Producing literature reviews of complex evidence for policymaking questions is a challenging methodological area. There are several established and emerging approaches to such reviews, but unanswered questions remain, especially around how to begin to make sense of large data sets drawn from heterogeneous sources. Drawing on Kuhn's notion of scientific paradigms, we developed a new method-meta-narrative review-for sorting and interpreting the 1024 sources identified in our exploratory searches. We took as our initial unit of analysis the unfolding 'storyline' of a research tradition over time. We mapped these storylines by using both electronic and manual tracking to trace the influence of seminal theoretical and empirical work on subsequent research within a tradition. We then drew variously on the different storylines to build up a rich picture of our field of study. We identified 13 key meta-narratives from literatures as disparate as rural sociology, clinical epidemiology, marketing and organisational studies. Researchers in different traditions had conceptualised, explained and investigated diffusion of innovations differently and had used different criteria for judging the quality of empirical work. Moreover, they told very different over-arching stories of the progress of their research. Within each tradition, accounts of research depicted human characters emplotted in a story of (in the early stages) pioneering endeavour and (later) systematic puzzle-solving, variously embellished with scientific dramas, surprises and 'twists in the plot'. By first separating out, and then drawing together, these different meta-narratives, we produced a synthesis that embraced the many complexities and ambiguities of 'diffusion of innovations' in an organisational setting. We were able to make sense of seemingly contradictory data by systematically exposing and exploring tensions between research paradigms as set out in their over-arching storylines. In some traditions, scientific revolutions were identifiable in which breakaway researchers had abandoned the prevailing paradigm and introduced a new set of concepts, theories and empirical methods. We concluded that meta-narrative review adds value to the synthesis of heterogeneous bodies of literature, in which different groups of scientists have conceptualised and investigated the 'same' problem in different ways and produced seemingly contradictory findings. Its contribution to the mixed economy of methods for the systematic review of complex evidence should be explored further.