lrberge
commented
10 months ago Hi, you can get that kind of information with coeftable. See ?coeftable.fixest_multi.
Arguments wide and long can be of interest.
base = setNames(iris, c("y", "x1", "x2", "x3", "species"))
est = feols(y ~ csw(x1, x2, x3) | species, base)
# default output
coeftable(est)
#> id rhs coefficient Estimate Std. Error t value Pr(>|t|)
#> 1 1 x1 x1 0.8035609 0.07139743 11.254759 0.007802281
#> 2 2 x1 + x2 x1 0.4322172 0.16130825 2.679449 0.115623419
#> 3 2 x1 + x2 x2 0.7756295 0.12654554 6.129252 0.025600825
#> 4 3 x1 + x2 + x3 x1 0.4958889 0.12061779 4.111242 0.054382298
#> 5 3 x1 + x2 + x3 x2 0.8292439 0.09701068 8.547965 0.013411238
#> 6 3 x1 + x2 + x3 x3 -0.3151552 0.10961007 -2.875239 0.102669993
# wide = TRUE
# => returns a list giving a wide table for each component
coeftable(est, wide = TRUE)
#> $coef
#> id rhs x1 x2 x3
#> 1 1 x1 0.8035609 NA NA
#> 2 2 x1 + x2 0.4322172 0.7756295 NA
#> 3 3 x1 + x2 + x3 0.4958889 0.8292439 -0.3151552
#>
#> $se
#> id rhs x1 x2 x3
#> 1 1 x1 0.07139743 NA NA
#> 2 2 x1 + x2 0.16130825 0.12654554 NA
#> 3 3 x1 + x2 + x3 0.12061779 0.09701068 0.1096101
#>
#> $tstat
#> id rhs x1 x2 x3
#> 1 1 x1 11.254759 NA NA
#> 2 2 x1 + x2 2.679449 6.129252 NA
#> 3 3 x1 + x2 + x3 4.111242 8.547965 -2.875239
#>
#> $pvalue
#> id rhs x1 x2 x3
#> 1 1 x1 0.007802281 NA NA
#> 2 2 x1 + x2 0.115623419 0.02560082 NA
#> 3 3 x1 + x2 + x3 0.054382298 0.01341124 0.10267
# long = TRUE
# => returns all the values stacked
coeftable(est, long = TRUE)
#> id rhs coefficient param value
#> 1 1 x1 x1 coef 0.803560901
#> 2 1 x1 x1 se 0.071397432
#> 3 1 x1 x1 tstat 11.254759089
#> 4 1 x1 x1 pvalue 0.007802281
#> 5 2 x1 + x2 x1 coef 0.432217209
#> 6 2 x1 + x2 x1 se 0.161308248
#> 7 2 x1 + x2 x1 tstat 2.679448904
#> 8 2 x1 + x2 x1 pvalue 0.115623419
#> 9 2 x1 + x2 x2 coef 0.775629458
#> 10 2 x1 + x2 x2 se 0.126545542
#> 11 2 x1 + x2 x2 tstat 6.129251549
#> 12 2 x1 + x2 x2 pvalue 0.025600825
#> 13 3 x1 + x2 + x3 x1 coef 0.495888938
#> 14 3 x1 + x2 + x3 x1 se 0.120617789
#> 15 3 x1 + x2 + x3 x1 tstat 4.111242167
#> 16 3 x1 + x2 + x3 x1 pvalue 0.054382298
#> 17 3 x1 + x2 + x3 x2 coef 0.829243912
#> 18 3 x1 + x2 + x3 x2 se 0.097010683
#> 19 3 x1 + x2 + x3 x2 tstat 8.547964933
#> 20 3 x1 + x2 + x3 x2 pvalue 0.013411238
#> 21 3 x1 + x2 + x3 x3 coef -0.315155173
#> 22 3 x1 + x2 + x3 x3 se 0.109610070
#> 23 3 x1 + x2 + x3 x3 tstat -2.875239220
#> 24 3 x1 + x2 + x3 x3 pvalue 0.102669993To access just one of the components, use the functions coef/tstat/pvalue. For example:
coef(est, long = TRUE)
#> id rhs coefficient estimate
#> 1 1 x1 x1 0.8035609
#> 4 2 x1 + x2 x1 0.4322172
#> 5 2 x1 + x2 x2 0.7756295
#> 7 3 x1 + x2 + x3 x1 0.4958889
#> 8 3 x1 + x2 + x3 x2 0.8292439
#> 9 3 x1 + x2 + x3 x3 -0.3151552On broom
I'm not a broom user, it's a different repo. It's a golden opportunity for you to contribute ;-) You may (or may not) use the function above to make it smooth.
broom::tidy()works great on regular oldfixestresult objects but fails onfixest_multi. It would be great if this worked out of the box rather than having to hack it together manually.