CDC SVI Documentation 2020
View print only PDF of CDC/ATSDR SVI 2020 Documentation [PDF – 671 KB]
CDC/ATSDR SVI 2020 Documentation – 8/5/2022
What is Social Vulnerability?
Every community must prepare for and respond to hazardous events, whether a natural disaster like a tornado or a disease outbreak, or an anthropogenic event such as a harmful chemical spill. The degree to which a community exhibits certain social conditions, including high poverty, low percentage of vehicle access, or crowded households, may affect that community’s ability to prevent human suffering and financial loss in the event of disaster. These factors describe a community’s social vulnerability.
What is CDC/ATSDR Social Vulnerability Index?
ATSDR’s Geospatial Research, Analysis, & Services Program (GRASP) created the Centers for Disease Control and Prevention and Agency for Toxic Substances and Disease Registry Social Vulnerability Index (CDC/ATSDR SVI or simply SVI, hereafter) to help public health officials and emergency response planners identify and map the communities that will most likely need support before, during, and after a hazardous event.
SVI indicates the relative vulnerability of every U.S. Census tract. Census tracts are subdivisions of counties for which the Census collects statistical data. SVI ranks the tracts on 16 social factors, including unemployment, racial and ethnic minority status, and disability, and further groups them into four related themes. Thus, each tract receives a ranking for each Census variable and for each of the four themes as well as an overall ranking.
In addition to tract-level rankings, SVI 2010, 2014, 2016, 2018, and 2020 also have corresponding rankings at the county level.
Notes below that describe “tract” methods also refer to county methods.
How can SVI help communities be better prepared for hazardous events?
SVI provides specific socially and spatially relevant information to help public health officials and local planners better prepare communities to respond to emergency events such as severe weather, floods, disease outbreaks, or chemical exposure.
SVI can be used to:
- Assess community need during emergency preparedness planning
- Estimate the type and amount of needed supplies such as food, water, medicine, and bedding.
- Decide how many emergency personnel are required to assist people.
- Identify areas in need of emergency shelters.
- Create a plan to evacuate people, accounting for those who have special needs, such as those without vehicles, the elderly, or people who do not speak English well.
- Identify communities that will need continued support to recover following an emergency or natural disaster.
Important Notes on SVI Databases
- SVI 2014, 2016, 2018, and 2020 are available for download in shapefile format from https://www.atsdr.cdc.gov/placeandhealth/svi/data_documentation_download.html. SVI 2014, 2016, 2018, and 2020 are also available via ArcGIS Online. Search for “CDC’s Social Vulnerability Index.”
- For SVI 2000 and 2010, keep the data in geodatabase format when downloading from https://www.atsdr.cdc.gov/placeandhealth/svi/data_documentation_download.html. Converting to shapefile changes the field names.
- ACS field names changed between SVI 2018 and 2020. Name changes are noted in the Data Dictionary below.
- For US-wide or multi-state mapping and analysis, use the US database, in which all tracts are ranked against one another. For individual state mapping and analysis, use the state-specific database, in which tracts are ranked only against other tracts in the specified state.
- Starting with SVI 2014, we’ve added a stand-alone, state-specific Commonwealth of Puerto Rico database. Puerto Rico is not included in the US-wide ranking.
- Starting with SVI 2014, we’ve added a database of Tribal Census Tracts (https://www.census.gov/newsroom/blogs/random-samplings/2012/07/decoding-state-county-census-tracts-versus-tribal-census-tracts.html). Tribal tracts are defined independently of, and in addition to, standard county-based tracts. The tribal tract database contains only estimates, percentages, and their respective margins of error (MOEs), along with the adjunct variables described in the data dictionary below. Because of geographic separation and cultural diversity, tribal tracts are not ranked against each other nor against standard census tracts.
- Tracts with zero estimates for total population (N = 645 for the U.S.) were removed during the ranking process. These tracts were added back to the SVI databases after ranking. The TOTPOP field value is 0, but the percentile ranking fields (RPL_THEME1, RPL_THEME2, RPL_THEME3, RPL_THEME4, and RPL_THEMES) were set to -999.
- For tracts with > 0 TOTPOP, a value of -999 in any field either means the value was unavailable from the original census data or we could not calculate a derived value because of unavailable census data.
- Any cells with a -999 were not used for further calculations. For example, total flags do not include fields with a -999 value.
- Whenever available, we use Census-calculated MOEs. If Census MOEs are unavailable, for instance when aggregating variables within a table, we use approximation formulas provided by the Census in Appendix A (pages A-14 through A-17) of A Compass for Understanding and Using American Community Survey Data here: https://www.census.gov/content/dam/Census/library/publications/2008/acs/ACSGeneralHandbook.pdf
If more precise MOEs are required, see Census methods and data regarding Variance Replicate Tables here: https://www.census.gov/programs-surveys/acs/data/variance-tables.html. For selected ACS 5-year Detailed Tables, “Users can calculate margins of error for aggregated data by using the variance replicates. Unlike available approximation formulas, this method results in an exact margin of error by using the covariance term.”
