AbstractTime-dependent flexural deformations for composite and noncomposite prestressed concrete beams and steel–concrete composite beams, which are subject to various loading and unloading events, are predicted based on several methods with varying levels of complexity. The considered prediction methods range from computational methods that provide curvatures at various points along the span as a function of time to widely used multiplier-based methods. The impact of various levels of analytical simplification on beam flexural deformation history is quantified. Pre-erection camber predictions based on various methods are compared with measured pre-erection camber values for 105 noncomposite prestressed concrete beams. Similarly, predictions for the full flexural deformation history of six composite prestressed concrete and six steel beams are compared with test data reported in the literature. All computational methods considered result in similar predictions of pre-erection camber in prestressed concrete beams despite fundamental differences in the methods and despite gross simplifications in creep behavior used in some methods. However, when long-term flexural deformations at service and prediction of beam rebound are considered, the effect of simplifying assumptions for creep behavior becomes more significant. The quantification of such differences in predictions may inform decisions during the design phase regarding which method should be employed for a particular application.