Amelia output with fixed effects regression

[flxflks]

I used Amelia to impute the missing data in my time-series cross-sectional data. However, when reading the vignette I couldn't find any example to use Amelia with fixed / random regressions. Notably, plm and fixest do not work with with for me. Is there a package that allows to model fixed effect regressions with an Amelia object?

[flxflks]

Crossposted on SO

[mattblackwell]

Yeah, the with() function for Amelia only works with models that don't need the data argument. For situations like this, you can use lapply() to do the fits like so:

library(tidyverse)
library(broom)
library(Amelia)
library(fixest)
library(plm)

# Load dataset
data("africa", package = "Amelia")

# Multiple imputation
a.out <- amelia(x = africa, m = 10, cs = "country",
                ts = "year", logs = "gdp_pc", p2s = 0)

# Run fixed effects regression
fits <- lapply(
  a.out$imputations,
  \(x) feols(gdp_pc ~ infl * trade, data = x)
)

mi.combine(fits)