multi_per_variable {sampcompR}

Returns a table based on the information of a multi_compare_object that indicates the proportion of biased variables. It can be outputted as html or LaTex Table, for example with the help of the stargazer function.

Description

Returns a table based on the information of a multi_compare_object that indicates the proportion of biased variables. It can be outputted as html or LaTex Table, for example with the help of the stargazer function.

Usage

multi_per_variable(
  multi_compare_objects,
  type = "coefs",
  label_df = NULL,
  lables_coefs = NULL,
  lables_models = NULL,
  ndigits = 1
)

Arguments

multi_compare_objects

A object returned by the biv_compare function.

type

The type of table, can either be "coefs", "models", or "complete". When coefs is chosen, the average bias of the coefficients is outputted, when models is chosen, the average bias for the models is outputted, and when complete is chosen, both is outputted.

label_df

A character vector containing labels for the data frames.

lables_coefs

A character vector containing labels for the coefficients.

lables_models

A character vector containing labels for the models.

ndigits

Number of digits shown in the table.

Value

A matrix, that indicates the proportion of bias for every individual coefficient or model for multivariate comparisons. This is given separately for every comparison, as well as averaged over comparisons.

Examples

require(wooldridge)
card<-wooldridge::card

south <- card[card$south==1,]
north <- card[card$south==0,]
black <- card[card$black==1,]
white <- card[card$black==0,]

## use the function to plot the data
multi_data1 <- sampcompR::multi_compare(df = north, 
                                        bench = south,
                                        independent = c("age","fatheduc","motheduc","IQ"),
                                        dependent = c("educ","wage"),
                                        family = "ols") 

multi_data2 <- sampcompR::multi_compare(df = black, 
                                        bench = white,
                                        independent = c("age","fatheduc","motheduc","IQ"),
                                        dependent = c("educ","wage"),
                                        family = "ols") 

table<-sampcompR::multi_per_variable(multi_compare_objects = c("multi_data1","multi_data2"))
noquote(table)