<p>Owing to their high load-bearing capacity and ease of assembly, tensile-bolted joints are widely used in applications ranging from conventional wooden houses to mass timber buildings. However, the current design method contains several critical gaps that may compromise structural integrity, especially for mass timber structures under cyclic loadings: (i) disregarding cyclic effects, (ii) equating compressive yielding with splitting, and (iii) applying the embedment stiffness formula for dowel-type fasteners. Therefore, the aim of this study is to address these gaps through cyclic bending tests. All tests were conducted on column-base connections using glulam members of the same grade. Experimental results showed three failure modes of timber: compressive yielding, bending failure, and splitting failure. In specimens with longer bolt anchorage lengths, the most ductile compressive yielding was observed. The load-bearing capacity determined by compressive yielding could be predicted by the bearing strength of the material. However, regarding the bending failure, the experimental values were smaller than those calculated by the design method. This result could be attributed to both stress concentration around the specimen and the effects of cyclic loading; the latter reduced the modulus of rupture. In the case of splitting failure, although the failure moments for compressive yielding and splitting were similar, the splitting failure was more brittle. Moreover, the specimen that failed in splitting had dimensions identical to that of the specimen that failed in compressive yielding. The rotational stiffness was approximately twice the calculated values for all specimens. This discrepancy can be attributed to the underestimation of embedment stiffness beneath the washers and in the column end grain, as the current calculation is based on dowel-type fasteners. These findings emphasize the need to revise the design method to more accurately account for embedment stiffness in such joints.</p>