Hello,
I am currently using synth on dataset of 21 countries with 12 variables to create a synthetic Hungary. When doing so, I wanted to check the statistical significance of my results and used synth_runner to get estimated effects results. When doing so I observed discrepancies between the graph generated by synth and the one generated using effect_graphs by synth_runner. Please see the images attached below.
I concluded this has to be due to one of two reasons: 1) either the treatment unit somehow changed or 2) synth_runner is a different method of calculating estimates.
Although both seem unlikely, the first one seems more unlikely than the second one having checked for it a few times.
This could very well be a shortcoming from my side but any help or suggestion would be appreciated!
Result from synth:
Result from synth_runner:
Command for synth_runner (treated unit is coded and the country code 4 refers to the same country in synth and synth_runner as far as I am aware):
I am currently using synth on dataset of 21 countries with 12 variables to create a synthetic Hungary. When doing so, I wanted to check the statistical significance of my results and used synth_runner to get estimated effects results. When doing so I observed discrepancies between the graph generated by synth and the one generated using effect_graphs by synth_runner. Please see the images attached below.
I concluded this has to be due to one of two reasons: 1) either the treatment unit somehow changed or 2) synth_runner is a different method of calculating estimates.
Although both seem unlikely, the first one seems more unlikely than the second one having checked for it a few times.
This could very well be a shortcoming from my side but any help or suggestion would be appreciated!
Result from synth:
Result from synth_runner:
Command for synth_runner (treated unit is coded and the country code 4 refers to the same country in synth and synth_runner as far as I am aware):
