One of the problems encountered in assessing the process dynamics of composting is the considerable variation observed between experimental replicates. The major objective of this study was to use ANOVA to assess the statistical differences between mixing treatments in bench-scale reactors. Two sets of mixing experiments were performed with 30-litre reactors using intervals between mixing of 0, 24, 96, and 192 h. Using ANOVA, mixing was found to increase the time required to achieve maximum temperatures, to prolong the time spent at elevated temperatures and to extend the period of high rates of 0, uptake. ANOVA for spatial gradients indicated a significant influence of mixing on the formation of moisture gradients; however it had little influence on temperature gradients. This study underscores the importance of improving experimental reproducibility in order to reach quantitative conclusions. (C) 2000 Elsevier Science Ltd. All rights reserved.