Single photon transmission (SPT) measurements offer a new approach for the determination of attenuation correction factors (ACF) in PET. A major drawback of this method is the high fraction of scattered photons in the transmission sinogram resulting in a marked underestimation of the ACFs. It was, therefore, the aim of the present work, to evaluate a scatter correction algorithm proposed by C. Watson by means of Monte Carlo simulations. SPT measurements with a Cs-137 point source were simulated for a whole-body PET scanner (ECAT EXACT HR+) in both the 2D and 3D mode. To examine the scatter fraction (SF) in the transmission data, the detected photons were classified as unscattered or scattered. The simulated data were used to determine (i) the spatial distribution of the SFs, (ii) an ACF sinogram from all detected events (ACFtot) and (iii) from the unscattered events only (ACFunscattered), and (iv) an ACFcor = (ACFtot)1 + kappa sinogram corrected according to the Watson algorithm. In addition, density images were reconstructed in order to quantitatively evaluate linear attenuation coefficients. A high correlation was found between the SF and the ACFtot sinograms. For the cylinder and the EEC phantom, similar correction factors kappa were estimated. The determined values resulted in an accurate scatter correction in both the 2D and 3D mode. The algorithm proposed by Watson allows an accurate correction of scattered radiation in SPT measurements. The correction factor kappa can by determined experimentally using simple phantoms and then applied to more complex objects. SPT measurements should be performed in the 3D mode, in order to increase the total numb of counts and/or to reduce the measurement time.