The problem with relying on dietary surveys: sociocultural correctives to theories of dietary change in the Pacific islands

Objective: Historical dietary surveys are frequently used as the basis for theorising nutritional change and diet-related non-communicable disease emergence (DR-NCD) in the Pacific islands. However, findings from historical survey data do not always align with ethnographic evidence. The aim of this paper is to examine the extent to which the two types of evidence can lead to similar conclusions and draw out the implications for current theories of nutritional change in the Pacific islands.

Design: Dietary surveys carried out on Nauru between 1927 and 2007 are reviewed, and their methods, assumptions and researcher motivations interrogated. These are compared with ethnographic evidence documented by anthropologists during the colonial and post-colonial periods, including through participant observation and oral histories.

Setting: The Republic of Nauru, between 1900 and 2011.

Subjects: The people of Nauru.

Results: Comparing historical survey data to ethnographic evidence reveals several shortcomings of survey data. The data collected and categories used reflect researcher and institutional biases. Historical survey data are often (re-)interpreted in ways that privilege researcher inference over the lived experiences of local populations. Nutritional issues considered to be relatively recent – such as high-fat, low-fibre diets and transition to imported foods – occurred a century ago in our analysis. There is limited evidence that caloric intake overall has increased significantly over this period of time in Nauru.

Conclusions: Theories of dietary change and diet-related non-communicable disease emergence (DR-NCD), and resulting interventions, could be improved through a more holistic approach to nutritional science that incorporates sociocultural and historical ethnographic evidence about the target population and the scientists doing the research, alongside historical surveys and other quantitative evidence.