Skip to main content

Table 2 Summary of the model selection procedure for finding the best-fit logistic model that predicts membership in a hotspot district using a data driven procedure where the least significant variable was dropped from the model regardless of the hypothesized relationship

From: Associations between Water, Sanitation and Hygiene (WASH) and trachoma clustering at aggregate spatial scales, Amhara, Ethiopia

ModelVariables analyzedModel fit
CleanLatrineH2O < 30iH2OEducationHC FacilityAICLRT
1−X**+X ns−X*−X ns−X ns−X ns51.3929.94
2−X**+X ns−X**−X ns−X ns65.5930.37
3−X**+X ns−X*−X ns−X ns50.6628.68
4−X**+X ns−X**−X ns63.9130.06
5−X**−X*−X ns−X ns−X ns49.8629.48
6−X**−X**−X ns−X ns48.8728.47
7−X**−X**−X ns63.6928.27
8−X**+X ns−X**−X ns−X ns52.4526.88
9−X**+X ns−X**−X ns52.8424.49
10−X**−X**−X ns50.8424.49
11−X**−X**−X ns−X ns50.7326.60
  1. Symbols: X, variable tested in model; –, variable not included in model; −, negative association; +, positive association; ns, not significant
  2. Notes: Variables: Clean (children who presented to examiners without nasal and ocular discharge); Latrine (household has a latrine of any type); H2O < 30 (the household has access to a water source within 30 minutes of the home); iH2O (the household has access to an improved water source where improved is defined using the sustainable development goals criteria for an improved source); Education (the percentage of households that reported having attended any type of formal education), HC Facility (the presence of any type of health facility within the surveyed village). Model fit: AIC (Akaikeʼs information criterion), LRT (Likelihood ratio test)
  3. *P ≤ 0.05, **P ≤ 0.01