To measure equity in America’s biggest metro areas, the ABC Owned Television Stations examined differences in outcomes across key quality of life indicators for people in the largest racial and ethnic groups.
We looked at differences in people’s experiences and outcomes in:
- Public education opportunities and barriers.
- Environmental risks they face where they live.
- Income and the opportunity to build wealth via homeownership.
- Healthy living.
- Policing.
The Equity Report shows the results in each of the 20 categories, but also provides a total number of categories where our analysis reveals inequities existing widely across a metro area. The report then shows in detail the differences in each of the categories for each metro area.
We studied metro areas rather than municipal city boundaries because that’s more consistent with how many people live their lives. We cross neighborhood, city, and county lines daily as we move between home, school, our jobs, and other places we need – or want – to go each day. The government tracks a lot of data this way for similar reasons.
For each measure, we looked at the gaps in outcomes between the largest racial and ethnic groups. In most measures, in almost every metro area, outcomes for white residents are better – and often significantly better – than those for Black people, Hispanic or Latino people, or other people of color.
Below, you’ll find a detailed explanation about how we analyzed each measure we chose. You will also find the original source of the publicly available raw data we used and links to the government web sites where the original data is published.
We acknowledge limitations and omissions in the source data and how we tried to mitigate them. When determining inequities, we adjusted the threshold for determining inequity by leaving some room for margins of error in the data sources or reporting mistakes made by police departments, schools and other government agencies who self-report some data.
We recognize there are many different, valid ways to measure these kinds of systemic differences in our communities and many alternate data sources. These are the ones we chose based on our experience, our research and in consultation with experts.
We welcome observations, concerns, and ideas from anyone who reviews this work. We want to hear from you about alternate data sources, different ways of examining these same data sources or any other feedback. You can reach us here.
We measured the number of school days missed because of suspensions. We calculated the number of days of school missed for every 100 white students, Black students, Latino students and a combination of all students of color, across all public schools located in each metro area.
We classified any difference of more than 25% in the rate of school days missed for white students compared to Black students, Latino students, or any students of color as inequity. We adjusted for that 25% difference to account for potential errors in the data that schools and districts self-report to the federal government.
We used data from the U.S. Department of Education’s Civil Rights Data Collection (online here), which collects its data directly from individual schools and districts. If a school's data is in error, it is most often because a school or district reported it incorrectly and has not corrected it even though the Department of Education has published the data online for years. The data accounts for the vast majority of public schools in the United States. The data is from the 2017-18 school year, the most recent year publicly released. We excluded alternative schools, juvenile-justice system schools and fully virtual schools.
We measured the share of school buildings in each metro area where data shows inequity in disciplinary actions exists, an indicator of how widespread the exposure to unequal discipline is for students and families in a metro area.
We classified a metro with more than 25 percent of schools having unequal discipline rates as broad inequity across the metro area.
We used data from the U.S. Department of Education’s Civil Rights Data Collection (online here), which collects its data directly from individual schools and districts. If a school's data is in error, it is most often because a school or district reported it incorrectly and has not corrected it even though the Department of Education has published the data online for years. The data accounts for the vast majority of public schools in the United States. The data is from the 2017-18 school year, the most recent year publicly released. We excluded alternative schools, juvenile-justice system schools and fully virtual schools.
We measured the share of students enrolled in advanced placement courses in high schools across the metro area, looking at differences in the rate of enrollment for white students compared to Black students, Latino students and all students of color.
We classified any difference of more than five percentage points between the percentage of Black students, Latino students, or students of color enrolled in advanced placement courses and the percentage of white students enrolled in all schools across the metro area as inequity.
We used data from the U.S. Department of Education’s Civil Rights Data Collection (online here), which collects its data directly from individual schools and districts. If a school's data is in error, it is most often because a school or district reported it incorrectly and has not corrected it even though the Department of Education has published the data online for years. The data accounts for the vast majority of public schools in the United States. The data is from the 2017-18 school year, the most recent year publicly released. We excluded alternative schools, juvenile-justice system schools and fully virtual schools.
We measured the share of white, Black, Latino and people of color across the metro area in households with a home computer and internet access.
We classified any difference of more than five percentage points between the percentage of white people with digital access and the percentage of Black people, Latino people, or people of color with digital access as inequity.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here). That is the most recent data released.
We measured the share of households that are white, Black, Latino and people of color and who owned their homes across the metro area.
We classified any difference of more than five percentage points in the rate of home ownership between white people and Black people, Latino people, and people of color as inequity. The gap was far wider in most cities. The difference was at least 40% in 21 of 100 metro areas we studied.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here). That is the most recent data released.
We measured the segregation of white people and all people of color across each metro area, using a well-known Dissimilarity Index employed by researchers for decades.
We calculated the dissimilarity index for the largest 100 metro areas in the U.S. The higher the index, the more segregated a city is.
It’s based on a 0-100 scale, with 0 indicating total integration of the metro area’s neighborhoods and 100 representing completely segregated neighborhoods.
