Analysis Plan and Project Code
Estimate baseline racial disproportionality in child welfare referrals and acceptances from July 2014 to June 2017.
- Using data from the most recent American Community Survey 5-year estimates (2012-2016), generate population estimates for children in Charlottesville by race and calculate proportions by race, incorporating the uncertainty represented by the 90% margins of error. See
prep_basedisp.R
for details. The detailed tables in the 5-year estimates were estimates for sex by age for
- The combined population (B01001)
- White alone population (B01001A)
- Black or African-American alone population (B01001B)
- American Indian and Alaska Native alone population (B01001C)
- Asian alone population (B01001D)
- Native Hawaiian and Other Pacific Islander alone population (B01001E)
- Some other race alone population (B01001F)
- Two or more races population (B01001G)
-
Using the CWS referral data from June 2014-July 2016 from DSS, generate counts of children referred to CWS by race for each year (with years defined as June-July). Calculate proportions by race. See prep_basedisp.R
for details.
-
Generate visualization of proportions of children by race in Charlottesville and proportions of children referred to CWS (see analysis_basedisp.R
).
-
Calculate racial disproportionatilty by race, the ratio of % of [racial cateogry] children referred to CWS to % of [racial category] children in population (e.g., % referred to CWS who are white/% children in Cville who are white). Generate point estimates and estimates for lower and upper bounds of estimated population proportions (see analysis_basedisp.R
).
- Pull Charlottesville Census tract data for child population and poverty rates; generate maps for child population, poverty rates, referral rates. See
prep_tract.R
.
Estimate racial disparity in post-referral decision points, modeling decision outcomes as a function of race and other characteristics.
- Reporter relation – racial proportions by reporter relation
- Screened in – referred children screened in (accepted as a case) as a function of race and covariates
- Investigation – accepted/screened in children assigned to investigation or assessment (alternative response) as a function of race and covaraites
- Substantiation – outcome of investigations for investigated cases as a function of race and covariates
- Level of substantiaion – level of finding among substantiated cases
See prep_postreferral.R
, figures_postreferral.R
, and analysis_postreferral.R
.
Estimate racial disparity in foster care outcomes, modeling outcomes as a function of race and other characteristics.
- Initial placement type – among children placed in out-of-home care, type of placement by race and other characteristics
- Number of placements – among children placed in out-of-home care, number of placements by race and other characteristics
- Time in care – among children placed in out-of-home care, duration of time in care by race and other characteristics
- Exit – among children placed in out-of-home care more than a year prior to June 30, 2018, exit from the system within year by race and other characteristics
- Nature of exit – among children in out-of-home care who have exited, how/why they exit by race and other characteristics