Youths of Mexican Descent of the Southwest: Exploring Differences in Ethnic Labels

Abstract

Knowledge of the factors that influence youths’ choice of racial or ethnic labels will help us understand intragroup diversity, suggest ways in which school social workers can support youth’s ethnic identity development, and learn if youth who choose different combinations of labels may be grouped together for research purposes. This study of 2,857 Mexican-origin youth in the U.S. Southwest found that linguistic acculturation, socioeconomic status, and educational aspirations are related to choice of ethnic labels. Implications for social work interventions in schools and for future research are offered.

The label one uses to self-identify is one aspect of ethnic identity. Ethnic identity is “one’s sense of belonging to a group and the feelings and attitudes that accompany this sense of group membership” (Phinney, Romero, Nava, & Huang, 2001, p. 136). Ethnic identity development is particularly salient for youths of color (Rumbaut, 1994), may begin as early as middle school for these youth (Tatum, 1997), changes over time, and is influenced by family and community contexts (Portes & Rumbaut, 2001). It is within the school context that youths often first encounter pressures to select a racial category (Chiong, 1998). This pressure transcends school forms, in that many children of color are forced to figure out their places within pre-established racial and ethnic groups. Because concentrations of Latinos now are found in many U.S. states (Guzmán, 2001) and Latinos in the United States are more likely than the general population to be under the age of 18 (Guzmán, 2001) school social workers across the nation are in unique positions to support the ethnic identity development of Latino youths.

The vast majority of studies that seek to examine whether there are relationships between the ethnicity of youths and various health and behavioral outcomes use youths’ self-selected ethnic or racial labels to categorize them. Researchers have faced the dilemma of deciding how to assign study participants who select multiple ethnic labels into only one category for purposes of analysis. Assigning youths to only one category, however, may obfuscate differences among youths who select various combinations of labels. If these differences are concealed, then practitioners in schools and other settings who develop interventions on the basis of research findings may offer interventions that do not meet the needs of many youth. In the case of Latino youths, understanding racial and ethnic self-identification presents unique issues in that Latinos of different nationality backgrounds, races, and socioeconomic and acculturation statuses tend to vary greatly in their self-identification.

This article attempts to address this issue by exploring whether gender, age, linguistic acculturation, family income, and educational expectations predict whether youths of Mexican heritage in the U.S. Southwest select only a label that describes their national origin or culture (that is Mexican/Mexican American/Chicano) or simultaneously choose this Mexican-origin label, umbrella labels (for example, Latino or Hispanic) and other labels (for example, other Hispanic, white/Anglo, American Indian). In addition, we consider three theoretical frameworks that may shed light on why selected factors (for example, linguistic acculturation) may affect youths’ choice of ethnic labels. Findings suggest factors that need to be taken into account in developing effective school-based interventions for Mexican-origin youths who claim a range of labels and can guide researchers in determining whether Mexican-origin youths in the Southwest who select different combination of labels can be grouped together for research purposes.

THEORETICAL FRAMEWORKS

Because of the exploratory nature of this study, we examine three different theoretical perspectives that have the potential to explain why selected factors may influence youths’ selection of ethnic labels. Although each framework offers a unique perspective, all recognize that youths may choose different labels over time and in different social contexts (LaFromboise, Coleman, & Gerton, 1993).

National versus Racial and Ethnic Identity Labels

U.S. Latino adults use national origin most often as a marker for self-identification (Latino National Political Survey, cited in Schmidt, Barvosa-Carter, & Torres, 2000). Second-generation Latino youths, however, tend to see their ethnicity in racial terms (Portes & Rumbaut, 2001), perhaps because the selection of a label may be influenced by how others perceive youths (Brunsma & Rockquemore, 2001) and because the dominant U.S. culture uses race as a way of understanding ethnic differences (Holleran, 2003). Applied to this study, youths who are less acculturated to the U.S. ideology would select only a Mexican/Mexican American/Chicano label. As youths become more acculturated to the U.S. ideology regarding race and ethnicity, they would be more likely to select additional labels such as white, black and indigenous, which may be seen to reflect the “racial” backgrounds of people of Mexican descent.

