[en] Purpose: Many articles have been published on the measurement of the dose to points outside the primary beam, often called the peripheral dose (PD), for instance, to the gonads, for specific treatment machines and/or techniques. We investigated the possibilities for developing a generalized method based on the data from several publications. Methods and Materials: The data from several publications were recalculated for a reference situation, then averaged, and the frequency distributions around the mean were determined. Published data were available for ⁶⁰Co, 4, 6, 8, and 10 MV, and 18 to 25 MV for a large variety of treatment machines. Furthermore, an analysis of possible corrections for depth dependence, field elongation, irregularly shaped fields, wedges, and shielding blocks was carried out. Results: The frequency distributions of all published PD values for square fields for photon energies of 4 MV to 25 MV showed a standard deviation of 33%. The PD values of ⁶⁰Co are significantly different with a standard deviation of 25%. A difference in the leakage radiation between cobalt machines and linear accelerators can possibly explain this difference, especially for large distances, where leakage radiation predominates. Taking the uncertainty of the risk factors into consideration, we conclude that the use of average values is justified. Although statistically not significant, the peripheral dose appears to be dependent on photon energy with a minimum around 6 MV. Conclusions: It is possible to estimate the peripheral dose for photon energies of 4 MV to 25 MV with an accuracy of ± 33%; for ⁶⁰Co, the accuracy is even better. The variation of the PD between different treatment machines is so small that it is justified to use average PD values, irrespective of the treatment machine