Recently, I commented online about a web page form I was required to use in order to obtain localized health services. Among several questions to determine my eligibility, there were a couple related to my ethnicity. Only that was not the term used in the form. What I originally noticed was the order of the answers, which seemed to be in descending order of a population ratio (Pareto, even). My take was that “white” should not be listed first in the choices, but it could be placed in alphabetical order. In response, online contacts questioned the way the question was asked, suggesting I had overlooked a larger issue.
Before we dig into the implications, a little background. I studied environmental engineering in college, then worked at federal and state level jobs, though the titles were, respectively, “sanitary engineer” and “public health engineer” since it takes time for civil service structures to rename positions. In these roles, I investigated different manners of pollution, from the localized dumping of wastewater or solid waste, to the more regional effects of air pollutant dispersal. The investigations were authorized based on then-existing statues, laws on the books as written by public legislators, and by policies and procedures created by my agencies to enforce the directives given us.
Often, my tasks were done on the border between industry and citizenry, true to the aptly titled environmental protection agency. We protected people from the risks of environmental contamination, in the ideal world. In the real world, though, powerful civic actions, or persuasive legal actions restricted our capability to take effective action. At times, I attended public meetings where proposed rules were presented so explanations and plans could be aired out. Each government level has established ways to collect public input, and uses designated resources and experts to plan future steps. I have listened to community activists complain we were not protecting them, and to industry flaks chastise us for hindering their progress.
In the background, health policies are written to not just improve overall status, but to address prior inequities in human health factors such as longevity, cancer rates, etc. When we see data showing specific population subsets have shorter life expectancy or disease prevalence, laws are written to attempt to correct these disparities. Here’s where it gets tricky: people descended from enslaved African ancestors have suffered from discrimination, lack of educational facilities and poor health care in America. To identify those needing the most assistance implies that we collect data showing who and where they are (below, I will relate historic documents guiding federal efforts). Unfortunately, the word used to distinguish one person from another is “race,” which as my SAP community peers pointed out, is scientifically unsound. We’re all of the human race.
Pull-down poll question on “Race*” listing white as the first entry.
Caveat onscreen: “The CDC requires us to capture this information”
If “race” is the wrong term, is that all? No. The collection of information saying this person is descended from people who lived in Africa, or who came from Asia, or other populations, can quite easily lead to data misuse, or worse.
At first, I had a glib answer to the data collection question: it is required by law. Yes, it is, and the laws were written to correct past and current inequities in health care service delivery. In my view, the data collection is not just “required by law,” it is in the best interest of our society. However, I support the contention that managing these data involve risk, and the way the questions are framed is critical to going beyond the statutory intent to positive side effects such as improved self-image, higher tolerance of others, and a less violent society.
The crux, or a primary one, is the term “race” in the context of genealogical or cultural differences. The federal documents I’ll reference later use the “race” and “racial” as an artificial separator, just as the incorrect social policies might have labeled public drinking fountains as “white only”, or “colored”.
A 1997 document said, “Development of the data standards stemmed in large measure from new responsibilities to enforce civil rights laws.”
The data form collection designer has an opportunity to correct the unscientific phrasing of past practices; but what about the forms consumer? Should they complain every time they see such injustice? Probably.
Fast forward to the 2020 pandemic year, also the year of the U.S. Census collection. For maximum confusion, the forms have Question 6 on “Hispanic origin” and Question 7 on “race.” And, as the page says, “For this census, Hispanic origins are not races.” Americans are familiar with these questions, having faced them for decades, if not longer, on official publications. Non-Americans, I learned, may not know about this, other than our somewhat related families in New Zealand and Australia.
The U.S. Bureau of Labor Statistics collects information on jobs, careers, etc. These are consumed in reports and analyses such as employment/unemployment trends, workforce characteristics, voting/redistricting, and access to banks/credit.
An article in the news today describes the disproportional impact of the current U.S. baby food shortage on women in minority groups. The sub-heading says “Black, Hispanic women facing a variety of hurdles,” then goes on relate the statistics collected by the CDC related to the topic. I see this as an instance of the maxim “you can’t fix what you don’t measure.”
