Brain structures are subjected to changes caused by genetic, environmental factors or disease and gender, age related reasons. Different methods have been developed to capture these changes, but their agreement over the same data may significantly vary. These variations are attributed to the methodological differences of the employed methods. Popular tools for measuring brain atrophy are SIENA and SIENAX that have been compared in previous studies but no work has carefully weeded out the possible bias and confounding that can be easily introduced to such analyses. The present work tackles the problem of bias, confounding and random variation factors that are inserted in any volumetric analysis methods from the selection of the input data. Evidence showed that normalized brain volume and brain volume change can be used to characterize group differences between healthy and non-healthy subjects in cases where it is not possible to circumvent the studied confounds, which can be introduced in any similar case-control study.