The code in this replication package recreates the main data set and results in the paper using R and Stata. To replicate outputs, run main.do. The replicator should expect the code to run in less than 20 minutes.
The raw data used in the paper can be found in data/raw. The two main raw datasets are mda.dta, containing data from MDA trials and created by script 01, and tt.dta, contaning data from test-and-treat trials and created by script 02. The data from the Campbell Systematic Reviews was provided by the authors of that study.
The data is licensed under an Unlicense license. See LICENSE.txt for details.
The code was last run on a 10-core Intel-based laptop with Windows 10 and 16GB RAM. The entire code ran in less than 20 minutes.
- Stata code was last run with version 18. The exact versions of all community-contributed packages last used are included in
code/ado. - R code was last run with version 4.2.3. The exact versions of all packages last used are listed in
renv.lock. To install these versions, launch the R projectdeworming.RProjand runrenv::restore().
Script 07_bayesian hierarchical models.R runs simulations and a random seed is set at line 89.
All scripts used for data creation and analysis are included in code. Scripts 01-03 create or import the raw data in data/raw. Scripts 04 and 05 reformat the raw data for use with different functions. Scripts 06-13 analyze the data. See main.do for a detailed description of inputs and outputs of each script.
The code is licensed under a Creative Commons Attribution 4.0 license. See LICENSE.txt for details.
- Edit line 11 in
main.doto adjust the file path to the root directory of this repository - Edit line 33 in
main.doto adjust the file path to the R installation in your computer - Run
main.doto run all steps in sequence. - The Excel file
output/Formatted tables.xlsxreads the csv files exported by the code and formats them as in the paper. It also adds some simple calculations as indicated in theNotescolumn of the table below.
The provided code reproduces all tables and figures in the paper, except for the items marked under "Note" below.
| Figure/Table # | Program | Line Number | Output file | Note |
|---|---|---|---|---|
| Table 1 | code/10_metan.do | 2 | output/tables/table1.csv | |
| Table 2 | code/10_metan.do | 47 | output/tables/table2.csv | |
| Table 3 | code/10_metan.do | 125 | output/tables/table3.csv | Columns 7-12 use estimates from GiveWell (2017), Kristjansson et al. (2007), Kristjansson et al. (2015a), Kristjansson et al. (2016), and Neumann (2003). |
| Figure 1 | code/06_forest_plots.R | output/figures/figure1.png | ||
| Figure 2 | code/06_forest_plots.R | output/figures/figure2.png | ||
| Figure 3 | code/06_forest_plots.R | output/figures/figure3.png | ||
| Figure 4 | code/06_forest_plots.R | output/figures/figure4.png | ||
| Figure 5 | code/11_roodman_graph.R | output/figures/metaforest--.png | ||
| Table S2 | code/10_metan.do | 165 | output/tables/tableS2.csv | Calculations for panel C are done directly in Excel |
| Table S3 | code/10_metan.do | 288 | output/tables/tableS3.csv | |
| Table S4 | code/10_metan.do | 320 | output/tables/tableS4.csv | |
| Table S5 | code/10_metan.do | 365 | output/tables/tableS5.csv | |
| Table S6 | code/10_metan.do | 400 | output/tables/tableS6.csv | |
| Table S7 | This table is created in Excel using values from tables 2 and 3 | |||
| Table S8 | code/08_clean_stan_output.do | output/tables/tableS8.csv | ||
| Figure S1 | code/13_compare_to_TMSDG.R | output/figures/figureS1.png | ||
| Table F | code/12_tableF.do | output/tables/tableF.csv | ||
| Table F1 | code/09_publication_bias.R | 162 | output/tables/andrews-kasy-bias-1.csv | |
| Table F2 | code/09_publication_bias.R | 252 | output/tables/andrews-kasy-bias-2.csv | |
| Figure F1 | code/09_publication_bias.R | 48 | output/figures/figureF1.png | |
| Figure F1 | code/09_publication_bias.R | 48 | output/figures/figureF1.png |