Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

p-value in the summary has a cap of 0.5 #56

Open
charlesdong1991 opened this issue Jun 30, 2022 · 5 comments
Open

p-value in the summary has a cap of 0.5 #56

charlesdong1991 opened this issue Jun 30, 2022 · 5 comments

Comments

@charlesdong1991
Copy link

Hi, thanks for making this nice package!

Small question regarding the p-value from summary: From the doc here , it seems the range of p-value is from 0 to 1.

But when trying to use the package and checking the p-value from summary, it seems the cap of p-value is 0.5?

@WillianFuks
Copy link
Owner

Hi @charlesdong1991 ,

You are correct, the upper limit for the p-value is indeed 50%. This happens because the computation of signal is symmetric due the min operation. So 50% is indeed the maximum value it can reach.

I'll improve the docs for this function soon. Thanks for pointing that out!

Best,

Will

@charlesdong1991
Copy link
Author

charlesdong1991 commented Jun 30, 2022

Thank you very much for your fast reply!! Appreciate it a lot!!

Small follow up question: is the interpretation of p-value in this case changed? How do you suggest to interpret the p-value here, e.g. when p-value=0.5, what does it mean?

Thanks again!! 👍

@WillianFuks
Copy link
Owner

The p-value has the same definition as in statistics. It represents the likelihood of obtaining the observed results supposing the null-hypothesis is true, in other words, if p=0.5 it means there's a 50% chance of observing the post-intervention period when there's actually no impact. As the value is high then we don't consider that there was a meaningful impact from the intervention.

But if you run the same test and find p=0, that means that the chance of observing the results by random chance without the intervention is effectively zero. In this case, as the response is considered too far away from what is normally expected, we then consider that there was a meaningful impact.

So in a nutshell it's really the usual definition of p-statistics as defined in the literature.

@charlesdong1991
Copy link
Author

Great thanks for your detailed explanation!! 👍

@marguegit
Copy link

Thank you @WillianFuks !

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants