Williams Latin square designs are used widely in consumer research to plan
complete block experiments. It is demonstrated how these designs are a subs
et of a wider class of cyclic designs and how alternative designs, for odd
numbers of samples p, are preferable both to obtain positional and carry-ov
er balance at twice as many Values of n the number of consumers. A conseque
nce is that the new designs give better robustness to the variable number o
f consumers who turn up at a central location test. An algorithm is describ
ed that is used to sample from the wider class of designs and is sufficient
ly fast to be used for the interactive construction of designs within a com
puter package. Other specialist designs for situations where there may be f
actorial structure among the consumers are also described. Tables are given
that list the first rows of cyclic designs generated by the algorithm, the
se give pairs, as well as triples of balanced Latin squares for between fiv
e and 19 products. Practical advice is given on how they may be used to con
struct suitable presentation order designs. (C) 2001 Elsevier Science Ltd.
All rights reserved.