Both sets of data are imperfect, and each with their strengths and limitations. Perhaps more importantly, the two reports used different criteria for claiming effects. Our study used accepted scientific criteria of applying statistical tests for significance.
Using these criteria the favourable effects reported in the NHS Health Scotland Report would be considered statistically non-significant, similar to our own findings.
We believe it is useful to have more than one evaluation of any policy and to use different methods. It is, however, overly optimistic to conclude, as does Mark Robinson, that the data in the NHS Health Scotland Report provide unambiguous evidence that the ban on multi-buys is effective in reducing alcohol consumption.
Overall, both studies show that the ban on multi-buy alone is not likely to achieve substantial changes in alcohol purchasing and, in turn, consumption. More encompassing regulation would be needed to achieve this.
Faculty of Medicine and Health Sciences
University of East Anglia