We tested the hypothesis that coffee systems with organic management have higher spider diversity by comparing a control (rainforest area) and two coffee systems, one with organic and the other with conventional management. Spiders were sampled every two weeks over three months during the dry season and three months during the rainy season in 2000. Spider alpha diversity was analyzed using Shannon and Simpson indices. We also used the Cody index for beta diversity and cluster analysis for analyzing changes in species abundance hierarchies. 2261 individuals were collected (including juveniles and adults) representing 20 families, 56 genera and 97 species. In most cases the alpha diversity indices showed no relation between management gradient and spider diversity. When compared across seasons, spider diversity differed significantly only in organic management. Species turnover among the three sites (Cody index) was highest between the two coffee farms but not so clearly in the dry vs. rainy season; the conventional management shared the fewest species with the forest. Cluster analysis showed changes in abundance hierarchy related to management type. Our results did not support the proposed hypothesis of a direct positive correlation between management gradient and alpha spider diversity. In contrast, beta diversity showed that management and seasons influenced species composition.