Table 3. Number of elephants (observed and predicted according to the latent class model) with each test result.
| Binary Input | 2.50% | 50% | 97.50% | Mean | Observed |
| 00000 | 69 | 82 | 92 | 81 | 84 |
| 00001 | 9 | 16 | 26 | 17 | 16 |
| 00010 | 0 | 3 | 8 | 3 | 1 |
| 00011 | 18 | 24 | 29 | 24 | 26 |
| 00100 | 1 | 6 | 13 | 6 | 8 |
| 00101 | 0 | 1 | 4 | 1 | 1 |
| 00110 | 0 | 0 | 2 | 0 | 1 |
| 00111 | 0 | 2 | 6 | 2 | 0 |
| 01000 | 0 | 4 | 11 | 4 | 4 |
| 01001 | 0 | 1 | 3 | 1 | 1 |
| 01010 | 0 | 0 | 2 | 0 | 0 |
| 01011 | 0 | 1 | 5 | 2 | 2 |
| 01100 | 0 | 0 | 2 | 0 | 0 |
| 01101 | 0 | 0 | 1 | 0 | 0 |
| 01110 | 0 | 0 | 2 | 0 | 0 |
| 01111 | 0 | 2 | 8 | 3 | 2 |
| 10000 | 0 | 0 | 4 | 1 | 0 |
| 10001 | 0 | 0 | 1 | 0 | 0 |
| 10010 | 0 | 0 | 1 | 0 | 0 |
| 10011 | 0 | 1 | 3 | 1 | 1 |
| 10100 | 0 | 0 | 2 | 0 | 0 |
| 10101 | 0 | 0 | 2 | 0 | 0 |
| 10110 | 0 | 0 | 2 | 1 | 0 |
| 10111 | 1 | 4 | 8 | 4 | 4 |
| 11000 | 0 | 0 | 2 | 0 | 0 |
| 11001 | 0 | 0 | 2 | 0 | 0 |
| 11010 | 0 | 1 | 3 | 1 | 1 |
| 11011 | 0 | 3 | 8 | 3 | 2 |
| 11100 | 0 | 1 | 2 | 1 | 1 |
| 11101 | 0 | 0 | 3 | 1 | 0 |
| 11110 | 0 | 2 | 5 | 2 | 2 |
| 11111 | 12 | 19 | 25 | 19 | 22 |
Evaluating the fit of the model by comparing the observed and expected number of elephants with different combinations of tests to see if the assumptions of the substantive model in Figure 2 were satisfied. 50% refers to the median estimate while the 2.5% and 97.5% quantiles define a 95% credible interval (CI).