In recent years Rutter's Malaise Inventory has been used increasingly to measure levels of stress among those who care for people with dependency needs. Although there has been an element of dispute in the literature about the unidimensional nature of the scale, later work (Bebbington and Quine 1987) tends to suggest that Malaise Inventory scores can be used as a single internally consistent measure of stress. In this study consideration is given to the results of a national sample survey of 527 members of the Association of Carers (now Carers: National Association) which suggests the existence of two valid dimensions to the Malaise Inventory across a wider range of carer groups. Implications for future development and administration of the scale as a measure of stress are discussed.