Minority Group, Extra Credit

Oxford University Press is collaborating with JSTOR to digitize, preserve and extend access to Social Forces.

http://www.jstor.org

White, Black, or Puerto Rican? Racial Self-Identification among Mainland and Island Puerto Ricans Author(s): Nancy S. Landale and R.S. Oropesa Source: Social Forces, Vol. 81, No. 1 (Sep., 2002), pp. 231-254 Published by: Oxford University Press Stable URL: http://www.jstor.org/stable/3086533 Accessed: 09-04-2015 20:00 UTC

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

White, Black, or Puerto Rican? Racial Self-Identification among Mainland and Island Puerto Ricans*

NANCY S. LANDALE, Pennsylvania State University R.S. OROPESA, Pennsylvania State University

Abstract

Recent studies have examined the implications of exposure to U.S. race relations for the racial and ethnic identities of migrants to the U.S. Most investigations are based exclusively on U.S. data. There arefew, if any, comparisons of the identities of migrants and their offspring to those of nonmigrants in their country of origin. Using data from a survey of Puerto Rican mothers in the U.S. and Puerto Rico, this study provides such a comparison. Responses to an open-ended race question show that mainland and island Puerto Ricans most often designate their "race" as Puerto Rican, but responses of women who do not self-identify as Puerto Rican diverge between the U.S. and Puerto Rico. Island women primarily identify themselves as white, black, or triguefa, while mainland women identify themselves as Hispanic/Latina, Hispanic American, or American. Mainland-island differences cannot be explained by parental ethnicity, skin tone, demographic factors, and socioeconomic status. The findings suggest that mainland Puerto Ricans more strongly reject the conventional U.S. notion of race than do their island counterparts.

During the last several years, both the meaning and the measurement of race have resurfaced as important academic and policy concerns. The growing diversity of the U.S. population and an increase in the number of mixed-race individuals have stimulated a reconsideration of the complexities of racial self-identification and racial classification systems (Hirschman, Alba & Farley 2000; Rodriguez 2000).

* The research reported in this article was supported by the National Institute of Child Health and Human Development, the Maternal and Child Health Bureau, and the Centers for Disease Control. Support services were provided by the Population Research Institute, Pennsylvania State University. Direct correspondence to Nancy S. Landale, Population Research Institute, 601 Oswald Tower, University Park, PA 16802. E-mail: [email protected].

? The University of North Carolina Press Social Forces, September 2002, 81 ( 1):231-254

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

232 / Social Forces 81:1, September 2002

Indeed, the U.S. government instituted significant changes in its racial classification

system in 1997, when the Office of Management and Budget issued revised federal standards for the collection of data on race and ethnicity. A fundamental revision to the standards is that individuals are now allowed to select more than one racial

category to identify their racial identity. A group that figures prominently in debates about the measurement of race is

the Hispanic population. Most data collection systems include separate items on race and Hispanic ethnicity. For example, the 2000 U.S. census designated six major racial categories (white, black or African American, Asian, American Indian or Alaskan native, native Hawaiian or other Pacific islander, and some "other" race) and two ethnic categories (Hispanic and non-Hispanic). Although Hispanics were allowed to self-identify with any racial group, including multiple groups if they desired, about 42% of Hispanics rejected the conventional race categories and instead selected "other" to represent their race (http://www.census.gov/Press-Release/ www/2001/cb01cn61.html). It is commonly argued that this pattern reflects the fact that race has a different meaning for Hispanics than it does for other ethnic

groups (Rodriguez 2000). In this article, we focus on racial self-identification among Puerto Ricans, a

Caribbean Hispanic group with a complex ancestry and considerable phenotypic variation. Using data collected from a representative sample of Puerto Rican mothers in the U.S. and Puerto Rico, we address several questions: What terms do Puerto Rican women use to describe their racial identities? How closely do these terms correspond to standard racial categories? What is the role of skin color in racial identity? What other personal attributes and achievements structure racial self-identification? As each of these questions is considered, a major focus is the difference between Puerto Rican women in the U.S. and Puerto Rico.

Background

Although the census and other federal data collection systems specify racial

categories other than black and white (i.e., Asian, American Indian or Alaskan native, native Hawaiian or other Pacific islander), very few Hispanics fall into such

groups. Thus, when Hispanics select their racial identity from closed-ended

questions modeled after the census, their choice set reduces to black, white, and other. The fact that Hispanics frequently select "other" indicates the inadequacy of

existing classificatory schemes for describing the racial identities of those who trace their origins to Latin America and the Spanish Caribbean (Denton & Massey 1989;

Rodriguez 2000). The lack of fit between a black-white conception of race and the views of

Hispanics is at least partially attributable to race relations in their countries of

origin (Safa 1998). In Latin America and the Spanish Caribbean, a long history of

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 233

intermarriage between indigenous populations, Spanish colonizers, and African-

origin slaves produced a racially mixed population. As a result, race is typically regarded as a multicategory spectrum on which black and white represent end-

points (Denton & Massey 1989; Fitzpatrick 1971; Mintz 1974; Rodriguez 2000; Wade 1997). Skin color influences social position in Latin American and Caribbean societies; however, individuals of mixed racial origins are not

automatically assigned to the subordinate group in a dichotomous classification, as typically occurs in U.S. society. As noted by Winn (1992),

Compared to Latin America and the Caribbean, the United States is not just racist, but color-blind: people are either black or white. In the other Americas, a more

complex consciousness of color sees black and white, but also recognizes many shades in between. Nor do the differences stop there. In the United States, any degree of African ancestry makes a person black, while in Latin America and the Caribbean any degree of non-African ancestry means that a person is not black. (277)

In addition, factors other than skin color influence racial classification in Latin America and the Caribbean. In particular, social class is based on skin tone, and other physical and social characteristics (e.g., facial features, hair texture, language use, demeanor) play a role (Duany 1985; Fitzpatrick 1971; Mintz 1971; Rodriguez 2000; Torres 1998). Thus, definitions of race are more flexible and ambiguous than is the case in the U.S.

Like other Latin American and Caribbean groups, Puerto Rico's population has a complex ancestry and a wide range of physical characteristics. Intermediate racial

categories are recognized, as indicated by terms such as trigueio and moreno

(Fitzpatrick 1971; Rodriguez 2000; Rogler 1944).3 At the same time, Puerto Rican islanders often view race as equivalent to nationality, culture, or birthplace (Mills, Senior & Goldsen 1950; Rodriguez 2000). The mismatch between the way race is viewed in Puerto Rico and the closed-ended categories typically used in the U.S. is evident in the exclusion of race questions in the U.S. census for Puerto Rico from 1950 through 1990 (Rivera-Batiz & Santiago 1996).

