This study deals with the statistical analysis of a data set of residential buildings to find simplified correlations for the assessment of the energy demand for space heating. The analyzed data come from a sample of 50 multi-family residential buildings of the local social housing company, built in different periods between the end of the 19th century and the end of the last century. Relationships are searched between the age of the building and its main geometric and thermophysical properties: the shape ratio, the volume, the opaque and glazed surfaces, their transmittances, the external and internal temperatures. These properties are suitably combined to find out statistical relations. Moreover, starting from standard deterministic equations, this study revises them and simplifies the most involved terms. This simplification is carried out making a statistical analysis of the energy quantities that affect the performance of a building. in such a way the solution of the energy balance is simplified, making the calculation easier and faster. Results are compared with those calculated using the rigorous Standard procedure, showing a good agreement. Moreover, the assessed energy demand is compared with the actual metered energy consumption. The present statistical model estimates in a simple and prompt way the energy demand for space heating of residential buildings, in particular the existing ones. It may represent a useful tool for their energy certification. (C) 2008 Elsevier B.V. All rights reserved.