Monday, August 08, 2011

Race and the U.S. Census -- from April 2010

I wrote this in April 2010, and tried to get it published without success as an Op-Ed. Printing it now in response to a debate some friends are having regarding race and the census. I may have a follow-up post on it soon.

More than sixty years ago, my grandparents left their home in a village in Pakistan, packed all their belongings in an ox-cart, and walked on foot for hundreds of miles – along with millions of other Sikhs and Hindus – across the border to a newly independent nation: India. Millions of Muslims from the other side of the border were going in the opposite direction, settling in Pakistan, where the immigrants from India are to this day identified by other Pakistanis as “Mohajirs” (“Migrants”), held to be somewhat separate from the old-timers who are really native to the region.

In 1973, my parents migrated again, this time from India to the United States. As my father still likes to brag, he and my mother came to this country with nothing but a medical degree from an Indian university, and about $300 in their pocket. My parents’ profile was not so different from that of thousands of other immigrants from various Asian countries, who came to the U.S. seeking economic opportunity after the reforms in the Immigration and Nationality Act of 1965. Four decades later, this wave of immigration has made the United States easily the most ethnically and racially diverse large nation in the world. As of 2000, the census recorded approximately 1.9 million Asian Indians living in the United States (counting people who checked two boxes on the Census for “race”), and this number is expected to be substantially higher as of 2010. By comparison in 2000, approximately 150,000 wrote in “Pakistani,” 43,000 wrote in “Bangladeshi,” and 20,000 wrote in “Sri Lankan” Unfortunately, the census data from 2000 gives us no way to ascertain how many immigrants from Pakistan, Bangladesh, and Sri Lanka might have simply defined themselves, racially, as “Asian Indian,” since that is the closest category on the survey that does not require a write-in answer.

While the race question on the Census has for the most part remained the same from 2000 to 2010, there is one significant change, which comes in the form of a “hint” that has been added in the 2010 census to the category “Other Asian.” The hint, in parenthesis, points to a write-in box suggesting several nationalities that may be written in, including, “Hmong, Laotian, Thai, Pakistani, Cambodian, and so on.” The addition of this small hint in the 2010 Census might seem small – perhaps it is meant merely as a clarification, and not as a substantive change – but it raises some difficult questions for Americans whose families hail from the Indian subcontinent.

Why is it that my parents and I are “Asian Indian,” while our friends who came (more recently) from Pakistan – people who share our complexion, languages, and much else – are given a separate slot? Similarly, while Bengali-speaking Indians will almost certainly fill out “Asian Indian” for Question 9, Bangladeshis, who again share the same language, appearance, and many deep cultural links with Bengalis on the Indian side of the border, might feel encouraged to follow the lead of Pakistanis in identifying themselves as of another “race,” again under “Other Asian.” And this confusion can be seen yet again regarding the ethnic groups from the southern part of the Indian subcontinent: Tamil Indians, who speak a language and have a distinctive way of life that resembles their Tamil-speaking cousins in Sri Lanka. If Tamil Indians are “Asian Indians,” how are Tamil Sri Lankans “Other Asian” – which by the logic of the new hint, they must be? All of these groups now have significant immigrant representation in the United States. It is puzzling that despite the many cultural and historical commonalities shared by people from the Indian subcontinent, they are seen by the Census as separate “racial” groups.

The confusion over the nationalities of the Indian subcontinent is only one among several signs that the Census remains hopelessly confusing on “race.” How is it that there is a distinct box for the Samoan “race,” but no boxes for “Arab” or “Persian/Iranian,” both of which have substantial populations in the U.S., and who are likely to be split between “white” and “some other race”? And what about Afghan-Americans? Are they “Asian,” “white,” or “some other race”?

After some consideration of these categories, it becomes clear is that while some boxes in Question 9 on the Census (the “race” question) actually refer to categories everyone would agree relate to “Race” (“white,” “black,” “American Indian or Alaska native”), there are several others – “Asian Indian,” “Japanese,” “Chinese,” “Korean,” “Filipino,” “Vietnamese,” and “Samoan,” that refer to what are really nationalities. While official Census statements and directives suggest that the survey should have one question that relates to “ethnicity” (Question 8, addressed to Hispanics), and one on “race,” the reality is that Question 9 in particular remains a mish-mash of social groupings that could variously be described as “racial,” “ethnic,” or “national.”

Admittedly, the changes to the approach to race taken by the census are not all bad; indeed, the current configuration is a great improvement over versions of the “race” and “ethnicity” questions in earlier Census surveys. (Before 1970, for instance, Asian Indians, along with others from the Indian subcontinent, were all simply expected to check “white,” even if it was far from clear that they were going to be recognized as such by ordinary Americans.) The Frequently Asked Questions section of the current Census.Gov website, in response to the question, “Why doesn't the race question include more categories?” refers to a 1997 directive from the White House’s Office of Management and Budget (OMB), setting up a general frame of reference that included five broadly defined racial groups. The 1997 OMB directive, which can be read in its entirety here [], also states that, “The categories represent a social-political construct designed for collecting data on the race and ethnicity of broad population groups in this country, and are not anthropologically or scientifically based.”

But this statement by the OMB raises more questions than it answers. If the racial and ethnic categories on the U.S. Census are not “anthropologically or scientifically based,” on what ideas or facts exactly are they based? While one understands the imperative to gather broad population data, wouldn’t that data be better if it were construed in keeping with terms and social groupings derived from scientific anthropology?

Let us briefly take a stab at a better system of categorization for Asians. Admittedly, there is no magic bullet – and certainly no simple way to balance the challenge posed by balancing ethnographical accuracy against the evident need to be brief and keep the form simple enough that it can be easily understood and filled out. Still, I would suggest that the various “Asian” nationalities could best be re-defined into three large, regional groups: South Asian, Southeast Asian, and East Asian. Though obviously culturally heterogeneous in many ways, Pakistanis, Indians, Bangladeshis, Sri Lankans, and Nepalis, come close to forming what ought to be an intelligible regional (if not “racial”) category – a category that anthropologists would call South Asian. Similarly, Japanese, Korean, and Chinese people might be identified as East Asian, while Filipinos, Burmese, Cambodians, Laotions, Thais, Indonesians, Fijians, Malaysians, and other Pacific Islanders, might all be Southeast Asians. Again, these three regional sub-categories of “Asian” are far from perfect, and each of the terms listed is abstract enough that it would probably need a list of “hints” along the lines of the hints currently given for “Other Asian.” While this change might not solve all of the analytic problems relating to Asian “racial” identity on the Census, it certainly would lead to better data than a survey that encourages Asian Indians and Pakistanis to see themselves as distinct “racial” groups.

1 comment:

omar ali said...

Minor point about Mohajirs. That term in Pakistan rather quickly came to mean Urdu-speaking migrants from North India, who mostly settled in the urban areas of Sindh (particularly Karachi). Migrants from East Punjab were assimilated as Punjabis and dont usually identify as Mohajir. Nobody would call Nawaz Sharif or Zia ul Haq Mohajirs, though both families came from East Punjab...