Non-orthogonal Multiple Access (NOMA) is a new generation multiple access technique in which multiple users simultaneously share the same frequency band. In the implementation of NOMA, the choice of user pairs that will use the same frequency is a determining factor for the NOMA superiority condition due to Successive Interference Cancellation (SIC) mechanism on the receiver. However, performance of NOMA depends on SIC mechanism which relies on the order of the user in a NOMA cluster. In the literature, it has been shown that utilizing distance based order of the users for user pairing outperforms over the random pairing and an analytical boundary distance is derived for NOMA superiority condition for a path loss based channel model. In this study, the path loss based NOMA superiority condition respect to the Orthogonal Multiple Access (OMA) was investigated in terms of the spectral efficiencies for Rayleigh channel model in a cellular uplink and the results were compared with an analytical boundary value derived for a path loss based channel model in the literature. In addition, the effect of power ratio of the users on the NOMA superiority condition is investigated and accuracy of the derived analytical model is validated with numerical results for both channel models. The results reveal that the NOMA superiority condition based on distance changes depend on the assumed channel model, but assuming the channel as Rayleigh fading or path loss based channel does not make a significant difference. In addition, it is observed that the accuracy of the derived NOMA superiority condition changes with the power ratios for both channel models and the gap between numerical and analytical results is larger in Rayleigh channel compared to path loss based channel model.