We have investigated the influence of sample size on the calibration of a vibrating sample magnetometer. As expected, when measurements and calibration are carried out on samples with the same shape the results agree within the experimental uncertainty. For two cylindrical samples of nickel, one with length and diameter 2.0 and 2.0 mm, and another with 5.3 and 1.0 mm respectively, the magnetometer signal varies with the length, in good agreement with numerical calculations.