Anthropology, University of Wisconsin-Milwaukee
Friday Afternoon, 301D
The invasive African green vervet monkey (Chlorocebus aethops) has pestered the agricultural system of St. Kitts since the monkeys were brought to the island with the slave trade over 300 years ago. This is the first study to systematically monitor monkey crop damage patterns. Crop damage data were collected on one-third of the registered farms in St. Kitts (n = 65, randomly selected) within half-acre grid cells (level 1) and for the entire farm (level 2) for 12 sequential months. I present a model that can predict a farm’s probability of incurring crop damage by monkeys. To develop the model, the data were randomly split into a training set (80%) and a testing set (20%). A binomial (damage or no damage) hierarchical generalized linear model (HGLM) was fit to the training data. Variable selection was performed via backward selection and a receiver operating characteristic (ROC) curve was used to determine a cutoff value to make predictions. Crop preference is the greatest predictor of crop damage (b = 0.259, p = 4.91e-10), followed by distance to water (b = -0.0105, p = 0.0026), number of neighbors (b = -0.451, p = 0.00288), distance to forest (b = -0.000549, p = 0.0782), and whether or not it is the mango season (b = -0.654, p = 0.0794). Predictions are made using a cutoff associated with achieving close to 80% for both sensitivity and specificity (0.0351). This model will be used to establish which farms should be slated for protection assistance.
This research was funded in part by the University of Wisconsin-Milwaukee's Golda Meir Library Scholar Award and the University of Wisconsin-Milwaukee's Dissertation Fellowship.