This study examines possible causes for variations in specific final energy use in new apartment buildings. The analysis is based on case studies of 22 new apartment buildings that were constructed as part of the 'Stockholm program for environmentally adapted buildings'. The buildings in the study were chosen because they share similar construction characteristics and similar energy systems but display unexpected large variations in specific energy use. Three causes were found to contribute to variations in monitored specific final energy use in the studied apartment buildings: (1) the time interval between the completion of construction work and the actual energy measurements, (2) the shape factor of the building and (3) the relative size of the common area. In addition, the buildings that participated in the Stockholm program failed to achieve the requirements for the specific final energy use, to a large extent, because of expectations based on the simulated values. The simulated specific final energy use predicted by the energy simulations were on average 19% lower than the monitored values, giving the impression that the buildings would fulfill the program's energy requirements. The reasons for the low simulated values were determined to be large uncertainties in the input data. (C) 2012 Elsevier B.V. All rights reserved.