Add information in stargazer with coeftest

Question

I am estimating a model using survey design and robust errors, and I am trying to display the results with stargazer. However, I haven't been able to add basic information, such as number of observations, R2, Adjusted R2, Residual Std. Error and F Statistic, at the end of the table.

My code is something like this (I abbreviated the model):

# Svy design.
svy <- svydesign(id = ~conglomerado, 
                 strata = ~estrato, 
                 check.strata = TRUE,
                 weights = ~fact_cal_esi, 
                 data = subset(completos, ocup_ref==1))

# Estimation.
short_model <- svyglm(ln_w ~ sex + age + factor(level),
                      design=svy)
short_model_r <- coeftest(short_model , vcov = vcovHC(short_model , type = "HC1"))

# Table.
stargazer(models,type="text", 
          column.labels = c("Short model"),
          title = 'Table 1: Short model')

But I do not know how to add those statistics.

Answer 1

The problem is that svyglm does not produce statistics like R-squared that can then be passed to stargazer. It seems that you would have to calculate these additional statistics of interest yourself and then pass them to stargazer using the add.lines parameter of stargazer. You can read about add.lines in the stargazer documentation here: https://cran.r-project.org/web/packages/stargazer/stargazer.pdf.

Here would be a reproducible example:

library(survey);library(stargazer)
set.seed(1)

data = data.frame(conglomerado=rnorm(100),
                  estrato=rnorm(100),
                  fact_cal_esi=abs(rnorm(100)),
                  sex=rnorm(100),
                  age=rnorm(100),
                  ln_w=rnorm(100),
                  level=rnorm(100))

# Svy design.
svy <- svydesign(id = ~conglomerado, 
                 strata = ~estrato, 
                 check.strata = TRUE,
                 weights = ~fact_cal_esi, 
                 data = data)

# Estimation.
options(survey.lonely.psu="remove")
short_model <- svyglm(ln_w ~ sex + age + factor(level),
                      design=svy)

#R-squared according to 
#https://stackoverflow.com/questions/73926485/survey-package-how-do-i-get-r-squared-from-a-svyglm-object
total_var <-svyvar(data, svy)
resid_var <- summary(short_model)$dispersion
rsq <- 1-resid_var/total_var

# Table.
stargazer(short_model,type="text", 
          column.labels = c("Short model"),
          title = 'Table 1: Short model',
          omit="level",
          add.lines = list(c("R-Squared","1")))

These other posts ask similar questions about calculating these statistics from svyglm() and may be helpful to you in doing this:

Survey-package: How do I get R-squared from a svyglm-object? ,

https://stats.stackexchange.com/questions/618248/how-to-calculate-f-statistic-from-a-svyglm-model-in-r ,

https://stats.stackexchange.com/questions/523152/different-ways-to-calculate-rsquared-after-regression-with-complex-survey-data-i.