The paper presents shear lag parameters for beam-to-column connections in steel box piers. Previous researches have analyzed beam-to-column connections in steel piers using a shear lag parameter ${eta}_o$ obtained from a simple beam model, which is not based on a reasonable design assumption. Instead, the current paper proposes a cantilever beam model and has proved the effectiveness through theoretical and experimental studies. The paper examines the inaccuracy of the previous researches by estimating the effective width, the width-span length ratio L/b, and the sectional area ratio S of a cantilever beam. Two different shear lag parameters are defined using the cantilever model and the results are compared each other. The first type of shear lag parameter ${eta}_c$ of a cantilever beam is derived using additional moments from various stress distribution functions while the other shear lag parameter ${eta}_{eff}$ of a cantilever beam is defined based on the concept of the effective width. An evaluation method for shear lag stresses has been investigated by comparing analytical stresses with test results. Through the study, it could be observed that the shear lag parameter ${eta}_{eff}$ agrees with ${eta}_c$ obtained from the $2^{nd}$ order stress distribution function. Also, it could be observed that the shear lag parameter ${eta}_c$ using the $4^{th}$ order stress distribution function almost converges to the upper bound of test results.