Soil carbon measurements are often highly variable, which makes the detection of statistically significant differences among different land uses and management systems difficult. A common mistake is to interpret a lack of statistical significance as evidence for the absence of differences. Ad-hoc power analysis is what should be used prior to conducting a field experiment or collecting soil samples in order to ensure high power of the statistical tests. Post-hoc power analysis should be used after the data are collected and analyzed if the results of statistical tests turn out to be not significant. I will present results of simulated and actual soil carbon measurements exploring effects of variability on sample/sub-sample requirements of the experimental studies that have monitoring soil carbon as one of their objectives.