Measurements involving responses on a left-right scales are common in market research practice. This paper investigates natural right and left-handed tendencies in ordering stimuli, and discusses the issues they raise in the use of preference and intensity ratings using such scales. In a first experiment, a series of six studies identified that in a preference rating task participants' natural tendency was to place preferred objects on the left in a 'bench scaling' task. Intensity measurement showed no such tendency. However, when participants were asked to rate objects for the presence or absence of an intensity related attribute the left- handed tendency did occur. The extent of this may be due to a preference for relative absence or presence. A second experiment investigated how this left-handed tendency affects the use of scales involving left-right responses. Four products were rated on a line scale for intention to purchase, with the 'buy' label on the right. Despite a visual example and verbal and written instructions there were cases where scaling was reversed, so that ratings were the opposite of what participants intended. This depended on the experimenter, indicating the importance of context on error rate. These experiments suggest that data from left-right scales methods should be checked for reversal errors, perhaps especially if the preferred term is placed on the right. The results also illustrate that intensity ratings may be unpredictably 'contaminated' by preference factors. Theoretical arguments for the use of behaviourally-based preference tests such as 'bench scaling' over more cognitively-oriented tasks employed in current preference tests are also presented.
