Comparison of RACE and Hispanic STATUS for National Tables 2019 to 2021

Andrew Beveridge over 1 year ago

Someone noticed that I had incorrectly computed (using Excel) t in the First Panel of the Table, where I report percent Total Race for the 2019 ACS. The table below reports corrected percents and corrected percent differences for that table. I should have checked this more carefully and I appreciate a user bringing this to my attention. Note that this error or the one I reported earlier affects the difference in the counts between the 2019 and 2021 ACS that is reported. (Those who have circulated this email and tables to others, please send them this corrected version.)

Someone pointed out that without having a crosswalk between the old and new way of coding race and Hispanic status, there is no way to possibly understand how much of the difference in the reported differences represent true change, and how much is due to the method of coding and the changing of the question to elicit fill-ins for white and black. How this will affect the Redistricting Office's Release of the Citizen of Voting Age Population, which is heavily used to understand the distribution of Citizens of Voting Age by various legislative districts, is still to be determined. Which multi-race and single race groups are affected also will require more work, but the complex race and Hispanici iterations of the PL94-171 file, which include every race (including some other race) in every combination are only present for the PL94-171 f file, so the only possible comparison would be the 2010 file.

Attached are comparison tables for race and race and Hispanic status for the 2019 and 2021 ACS. Overall the number of multi-race (two or more race households) has grown from about 11.3 million to 41.9 million between the two ACS's. Among non-Hispanic the increase is about 6 million, and for Hispanic about 24.6 million.

At this writing there is no way to reconcile the numbers or make them consistent so that one can actually look at a trend. I realize the Census Estimates Division, so far, has not adopted this approach. Much of the change in these numbers is due the decision by the Bureau to conduct a content analysis of the fill-ins for the race question answers, and for the Hispanic question. Up to 200 characters were scanned looking for evidence of other races, aside from the one checked. If such a race was found, then the person was coded into that race as well. This change in the 2020 Census handling of race and Hispanic status was ported to the ACS. However, there is no way to know, nor has such information been reported on how many individuals in what race/Hispanic configuration were shifted in this way. Nor is there any available information on what the results would have been if they had been processed, as they were in 2010, and the white and black fill-ins had not been added. In short we have no way to interpret change in race and Hispanic status over time, nor can these numbers be used for comparison purposes when the OMB mandated that other agencies along with the Census, adopted the "Check one or more races" which was first recorded in the 2000 Census.

Having this massive change with no crosswalk for either the ACS or the Census, will make it virtually impossible to do meaningful comparisons with the race and Hispanic data from all of the other sectors of society that use such data to monitor race and Hispanic disparities. This includes, the employment sector through the EEOC, the criminal justice system, the courts, allocation of housing and housing subsidies, birth and death certificates, the health system and virtually any other area that through law or custom or both require monitoring of race and Hispanic disparities. Having used the ACS in court many times to compare housting, employment or jury composition, as well as the operation of law enforcement with the underlying numbers derived from the ACS, I am personally quite concerned. Nor do I know how to give guidance to the users of Social Explorer, which include students at over 500 colleges and universities, and many users from the state, local and federal government, as well as the non-profit and business sectors.

We plan to do more work regarding this issue.

Andy

American Community Survey 2019 National Race and Hispanic Distributions (Note: % Changed in Number Revised for Hispanic by Race Revised)
Corrected September 18, 2022
Population by Race
Group	2019		2021		Difference	% Differ- ence	% Change in Number
Total Population	328,239,523	*	331,893,745		3,654,222	*	1.11%
White Alone	236,475,401	72.04%	202,981,791	61.16%	-33,493,610	-10.88%	-14.16%
Black or African American Alone	41,989,671	12.79%	40,194,304	12.11%	-1,795,367	-0.68%	-4.28%
American Indian and Alaska Native Alone	2,847,336	0.87%	3,158,694	0.95%	311,358	0.08%	10.94%
Asian Alone	18,636,984	5.68%	19,157,288	5.77%	520,304	0.09%	2.79%
Native Hawaiian and Other Pacific Islander Alone	628,683	0.19%	612,448	0.18%	-16,235	-0.01%	-2.58%
Some Other Race Alone	16,352,553	4.98%	23,902,781	7.20%	7,550,228	2.22%	46.17%
Two or More Races	11,308,895	3.45%	41,886,439	12.62%	30,577,544	9.18%	270.38%

	* Corrected Column
Population by Race and Hispanic Status
Group	2019		2021		Difference	% Differ- ence	% Change in Number
Total Population	328,239,523		331,893,745		3,654,222		1.11%
Not Hispanic or Latino:	267,757,777	81.57%	269,364,681	81.16%	1,606,904	-0.41%	0.60%
White Alone	196,789,401	59.95%	192,753,887	58.08%	-4,035,514	-1.88%	-2.05%
Black or African American Alone	40,596,040	12.37%	39,269,297	11.83%	-1,326,743	-0.54%	-3.27%
American Indian and Alaska Native Alone	2,236,348	0.68%	1,749,871	0.53%	-486,477	-0.15%	-21.75%
Asian Alone	18,427,914	5.61%	18,889,050	5.69%	461,136	0.08%	2.50%
Native Hawaiian and Other Pacific Islander Alone	565,473	0.17%	558,717	0.17%	-6,756	0.00%	-1.19%
Some Other Race Alone	839,270	0.26%	1,845,426	0.56%	1,006,156	0.30%	119.88%
Two or More Races	8,303,331	2.53%	14,298,433	4.31%	5,995,102	1.78%	72.20%
Hispanic or Latino:	60,481,746	18.43%	62,529,064	18.84%	2,047,318	0.41%	3.39%
White Alone	39,686,000	12.09%	10,227,904	3.08%	-29,458,096	-9.01%	-74.23%
Black or African American Alone	1,393,631	0.42%	925,007	0.28%	-468,624	-0.15%	-33.63%
American Indian and Alaska Native Alone	610,988	0.19%	1,408,823	0.42%	797,835	0.24%	130.58%
Asian Alone	209,070	0.06%	268,238	0.08%	59,168	0.02%	28.30%
Native Hawaiian and Other Pacific Islander Alone	63,210	0.02%	53,731	0.02%	-9,479	0.00%	-15.00%
Some Other Race Alone	15,513,283	4.73%	22,057,355	6.65%	6,544,072	1.92%	42.18%
Two or More Races	3,005,564	0.92%	27,588,006	8.31%	24,582,442	7.40%	817.90%

www.socialexplorer.com

andy@socialexplorer.com

Professor Emeritus of Sociology

Queens College and Graduate Center CUNY

72 Pondfield Road West, Apt 3A

Bronxville, NY 10708

Phone 914-337-6237 or Mobile 914-522-4487

Andrew A. Beveridge (he, him, his)

Co-Founder and President

Social Explorer, Inc.