minimal reproducible example explanation of change for influence functions

[ajdamico]

i feel like it'd be helpful for me if you prepared an illustrative example of the broader change?

so starting with any example design or dataset (for example: https://github.com/ajdamico/convey/blob/master/R/svygini.R#L33)

then pulling the contents of any of our functions/methods (for example: https://github.com/ajdamico/convey/blob/master/R/svygini.R#L109)

and (without removing the current steps) making the revised calculations side by side in the code

and finally printing the results both overall (which i believe does not change) and by subgroup (which does)?

(will this change affect replication designs?)

does what i'm asking make sense?

thanks!!

[ajdamico]

does it make sense to write this minimal reproducible example with svygini? probably the easiest for readers to understand unless there's a strong reason to show it with a different function..

[ajdamico]

https://github.com/guilhermejacob/context/commit/088f75c3d566ae091c951d15729db4fe98a09d13

[ajdamico]

this is a really nice explanation, thank you!