In vivo radiobioassay is integral to many health physics and radiological protection programs dealing with internal exposures. The Bottle Manikin Absorber (BOMAB) physical phantom has been widely used for whole-body counting calibrations. However, the shape of BOMAB phantoms-a collection of plastic, cylindrical shells which contain no bones or internal organs-does not represent realistic human anatomy. Furthermore, workers who come in contact with radioactive materials have rather different body shape and size. To date, there is a lack of understanding about how the counting efficiency would change when the calibrated counter is applied to a worker with complicated internal organs or tissues. This paper presents a study on various in vivo counting efficiencies obtained from Monte Carlo simulations of two BOMAB phantoms and three tomographic image-based models (VIP-Man, NORMAN and CNMAN) for a scenario involving homogeneous whole-body radioactivity contamination. The results reveal that a phantom's counting efficiency is strongly dependent on the shape and size of a phantom. Contrary to what was expected, it was found that only small differences in efficiency were observed when the density and material composition of all internal organs and tissues of the tomographic phantoms were changed to water. The results of this study indicate that BOMAB phantoms with appropriately adjusted size and shape can be sufficient for whole-body counting calibrations when the internal contamination is homogeneous.