Regardless of the simplicity of the test layout, interpretation of the tensile test results may be inadequate. Typically, a test of the reinforced concrete tie provides measurements of average deformations of the internal reinforcing bar and the concrete surface. The experimental evidence, however, often contradicts with the general assumption of the similarity of mean strains of the reinforcement and concrete. Analysis of the sources influencing the scatter of the experimental results motivated this investigation of the representativeness of the specimens. In this study, the representativeness is understood as a property of a tie to isolate the investigated parameters (e.g., the deformation components) from other uncontrolled effects typical of test with limited sample sizes. This study aims at determining the representative geometry of the test samples that would enable the reduction of the end effect. A consistent procedure is proposed for identifying the cracking parameters of the ties with different testing layouts.