I am using pretty much the exact same code as provided in the example for tidysynth, however, the graph I am receiving has the incorrect unit where the intervention occurred. The unit of intervention I assigned is "Colombia", however, it is returning "Argentina" for all graphs.
`code:
all_data <- correct_missing %>% synthetic_control(outcome = loghomicides, # outcome
unit = Country, # unit index in the panel data
time = Year, # time index in the panel data
i_unit = "Colombia", # unit where the intervention occurred
i_time = 2015, # time period when the intervention occurred
generate_placebos=T # generate placebo synthetic controls (for inference)
) %>%
generate_predictor(time_window = 2005:2015,
prison_pop = mean(logprison_population, na.rm = T),
gini_ind = mean(loggini_index, na.rm = T),
unemployment_log = mean(logunemployment, na.rm = T),
serious_assault = mean(logserious_assault_rate, na.rm = T)
) %>%
generate_predictor(time_window = 2010,
homicides_2010 = loghomicides) %>%
generate_predictor(time_window = 2006,
homicides_2006 = loghomicides) %>%
generate_predictor(time_window = 2015,
homicides_2015 = loghomicides) %>%
generate_weights(optimization_window = 2005:2015,
margin_ipop = .02,sigf_ipop = 7,bound_ipop = 6) %>% generate_control()`
I am using pretty much the exact same code as provided in the example for tidysynth, however, the graph I am receiving has the incorrect unit where the intervention occurred. The unit of intervention I assigned is "Colombia", however, it is returning "Argentina" for all graphs.
`code: