3 Example Datasets
3.1 WASH Benefits Bangladesh Study
The example data come from a study of the effect of water quality, sanitation, hand washing, and nutritional interventions on child development in rural Bangladesh (WASH Benefits Bangladesh), a cluster randomized controlled trial (Tofail et al. 2018). The study enrolled pregnant women in their first or second trimester from the rural villages of Gazipur, Kishoreganj, Mymensingh, and Tangail districts of central Bangladesh, with an average of eight women per cluster. Groups of eight geographically adjacent clusters were block randomized, using a random number generator, into six intervention groups (all of which received weekly visits from a community health promoter for the first 6 months and every 2 weeks for the next 18 months) and a double-sized control group (no intervention or health promoter visit). The six intervention groups were:
- chlorinated drinking water;
- improved sanitation;
- handwashing with soap;
- combined water, sanitation, and handwashing;
- improved nutrition through counseling and provision of lipid-based nutrient supplements; and
- combined water, sanitation, handwashing, and nutrition.
In the handbook, we concentrate on child growth (size for age) as the outcome of interest. For reference, this trial was registered with ClinicalTrials.gov under registration number NCT01590095.
library(readr) # read in data via readr::read_csv dat <- read_csv( paste0( "https://raw.githubusercontent.com/tlverse/tlverse-data/master/", "wash-benefits/washb_data.csv" ) )
For instructional purposes, we start by treating the data as independent and identically distributed (i.i.d.) random draws from a large target population. We could account for the clustering of the data (within sampled geographic units), but, we avoid these details in this handbook for the sake of clarity of illustration. Modifications of TL methodology for biased samples, repeated measures, and related complications, are readily available.
We have 28 variables measured, of which a single variable is set to
be the outcome of interest. This outcome, \(Y\), is the weight-for-height Z-score
dat); the treatment of interest, \(A\), is the randomized treatment
dat); and the adjustment set (potential baseline confounders),
\(W\), consists simply of everything else. This results in our observed data
structure being \(n\) i.i.d. copies of \(O_i = (W_i, A_i, Y_i)\), for \(i = 1, \ldots, n\).
skimr package, we can
quickly summarize the variables measured in the WASH Benefits data set:
|Number of rows||4695|
|Number of columns||28|
|Column type frequency:|
Variable type: character
Variable type: numeric
A convenient summary of the relevant variables appears above, complete with a sparkline visualizations describing the marginal characteristics of each covariate. Note that the asset variables reflect socioeconomic status of the study participants. Notice also the uniform distribution of the treatment groups (with twice as many controls) – this is, of course, by design.