narrative is by inspecting the correlation between racial and national identifications (Hypothesis 3). If the Rainbow Nation narrative is descriptively accurate, then we should expect to find evidence of an equally positive correlation between one’s ethnic/racial and national identities within all major ethnic/racial groups in South Africa. To address this question, we regressed South African identity on racial identity within each of the four racial categories while using the four background variables as controls (see Table 6).

Table 6.

Simultaneous Multiple Regression With National Identification as a Function of Four Background Variables and Racial Identification Within Each Racial Group (Entries Are Unstandardized Regression Coefficients).

	Blacks	Whites	Coloreds	South Asians
Variables
Gender	0.05	−0.22	−0.10	0.30
Age	0.01**	0.02**	0.01	0.03
Academic year	0.00	−0.05	0.07	−0.03
Social class	0.01	−0.03	−0.07	0.09
Racial identity	0.39***	0.25**	0.38***	0.28*
R ² _adj	.21***	.11***	.15**	.10

p < .05. **p < .01. ***p < .001.

Inspection of Table 6 reveals that the relationship between national and racial identities was significant and positive among all four major ethnic/racial groups. Thus, the more one identified with one’s racial subgroup, the more one identified with the South African nation. This finding seems clearly supportive of the “Rainbow Nation” narrative.

However, although the answer to our third question offers some support for the “Rainbow Nation” narrative, the support is somewhat qualified. Namely, although this racial identity/national identity interface was positive among all ethnic/racial groups, this interface appeared to be stronger for some groups than for others. For example, the interface between racial and national identities was strongest among Blacks (b = 0.39, p < .01), and although still significantly positive, appeared to be substantially weaker among Whites (b = 0.25, p < .01) and South Asians (b = 0.28, p < .05).

To explore whether these slope differences were significant, we employed OLS regression analysis and simple slope analyses. Once again assigning Blacks as the reference category, we computed three two-way interactions for the regression of national identification on racial identification for Blacks versus each of the other three ethnic/racial categories. Including the control variables, these analyses showed that the slopes for the relationship between national and racial identities were significantly stronger among Blacks than among South Asians (b = −0.19, t = −2.24, p < .05, 95% CI = [−0.36, −0.02]) and Whites (b = −0.13, t = −2.39, p < .05, 95% CI = [−0.36, −0.02]). The slope contrast for Blacks versus Coloreds was not significant (b = −0.02, t = −0.29, p = .77, 95% CI = [−0.15, 0.11]; see Table 7).

Table 7.

Analyses Comparing Relationship Strengths of National Identity Regressed on Racial Identity for Whites, Coloreds, and South Asians vs. Blacks.

Variables	Whites vs. Blacks	Coloreds vs. Blacks	South Asians vs. Blacks
Gender	−.01	.03	.06
Age	.01***	.01***	.01***
Academic year	−.01	.01	.00
Social class	.00	.00	.02
Racial identity	.40***	.40**	.40***
Whites vs. Black	.38	—	—
Coloreds vs. Black	—	.13
South Asian vs. Black	—	—	.96
Whites vs. Blacks × Racial identity	−.13*	—	—
Coloreds vs. Blacks × Racial identity	—	.02	—
South Asians vs. Blacks × Racial identity	—	—	−.19**

p < .05. **p < .01. ***p < .001.

Fourth, and finally, we assessed the validity of the “Rainbow Nation” narrative by examining the relationship between national identification and preference for group-based hierarchy and inequality (e.g., SDO; Hypothesis 4). In contrast to what has been found in the United States (Sidanius et al., 1997), where national identification was positively associated with SDO among Whites, yet negatively correlated with SDO among Blacks, if the notion of the “Rainbow Nation” is descriptively accurate of South Africa, identification with the nation should be unrelated, or even negatively related, to SDO within each of the major ethnic/racial categories. We might expect a negative relationship between SDO and national identity because those endorsing the “Rainbow Nation” narrative should also reject endorsement of group-based hierarchy and social inequality. To examine this question, we regressed national identity on SDO within each of the four major racial categories while controlling for the four backgrounds variables (i.e., age, gender, social class, and academic year; see Table 8).

Table 8.

South African Identity as a Function of Racial Identification, Control Variables, and SDO Within Each Racial Category.

	Blacks	Whites	Coloreds	South Asians
Variables
Gender	0.05	−0.14	−0.12	0.30
Age	0.01**	0.02**	0.01	0.03
Academic year	0.00	−0.06	0.07	−0.03
Social class	0.01	−0.03	−0.07	0.09
Racial identity	0.39***	0.28***	0.39***	0.28*
SDO	0.01	−0.17**	−0.09	0.04
R ² _adj	.21***	.11***	.15**	.08

Note. SDO = social dominance orientation.

p < .05. **p < .01. ***p < .001.

