BACKGROUND ON THE CENSUS ADVISORY COMMITTEE
The 2010 Census Advisory Committee (CAC) provides advisory input on the design of the 2010 census, the American Community Survey, and related programs. Committee members represent a range of census stakeholders, and APDU’s seat on the Committee provides a channel for APDU members to comment from the data user perspective.
Ken Hodges is your APDU representative on the Census Advisory Committee. This report describes the most recent meeting of that Committee. Reports on these meetings are designed to keep APDU members informed on census activities, and to encourage feedback. Contact Ken anytime at khodges@claritas.com with comments, questions, or suggestions.
NOVEMBER 30, 2006 JOINT MEETING OF THE CENSUS ADVISORY COMMITTEES
Background on the One Day Joint Meeting
The November 30 meeting was a joint meeting including the 2010 Census Advisory Committee (on which APDU has a seat), as well as the Race and Ethnic Advisory Committees, and the Advisory Committee of Professional Associations.
The Census Bureau initiated the meeting to help finalize content decisions for the 2010 census by January 2007. The early determination date reflects the desire to achieve comparability with the American Community Survey (ACS), the approaching 2008 Dress Rehearsal, and advance time needed for software development. In advance of the meeting, the Census Bureau had provided materials describing the 2005 content test, and the Bureau’s proposals for 2010 content and question wording. These materials are available on the Census Bureau website (under Advisory Committees). The committees deliberated in advance via conference calls, and were asked to finalize and present their recommendations at the November 30 meeting.
Summary of Census Bureau Proposals
Many of the proposed changes reflect the kind of incremental improvements we expect from each new census. For example, the tenure (owner vs. renter) question has asked if the unit is occupied without payment of “cash rent.” But some renters who pay by check are confused by this term, so the Bureau proposes to ask simply about “rent.” And because some respondents with home equity loans report owning “free and clear,” the proposed question would instruct respondents to “include home equity loans” as a form of mortgage.
The relationship to householder question would change “Natural-born son/daughter” to “Biological son or daughter” because some adoptive parents have an unfavorable reaction to “Natural-born,” and some respondents misinterpret it to exclude births involving drugs or caesarian delivery. Space considerations mandate that one of the relationship categories be dropped, and the Census Bureau proposes dropping “Foster child,” because it had the fewest responses in the 2000 census.
Changes such as these are relatively uncontroversial, but the race and ethnicity questions always generate intense interest and debate. Consideration of the race and ethnicity questions figured to be the meeting’s main event. Of most interest was a “three question option,” in which the 2010 (short form only) census would ask not only race and Hispanic ethnicity, but a third question on ancestry, which has previously been on the long form. It was hoped that the write-in ancestry responses would help in imputing race, where necessary, and allow for the abbreviation of the race and ethnicity questions. Specifically, the race question would no longer offer detailed response options for Asian and Pacific Islander race, and the detailed Hispanic categories also would be dropped. Six alternative “panels” for race and ethnicity were tested. Panels 1 and 6 were controls, retaining the two question format from previous censuses. Panel 1 was very similar to Census 2000, while Panel 6 included revised instructions and formatting. Panels 2 through 5 were variations on the three question option. The panels also varied in wording and the “examples” offered for various categories (e.g., Argentinean and Dominican might be offered as examples of “Other Hispanic” ethnicity. The 2005 tests indicated little impact on the totals for the basic Hispanic, Asian and Pacific Islander categories, but a sharp drop in the numbers reporting detailed nationalities within these groups, so the Census Bureau proposes to stay with the two-question arrangement used in recent censuses.
The Meeting Itself
Teresa Angueira, the designated federal official, started the meeting by introducing the new representatives, and federal officials in attendance. Census director C. Louis Kincannon made welcoming remarks, including a “wow” for the new Census Bureau building, and the recently opened conference facility. He noted that with this facility, all advisory committee meetings can be held at the Census Bureau, thus allowing more interaction with Census staff. Kincannon commented on the importance of finalizing 2010 content, said he anticipated robust debate and valuable input from the meeting, but cautioned that wholesale changes are unlikely, as the proposals already reflect much work and input. Kincannon mentioned his recent resignation (and that of deputy director Hermann Habermann) only briefly, but reassured that most of the major decisions about the 2010 census have already been made. Kincannon then introduced Cynthia Glassman, the new under secretary for Economic Affairs at the Department of Commerce. Glassman remarked on the importance of the census, and expressed confidence in the Bureau’s post-Kincannon/Habermann management team.
Next, Howard Hogan, the Census Bureau’s Associate Director for Demographic Programs, lead a Q and A session, soliciting technical or clarifying questions on the Census proposals. But such questions took a back seat as committee reps took the opportunity to express concerns. One rep thanked Kincannon for his service, expressed concern with the recent resignations, and noted the importance of finding replacements with solid knowledge of census issues. Under secretary Glassman reminded the attendees that the Census director is a presidential appointee, and assured that the replacement will be highly qualified. The other positions are internal hires, and she assured that Commerce would work to ensure they are filled by highly qualified individuals.
