This paper presents a novel analytical framework to assess soil quality, incorporating bacterial community structure data derived from PCR-DGGE analysis into a principal component analysis, along with a total of 23 other physical, chemical, and biotic soil variables. This framework was applied to a slash-and-burn agricultural system in the Atlantic forest region of Rio de Janeiro. Cropped fields had significantly reduced soil aggregation indices and organic carbon contents, as well as increased impacts on their soil bacterial community structure, relative to forest fragments. The results demonstrate the applicability of the soil quality analytical framework proposed, and reveal that traditional shifting cultivation practices in the area studied are compatible with conservation of soil quality, provided that the fallow plots are maintained for periods longer than 5 years. © 2007 Elsevier B.V. All rights reserved.