Warning about CI length

[molivov]

I am running HonestDiD on a large dataset. I am getting the following warnings when running HonestDiD::createSensitivityResults_relativeMagnitudes

CI is open at upper end-point; CI length may not be accurate.

When I run warnings() I get multiple lines of

In .ARP_computeCI(betahat = betahat, sigma = sigma, numPrePeriods = numPrePeriods, ... :
CI is open at one of the endpoints; CI length may not be accurate

Is that something to worry about?

[molivov]

Turns out this happens when the variance-covariance matrix of coefficients has very small entries. The variance-covariance matrix I have comes from estimation on a large admin dataset.

[jonathandroth]

HonestDiD computes a confidence interval by inverting tests over a grid. The warning indicates that the CI includes either the upper or lower point in the default grid, so you may be missing some points. You can change the grid.lb and grid.ub arguments, which may eliminate the warning

…

On Fri, Oct 18, 2024 at 2:47 AM molivov ***@***.***> wrote: I am running HonestDiD on a large dataset. I am getting the following warnings when running HonestDiD::createSensitivityResults_relativeMagnitudes CI is open at upper end-point; CI length may not be accurate. When I run warnings() I get multiple lines of In .ARP_computeCI(betahat = betahat, sigma = sigma, numPrePeriods = numPrePeriods, ... : CI is open at one of the endpoints; CI length may not be accurate Is that something to worry about? — Reply to this email directly, view it on GitHub <#62>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AE6EXFGFKPIVQ5NOTZSZK5LZ4CVIZAVCNFSM6AAAAABQFIOEXOVHI2DSMVQWIX3LMV43ASLTON2WKOZSGU4TMNJZGE2TANI> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>