Turning to the day’s topic, a rep asked if a Spanish language version of the relationship question had been tested. Angueira explained that there was a Spanish language version of the forms, but that language was not part of the test, and that later cognitive tests would focus on the Spanish translation. The rep argued the importance of including language as part of such tests, rather than making decisions, and addressing translation issues later. This recommendation was echoed throughout the day by several other reps.
Following relatively brief discussion of the tenure, relationship and age questions, Hogan turned to the race and ethnicity recommendations. A rep from the African American Advisory Committee expressed concern that the race question lists only three terms for this population – Black, African American, and Negro. The rep criticized the use of “Negro” as offensive, and argued for the addition of terms such as Ethiopian, Jamaican, and Haitian, which some Blacks identify with more readily. The rep expressed frustration with the Census Bureau’s lack of responsiveness to this recommendation, which apparently dates back many years. Claudette Bennett, with the Census Bureau’s Population Division, explained that there is a sizable population (about a half million) that identifies with neither Black nor African American, but does identify with Negro, and observed that concepts such as Ethiopian and Jamaican are part of the ancestry question. Others from the African American Advisory Committee agreed with the use of the term Negro.
Another rep broadened the question – asking why Asian and Hispanic populations have detailed subcategories as check boxes, while the Black or African American category does not. Other reps asked variations of this question. Hogan was still hoping for technical and clarifying questions, but this became the central question of the meeting – why do some race and ethnic groups get detailed ancestry/nationality response options on the race (or ethnicity) question, while others do not? Such response options, or check boxes, take up a lot of room on the census form – contributing to what Jay Waite (Associate Director for Decennial Census) calls the short form’s “real estate problem,” and adding all options that might be helpful is not feasible. Waite explained that they had hoped the three question option (ancestry on the short form) would allow the census to drop all subcategory boxes, but that this did not work out, as some subpopulations experienced large drops.
Given the competition for scarce space on the census form, one might have expected tension between the race and ethnic group reps, but the reps were respectful of the needs of other groups. Still, much of the Q and A session consisted of reps explaining why the populations they represent need extra response options (boxes) in order to achieve an accurate count.
Committee Concurrent Sessions
The Q and A session showed no sign of concluding, but mindful of our schedule commitment, Hogan brought it to a close, so the committees could meet in concurrent sessions to finalize their recommendations.
The 2010 Census Advisory Committee had an advance conference call, but only a few reps participated (your APDU rep was one), so we were almost starting from scratch. And with reps reflecting diverse stakeholder groups, one could argue that the 2010 Committee had a more challenging task than the others. The race and ethnic advisory committees met separately (e.g., the Asian Committee met on its own), and could be very focused. Even the Advisory Committee of Professional Associations met separately by association. The 2010 Committee brought a wide range of perspectives into one meeting, and had a predictably wide ranging discussion.
We started with tenure—agreeing on the need to consider how the question translates to other languages, and that this is just one example of the need to integrate language into the content testing process. We even considered why tenure is on the short form, and in response, Census staff explained that it is a fundamental housing measure, used to weight federal surveys, and serves as a powerful indicator of census response rates and coverage. With respect to relationship to householder, many were concerned with the proposed elimination of the “Foster child” category, since many in foster care are African American, and are now aging out of foster care and into poverty. Some wondered if “Roomer or boarder” might be a better category to drop, but Census staff, without advocating one way or another, explained that responses to this category are useful as an indication of income potential.
Moving to the race and ethnicity questions, Clark Bensen (POLIDATA Political Data Analysis) asked about the history of the detailed Asian and Pacific Islander subcategories. Karen Narasaki (Asian American Justice Center) explained that the response options recognize the fact that two thirds of the Asian American population is foreign born, and that omitting these options yields a less accurate count of the Asian and Pacific Islander populations.
Committee Vice Chair Lee Adams (Little Rock Missionary Baptist Church) made a case for keeping the race and ethnicity questions simple – keeping in mind the perspectives of the typical respondent and census field worker.
Committee Chair A. Mark Neuman then asked Helen Samhan (American-Arab Anti-Discrimination Committee) to share her thoughts. Samhan represents a population that is not identified by the OMB race standards, and is not recognized by federal law as in need of civil rights protection. As such, there is no Arab American Advisory Committee, and data for this population depend completely on long form (and now ACS) ancestry data. And with ancestry data now provided only down to census tract level, the inclusion of ancestry on the short form census would make a big difference for this population. Samhan expressed appreciation for the test of the three question option, but was critical of the either/or presentation (ancestry added only if Asian and Hispanic detail are dropped). She pointed out that many Arab Americans do not identify with the race response options provided, and asked why we should have to eliminate Asian and Hispanic response options just to get a compact ancestry question on the census.
There was broad sympathy for Samhan’s position, but the Committee did not adopt her recommendation that ancestry be added without eliminating the detailed race response options. As part of the discussion, committee chair Neuman observed that the history of disenfranchisement for the Black, Asian, and Hispanic populations has been different than that for the Arab American population, and Bensen noted that ancestry is not needed for redistricting, and would add to the already cumbersome tables used in that application. Noting that Asian and Hispanic detail is not provided on redistricting files, and the minor impact (of the three question option) on the broader categories that are used in redistricting, your APDU rep asked for clarification on the need for the detailed response options on the short form. Karen Narasaki and Kim Brace (Election Data Services) explained that there are legislative applications other than the Voting Rights Act that make use of the detailed categories.
Ilene Jacobs (California rural Legal Assistance) expressed the frustration shared by many that we had been given a narrow set of options to consider, forcing us to make bad choices. As time ran out, we were discussing the possibility of adding examples (to clarify race categories) even at this late date in the census planning process.
Presentation of Written Recommendations
The committees then reconvened in the main conference center to present their recommendations.
Census Advisory Committee of Professional Associations
The Population Association of America expressed general agreement with the Census Bureau’s proposed content, including the use of the Panel 6 version of the race and ethnicity questions (two questions with revised wording and instructions). The PAA reps are troubled by the lack of non-English testing, and argued that—with the ACS in the field throughout the decade—testing and question revision should be an ongoing process.
The American Statistical Association expressed agreement with most of the Census Bureau’s findings and proposals, and cautioned against the use of untested options. ASA also called for elevating the importance of non-English testing, and expressed a preference for the Panel 6 version of the race and ethnicity questions.
The American Marketing Association recommended that the 2010 census proceed as proposed by the Census Bureau. They stressed the risk involved in major changes, and stressed that changes to the race and ethnicity questions should focus only on their contribution to the broader OMB categories.
The American Economic Association sent no representatives to the joint meeting, but in a brief prepared statement, indicated that they had no recommendations or comments on the Census Bureau’s proposals.
Census Advisory Committee on the Asian Population
This committee expressed support for the Panel 6 race and ethnicity questions, citing their superior performance relative to the other options. The committee also strongly opposed Panels 2 through 5 (the three question options) due to their severe negative impact on the count of Asians and Hispanics reporting detailed origins. The committee also called on the Census Bureau to put more into the advertising and partnership efforts related to these population groups.
Census Advisory Committee on the American Indian and Alaska Native Populations
This committee also supports the Panel 6 race and ethnicity questions, but recommends more inclusive examples of origins (such as Central American origins). The committee also called on the Census Bureau to revise its testing process to include American Indian reservations, and to hold advisory meetings two times per year. The recommendation for two meetings per year (as opposed to the present schedule of one) was shared by many of the advisory committees.
Census Advisory Committee on the African American Population
This committee indicated acceptance of the Census Bureau proposals for the tenure, sex and age questions, but recommended that the “Foster child” category be retained in the relationship question – noting that 60 percent of those in foster care are Black. They recommend the Panel 1 version of the race and ethnicity questions, but with added examples for Black or African American race (such as African, West Indian and Caribbean).
Census Advisory Committee on the Hispanic Population
The Hispanic committee expressed unanimous support for the Panel 6 race and ethnicity questions, citing the more complete data on detailed origin and the high rate of self-identification. The committee strongly opposes the Panel 2 through 5 options, which involved shortened race and Hispanic questions combined with a short form ancestry question.
Census Advisory Committee on the Native Hawaiian and Other Pacific Islander Populations
The committee leans to the Panel 6 race and ethnicity questions, but recommended expanded response options for the Pacific Islander population – such as Marshallese, Fijian and Tongan. They also recommended further research to better understand the “Other Pacific Islander” population, and increased advertising and partnership efforts relative to this population.
2010 Census Advisory Committee
This is the committee with APDU representation, and as noted above, it is unique among the advisory committees in the diversity of the interests it represents. Committee chair A. Mark Neuman described the committee’s discussions and recommendations on a variety of topics, including the need to devote more attention to translation of concepts (such as home equity loans) to Spanish and other non-English languages. Neuman also described the committee’s concern that “Foster child” be retained as a relationship category given the many African American children aging out of foster care, and into poverty. With respect to race and ethnicity, the committee expressed frustration with the either/or choice between a short form ancestry question and Hispanic and Asian detail on the race and ethnicity questions. Neuman reported that some on the committee still strongly favor the inclusion of ancestry on the short form (as in Panels 2 through 5), but that the overall preference was for Panel 6 – modified to include Panel 1’s instruction to mark one or more “races” as opposed to one or more “boxes.” To the extent feasible, the committee recommended the inclusion of more examples in the race and ethnicity questions, and recommended that the census needs to determine better ways to identify indigenous populations from Central and South America. With its diverse composition and perspectives, the 2010 Committee’s recommendations remain a work in progress, and committee members were revising and refining even as the meeting was concluding.
Following the committee presentations, Jay Waite brought the meeting to a close by thanking the committee members for their work, and expressing appreciation for the many good suggestions the committees had offered.