La Raza vs. Racial/Ethnic Identity Labels

The stated ideology of Mexico and many other Latin American countries, La Raza Cósmica, denotes acceptance of both Spanish and indigenous (that is, “Indian”) and, in some cases African origins and is indicative of a sense of pride (Marsiglia, 1991). It remains unclear, however, how pervasive the La Raza ideology in Mexico. Vila (1999) wrote of the “official rhetoric” of Mexico that embraces its indigenous roots while an anti-Indigenous discourse persists. Applied to this study, both less- and more-acculturated youths would claim multiple labels: Less-acculturated Mexican-origin youths may select white, black, or indigenous racial labels (in addition to Mexican) that reflect the La Raza ideology; youths who are more acculturated to the dominant U.S. view of race would select racial labels (for example, white) in addition to Mexican or Mexican American that reflect the U.S. racial ideology.

Conflict Theory

Mexican-origin youths live within a context of nativist and racist sentiment that has led to increased legislative and other activities aimed at reducing immigration from Mexico (Massey, 1995). In such an environment, people who are U.S.-born —particularly white people—often create and accept negative stereotypes about Mexican-origin communities and talk derogatorily about “Mexicans” (Holleran, 2003). Massey and Denton (1992) asserted that experiencing discrimination reinforces a racially distinct identity among Mexican immigrants when they realize that “they will not be accepted as whites” (p. 240). Negative stereotypes may be more likely to be assigned to people who are perceived to be low-incomes, speak Spanish, or speak English with a Mexican accent. Because members of oppressed groups may proudly claim labels that dominant groups use to subordinate them as an act of resistance (Massey & Denton, 1992), lower income and Spanish-speaking youths may select only the Mexican/Mexican American/Chicano label for themselves.

LITERATURE REVIEW

Mexican-Origin Youth and Ethnic Labels

Youths of Mexican descent in the United States may choose—or be assigned to—many possible ethnic labels. These labels fall into several (sometimes overlapping) categories: panethnic (that is, Hispanic, Latino), ethnic (that is, Mexican, Mexican American), racial (for example, white, indigenous), and national (that is, American, Mexican). A Los Angeles area study asked high school students who were of primarily Mexican descent to circle terms that reflected their identities (Reese, Gallimore & Zarate, 2003). Many students reported not seeing the difference between labels such as Latino, Hispanic, or Mexican. Although 78.5% of the sample was born in the United States, only one-third selected the American label and no students chose the white label. Student interviews indicated that they did not claim an American identity because of their language, because it was not connected to their parents, and because the American label was associated with being white. Research indicates that ethnic labels alone are not very good predictors of health behaviors for Mexican American students in the Southwest, but ethnic labels combined with ethnic identity measures have been identified as much more robust predictors of health outcomes (Marsiglia, Kulis & Hecht, 2001).

Age, Gender, and Ethnic Labels

Youths’ selection of ethnic labels may change as they age. A longitudinal study of 5,262 second-generation youths, more than half of whose parents were from Cuba, Mexico, or the Philippines, found that self-selected ethnic labels changed as youths aged from middle to high school (Portes & Rumbaut, 2001). Social groupings and larger social contexts become more important for older adolescents (Macias, 2004). Many studies fail to find gender differences in ethnic identity (Phinney et al., 2001), though one longitudinal study found that girls in junior high were less likely than were boys to select only an American identity label; this gender relationship disappeared when the survey was readministered in high school as boys moved toward panethnic and national identifications (Portes & Rumbaut, 2001). Portes and Rumbaut (2001) attribute changes in ethnic label selection to the malleability of ethnicity, noting that as youths mature labels are less influenced by individual and family characteristics. Consistent with the national origin versus racial-ethnic identity framework, some researchers have suggested that the gender difference may be due to the tendency of parents to be more protective of girls, keeping them closer to home, which could lead to them identify more closely with labels reflective of Mexican descent (Suárez-Orozco, 2001).

