This paper describes the result of long-term crack-inducing experiment on half-scaled four RC wall specimens. Compared to conventional crack control joints, the proposed crack control joints with crack-inducing bars could contribute to inducing shrinkage cracks within the joints more uniformly. As a result, the maximum crack width in the joints with crack-inducing bars was smaller than that in the conventional joints. A crack width estimation method proposed by Ohno et al. could accurately estimate the maximum crack widths of the specimens.