Uso de modelos lineales generalizados en el conteo de Leptopharsa gibbicarina (Hemiptera: Tingidae) en palma de aceite

The analysis of data coming from the counting of events must consider tools that are appropriate to deal with this type of variable. A valid approach to cope with this task consists of using the Generalized Linear Model (GLM). The use of GLM allows working with density functions such as Poisson or Negative Binomial. This work presents the results of analyzing the data from an experiment carried out at the Research Station of Cenipalma located in the municipality of Barrancabermeja (Colombia). The experiment was established to evaluate, at the field level, the entomopathogenic fungus Purpureocillium lilacinum as a strategy to control the population of the lace bug, Leptopharsa gibbicarina. The experiment was carried out in two plots, in one of them, the control of L. gibbicarina with P. lilacinum was implemented as suggested by the Entomology research program of Cenipalma and, in the other plot there was no control strategy. The data were analyzed by using GLM and two different density functions (Poisson and the negative binomial). Results showed statistically significant differences between the treatments. The estimator of the maximum likelihood associated with the P. lilacinum treatment indicated that the application of the entomopathogenic fungus reduced the populations of L. gibbicarina with a rate of incidence of 0.28 times more than the control.