A comparison of local linear power densities measured with a gamma thermometer string (GTS) and with a fission chamber shows important discrepancies at the extremities of the assembly. The cause was revealed by a simulation of axial power distribution. These discrepancies appear to arise from inaccurate knowledge of the axial position of the GTS in the assembly. Using the method of fictitious displacement of a GTS, it was possible to reduce these discrepancies to <3.7%. One method of reducing this disadvantage could be to include the GTS in the fuel assemblies in the factory. By this way, it would be possible to eliminate passages for the string through the bottom of the core vessel.