Estimates of peak spinal compression in manual materials handling were compared between a state-of-the-art laboratory technique and a method applicable at the worksite. Nine experienced masons performed seven simulated tasks in a mock-up in the laboratory and nine matched masons were studied during actual performance of the same tasks at the worksite. From kinematic and kinetic data obtained in the laboratory, compression forces on the L5S1 joint were calculated. In addition, compression forces were estimated from the horizontal and vertical position of the blocks handled relative to the subject measured at the worksite. Comparison of group-averaged values showed that the worksite method underestimated peak compression by about 20%. Rank ordering of tasks for back load was, however, consistent between methods, supporting validity of the worksite method to compare different tasks or to determine the effects of ergonomic interventions with regard to mechanical back load. STATEMENT OF RELEVANCE: This study validated a method that can be used by ergonomists to determine the effects of (characteristics of) manual materials handling tasks on back load at the worksite.