Abstract
THIS PAPER SETS OUT GUIDELINES BASED ON THE AUTHORS' EXPERIENCES REGARDING THE SATISFACTORY DESIGN OF STATED PREFERENCE EXPERIMENTS. IN PARTICULAR, THE IMPORTANCE OF INCORPORATING APPROPRIATE 'BOUNDARY VALUES' FOR THE TRADE-OFFS BETWEEN ATTRIBUTES IS STRESSED. THIS PROCESS SHOULD TAKE ACCOUNT OF POTENTIAL TASTE VARIATION AMONGST RESPONDENTS. THE HYPOTHETICAL CHOICES SHOULD BE MEANINGFUL AND SHOULD NOT INVOLVE VERY SMALL DIFFERENCES IN ATTRIBUTE LEVELS. DESIGNS PREPARED IN THE WAY DESCRIBED SHOULD THEN BE SUBJECTED TO A SIMULATION EXERCISE USING SYNTHETIC DATA COVERING A RANGE OF POSSIBLE VALUES FOR THE POPULATION PARAMETERS.(Author/publisher).