TABLE 7.
Results of CART analyses for the Simpson index
| Root node split criteriona | Secondary split criteriona | Simpson index (mean ± SD) | No. of samples | % of samples positive for the following pathogen: |
||||
|---|---|---|---|---|---|---|---|---|
| Salmonella | Campylobacter | E. coli O157:H7 | Cryptosporidium | Giardia | ||||
| Yr = 2004, 2006-2007 | SHREVE, ≤572 | 0.209 ± 0.127 | 23 | 5 | 10 | 0 | 45 | 30 |
| SHREVE, >572 | 0.121 ± 0.069 | 181 | 13 | 32 | 0 | 47 | 27 | |
| Yr = 2005 | NA | 0.193 ± 0.104 | 102 | 11 | 10 | 0 | 52 | 12 |
a
Regression tree split criteria for terminal nodes (data groupings where no further splitting occurs) are given. Root node split criterion, the variable and condition by which all the data were divided into two nodal groupings (child nodes); secondary split criterion, the variable and condition by which the child nodes derived from the root nodal split were divided. NA, not applicable. Within-nodal data descriptions are given as Simpson indices.