Interactions with Plausible Values

[Rafaela Andres]

Hey!
I'm using the pv prefix in order to correctly work with PIAAC data. So far all regression outputs make sense, but I don't know how to include interactions with plausible values.
For example I want to know if the returns to numeracy skills for women are different to those of men.

Here is the command without interaction term (lwage as the dependent and some independent variables as gender, age etc. are in the macro called "base"):

pv, pv(PVNUM*) jrr jk(1) weight(SPFWT0) rw(SPFWT1 - SPFWT80): regress lwage $base @pv [aw = @w]

Does anybody know how to deal with interactions with plausible values? Any idea would be highly appreciated!

[Clyde Schechter]

-pv- is not an official Stata command. Your chances of getting a helpful response will improve if you tell us where it comes from.

[daniel klein]

I assume you have downloaded pv from the SSC.

From a quick glance at the code it seems that @pv is merely substituted for the different variable names, thus you might be able to to get results by treating @pv just as you would treat any regular variable name. However, unless someone on the list has experience with the issue, I would contact the author directly with this question.

From a statistical point of view, you should be aware that you are likely to underestimate any existing interactions if those interactions have not been accounted for during the imputation of the plausible values, which might or might not be documented in the PIAAC technical reports.

Best
Daniel

[Rafaela Andres]

Yes I downloaded pv from SSC.
Thank you both for your answers!

It is also right that @pv is written instead of the plausible values, but unfortunately I can't just include it with an interaction. If I for example type

pv, pv(PVNUM*) jrr jk(1) weight(SPFWT0) rw(SPFWT1 - SPFWT80): regress lwage $base @pv*female @pv [aw = @w]

Stata tells me that variable PVNUM1*female (which is the first plausible value times the female dummy) was not found.


Do you have any other idea?

Greetings from Germany,
Rafaela

[daniel klein]

Well, using an asterisk is not the proper way to denote interaction effects. You would not do this with a regular variable name, either. Compare

sysuse auto , clear
regress price mpg*foreign

with

sysuse auto , clear
regress price c.mpg##i.foreign

and see

help fvvarlist

for more. Then try applying the concept to your problem. As mentioned, not receiving an error message does not necessarily mean that everything went fine. Ask the author and consult the PIAAC doumentation, too.

Best
Daniel

[Rafaela Andres]

Wow I'm sorry, that was a bad mistake! I'm a newbie to Stata...
Thank you for you fast replies!
And yes I acknowledged the underestimation problem and will ask the author!

Greetings from Germany
Rafaela