MIGRATION AND RACIAL IDENTITY

A number of scholars maintain that the construction of a racial and ethnic identity is an ongoing process that involves negotiation between an individual and others (Omi & Winant 1994; Portes & Rumbaut 2001; Rodriguez & Cordero-Guzman 1992; Rodriguez 2000; Waters 1999). Self-definitions shift over time and across social contexts. Waters (1999) notes the situational flexibility of identities: In the course of their everyday lives, individuals may emphasize a certain identity in one situation and a different identity in another. In addition, racial and ethnic identities

may be challenged when individuals change their place of residence. In particular,

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

234 / Social Forces 81:1, September 2002

migrants often encounter different definitions of their racial identities in origin and destination locales.

The influence of social context on identity is especially salient for Puerto Ricans because of the frequency of migration between Puerto Rico and the U.S.

During the latter half of the twentieth century, Puerto Rico experienced a very high rate of outmigration, and most outmigrants settled in the U.S. (Rivera-Batiz &

Santiago 1996). As a result, roughly 40% of mainland Puerto Ricans were born in Puerto Rico.

One of many adjustments required of Puerto Rican migrants to the mainland is finding their place in the U.S. system of race relations. Not only must migrants adapt to a change in their status from majority group member to minority group member; they also face pressure to redefine themselves in terms of the black-white

dichotomy that delineates race relations in the U.S. (Denton & Massey 1989; Fitzpatrick 1971). In many encounters, relatively dark-skinned Puerto Ricans are

regarded as black because of their mixed racial origins. Yet they are often reluctant to assume a black identity because they recognize that black people are stigmatized in the U.S.

Puerto Rican migrants frequently respond to this dilemma by rejecting the U.S. racial dichotomy altogether. Consistent with the way race is often conceptualized in Puerto Rico, many migrants emphasize a national identity as their "race"; that is, they define their race as Puerto Rican. Others develop a Hispanic/Latino identity as they recognize the similarities between their social position and that of other U.S. Hispanic subgroups (Nelson & Tienda 1985; Portes & Rumbaut 1990). As stated by Rodriguez (2000), "Latinos' choice of an ethnic descriptor, as opposed to a racial descriptor, reflects the disinclination of many Hispanics to identify as black"

(148).

ASSIMILATION AND RACIAL IDENTITY

Like Puerto Ricans, immigrants from a variety of countries define themselves

primarily in terms of their national origins (Portes & Rumbaut 2001; Rodriguez 2000; Waters 1999). In contrast, their U.S.-born offspring have more variable racial identities that depend on how they are incorporated into U.S. society (Portes & Rumbaut 2001). Second-generation residents who are oriented toward their origin culture and well integrated into ethnic communities are more likely to retain a national identity than are second-generation residents who are more assimilated into U.S. society. At the same time, the implications of assimilation are not

straightforward. Migrants to the U.S. encounter different contexts of reception. After

considering factors such as their skin color, their socioeconomic resources, and the strength of their ethnic communities, they may assimilate into various segments of the U.S. social structure. This, in turn, defines the social context in which racial identities develop and change.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 235

Among Puerto Ricans, several factors critical to the assimilation process are

likely to influence racial self-identification. First is their mixed racial heritage, reflected in skin tones that range from very light to very dark. Skin tone is important for racial self-identification because it is the most salient feature that others use to define race (Brown, Dane & Durham 1998). Puerto Ricans with relatively dark skin are more likely than their light-skinned counterparts to be defined by others as black. Skin tone is also important because of its implications for socioeconomic assimilation. Numerous recent studies show that skin tone has an influence on socioeconomic outcomes, net of socioeconomic background (Hughes & Hertel 1990; Hunter 1998; Keith & Herring 1991). While most of this research focuses on African Americans, similar findings have been reported for Mexican Americans

(Murguia & Telles 1996; Telles & Murguia 1990). Thus, it is likely that Puerto Ricans with dark skin face greater barriers to socioeconomic mobility than those with light skin.

The other key factor that shapes the assimilation process is the relatively small human and financial capital of Puerto Rican migrants (Bean & Tienda 1987; Rivera-Batiz & Santiago 1996). Because they have few economic resources, Puerto Rican migrants frequently settle in central-city areas where they encounter

inadequate employment opportunities, discrimination, and other disadvantaged minority groups. Both their struggle with difficult conditions and their exposure to other groups that have experienced discrimination may erode the optimism of Puerto Rican migrants (Portes & Rumbaut 2001). While members of the first

generation are often partially protected by their strong coethnic ties and national

identity, their U.S.-born offspring are less inclined to use the origin country as a frame of reference and more inclined to interact extensively with African Americans and other minority groups. Thus, the U.S.-born more often develop what has been called an "adversarial stance" toward the dominant white society (Portes & Zhou 1993; Waters 1999). Their racial identity, shaped by this perspective, becomes a way of rejecting the black or white identity imposed by others. For some, this entails adoption of a panethnic identity (e.g., Hispanic, Latino), while for others it entails rejection of the notion of race.

PREDICTORS OF RACIAL IDENTITY

The foregoing discussion suggests that place of residence and birthplace will be

key predictors of racial identity. Puerto Ricans living in the U.S. should be less

likely than their counterparts in Puerto Rico to identify their race as Puerto Rican. Moreover, U.S.-born mainland Puerto Ricans should be the most likely to move

away from a Puerto Rican racial identity toward a panethnic or American identity. In addition, parental ethnicity may influence an individual's racial identity. Identification with Puerto Rico should be stronger among those with two Puerto Rican parents than among those with a mixed ethnic background.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

236 / Social Forces 81:1, September 2002

Another critical factor in racial self-identification is the color of one's skin.

Relatively dark-skinned Puerto Ricans should be more likely than light-skinned Puerto Ricans to reject standard racial categories because acceptance would imply adopting a black identity. In addition, because those with dark skin assimilate less

readily than those with light skin, the former group should be less likely to identify as American than the latter. Their rejection of a black identity and their experience with discrimination may instead encourage dark-skinned individuals to retain a national identity or to adopt a panethnic identity (Portes & Rumbaut 2001).

Racial identities also may be influenced by individuals' skills and economic resources. In particular, those with relatively high educational attainment and family income are unlikely to base their racial identities on renunciation of the dominant

society and its conceptions of race. Accordingly, they should be less inclined to

identify as Hispanic or Latino and more likely to identify as white than others. Furthermore, language skills may be related to the racial identities of mainland Puerto Ricans. Portes and Rumbaut (2001) indicate that Spanish-dominant individuals are more likely to retain a national identity than are English-dominant or bilingual individuals. English-dominant persons generally interact more

extensively outside their ethnic communities than those who use Spanish most of the time. Because they are more assimilated, the English-dominant are more likely to take on an Americanized identity. Coethnic ties should also influence racial

identity: Mainland Puerto Ricans who associate extensively with Puerto Ricans or other Hispanics should be more likely to adopt a national or panethnic identity and less likely to adopt an American identity than those with fewer associations with Hispanics.

RESEARCH ISSUES

Recent studies have contributed greatly to our knowledge of how non-European immigrant groups adapt to the U.S. racial classification system (Portes & Rumbaut 2001; Waters 1999). However, none has compared the racial identities of U.S.

migrants to those of nonmigrants in their origin country. Such a comparison is

pivotal for understanding how self-definitions change with exposure to race relations in the U.S. To address this research gap, we investigate the racial identities of Puerto Rican women in the U.S. and Puerto Rico.