Given the spirit of the “Rainbow Nation” narrative, and in contrast to what has been found in the United States, one should expect no positive associations between attachment to the nation and endorsement of and desire for group-based inequality and group-based hierarchy. Employing simple OLS regression analyses, we computed the relation between national attachment and SDO, controlling for the same four background variables used previously. In contrast to what has been found in the United States (e.g., Peña & Sidanius, 2002; Sidanius et al., 1997), the South African data showed a tendency for identification with the nation to be largely independent of support for group-based inequality or SDO. In the case of South African Whites, there was even a significantly negative relationship between these two constructs (i.e., b = −0.17, t = −2.64, p < .01, 95% CI = [−0.29, −0.04]). In other words, among White South Africans, the more one rejected ideologies of group-based dominance and inequality, the greater one’s identification with the South African nation. This indicates that, at least for South African Whites, the Rainbow Nation narrative appears to hold.

Summary and Discussion

This article examined the degree to which the dynamics of ethnic/racial and national identities in South Africa were congruent with the expectations of racial pluralism, or what has been described as the “Rainbow Nation.” Racial pluralism, or the “Rainbow Nation” ideal, is not simply a matter of a society being multiracial, but a society in which all constituent communities also share a common national identity and are regarded as coequal and legitimate partners in the national project. Past evidence has revealed that although a number of societies are multiracial, to one degree or another, the relationships among the nation’s racial constituencies are one of group-based hierarchy and racial exclusion rather than racial pluralism or anything resembling a “Rainbow Nation” (e.g., the United States, the United Kingdom, Myanmar, Russia, France, Israel, Persia, and Nigeria; see also Devos & Banaji, 2005; Sidanius et al., 1997; Sidanius & Petrocik, 2001; Sinclair et al., 1998; Staerklé et al., 2010).

Given the fact that the founding ideology of the postapartheid South African state is explicitly antiracialist and inclusionary, we were interested in whether the ideals of the “Rainbow Nation” could be said to genuinely describe the dynamics of South African identity. We have assessed the degree to which the “Rainbow Nation” narrative has been achieved by the use of four analyses: (a) assessing the degree to which all citizens, regardless of ethnic/racial category, identify equally strongly with the nation as a whole; (b) exploring whether or not national identity is a more compelling social identity than racial identity, and is so among all ethnic/racial categories; (c) observing the degree to which there is a positive rather than a negative relationship between identification with one’s ethnic/racial subgroup and identification with the nation, across all ethnic/racial categories; and (d) determining whether or not there is a negative, rather than a positive, relationship between identification with the nation, on one hand, and the endorsement of exclusionary and hierarchical relationships between groups.

Somewhat in contrast to the findings of Sidanius et al. (1997), Sidanius and Petrocik (2001), and Staerklé et al. (2010), our study here did find some support for the “Rainbow Nation” trope. Although Blacks were more identified with the nation than were Whites and Coloreds, consistent with data from Cuba (e.g., Sawyer et al., 2004), there was only the slightest difference in national attachment between Blacks and South Asians. These latter findings were congruent with a weak version of the “Rainbow Nation” discourse.

Furthermore, in support of the expectations of racial pluralism was the fact that national identity seemed to be the more compelling social identity than racial identity, and among all four ethnic/racial groups. One could reasonably argue that this greater commitment to the superordinate, national identity, rather than to a more parochial, ethnic/racial attachment is critically important for the stability and coherence of a multiracial society and is clearly consistent with the spirit of the “Rainbow Nation.”

However, these results did not present entirely unequivocal support for the “Rainbow Nation” narrative. Although national identity clearly appeared to be the more important social identity among all major ethnic/racial groups in South Africa, the degree to which national identity trumped racial identity was not completely uniform across racial categories. Specifically, the “compellingness” of nation over race was noticeably weaker among the racial majority group (i.e., Black South Africans) than among the other three ethnic/racial groups. Taking the “Rainbow Nation” narrative seriously, the relatively weak compellingness of nation over race among the majority Black population could be regarded as somewhat unsettling.

Be that as it may, a third piece of evidence that is largely congruent with the “Rainbow Nation” narrative is the nature of the covariation between national and racial identities across racial groups. Unlike the findings of Sidanius et al. (1997), who found an asymmetrical relationship between racial and national attachment (i.e., positive among Whites vs. negative among Blacks), in this South African sample, the relationship between racial identity and national identity was positive across all four racial categories. This symmetry in the relationship between these two social identities would also seem to be consistent with the “Rainbow Nation” narrative, in that, feeling identified with one’s racial particularism is consonant rather than discordant with attachment to the nation. However, here as well, this support for the “Rainbow Nation” narrative was qualified by the fact that the degree of overlap between national and racial identities was not of equal magnitude across all four ethnic/racial groups. The degree of overlap between these two identities was strongest among Blacks and weakest among South Asians. The fact that this identity overlap was strongest among South African Blacks is not surprising given that Blacks represent, by far, the largest portion of the South African population.

