The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.

Multiple Treated

Background

In this vignette I show how {bsynth} can be used for causal estimation when you have multiple treated accounts. In particular, I use the synthetic data exaple from Xu, Yiqing, 2017. In this example, the treatment starts in period 21 and increases by one each period (e.g. the effect is 5 period 25 and 10 in period 30).

Figure 1: Xu, Yiqing, 2017

Figure 1: Xu, Yiqing, 2017

Bayesian Synthetic Control with Covariates

library(bsynth)
ci_width <- 0.95
data(gsynth, package = "gsynth")
dplyr::glimpse(simdata)
#> Rows: 1,500
#> Columns: 15
#> $ id    <dbl> 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101, 101,…
#> $ time  <dbl> 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 1…
#> $ Y     <dbl> 6.2109977, 4.0271059, 8.8771874, 11.5153460, 5.9715262, 8.237905…
#> $ D     <dbl> 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1…
#> $ X1    <dbl> 0.3776736, 1.7332009, 1.8580159, 1.3943369, 2.3636963, 0.5370867…
#> $ X2    <dbl> -0.1732470, -0.4945009, 0.4984432, 1.1272713, -0.1535215, 0.8774…
#> $ eff   <dbl> 0.0000000, 0.0000000, 0.0000000, 0.0000000, 0.0000000, 0.0000000…
#> $ error <dbl> 0.29822759, 0.63656967, -0.48378063, 0.51686204, 0.36896453, -0.…
#> $ mu    <dbl> 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5…
#> $ alpha <dbl> -0.06052364, -0.06052364, -0.06052364, -0.06052364, -0.06052364,…
#> $ xi    <dbl> 1.1313372, -1.4606401, 0.7399475, 1.9091036, -1.4438932, 0.70178…
#> $ F1    <dbl> 0.25331851, -0.02854676, -0.04287046, 1.36860228, -0.22577099, 1…
#> $ L1    <dbl> -0.04303273, -0.04303273, -0.04303273, -0.04303273, -0.04303273,…
#> $ F2    <dbl> 0.005764186, 0.385280401, -0.370660032, 0.644376549, -0.22048656…
#> $ L2    <dbl> -0.8804667, -0.8804667, -0.8804667, -0.8804667, -0.8804667, -0.8…

outcome_data <- simdata %>% 
  dplyr::select(time, id, D, Y)

covariates <- simdata %>% 
  dplyr::select(time, id, X1, X2)

synth <-
  bsynth::bayesianSynth$new(
    data = outcome_data,
    time = time,
    id = id,
    treated = D,
    outcome = Y, 
    ci_width = ci_width,
    covariates = covariates
  )

synth$timeTiles +
  ggplot2::theme(text = ggplot2::element_text(size=6))

Fit

synth$fit()

Visualize the synthetic controls for each treated unit

synth$synthetic

### Visualize the treatment effect

synth$effectPlot(subset = c("Average"), facet = FALSE) + 
  ggplot2::scale_y_continuous(breaks=seq(-2,12,2))

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.