Skip to content

Data for difference in difference analysis on Bioko Island

Notifications You must be signed in to change notification settings

d-hergott/DID-public

Repository files navigation

DID-public

Data and code for difference in difference analysis on Bioko Island. Results published in Nature Communications

Hergott, D.E.B., Guerra, C.A., García, G.A. et al. Impact of six-month COVID-19 travel moratorium on Plasmodium falciparum prevalence on Bioko Island, Equatorial Guinea. Nat Commun 15, 8285 (2024). https://doi.org/10.1038/s41467-024-52638-2

Overview

This repository contains a data set used to run the DID travel analysis as published in Hergott, D. et al.

bioko_did_analytic_dataset.csv has blinded, individual level data that was fed into the models. Bioko_DID_travel_code.R contains the code for running the main DID analysis, land use sensitivity analysis, and generating Table 2 and Table 3 in the paper.

bioko_did_all_years_paralle_trends_data.csv has blinded, individual level data for 2015-2020 used to test the parallel trends assumption. Bioko_DID_parallel_trends_test.R contains the code for running the test of parallel trends and visualizing data.

Data Dictionary

The following variables are available in the analytic data set.

Variable Description Values
stratum Stratum assignment for EA. 1 = stratum 1 (low population density/high residual transmission); 2= stratum 2 (high population density/low residual transmission)
gender Gender of respondent 1= male; 2= female
falc_pos Binary variable indicating if respondent had Pfalicarum infection by RDT 1= positive; 0=negative
year Numeric Year 2019; 2020
year.b Transformed binary year variable for DID analysis 0=2019; 1= 2020
inbefore7 Binary variable indicating if respondent went into their home for the evening before 7pm night before survey 1=yes; 0=no
spry_perc Proportion of households in EA sprayed during spray round 0-1
trav Binary variable indicating if EA is in high or low travel group 0= low travel; 1= high travel
trv_prev Smoothed prevalence of proportion of respondents reporting traveling to mainland in past 8 weeks (2015-2018 average) 0-1
wt Survey weight reflecting sampling scheme of MIS 1=4.17 (5%); 2=20.8 (24%)
travelledisland Binary variable indicating if individual spent the night in a district other than their own (but within Island) in past 8 weeks 1=yes; 0=no
treat Binary indicator asking if individual sought treatment when sick. Only asked of those who were sick in past 14 days 1=yes; 0=no
prop_treat EA level variable, calculate as number of sick individual who sought treatment 0-1
aircon Binary level variable indicating if household has airconditioning 1=yes; 0=no
land_use_change Flag for EAs with known land use changes in study period 1= land use change EA
parent_key Household identifier used in creating survey object
psuId Enumeration area identifier used in creating survey object

Code Questions

For any questions on code or data, please contact us using Github or email: dhergott@uw.edu.

About

Data for difference in difference analysis on Bioko Island

Resources

Stars

Watchers

Forks

Packages

No packages published

Languages