There has been much effort in recent years to quantify manual handling capabilities. Four main techniques have been used to this end; biomechanical modelling; the measurement of intra-abdominal pressure; psychophysics; and metabolic/physiological criteria. The aim of this study was to compare quantitatively the data produced from the first three techniques. The comparisons were limited to bimanual, sagittal plane lifting, which of all manual handling activities has been studied the most comprehensively, except that pushing and pulling data were compared from the psychophysics and intra-abdominal pressure ('force limits') databases. It was found that the data from 'force limits' proposed weights for bimanual lifting in the sagittal plane which [corrected] are lower than those reported to be psychophysically acceptable except for lifting close to and around the shoulder. The closest agreement between the databases was for lifting from an origin above knuckle height. The 'force limits' data were found to propose weights of lift which are at a minimum when lifting with a freestyle posture from the floor whereas the psychophysical technique proposes weights which are at a maximum when lifting from the floor. The psychophysical data were found to generate compressive forces at L5/S1 according to a static sagittal plane biomechanical model about 10% in excess of the NIOSH action limit (NIOSH 1981) when lifting from the floor, although over other lifting ranges the compressive forces were less than the NIOSH action limit. Lifting the 'force limits' weights generated compressive forces which were on average 55% less than AL (range 45 to 60%) when lifting in an erect posture. The data for pushing according to the psychophysical and 'force limits' database showed good agreement, but for pulling the 'force limits' weights were considerably greater than those selected psychophysically. The implications of these findings are discussed.