Neglecting the real joint behaviour in frame analysis may result in unrealistic predictions of the response and reliability of steel frames. The reliability of the prediction of main joint properties according to the component method (Eurocode 3-Part 1.8) still remains open to further investigation. The first step toward the solution is to compare the theoretical expressions given in EN 1993-1-8 and the experimental results. With that goal in mind six nominally the same, but really different specimens of welded beam-to-column joints subjected to static load were tested. The specimens present a combination of nominally identical structural elements produced in different European mills. This paper provides these tests, as well as their detailed evaulation and interpretation. All three joint structural properties (rotational stiffness, moment resistance and rotation capacity) have been considered. Four models for determining the plastic resistance out of experimental Mj-${phi}$ curves have been applied. The results that have been discussed in detail, point to the fact that EN 1993-1-8 underestimates the real structural properties of the tested type of joint, as well as to the conclusion that detailed research of this problem needs to be conducted using the probabilistic reliability methods.