Designing sampling programmes and experiments requires great care and logical thought to ensure that appropriate data are obtained, valid statistical procedures are possible and robust interpretations are possible. Examples are given of problems of sampling through time (where overlapping time-scales influence ecological processes). Problems with spatial scales of processes and long time courses of ecological change are noted. Particular issues are presented in analyses of preferences for microhabitat or among different types of prey. Experimental designs for analyses of complex indirect interactions and for transplantations in behavioural studies are also explored.
