This paper opens with two stories-about the ''joking relationship'' discovered by anthropologist Radcliffe-Brown Structure and function in primitive society. New York: The Free Press (1965), and an incident in the life of Rousseau-which illustrate that sometimes a story needs to be true, and that the opinion of an outsider can be preferable to that of the insider whose story it is. Recent papers by Carter and Polkinghorne Educational Researcher, 22(1), 5-12 (1993) and International Journal of Qualitative Studies in Education, 8(1), 5-23 (1995) are analysed; it is argued that a clear and enticing plot is no indicator of a story's truth, furthermore the need to be true can hold both of ''analysis of narratives'' and ''narrative analysis''. Copyright (C) 1997 Elsevier Science Ltd.