We classified cities where the Dissimilarity Index was above 30 as segregated.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here)
We also measured the segregation of white people and people of color across each metro area’s approved loan applications, a measure of the potential progress toward desegregation of a city’s neighborhoods, using the same Dissimilarity Index.
We calculated the dissimilarity index for the home loans across largest 100 metro areas in the U.S. The higher the index, the more segregated the city’s most recent homebuyers.
It’s based on a 0-100 scale, with 0 indicating total integration of the metro area’s home purchases and 100 representing completely segregated home purchases. The index only includes homes purchased via loans and not cash purchases.
We calculated the index for approved loans for the largest 100 metro areas in the U.S.
We classified recent home-lending in cities where this home-buying dissimilarity index was above 30 as segregated.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here), and from the Home Mortgage Disclosure Act, specifically 2019 loan applications (online here). That is the most recent data released.
We measured the share of home-loan applications approved for white people, Black people, Latino people and people of color.
We classified any difference of more than five percentage points in the rate of home-loan approvals for white borrowers and applicants of any other racial or ethnic group as inequity.
We used data from the Home Mortgage Disclosure Act, specifically from a combination of the 2018 and 2019 home loan applications (online here). That is the most recent data released by the U.S. government.
We measured the median household income of white people, Black people, Latino people, and people of color across each metro area.
We classified any difference greater than 5% in median income between white households and households of any other racial or ethnic background as inequity. The gap is more than 30% in 93 of 100 metro areas.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here). That is the most recent data released.
We measured the share of white, Black, Latino and nonwhite people in a metro area with health insurance coverage.
We classified any difference greater than five percentage points as inequity.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here). That is the most recent data released.
We measured the life expectancy in thousands of neighborhoods in each metro area and the racial and ethnic makeup of those neighborhoods. We calculated the percentage of white, Black, Latino and people of color living in neighborhoods where the life expectancy was at least 10% lower than the norm statewide. We treated Census tracts as neighborhoods.
We classified any difference greater than five percentage points as inequity.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here), to determine the racial makeup of neighborhoods. We used the Centers for Disease Control’s life expectancy data by tract (online here). In both cases, we used the most recent data publicly released.
We measured the share of white, Black, Latino and nonwhite households in a metro area living in a neighborhood the U.S. Department of Agriculture has deemed a food desert. The federal government defines a food desert as a neighborhood without a major food store within one half-mile of where a person lives in urbanized areas or within 10 miles in rural areas.
We classified any difference greater than five percentage points of people living in a food desert as inequity.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here), and from the U.S. Department of Agriculture’s latest food availability atlas (online here). Both are the most recent data available.
We measured the rate at which white, Black, Latino and nonwhite people in a metro area arrested for any crime in 2019, the most recent, consistent data available nationwide.
We measured the number of arrests per 100 white, Black and Latino people as well as a combination of all people of color.
We aggregated the arrests for all reporting police departments across a metro area, in each of the largest 100 metro areas in the United States. We used the population of the metro area to account for typical day-to-day movements of people across police jurisdictions in a region in order to commute to work, school and other personal and leisure activities.
We classified any difference in the rate of being arrested as inequity, allowing for a 25% margin to account for potential errors in data reported by police departments to the public and the FBI.
We used population data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here), and arrest data from the FBI’s Uniform Crime Reports system (online here), which is the most comprehensive available and is reported directly to the FBI by the police agencies themselves every year.
Where large departments at the core of a metro area fail to report arrest data to the FBI (examples: Chicago, New York City, Washington, D.C., Atlanta and all agencies in Florida), we used comparable data from agencies or states. For the Chicago metro area, because Illinois agencies do report no crime or arrest data to the FBI, the analysis includes only data published by the Chicago Police Department. For Atlanta, we used the 12 most recent months the police department released, which cover months in 2018 and 2019.
We measured the rate at which white, Black, Latino and nonwhite people in a metro area arrested for drug-related offenses in 2019, in an attempt to take a closer look at arrests for a nonviolent crime where researchers have repeatedly found disproportionate enforcement, particularly since government surveys show illegal drug use is similar across racial groups.
We measured the number of drug arrests per 100 white, Black and Latino people as well as a combination of all people of color.
We aggregated the arrests for all reporting police departments across a metro area, in each of the largest 100 metro areas in the United States. We used the population of the metro area to account for typical day-to-day movements of people across police jurisdictions in a region in order to commute to work, school and other personal and leisure activities.
We classified any difference in the rate of being arrested as inequity, allowing for a 25% margin to account for potential errors in data reported by police departments to the public and the FBI.
We used population data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here), and drug-related arrest data from the FBI’s Uniform Crime Reports system (online here), which is the most comprehensive available and is reported directly to the FBI by the police agencies themselves every year.
Where large departments at the core of a metro area fail to report arrest data to the FBI (examples: Chicago, New York City, Washington, D.C., Atlanta and all agencies in Florida), we used comparable data from agencies or states. For the Chicago metro area, because Illinois agencies do report no crime or arrest data to the FBI, the analysis includes only data published by the Chicago Police Department. For Atlanta, we used the 12 most recent months the police department released, which cover months in 2018 and 2019.
