Plot CATT or event-study estimates from a fitted FETWFE / ETWFE / BETWFE

Returns a ggplot object showing either event-study coefficients (default; type = "event_study") or per-cohort average treatment effects (type = "catt") from a fitted estimator. Mirrors the visualization style of did::ggdid() from the Callaway-Sant'Anna did package, providing a single-call route from a fitted object to a publication-ready visualization.

For fetwfe / betwfe (the bridge-penalty estimators) in the CATT view, points are shape- and color-coded by whether the bridge penalty left that cohort's ATT nonzero (selected = TRUE) or zeroed it out (selected = FALSE). For etwfe (no selection), all points are uniformly styled.

Usage

# S3 method for class 'fetwfe'
plot(x, type = c("event_study", "catt"), conf_int = TRUE, alpha = NULL, ...)

Arguments

x

A fitted object from fetwfe(), etwfe(), or betwfe(). (twfeCovs is not currently supported – see GitHub issue #58 for the broader treatment of twfeCovs class methods.)

type

Character; either "event_study" (default; event-time pooled coefficients from eventStudy()) or "catt" (per-cohort ATTs from result$catt_df).

conf_int

Logical; if TRUE (default), include confidence interval error bars.

alpha

Numeric; overrides the fit's alpha for CI computation. NULL (default) uses the fit's alpha (so 95% CIs when the fit's alpha is 0.05). With the default NULL, both views show the fit's ci_type band (simultaneous by default): the "catt" view reads the bounds stored in catt_df, and the "event_study" view re-derives them via eventStudy() (which inherits the fit's ci_type). Asymmetry under an explicit alpha: for the "event_study" view, the explicit alpha is forwarded to eventStudy(), which re-runs the joint machinery and so still produces a simultaneous band at that alpha (when the fit's ci_type is "simultaneous"); for the "catt" view, an explicit alpha recomputes ci_low / ci_high from estimate +/- qnorm(1 - alpha/2) * se, i.e. a pointwise band at that alpha (the stored simultaneous bounds are at the fit's alpha and are not re-derived at a new alpha for the catt view). To plot simultaneous catt bands at a different alpha, refit at that alpha or call simultaneousCIs() directly.

...

Currently unused; reserved for future arguments.

Value

A ggplot object. Users can customize further via standard ggplot layer-addition syntax (e.g., plot(res) + ggplot2::theme_classic()).

See also

cohortStudy() for the per-cohort accessor; eventStudy() for the event-time accessor; plot.etwfe(), plot.betwfe() for the parallel methods.

Examples

if (FALSE) { # \dontrun{
  coefs <- genCoefs(G = 3, T = 6, d = 2, density = 0.5, eff_size = 2)
  dat <- simulateData(coefs, N = 120, sig_eps_sq = 1, sig_eps_c_sq = 0.5, seed = 123)
  res <- fetwfeWithSimulatedData(dat)
  if (requireNamespace("ggplot2", quietly = TRUE)) {
    plot(res)                          # default: event-study coefficients
    plot(res, type = "catt")           # per-cohort ATTs
    plot(res, conf_int = FALSE)        # point estimates only
    plot(res, alpha = 0.1)             # 90% CIs
  }
} # }