The conversion of reactant orbital angular momentum to product rotational angular momentum is usually discussed for heavy + heavy-light A + BC reactive collisions. We argue that this will often be largely the case also for other mass combinations due to the steric requirements of the reaction. In particular, collinearly dominated reactions (e.g., H + H2) would exhibit considerable kinematic polarization of the products. Computational results are provided for a number of reactive collisions (H + H2, O + HCI, H + O2, Li + HF), where due to the stereochemistry, the kinematic effects can be different from those due to the mass change alone. The implications of the discussion are that (i) for collinearly dominated reactions, the reaction probability (the "opacity function") can be related to the measured product rotational state distribution, (ii) the kinematic polarization of the reaction products can be more extreme than is often recognized, and (iii) surprisal analysis of product rotation can be understood in terms of kinematic constraints.