Cracking is one of the predominant distresses occurring in flexible pavements, especially in old pavements that were rehabilitated with an asphalt overlay. In such cases asphalt mixtures should be designed to ensure high resistance to reflective cracking because new asphalt layers are exposed to existing cracks of the old pavement. The nature of these cracks can be various (transverse, longitudinal as well as crazy cracking). One factor that minimizes this type of distress is the proper mix design process, which should involve selection of specific bitumen binder and mineral mix gradation. However, still there is no universally adopted laboratory test method that would allow to clearly assess resistance of asphalt mixtures to reflective cracking. This paper describes the usage of one of the devices developed to test asphalt mixtures in terms of such distress – Texas Overlay Tester. For this test, samples prepared in laboratory conditions (i.e. compacted with the use of Superpave Gyratory Compactor) as well as obtained in the field (by core drilling) can be used. The results are obtained not only quickly and easily, but also with sufficient repeatability. The described method characterizes both crack initiation and crack propagation properties of asphalt mixtures. In this work one type of mineral mixture was tested with 4 different types of bitumen (one neat bitumen, two ordinary polymer-modified and one polymer-modified with high polymer content). For selected cases extra additives (rubber and loose fibres) were also tested. In total, six asphalt mixtures were tested. A ranking of the used binders was created on the basis of the results in order to conclude which bitumen would ensure the best performance characteristics in terms of reflective cracking. The results have clearly shown that deliberate choice of the binder used in the asphalt mixture for the overlay will significantly improve its reflective cracking resistance or even fatigue resistance.