Neighbour design and analysis

In most experiments, we assume that the response on each experimental unit is affected by the treatment on that unit and by inherent features of that experimental unit. However, tall sunflowers may shade shorter neighbouring varieties, or disease may spread from an untreated plot to a neighbouring plot. Neighbour designs allow for such neighbour effects.

A different departure from the usual assumptions is that responses on neighbouring plots are correlated. It turns out that neighbour designs are good for this situation too. However, the appropriate data analysis is quite different in the two situations.

TopicsSome of my publications
Neighbour effects in one dimension or around a circle
  • J.-M. Azaïs, R. A. Bailey and H. Monod: A catalogue of efficient neighbour-designs with border plots. Biometrics 49 (1993), 1252–1261.
  • R. A. Bailey: Design of experiments with edge effects and neighbour effects. In The Optimal Design of Forest Experiments and Forest Surveys (eds. K. Rennolls and G. Gertner), University of Greenwich, London, 1993, pp. 41–48.
  • R. A. Bailey, M. A. Ollis and D. A. Preece: Round-dance neighbour designs from terraces. Discrete Mathematics 266 (2003), 69–86. doi: 10.1016/S0012-365X(02)00799-9 [Maths Reviews 1991707 (2004k: 05043)]
  • R. E. L. Aldred, R. A. Bailey, Brendan D. McKay and Ian M. Wanless: Circular designs balanced for neighbours at distances one and two. Biometrika, 101 (2014), 943–956. doi: 10.1093/biomet/asu036
  • Neighbour effects in one dimension in one direction only
  • R. A. Bailey: Designs for one-sided neighbour effects. Journal of the Indian Society of Agricultural Statistics 56 (2003), 302–314. [Maths Reviews 2059549]
  • R. A. Bailey, Peter J. Cameron, Katarzyna Filipiak, Joachim Kunert and Augustyn Markiewicz: On optimality and construction of circular repeated-meaurements designs. Statistica Sinica, 27, (2017), pp.1–22. doi: 10.5705/ss.202015.0045
  • Neighbour effects in two dimensions
  • E. R. Williams and R. A. Bailey: A note on designs for neighbor configurations. Mathematical Biosciences 56 (1981), 153–154. [Maths Reviews 0627089 (82h: 92050)]
  • R. A. Bailey: Quasi-complete Latin squares: construction and randomization. Journal of the Royal Statistical Society, Series B 46 (1984), 323–334. [Maths Reviews 0781893 (86i: 62161)]
  • Neighbour effects of only one factor
  • H. Monod and R. A. Bailey: Two-factor designs balanced for the neighbour effect of one factor. Biometrika 80 (1993), 643–659. [Maths Reviews 1248028 (95e: 62087)]
  • How to randomize without destroying neighbour properties
  • R. A. Bailey: Quasi-complete Latin squares: construction and randomization. Journal of the Royal Statistical Society, Series B 46 (1984), 323–334. [Maths Reviews 0781893 (86i: 62161)]
  • R. A. Bailey: Restricted randomization for neighbour-balanced designs. Statistics and Decisions Supplement 2 (1985), 237–248. [Maths Reviews 0829496 (87e: 62073)]
  • R. A. Bailey and C. A. Rowley: Valid randomization. Proceedings of the Royal Society, Series A 410 (1987), 105–124.
  • Design and analysis when responses on neighbouring plots are correlated
  • R. A. Bailey, J.-M. Azaïs and H. Monod: Are neighbour methods preferable to analysis of variance for completely systematic designs? `Silly designs are silly!'. Biometrika 82 (1995), 655–659. [Maths Reviews 1366290 (96h: 62141)]
  • H. Monod, J.-M. Azaïs and R. A. Bailey: Valid randomisation for the first difference analysis. Australian Journal of Statistics 33 (1996), 91–106. [Maths Reviews 1410805 (97f: 62120)]
  • J.-M. Azaïs, H. Monod and R. A. Bailey: The influence of design on validity and efficiency of neighbour methods. Biometrics 54 (1998), 1374–1387.
  • Page maintained by R. A. Bailey