As noted earlier, we focus on several key issues. We first examine the way race is viewed by the women themselves. This topic is addressed using data from an

open-ended survey item on racial identity. We provide descriptive information on

responses to the open-ended item and compare those responses to women's classifications on a standard closed-ended race question. In addition, because of the centrality of physical appearance to racial classifications, we investigate the

relationship between skin tone and racial self-identification.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 237

Throughout the descriptive analysis, we emphasize comparisons between Puerto Rican women in the U.S. and Puerto Rico. Such comparisons are central to

understanding how the social context of race relations affects women's racial self- identification. Differences between the mainland and the island are further addressed in a multivariate analysis that also evaluates the role of personal attributes and achievements in racial self-identification. Among the characteristics considered are birthplace, parental ethnicity, skin tone, and socioeconomic status.

A final set of issues pertains to the implications of the assimilation process for racial self-identification among mainland Puerto Ricans. Previous research

suggests that women's identities will vary with generational status, language use, and the extent to which they are embedded in a coethnic community. The role of these factors in the racial identities of Puerto Rican women in the U.S. is assessed.

Data, Measures, and Methods

DATA

Our analysis is based on data from the Puerto Rican Maternal and Infant Health

Study (PRMIHS), a study of maternal and infant health outcomes among Puerto Ricans in the U.S. and Puerto Rico (Landale, Oropesa & Davila 2000).4 In-person interviews were conducted with 2,763 mothers of infants sampled from the 1994 and 1995 birth and infant death records of six U.S. vital statistics reporting areas (Connecticut, Florida, Massachusetts, New Jersey, New York City, and Pennsylvania5) and the Commonwealth of Puerto Rico. Roughly two-thirds of the interviews were with mothers of infants sampled from the computerized birth certificate files maintained by the states indicated and Puerto Rico. The present analysis is restricted to the birth sample.6

Infants in the U.S. were eligible for inclusion in the birth sample if the Hispanic ethnicity of the mother was designated as Puerto Rican on the birth certificate. Information on ethnicity is not included on birth certificates in Puerto Rico because an extremely high percentage of island residents are of Puerto Rican descent. To avoid inclusion of non-Puerto Rican infants in the study, a question on whether the focal infant was of Puerto Rican descent was included to screen for eligibility. Mothers who answered that their infant was not of Puerto Rican descent were excluded from the study. This screening question was used in both Puerto Rico and in the U.S. states.

Mothers of the sampled infants were located from the address information provided on the vital records and were asked to participate in a computer-assisted personal interview (CAPI). All PRMIHS interviewers were bilingual, and the questionnaire was available in both Spanish and English. The response rate for the birth sample was 79.8%. Sample selection bias due to nonresponse was minimal: Nonrespondents did not differ from respondents on a variety of socioeconomic

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

238 / Social Forces 81:1, September 2002

characteristics (Oropesa & Landale 2002). The weighted birth sample therefore can be considered representative of 1994-95 births to Puerto Rican women residing in the study areas. By extension, the birth sample represents Puerto Rican mothers of infants born in the specified areas and period of time.7

MEASURES

Race

Two measures of race are examined. As noted above, the PRMIHS sample was drawn from a complete listing of birth certificates in the study areas. Information from the birth certificate, including the mother's reported race, is included in the PRMIHS data set. The measure of race from the birth certificate is a closed-ended item that reflects the response of an informant, usually the mother (U.S. Department of Health and Human Services 1999). Possible response categories included white, black, five different Asian groups, American Indian, and other. Because all but one of the Puerto Rican women in our sample selected white, black, or other, we consider only those categories in our analysis.8

The PRMIHS survey included an open-ended question on race: "What race do

you consider yourself?" Interviewers were instructed not to provide any interpretation or help on this question. The responses given by the women were coded into the following categories: white, black, trigueia, Puerto Rican, Hispanic, Latina, Spanish, Hispanic American, American, and "other." Information on the full set of categories is presented in the descriptive analysis. A collapsed set of

categories is used in the multivariate analysis: white, black/triguefia, Hispanic/ Latina/Spanish, American/Hispanic American, and other.9

Place of Residence and Birthplace

The respondent's place of residence at the time of the survey is of central interest in our study. Women living in the U.S. (coded as 1) are contrasted with women

living in Puerto Rico (coded as 0). In addition, women who were born in the U.S.

(coded as 1) are compared with women who were born in Puerto Rico (coded as 0). In preliminary models, we tested the interaction between place of residence and

birthplace. Because the interaction was nonsignificant in all models, it was not included in the final analysis.

Parental Ethnicity

In separate questions, respondents were asked to indicate whether their father and their mother were of Puerto Rican ancestry. Women with two Puerto Rican parents (coded 1) are contrasted with women with only one Puerto Rican parent (coded 0).

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 239

Skin Tone

Consistent with a number of recent investigations of the influence of skin color on the socioeconomic outcomes of Hispanics (e.g., Murguia & Telles 1996; Telles & Murguia 1990), our measure of skin tone is based on the interviewer's rating of whether the respondent's skin was very light, light brown, medium brown, dark brown, or very dark brown. In the multivariate analysis, the skin tone variable is coded from 1 (very light) to 5 (very dark brown).

Socioeconomic and Demographic Characteristics

Socioeconomic and demographic variables include the mother's age, union status, education, family income, and employment status. With the exception of family income, these variables reflect the mother's status at the time of the interview, when her racial self-identity was measured. Maternal age is measured in completed years. Union status includes three categories: single, cohabiting, and married. Educational attainment is measured as the highest grade of school completed by the respondent. To measure family income, we asked women to indicate which of 13 categories best represented their total household income around the time the focal baby was born. The categories were recoded to their midpoints and income is treated as a continuous variable. Finally, a dichotomous measure indicates whether the

respondent was employed (1 = yes; 0 = no).10

Language Use Index

The language use index is an additive index constructed from three questions regarding the respondent's use of English versus Spanish at home, with friends, and when watching TV. The responses to each question (English most of the time; Spanish and English about equally; Spanish most of the time) were coded from one to three, with three indicating Spanish dominance. The resulting index ranges from three to nine.

Latino Friends and Neighbors

We include separate measures of the ethnic composition of the respondents' friends and neighborhoods. Responses range from 1 to 5, with 1 indicating"all non-Latinos" and five indicating "all Latinos.""

