The use of scale models is widely adopted as an efficacy method for determining the daylight assessment of interior areas of buildings. Scale models can be tested under real sky or sky simulators. In both cases, it is commonly accepted among the researchers that the use of scale models leads to an overestimation of experimental data with respect to real scale test rooms. Starting from this assumption, the authors have carried out an experimental analysis under real sky on a reduced scale model of a test room equipped with a window and on two scale models of a room equipped with a double light pipe in order to define the influence of the scale factor on experimental data both with traditional and innovative daylight sources. The analysis leads in particular to identify the influence of direct solar radiations on the results. Although the authors agree with the majority of researchers and confirm the common opinion that daylight performances are generally overestimated in scale model analysis, this study evidences that the presence of intense direct radiations can produce exceptions and the opposite case can be experienced. (C) 2014 Elsevier B.V. All rights reserved.