Closed garland-culbreth closed 10 months ago
The problem here is that 'true value' has not been defined. Does the following for you and Sarah ?
This corresponds to the simulation value for all the model variables and covariates. This can be thought of as the true value for the average of the integrand that is measured.
Note that the concept of true value gets becomes unclear when you have model miss-specification.
That works perfectly! Updated the text in the most recent commit.
If you look at https://github.com/bradbell/at_cascade/commits/master
You will notice that all my log entries start with the branch that the edits are done on. E.g. my changes to pull request #3 start with pull_3:
. It would help if your log entries started with a_short_name_that_you_choose: ; e.g. in this case it could have been meas_mean:
.
We thought adding this extra line to this description would be useful for users, to make it extra clear that the meas_mean is also equivalent to the truth used for the sim.