By Terri Ann Lowenthal, Senior Adviser, Funders Census Initiative 2020

When Americans fill out their 2020 Census forms, they will have more choices for identifying their race, ethnicity, and national origin than ever before. It is almost a cliché: we are an increasingly diverse nation — a factor that permeated much of the 2016 election dialogue. No wonder, then, that the census questions on race and ethnicity generate more interest, scrutiny, criticism, and debate than any others. Given the complexity of the topic, it’s also not surprising that the U.S. Census Bureau began researching ways to improve these questions for the 2020 Census literally during the 2010 count!

The collection of accurate race and ethnicity data in the census is vital for many reasons. Topping the list is the implementation, enforcement, and monitoring of civil rights laws that promote equality of access and opportunity in voting rights and representation, education, housing, the workforce, health care, and other sectors. FCCP members and their grantees rely on these data not only for their civic engagement work, but also to inform the broader investments their foundations make to promote social and economic justice.

The collection of race and ethnicity data in the census occurs within a framework established by the U.S. Office of Management and Budget (OMB). OMB’s Office of Information and Regulatory Affairs is responsible for federal policy on race and ethnicity statistics, formally known as the Standards for the Classification of Federal Data on Race and Ethnicity (the “Standards”). The Standards, last overhauled in 1997, set forth five minimum racial categories (American Indian or Alaska Native; Asian; Black or African American; Native Hawaiian or Other Pacific Islander; and White) and one ethnicity category (Hispanic or Latino) that federal agencies must use to report data. The current Standards favor a two-question design for gathering race and ethnicity data. Federal agencies are encouraged to collect more detailed data whenever possible, and the census represents the most visible effort to do just that, providing the richest portrait of our diverse population available. The 1997 Standards also gave people the option of choosing more than one race for the first time.

The Census Bureau’s decennial review of the race and ethnicity questions began with the 2010 Census Alternative Questionnaire Experiment (AQE) program, which tested 17 versions of the questions among a sample of half a million addresses. After releasing initial AQE results in 2012, the Bureau continued its research, testing, and consultation with stakeholders about ways to improve the measurement of race and ethnicity, including the collection of detailed subgroup data. Worth noting among the challenges the Census Bureau sought to address: (1) almost half of Hispanics selected “Some Other Race” in the 2010 Census, a category that Congress has required for inclusion in recent censuses, but one that nevertheless requires the Bureau to reallocate responses to one of the race categories included in the OMB Standards for some purposes; and (2) recent censuses did not collect detailed subgroup data for the White or Black race categories. Civil rights stakeholders, in particular, also cautioned the Bureau against reducing the accuracy and granularity of data compared to the 2010 Census, especially by replacing subgroup checkboxes with write-in space.

The additional review led to the seminal 2015 National Content Test (NCT), a self-response-only sample of 1.2 million addresses designed to evaluate refined versions of race and ethnicity questions, both for web-based and paper census forms. Primary goals for the test were to continue evaluating a combined race and ethnicity question, which prior testing had shown was likely to reduce responses in the “Some Other Race” category substantially; test a new category for persons of Middle Eastern and North African (MENA) origin; and assess alternative terminology for eliciting the most accurate, complete responses (for example, “race,” “ethnicity,” “origin,” or simply “categories”).

The Census Bureau released preliminary findings from the NCT in October 2016 and tentatively plans to release the detailed report in early 2017. The results led the research team to recommend two significant changes for the 2018 census dry run:

1. Use a combined race and ethnicity question, with checkboxes (rather than write-in boxes) to collect detailed subgroup data for all categories; and

2. Include a new dedicated MENA response category.

The Bureau will continue evaluating NCT results on question instructions and terminology, as well as the order of subgroup checkboxes and other formatting issues; the 2017 Census Test (scheduled for this spring) provides a final opportunity for real-time assessment.

Meanwhile, OMB has relied on the Census Bureau’s extensive research on improving the measurement of race and ethnicity to propose a review and possible “limited revisions” to the official federal Standards. On September 30, 2016, OMB asked for public comments on the need to examine and revise selected parts of the Standards, including whether the policy should favor collecting race and ethnicity data through a combined question, whether the minimum categories should be expanded to include Middle Eastern and North African, and how to encourage federal agencies to go beyond the minimum reporting categories whenever feasible. OMB is reviewing the thousands of comments it received during the 30-day reply period and is expected to issue a second Federal Register notice, with a longer comment period, in the coming weeks or months (the timetable is uncertain), possibly setting forth specific proposed revisions to the Standards. Beyond the primary areas of review outlined above, OMB must develop guidance on comparing race and ethnicity data over time (called “bridging”) — an important component of the Standards for civil rights data users, including federal agencies.

Over the coming year, funders should monitor final OMB and Census Bureau decisions on the collection of race and ethnicity data, to help ensure that the 2020 Census collects accurate, useful information to guide their activities for the next decade and beyond. They also should begin to map out investment strategies for “get out the count” campaigns that will promote full participation and accurate responses in America’s diverse communities.