Users can now supply a list of vcov arguments to compare multiple SE adjustments on the fly. (#65)
est = est = feols(flipper_len ~ bill_len | species, penguins)
ggcoefplot(est, vcov = list("iid", "hc1", ~species))
keep and drop arguments now work correctly with a list of models.
Thanks to @femdias for the report. (#60)ggcoefplot or ggiplot
then the original order is preserved for grouping and facet behaviour (#63)ggplot2 dependency to v4.0.0 and update test snapshots. (#55, #59)fixest dependency to v0.13.0 and update tests. (#58, #59)aggr_es(..., period = "diff") convenience keyword argument allows
users to estimate the difference between the (mean) post- and pre-treatment
periods. Thanks to @FBrunamonti for the suggestion. (#52).svglite dependency version and update test snapshots. (#51)aggr_es objects. (#43)ggh4x dependency with legendry. (#41 @teunbrand)First CRAN release!
aggr_es function now supports numeric sequences for aggregating a
specific subset of periods, in addition to the existing keyword strings like
"pre" or "post". This functionality also passes through to the higher order
plotting functions that call aggr_es under the hood. For example,
ggiplot(est, aggr_eff = 6:8). (#33)ggcoefplot(est, vcov = "hc1").
These on-the-fly adjustments are done via summary.fixest, and so the effect is
just the same as passing an adjusted object directly, e.g.
ggcoefplot(summary(est, vcov = "hc1")). However, it may prove more convenient
for simultaneously adjusting a list of multiple models, e.g.
ggcoefplot(list(est1, est2, est3), vcov = "hc1"). (#35)ggcoefplot, a ggplot equivalent of coefplot (#28).pt.size argument for controlling the size of point markers (#27).
Thanks @jcvdav.keep and drop arguments for subsetting coefficients (#22).iplot() (#e5cf0b0).log(y
(#20).marginaleffects::hypotheses() internally for aggr_es() to match
the upstream changes in marginaleffects.