Table 5 Multivariable weighted stepwise linear regression models of subject‐specific means of log transformed styrene and SO levels in air (weights are the frequencies of subject‐specific repeated measurements of styrene and SO, respectively).
| Response variable (n = 325) | Overall r2 | Predictor variables | Parameter estimates | Δr2 | 95% CI |
|---|---|---|---|---|---|
| ln(Styrene in air, ppm) | 0.70 | Intercept | 1.37 | 1.01 to 1.72 | |
| Laminator | 1.77 | 0.46 | 1.58 to 1.96 | ||
| RV production | 1.41 | 0.21 | 1.21 to 1.62 | ||
| Years on job | −0.024 | 0.02 | −0.041 to −0.008 | ||
| Age (years) | −0.012 | 0.01 | −0.020 to −0.003 | ||
| Grinder/sander | 0.286 | 0.01 | 0.033 to 0.538 | ||
| ln(SO in air, ppb) | 0.63 | Intercept | 1.64 | 1.26 to 2.01 | |
| RV production | 1.43 | 0.21 | 1.22 to 1.65 | ||
| Laminator | 1.40 | 0.18 | 1.20 to 1.59 | ||
| Truck manufacture | 0.985 | 0.14 | 0.773 to 1.20 | ||
| Grinder/sander | 1.08 | 0.08 | 0.824 to 1.34 | ||
| Age (years) | −0.012 | 0.02 | −0.020 to −0.003 | ||
| Years on job | −0.020 | 0.01 | −0.037 to −0.003 |