Linguistic Acculturation and Ethnic Labels

Language use accounts for the greatest portion of variance on acculturation instruments (Epstein, Botvin, Dusenbury, & Diaz, 1996) and thus is a commonly used proxy for acculturation. Speaking English at home has been found to be associated with selecting an American label (Portes & Rumbaut, 2001). For youths of Mexican descent the relationship between English language usage and racial label preference appears to be influenced by nation of birth: As English language usage increases, U.S.-born Mexican American youths are less likely to select a white label; however, Mexican-born youths are more likely to do so (Massey & Denton, 1992). This latter finding lends support to the national origin versus racial-ethnic identity framework.

Social Class and Ethnic Labels

The limited research about how social class influences ethnic labels is inconsistent. Some research suggests that economic privilege may lead to stronger ethnic retention or identification with their parents’ nationalities, with more affluent families having the economic resources to promote ethnic retention (Phinney et al., 2001; Rumbaut, 1994). One study found that lower income junior high youths were more likely than wealthier youths to identify as panethnic or mixed label rather than only American (Portes & Rumbaut, 2001). However, this relationship changed as adolescents matured and more affluent youths also shifted to panethnic self labels. Consistent with the national origin versus racial-ethnic identity framework, Portes and Rumbaut (2001) suggested that this stems from a growing awareness of externally defined ethnic categories among older adolescents, though conflict theory also may explain youths’ reluctance to select only an American label. Some research has found no relationship between socioeconomic status and ethnic labels or ethnic identity (Pellebon, 2000; Phinney & Devich-Navarro, 1997).

Educational Aspirations and Ethnic Labels

The vast majority of first- and second-generation immigrant middle school youths aspire to graduate from college, though their expectations of doing so are lower than their aspirations (Portes & Rumbaut, 2001; St. Hilaire, 2002). One study found that youths who selected panethnic identities in junior high had lower educational aspirations than did youths who selected other ethnic self identities (Portes & Rumbaut, 2001); another study found no relationship between ethnic labels and educational aspirations (St. Hilaire, 2002). The relationship, if any exists, between educational aspirations and ethnic labels may be affected by socioeconomic status and language skills. Studies of Mexican-descent youths have found that youths from families with higher socioeconomic status have higher educational aspirations than do their peers from families with lower socioeconomic status (Portes & Rumbaut, 2001; St. Hilaire, 2002) and that youths who are fluent in both Spanish and English have higher expectations than those who are not fluent in both languages (St. Hilaire, 2002).

In summary, previous research indicates that Mexican-origin youths may not see themselves as white or American, labels are influenced by language use at home, the relationship between socioeconomic status and choice of labels is unclear, and Mexican-descent youths who are fluent in both Spanish and English may have higher education expectations than such youths who are not fluent in both languages. It appears that existing research offers support for the national origin versus racial-ethnic identity framework, for conflict theory, or for both. Because research has led to inconsistent findings in many of these areas and some research lacks an explicit theoretical framework, many questions remain. This study addresses three primary questions. Are there differences in linguistic acculturation and income between Mexican-origin youths who select one versus multiple labels? Do linguistic acculturation, family income, educational expectations, age, and gender predict the odds of selecting one versus multiple labels, including specific labels (for example, other Hispanic, white)? Do national versus racial/ethnic identity, La Raza, or conflict theory frameworks aid us in understanding any of these relationships, if any exist?

METHOD

Context of Study

The sample for this study came from 35 middle schools located in the southern and central neighborhoods of Phoenix, Arizona. Latino enrollment in these schools ranged from 14 percent to 99 percent. Phoenix has a predominately non-Hispanic white population, and over one-third of the residents are Latino (Guzmán, 2001). The vast majority of Latinaos in Phoenix are of Mexican descent, ranging from recent immigrants to those whose ancestors have lived in the area for generations. Approximately one-fourth of residents in the county reported speaking a language other than English in the home (U.S. Census Bureau, 2006). The 2000 U.S. census marked the first census in which individuals were able to select more than one race, and 2.9 percent of residents in Arizona did so. Of this group, 47.9 percent also identified as Hispanic (Jones & Smith, 2001).

Anti-immigrant sentiment was apparent at the time of this study. During 1999 the Associated Press (1999) voted undocumented or “illegal” immigration as one of the top 10 news stories in the state. This sentiment led to voter approval in 2000 of a proposition that aimed to end bilingual education in public schools.

Study Participants and Questionnaire

Questionnaires were administered in fall 1999 to 4,630 seventh graders. The survey instrument offered English and Spanish versions back-to-back, and students could opt to complete the survey in either language. Youths were asked to answer “Yes” or “No” to items asking whether they were of Mexican origin (that is, Mexican/Mexican American/Chicano, hereafter referred to as MMAC), Other Hispanic, American Indian, black or African American, Asian or Pacific Islander, white, or other ethnicities or races. (These labels were used because they are commonly used in the U.S. Southwest.) We omitted the modeling of Asian labels as the number of students who chose the label, 37 or 1.3 percent, was too small to provide a reliable estimate. This analysis comprises all youths who self-identified as MMAC (N = 2,857), whether they selected one or more ethnic labels. The majority of youths (82.7 percent) selected only the MMAC label; the other 17.3 percent, in addition to selecting the MMAC label, selected other labels. For example, 7.4 percent (n = 212) of the 2,857 youths in the study also selected the white label (see Table 1). Two items about youths’ language preferences with family and with friends were used to assess linguistic acculturation. The mean for language preference with family fell close to “both English and Spanish equally” and that of language preference with friends was between “mostly English” and “both English and Spanish equally” (see Table 1).

Table 1.

Descriptive Statistics for Dependent and Independent Variables (N = 2,857)

Variable	n (%)	M (SD)
One vs. Multiple Labels
Mexican/Mexican American/Chicana(o) only	2,362 (82.7)
Mexican/Mexican American/Chicana(o) and Other Label(s)	495 (17.3)
Number of Labels Selected
1	2362 (82.7)
2	366 (12.8)
3	91 (3.2)
4 or more	38 (.14)
Additional Labels Selected
White	212 (7.4)
Other Hispanic	175 (6.1)
American Indian	181 (6.3)
Black or African American	75 (2.6)
Asian or Pacific Islander	37 (1.3)
Gender
Female	1,392 (48.7)
Male	1,465 (51.3)
Age		13.11 (.73)
11 years	109 (3.8)
12 years	181 (6.3)
13 years	1,865 (65.2)
14 years	614 (21.4)
15 years	81 (2.8)
16 years	5 (.17)
17 years	2 (.07)
Language Preference with Family		3.10 (1.26)
1 = English only	425 (14.9)
2 = Mostly English	326 (11.4)
3 = Both English and Spanish equally	1,202 (42.1)
4 = Mostly Spanish	357 (12.5)
5 = Spanish only	547 (19.1)
Language Preference with Friends		2.48 (1.16)
1 = English only	747 (26.1)
2 = Mostly English	597 (20.9)
3 = Both English and Spanish equally	1,119 (39.2)
4 = Mostly Spanish	178 (6.2)
5 = Spanish only	216 (7.6)
Get Free or Reduced Cost Lunch?
0 = Neither free nor reduced cost lunch	260 (9.1)
1 = Free or reduced cost lunch	2,597 (90.9)
How Far Will Get in School		3.98 (1.27)
1 = Finish 8th grade	89 (3.1)
2 = Finish high school	569 (19.9)
3 = Finish trade/vocational school program	77 (2.7)
4 = Finish 2 years of college	694 (24.3)
5 = Finish 4 years of college	1428 (50.0)

Family income was measured with an item asking whether youths received a free or reduced-cost lunch. Students usually do not accurately know their parents’ income, so free or reduced lunch status is an effective proxy for parental income (Bankston and Caldas, 1996; Gerard and Buehler, 1999). The vast majority of youths lived in low-income families (see Table 1) (over 90 percent received free or reduced-cost lunches). Educational expectations were measured by asking youths how far they would go in school. Almost half of the youth in this sample indicated that they expect to complete college (see Table 1). In the analyses, we combined the vocational/technical and two years of college categories into a single group because these are likely to be similar expectations of schooling.

