New statistical tests determining the likely reality of a connection between stone circles and astronomical phenomena
Gail Michele Higginbottom, Roger Clay
Abstract Ruggles, in his major 1984 research project in Scotland chose to dismiss ‘from further consideration any on-site indications involving stone rings’ (C. Ruggles, Megalithic astronomy 1984: p. 61.). This was because astronomical hypotheses involving sightings across stone rings are dependent upon other variables, such as whether or not the ‘site fits a particular geometrical construction’ and therefore did not fit with his general research questions at the time. Further, no statistical test had yet been determined to deal with the associated probability issues connected to investigations of looking at orientations within a single circle. Such a test involves separate determinations of the likelihood of various statistical errors, including errors in orientation due to archaeological alignment uncertainties and the uncertainty of which part of the astronomical phenomenon was of interest as it crossed the horizon (e.g. when it first touches the horizon or its final disappearance; thus testing the intentions of the builders). Also, circles with large numbers of outer stones increased the likelihood of hitting an astronomical object by chance, increasing the statistical errors, and therefore reducing the level of probability at which one can reject the null hypothesis. No-one since then has tried to develop a purpose built statistic to deal with these issues until our own attempt. We have now created a test to empirically assess the possible astronomical associations of stone circles, which takes in account these concerns. This opens the way forward for more considered methodologies involving orientation issues connected to monuments, the landscape and/or astronomy and is thus this innovative statistic assists in ensuring more secure results for landscape archaeological research.
