Cone calorimeter sample size sensitivity and the influence of sample size on the results were studied in order to investigate the possibility to use smaller samples than prescribed by the standard. The influence of the position of the sample on the sample holder of the decreased samples was also studied. Samples of birch plywood, industrial two-component polyurethane adhesive and polypropylene wax were tested. The results from standard size sample experiments, 100 cm 2, and decreased sample size, 50 and 25 cm2, were compared, in order to investigate the sample size influence on the results. The experiments were performed using an external heat flux of 50kW/ m2. The results show differences between the sample types. The measured heat release rates for plywood and polyurethane adhesive indicate that the sample size is possible to down-scale, while the polypropylene wax results show the opposite. All sample types and sizes show constant results for the effective heat of combustion, indicating down-scalability. For in-house testing and screening of materials, downscaling can be used for most sample types. However, results from experiments carried out with decreased sample size should not be compared to results from experiments executed according to the standard procedure. Copyright © 2011 John Wiley & Sons, Ltd.