Abstract
In light of the United Nations Millennium Development Goals, a multivariate linear model was built to estimate the percent of each countries’ population having reasonable access to clean drinking water. The statistical program
ARC [1] was used to determine which adequately transformed variables were most significant in predicting water accessibility. In order of importance, these variables were: gross national income per capita, population density, domestic water withdrawal per capita, and the corruption perceptions index. Interestingly, the water resource variables used here were statistically orthogonal to the water accessibility model. In other words, the problem of water accessibility does not depend on water supply or its scarcity, at a country-wide level. These findings were also confirmed geographically. This model is a step forward, and invites academics to improve the model and developing countries to improve their reported data.
