The reference shift hypothesis of unilateral neglect holds that spatial bias in left neglect stems from a rightward deviation of patients' egocentric frame of reference (ER). Twenty five unselected right brain-damaged patients participated in a straight-ahead pointing task to assess the position of their ER (Experiment 1). A rightward ER shift emerged only in the subgroup of patients with extensive parietal lesions. In Experiment 2, we found that the position of the ER did not predict the outcome of various visuospatial neglect tests (r = 0.07 to 0.27). In Experiment 3, no significant positive correlation emerged between the ER position and visual (r = 0.26) or tactile (r = -0.48) extinction. Two further experiments examined the relationships between the ER position and patients' performance on a reaction time test of directional motor bias (Experiment 4), and on a test of response times to lateralised visual stimuli (Experiment 5). Results showed that the ER position did not predict the distribution of accuracy scores or response times in either task (Experiment 3: accuracy: r = 0.06; response times: r = 0.16; Experiment 4: accuracy: r = 0.09; response times: r = 0.04). We concluded that the position of the ER plays no crucial role in the behavioural consequences of spatial bias induced by right hemisphere lesions.