SAMPLE DESIGN AND MISSING DATA

The PRMIHS was based on a complex sample design that involved stratification by state, month, and birth outcome. Consequently, our analysis was conducted with SUDAAN, which adjusts coefficients and standard errors to take the sampling

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

240 / Social Forces 81:1, September 2002

TABLE 1: Racial Classification of Puerto Rican Mothers in the United States and Puerto Rico

United Puerto States Rico

Race of mother from birth certificate White 88.3 93.9* Black 3.2 5.9* Races other than white or black 8.5 .2*

100.0 100.0

"What race do you consider yourself?" White 2.0 17.2* Black 1.0 5.5*

Triguefia .2 7.2* Puerto Rican 56.6 52.5

Hispanic 20.5 3.7* Latina 7.3 5.2

Spanish 3.6 .7* American 4.2 2.7

Hispanic American 1.1 .6 Other 3.5 4.7

100.0 100.0

Unweighted N 1,256 669 *

Percentage significantly different from that for Puerto Rican women in the U.S., p < .05

design into account. All models are based on weighted data, using the final birth

sample weights. The weights were adjusted to retain the original sample size. Cases with missing data are not excluded from the analysis to avoid erroneous

inferences that can stem from the rejection of cases in which data are not missing completely at random. Instead, we employed Bayesian procedures for the multiple imputation of missing data (Schafer 1997, 1998). Five imputations were made to

generate plausible values for missing data, and the five imputed data sets were

analyzed with standard complete-data methods. The results were combined to yield estimates, standard errors, and p-values that incorporate uncertainty about missing data.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 241

TABLE 2: Cross-classification of Race As Reported on Birth Certificate and in Open-ended Survey Question, Puerto Rican Mothers in the United States and Puerto Rico

United States Puerto Rico Race of Mother Race of Mother

from Birth Certificate from Birth Certificatea

White Black Other White Black

"What race do you consider yourself?"

White 2.2 0.0 .2 18.2 .9 Black .8 8.3 0.0 4.5 20.4

Triguefia .2 0.0 0.0 7.0 10.2 Puerto Rican 57.4 41.0 53.9 52.2 50.8

Hispanic 19.6 24.8 28.4 3.6 5.4 Latina 6.9 16.4 8.0 5.2 5.5

Spanish 4.1 0.0 .3 .7 1.0 American 4.2 0.0 5.8 2.8 0.0

Hispanic American 1.3 0.0 0.0 .6 0.0 Other 3.3 9.7 3.4 5.1 5.8

100.0 100.0 100.0 100.0 100.0

Unweighted N 1,098 53 105 621 47 a Too few women identified themselves as "other" in Puerto Rico to allow for analysis.

Findings

How Do PUERTO RICANS DESCRIBE THEIR RACIAL IDENTITIES?

Table 1 provides descriptive information on responses to the closed-ended race

question on the birth certificate and the open-ended race question on the PRMIHS

survey. Separate distributions are presented for women in the U.S. and Puerto Rico. The majority of Puerto Rican women in the U.S. and Puerto Rico are classified

as white on the birth certificate. About 88% of mainland Puerto Ricans and 94% of Puerto Rican islanders are identified as white. A notable difference between mainland and island respondents is that the mainlanders are more likely to be classified as other (8%) than the islanders (.2%).

In contrast to responses to the closed-ended question, only a minority of mainland and island Puerto Ricans report their race as white on the open-ended

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

242 / Social Forces 81:1, September 2002

TABLE 3: Cross-classification of Race and Skin Tone, Puerto Rican Mothers in the United States and Puerto Rico

United States Puerto Rico Interviewer-Recorded Skin Tone Interviewer-Recorded Skin Tone

Very Very Light Light Medium Dark Light Light Medium Dark

"What race do you consider yourself?"

White 2.8 .9 2.4 0.0 31.0 14.7 .4 11.3 Black 0.0 2.3 1.6 .6 .7 3.7 10.2 16.4

Triguefia .2 .4 0.0 0.0 2.0 13.8 6.0 6.4 Puerto Rican 57.3 56.8 56.8 47.5 52.1 48.9 62.3 47.6

Hispanic 19.8 19.2 25.3 21.8 3.2 4.5 2.8 4.3 Latina 6.3 7.3 6.9 18.8 5.1 5.7 4.6 5.4

Spanish 2.3 6.2 2.3 4.1 0.0 2.0 .4 0.0 American 5.9 3.2 1.6 2.1 .9 4.3 3.0 2.7

Hispanic American 1.3 .8 1.5 0.0 .8 0.0 1.5 0.0 Other 4.2 2.9 1.8 5.1 4.1 2.3 8.9 5.9

100.0 100.0 100.0 100.0 100.0 100.0 100.0 100.0

Unweighted N 581 403 205 67 217 223 152 77

question. However, there is a large difference between the percentage of respondents reporting themselves as white in the U.S. and Puerto Rico. Only 2% of mainland women report their race to be white, compared to 17% of island women. Similarly, mainland women are less likely than island women to report their race as black (1% vs. 6%). It is apparent that mainland Puerto Rican women are disinclined to define themselves in terms of the black-white dichotomy that predominates in the U.S. In addition, mainland women are unlikely to use the intermediate racial

category of triguefia. Less than 1% of mainland women classified themselves as

triguefia, compared to 7% of island women. Mainland Puerto Ricans are slightly more likely than island Puerto Ricans to

indicate that their race is Puerto Rican (57% vs. 52%), although the difference is not statistically significant and Puerto Rican is the modal category for both groups. The groups diverge more with respect to the percentage reporting their racial identity to be Hispanic. About 20% of mainland Puerto Rican women report their race to be Hispanic, compared to 4% of women in Puerto Rico. Other differences between the groups are relatively slight.

Overall, it is clear that Puerto Rican women in the U.S. do not identify with the

categories typically used to define race. Instead, they define themselves in terms of their national origins or panethnic categories such as Hispanic or Latina. Islanders also frequently adopt a national racial identity, but they are less prone to reject the usual U.S. conceptualization of race.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 243

How ARE RESPONSES TO THE Two RACE QUESTIONS RELATED?

Table 2 shows the cross-classification of the two race variables. Within each location, the racial identities of women defined as white on the birth certificate parallel those for the overall sample. This is to be expected because of the predominance of white

respondents. Mainland women defined as other on the birth certificate provided responses to the open-ended question that are roughly similar to those of whites, except that a higher percentage of women defined as other classify themselves as

Hispanic (28% vs. 20%). In both the U.S. and Puerto Rico, the pattern for women classified as black on the birth certificate diverges most from that for the overall

sample. In the U.S., only 8% of women defined as black on the birth certificate report

their race as black on the open-ended question. Instead, blacks are more likely to identify their race as Puerto Rican (41%), Hispanic (25%), or Latina (16%). Interestingly, mainland women who are classified as black on the birth certificate are considerably more likely to report themselves to be Latina on the open-ended question than mainland women who are classified as white on the birth certificate (16% vs. 7%). In addition, a nontrivial percentage (10%) of mainland women defined as black on the birth certificate provided answers to the open-ended question that could only be classified as other. Typically, these responses indicated a mixed racial/ethnic background. For example, several respondents wrote that they were half black and half Puerto Rican, and another respondent indicated that she was multiethnic.

In Puerto Rico, women defined as black on the birth certificate are more likely than their U.S. counterparts to self-identify as black (20% vs. 8%) and less likely to self-identify as Hispanic (5% vs. 25%) or Latina (6% vs. 16%). In addition, about 10% of islanders classified as black on the birth certificate self-identify as triguefia.

