The importance of accounting for larval detectability in mosquito habitat-association studies

Malaria Journal

Table 2 Mean ± standard deviation of the posterior distribution (with 95 % credible intervals) for coefficients fitted in the full presence-detection mixture model

For each coefficient the proportion of the posterior distribution that lies above (or below) zero is also shown as the ‘effect probability’: this is the probability that the effect of the parameter on larval presence or detection is in the direction specified by the sign in front of the coefficient (i.e., complete certainty = 1; complete uncertainty = 0.5). For example, there is a 98.2 % probability that water depth has a negative effect on larval presence and a 99.9 % probability that water temperature has a positive effect on detection. See Table 1 for definition of parameters. See Additional file 4: Table S4 for coefficient estimates and effect probabilities when terms with high overlap with zero are dropped from the model

ISSN: 1475-2875