This study investigates the scale effects on the experimental shear strength of earthen walls, a critical parameter influencing the seismic performance of adobe and rammed-earth (RE) buildings. Recognized for their historical significance and sustainable construction practices, earthen structures require a comprehensive understanding of their mechanical behavior under shear loads to ensure effective design and preservation. This research compiles data from over 120 in-plane shear wall tests (adobe and RE), nearly 20 direct shear tests from the scientific and technical literature, and new cyclic direct shear tests performed on large cubic specimens (300 mm side length) made from the same material as a previously tested two-story RE wall. Based on the findings, this study recommends a minimum specimen cross-sectional area of 0.5 m(2) for reliable shear strength testing of earthen walls in structural laboratories. This recommendation aims to prevent the unconservative overestimation of shear strength commonly observed in smaller specimens, including direct shear tests. Furthermore, the Mohr-Coulomb failure criterion outlined in the AIS-610 Colombian standard is validated as a conservative lower bound for all compiled shear strength data. Cyclic direct shear tests on nine 300 mm cubic specimens produced a Mohr-Coulomb envelope with an apparent cohesion of 0.0715 MPa and a slope of 0.66, whereas the full-scale two-story wall (5.95 x 6.20 x 0.65 m) constructed with the same material exhibited a much lower cohesion of 0.0139 MPa and a slope of 0.26. The analysis reveals significant scale effects, as small-scale specimens consistently overestimate shear strength due to their inability to capture macro-structural behaviors such as compaction layer interactions, construction joint weaknesses, and stress redistributions. Based on the analysis of the compiled data, the novelty of this study lies in defining a strength reduction factor for direct shear tests (3.4-3.8 for rammed earth, similar to 3.0 for adobe) to align with full-scale wall behavior, as well as establishing a minimum specimen size (>= 0.5 m(2)) for reliable in-plane shear testing of earthen walls, ensuring accurate structural assessments of shear strength. This study provides a first approach to the shear behavior of unstabilized earth. To expand its application, future research should explore how the scale of specimens with different stabilizers affects their shear strength.