A recurring issue in any discussion of forest carbon inventories is the measurement and monitoring of soil carbon stocks. While soil contains a large reservoir of carbon, the composition and size of this pool can be difficult to characterize. The high spatial variability of soil carbon presents an ongoing challenge when planning and implementing soil carbon surveys. This project measured soil organic carbon stocks in treated and untreated plots from seven long-term forest management experiments in six northern states; treatments vary across sites and range from clearcuts to light thinnings. This study has two main objectives, the first of which is to investigate the effects of thinning and harvesting treatments on mineral soil carbon pools in these long-term studies. The second is to obtain carbon stock estimates from the untreated plots, to serve as benchmarks and provide an opportunity to search for underlying patterns or trends on which to base generalized estimates. This paper presents results from this regional study, with a special emphasis on the variability within and between treatments at the different study sites. Also discussed are the tradeoffs between different sampling approaches for soil carbon studies.