partialling_out: partialling out variable of interest and main

Creates a data.frame of the residualised main explanatory variable and, if wanted, variable of interest of a linear or fixed effects model

Usage

partialling_out(model, data, weights, both, na.rm, ...)

Arguments

model: object for which we want to residualise variables
data: data.frame used in the original model. Using different data will return unexpected results or an error.
weights: a numeric vector for weighting the partial models. Length must be equal to number of rows of data
both: if TRUE will residualise both the variable of interest and the first explanatory variable in the model. If FALSE, only the latter. Set to TRUE by default
na.rm: if TRUE will remove observations with NA before any models are run. If FALSE, the underlying lm, feols, or felm will remove NA values but errors may arise if weights are used.
...: Any other lm, feols, or felm parameters that will be passed to the partial regressions

Value

a data.frame with the (residualised) variable of interest and residualised main explanatory variable

Details

The function regresses the main (i.e. first in the model) explanatory variable and the variable of interest (if parameter both is set to TRUE) against all other control variables and fixed effects and returns the residuals in a data.frame

Will accept lm, felm (lfe package), and feols (fixest package) objects

Examples

library(palmerpenguins)
#> 
#> Attaching package: ‘palmerpenguins’
#> The following objects are masked from ‘package:datasets’:
#> 
#>     penguins, penguins_raw
library(fixest)
model <- feols(bill_length_mm ~ bill_depth_mm | species + island,
               data = penguins)
#> NOTE: 2 observations removed because of NA values (LHS: 2, RHS: 2).
partial_df <- partialling_out(model, penguins, both = TRUE)