Purpose: To facilitate the use of the collimator scatter correction factor, S(c), parametrization methods that relate S(c) to the field size by fitting were investigated. Materials and methods: S(c) was measured with a mini-phantom for five types of dual photon energy accelerators with energies varying between 6 and 25 MV. Using these S(c)-data six methods of parametrizing S(c) for square fields were compared, including a third-order polynomial of the natural logarithm of the field size normalized to the field size of 10 cm2. Also five methods of determining S(c) for rectangular fields were considered, including one which determines the equivalent field size by extending Sterling's method. Results: The deviations between measured and calculated S(c)-values were determined for all photon beams and methods investigated in this study. The resulting deviations of the most accurate method varied between 0.07 and 0.42% for square fields and between 0.26 and 0.79% for rectangular fields. A recommendation is given as to how to limit the number of fields for which S(c) should be measured in order to be able to accurately predict it for an arbitrary field size.