Replication Materials

This directory contains the replication data and scripts for Toni Rodon, Tom Paskhalis. "Parliamentary Power in Early Representative Assemblies: Evidence from XVII century England" published in the Journal of Historical Political Economy.

Main Script

analysis_and_write_up.R - Complete replication script that combines all analysis steps and generates all figures and tables.

Required Input Data

The script requires the following data files in the ./data/ directory (within the replication folder):

Essential Data Files

These files must be present in ./data/:

1. Parliamentary Journals:

   • commons_df.rds - Journal of the House of Commons
   • lords_df.rds - Journal of the House of Lords

2. Parliamentary Sessions & Acts:

   • parliament_sessions.csv - Parliamentary session dates and monarchs
   • parliament_royalcontrol.csv - Royal control indicators
   • parliament_acts_smith.csv - Parliamentary acts from Smith (1999)
   • parliament_acts_hoppit.csv - Parliamentary acts from Hoppit (2017)
   • parliament_acts_interregnum_bho.csv - Interregnum acts

3. Kings' Speeches:

   • kings_text.rds - Kings' speeches text

4. Members of Parliament:

   • mps_data.xlsx - MPs data by parliamentary session (for surnames analysis)

5. Constituencies:

   • constituencies_infos.csv.bz2 - Constituency information
   • constituencies_list.csv.bz2 - List of constituencies

6. Economic Data:

   • governmental_income_chandaman.csv - Governmental income data
   • national_income_obrien.csv - National income data

Helper Code Files

The script also requires these files from ./code/ directory:

• ./code/helper_functions.R - LDA fitting and analysis functions
• ./code/plotting_functions.R - Plotting utilities
• ./code/latin_stopwords.R - Latin stopwords for text preprocessing

Note: All input data and code files are in subdirectories within the replication/ folder. The script is self-contained and does not require access to parent directories.

What the Script Does

The script performs the following steps:

1. Data Loading - Loads all necessary datasets (Commons, Lords, Acts, Kings' speeches, MPs data, etc.)

2. Descriptive Analysis - Calculates summary statistics for Commons and Lords data

3. LDA Model Fitting - Fits topic models with 10 and 30 topics for both Commons and Lords journals

4. Validation & Entropy - Calculates entropy measures (parliamentary power) and combines with acts data

5. Kings' Speeches Analysis - Calculates cosine similarity between Kings' speeches and parliamentary journals

6. Surnames Analysis - Analyzes surname diversity among MPs as a measure of political competition

7. Constituency Analysis - Counts mentions of constituencies in parliamentary texts

8. Regression Analysis - Fits regression models examining the relationship between policy diversity and acts

9. Figure Generation - Generates all main manuscript figures

10. Table Generation - Generates all main manuscript tables

Running the Script

Using renv for Package Management

IMPORTANT: This project uses renv for reproducible package management. The script automatically activates the renv environment from the parent directory (../renv/) which contains all required package versions used in the original analysis.

Recommended approach:

• The script will automatically load the renv environment from the parent directory
• All package dependencies are managed through renv, ensuring reproducibility
• No need to manually install packages - renv will restore the exact versions used

If you need to restore packages manually or set up the environment for the first time:

# From the main project directory (parent of jhpe_parliament_england_xvii/)
renv::restore()

Prerequisites

The following R packages are required and managed by renv:

# Packages managed through renv (auto-installed):
# "readr", "readxl", "dplyr", "tidyr", "stringr", "purrr",
# "lubridate", "quanteda", "quanteda.textstats", "topicmodels",
# "ggplot2", "gridExtra", "modelsummary", "fixest", "stopwords"

Manual Installation (Alternative):

If you prefer to work outside of renv or encounter issues with the renv environment, you can manually install all required packages:

install.packages(c(
  "readr", "readxl", "dplyr", "tidyr", "stringr", "purrr",
  "lubridate", "quanteda", "quanteda.textstats", "topicmodels",
  "ggplot2", "gridExtra", "modelsummary", "fixest", "stopwords"
))

Note: Manual installation may result in different package versions than those used in the original analysis, which could lead to minor variations in results. For exact replication, use renv.

Execution

From R or RStudio:

# Navigate to replication directory
setwd("jhpe_parliament_england_xvii")

# Run the script (automatically activates renv from parent directory)
source("analysis_and_write_up.R")

From the command line:

cd /path/to/jhpe_parliament_england_xvii
Rscript analysis_and_write_up.R

What happens when you run the script:

• Automatically activates the renv environment from the parent directory (renv::load(".."))
• Reads input data from ./data/
• Sources helper functions from ./code/
• Writes all outputs to ./figures/, ./results/, ./tables/, and ./log/

Output

The script generates the following outputs (all within the replication/ directory):

Figures (./figures/)

