Various population indices are widely used to monitor relative population size of many pest and game species to aid their management. However, information on the level of uncertainty associated with their estimates is rarely available. Here we explore sampling and systematic error associated with the index of re-opened burrow entrances which is used extensively in central Europe for surveying common vole populations. We found that relative sampling errors were enormous for low-density populations, attaining almost 400%. However, in high-density populations and for large sample sizes, the relative error dropped below 10% and the estimate is quite reliable. The dispersion of burrow entrances became more clumped at low population densities enhancing further the sampling variation. We demonstrated that the index is related to population size in a non-linear fashion, overestimating the population change at high densities. Consequently, population dynamics of the common vole described by the untransformed burrow index appear more variable than they are in reality.