2.5. Statistical analyses
Independent-sample t-tests or covariance analyses were used to determine the
differences between the comparison groups, and non-parametric analyses were
used for data with skew distributions. P-values were calculated using Chi-square
test for categorical data. Spearman rank correlation analysis was used to evaluate
the relationships between BLLs and some physical indices. BLLs or BCLs were taken
as dependent variables and the investigated 25 items with self-designed
questionnaire as independent variables; the correlations between them were also
performed with Spearman rank correlation analysis. The data of blood lead or
cadmium concentration were divided into two groups: high-level group and lowlevel
group taking 10 mg/dL as the boundary for BLLs and 2 mg/L as the boundary for
BCLs, respectively. To explore some risk or protective factors for high BLLs and
BCLs, logistic regression analyses were used for further discussion. All the analyses
were conducted with SPSS statistical software version 13.0 (SPSS, Inc., Chicago, IL,
USA). The level of significance was set at 0.05.