Description
Synthetic Control Methods with Micro- And Meso-Level Data.
Description
A generalization of the 'Synth' package that is designed for data at a more granular level (e.g., micro-level). Provides functions to construct weights (including propensity score-type weights) and run analyses for synthetic control methods with micro- and meso-level data; see Robbins, Saunders, and Kilmer (2017) <doi:10.1080/01621459.2016.1213634> and Robbins and Davenport (2021) <doi:10.18637/jss.v097.i02>.
README.md
microsynth
The goal of microsynth is to implement the synthetic control method for micro-level data as outlined in Robbins, Saunders, and Kilmer (2017). is designed for use in assessment of the effect of an intervention using longitudinal data. However, it may also be used to calculate propensity score-type weights in cross-sectional data. is a generalization of (see Abadie and Gardeazabal (2003) and Abadie, Diamond, Hainmueller (2010, 2011, 2014)) that is designed for data at a more granular level (e.g., micro-level). For more details see the help vignette: .
Installation
You can install microsynth from github with:
# install.packages("devtools")
devtools::install_github("ssdavenport/microsynth")
Example
This is a basic example which shows you how to run a simple microsynth model, using block-level panel data to evaluate a crime intervention:
library("microsynth")
# Use seattledmi, block-level panel data, to evaluate a crime intervention.
# Declare time-variant (outcome) and time-invariant variables for matching
cov.var <- c('TotalPop', 'BLACK', 'HISPANIC', 'Males_1521',
'HOUSEHOLDS', 'FAMILYHOUS', 'FEMALE_HOU', 'RENTER_HOU', 'VACANT_HOU')
match.out <- c('i_felony', 'i_misdemea', 'i_drugs', 'any_crime')
set.seed(99199) # for reproducibility
# Perform matching and estimation, without permutations or jackknife
# runtime: < 1 min
sea1 <- microsynth(seattledmi,
idvar="ID", timevar="time", intvar="Intervention",
start.pre=1, end.pre=12, end.post=16,
match.out=match.out, match.covar=cov.var,
result.var=match.out, omnibus.var=match.out,
test="lower")
#> Calculating weights...
#> Created main weights for synthetic control: Time = 1.68
#> Matching summary for main weights:
#> Targets Weighted.Control All.scaled
#> Intercept 39 39.0002 39.0000
#> TotalPop 2994 2994.0519 2384.7477
#> BLACK 173 173.0010 190.5224
#> HISPANIC 149 149.0026 159.2682
#> Males_1521 49 49.0000 97.3746
#> HOUSEHOLDS 1968 1968.0340 1113.5588
#> FAMILYHOUS 519 519.0108 475.1876
#> FEMALE_HOU 101 101.0010 81.1549
#> RENTER_HOU 1868 1868.0203 581.9340
#> VACANT_HOU 160 160.0115 98.4222
#> i_felony.12 14 14.0000 4.9023
#> i_felony.11 11 11.0002 4.6313
#> i_felony.10 9 9.0000 3.0741
#> i_felony.9 5 5.0000 3.2642
#> i_felony.8 20 20.0000 4.4331
#> i_felony.7 8 8.0000 3.7617
#> i_felony.6 13 13.0000 3.0012
#> i_felony.5 20 20.0007 3.1549
#> i_felony.4 10 10.0000 4.0246
#> i_felony.3 7 7.0000 3.3693
#> i_felony.2 13 13.0002 3.2803
#> i_felony.1 12 12.0000 3.4381
#> i_misdemea.12 15 15.0002 4.2470
#> i_misdemea.11 12 12.0000 4.6070
#> i_misdemea.10 12 12.0000 4.0772
#> i_misdemea.9 14 14.0000 3.7414
#> i_misdemea.8 12 12.0000 3.9680
#> i_misdemea.7 20 20.0000 4.2551
#> i_misdemea.6 16 16.0005 3.5594
#> i_misdemea.5 24 24.0000 3.5635
#> i_misdemea.4 21 21.0002 4.3360
#> i_misdemea.3 21 21.0000 4.3846
#> i_misdemea.2 14 14.0000 3.5352
#> i_misdemea.1 16 16.0000 4.1540
#> i_drugs.12 13 13.0000 1.6543
#> i_drugs.11 8 8.0000 1.5128
#> i_drugs.10 3 3.0000 1.3227
#> i_drugs.9 4 4.0000 0.9788
#> i_drugs.8 4 4.0000 1.1123
#> i_drugs.7 10 10.0000 1.0516
#> i_drugs.6 4 4.0000 1.2377
#> i_drugs.5 2 2.0000 1.2296
#> i_drugs.4 1 1.0000 1.1245
#> i_drugs.3 5 5.0000 1.3550
#> i_drugs.2 12 12.0000 1.1366
#> i_drugs.1 8 8.0002 1.3591
#> any_crime.12 272 272.0012 65.3398
#> any_crime.11 227 227.0017 64.2396
#> any_crime.10 183 183.0010 55.6929
#> any_crime.9 176 176.0005 53.2377
#> any_crime.8 228 228.0005 55.8143
#> any_crime.7 246 246.0024 55.8062
#> any_crime.6 200 200.0010 52.8292
#> any_crime.5 270 270.0014 50.6531
#> any_crime.4 250 250.0010 57.2946
#> any_crime.3 236 236.0010 58.8681
#> any_crime.2 250 250.0012 51.5429
#> any_crime.1 242 242.0010 55.1145
#> Calculation of weights complete: Total time = 2.06
#>
#> Calculating basic statistics for end.post = 16...
#> Completed calculation of basic statistics for end.post = 16. Time = 3.76
#>
#> Calculating survey statistics for end.post = 16...
#> Completed survey statistics for main weights: Time = 9.22
#> Completed calculation of survey statistics for end.post = 16. Time = 9.22
#>
#> microsynth complete: Overall time = 17.73
sea1 # View description of microsynth object and results
#> microsynth object
#>
#> Scope:
#> Units: Total: 9642 Treated: 39 Untreated: 9603
#> Study Period(s): Pre-period: 1 - 12 Post-period: 13 - 16
#> Constraints: Exact Match: 58 Minimized Distance: 0
#> Time-variant outcomes:
#> Exact Match: i_felony, i_misdemea, i_drugs, any_crime (4)
#> Minimized Distance: (0)
#> Time-invariant covariates:
#> Exact Match: TotalPop, BLACK, HISPANIC, Males_1521, HOUSEHOLDS, FAMILYHOUS, FEMALE_HOU, RENTER_HOU, VACANT_HOU (9)
#> Minimized Distance: (0)
#>
#> Results:
#> end.post = 16
#> Trt Con Pct.Chng Linear.pVal Linear.Lower Linear.Upper
#> i_felony 46 68.22 -32.6% 0.0109 -50.3% -8.4%
#> i_misdemea 45 71.80 -37.3% 0.0019 -52.8% -16.7%
#> i_drugs 20 23.76 -15.8% 0.2559 -46.4% 32.1%
#> any_crime 788 986.44 -20.1% 0.0146 -32.9% -4.9%
#> Omnibus -- -- -- 0.0011 -- --