Ecosystem and soil scientists frequently use whole soil carbon:nitrogen (C : N) ratios to estimate the rate of N mineralization from decomposition of soil organic matter (SOM). However, SOM is actually composed of several pools and ignoring this heterogeneity leads to incorrect estimations since the smaller pools, which are usually the most active, can be masked by the larger pools. In this paper, we add new evidence against the use of C : N ratios of the whole soil: we show that a disturbance can decrease the whole-soil C : N ratio and yet increase C : N ratios of all SOM pools. This curious numerical response, known as Simpson's paradox, casts doubt on the meaning of frequently reported whole-soil C : N changes following a disturbance, and challenges the N mineralization estimates derived from whole-soil C : N ratio or single-pool modeling approaches. Whole-soil C : N ratio may not only hide features of the labile SOM pool, but also obscure changes of the large recalcitrant SOM pools which determine long-term N availability.