36 Nicola Terrenato 4. Sample Size Matters! The Paradox of Global Trends and Local Surveys Nicola Terrenato Survey archaeologists have always had to deal with a tantalizing mismatch, in terms of orders of magnitude, between their ambitions to reconstruct large-scale trends and processes and the material constraints inherent in the practice of field survey. Paradoxically, with the evolution of the discipline in the twentieth century, the contradiction has only become more acute and distressing. On the one hand, recent theorizing has brought survey archaeologists in contact with global concepts, such as world systems or core-periphery models (e.g., Stein 1999; Champion 1989), but on the other, improved field methodologies have inevitably contracted the size of survey areas. Just as we are moving down from the hundreds into the tens of square kilometers surveyed, we contradictorily yearn to transcend local issues and take part in the debate with other special- ists, and especially historians, who may not even conceive anything smaller than the territory of a city-state or an administrative region as a unit of study. More and more, survey people need a critical mass of data of considerable absolute size even to sit at the table where the big issues are under discussion. Spatial sampling seemed to offer some measure of relief for the problem; but it ended up producing a whole autonomous diatribe of its own, which is giving no sign yet of dying down and has done little to improve the wider credibility of field survey archaeo- logists. (It is enough to recall how the final discussion on sampling in Fish and Kowalewski [1990] still sounds very much like Flannery’s [1976: 131–36] classic spoof of 15 years earlier.) It is natural, thus, to see as the only hope the combin- ation of results from different surveys, as well as the improvement and standardization of the quality of their results. Only by pooling the resources of different teams, operating over a long period of time, can one hope to gather data in sufficient quantity to address the global issues that have been raised. And this, precisely, was the dream of those new archaeologists who, from the 1960s onwards, strove to codify an optimal approach to regional studies (most famously, Binford 1964; Flannery 1976). This was obviously part and parcel of the revolution in methods and theory that was being advocated at the time, but it is important to remark that field survey had a specific and prominent role in all this. Much more than excavation, survey was seen as the truly scientific approach to settle- ment archaeology, at least potentially. Its practice, in fact, could be seen as a repeatable experiment that would allow the objective testing of different strategies and approaches. Through systematic, controlled coverage of regions (or samples thereof), archaeologists would finally be able to obtain data in a standard format, thus susceptible of being analyzed statistically. Survey would provide, as it were, a ‘clinical epidemiology’ approach, which would allow the identification of global trends and would escape the inherent case-by-case variability and ad hoc decision- making of archaeological ‘surgery’ (that is, excavation). The large-scale work carried out in Mexico and in the American Southwest seemed to confirm the validity of this approach, as valley after valley was surveyed and the data consolidated in grand syntheses (Blanton 1981 et al.; Euler and Gummerman 1978). Another exciting develop- ment was the use of spatial analysis, which would finally bring archaeologists on a par with the other envied social scientists (e.g., Johnson 1977). Thus came into being a great chimera that proved as difficult to pursue as to abandon. Even future anti-positivist guru Ian Hodder was guilty of sowing a spatial wild oat in his youth (Hodder and Orton 1976). At the time Hodder and Orton’s book Spatial Analysis in Archaeology was published, however, another young British archaeologist (also destined for a remarkable career) found himself in a position to evaluate and assess many of the strongest claims and hopes that were being built on survey at the time, especially in Europe. Stephen Shennan, later to become Professor of Theoret- ical Archaeology at University College London, was wading through East Hampshire with a ‘Job Creation’ team (presumably composed of victims of early Thatcher- ite cuts). His idea was to take this opportunity to assess quantitatively the validity of field survey as a scientific data collection procedure. The results were not particularly encouraging: significant biases were being introduced by