2. Methodology

This section provides an overview of the data sources and analysis for the OC Census Atlas. This is just one approach to analyzing the wealth of data available from the 2020 Census and we invite others to build on this framework. For further details about the methodology, contact the Good Work Collaborative at info@ourgoodwork.co.

Self-Response Rates (SRR)

Self-Response Rates (SRR) reflect the percentage of all known housing units in a given geography (such as a county, city, census tract) that self-responded to the census by completing the questionnaire online, by phone, or by mail between March and October 2020. The SRR in the OC Census Atlas reflects the U.S. Census Bureau’s updates on SRR on January 29, 2021.

SRR does not reflect the overall share of people counted in the census because some populations were counted through special enumeration methods. People living in group quarters (such as prisons, student housing, or assisted living facilities) were counted via Group Quarters enumeration, and unhoused people were counted via Service Based Enumeration and Targeted Non-Sheltered Outdoor Locations operations, both facilitated by the U.S. Census Bureau. Households enumerated in the Census Bureau’s Nonresponse Followup (NRFU) operations are also not reflected in SRR.

Throughout the Census Atlas, there are graphs that demonstrate weekly SRR progress as well as change in SRR rates. Upticks in SRR in mid-August are notable because they are the likely result of partner outreach efforts while the NRFU operation was taking place. Some jurisdictions also demonstrated an uptick on the last day of the self-response period (October 15), which reflects final partner outreach efforts as well as the U.S. Census Bureau’s final processing of mailed questionnaire responses and other data cleanup before their January 2021 update.

Demographic Data Sources

Demographic, housing and socioeconomic information at the county, city and census tract level reflect 2015-2019 American Community Survey (ACS) 5-year estimates and not the population count from the 2020 Census. This is to reflect the information that was publicly available to outreach partners at the time of the decennial census campaign.

Population and race/ethnic estimates will differ from data in the PL94-171 redistricting data file released by the Census Bureau in August 2021, although many numbers are similar.

Race/Ethnic Data

The U.S. Census Bureau collects race/ethnic data using five mutually exclusive categories in accordance with the 1997 Office of Management and Budget (OMB) standards: American Indian and Alaska Native, Asian, Black or African American, Hispanic or Latino (of any race), Native Hawaiian and Other Pacific Islander alone, and White. Data is based on how people self-identify in their responses, and people were only able to select multiple Race/Ethnic categories starting with the 2000 Census.

The OC Census Atlas disaggregates these race categories by cross-referencing demographic information from the 2015-2019 American Community Survey (ACS) 5-year estimates referenced in the section above. There are 7 ethnic categories for the Asian category (Asian Indian, Chinese, Filipino, Japanese, Korean, Vietnamese, “Other Asian”) and 4 ethnic categories for the Hispanic or Latino (of any race) category (Mexican, Puerto Rican, Cuban, “Other Hispanic/Latino”).

Note that while the main race categories are mutually exclusive, the disaggregated Asian categories include people of Hispanic or Latino origin. For example, the “ASIAN” data category does not include Asians of Hispanic or Latino origin, but the “Korean” category does. In other words, the data presented in the “Korean” category reflects Koreans of both non-Hispanic or Latino and Hispanic or Latino origin.

There was no public data available to allow a disaggregated analysis of the Black or African American population. Given the methodology behind the calculations (see: SRR Analysis section), the populations of American Indian and Alaska Native and Native Hawaiian and Other Pacific Islander alone are too small to be disaggregated and still present accurate conclusions.

For simplicity and ease of presentation, the OC Census Atlas refers to each of the groups as the following: American Indian / Alaska Native, Asian, Black/African American, Hispanic/Latino of Any Race, Native Hawaiian / Pacific Islander, and White.

SRR Analysis

The Census Atlas provides analysis on SRR for various characteristics such as race/ethnic groups, renter-occupied households, and children under 5. These numbers are the calculated self-response rate for all census tracts where the characteristic’s estimated percentage is greater than or equal to the estimated percentage for the total geography. The number does not represent the actual SRR of a characteristic, such as the Hispanic/Latino population.

This analysis is conducted for approximately 600 Orange County census tracts that have a published 2020 Census SRR. A few non-residential tracts (i.e. fully industrial or commercial areas) are excluded as they contain 0 households. The Census Atlas provides analysis on the demographic characteristics of the bottom 20% lowest responding census tracts (SRR below 71%) as well as the top 20% highest responding census tracts (SRR above 85%) in the county.

Because the U.S. Census Bureau did not release information about specific households that self-responded to the 2020 Census, this analysis presents the overall characteristics of the census tracts based on 2015-2019 American Community Survey (ACS) 5-year estimates referenced in the section above. Tract-level estimates are aggregated for each characteristic and divided by the sum of the applicable denominator (i.e. total estimated population or total number of households). For example, the estimated number of renter-occupied households in the lowest responding tracts is summed and divided by the total number of occupied housing units in the same tracts.

The U.S. Census Bureau released SRR data at several levels of geography, including planned 2020 tract geography. The boundaries of these census tracts are different from those defined in 2010 that are used for ACS estimates up to 2019. Subsequently, the Census Atlas utilizes 2010 tract geography in order to incorporate 2019 ACS data into the analysis. This is accomplished by applying the Census Bureau’s file that outlines the conversion of 2020 SRR data into 2010 tract geography.

Note that not all of Orange County’s 34 cities cover one or more of the highest and/or lowest-responding census tracts. For example, all census tracts that overlap the City of Villa Park had a SRR above 71% (the bottom 20% threshold) and below 85% (the top 20% threshold). The analysis also generally excluded tracts that cover only a very small area of a city.

Assigning Census Tracts to Cities

Census tracts are the building blocks of census geography. Tracts are contained within counties, but they can overlap the boundaries of cities (also known as incorporated places), zip codes, legislative districts, and other geographies. Some cities are so small they are contained within one tract.

The OC Census Atlas provides analysis at the county level as well as tract-level estimates for all 34 cities in Orange County based on the approximate city-tract overlap. For example, if 100% of the tract is within a particular city’s boundaries, the tract’s entire population and other characteristics were part of the analysis for that city. If 50% of the census tract’s geography falls within the city’s boundaries, one-half of the tract’s population and other characteristics were part of the analysis for that city. The same methodology applies to the other calculated characteristics such as total number of households.

Unincorporated areas are excluded from the analysis due to the complexity of accurately aggregating data for multiple proportions of census tracts. Unlike cities, the census tracts of unincorporated areas do not follow the boundaries as closely because they are separate from official municipalities.