From preliminary samples of the coffee berry borer (Hypothenemus hampeiFerr.) in Chiapas, Mexico, various methods of calculating the number of samples needed for assessing the success of a parasitoid introduction scheme were reviewed, including those based on binomial, distribution‐free and Taylor's power law methods. The final choice of a method depends on the area to be sampled and time of year. Because of limitations on time for sampling, the high variance of the preliminary samples and the necessity of evaluating parasitoid impact over a wide area, it seems unlikely to be able to detect a significant drop of less than 20% in the borer population.