On the whole, there is a striking lack of correspondence between the way Puerto Rican women define their race and their classifications on the standard race question included on the birth certificate. This finding corresponds to the general conclusions of other researchers regarding the inadequacy of the standard race question for Hispanics (Denton & Massey 1989; Rivera-Batiz & Santiago 1996; Rodriguez 2000), but our study goes beyond prior research in providing detailed responses to an open-ended question on race.

DOES SKIN COLOR INFLUENCE RACIAL SELF-IDENTIFICATION?

A key issue in the study of racial self-identification is the extent to which individuals' self-identities are influenced by the color of their skin. In Table 3, we cross-classify the open-ended race question by respondent skin tone. In the U.S., women with very light skin, light skin, or medium skin have similar responses to the race question. A noteworthy exception to this generalization is that medium-skinned

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

244 / Social Forces 81:1, September 2002

TABLE 4: Odds Ratios from Multinomial Logit Models of Racial Identity, Puerto Rican Mothers in the United States and Puerto Rico

Black/ Hispanic/ White Triguefia Latina American Other

vs. vs. vs. vs. vs. Puerto Puerto Puerto Puerto Puerto Rican Rican Rican Rican Rican

Model 1 U.S. resident .12*** .11*** 3.20*** .74 .62 U.S. born .70 .60 .88 4.45*** 1.28

Model 2 U.S. resident .06*** .08*** 3.47*** .88 .47 U.S. born .56 .47 .79 3.79*** .83 Both parents Puerto Rican .08*** .05*** .32* .20** .06*** Skin tone .52** 1.50** 1.18 .98 1.21

Age 1.03 1.00 1.00 .99 1.02 Union status

Single - -

Cohabiting .65 .67 .97 1.39 .39 Married .51 .50 1.15 1.04 .53

Education 1.05 1.10 1.08* 1.20 1.20**

Family income 1.02* 1.01 1.00 .99 1.01

Employed 1.19 .77 1.08 1.15 .91

(N= 1,925)

* p < .05

** p < .01

*** p < .001

women are more likely to identify as Hispanic than the two groups of lighter- skinned women. In addition, there is a monotonic decline in the percentage identifying their race as American as we move from very light skin to medium skin tone.

Dark-skinned mainland women are less likely to self-identify as Puerto Rican than the other groups (48% compared to 57% for each other group). A relatively high percentage of dark-skinned mainland women designate their race as Latina, 19% compared to 6-7% for the lighter-skinned groups. The distinction between

Hispanic and Latina is noteworthy. Although the two terms are frequently used

interchangeably, they may have a somewhat different meaning for the women themselves.'4 It is also significant that less that 1% of the darkest-skinned mainland women indicate that their race is black.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 245

In Puerto Rico, a very different pattern is evident. First, there is a monotonic increase in the percentage identifying their race as black as skin tone becomes darker. While less than 1% of very light-skinned island women self-identify as black, fully 16% of the darkest-skinned women do so. There is also a decline in the percentage self-identifying as white as skin tone darkens, but the pattern is somewhat irregular. Additionally, light-skinned island women are the most likely to identify their race as triguefia.

In summary, the descriptive results in Tables 1 through 3 indicate that race is often viewed differently by Puerto Rican women in the U.S. and Puerto Rico.

Slightly more than half the women in each setting consider their race to be Puerto Rican. However, the distribution of the remaining women varies greatly between the two locations. In the U.S., a very small share of Puerto Rican women (3%) identify their race as white, black, or triguefia, while in Puerto Rico roughly 30% do so. Instead, mainland women are likely to adopt a panethnic identity: About 30% indicate that they are Hispanic, Latina, or Spanish, compared to 10% in Puerto Rico. Among mainland women, the likelihood of self-identification as white, black, or triguefia varies little by skin tone; however, the tendency to adopt a panethnic identity is greatest among those with the darkest skin. In contrast, in Puerto Rico, dark-skinned women are considerably less likely than light-skinned women to

identify as white and considerably more likely to identify as black.

WHAT PERSONAL CHARACTERISTICS STRUCTURE RACIAL SELF-IDENTIFICATION?

Pooled Mainland and Island Samples

We further examine mainland- island differences in racial self-identification in a multivariate framework in Table 4. Multinomial logistic regression models are used to estimate the odds of self-identification as (1) white, (2) black/triguefia, (3) Hispanic/Latina/Spanish, (4) American/Hispanic American, or (5) other versus Puerto Rican. We assess whether the differences by place of residence shown in the

descriptive tables can be explained in terms of respondent skin tone, parental ethnicity, demographic characteristics (age, union status), and socioeconomic status (education, family income, and employment status).

Model 1 in Table 4 includes only place of residence and birthplace as predictors. It is apparent that place of residence is generally more important for racial self- identification than birthplace. The odds of self-identification as white versus Puerto Rican are only 12% as high for mainland women as for island women. A similar odds ratio is observed for the comparison between black/triguefia and Puerto Rican. In contrast, the odds of self-identification as Hispanic/Latina versus Puerto Rican are three times as high for mainland women as for island women. The only contrast in which birthplace is a significant predictor is that between American/ Hispanic American and Puerto Rican. U.S.-born women are more than four times

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

246 / Social Forces 81:1, September 2002

TABLE 5: Odds Ratios from Multinomial Logit Models of Racial Identity, Puerto Rican Mothers in the United States

American/ Hispanic Hispanic Latina American Other

vs. vs. vs. Puerto Puerto Puerto Rican Rican Rican

Model 1 U.S. born .96 6.23** 2.49*

Model 2 U.S. born .95 4.27* 1.85 Both parents Puerto Rican .39* .19** .12*** Skin tone 1.19 .71 1.00 Age 1.01 1.05 1.05 Union status

Single Cohabiting .68 1.10 .25** Married 1.00 .87 .14**

Education 1.01 1.10 1.09 Familyincome 1.01 .99 1.02 Employed 1.03 1.15 .80 Language use index 1.10 .80* 1.06 Latino friends .84 1.12 .97 Latino neighbors .87 1.32 .91

(N = 1,256)

* p <.05 **p < .01

*** p < .001

as likely to self-identify as American or Hispanic American as women born in Puerto Rico.

Differences between mainland- and island-resident women are strengthened with the inclusion of the full set of predictors in model 2. Net of the included

predictors, the odds of self-identification as white or black/triguefia versus Puerto Rican are only 6-8% as high for U.S.-resident women as for women in Puerto Rico. In addition, having two Puerto Rican parents substantially reduces the likelihood of identification as white or black/triguefia (vs. Puerto Rican). Skin tone is also a significant predictor. The darker the skin tone, the less likely a women is to consider her race to be white and the more likely she is to consider her race to be black or triguefia. None of the other covariates is significantly related to the contrast between black/trigueia and Puerto Rican. However, family income is

positively related to taking on a white identity.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 247

As noted earlier, U.S.-resident women are more than three times as likely to

identify their race as Hispanic/Latina versus Puerto Rican as women living in Puerto Rico. This pattern remains in the full model. Again, we see that women with two Puerto Rican parents are less likely to depart from a national "racial" identity than others. They are about a third as likely as women with only one Puerto Rican parent to self-identify as Hispanic/Latina. Skin tone is a marginally significant (p = .08) positive predictor of self-identification as Hispanic/Latina. The only other significant predictor of taking on a panethnic identity is educational attainment. The higher the education, the more likely a woman is to designate her race as Hispanic or Latina.

