The OCA Opportunity Index is formed by converting nine “indicators of opportunity” into the domains of Education, Employment and Neighborhood Factors and averaging these domains to give each Census Tract in Connecticut an “Opportunity Level.”

The processes described below and the data used are all contained in the OCA Opportunity Model Workbook and the results are mapped in the Opportunity Data Portal.

The first step is to bring in the data, convert it from municipal level to census tract level, if necessary, and arrange the data so that higher numbers indicate higher opportunity. 

• Median Income: This data is available by census tract from the American Community Survey (ACS), (https://data.census.gov/cedsci/). To increase the sample size, data from the five years spanning 2014 to 2018 are used. Higher median income indicates higher opportunity.
• School Performance: School performance data is available by school district from the CT State Department of Education, and is based on an index calculated by EdSight called Next Generation Accountability, (http://edsight.ct.gov/). Districts generally conform to town boundaries, however since there are regional school districts encompassing multiple towns, the regional school district scores are averaged with individual town districts to get the combined score, which is applied to all the census tracts in the town.  For example, the Score applied to all census tracts in Hebron is the total points achieved for Hebron School District Score and the Regional School District 8 over the total points possible for the Hebron School District Score and the Regional School District 8. In some cases, the Regional School District is the only public system that serves a town and in such cases, that score is used.
• Unemployment: This data is available by census tract from the (ACS) by census tract, (https://data.census.gov/cedsci/). To increase the sample size, data from the five years spanning 2014 to 2018 are used. Since lower unemployment indicates higher opportunity, the inverse of the unemployment rate is calculated for use in the index.
• Job Access: This data is available by census tract from the HUD Location Affordability Index using American Community Survey data from 2012-2016 and Longitudinal Employer Household Dynamics (LEHD) Data from 2014, (https://hudgis-hud.opendata.arcgis.com/datasets/b7ffe3607e8c4212bf7cf2428208dbb6_0?geometry=-159.821%2C-0.614%2C160.452%2C76.538). The variable used for Job Access is “Job Gravity” which represents the number of jobs accessible to each census tract, weighted by each job’s distance from that census tract.
• Retail Job Access: This data is available by census tract from the HUD Location Affordability Index using American Community Survey data from 2012-2016 and Longitudinal Employer Household Dynamics (LEHD) Data from 2014, (https://hudgis-hud.opendata.arcgis.com/datasets/b7ffe3607e8c4212bf7cf2428208dbb6_0?geometry=-159.821%2C-0.614%2C160.452%2C76.538). The variable used for Retail Job Access is “Retail Gravity” which represents the number of retail jobs accessible to each census tract, weighted by each job’s distance from that census tract.
• Job Growth: This data is available by town from the CT Department of Labor’s Quarterly Census of Earnings and Wages, a near census of employment and wage information, (https://www1.ctdol.state.ct.us/lmi/202/202_annualaverage.asp). To calculate job growth, the growth rate in Annual Average Employment (AAE) is calculated between 2016 and 2019 using the function (AAE2019 – AAE2016)/AAE2016. Since this is town-wide data, each census tract receives the value for town which contains it.
• Poverty Rate: This data is available by census tract from the (ACS) by census tract, (https://data.census.gov/cedsci/). To increase the sample size, data from the five years spanning 2014 to 2018 are used. Since lower poverty rate indicates higher opportunity, the inverse of the poverty rate is calculated for use in the index.
• Homeownership / Tenure: This data is available by census tract from the 2010 Decennial Census by census tract, (https://data.census.gov/cedsci/). To reduce error due to the ACS sample size estimation, only one ACS variable is used in each domain. Poverty Rate is the ACS variable in the neighborhood factors domain, so the Decennial Census is used for Tenure.
• Crime Rate: This data is available from the CT Department of Public Safety by town, (https://www.dpsdata.ct.gov/dps/ucr/ucr.aspx). The most recent year data is available, 2017, is used. The Crime Rate is calculated by adding the Violent Crime Offenses that occurred by town in 2017 and dividing by that town’s population. Violent Crime Offenses consist of murder, rape, attempted rape, robbery, and aggravated assault.

Each of the 9 variables are then converted into z-scores, a standardized measurement of the distance of each value from the mean of the set of values. Z-scores represent the distance of a census tract from the average for all census tracts for each variable.

The z-scores are then grouped into domains to equalize the impact of Educational, Employment and Neighborhood factors on the Opportunity Index. The Education Domain contains Median Income and School Performance. The Employment Domain contains Job Access, Retail Job Access, Unemployment and Job Growth. The Neighborhood Domain contains Poverty Rate, Crime Rate and Homeownership. The z-scores in each domain are averaged and the resulting domain scores are averaged to calculate the final opportunity index.

The final step is assigning an Opportunity Level to each census tract based on percentile. Very High opportunity tracts are those in within the 20% of tracts with the highest opportunity index. High opportunity tracts are those within the 60% to 80% range of highest opportunity index and so on.