Soil structure describes the arrangement of soil particles in aggregates and the space between them, and is a primary attributes of soil quality. Soil aggregation is often reported as indices such as the mean weight diameter and fractal dimensions. Both mean weight diameter and fractal dimensions could lead to serious misinterpretations, and, following the principles of compositional data analysis, developed the Aitchison distance aggregation index. In this paper, we compare mean weight diameter, fractal dimensions and Aitchison distance aggregation index in order to detect potential biases and limitations. The balance aggregation index, an alternative index that avoids limitations in current indices, is also evaluated in a case study. We found that the balance aggregation index was closely correlated with the fractal dimensions (correlation = -0.99 in theoretical gridded data and -0.96 in the case study), and could be used as robust index to assess soil aggregation.