The importance of birthplace for taking on an American or Hispanic American

identity (vs. a Puerto Rican identity) is indicated by the significant odds ratio of almost four for U.S. birth. As with the other racial contrasts, having both a Puerto Rican mother and a Puerto Rican father decreases the odds of self-identifying as American or Hispanic American versus Puerto Rican. Neither the demographic nor the socioeconomic characteristics included in the model are related to self- identification as American or Hispanic American

Overall, it is apparent that the differences between the racial identities of Puerto Rican women in the U.S. and Puerto Rico cannot be explained in terms of differences in parental ethnicity, skin tone, demographic characteristics, or socioeconomic status. Controls for the complete set of predictors strengthen rather than attenuate the relationship between place of residence and racial identity. Clearly, there are differences in the social construction of racial identity that go beyond the individual-level variables in our models.

MAINLAND SAMPLE ALONE

Although comparisons of the racial identities of migrant groups to those of

nonmigrants in their country of origin are rare, a number of studies have examined the implications of the assimilation process for racial/ethnic identity among immigrant groups in the U.S. (e.g., Portes & Rumbaut 2001; Waters 1999). These studies suggest that key factors for understanding racial/ethnic identity within

immigrant groups are nativity and indicators of assimilation, such as language use and the extent to which interactions are confined to a coethnic community. In Table 5, we present an analysis restricted to mainland Puerto Rican women. The

analysis is similar to that presented in Table 4 except that place of residence is omitted and three indicators of the degree of assimilation are added: the language use index and measures of the extent of interaction with Latino friends and

neighbors. It should be noted that the racial identity variable was recoded into four

categories for this analysis because there was an insufficient number of whites and

blacks/triguefias to analyze as separate categories. Thus, those groups are combined with the "other" category for the mainland analysis. Although results for the "other"

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

248 / Social Forces 81:1, September 2002

category are shown, they are not discussed because the category is too diverse to be

substantively meaningful. Model 1 in Table 5 shows that U.S. birth is unrelated to self-identification as

Hispanic or Latina (versus Puerto Rican), but the odds of taking on an American or Hispanic American identity are six times as high for Puerto Rican women born in the U.S. as for those born in Puerto Rico. This pattern persists in Model 2, although the odds ratio is reduced from about six to about four with controls for the included predictors.

In model 2, the only variable that is significantly related to taking on a Hispanic/ Latina identity versus a Puerto Rican identity is the ethnicity of the respondent's parents. The odds of identifying as Hispanic or Latina are only about 40% as high for women with two Puerto Rican parents as for other women. An even stronger relationship between parental ethnicity and identity is found for the contrast between American/Hispanic American and Puerto Rican. The odds of identifying as American or Hispanic American are less than one-fifth as high for respondents with two Puerto Rican parents as for respondents with a non-Puerto Rican mother or father.

One other variable is related to taking on an American or Hispanic American racial identity - the language use index. The greater the use of Spanish versus

English, the less likely a woman will self-identify as American or Hispanic American instead of Puerto Rican. Language use may both shape and reflect an individual's

identity. Puerto Rican women who use Spanish most of the time are probably more immersed in interactions with ethnically similar individuals than Puerto Rican women who primarily use English. Such interactions are likely to reinforce their Puerto Rican identity. The fact that the ethnic composition of the respondents' friends and neighbors is unrelated to self-identification as American/Hispanic American versus Puerto Rican may have to do with the wording of the survey questions. The questions asked respondents about the proportions of their friends and neighbors who were Latino and did not distinguish between Latinos of Puerto Rican descent and those with other national origins.

Conclusion

The concept of race has become increasingly difficult to delineate and measure in a straightforward fashion. While distinctions between racial groups were once

thought of as biologically determined, it is now widely recognized that racial

categories are socially constructed and differ across social settings (Etzioni 2001; Omi & Winant 1994). Moreover, the concept of race is often inseparable from the

concept of ethnicity. Although sociologists have generally used race to refer to distinctions based on physical appearance and ethnicity to refer to distinctions based on culture, language, or descent (Hirschman, Alba & Farley 2000), these concepts are increasingly blurred in the self-definitions of individuals. Hispanics provide

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 249

the classic example of this phenomenon. When asked to select their race from the standard response categories, Hispanic individuals frequently select "other." If given the option to elaborate, many Hispanics provide a national origin or a panethnic descriptor.

The present study analyzed racial self-identification among Puerto Rican women. Several features of the study are noteworthy. First, responses to an open- ended question on racial self-identification allowed us to examine how the women themselves view race. Second, the sample included respondents from both the U.S. and Puerto Rico, allowing us to investigate how the women's social context influences racial identity. Third, we collected data on skin color, one of the major observable characteristics used by others to define one's race. Finally, we were able to compare responses from the open-ended question to those from a closed-ended race item drawn from the birth certificate for a recent birth of the respondent. This

comparison provided useful information on the adequacy of closed-ended questions for capturing the meaning of race among Puerto Ricans.

Among mainland and island Puerto Ricans, the most common response to the open-ended race question is Puerto Rican. Slightly more than half of each group provided a national racial identity. At the same time, the responses of the women who did not choose Puerto Rican diverged dramatically across locales. Though the

majority of the women in Puerto Rico who did not self-identify as Puerto Rican identified as white, black, or triguefia, almost none of the mainlanders did so. Instead, mainland women were most likely to select a panethnic identity (Hispanic or Latina), followed by an American or Hispanic American identity. We cannot

explain the island-mainland differences in terms of the individual-level characteristics in our data. In fact, differences between mainland and island women were strengthened when we controlled for parental ethnicity, skin tone, demographic factors, and socioeconomic status.

Recent studies suggest that the racial and ethnic identities adopted by contemporary immigrants depend in part on their experiences with discrimination.

