Water repellency is a relevant topic in soil-science research due to its effects on vegetation growth, occurrence of surface runoff, infiltration, and erosion. Different methods have been adapted for the assessment of soil wettability by measuring the contact angle (CA), like the capillary-rise method (CRM), where the liquid penetration dynamics into a dry sample is analyzed. However, questions related to sample preparation and the use of a suitable reference liquid in order to improve the reproducibility in such heterogeneous materials are still open. Different methods use ethanol as a reference liquid to quantify the degree of water repellency, like the molarity of ethanol droplets (MED), whereas other methods (CRM) suggest in addition n-hexane as reference. To date, no generally accepted protocol has been invented to apply the CRM to soil particles. By using model porous materials with defined and stable levels of water repellency (silt, sand, and glass beads), CA results were compared for different initial settings of the sample. The main objective was to prove the CRM as a reliable and reproducible method to characterize soil wettability and to specify general guidelines for its application for granular materials in terms of sample size, sample-packing procedure, and reference liquid. Results showed that a sample weight of ≈ 2 g has a relatively lower CA variation between replications. The packing procedure showed erratic results in CA, proving to be a critical factor in reproducibility. A uniform criterion of samples packing is recommended regarding application of CRM in soils. Regarding the reference liquid, n-hexane should be preferred instead of ethanol for dynamic CA determination, because ethanol increased the tendency of CRM to overestimate angles due to dynamic effects, especially in finer-textured materials (i.e., silt).