Access Denied: How we analyzed the data

By Libby Seline

For the first story in the “Access Denied” investigative series, the San Antonio Express-News analyzed historical census data between 1970 and 2020 that covered three topics: education atainment, Hispanic ethnicity and median family income. The data is archived by IPUMS National Historical Geographic Information System. 

Additionally, we analyzed data from the City of San Antonio’s COVID-19 dashboard and redlining data from the University of Richmond. 

The education data includes people 25 years old or older into three categories: those who have not attended a form of higher education, people who attended some college and those who earned a bachelor’s degree or above. Those who attended some college had been to college for at least a year but had not received a bachelor’s degree. In 1970 and 1980, the “bachelor’s degree or above category” included those who had attended college for four or more years. 

The median family income for each decennial census is based on the amount of money a family made the previous year. So, the median family income for 1980 is based on the amount of money a family made in 1979. Data from 2012 and 2019 use the past five years worth of data and count the median family income over the past 12-months. That information is presented in inflation-adjusted dollars based on the last year of the period. 

Information from the census was presented by an area called a “census tract,” which represents areas throughout Bexar County. To analyze tracts in the South Side, the Express-News looked at tracts south of US-90 and US-87. 

The city’s COVID-19 vaccination and death data — broken down by ZIP code — from October 2020 was interjoined to the University of Richmond’s redlining data in a mapping program called QGIS. By combining the data, the Express-News determined the number of deaths in the redlining areas. 

For reporting and comparison purposes, the Express-News used the data points listed below to create an index measurement that would examine healthcare access and need. Each factor was assigned a weight, which determined the importance of the factor in determining healthcare access and need. The weights assigned add up to 100. 

• Hospital counts (Source: Southwest Texas Regional Advisory Council) 
  Weight 10/100
• Diabetes prevalence rates in 2019 (Source: CDC PLACES data)
  Weight 25/100
• Obesity prevalence rates in 2019 (Source: CDC PLACES data)
   Weight 25/100
• Percentage of adults who received routine checkups in 2019 (Source: CDC PLACES data)
  Weight 25/100
• Percentage of residents 25 and older whose highest level of education was high school/GED or below in 2019 (Source: Census ACS 5-year 2015-2019)
  Weight 5/100
• Median family income in 2019 (Source: Census ACS 5-year 2015-2019)
  Weight 10/100

Using this ranking, we were able to compare areas with low healthcare access and use with areas that had higher COVID case and death rates.

Methodology: Everything

The San Antonio Express-News analyzed historical census data between 1970 and 2020 that covered four topics: education attainment, race, Hispanic ethnicity and median family income. The data is archived by IPUMS National Historical Geographic Information System. Using this data, we created categories that we could look at over time. 

For education, we separated people 25-years-old or older into three categories: those who have not attended a form of higher education, people who attended some college and those who earned a bachelor’s degree or above.

Those who attended some college had been to college for at least a year but had not received a bachelor’s degree. In 1970 and 1980, the “bachelor’s degree or above category” included those who had attended college for four or more years. 

The U.S. Census changed how it examined race between 1970 and 2020. Before 1970, the census-takers would determine a person’s race, but starting with the 1970 survey, people were allowed to choose their own race. In 2000, the government added a “two or more races” category, allowing people to mark down multiple races. That same year, the government asked people to label whether or not they had a Hispanic or Latinx origin, which allowed the government to designate a person’s race by their Hispanic origin. So, people who identified as white could be labeled as white with a Hispanic/Latinx origin or white without a Hispanic/Latinx origin. Prior to that year, they had people mark whether they had a Hispanic/Latinx origin but a category of people in a race who were not Hispanic or Latinx was not available. 

The political climate often affected the information a respondent would provide the government, including information about their race or ethnicity. Additionally, funding and initiatives to count every person in the U.S. affects the number of people of color counted in an area. A study by the Urban Institute said Black and Hispanic/Latinx were undercounted in the 2020 U.S. Census and have been undercounted in previous U.S. Censuses. 

To account for as many people of color as possible, the Express-News created a category called “people of color” by subtracting the area’s white population from an area’s total population. 

The median family income for each decennial census is based on the amount of money a family made the previous year. So, the median family income for 1980 is based on the amount of money a family made in 1979. Data from 2012 and 2019, use the past five years worth of data and count the median family income over the past 12-months. That information is presented in inflation-adjusted dollars based on the last year of the period. 

Information from the census was presented by an area called a “census tract,” which represents areas throughout Bexar County. To find tracts in the Southside, the Express-News looked at tracts south of US-90 and US-87. 

For reporting and comparison purposes, the Express-News used the data points listed below to create an index measurement that would examine healthcare access and need. Each factor was assigned a weight, which determined the importance of the factor in determining healthcare access and need. The weights assigned add up to 100. 

Hospital counts (Source: Southwest Texas Regional Advisory Council) 
Weight 10/100
Diabetes prevalence rates in 2019 (Source: CDC PLACES data)
Weight 25/100
Obesity prevalence rates in 2019 (Source: CDC PLACES data)
 Weight 25/100
Percentage of adults who received routine checkups in 2019 (Source: CDC PLACES data)
Weight 25/100
Percentage of residents 25 and older whose highest level of education was high school/GED or below in 2019 (Source: Census ACS 5-year 2015-2019)
Weight 5/100
Median family income in 2019 (Source: Census ACS 5-year 2015-2019)
Weight 10/100

Using this ranking, we were able to compare areas with low healthcare access and use with areas that had higher COVID case and death rates.

Healthcare locations were provided by the Southwest Texas Regional Advisory Council.

[This article was originally published by San Antonio Express-News.]

Did you like this story? Your support means a lot! Your tax-deductible donation will advance our mission of supporting journalism as a catalyst for change.