We critically review experiments that reported that the pulse yield of CsI depends on the state of polarization of the incident X-rays. We show that such experiments, and possibly other polarization dependent experiments, are susceptible to serious systemic errors, particularly at small grazing incidence anges. We present a model that accounts for the different systematic errors, and measurements that demonstrate the validity of the model. We analyze the particular experimental parameters of some of the pulse yield experiments and conclude that their results were affected by the systematic errors and hence have to be retracted. According to a set of measurements designed to avoid these systematic errors the pulse yield of CsI does not depend on the polarization state of the incident X-rays.