- FIPS codes are generally defined as text to preserve leading zeros (0s). While working with csv files, leading 0s are required to properly join or merge tables. ArcGIS maintains leading 0s in the FIPS code fields of csv files. To preserve leading 0s and create an Excel file in Excel for Office 365, follow these steps:
- Open a blank worksheet in Excel.
- Click Data in the menu bar and choose the icon From Text/CSV
- Navigate to the csv file and choose to Import
- In the dialog box that opens, choose to Transform Data
- In the Power Query Editor dialog box, for each of the FIPS columns (ST, STCNTY, FIPS for tracts and ST, FIPS for counties), right click the column name and choose to Change Type to Text.
- As prompted in the Change Column Type dialog box, choose to Replace current. Click Close and Load.
- Save As an Excel xlsx file.
- See the Methods section below for further details.
- Questions? Please visit the SVI website for additional information or email the SVI Coordinator at firstname.lastname@example.org.
American Community Survey (ACS), 2016-2020 (5-year) data for the following estimates:
Text version of overall vulnerability image:
- Socioeconomic Status
- Below 150% Poverty
- Housing Cost Burden
- No High School Diploma
- No Health Insurance
- Household Characteristics
- Aged 65 & Older
- Aged 17 & Younger
- Civilian with a Disability
- Single-Parent Households
- English Language Proficiency
- Racial & Ethnic Minority Status
- Hispanic or Latino (of any race); Black and African American, Not Hispanic or Latino; American Indian and Alaska Native, Not Hispanic or Latino; Asian, Not Hispanic or Latino; Native Hawaiian and Other Pacific Islander, Not Hispanic or Latino; Two or More Races, Not Hispanic or Latino; Other Races, Not Hispanic or Latino
- Housing Type & Transportation
- Multi-Unit Structures
- Mobile Homes
- No Vehicle
- Group Quarters
For SVI 2020, adjunct variables were included:
- An estimate of daytime population derived from LandScan 2020 estimates
- 2016-2020 ACS estimates for households without a computer with a broadband Internet subscription
- 2016-2020 ACS estimates for Hispanic/Latino persons, Not Hispanic or Latino Black/African American persons, Not Hispanic or Latino Asian persons, Not Hispanic or Latino American Indian and Alaska Native persons, Not Hispanic or Latino Native Hawaiian and Other Pacific Islander persons, Not Hispanic or Latino persons of two or more races, and Not Hispanic or Latino persons of some other race
These adjunct variables are excluded from SVI rankings. We include these variables as adjunct variables because they can be helpful to explain more about the local areas in certain circumstances, and we want to make them easily accessible.
Raw data estimates and percentages for each variable, for each tract, are included in the database. In addition, the margins of error (MOEs) for each estimate, at the Census Bureau standard of 90%, are also included. Confidence intervals can be calculated by subtracting the MOE from the estimate (lower limit) and adding the MOE to the estimate (upper limit). Because of relatively small sample sizes, some of the MOEs are high. It is important to identify the amount of error acceptable in any analysis.
We ranked Census tracts within each state and the District of Columbia, to enable mapping and analysis of relative vulnerability in individual states. We also ranked tracts for the entire United States against one another, for mapping and analysis of relative vulnerability in multiple states, or across the U.S. as a whole. Tract rankings are based on percentiles. Percentile ranking values range from 0 to 1, with higher values indicating greater vulnerability.
For each tract, we generated its percentile rank among all tracts for 1) the 16 individual variables, 2) the four themes, and 3) its overall position.
Theme rankings: For each of the four themes, we summed the percentiles for the variables comprising each theme. We ordered the summed percentiles for each theme to determine theme-specific percentile rankings.
The four summary theme ranking variables, detailed in the Data Dictionary below, are:
- Socioeconomic Status – RPL_THEME1
- Household Characteristics – RPL_THEME2
- Racial & Ethnic Minority Status – RPL_THEME3
- Housing Type & Transportation – RPL_THEME4
Overall tract rankings: We summed the sums for each theme, ordered the tracts, and then calculated overall percentile rankings. Please note taking the sum of the sums for each theme is the same as summing individual variable rankings. The overall summary ranking variable is RPL_THEMES.
Tracts in the top 10%, i.e., at the 90th percentile of values, are given a flag value of 1 to indicate high vulnerability. Tracts below the 90th percentile are given a flag value of 0.
