Understanding pesticide risks requires characterizing pesticide exposure within the environment in a manner that can be broadly generalized across widely varied conditions of use. The coupled processes of sorption and soil degradation are especially important for understanding the potential environmental exposure of pesticides. The data obtained from degradation studies are inherently variable and, when limited in extent, lend uncertainty to exposure characterization and risk assessment. Pesticide decline in soils reflects dynamically coupled processes of sorption and degradation that add complexity to the treatment of soil biodegradation data from a kinetic perspective. Additional complexity arises from study design limitations that may not fully account for the decline in microbial activity of test systems, or that may be inadequate for considerations of all potential dissipation routes for a given pesticide. Accordingly, kinetic treatment of data must accommodate a variety of differing approaches starting with very simple assumptions as to reaction dynamics and extending to more involved treatments if warranted by the available experimental data. Selection of the appropriate kinetic model to describe pesticide degradation should rely on statistical evaluation of the data fit to ensure that the models used are not overparameterized. Recognizing the effects of experimental conditions and methods for kinetic treatment of degradation data is critical for making appropriate comparisons among pesticide biodegradation data sets. Assessment of variability in soil half-life among soils is uncertain because for many pesticides the data on soil degradation rate are limited to one or two soils. Reasonable upper-bound estimates of soil half-life are necessary in risk assessment so that estimated environmental concentrations can be developed from exposure models. Thus, an understanding of the variable and uncertain distribution of soil half-lives in the environment is necessary to estimate bounding values. Statistical evaluation of measures of central tendency for multisoil kinetic studies shows that geometric means better represent the distribution in soil half-lives than do the arithmetic or harmonic means. Estimates of upper-bound soil half-life values based on the upper 90% confidence bound on the geometric mean tend to accurately represent the upper bound when pesticide degradation rate is biologically driven but appear to overestimate the upper bound when there is extensive coupling of biodegradation with sorptive processes. The limited data available comparing distribution in pesticide soil half-lives between multisoil laboratory studies and multilocation field studies suggest that the probability density functions are similar. Thus, upper-bound estimates of pesticide half-life determined from laboratory studies conservatively represent pesticide biodegradation in the field environment for the purposes of exposure and risk assessment. International guidelines and approaches used for interpretations of soil biodegradation reflect many common elements, but differ in how the source and nature of variability in soil kinetic data are considered. Harmonization of approaches for the use of soil biodegradation data will improve the interpretative power of these data for the purposes of exposure and risk assessment.