Fourth, in contrast to the findings in the United States and Israel (see Sidanius et al., 1997; Sidanius & Petrocik, 2001), these analyses showed no evidence that attachment to the nation was associated with ideologies of racial exclusion, antiegalitarianism, or preference for group-based hierarchy (see also Guimond et al., 2013; Verkuyten, 2005).

Altogether, keeping in mind Sawyer et al.’s (2004) findings in Cuba, these South African data seem to suggest that the notion of the “Rainbow Nation” is not simply a comforting delusion, but is, at least to some meaningful extent, an actual lived reality. These are important findings because they imply that racially diverse societies can accommodate a positive-sum, rather than a zero-sum, relationship between one’s national and ethnic/racial identities, even within a nation that has had a recent history of brutal racial oppression. Furthermore, these results imply that national identity within racially diverse societies need not necessarily imply ideologies of exclusion, ethnoracial parochialism, and group-based social inequality. At the same time, evidence seems to suggest that if subgroups are treated with dignity and respect, superordinate social cohesion does not have to come at the cost of suppressing or abandoning subgroup loyalty (see also Berry, 1991; Huo, Smith, Tyler, & Lind, 1996; Moghaddam & Solliday, 1991; Subhasish, 2018).

Some Caveats

Be that as it may, there are at least five reasons to embrace our conclusions with a certain degree of circumspection. First, we must remind ourselves that South Africa is a considerably more ethnoracially complex society than could be accounted for in this study using the four officially recognized ethnic/racial categories. South Africa is far more ethnically complex than “racially” complex, consisting of some 14 separate ethnic groups and embracing 11 official languages. It is not certain that our major conclusions would be largely invariant had we been able to examine the interface between nation and ethnicity in addition to the interface between nation and race. Extending these inquiries into the intersection of ethnicity and national identity could suggest an array of questions that should be tackled in future research.

Second, we have utilized a university sample at one university. Naturally, it is not clear whether the findings here would generalize to a probability sample from the entire population of South African adults. However, the use of a university sample can be informative and important because today’s university students will comprise tomorrow’s local, regional, and national elites, and, as such, will have a disproportionate impact on the dynamic relations between national and social identities.

Third, we must keep in mind that the data here provide us with only a snapshot of an ongoing relationship between national and racial identities. The postapartheid, democratic regime of South Africa is less than a quarter of a century old. It is, thus, not at all clear how stable this fragile racial democracy will be over time (see also Gibson, 2004).

Fourth, there is some reason to believe that the degree to which the interface between ethnic/racial identity and national identity is asymmetric across racial and ethnic subgroups will very much depend on precisely how one conceptualizes these social identities. For example, in recent work, Hamamura (2017) found that the relationship between national identity and diversity attitudes depended on whether one was conceptualizing social identity in collective or relational terms.

Finally, although we feel that our findings are encouragingly consistent with the “Rainbow Nation” ideal, it is also clear that we cannot unambiguously attribute this success to the founding ideals of the postapartheid South African state. Such an attribution would be pure conjecture and would need to be adjudicated by analyses beyond the scope of this article.

Nonetheless, whatever the cause, for the present at least, the previously apartheid nation of South Africa seems to have come a substantial distance in reaching the gateway to the “Rainbow Nation.”

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Ariely

Davidov

(2012). Assessment of measurement equivalence with cross-national and longitudinal surveys in political science. European Political Science, 11, 363-377.

Berry

J. W.

(1991). Understanding and managing multiculturalism: Some possible implications of work in Canada. Psychology and Developing Societies, 3, 17-49.

Central Intelligence Agency. (2018). The world factbook. Retrieved from https://www.cia.gov/library/publications/the-world-factbook/geos/rs.html

Colin

(2016, March). Raw tensions over race fester in South Africa. The New York Times. Retrieved from https://www.nytimes.com/2016/03/02/world/africa/raw-tensions-over-race-fester-in-south-africa.html

de la Fuente

(1998). Race, national discourse and politics in Cuba. Latin American Perspectives, 25, 43-69.

Devos

Banaji

M. R.

(2005). American = White? Journal of Personality and Social Psychology, 88, 447-466.

