Table 6.
Multivariable modeling of well construction risk factors as related to detection probabilities and concentrations of coliforms, any fecal-associated marker, and nitrate in private wells.
| Contaminant and outcome measurement (n)a | Risk factor | Univariable model -value |
Multivariable model | |||
|---|---|---|---|---|---|---|
| Risk factor medianb | Risk factor rangeb | Coefficient or trendc,d | -valuee | |||
| Any fecal markerf detection (83) | Casing depth | 0.15 | 17.7 | 12.2–48.2 | Negative | 0.31 |
| Open interval length | 0.13 | 29.0 | 2.1–79.6 | Positive | 0.24 | |
| Bedrock depth | 0.027 | 4.6 | 0–46.6 | 0.26 | ||
| Coliforms concentration, recharge (47) | Well depth | 0.047 | 48.8 | 18.3–100.6 | Negative | 0.59 |
| Casing depth | 0.057 | 18.9 | 12.2–80.2 | None | 0.91 | |
| Groundwater depth at construction | 0.0004 | 12.2 | 1.8–36.6 | Negative | 0.0038 | |
| Well age | 0.0042 | 24 | 5–49 | 0.04 | 0.016 | |
| detection, recharge (201) | Casing length below water table | 0.040 | 8.5 | –36–58.8 | 0.13 | |
| Bedrock depth | 0.0028 | 6.4 | 0–55.2 | Negative | 0.28 | |
| concentration, recharge (124) | Well age | 0.15 | 22 | 2–80 | Positive | 0.16 |
| Bedrock depth | 0.0063 | 4.7 | 0–31.4 | Negative | 0.11 | |
| detection, no-recharge (251) | Casing depth | 0.12 | 18.9 | 6.1–126.5 | 0.01 | 0.65 |
| Casing length below water table | 0.02 | 8.8 | –36–117.3 | 0.33 | ||
| Bedrock depth | 10.1 | 0.3–54.3 | Negative | 0.07 | ||
| concentration, no-recharge (127) | Casing depth | 0.043 | 18.0 | 6.1–126.5 | 0.57 | |
| Casing length below water table | 0.054 | 5.5 | –19.8–117.3 | Negative | 0.74 | |
| Bedrock depth | 0.0019 | 6.7 | 0.3–49.4 | Negative | 0.0088 | |
Note: Univariable model -values used for selecting risk factors are included for reference; complete univariable statistics are provided in Tables S11, S12, and S13. Risk factor eligibility for inclusion in multivariable models is described in statistical methods.
in multivariable model.
Units for length and depth are meters; age is in years.
In lieu of reporting multiple coefficients for spline-represented variables we report the overall trend (positive or negative).
Interpretation of coefficient linear terms: change in ln(detection probability) or change in ln(concentration) for a unit change in the risk factor.
The composite -value is reported for spline-represented variables.
“Any fecal marker” includes all microorganisms regardless of host specificity.