We measured the share of police officers who were white, Black, Latino, Asian or who identified as any racial or ethnic background other than white and compared the diversity of officers to the diversity of people living in the metro area they serve.
We classified any difference of more than 5 percent between the racial makeup of officers and any racial or ethnic group across the metro area as inequity.
We used data from the U.S. Census Bureau’s American Community Survey, specifically from the 2019 five-year estimates (online here) and the Equal Employment Opportunity tabulation released in early 2021, which represented estimates from 2018 (online here). The Census Bureau provides occupation data only for counties with at least 50,000 residents. For each metro region, we were only able to include counties with at least 50,000 people.
We do not use staffing data from individual departments because it is often incomplete or inconsistently kept, or sometimes not kept at all. We looked at data across metro areas because that is how most people encounter police as they move between home, work, schools, and other locations throughout the day, often commuting across many police jurisdictions. Few people encounter police officers from a single jurisdiction only in the city where they live.
We measured the percentage of Black, Latino, white and people of color who reside in neighborhoods that the Environmental Protection Agency determined are in the 80th percentile for cancer risk from toxic pollutants released into the air. We treated Census Block Groups as neighborhoods.
We classified any difference of more than five percentage points in people living in those highest-risk neighborhoods as inequity.
We used data from the U.S. Environmental Protection Agency’s Environmental Justice screening system (online here), which examines exposures and risks from pollutants, toxins, and other environmental dangers by pulling from the agency’s list of regulated sites and looking at releases and dangers in the context of the socioeconomic makeup of neighborhoods. The latest calculations were released in 2020. They rely upon data collected by EPA and states over five previous years.
We measured the percentage of Black, Latino, white and people of color who reside in neighborhoods that the Environmental Protection Agency has determined are in the 80th percentile for lifetime respiratory dangers from chemicals released into the air. We treated Census Block Groups as neighborhoods. The 80th percentile and above neighborhoods are those that the government’s scientists identified as the 20% statewide with the highest relative risk.
We classified any difference of more than five percentage points in people living in those highest-risk neighborhoods as inequity.
We used data from the U.S. Environmental Protection Agency’s Environmental Justice screening system (online here), which examines exposures and risks from pollutants, toxins, and other environmental dangers by pulling from the agency’s list of regulated sites and looking at releases and dangers in the context of the socioeconomic makeup of neighborhoods. The latest calculations were released in 2020. They rely upon data collected by EPA and states over five previous years.
We measured the percentage of Black, Latino, white and people of color across who reside in a neighborhood that the Environmental Protection Agency has determined are in the 80th percentile for the share of homes built before 1960. The reason the government looks at that measure is because it’s indicative of buildings more likely to have lead-based paint hazards. We treated Census Block Groups as neighborhoods. The 80th percentile and above neighborhoods are those that the government’s scientists identified as the 20% statewide with the highest relative risk.
We classified any difference of more than five percentage points in people living in those highest-risk neighborhoods as inequity.
We used data from the U.S. Environmental Protection Agency’s Environmental Justice screening system (online here), which examines exposures and risks from pollutants, toxins, and other environmental dangers by pulling from the agency’s list of regulated sites and looking at releases and dangers in the context of the socioeconomic makeup of neighborhoods. The latest calculations were released in 2020. They rely upon data collected by EPA and states over five previous years.
We measured the percentage of Black, Latino, white and people of color who reside in neighborhoods that the Environmental Protection Agency determined are in the 80th percentile for being near facilities so potentially hazardous that they are required to file Risk Management Plans with government regulators documenting the dangers to the public. We treated Census Block Groups as neighborhoods. The 80th percentile and above neighborhoods are those that the government’s scientists identified as the 20% statewide with the highest relative risk.
We classified any difference of more than five percentage points in people living in those highest-risk neighborhoods as inequity.
We used data from the U.S. Environmental Protection Agency’s Environmental Justice screening system (online here), which examines exposures and risks from pollutants, toxins, and other environmental dangers by pulling from the agency’s list of regulated sites and looking at releases and dangers in the context of the socioeconomic makeup of neighborhoods. The latest calculations were released in 2020. They rely upon data collected by EPA and states over five previous years.
We measured the percentage of Black, Latino, white and people of color who reside in neighborhoods that the Environmental Protection Agency determined are in the 80th percentile for the amount of toxic pollutants discharged into streams and waterways, with an emphasis on pollutants upstream from neighborhoods. We treated Census Block Groups as neighborhoods. The 80th percentile and above neighborhoods are those that the government’s scientists identified as the 20% statewide with the highest relative risk.
We classified any difference of more than five percentage points in people living in those highest-risk neighborhoods as inequity.
We used data from the U.S. Environmental Protection Agency’s Environmental Justice screening system (online here), which examines exposures and risks from pollutants, toxins, and other environmental dangers by pulling from the agency’s list of regulated sites and looking at releases and dangers in the context of the socioeconomic makeup of neighborhoods. The latest calculations were released in 2020. They rely upon data collected by EPA and states over five previous years.