• figure1.pdf/png - Overview of Commons and Lords data over time
• figure2.pdf/png - Yearly parliamentary acts by house
• figure3a.pdf/png - Scatter plot of power vs acts
• figure3b.pdf/png - Regression coefficient plot
• figure4.pdf/png - Cosine similarity between Kings' speeches and journals
• figure5.pdf/png - Surname diversity and parliamentary power

Results (./results/)

• LDA models (in ./results/lda/ subdirectory):
  • lda/commons_10/ - Commons 10-topic models by year (entropies.rds, topics.rds, yearly .rds files)
  • lda/commons_30/ - Commons 30-topic models by year
  • lda/lords_10/ - Lords 10-topic models by year
  • lda/lords_30/ - Lords 30-topic models by year
• Regression models (in ./results/ols/ subdirectory):
  • ols/acts_10_md1.rds through ols/acts_10_md6.rds - 10-topic regression models
  • ols/acts_30_md1.rds through ols/acts_30_md6.rds - 30-topic regression models
• Aggregate results:
  • acts.csv - Combined parliamentary acts data
  • xvii.csv - 17th century data with entropy measures (parliamentary power)
  • commons_kings_similarities.csv - Cosine similarity between Kings' speeches and Commons journals
  • surnames_df.csv - Surname diversity index merged with power measures

Tables (./tables/)

• table1.tex - Summary of 17th century parliaments
• table2.tex - Top words by topic for key years (1625, 1642, 1662, 1689)
• table3.tex - Regression models of parliamentary power

Intermediate Data (./data/)

In addition to input files, intermediate datasets generated during analysis:

• meta_commons.csv, meta_lords.csv - Descriptive statistics per year
• constituencies_mentions_xvii.csv - Frequency of constituency mentions
• full.csv - Complete merged dataset for regression analysis

Log (./log/)

• analysis_YYYYMMDD_HHMMSS.log - Timestamped log of the analysis run with progress and results

Computation Time

The full script may take several hours to complete due to:

• LDA model fitting (most time-consuming step)
• Large corpus processing

For testing purposes, you can modify the LDA control parameters to reduce iterations:

control = list(alpha = 5, verbose = 25L, seed = 1234,
               burnin = 100, thin = 10, iter = 200)  # Reduced iterations

Directory Structure

The replication directory structure is:

jhpe_parliament_england_xvii/
├── analysis_and_write_up.R  ← Main replication script
├── README.md
├── code/                    ← Helper functions
│   ├── helper_functions.R
│   ├── plotting_functions.R
│   └── latin_stopwords.R
├── data/                    ← Input data & intermediate files
│   ├── commons_df.rds
│   ├── lords_df.rds
│   ├── parliament_sessions.csv
│   ├── parliament_acts_*.csv
│   ├── kings_text.rds
│   ├── mps_data.xlsx
│   ├── governmental_income_chandaman.csv
│   ├── national_income_obrien.csv
│   ├── constituencies_infos.csv.bz2
│   ├── constituencies_list.csv.bz2
│   ├── meta_commons.csv     ← Generated
│   ├── meta_lords.csv       ← Generated
│   ├── constituencies_mentions_xvii.csv  ← Generated
│   └── full.csv             ← Generated
├── figures/                 ← Generated figures (PDF/PNG)
│   ├── figure1.pdf/png
│   ├── figure2.pdf/png
│   ├── figure3a.pdf/png
│   ├── figure3b.pdf/png
│   ├── figure4.pdf/png
│   └── figure5.pdf/png
├── results/                 ← Generated results
│   ├── lda/                 ← LDA topic models
│   │   ├── commons_10/
│   │   ├── commons_30/
│   │   ├── lords_10/
│   │   └── lords_30/
│   ├── ols/                 ← Regression models
│   │   ├── acts_10_md1.rds through acts_10_md6.rds
│   │   └── acts_30_md1.rds through acts_30_md6.rds
│   ├── acts.csv
│   ├── xvii.csv
│   ├── commons_kings_similarities.csv
│   └── surnames_df.csv
├── tables/                  ← Generated LaTeX tables
│   ├── table1.tex
│   ├── table2.tex
│   └── table3.tex
└── log/                     ← Analysis logs
    └── analysis_YYYYMMDD_HHMMSS.log

Notes

• Package Management: The project uses renv for reproducible package management. The script automatically loads the environment from the parent directory (renv::load(".."))
• All paths are relative to the replication/ directory
• The script is self-contained and does not require access to parent directories (except for renv activation)
• All input data files are in ./data/
• Helper functions are in ./code/
• All output directories (figures/, results/, results/lda/, results/ols/, tables/, log/) are created automatically if they don't exist
• Progress and results are logged to both console and timestamped log file in ./log/
• The original data files in ./data/ are NOT modified (except for generated intermediate files)
• Log files are timestamped (e.g., analysis_20260302_143025.log) to preserve multiple runs

Questions or Issues

If you encounter any issues running the script, check:

1. All required packages are installed
2. The working directory is correctly set
3. All data files are in their expected locations
4. You have write permissions for output directories