Negative One-Sided P-Values

[RahulDaswani]

Preliminaries

Before submitting an issue, please check (with x in brackets) that you:

• Are using the newest release (see here for latest release version number).
• [x ] Have checked that the examples in the help work.
• [x ] Have read the help (HTML version).
• [ x] Have checked that there is not already an existing issue for what you are reporting.
• [ x] Have check running synth directly for one of the treated units works.

Hello,

Firstly, I just wanted to say a huge thanks for this tool. I am a final-year undergraduate Economics student and this has really helped with my dissertation.

I was looking to conduct one-sided p-values for a negative effect (in my case, all 'effects' in the post-intervention period are negative). However, when I enter the pvals1s code in the appropriate manner, as instructed by the help book, the outputted p-values are the same as if I would have excluded it. Ideally, I would like to use the tool to plot the one-sided p-values as in the same manner the 'pvals_graphs' command works.

As a note:

-I am using a Mac
-My Stata version is 15.1
-My synth_runner package is fully updated.

Thanks a lot! I hope you will be able to come back to me soon as my deadline is in 2 days!

Best,

Rahul

[bquistorff]

When you add the pvals1s option then a new return value of e(pvals_1s) should be included. Are you seeing that?