Data Analysis

To compare and contrast mono-label and multi-label MMAC youths, chi-square tests of independence were used to compare responses on linguistic acculturation and family income variables. Multilevel logistic regression was used to explore whether linguistic acculturation, family income, educational expectations, age, and gender predicted the odds of selecting one or multiple ethnic labels. We do not include any school-level variables in the models because our hypotheses focus on the individual-level determinants of ethnic label choice. The multilevel procedures, however, adjust for the clustering in the data due to the sampling of students through schools (Raudenbush & Bryk, 2002). We used SAS 9.1 to estimate our models with random intercepts. Thus a significant school-level variance in this intercept would indicate that there is additional unobserved variation between schools in the tendency of students to choose ethnic labels.

RESULTS

Single Ethnic Label versus Multiple Ethnic Labels Self-Identification

Linguistic Acculturation

Youths who selected multiple labels live in families that are significantly more likely to be English dominant (see Table 2, panel A). More than one-half (n = 249, 51 percent) of youths who selected multiple ethnic labels reported that they prefer to speak “mostly English” or “English Only” with their families, compared to about one-fifth (n = 502, 21 percent) of youths who selected only one label. In addition, the percentage of single-label youths in Spanish-dominant families (n = 810, 34 percent) is almost double that of multiple-label youths (n = 94, 19 percent).

Table 2.

Crosstabulations of Self-Identification as Single vs. Multiple Labels with Language Preferences and Free/Reduced-Cost School Lunch

Panel A	Youth’s Language Preference with Family
Ethnic Self-Identification	English only n (%)	Mostly English n (%)	English and Spanish Equally n (%)	Mostly Spanish n (%)	Spanish Only n (%)	Total n (%)
MMAC label only	268 (11)	234 (10)	1050 (44)	319 (14)	491 (21)	2362 (100)
MMAC plus other label(s)	157 (32)	92 (19)	152 (31)	38 (8)	56 (11)	495 (100)
Total	425	326	1202	357	547	2857

Panel B	Youth’s Language Preference with Friends
Ethnic Self-Identification	English only n (%)	Mostly English n (%)	English and Spanish Equally n (%)	Mostly Spanish n (%)	Spanish Only n (%)	Total n (%)
MMAC label only	548 (23)	486 (21)	985 (42)	154 (7)	189 (8)	2362 (100)
MMAC plus other label(s)	199 (40)	111 (22)	134 (27)	24 (5)	27 (5)	495 (100)
Total	425	326	1202	357	547	2857

Panel C	Does Youth Get a Free or Reduced Cost Lunch?
Ethnic Self-Identification	Yes n (%)		No n (%)		Total n (%)
MMAC label only	2202 (93)		160 (7)		2362(100)
MMAC plus other label(s)	395 (80)		100 (20)		495 (100)
Total	2597		260		2857

Χ²=189.870, df=4, p<.001

χ²=73.624, df=4, p<.001

χ²=89.203, df=1, p<.001

Note: MMAC = “Mexican/Mexican American/Chicana/o”

Chi-square tests of independence also revealed significant differences in language preferred with friends between single- and multiple-label youths (see Table 2, panel B). As with family language use, multiple ethnic-label youths have stronger English language preference with friends than do single ethnic-label youths. With both groups of youths, however, there is greater use of English among friends than family.

Family Income

Statistically significant differences also were found between single- and multiple-label youths in whether they received free or reduced-cost lunches (see Table 2, panel C). The rate of free or reduced-cost lunches was about 13 percent higher for single-label youths.

Comparing Single-Label and Specific Multiple-Label Self-Identifications

The results from multilevel logistic regression models are presented in Table 3. The coefficients are odds ratios, which are the exponentiated logistic regression coefficients. An odds ratio greater than one is a positive effect, and an odds ratio less than one is a negative effect. An odds ratio equal to one or not significantly different from one is a null effect.

Table 3.

