Extracting information from R objects and importing it to a modelsummary table-CodePudding

I am trying to make a statistical model summary table using the modelsummary package.

I ran 6 regressions. I conducted joint hypothesis tests for 2 of them, which include certain variables (specificaly, educ and exper) as explanatory variables. I want to include the F statistics from those tests in the table, but have trouble extracting information from the tests and importing it into the table.

Here are some more details.

(1) What I did, part 1: Packages, data, and analyses

library(tidyverse)
library(wooldridge)   # Data
library(estimatr)     # Estimation
library(car)          # F test
library(modelsummary) # Table
library(kableExtra)   # Table

wage2 <- wage2 %>% 
  mutate(lwage = log(wage))

model1 <- lm_robust(lwage ~ educ, 
                    data = wage2, se_type = "HC1")
model2 <- lm_robust(lwage ~ exper, 
                    data = wage2, se_type = "HC1")
model3 <- lm_robust(lwage ~ educ   exper, 
                    data = wage2, se_type = "HC1")
F3 <- linearHypothesis(model3, test = "F", 
                       c("educ = 0", "exper = 0"))
model4 <- lm_robust(lwage ~ educ   tenure, 
                    data = wage2, se_type = "HC1")
model5 <- lm_robust(lwage ~ exper   tenure, 
                    data = wage2, se_type = "HC1")
model6 <- lm_robust(lwage ~ educ   exper   tenure, 
                    data = wage2, se_type = "HC1")
F6 <- linearHypothesis(model6, test = "F", 
                       c("educ = 0", "exper = 0"))

The results of the joint hypothesis tests are as follows.

F3

## Linear hypothesis test
## 
## Hypothesis:
## educ = 0
## exper = 0
## 
## Model 1: restricted model
## Model 2: lwage ~ educ   exper
## 
##   Res.Df Df      F    Pr(>F)    
## 1    934                        
## 2    932  2 67.715 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

F6

## Linear hypothesis test
## 
## Hypothesis:
## educ = 0
## exper = 0
## 
## Model 1: restricted model
## Model 2: lwage ~ educ   exper   tenure
## 
##   Res.Df Df      F    Pr(>F)    
## 1    933                        
## 2    931  2 63.592 < 2.2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

(2) What I did, part 2: Table

regs <- list()
regs[['Model 1']] <- model1
regs[['Model 2']] <- model2
regs[['Model 3']] <- model3
regs[['Model 4']] <- model4
regs[['Model 5']] <- model5
regs[['Model 6']] <- model6

gm <- tribble(
  ~raw, ~clean, ~fmt,
  "adj.r.squared", "$\\bar{R}^2$", 3,
  "nobs", "Sample size", 0
  )

rows <- tribble(
  ~term, ~'Model 1', ~'Model 2', ~'Model 3', 
  ~'Model 4', ~'Model 5', ~'Model 6', 
  '$F$ statstics', '', '', '', '', '', '',
  '$H_0: \\beta_{\\rm{educ}} = 0, \\beta_{\\rm{exper}} = 0$', 
  '', '', '67.715***', '', '', '63.592***',
  '', '', '', '(0.000)', '', '', '(0.000)'
)
attr(rows, 'position') <- c(9:11)

table <- msummary(regs, 
         estimate = "{estimate}{stars}",
         gof_map = gm, 
         notes = list("*** p < 0.01, ** p < 0.05, * p < 0.1. 
                      Heteroskedasticity-consistent 
                      standard errors in brackets. 
                      P values in brackets for F statistics."),
         add_rows = rows) %>% 
  row_spec(c(1, 3, 5, 7, 9, 10), 
           extra_css = "border-bottom: hidden") 
table

The above code gave me a nice table. (Unfortunately, I do not seem to be able to post it because I have not earned enough reputation.)

(3) What I want to know

In the above, I manually typed in the F statistics, stars, and p-values in rows to obtain the table I wanted, but I would like to find out if there is a way to extract information from F3 and F6 and import it to the table.

CodePudding user response：

If I understand correctly, you want to:

Add two rows to the bottom of the table: F statistic and associated p-value
These rows should be added only to some models, which you pick manually

The simplest way to achieve this is to use the add_rows argument you use in your example. Your approach is correct!

If you want a more general and automated strategy,

CodePudding user response：

If I understand correctly, you would like to export data from the model. Just call it out:

F3$F

and

F3$`Pr(>F)`