Groups that encounter significant discrimination are the most likely to reject the traditional concept of race and adopt a national or panethnic racial identity. Our data do not include direct measures of discrimination. Nonetheless, the discrimination argument suggests that the darkest-skinned Puerto Rican women

might be the least likely to self-identify as white, black, or triguefia. Though there are very slight differences in our cross-tabulations for mainland women that are consistent with this idea, what is most striking is that almost all mainland women

reject the traditional black-white conceptualization of race. This is true of both the

lightest-skinned women and darkest-skinned women. Indeed, there are so few mainland respondents in the white and black/triguefia categories that it was not

possible to examine those groups within a multivariate framework in the mainland-

only analysis.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

250 / Social Forces 81:1, September 2002

Several explanations can be offered for mainland Puerto Ricans' disinclination to use standard racial categories when describing their racial identities. First, the

widespread use of the terms Hispanic and Latino may play a role in Puerto Ricans'

self-conceptions. In particular, it is increasingly common for the U.S. population to be described in terms of a five-category scheme that mixes race and ethnicity: non-Hispanic white, non-Hispanic black, non-Hispanic Asian, non-Hispanic Indian, and Hispanic (Hirschman, Alba & Farley 2000). This practice may contribute to a perception that the panethnic label Hispanic represents a "racial"

group. Thus, the greater tendency of mainland Puerto Ricans, relative to island Puerto Ricans, to define their race as Hispanic or Latino may to some extent reflect

labeling by the mainstream society. A second explanation emphasizes a more active role for individuals in the construction of their racial identities. It is not necessarily antithetical to the former perspective in that it recognizes that racial categories are learned and identities are formed in interaction. However, it suggests that individuals actively create and manage their identities as they respond to the

exigencies of the social context. We think that the second explanation more accurately reflects the situation of

Puerto Ricans than the first does. In Puerto Rico, racial classifications are more

complex and more flexible than in the mainland U.S. Many racial categories are

recognized, and physical characteristics that suggest a mixed heritage do not

generally result in placement at the bottom of the racial hierarchy (Rodriguez 2000; Rodriguez & Cordero-Guzman 1992). Puerto Rican migrants to the U.S. face a

very different racial climate. From their perspective, they are neither black nor white, but they are living in a society in which race traditionally has been conceived as a black-white dichotomy. They must grapple with redefining their racial identity in a social context that is relatively intolerant of nonwhites. It is not surprising that mainland Puerto Ricans, even those with dark skin, are highly unlikely to identify as black. However, the fact that very few mainland women identify as white - and the fact that mainland women are strikingly less likely to do so than island women - suggest that mainland Puerto Ricans are opting out of the U.S. racial dichotomy because it does not fit with their conceptual framework and is less affirming than an emphasis on nationality or panethnicity.

These results have implications for recent discussions of the emergence of a new racial dualism. Some scholars argue that the growth of the Hispanic and Asian

populations may stimulate a shift in the U.S. racial hierarchy from a black-white

dichotomy to a black-nonblack dichotomy. Specifically, Gans (1999) argues that the current racial hierarchy could be replaced by"a dual or bimodal one consisting of'nonblack' and 'black' population categories, with a third, 'residual' category for the groups that do not, or do not yet, fit into the basic dualism" (371). He further

suggests that Caribbean blacks and dark-skinned Hispanics are likely to be classified with American blacks in the new racial system, especially given the disadvantaged socioeconomic circumstances shared by these groups. Gans's provocative discussion

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 251

acknowledges that the reconstruction of the racial hierarchy is a "work in progress" that may take a generation to resolve itself. During that time, the practice of

regarding nonblack Hispanics as a quasi racial-ethnic group may continue. Although discussions of the new racial dualism typically emphasize the racial

definitions imposed by the dominant population, our study of Puerto Ricans' racial identities sheds light on the perceptions of those who occupy an ambiguous racial status. Mainland Puerto Rican women's reluctance to identify themselves as white or black suggests that they are ill at ease with features of both the old dualism and the new dualism. This reluctance is evident regardless of skin tone and suggests that dark-skinned Puerto Ricans are likely to continue to resist being placed in the same category as American blacks. While both light-skinned and dark-skinned Puerto Ricans may be willing to think of themselves as nonblack, it is an open question whether those with dark skin will be regarded as such by other members of society. Our findings strongly indicate that Puerto Ricans continue to emphasize their national and ethnic origins. This suggests that multifaceted constructions of race that combine "racial" and "ethnic" distinctions are likely to remain a part of the identities of Puerto Ricans for some time to come.

In closing, we return to the practical implications of our findings. We underscore that there is almost no correspondence between Puerto Rican women's racial classifications on the birth certificates and their responses to our open-ended survey question on race. This is especially the case for mainland women. The lack of concurrence between the race items raises questions about the meaning of traditional closed-ended race classifications for this population. Hispanics have been voicing the lack of fit between the usual racial categories and their racial identities for many years by selecting the "other" category on standard race questions. Our study reinforces this message by demonstrating that Puerto Ricans think of race much more in terms of their national origins or ties with other Hispanic groups than they do in terms of the color of their skin. Accordingly, we recommend that closed-ended race questions be disregarded or used very cautiously for the Puerto Rican population.

Notes

1. A summary of the implications of the revised federal standards for the racial and ethnic classifications used in the 2000 U.S. Census can be found at http://www.census.gov/ population/www/socdemo/race/racefactcb.html.

2. Individuals who indicated that they were Hispanic also were asked to designate their specific Hispanic origin. The options provided were (1) Mexican, Mexican American, Chicano; (2) Puerto Rican; (3) Cuban; and (4) other, with a space to write in the name of the other origin group.

3. The terms trigueno and moreno are used to describe mixed-race individuals. In general, trigueio refers to individuals with bronze- or wheat-colored skin and moreno refers to individuals who are dark-skinned (Rodriguez 2000). However, Rogler (1944) notes that

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

252 / Social Forces 81:1, September 2002

the term trigueno is used to describe mixed-race individuals who have a wide range of skin tones.

4. The data were collected by the Institute for Survey Research at Temple University under a subcontract from the Population Research Institute, Pennsylvania State University.

5. The U.S. states included in the PRMIHS are those with the greatest number of births to Puerto Rican women each year. In 1994 and 1995, 72% of all births to mainland Puerto Rican women occurred in these six states. The state of New York is divided into two separate vital statistics reporting areas, New York City and the remainder of the state. New York City granted permission to conduct the survey, but permission could not be obtained from the state of New York. New York cases are therefore restricted to births and deaths occurring in New York City.

6. Twenty-one of the 1,946 cases in the birth sample were excluded from the present analysis because the mother was born outside the U.S. and Puerto Rico.

7. Because our study is based on a sample of mothers, our findings cannot be generalized to all Puerto Rican women. However, we do not have any reason to suspect that self- selection into motherhood is related to racial self-identification.

8. One respondent's race was classified as American Indian on the birth certificate. This case was recoded as other.

9. Although it would be desirable to retain the full set of uncollapsed categories in the multivariate analysis, it was not possible to do so because of the small number of respondents in some groups. For example, women who identified their race as triguefna were combined with those who identified their race as black because only 2-3% of cases were in each separate category. Among U.S. residents, less than 1% considered themselves to be trigueia and only 1% identified their race as black. Additionally, Hispanic Americans could not be analyzed as a separate category because they compose less than 1% of cases. Whether the Hispanic Americans should be combined with Hispanics or with Americans is an open question. All analyses were conducted both ways. The results are not sensitive to our decision to combine Hispanic Americans with Americans rather than with Hispanics.