Oversimplifying the secondary question about ancestry is the Hispanic question. I speculate that the intent is to address “Spanish as a second language”, or perhaps “English is not the primary language at home,” resulting from increased U.S. population from countries in Latin and South America. In Canada, similar challenges revolve around using French and English in official sources, whether road signs or statutes.
As I described in an earlier post about gender identification, if you as a form designer can push back on requirements to collect data that could introduce bias, or is just not necessary for the intended purpose, leave those questions off. Don’t gather the data, don’t store it, and thus have no worries about future data maintenance. For legally required content, you have fewer options. Or none.
Is there a better way to ask the question about ancestry and heritage than what is usually done? I welcome your comments. It is clear from the online conversation that the term “race” is ambiguous at best. Very far down in the 2016 Federal Register link. for example, is the claim that race/ethnic categories should not be interpreted as “scientific or anthropological.” Yet the link text uses these terms specifically.
The primary complaints I heard were “why collect these” and “these could be used badly.” I think I addressed the first; what about the second? It is unsettling, after years of data breaches and personal information collection, to think there could be negative outcomes from good intentions (I heard the road to Hell was paved that way).
[short version: visit www.regulations.gov]
I am not going to link to the address in the image of above, as it seems to already have the fixed notion that race and ethnicity are the terms to be used without appeal or discussion.
One startling observation: in the 11 principles set forth as recently as 2016, the risks of managing these data is not portrayed. Nothing about privacy or worse, other than ” Respect for individual dignity.”
As I wrap this up, I wanted to thank my European network colleagues for calling my attention to this topic. I promised I’d expand on my thoughts. Please share yours.
For those not as familiar with recent American history, I recommend reading “Sundown Towns” by James Loewen, “The Black Butterfly” by Lawrence T Brown, and “Lorraine for Lorraine” by Imani Perry. Further back, anything about the The National Museum of African American History and Culture complex, and also anything about “Brown versus Board of Education [of Topeka Kansas]”. Even Britannica uses the term “racially segregated public schools” – britannica.com/event/Brown-v-Board-of-Education-of-Topeka
The federal documents I located online are from 1978 forward. I dug these up to see for myself if there is discernible bias in using the term “race” as a population descriptor. Was this addressed with any sensitivity? While there are plenty of listed experts of the PHD, MD, and similar standing, not much about the “why” of the term. It is as if “we all know there is a problem” is the start of the discussion. Yes, there are definite problems if different people have different health challenges and life expectancy solely based on who their parents are.
One of the (chemical) industrialized areas I inspected in Baltimore City has an adjacent neighborhood, which has high reported cancer rates. Even without knowing who lives there, these statistics tell a story that poorer people are more likely to be forced to live in these less desirable areas.
“Current principles for the categorization of “race” and “ethnicity” in federal statistics are specified in the Office of Management and Budget’s (OMB) Directive 15, “Race and Ethnic Standards for Federal Statistics and Administrative Reporting” developed in 1978 by OMB’s Office of Federal Statistics Policy and Standards (9).”
obamawhitehouse.archives.gov/omb/fedreg_1997standards – Federal Register Notice
October 30, 1997
1. The racial and ethnic categories set forth in the standard should not be interpreted as being scientific or anthropological in nature.
Electronic Availability: This document is available on the Internet on the OMB Web site at: https://www.whitehouse.gov/sites/default/files/omb/inforeg/directive15/race-ethnicity_directive_2016FRN1.pdf.
Due to changes in government web pages after President Obama left office, the above document link is stale; the current  location is:
Acronyms & Abbreviations
- BLS – Bureau of Labor Statistics (Labor)
- CDC – Centers for Disease Control and Prevention (Health)
- EOP – Essentially, the U.S. executive office (one of the 3 branches, others being legislative and judicial), or White House (there it is again)
- .gov – Top level internet domain (Government of the United States)
- NIH – National Institutes of Health (Health)
- OMB – Office of Management & Budget (Administration)
Should I post about the “American Indians?”