Floor vibration is widely recognized as an important limit state in the design of steel-framed floors, and various evaluation methods have been developed over the years. These methods range from simplified methods that are suitable for routine usage during the design phase to complex computerized or probabilistic models. The former are in common usage in the structural design community. It is critically important to structural engineers that the accuracies of the commonly used methods are known; however, a study of these methods based on a large database of recorded observations is not available in the literature. Therefore, the authors investigated the evaluation accuracy of four well-known simplified methods by comparing predicted and observed acceptability (whether or not occupants complained) of 50 floor bays in real buildings framed with W-shaped members subjected to walking excitations. The percentage of correct predictions is used to judge the accuracy of each method. It is observed that the AISC Design Guide 11 accurately predicted acceptability of floors; two methods from the SCI P354 (Simplified Method and Vibration Dose Value Method) fairly accurately predicted acceptability of floors. The HIVOSS method provided unconservative predictions. In addition, modified version of the P354 method was investigated in an attempt to improve the accuracy of these methods.