A total of 364 samples of vegetable- and meat-based meals were collected at three processing steps: step I, preparation of raw ingredients; step II, processing and cooking; and step III, finished meals. Microbiological quality was evaluated by using data for the prevalence and concentration of the aerobic plate counts, total coliforms (TC), fecal coliforms (FC), and Escherichia coli. The data were analyzed for differences between cafeterias, seasons, raw materials, and processing steps. Fourteen (15.2%) of the 92 finished meal samples were microbiologically unsatisfactory. Neither cafeteria nor season was significantly associated with microbiological quality (P > 0.05). However, the type of raw ingredients and processing steps were significantly associated with differences in microbiological quality. Vegetable-based meals had higher TC concentrations than meat-based meals because salad and seasoned and fermented vegetables are not cooked, unlike heat-processed meat products. Microbial counts tended to decrease through the processing steps, and E. coli, which could only be enumerated on uncooked chicken breast (1.6 log CFU/g) and sliced pork (2.6 log CFU/g), was totally eliminated by boiling and roasting. However, the presence of FC was not completely eliminated, even by cooking, and so this group of organisms should be considered as an important indicator of hygienic meal preparation in cafeterias. Although pathogenic E. coli was not isolated in this study, continuous microbiological monitoring of composite foods served in cafeterias should be performed as the presence of TC and FC in finished meals indicates the potential for contamination by pathogenic E. coli.