Compute marginal effect of a covariate — tobs_marginal

Returns the fitted response over the range of one covariate, holding the others at their means, on the process's response scale (occupancy / detection probability, or abundance intensity).

Usage

tobs_marginal_effect(
  object,
  covariate,
  process = c("occupancy", "detection", "abundance"),
  n_points = 100L
)

Arguments

object: A tobs_fit object.
covariate: Name of covariate.
process: "occupancy" (default), "detection", or "abundance" (the state process of a count family, on the intensity scale).
n_points: Number of prediction points (default 100).

Value

A data.frame with covariate values and the predicted response.

Examples

# \donttest{
sim <- simulate_abun(N = 100, J = 4, n_abund_covs = 2, n_det_covs = 1,
                     seed = 1)
fit <- tobs(~ abund_cov1 + abund_cov2, data = sim$data,
            family = abun(K_max = 50), detection = ~ det_cov1, y = sim$y,
            control = list(verbose = FALSE))
me <- tobs_marginal_effect(fit, "abund_cov1", process = "abundance")
head(me)   # estimate is on the abundance (lambda) scale, not a probability
# }