Despite numerous studies (e.g., Fewtrell et al. 2005, Sobsey 2002, Quick et al. 2002, and Clasen et al. 2006) establishing the fact that household level safe water behaviors improve in-house drinking water quality and reduce diarrheal illness, there remain important gaps in our knowledge about the determinants of safe in-home drinking water. One of these gaps is the role of households' perceptions of source water quality in producing safe in-home water. Households' safe water, sanitation, and hygiene (WSH) practices depend on perceived water quality since actual chemical and microbial water quality is unobservable to the households. Households rely on their water's observable properties (e.g., taste, odor, clarity, stains on cloths or teeth, corrosion of pipes or storage tanks), as well as their knowledge, attitudes, and beliefs, to form perceptions about water quality (e.g., about their water's role in household disease) and determine behaviors. Understanding these perceptions and their link to behaviors is central to understanding households' behaviors and how they might be influenced through commercial or social marketing campaigns focused on adoption of point-of-use water treatment products.The objective of this study is to examine the link between water quality perceptions and their relationships to objective measures of source water quality, self reports of household water treatment practices, and measures of in-home drinking water quality. We build on Patil and Pattanayak's (2007) empirical study of whether unsafe water treatment, handling and storage behaviors are associated with higher levels of microbial contamination of household drinking water. They use data on microbial contamination of household drinking water (measured by the concentration of total coliforms and E. coli), microbial contamination of source water, and pointof-use household water treatment, storage, and handling behaviors from 5,900 households in rural Maharashtra, India. Using an index of households' self-reported behaviors (including safe water treatment, handling, and storage behaviors) they find that when households add one safe water behavior, they can reduce the concentration of total coliforms in in-home water by more than 25 percent. They also find that the concentration of total coliforms in in-home water is 40 percent lower when households use piped water, though the improvement of other types of water sources appears to have a weaker effect on in-home water quality. Finally, they find that households who perceive source water to be “safe” are less likely to engage in safe water practices, by using a single indicator of water quality perceptions – whether or not the respondent perceived their drinking water source to be “safe.”We extend their study in two ways. First, we examine how a more complete set of water quality perceptions are associated with source water quality, safe water management behaviors, and inhome drinking water quality. In particular, we will examine the respondents' perceptions about clarity, odor, and taste as well as their perceptions about the reliability and overall safety of the source. By using a more detailed set of water quality perception indicators, we can examine the link between these perceptions and their role in triggering safe water behaviors.Second, we will add data collected in rural and urban areas of Andhra Pradesh to the Maharashtra data analyzed by Patil and Pattanayak to develop a dataset of about 6,900 households and nearly 3,000 water samples. As with the Maharashtra data, these additional datasets include data on total coliforms and E. coli in source water samples and in-home water samples, survey data on source water quality perceptions, and survey data on safe WSH behaviors. Table 1 summarizes the average microbial contamination for the Maharashtra and Andhra Pradesh samples. These data represent a broad range of WSH, socioeconomic, and environmental conditions and offer an opportunity to conduct robust data analyses of the relationships between source water quality, water quality perceptions, safe water practices, and in-home water quality. Table 1.
| Number of households | Number per 100mL in Source Water Sample | Number per 100mL in Household Water Sample |
Maharashtra | | | |
E. Coli | 5946 | 23 (5) | 17 (4) |
Total Coliform | | 106 (87) | 169 (68) |
Andhra Pradesh | | | |
E. Coli | 1370 | 2 | 1 |
Total Coliform | | 29 | 32 |
Using these data, the study addresses the following research questions:(1) What is the relationship between actual (chemical and microbial) water quality and perceived water quality?; and(2) What is the relationship between households' water quality perceptions and safe water handling behaviors?.These analyses suggest that, in most cases, actual water quality is not correlated with households' perceptions of water quality. When perceived water quality and actual water quality are correlated, they move in opposite directions. On the other hand, water quality perceptions are correlated with household water treatment. In combination, these results suggest that many households are treating water that is not microbiologically contaminated.Water quality perceptions are influenced by several household characteristics, including socioeconomic status, awareness of the water-health link, and perceptions of community environmental sanitation.