10. Age, education, and income are treated as continuous rather than categorical variables. In preliminary analyses, alternative specifications of these variables were examined. The findings were very similar to those shown.

11. The language use index and the measures of Latino friends and neighbors are included only in the analysis restricted to mainland Puerto Ricans. Given that Spanish language use and association with Latinos are universal in Puerto Rico, these variables are not applicable to island women.

12. None of the variables had missing values for more than 5% of the cases. Fewer than 1 percent of cases were missing on the primary dependent variable, the open-ended survey question on race.

13. Dark brown and very dark brown are combined into the category "dark" in Table 2 because of the relatively small number of cases in each of the categories.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

Mainland and Island Puerto Ricans / 253

14. Some (e.g., Ellison 2001) have noted that the term Latino refers to persons whose ancestors came from the central and southern parts of the Western Hemisphere (i.e., Mexico, Puerto Rico, Cuba, the Dominican Republic, Central and South America), while the term Hispanic is more inclusive and also refers to persons from Spain. Murguia (1991) argues that the term Hispanic is more integrationist, while the term Latino implies a greater commitment to cultural pluralism.

References

Bean, Frank D., and Marta Tienda. 1987. The Hispanic Population of the United States. Russell Sage Foundation.

Brown, Terry D., Francis C. Dane, and Marcus D. Durham. 1998. "Perception of Race and Ethnicity." Journal of Social Behavior and Personality 13:295-306.

Denton, Nancy A., and Douglas S. Massey. 1989. "Racial Identity among Caribbean Hispanics: The Effect of Double Minority Status on Residential Segregation." American Sociological Review 54:790-808.

Duany, Jorge. 1985. "Ethnicity in the Spanish Caribbean: Notes on the Consolidation of Creole Identity in Cuba and Puerto Rico." Ethnic Groups 6:99-123.

Ellison, Keith D. 2001. "Hispanic, Latino, Person of Color: What's the Right Label?" http:// nextstepmag.com/race_colorlabel.htm.

Etzioni, Amitai. 2001. The Monochrome Society. Princeton University Press.

Fitzpatrick, Joseph P. 1971. Puerto Rican Americans: The Meaning of Migration to the Mainland. Prentice-Hall.

Gans, Herbert J. 1999. "The Possibility of a New Racial Hierarchy in the Twenty-First-Century United States." Pp. 371-90 in The Cultural Territories of Race: Black and White Boundaries, edited by Michele Lamont. University of Chicago Press.

Hirschman, Charles, Richard Alba, and Reynolds Farley. 2000. "The Meaning and Measurement of Race in the U.S. Census: Glimpses into the Future." Demography 37:381-94.

Hughes, Michael, and Bradley R. Hertel. 1990. "The Significance of Color Remains: A Study of Life Chances, Mate Selection, and Ethnic Consciousness among Black Americans." Social Forces 68:1105-20.

Hunter, Margaret L. 1998. "Colorstruck: Skin Color Stratification in the Lives of African American Women." Sociological Inquiry 68:527-35.

Keith, Verna M., and Cedric Herring. 1991. "Skin Tone and Stratification in the Black Community." American Journal of Sociology 97:760-78.

Landale, Nancy S., R.S. Oropesa, and Ana Luisa Davila. 2000. The Puerto Rican Maternal and Infant Health Study [Computer file]. University Park, Penn.: Penn State Population Research Institute [producer and distributor].

Mills, C. Wright, Clarence Senior, and Rose Kohn Goldsen. 1950. The Puerto Rican Journey: New York's Newest Migrants. Harper.

Mintz, Sidney W. 1971. "Groups, Group Boundaries, and the Perception of Race." Comparative Studies in Society and History 13:437-43.

. 1974. Caribbean Transformations. Aldine.

Muiguia, Edward. 1991. "On Latino/Hispanic Ethnic Identity." Latino Studies Journal 2:8-18.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions

254 / Social Forces 81:1, September 2002

Murguia, Edward, and Edward E. Telles. 1996. "Phenotype and Schooling among Mexican Americans." Sociology of Education 69:276-89.

Nelson, Candace, and Marta Tienda. 1985. "The Structuring of Hispanic Ethnicity: Historical and Contemporary Perspectives." Ethnic and Racial Studies 8:49-74.

Omi, Michael, and Howard Winant. 1994. Racial Formation in the United States: From the 1960s to the 1990s. Routledge & Kegan Paul.

Oropesa, R.S., and Nancy S. Landale. 2002. "Nonresponse in Samples from 'Rare' and 'Not-So- Rare' Populations: Findings on Mainland and Island Puerto Ricans from a Maternal and Infant Health Survey." Maternal and Child Health Journal 6:49-58.

Portes, Alejandro, and Ruben Rumbaut. 1990. Immigrant America: A Portrait. University of California Press.

.2001. Legacies: The Story of the Immigrant Second Generation. Russell Sage Foundation and University of California Press.

Portes, Alejandro, and Min Zhou. 1993. "The New Second Generation: Segmented Assimilation and Its Variants." Annals of the American Academy of Political and Social Sciences 530:74- 96.

Rivera-Batiz, Francisco L., and Carlos E. Santiago. 1996. Island Paradox: Puerto Rico in the 1990s. Russell Sage Foundation.

Rodriguez, Clara E. 2000. Changing Race: Latinos, the Census, and the History of Ethnicity in the United States. New York University Press.

Rodriguez, Clara E,. and Hector Cordero-Guzman. 1992. "Placing Race in Context." Ethnic and Racial Studies 15:523-41.

Rogler, Charles. 1944. "The Role of Semantics in the Study of Race Distance in Puerto Rico." Social Forces 22:448-53.

Safa, Helen I. 1998. "Introduction." Latin American Perspectives 25:3-20. (Special issue: Race and National Identity in the Americas.)

Schafer, Joseph L. 1997. Analysis of Incomplete Multivariate Data. Chapman & Hall.

.1998. "Multiple Imputation: A Primer." Unpublished manuscript, Pennsylvania State University.

Telles, Edward E., and Edward Murguia. 1990. "Phenotypic Discrimination and Income Differences among Mexican Americans." Social Science Quarterly 71:682-96.

Torres, Arlene. 1998. "La Gran Familia Puertorriquena 'Ej Prieta de Belda."' Pp. 285-306 in Blackness in Latin America and the Caribbean: Social Dynamics and Cultural Transformations, edited by Arlene Torres and Norman E. Whitten Jr. Indiana University Press.

U.S. Department of Health and Human Services. 1999. Vital Statistics of the United States, 1997. Hyattsville, Md.

Wade, Peter. 1997. Race and Ethnicity in Latin America. Pluto Press.

Waters, Mary C. 1999. Racial Identities: West Indian Immigrant Dreams and American Realities. Russell Sage Foundation and Harvard University Press.

Winn, Peter. 1992. Americas: The Changing Face of Latin America and the Caribbean. Pantheon Books.

This content downloaded from 144.13.128.131 on Thu, 09 Apr 2015 20:00:29 UTC All use subject to JSTOR Terms and Conditions