Characterizing the interrelationships of commonly used water measures

There are numerous commonly used measures of household water access that capture various dimensions of the household water experience, but it is unclear if individual variables are complementary, redundant, or uniquely informative. We characterized how measures of household water access are statistically and conceptually interrelated. Using data collected from 861 households in Beira, Mozambique we used correlation analyses to characterize relationships between measures of household safe water access, and factor analyses to characterize interrelationships and constructs being measured. In a subset of 75 households, we assessed the parameterization and classification of water quality variables, and how variable definitions impact correlations between variables. Correlation and factor analysis showed that the component variables that make up the WHO/UNICEF Joint Monitoring Programme (JMP) service ladder measured distinct elements related to water access, adequacy, and quality, and that these component variables offer unique information that may not be portrayed when just presenting the JMP water service ladder by itself. User satisfaction was broadly correlated with many variables, even with microbiological water quality, suggesting “satisfaction” captures a breadth of safe household water access. In the factor analysis, the user satisfaction loaded in distinct constructs, indicating potential to reduce these multiple satisfaction variables. How water quality variables were classified (e.g., binary, raw, or log-transformed) strongly contributed to how they correlated with each other. There was little correlation between E. coli and various pathogens in water. Our research highlights the complexities of relationships between household water access variables, and characterization of these interrelationships inform how to use and interpret measures in research and practice.