The effect of deformation history (hysteresis) on transient capillary rheometric data was studied compared to conventional assumptions regarding steady state data. The factors studied were: the position instrumentation, the pressure instrumentation, entrance and exit effects, polymer melt compressibility, pressure dependence of the viscosity, and polymer melt viscous heating. Statistical analysis of variance was performed to statistically determine the sources of variance to specific degrees of confidence. The polymer melt compressibility, pressure dependence, and viscous heating were found to be statistically significant contributors of the observed variation at the 95% confidence level; the capillary length and instrumentation were not found to be significant. The results indicate that the transient behavior can vary the modeling of the apparent viscosity in a significant manner such that the model fidelity and model coefficients may vary substantially. Hence, polymer melt compressibility, pressure dependence, and viscous heating should be considered during rheological model fitting to increase model fidelity and predictive accuracy in end use. (C) 2016 Society of Plastics Engineers