Hispanic\Latino Ethnicity and Identifiers

Jorge Chapa in Encyclopedia of the U.S. Census. Margo Andersen, editor-in-chief, Jorge Chapa, Connie Citro and Joe Salvo, editorial board. (Congressional Quarterly Press: Washington, DC, 2000).

Introduction

The Census concept of Hispanic ethnicity and the various identifiers by which the conceptualization was made concrete have changed many times since the first crude effort to conceptualize and identify Hispanics who were not immigrants or the children of immigrants made in the 1930 census. These changes reflect the substantial changes in the composition of this population that have occurred over the course of the twentieth century. They also reflect the increase in the size and status of the population and in the Census Bureau’s and the nation’s awareness and knowledge of this group. During the early part of this century, almost all of the population now identified as Hispanic were people of Mexican origin who were largely concentrated in a few southwestern states. In contrast, the 2000 census will report data on more than twenty categories of Hispanics who are to be found in increasingly large numbers in all fifty states. Since the composition of this population and its conceptualization have changed tremendously, this article will use “Hispanic” in its contemporary sense and note how earlier identifiers and conceptualizations differed from the contemporary meaning.

1850-1930 – Conflating Immigrants, Race and National Origin

The census has counted the by race since its inception and has kept track of immigrants since 1850 and their children from 1880 to 1970 by the use of nativity and parentage questions. Censuses from 1910 through 1970 excluding 1950 determined the language respondents spoke at home as a child, also known as their mother tongue. The mother tongue questions were typically determined and/or tabulated only for immigrants or the children of immigrants. Thus the enduring Census concerns with race and immigration were not a fully appropriate precedent for enumerating Hispanics. The first significant population of Hispanics in the US was found among the residents of Texas when it became a state in 1845. The lands annexed in 1848 as a result of the Mexican War added substantially to the total Hispanic population. The history of the Southwest shows that the descendants of the original Hispanic residents of the annexed areas and those of the Mexican nationals who continued to migrate there were typically relegated to a subordinate status in a largely segregated social world. So the typical immigrant notion did not apply. The census concept of race did not fit well either. In varying degrees, the populations of Central and South America are largely mestizos -- the result of the racial mixture of European colonizers and the indigenous residents. Motivated by the generally xenophobic concerns, the 1930 census attempted to enumerate these Hispanics by using the concept of a Mexican race.

There were many serious problems with this approach. Many of these Hispanics were US citizens and the US-born children of US-born parents. The Mexican identifier appropriately applies to citizens of Mexico. Additionally, many of these Hispanics did not want to be identified as members of a socially subordinate group commonly referred to as Mexicans, regardless of their nativity or how many generations their ancestors had resided in the US. The preferred and polite term used as an alternative at the time was Latin. For example, the name of the organization known as LULAC, the League of Latin American Citizens founded in Texas in 1929 is an example of this use. The name also emphasizes the US citizenship of many Hispanics. Unfortunately, LULAC’s name did not send its message as clearly as it could since Latin American Citizens can also be understood to refer to citizens of Latin America, a name which has been used at least since the 1890s in the same sense it is today. One of the clear indications of the inadequacy of the Mexican race approach is that many people were identified as being of Mexican birth or parentage, but not of Mexican race. This highlights another problem with approach. This identifier depended on the judgement of the enumerator which apparently was neither consistent nor reliable. Finally and perhaps most important, being racially designated as Mexican excluded the possibility of being classified as white. At the time many rights and privileges, including the right to become a US citizen, were explicitly available to whites only. Because of these problems and in response to protest and litigation, the Census Bureau dropped the use Mexican race identifier after 1930. This experience also set the precedent for the current practice of separating race and Hispanic ethnicity into two items on the census questionnaire.

1940-1970 – Objective Identifiers and Post-Enumeration Attribution

There was not another specific effort to identify and enumerate the Hispanic population until 1950 census. Puerto Ricans started to migrate to the US in large numbers after WWII. Even though all Puerto Ricans were made US citizens by the Jones Act passed in 1917, the 1950 census was the first to identify Puerto Rico as a response to the place of birth question typically used to determine foreign places of birth or parentage.

Also for the first time in 1950, the last names of respondents in the five southwestern states that had originally been part of Mexico (Arizona, California, Colorado, New Mexico, and Texas) were also compared to a list of Spanish surnames. The main reason for restricting the use of the Spanish surname identifier to the Southwest points to one of the identifiers major limitation. Many Spanish surnames are also found among people whose ancestors were from European countries where one of the various Latin-based Romance languages were spoken such as Italy and Portugal. Given their migration and settlement history, these people were more likely to be found outside the Southwest and Hispanics with Spanish surnames were more likely to be found in the southwestern states.

Even if all Hispanics did at one time have Spanish surnames at birth, intermarriage would tend to make this identifier less precise and useful because the marriage a Spanish-surnamed woman to a non-Spanish-surnamed man would typically result in the woman using the man’s surname in her family name. Any children of such a union would also typically be given the non-Spanish surname. The Hispanic mother and her partly Hispanic children would not be detected by the use of the Spanish surname list. The marriage of a Hispanic Spanish-surnamed male with a non-Hispanic female would typically result in the woman being mistaking identified as Hispanic. However the children of this union would normally have the father’s Spanish surname.

Despite these problems, a continuously improved Spanish surname list was used as a Hispanic identifier from 1950 to 1980. This same procedure was also used to identify Hispanics in research based on birth and death certificates and other records and files. One of the appealing aspects of the surname identifier derives from the fact that it could be used retrospectively on information collected without any other means of identifying Hispanics.