Falk

P. S.

(1987, Summer). Cuba in Africa. Foreign Affairs. Retrieved from https://www.foreignaffairs.com/articles/africa/1987-06-01/cuba-africa

Gaertner

S. L.

Dovidio

J. F.

Anastasio

P. A.

Bachman

Rust

M. C.

(1993). The common ingroup identity model: Recategorization and the reduction of intergroup bias. European Review of Social Psychology, 4, 1-26.

Gibson

J. L.

(2004). Overcoming apartheid: Can truth reconcile a divided nation? New York, NY: Russell Sage Foundation.

10.

Guimond

Crisp

De Oliveira

Kamiejski

Kteily

Kuepper

Zick

(2013). Diversity policy, social dominance and intergroup relations. Journal of Personality and Social Psychology, 104, 941-958.

11.

Hamamura

(2017). Social identity and attitudes toward cultural diversity: A cultural psychological analysis. Journal of Cross-Cultural Psychology, 48, 184-194. doi:10.1177/0022022116681845

12.

Sidanius

Kteily

Sheehy-Skeffington

Pratto

Henkel

K. E.

Stewart

A. L.

(2015). The nature of social dominance orientation: Theorizing and measuring preferences for intergroup inequality. Journal of Personality and Social Psychology, 109, 1003-1028. Retrieved from http://psycnet.apa.org/doi/10.1037/pspi0000033

13.

Huo

Y. J.

Smith

H. J.

Tyler

T. R.

Lind

E. A.

(1996). Superordinate identification, Subgroup identification: Is separatism the problem; is assimilation the answer? Psychological Science, 7, 40-45.

14.

Mandela

(1994, May 10). Inaugural presidential speech, Union building. Pretoria. Retrieved from http://www.mandela.gov.za/mandela_speeches/1994/940510_inauguration.htm

15.

Moghaddam

F. M.

Solliday

E. A.

(1991). Balanced multiculturalism and the challenge of peaceful coexistence in pluralistic societies. Psychology and Developing Societies, 3, 51-72.

16.

Orthofer

(2016). Wealth inequality—Striking new insights from tax data. Retrieved from http://www.econ3x3.org/article/wealth-inequality-%E2%80%93-striking-new-insights-tax-data

17.

Peña

Sidanius

(2002). American patriotism and ideologies of group dominance: A tale of asymmetry. Journal of Social Psychology, 142, 782-790.

18.

Pratto

Sidanius

Levin

(2006). Social dominance theory and the dynamics of intergroup relations: Taking stock and looking forward. European Review of Social Psychology, 17, 271-320.

19.

Sawyer

M. Q.

Peña

Sidanius

(2004). Cuban exceptionalism: Group based hierarchy and the dynamics of patriotism in Puerto Rico, the Dominican Republic and Cuba. The Dubois Review, 1, 93-114.

20.

Sidanius

Feshbach

Levin

Pratto

(1997). The interface between racial and national attachment: Racial pluralism or racial dominance? Public Opinion Quarterly, 61, 103-133.

21.

Sidanius

Petrocik

J. R.

(2001). Communal and national identity in a multiracial state: A comparison of three perspectives. In Ashmore

R. D.

Jussim

Wilder

(Eds.), Social identity, intergroup conflict, and conflict reduction (pp. 101-129). New York, NY: Oxford University Press.

22.

Sidanius

Pratto

(2001). Social dominance: An intergroup theory of social hierarchy and oppression. New York, NY: Cambridge University Press.

23.

Sinclair

Sidanius

Levin

(1998). The interface between racial and social system attachment: The differential effects of hierarchy-enhancing and hierarchy attenuating environments. The Journal of Social Issues, 54, 741-757.

24.

Staerklé

Sidanius

Green

Molina

(2010). Racial minority-majority asymmetry in national attitudes around the world: A multilevel analysis. Political Psychology, 31, 491-519.

25.

Statistics South Africa. (2016). Mid-year population estimates, 2016. Retrieved from https://www.statssa.gov.za/publications/P0302/P03022016.pdf

26.

Steenkamp

J. E. M.

Baumgartner

(1998). Assessing measurement invariance in cross-national consumer research. Journal of Consumer Research, 25, 78-90.

27.

Subhasish

(2018). Ethnic inequality and national pride. Political Psychology, 39, 263-280. doi:10.1111/pops.12406

28.

Verkuyten

(2005). Ethnic group identification and group evaluation among minority and majority groups: Testing the multiculturalism hypothesis. Journal of Personality and Social Psychology, 88, 123-138.

29.

Vickery

K. P.

(1974). “Herrenvolk” democracy and egalitarianism in South Africa and the U.S. South. Comparative Studies in Society and History, 16, 309-328.