DESIGNING CONSUMER TRIALS BALANCED FOR FIRST AND HIGHER ORDERS OF CARRY-OVER EFFECT WHEN ONLY A SUBSET OF K SAMPLES FROM T MAY BE TESTED

The problem of balancing out carry-over effects of preceding samples i n consumer trials when each consumer only receives k out of a possible t products is discussed. For large trials an all-possible-combination s approach can balance all higher order effects. Mutually orthogonal l atin squares can achieve this with much smaller numbers of consumers w hen t is a prime number or an integer power of a prime In general subs ets-designs balanced just for first-order carry-over may easily be con structed either by unfolding a full design or by combining them with i ncomplete block designs. Tables of designs from 4 to 11 products are 9 given, general methods are described that give routes to designs with higher numbers of samples.