Solid-state nuclear track detectors are commonly used for the detection of indoor radon levels. However, despite numerous advantages, this technique still presents many unsolved problems. An important source of error is represented by the reduction in the detection efficiency due to overlapping tracks, which results in a sensible underestimation of the radon levels. This paper presents a new experimental procedure to address the effect of overlapping tracks by establishing a relationship between the detection efficiency and the number of detected tracks. Experimental data have been collected at the radon chamber of the Istituto Nazionale di Metrologia delle Radiazioni Ionizzanti (INMRI), at the ENEA centre in Casaccia, using CR-39 detectors provided by Radosys Ltd. It has been proved that the method, applied to the experimental settings of this study, gives reliable results up to 7607 kBq h m(3). Finally, the method has been validated through a Monte Carlo simulation, exploring a wide range of radon exposure.