Analysis of Factors Affecting the Arsenic Contamination in Toenails Using Loglinear Models

Abstract

Arsenic contamination in human is possibly affecting one’s body from several hazard factors. Public health and epidemiology works concern and monitor to this issue. One way to tackle this problem is to investigate several effecting factors that may associated with Arsenic cumulatively contamination in toenails. This research aims to study effective factorss on Arsenic contamination in toenails data obtained from website http://lib.stat.cmu.edu (8 March, 2010), on “Toenails Samples as an Indicator of Drinking Water Arsenic Exposure”. The factors consist of 6 categorical variables: Age is categorized into 3 levels, similarly, for Sex is categorized into 2 levels, Drinkwater into 5 levels, Cookwater into 5 levels, Arswater into 2 levels, and Arsnails into 2 levels. Statistical analyses based on Fisher’s exact test and loglinear modeling are performed using SAS version 9.1. The results under Fisher’s exact test and those from log-linear modeling are concluded that for Fisher’s tests, each Age, Sex, Cookwater, Drinkwater, is not associated with Arsnails under the level of significance at 0.05. However, it is found that the Arswater is associated with Arsnails significantly (p-value = 0.026). The analysis using loglinear models reveals that there is significantly associated between the Arsnails and Arswater at α = 0.05. The estimated loglinear model, log(_ij) =  + _i ^arswater + _j ^arsnails + _ij^{arsnails*arswater} has adequate of fit (P = 0.999), of which the predicted values (_ij), the corresponding probabilities (_ij), and odds ratios can be obtained and further studied.