I've been attempting to work with Amelia to multiply overimpute values for
a specific variable that have been measured with error. I know which
observations were measured with error (they all have a value in a
particular range) but I cannot use moPrep to properly prepare the data for
overimputation. I want to use the "subset" and "gold.standard"
arguments,
but depending on how I write these, some error is always returned.
I cannot find a simple example of sample code that uses this means of
prepping the data, rather than specifying the error proportion or sd (which
I don't have/know). Can you either point me to an example, or suggest a
simple line of code, using the following setup:
dataframe= data
variable with error: A
subset with error: where A<5
Thanks for any help or guidance you can give. All the best,
Sean