If variations and uncertainty in building acoustic measurements can be controlled, construction costs can potentially be reduced since the building will not have to be acoustically over-dimensioned. Field measurements of impact and airborne sound insulation were carried out for an industrially prefabricated cross-laminated timber (CLT) system of plate elements. The results from 18 rooms, forming three groups with respect to size, were compared to a similar study dealing with a volume based building system. Large variations were found at frequencies below 100 Hz which is crucial for the low frequency adaptation terms connected to the weighted sound insulation indices. The measurement uncertainty was controlled by analysing the repeatability, measurement direction and the time dependence of the sound source. The variations due to the measurement procedure were found to be small compared to the total variations. It was also indicated that the variations in sound insulation seem to decrease with the amount of workmanship required at the building site