The Unequal Landscape of Civic Opportunity in America

Authors: Milan de Vries, Jae Yeon Kim, and Hahrie Han (2023)

Session information

• R version 4.2.2 (2022-10-31)
• Platform: aarch64-apple-darwin20 (64-bit)
• Running under: macOS Ventura 13.4.1

Data

Raw data

• We collected the IRS tax returns and related website data (i.e., their "About" pages) using our own R package, called "MapAgora" (ver 0.08), developed by de Vries and Kim.

• We classified the organizational types based on the above data using our own R package, called "autotextclassifier" (ver 0.05), also developed by de Vries and Kim.

Other data sources

• Kyne and Aldrich (2020)

• Chetty, et al (2023)

• Pierri, et al. (2022)

• Rupasingha, et al (2006, with updates) (Penn State Index)

• Schlozman et al (2014) (Washington Representative Study)

Replication data

All replication data are available at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/TCXRTM

• mma_sc_demo.csv (3127 * 8)

  • FIPS (character): County FIPS code
  • opc (numeric): Civic opportunity scores per capita
  • socialcap (numeric): Social capital measure from Kyne & Aldrich
  • sk2014 (numeric): Social capital measure from Rupasingha, et al.
  • civicorganization_county (numeric): Civic organization measure from Chetty, et al.
  • race_per_white_nonhispaic (numeric): The proportion of non-white Hispanic populations in counties
  • per_poverty (numeric): The proportion of people below the federal poverty line in counties
  • college_educ (numeric): The proportion of college educated people in counties

• unit_cor_df.csv (4 * 5)

  • term (character): Civic opportunity dimensions (i.e., "Taking action," "Volunteering," "Membership," "Holding events")
  • Binary index (numeric): Binary index civic opportunity scores
  • Mean index (numeric): Mean index civic opportunity scores
  • Inverse covariance matrix index (numeric): Inverse covariance matrix index civic opportunity scores
  • Principal component first factor index (numeric): PCA first factor index civic opportunity scores

• map_trade_off.csv (51 * 3)

  • State (character): U.S. States
  • pct_higher_cnts (numeric): The proportion of high civic opportunity counties in states
  • pct_lower_cnts (numeric): The proportion of lower civic opportunity counties in states

• county_opc.rds (3108 * 16)

  • opc_tile (numeric): Civic opportunity ranks (1-5)
  • The remaining columns are from the US shapefile, which is based on the tigris package in R (ver 2.03).

• la_zip_shp.shp: LA TIGER/Line shapefile (2018)

• la_zcta_opc.csv (300 * 2)

  • ZCTA (character): 5 digit Zip Code
  • opc_tile (numeric): Civic opportunity ranks (1-5)

• matched_mma_wa.csv (5,853 * 5)

  • ein (numeric): Employment Identification Number (IRS-assigned-organization IDs)
  • state (character): The state in which the organization is located
  • lobby (dummy): 1 = "lobbying organization," 0 = "non-lobbying organization"
  • predicted (character): 15 predicted categories or NA
  • ruling (numeric) = IRS ruling year-month (e.g., 200303)

• matched_mma_wa_summary.csv (30 * 6)

  • class (character): 15 predicted categories
  • n (numeric): The number of classified organizations in each category
  • freq (numeric): The frequency of classified organizations in each category
  • typr (character): "DC Organizations" or "Civic Organizations"

• org_flow_over_time.csv (30 * 6)

  • period (character): "Pre-1960" or "Post-2010"
  • class (character): 15 predicted categories
  • n (numeric): The number of classified organizations in each category
  • freq (numeric): The frequency of classified organizations in each category
  • flow_change (numeric): The frequency difference between the post-2010 and the pre-1960 cohorts
  • dir (character): "Increase" or "Decrease"

• org_volume_over_time.csv (15 * 4)

  • class (character): 15 predicted categories
  • all (numeric): The total number of classified organizations in each category
  • Post-2010 (numeric): The number of classified organizations in each category from the post-2010 cohort
  • Pre-1960 (numeric): The number of classified organizations in each category from the pre-1960 cohort

• pred_eval.csv (135 * 5)

  • .metric: "Accuracy," "Balanced accuracy," or "F-Score"
  • .estimator: "Binary"
  • .estimate: Predicted probability
  • model (classifier): "Lasso," "Random forest", or "XGBoost"
  • class (character): 15 predicted categories

• regression_mutual_aid.csv (3112 * 17)

  • FIPS (character): County FIPS code
  • has_hubs (numeric): Presence of COVID-19 mutual aid hubs in the county
  • hubs (numeric): Number of COVID-19 mutual aid hubs in the county
  • civic_opportunity_per_capita (numeric): Civic opportunity scores per capita
  • socialcap (numeric): Social capital measure from Kyne & Aldrich
  • civic_organizations_county (numeric): Civic organization measure from Chetty, et al.
  • sk2014 (numeric): Social capital measure from Rupasingha, et al.

• regression_vaccine_hesitancy.csv (3107 * 16)

  • FIPS (character): County FIPS code
  • vac.acceptance (numeric): County vaccine acceptance rate. See Pierri, et al.
  • civic_opportunity_per_capita (numeric): Civic opportunity scores per capita
  • sk2014 (numeric): Social capital measure from Rupasingha, et al.
  • misinfo (numeric): COVID-19 misinformation. See Pierri, et al.
  • covidmortality (numeric): County-level COVID-19 mortality rate

• regression_vaccine_uptake.csv (3107 * 16)

  • FIPS (character): County FIPS code
  • vpt.all (numeric): County vaccine uptake rate as defined by the CDC
  • civic_opportunity_per_capita (numeric): Civic opportunity scores per capita
  • sk2014 (numeric): Social capital measure from Rupasingha, et al.
  • Recip_State (numeric): U.S. States
  • covidmortality (numeric): County-level COVID-19 mortality rate

• regression_zcta_mn.csv (869 * 14)

  • GEOID (character): ZCTA Geoid (Zip Code)
  • vaccination_rate (numeric): ZCTA vaccination rate
  • civic_opportunity_per_capita (numeric): Civic opportunity scores per capita
  • d.share (numeric): Democratic vote share, 2020 election

• regression_zcta_tx.csv (1855 * 14)

  • GEOID (character): ZCTA Geoid (Zip Code)
  • vaccination_rate (numeric): ZCTA vaccination rate
  • civic_opportunity_per_capita (numeric): Civic opportunity scores per capita
  • d.share (numeric): Democratic vote share, 2020 election

• regression_zcta_ny.csv (1793 * 14)

  • GEOID (character): ZCTA Geoid (Zip Code)
  • vaccination_rate (numeric): ZCTA vaccination rate
  • civic_opportunity_per_capita (numeric): Civic opportunity scores per capita
  • d.share (numeric): Democratic vote share, 2020 election

Code

• Wrangling and joining the MMA and other outcomes and covariates
• Analyzing the MMA and other outcomes and covariates (Figures 1-3, Figures S1-2)
• Matching the MMA and Washington data
• Analyzing the MMA and Washington data (Figure 4, Tables S15-16)
• Regression tables
• Classification evaluation tables (Tables S2-3)

Tables and Figures

• All tables are in /tables
• All figures are in /figures