All statistical analyses were performed with Stat View (version 5.0,
SAS Institute) and SPSS (version 18.0, SPSS Inc.). Before statistical analysis,
one-half of concentration of the respective limit of detection was
substituted for those values below the limit of detection. All data were
tested for goodness of fit to a normal distribution with a Kolmogorov–
Smirnov's one sample test. Concentrations of water arsenic and urinary
arsenic compounds were log-transformed because these data did not
show normal distributions. Regional differences in arsenic concentrations
in water and urine samples were assessed by analysis of variance
(ANOVA), followed by Tukey–Kramer post hoc tests. Effects of sex, age,
and arsenic exposure level (based on concentrations of arsenic in filtered
groundwater: all subjects were divided into two groups, b50 μg/L
and ≥50 μg/L) on arsenic concentration in urine and onmethylation capacities
(MMA/IA and DMA/MMA) were examined by multi-way
ANOVA. A linear regression analysis was used to measure the strength
of the association between variables. In this study, a p value of less
than 0.05 was considered to indicate statistical significance.