With almost 30 years behind it, additive manufacturing (AM) continues to grow. An uptick in interest in the mid 2000s has caused the number of available printers and materials to rapidly increase. Despite this interest, increased adoption of AM is limited by lack of expertise and poor part reliability, both problems that may be addressed with improved fundamental knowledge of the systems and materials. One challenge in material extrusion (MatEx) AM is strain after annealing. Annealing is a popular method for removing residual stress, but MatEx parts warp in difficult to predict ways when annealed. Study of this irreversible thermal strain (ITε) has demonstrated layer thickness dependence, but that is just one print parameter. We studied ITε is tied to raster angle, material choice, and print temperature. Our work has shown ITε over 40% with dependence on all studied variables. Using raster angle and layer thickness dependence and single road annealing studies, we demonstrated ITε directionality aligns with road direction. Applying micromechanical modeling to this system also shows that changing raster angle only changes strain direction, not strain magnitude. With this it is possible to predict ITε for a part, which may enable annealing of MatEx parts. ITε was also observed to increase with decreasing print temperature. Improving our understanding of this phenomena will not only improve MatEx part annealing, but also enable future shape memory applications with commodity polymers such as ABS.