A critical evaluation of step strain flow experiments on entangled, linear polymer liquids is performed. Roughly one-half of the published shear stress relaxation modulus data for these systems are consistent with the predictions of the well-known tube model. A model of step strain flow experiments is developed to determine whether the remaining published data, which are qualitatively different from tube model predictions, are simply artifacts caused by slip, an imperfect strain step strain history, or transducer compliance. Modeling results suggest that these factors are capable of producing the types of behavior observed in experiments deemed anomalous. New criteria based on the retraction time tau(R), or longest Rouse relaxation time, for avoiding anomalies caused by imperfect strain step strain history and transducer compliance are proposed. These simple criteria are, in a majority of cases, found to be capable of predicting the type of observed stress relaxation behavior for 60 published step strain experiments on entangled, linear polymer liquids. (C) 2005 The Society of Rheology.