Line bisection has long been a routine test for unilateral neglect, along with a range of tests requiring cancellation, copying or drawing. However, several studies have reported that line bisection, as classically administered, correlates relatively poorly with the other tests of neglect, to the extent that some authors have questioned its status as a valid test of neglect. In this article, we re-examine this issue, employing a novel method for administering and analysing line bisection proposed by McIntosh et al. (2005). We report that the measure of attentional bias yielded by this new method (EWB) correlates significantly more highly with cancellation, copying and drawing measures than the classical line bisection error measure in a sample of 50 right-brain damaged patients. Furthermore when EWB was combined with a second measure that emerges from the new analysis (EWS), even higher correlations were obtained. A Principal Components Analysis found that EWB loaded highly on a major factor representing neglect asymmetry, while EWS loaded on a second factor which we propose may measure overall attentional investment. Finally, we found that tests of horizontal length and size perception were related poorly to other measures of neglect in our group. We conclude that this novel approach to interpreting line bisection behaviour provides a promising way forward for understanding the nature of neglect.