Abstract: For several decades, it has been universally accepted that the health of population improves with higher education levels, measured in terms of years of formal schooling or adult literacy rates. The most common health indicator used to demonstrate the link is the reduced rate of child mortality, and the effect persists even after controlling for the positive effects that education has on income and employment (Hobcraft 1993). Higher education levels, however, have a much more pronounced effect on child survival and health when accompanied by improved access to primary healthcare services, especially maternal and infant healthcare (Caldwell & Caldwell 1995). It is also well known that Aboriginal peoples and Torres Strait Islanders receive much less formal schooling and have much lower levels of literacy than non-Indigenous people (DEST 2005). Nevertheless, until the Cooperative Research Centre for Aboriginal and Tropical Health (CRCATH) embarked on its Health and Education Research Program in 1997, only one study (Gray 1988) had examined the education–health link hypothesis in relation to Indigenous peoples in Australia. This paper begins with a brief review of international research, which demonstrates the link between education and health, followed by an account of the quantitative, qualitative and theoretical studies carried out by the CRCATH. We acknowledge at the outset that the interpretation of empirical data on education–health links is highly contested. Sometimes, as we show, the data itself is ambiguous, or incomplete, especially the Australian data. But it is more difficult to arrive at a shared theoretical understanding of the two major variables: health and education. This is not as straightforward as simply examining their empirical ‘markers’—such as child survival rates, years of schooling and literacy levels. Why? Because both health and education are cultural ‘artefacts’ or ‘constructs’, which mean different things to different people at different times. The confusion is magnified by the assumption of both health and education professionals that the paradigms of their own field transfer in a straightforward way to the other. The confusion becomes almost overwhelming when a cross-cultural dimension is added, because both Indigenous and non-Indigenous peoples have such diverse experiences of education from which to draw an understanding of it. The second part of this paper is a preliminary report into the progress we have made trying to achieve greater clarity about the meanings of the education–health link, and what action should flow from this in the specific context of Indigenous peoples’ health. Our methodology was to undertake a series of semi-structured dialogues between two academically trained researchers—one an educationalist, the other a health professional—and a small number of Indigenous health leaders, including our co-author and project leader.1 These dialogues helped us to clarify meaning, and also addressed a key finding of the social determinants literature, namely that power and control are at the heart of health inequalities. The third part of the paper draws together these first two elements—the review of the research literature and the dialogue with the health leaders—to build some illustrative models of the way in which education can be better understood as both a determining factor in the reproduction of health inequalities, and as an active intervention into overcoming them. In the final section, we suggest some future directions for research and program development.