Multilevel Logistic Regression Models of Ethnic Labels

	Model 1: Selected Multiple Labels vs. Single Label	Model 2: Selected Other Hispanic vs. Did Not	Model 3: Selected White vs. Did Not	Model 4: Selected Black vs. Did Not	Model 5: Selected American Indian vs. Did Not	Model 6: Selected Asian vs. Did Not
Spanish Language Use
With Family	0.63*** (−8.15)	0.89 (−1.43)	0.62*** (−5.70)	0.48*** (−5.05)	0.50*** (−7.85)	0.70* (−2.20)
With Friends	1.05 (0.76)	1.08 (0.87)	1.06 (0.68)	1.15 (0.93)	1.05 (0.53)	1.08 (0.41)
Free or Reduced Lunch	0.44*** (−5.24)	0.72 (−1.36)	0.48*** (−3.64)	0.44** (−2.67)	0.54** (−2.96)	0.58 (−1.43)
Educational Expectations	0.97 (−0.79)	0.93 (−1.23)	1.10 (1.47)	1.04 (0.37)	0.82** (−3.28)	1.02 (0.17)
Age	0.95 (−0.69)	0.89 (−1.06)	0.86 (−1.36)	0.91 (−0.51)	0.83 (−1.67)	1.15 (0.65)
Female	1.20 (1.77)	1.28 (1.60)	1.33 (1.88)	1.17 (0.65)	1.25 (1.46)	0.67 (−1.29)
Intercept	3.29 (1.14)	0.56 (−0.38)	2.08 (0.49)	0.75 (−0.12)	14.33 (1.79)	0.01 (−1.59)
School-level variance	0.19**	0.29*	0.24*	0.24	0.17*	0.85*
N	2857	2857	2857	2857	2857	2857

^*p<.05,

^**p<.01,

^***p<.001, two-tailed tests

Note: Coefficients are odds ratios, t-statistics are in parentheses

We first examined predictors of selecting multiple labels, followed by predictors of selecting specific racial labels. In model 1, significant predictors of selecting multiple ethnic labels rather than MMAC only were Spanish language use preferred with family and receiving a free or reduced-cost lunch. Specifically, every unit increase in preference for speaking Spanish with family decreased the odds of selecting multiple labels by 38 percent (1 − .62 = .38). For students who received reduced-cost or free lunches, the odds of selecting multiple ethnic labels were 55% lower than for students not receiving free or reduced cost lunches (1 − .45 = .55). Students with educational expectations greater than less than a high school degree (the reference group) had significantly lower odds of choosing multiple labels. In model 2, only the educational expectations significantly predict whether or not MMAC youths chose the other Hispanic label. In models 3 and 4, similar patterns were found to predict whether MMAC youths select additional white or black labels. Spanish language use with family and free or reduced-cost lunch status significantly decreased the odds of choosing white and black labels, although these effects were larger in magnitude (coefficients are farther away from 1.0) for the selection of black labels (model 4). A difference between the two models, however, is the role of educational expectations. In model 3, students who expect a high school degree or a vocational or two-year college experience had significantly lower odds of choosing a white label. In model 4, however, there was no association between educational expectations and the odds of selecting a black label. Model 5 examines the selection of the American Indian label. Again, Spanish language use with family and free or reduced-cost lunches decreased the odds of the American Indian label. Greater educational expectations also decreased the odds of selecting American Indian ethnicity.

We also note that the school-level variance was significant in all but model 4 (selecting Black label versus not). This suggests that for these models, there is significant school-level variation in the odds of ethnic label choice. Although our models have examined the individual-level predictors of these choices, these significant random effects (the school-level variances) suggest that there may be factors beyond the individual, that is, at the level of the school, that also influence how students choose ethnic labels

DISCUSSION

Consistent with previous research (Portes & Rumbaut, 2001), this study found that among Mexican-descent youths in the U.S. Southwest, linguistic acculturation is related to selecting one versus more than one label. This finding appears to lend support to the national origin versus racial-ethnic identity perspective. That is, as youths move from Mexico to the United States, they identify primarily as being of Mexican origin and thus select only one label: MMAC. As part of the acculturation process they learn the U.S. ideology that emphasizes race rather than culture or nationality and come to self-identify as members of racial groups such as white, in addition to claiming a nationality or cultural group label.

