How to export tables from R depends on what word processor you use. This tutorial focuses on Word. If you use LaTeX, there are many existing R packages and tutorials that will get you started, including xtable and stargazer.
To export tables to Word, follow these general steps:
- Create a table or data.frame in R.
- Write this table to a comma-separated .txt file using
write.table(). - Copy and paste the content of the .txt file into Word.
- In Word,
- select the text you just pasted from the .txt file
- go to Table → Convert → Convert Text to Table…
- make sure “Commas” is selected under “Separate text at”, click OK
You’ll now have a basic table that you can format in Word. Below are three examples of how to use this process to create crosstabs, tables for summary statistics, and regression tables.
Data and Packages
Before we get started, read in a dataset on U.S. states (codebook here) into R:
states <- read.csv("states.csv")Also install and load packages dplyr, tidyr, and broom:
pkgs <- c("dplyr", "tidyr", "broom")install.packages(pkgs) #install sapply(pkgs, require, character.only = T) #load ## dplyr tidyr broom## TRUE TRUE TRUECreate a table showing the proportion of states that supported Bush in 2000, by region (South versus Non-South):
# Create tablet <- with(states, table(south, gb_win00))t <- prop.table(t, margin = 1)t #large majority of southern states supported Bush in 2000: ## gb_win00## south Bush win Gore win## Nonsouth 0.4705882 0.5294118## South 0.8750000 0.1250000# Write this table to a comma separated .txt file: write.table(t, file = "bush_south.txt", sep = ",", quote = FALSE, row.names = F)The .txt file will end up in your working directory. Now follow steps 3 and 4 in the Overview section above to create the crosstab in Word.
Here’s another example that again uses the states.csv dataset. Say we wanted to create a table with summary statistics for five of the variables in this dataset:
sumstat <- states %>% # Select and rename five variables select( `Black (%)` = blkpct, `Attend church (%)` = attend_pct, `Supported Bush in 2000 (%)` = bush00, `Supported Obama in 2008 (%)` = obama08, `Women in State Legislature (%)` = womleg ) %>% # Find the mean, st. dev., min, and max for each variable summarise_each(funs(mean, sd, min, max)) %>% # Move summary stats to columns gather(key, value, everything()) %>% separate(key, into = c("variable", "stat"), sep = "_") %>% spread(stat, value) %>% # Set order of summary statistics select(variable, mean, sd, min, max) %>% # Round all numeric variables to one decimal point mutate_each(funs(round(., 1)), -variable)sumstat## variable mean sd min max## 1 Attend church (%) 38.9 9.4 22.0 60.0## 2 Black (%) 10.3 9.7 0.4 36.8## 3 Supported Bush in 2000 (%) 50.4 8.7 31.9 67.8## 4 Supported Obama in 2008 (%) 50.5 9.5 32.5 71.8## 5 Women in State Legislature (%) 23.2 7.3 8.8 37.8# Write to .txtwrite.table(sumstat, file = "sumstats.txt", sep = ",", quote = FALSE, row.names = F)Again, the sumstats.txt file will end up in your working directory, and you can use steps 3 and 4 from the Overview section above to import this file into Word.
Exercise
Create a table of summary statistics in Word for vep04_turnout, vep08_turnout, unemploy, urban, and hispanic. The table should include the number of observations (n), mean, median, 10th percentile, and 90th percentile of each of the variables. Put the variables in the rows of the table and the summary statistics in the columns, like we did in the example above. Format your table in Word to make it look similar to this table.
Say we wanted to run three OLS models to predict state-level support for Bush in 2000, where each model adds a predictor to the preceding model. We can create a regression table with all three models like so:
m1 <- tidy(lm(bush00 ~ blkpct, states))m2 <- tidy(lm(bush00 ~ blkpct + south, data = states))m3 <- tidy(lm(bush00 ~ blkpct + south + womleg, data = states))# Note that tidy() from the broom package is used to convert each model to a data frameall_models <- rbind_list( m1 %>% mutate(model = 1), m2 %>% mutate(model = 2), m3 %>% mutate(model = 3))all_models## Source: local data frame [9 x 6]#### term estimate std.error statistic p.value model## (chr) (dbl) (dbl) (dbl) (dbl) (dbl)## 1 (Intercept) 50.92242670 1.8269042 27.873617 2.674311e-31 1## 2 blkpct -0.04645116 0.1295857 -0.358459 7.215717e-01 1## 3 (Intercept) 51.26374042 1.7392966 29.473834 5.855540e-32 2## 4 blkpct -0.35982687 0.1753724 -2.051788 4.578488e-02 2## 5 southSouth 9.04563705 3.6085132 2.506749 1.570125e-02 2## 6 (Intercept) 66.01152726 4.5218532 14.598335 6.949435e-19 3## 7 blkpct -0.41579167 0.1585744 -2.622061 1.181097e-02 3## 8 southSouth 6.03572838 3.3595138 1.796608 7.896601e-02 3## 9 womleg -0.56807999 0.1634218 -3.476157 1.121456e-03 3# Now make this data frame look more like a regression table ols_table <- all_models %>% select(-statistic, -p.value) %>% mutate_each(funs(round(., 2)), -term) %>% gather(key, value, estimate:std.error) %>% spread(model, value)ols_table## Source: local data frame [8 x 5]#### term key 1 2 3## (chr) (chr) (dbl) (dbl) (dbl)## 1 (Intercept) estimate 50.92 51.26 66.01## 2 (Intercept) std.error 1.83 1.74 4.52## 3 blkpct estimate -0.05 -0.36 -0.42## 4 blkpct std.error 0.13 0.18 0.16## 5 southSouth estimate NA 9.05 6.04## 6 southSouth std.error NA 3.61 3.36## 7 womleg estimate NA NA -0.57## 8 womleg std.error NA NA 0.16# Exportwrite.table(ols_table, file = "olstab.txt", sep = ",", quote = FALSE, row.names = F)Again, follow steps 3 and 4 from the Overview section above to import the content of the .txt file into Word.
