Extensive/intensive first-difference decomposition for a hurdle- or ZI-CPB

The effect of moving variable from one value to another on the participation (extensive) margin, the intensity (intensive) margin, and the marginal expected count, holding other covariates at their means (or at the values at gives them). The two margins are reported side by side; their changes do not add up to the marginal change (see first_difference.count). With B = 0 the intervals follow the package's one rule (see first_difference()): the stored bootstrap draws of a zi_cpb fit, and for a hurdle the delta method over its two margins' covariances (the participation model's, and the intensity's stored bootstrap draws). B > 0 refits the whole model on resampled data instead.

Usage

# S3 method for class 'hurdle_cpb'
first_difference(
  object,
  variable,
  from,
  to,
  B = 0,
  level = 0.95,
  data = NULL,
  stage = c("both", "participation", "intensity", "zero", "count"),
  cores = 1L,
  at = NULL,
  ...
)

# S3 method for class 'zi_cpb'
first_difference(
  object,
  variable,
  from,
  to,
  B = 0,
  level = 0.95,
  data = NULL,
  stage = c("both", "participation", "intensity", "zero", "count"),
  cores = 1L,
  at = NULL,
  ...
)

Arguments

object

A "hurdle_cpb" or "zi_cpb" object.

variable

Name of a model-matrix column to vary (in either margin).

from, to

The two values to contrast.

B

Refit-bootstrap replicates for percentile intervals; 0 (default) uses the fit's own inference instead (see Details). For zi_cpb the refit bootstrap is slow (each replicate refits the mixture).

level

Confidence level.

data

The original data frame; required when B > 0. Each replicate refits the model with the fit's own settings (fixed effects, link, offset, weights, support, and the like), resampling whole fixed-effects units (or the fit's clusters) when it has them; an offset or weights given to the fit as a vector must then be a column of data instead.

stage

Which equation(s) the covariate moves in: "both" (default), the binary stage only ("participation"/"zero"), or the count stage only ("intensity"/"count"), holding the covariate at its reference in the other equation. All three component rows are always returned.

cores

Worker processes for the bootstrap refits (default 1); see cpb().

at

Optional named numeric vector fixing other model-matrix columns at given values instead of their means (see first_difference()).

...

Unused; unknown arguments error.

Value

A "ud_fd" data frame (the package-wide first-difference contract): columns component, from, to, diff, lower, upper, method, one row per margin level.