A study was conducted in order to replicate a scale which was previously reported for assessing the feeding difficulty of elderly patients with dementia. The clinical significance of the scale was also investigated. It was demonstrated that the scale was replicable in two separate nursing units. Also, it was demonstrated that, according to criteria such as its correlation with levels of nursing intervention, for example physical help, and with an indicator of feeding difficulty, namely leaving food on the plate at the end of a meal, the scale was clinically significant. The consequences and limitations of the present investigation are discussed and recommendations for future lines of enquiry are suggested. Specifically, the possibility of analysing data from such studies by the use of multivariate statistical techniques is raised.