In contrast to the findings of Phinney and Devich-Navarro (1997), we found that family income is associated with youths’ choice of ethnic labels. Youth from lower income families were more likely to select only the MMAC label. Conflict theory may partially explain this finding. If nativist phrases (for example, “Mexicans are threatening American culture,” “Mexicans are more likely to be involved in criminal activity”) are attached more strongly to lower income Mexican immigrants and Mexican Americans, then it is possible that youths in this study react to nativism by proudly embracing the label intended to subordinate them. Furthermore, if youths perceive that white people are leading the efforts to exclude them from full participation in U.S. society, they may be less likely to see themselves as white.

A higher percentage of youths spoke English with friends than with family. Furthermore, although more single-label than multiple-label youths reported using Spanish with friends, over 40 percent of single-label youths preferred to speak English with friends. These findings indicate that although there is much diversity in language use within the single-label group, immigrant youths are adopting English in some settings as their language of choice. Given that use of Spanish is associated with a stronger ethnic identity among Mexican-origin youths (Phinney et al., 2001) and that stronger ethnic identity is associated with positive health outcomes (for example, see Lorenzo-Hernandez & Ouellette, 1998; Martinez & Dukes, 1997; Phinney, Cantu, & Kurtz, 1997; Rotheram-Borus, Dopkins, Sabate, & Lightfoot, 1996), this decreasing use of Spanish may be cause for concern.

Students in this study with higher educational expectations were less likely to choose multiple labels, and the other Hispanic, white, and American Indian labels specifically. Our finding in this area may be related to that of Portes and Rumbaut (2001) that junior high youths who selected panethnic identities had lower educational aspirations than youths who selected other identities.

In summary, youths who spoke more Spanish with their friends, who were from lower income families, and who had higher educational expectations were more likely to identify only as MMAC. Their self-identification appears to be connected to their family’s nationality and culture. As youths acculturate linguistically and perhaps increase their contact with mainstream U.S. society, they start to incorporate other labels such as white, denoting a racial rather than cultural approach. In a more racialized U.S. context, the label “Mexican” transitions from being an ethnicity-national origin to becoming a racial label. In addition, higher income evidently made choices such as “white” more feasible. Selection of a single Mexican label on the part of youths from lower income families and those with higher educational expectations could also represent resistance to oppression. None of our findings support the perspective that Mexican-origin youths in this sample embrace the La Raza ideology.

Limitations

Several study limitations are important to consider before offering implications. The study used cross-sectional data, preventing analysis of how acculturation, educational expectations, family income, and preferred labels change over time. Data were collected in 1999 and it is possible that recent changes in the political climate have changed youths’ perceptions. An anti-immigrant sentiment has been present in the U.S. Southwest for many years, however, so the data likely remain useful. This study is envisioned as a needed baseline for ongoing research undertaken by the team. The study would have been enhanced had youths been able to identify separately as Mexican, Mexican American, or Chicano. Generational status was not included as a variable, although linguistic acculturation is a useful indicator of acculturation, regardless of generational status. Information was not available about the parents’ ethnicities. This information might shed light on why youths chose various combinations of ethnic labels, although data about parental race and nationality cannot be assumed to lead to accurate conclusions about the identity of multiethnic youths (Jackson, 2007). That is, those who are adopted or who acknowledge step-parents or guardians as parents may be placed into an inaccurate ethnic category by the researcher. Finally, information about physical appearance may have enriched the analyses, as it is possible that youths’ perception of their skin color and other aspects of ethnic appearance may affect their experiences and preferred ethnic labels.

Implications for Social Work in Schools

The finding that there are differences among youth who choose single multiple labels confirms that there is much diversity among Mexican-origin youths. Social workers thus must address a wide variety of perspectives, experiences, and needs within the commonly used “Mexican American” category.

At the very minimum, practitioners must allow youths to name themselves using labels they self-select. To empower youths, social workers can assist them in considering how language influences thinking and how it constricts or expands opportunities for some. In this way, youths might become more intentional in the language they use to describe themselves and others.

