Multilingual Moral Political Dictionary (mMPD)

Kristina Bakkær Simonsen & Tobias Widmann 2023

Repo containing code and dictionary files for the multilingual moral political dictionary measuring moral language in political text in 6 different languages. Details are described in the following article:

Simonsen, K. B., & Widmann, T. (2023). When Do Political Parties Moralize? A Cross-National Study of the Strategic Use of Moral Language in Political Communication on Immigration. OSF Preprints. https://doi.org/10.31219/osf.io/hxd5a

Please start by reading this article which contains information about the creation and performance of the dictionary. The dictionary files are free to use for academic research. In case you use one or multiple of the dictionary files, please always cite the article above.

Languages included in the mMPD are: Danish, Dutch, English, German, Spanish, and Swedish.

In order to obtain all necessary files, start by downloading this repo. The folder contains all dictionary files and a R-script to apply the dictionary files. Code is shown below.

Code to apply dictionary

The mMPD is provided in YAML format and can be applied via the quanteda package.

# Packages
library(quanteda)
library(tidyverse)

# Set working directory to the downloaded folder called "mMPD-main"
setwd("./mMPD-main")

# Load Dictionary Files

#danish
mMPD_da <- dictionary(file = "./dictionary_files/mMPD_da.yml",
                      format = "YAML")
#german
mMPD_de <- dictionary(file = "./dictionary_files/mMPD_de.yml",
                  format = "YAML")
#english
mMPD_en <- dictionary(file = "./dictionary_files/mMPD_en.yml",
                      format = "YAML")
#spanish
mMPD_es <- dictionary(file = "./dictionary_files/mMPD_es.yml",
                      format = "YAML")
#dutch
mMPD_nl <- dictionary(file = "./dictionary_files/mMPD_nl.yml",
                      format = "YAML")
#swedish
mMPD_sv <- dictionary(file = "./dictionary_files/mMPD_sv.yml",
                     format = "YAML")


#  Choose language file
dic <- dictionary(mMPD_en) #run this line for English dictionary
dic <- dictionary(mMPD_de) #run this line for German dictionary
dic <- dictionary(mMPD_da) #run this line for Danish dictionary
dic <- dictionary(mMPD_sv) #run this line for Swedish dictionary
dic <- dictionary(mMPD_nl) #run this line for Dutch dictionary
dic <- dictionary(mMPD_es) #run this line for Spanish dictionary


# Create application function
apply_dic <- function(data){
  #Create a corpus from your data frame
  corp <- corpus(data)
  
  #Tokenize corpus and pre-process (remove punctuations, numbers, and urls)
  toks <- tokens(corp, remove_punct = TRUE, remove_numbers = TRUE, remove_url = TRUE)
  
  #Create DFM 
  terms_dfm <- dfm(toks)
  
  #Apply dictionary
  dict_dfm_results <- dfm_lookup(terms_dfm, dic)
  
  #Convert results back to data frame
  results_df <- cbind(data, convert(dict_dfm_results, to = 'data.frame'))
  
  #Assign length to each documents
  results_df$terms <- ntoken(terms_dfm)

  return(results_df)
}

#Now you can use the function on your data; simply enter a data frame with a column called "text" including the text data
results <- apply_dic(cc_english_prepared)

#Take a look at the resulting dataframe
glimpse(results)