For a theme, the flag value is the number of flags for variables comprising the theme. We calculated the overall flag value for each tract as the number of all variable flags.
For a detailed description of SVI variable selection rationale and methods, see A Social Vulnerability Index for Disaster Management (https://www.atsdr.cdc.gov/placeandhealth/svi/img/pdf/Flanagan_2011_SVIforDisasterManagement-508.pdf).
Caveat for SVI State Databases
The order of overall SVI rankings and SVI theme rankings of census tracts and counties may differ between the U.S. and state SVI databases. A detailed explanation follows.
Overall and theme rankings are based on cumulative values that are relative to the number of census tracts or counties being compared. Thus, differences between the order of overall and theme rankings in the U.S. database and that of state databases may arise from the accumulation of differences in summing the percentile ranks for the individual SVI variables.
For example, using the 2018 Georgia SVI database, Fulton County has an overall SVI score of 0.2658 with a ranking of 117 out of 159 Georgia counties. However, using the 2018 U.S. SVI database, Fulton County has an overall SVI score of 0.5268, giving Fulton County a ranking of 125 out of the 159 Georgia counties. The ranking differences between the two databases are due to differences in summed percentile ranks caused, in turn, by differences in the number of counties being compared in the U.S. database versus Georgia database.
In short, because Georgia (or any state) has far fewer census tracts and counties than does the nation, differences in one or more variable percentages from one census tract or county to another are more pronounced at the state level than at the national level. Such differences, when summed across all variables, will in some cases result in a rank order change between the two databases.
If there are any questions, please contact the SVI Coordinator at email@example.com.
SVI 2020 Updates
As our understanding of social vulnerability evolves over time, SVI must evolve as well. Beginning with SVI 2020, we made modifications to SVI theme names, individual SVI indicators, and adjunct data. We modified the name of Theme 2 from Household Composition & Disability to Household Characteristics, and we modified the name of Theme 3 from Minority Status & Language to Racial & Ethnic Minority Status. Within Theme 1 Socioeconomic Status, we modified the Below Poverty variable from the 100% federal poverty level to the 150% federal poverty level, considering the federal poverty line thresholds established for several federal health coverage policies.1 Similarly, we included a No Health Insurance variable in Theme 1 Socioeconomic Status as a lack of health insurance coverage is increasingly considered a marker of lower socioeconomic status and a barrier to healthcare access.2 Also, within Theme 1 Socioeconomic Status, we exchanged the Per Capita Income variable for Housing Cost Burden, which are households that spend 30% or more of annual income on housing costs. Recent studies have emphasized the importance of examining housing cost burden as opposed to per capita income as a better indicator of insufficient disposable income among households.3,4 Further, we moved the English Language Proficiency variable from Theme 3 Racial & Ethnic Minority Status to Theme 2 Household Characteristics because the ACS variables are based on language spoken at home and are better suited in the Household Characteristics theme. Additionally, although people in racial and ethnic minority groups are overall more likely to have limited English language proficiency than non-Hispanic whites, most (90.9%) are English language proficient.5 Thus, we moved the English Language Proficiency out of the Minority theme because it may have adversely affected the vulnerability ranking of communities in high minority areas of the country. Lastly, we included new adjunct variables: households without a computer with a broadband Internet subscription, and breakdowns of racial and ethnic minority populations. The coronavirus disease 2019 pandemic has underscored the importance of broadband Internet access as a social determinant of health, justifying the inclusion of data on the lack of broadband Internet access as an adjunct variable.6 While we aggregate all racial and ethnic minority persons in Theme 3 Racial & Ethnic Minority Status, we recognize that SVI users may be interested in its component populations. A thorough literature review and internal validation were conducted to finalize the construction of SVI 2020.
- McMaughan DJ, Oloruntoba O, Smith ML. Socioeconomic status and access to healthcare: Interrelated drivers for healthy aging. Front Public Health. 2020;8:231. doi:10.3389/fpubh.2020.00231
- Hernández D, Swope CB. Housing as a platform for health and equity: Evidence and future directions. Am J Public Health. 2019;109(10):1363-1366. doi:10.2105/AJPH.2019.305210
- Swope CB, Hernández D. Housing as a determinant of health equity: A conceptual model. Soc Sci Med. 2019;243:112571. doi:10.1016/j.socscimed.2019.112571
- U.S. Census Bureau; American Community Survey (ACS), Five-Year Public Use Microdata Sample (PUMS), 2016-2020; accessed via MDAT; ; (27 July 2022).
- Benda NC, Veinot TC, Sieck CJ, Ancker JS. Broadband Internet Access Is a Social Determinant of Health! Am J Public Health. 2020;110(8):1123-1125. doi:10.2105/AJPH.2020.305784