Although we did not include any school-level variables, our analyses show that unmeasured school-level factors may be associated with youths’ choice of labels. This suggests that schools have opportunities to shape youths’ identities. School social workers play pivotal roles in the process of cultural transmission and negotiation in school environments (Blair, 2002). A social worker could hold training sessions, provide written materials, and advise school staff and administrators on ways to support multicultural youths (Wardle, 1991). Social workers should also review school forms to make sure they do not inadvertently discriminate against those students who self-identify with more then one race, ethnicity, or culture. This form of advocacy, though indirect, could have long-standing effects on the identity of multiethnic or multicultural youths who are made to feel displaced by certain questions of ethnic-racial membership (Jackson, 2007).

The relationship between educational expectations and ethnic labels suggests that youths who claim both Mexican-origin and other identities would benefit from efforts to raise their expectations. Perhaps university-middle school partnerships that allow youths to visit college campuses, be mentored by college students and community professionals from within their ethnic communities, and offer other opportunities for raising expectations would be beneficial. Of course, policies that expand funding for higher education also are needed if expectations are to become reality for these disproportionately low-income youths.

Implications for Research

“all that apply” obviously is needed, as evidenced by intergroup differences. There were many similarities, as evidenced by the relatively narrow ranges of means on some outcomes, yet this large sample allowed us to identify several differences. How can researchers group study participants in ways that make sense to students? Adding a multiracial or multiethnic label may help acknowledge those youths who do identify with more than one race or ethnic group.

Our findings indicate the need for further studies to understand how these youths think about race, ethnicity, nationality, and culture. Do youths from different socioeconomic and acculturation levels think differently about these constructs? Qualitative studies would provide insight into the relationship between ethnic label selection and the national origin and conflict perspectives explored in this study. Narrative methods in particular have been found to support a multiracial individual’s own creation and expression of identity (Jackson, 2007).

Our finding that educational expectations predicted choice of labels is particularly important for future researchers to examine. Ajzen’s (1991) theory of planned behavior might be a useful framework for such research in understanding how attitudes, subjective norms, and perceived control shape behavior in the context of culture and race. Does the internalization of ethnic labels mean that youths also internalize specific attitudes and norms related to educational expectations?

Future research needs to explore how contextual factors, including specific school-level variables, affect the meaning youths attach to ethnic labels, the selection of ethnic labels, and ultimately identity development. This study was conducted in an environment wrought with anti-immigrant sentiment. Since then these sentiments have led to policies such as the passage of Proposition 200 which requires Arizonans to show proof of citizenship in order to receive public services (Hawley, 2005), new employer sanctions against those who hire undocumented workers, the use of police deputies to arrest immigrants who are undocumented, and the establishment of a hotline that allows the public to report suspected undocumented immigrants (Gonzalez, 2007). In this climate, Mexican immigrant mothers report their children are at times afraid to go to school for fear that they or their parents might be deported (Moya Salas, 2007). Hence, future researchers must examine how living in a nativist environment affects adolescents.

Finally, the questionnaire used the common term “American Indian.” Future studies might want to add the term “Indigenous,” along with the example of “Aztec,” to similar questionnaires to determine whether youths of Mexican origin have incorporated this aspect of their heritage. In addition, three Mexican-origin labels (Mexican, Mexican American, and Chicano) were combined this study. Future studies allowing youths to select one or more of these labels may help us better understand the educational and other implications of these labels.

CONCLUSION

The numbers of Mexican-origin residents are increasing across the United States (Guzmán, 2001). Because this population has, on average, higher numbers of household members under age 18 than does the general U.S. population (Guzmán, 2001), school social workers are increasingly likely to work with youths of Mexican descent. School-based interventions that recognize the diversity among students who claim Mexican, Mexican American, Chicano, or other labels will allow school social workers to develop interventions that support the ethnic identity development of Mexican-origin youths. Forcing youths to select only one ethnic or racial label—or grouping together all youth who have selected a Mexican-descent label regardless of whether they have selected other ethnic labels—prevents researchers from seeing variation among youths within this large category. This study confirms that having the option of checking