The 1970 Census collected and analyzed six different Hispanic identifiers on two different sample or long-form questionnaires in an effort to determine how to the best count the Hispanic population. These included country of foreign birth or parentage, expanded application of Spanish language and Spanish mother tongue, Spanish Surname, Spanish Heritage and, for the first time, self-identification. The number and diversity of Hispanic identifiers found in the 1970 Census mirrors the growth in the size and diversity of the Hispanic population and the strength of Census Bureau’s urge to get a good grasp on this population. By this time, the exodus from Cuba had resulted in a large increase in the populations of Cubans; the number of Puerto Ricans resident in the US continued to grow; and, due to the initiation and termination of the Bracero program, the number of recent Mexican immigrants swelled all on top of the growing numbers of US-born children with Hispanics ancestry.

In 1970 the Spanish language and Spanish mother tongue variable were determined for all of the respondents in the 15 percent sample including the respondents who were the US-born children of US-born parents. Spanish language was made a household rather than an individual attribute. The responses were tabulated so that all of the persons in a household where either the head or the spouse had spoken Spanish as a child were counted as having a Spanish language background.

Spanish Heritage was a new composite identifier created during post-enumeration data processing with different operational definitions in different parts of the country. In the five Southwestern states it was the combination of people who either had a Spanish surname or who had lived in a Spanish language household. In the three Middle Atlantic states of New York, New Jersey and Pennsylvania anyone who was born in Puerto Rico or who Was of Puerto Rican parentage was deemed as being of Spanish Heritage. Finally, in the rest of the states, it was equated with the expanded concept of Spanish language.

1970-2000 – Standardizing Subjective Self-Identification as Spanish/Hispanic Origin

The 1970 Census 5 percent long form questionnaire asked, “Is this person of Spanish/Hispanic origin?” The possible responses were, “Mexican,” “Puerto Rican,” “Cuban,” Central or South American,” “Other Spanish.” and, “No. none of these.” The extensive analysis of all of the Spanish identifiers used in 1970 showed that this identifier was the best because it was the most consistent, it distinguished among Mexicans, Puerto Ricans, Cubans, etc., and it worked for respondents who were neither foreign born nor had foreign parentage. The demographic advantages of this identifier coincide with political and legal considerations. In 1976, Congress passed Public Law 93-311, known as the Roybal Resolution requiring the use of a self-identified Hispanic question on federal censuses and surveys. The use of this identifier was further promulgated in the OMB Directive 15, first released in 1977. However, it is worth noting that Directive 15 permits the use of a combined race and Spanish origin question. The data collected from a combined question are significantly different than data collected using separate race and Hispanic questions. Self-identification has now become the accepted standard for determining Hispanic origins. Slightly modified and improved versions of the question have been part of the 100 percent questionnaire in 1980-2000. One modification in these subsequent Censuses was to make the “Mexican” origin response category more inclusive by changing it to, “Mexican, Mexican American, Chicano.”

Despite its advantages there were several problems with this question. One was that there was a large rate of non-response. Based on the analysis of other characteristics, many of the non-respondents were non-Hispanics. In subsequent Censuses, this problem was addressed by placing the negative response, “No, not Spanish/ Hispanic origin,” first since it applied to most of the population. In 2000, non-response was further addressed by asking the Hispanic question before the race question. Test showed that many non-Hispanics did not answer the Hispanic question when it was asked after the race question because they felt that their response to the race question also responded to the Hispanic question.

Another problem with the 1970 question is many of the non-Hispanic residents of Alabama, Arkansas, Georgia, Mississippi, North Carolina, South Carolina, and Tennessee identified themselves as being of Central or South American origin because they were Americans (US citizens) living in the Southern or Central part of the US. Subsequent Censuses dropped this specific response.

One other problem was that it was unclear who the “Other Spanish/Hispanic” respondents were. This problem has been solved in 1990 and 2000 by allowing those who identified as other Hispanics to write in a more specific group.

In 2000 in order to reflect its growing popularity of the use of Latino, all of the residents of the U.S. were asked if they were “Spanish/Hispanic/Latino.” Many writers claim that the Census has always preferred to use objective rather than subjective measures. The history of the Hispanic identifiers that preceded self-identification can be seen as steps in a search for an objective measure. The fact that the best solution was the subjective Spanish/Hispanic/Latino identifier suggests that will become a central aspect of the Census and of American society in a manner similar to the other major subjective Census question – race.

References

Anderson, Margo and Stephen E. Fienberg. 1999. Who Counts? The Politics of Census-Taking in Contemporary America. (New York: Russell Sage Foundation).

Bean, Frank and Martha Tienda. 1987. The Hispanic Population of the United States. (New York: Russell Sage Foundation).

del Pinal, Jorge. 1996. “Treatment and Counting of Latinos in the Census.” The Latino Encyclopedia, (New York: Marshall Cavendish).

del Pinal, Jorge and Audrey Singer. 1997. “Generations of Diversity: Latinos in the United States.” Population Bulletin, (V. 52, No. 3, October).

Hayes-Bautista, David and Jorge Chapa. 1987. "Latino Terminology: Conceptual Basis for Standardized Terminology.” Journal of the American Public Health Association, (V.77, No. 1: 61-68, January).

Hernandez, Jose, Leo Estrada and David Alivirez. 1973. “Census data and the Problem of Conceptually Defining the Mexican American Population.” Social Science Quarterly. (V.53, No.4, 671-687).

Siegel, Jacob S. and Jeffrey S. Passell. 1979. “Coverage of the Hispanic Population in the 1970 Census: A Methodological Analysis.” Current Population Reports, special studies, series P-23, no. 82, Washington, DC: US Government Printing Office.

Teller, Charles H., Jose Hernandez, Leo Estrada and David Alivirez. 1977. Cuantos Somos: A Demographic Study of the Mexican-American Population. Monograph No. 2, Center for Mexican American Studies, The University of Texas at Austin.