Objective. To examine the factor structure of the 4- and 10-item Perceived Stress Scale (PSS) among 1 310 Mexican women participating in a prospective cancer cohort study. Materials and methods. We performed exploratory and confirmatory factor analyses in two sub-samples of the Mexican Teachers’ Cohort. We evaluated internal consistency, estimated the correlation between PSS-4 and PSS-10, and assessed their correlation with a depressive symptoms scale. Results. Two-factor models were the solutions with the best fit to the data for both PSS-4 and -10, exhibiting strong factor loadings (0.39 to 0.75) and high internal consistency (Cronbach’s alpha 0.72 and 0.83). The correlation between PSS-4 and PSS-10 was r=0.91 and the correlations of these two scales with a depressive symptoms scale were r=0.41 and r=0.46, respectively. Conclusions. PSS might be an adequate scale to assess perceived stress in this prospective cancer cohort study. PSS-4 may be advantageous due to its simplicity, low cost, and short application time in multicountry studies on stress and cancer.