The aim of this work was to investigate the effect of patient and cohort size on the overall uncertainty associated with dose audit using radiography of the abdomen as the exemplar. Water equivalent diameter Dw was used as the surrogate for patient size and its distribution (σ(Dw)) was used to quantify the effect of sample size. The more precise the kerma area product calibration, the more patients are required in the cohort to have the same impact on the overall uncertainty. Patient sample sizes of 300–400 will result in expanded uncertainties approaching the theoretical limit of double the measurement uncertainty when audits are performed with instruments having measurement uncertainties equal to ±7%, ±10% or ±12.5%. By way of example, for a field instrument with a measurement uncertainty of ±10%, a minimum sample size of 350 is required to achieve a total expanded uncertainty of ±21%. In the case of instruments with associated measurement uncertainty of ±3.5%, patient sample sizes of 300–400 will result in expanded uncertainties of approximately ±10%. From review of the literature and comparison with the results obtained here, it is conjectured that for radiographic dose audits of all parts of the trunk the contribution to overall uncertainty due to patient and sample size could be predicted using an indicative value for σ(Dw) of 3.4 where local data is not available.