Extracts the residuals from a fixest_multi object — resid.fixest

Utility to extract the residuals from multiple fixest estimations. If possible, all the residuals are coerced into a matrix.

# S3 method for class 'fixest_multi'
resid(
  object,
  type = c("response", "deviance", "pearson", "working"),
  na.rm = FALSE,
  ...
)

# S3 method for class 'fixest_multi'
residuals(
  object,
  type = c("response", "deviance", "pearson", "working"),
  na.rm = FALSE,
  ...
)

Arguments

object: A fixes_multi object.
type: A character scalar, either "response" (default), "deviance", "pearson", or "working". Note that the "working" corresponds to the residuals from the weighted least square and only applies to feglm models.
na.rm: Logical, default is FALSE. Should the NAs be kept? If TRUE, they are removed.
...: Not currently used.

Value

If all the models return residuals of the same length, a matrix is returned. Otherwise, a list is returned.

Examples


base = iris
names(base) = c("y", "x1", "x2", "x3", "species")

# A multiple estimation
est = feols(y ~ x1 + csw0(x2, x3), base)

# We can get all the residuals at once,
# each column is a model
head(resid(est))
#>            [,1]        [,2]        [,3]
#> [1,] -0.6444588  0.10583516  0.08458424
#> [2,] -0.9561394  0.20359754  0.21000282
#> [3,] -1.1114672 -0.06831541 -0.04925142
#> [4,] -1.2338033 -0.20314694 -0.22599409
#> [5,] -0.7221227 -0.05371731 -0.08049948
#> [6,] -0.2551144  0.02604925  0.02280632

# We can select/order the model using fixest_multi extraction
head(resid(est[rhs = .N:1]))
#>             [,1]        [,2]       [,3]
#> [1,]  0.08458424  0.10583516 -0.6444588
#> [2,]  0.21000282  0.20359754 -0.9561394
#> [3,] -0.04925142 -0.06831541 -1.1114672
#> [4,] -0.22599409 -0.20314694 -1.2338033
#> [5,] -0.08049948 -0.05371731 -0.7221227
#> [6,]  0.02280632  0.02604925 -0.2551144