Evaluating English Accents WorldWide: Methods
The methodology used in this research is one with a long pedigree, stretching back to Lambert and Anisfeld's Canadian research of the late 1950s and early 1960s. This has gone on to be further refined and elaborated by Giles, Bourhis, and many others. It should be noted that one of the important features of this techniquethe matched guise, which employs the same person speaking with two different accentswas not used in this study. It seemed to us impossible to find speakers of each gender who could give convincing renderings of all four accents without falling into the trap of projecting a stereotype rather than the genuine article. However, every attempt was made to hold paralinguistic variables like reading speed and number of hesitations to the minimum:
Despite this care, paralinguistic variation does exist, and it is important to bear this in mind when considering the low ratings assigned to the male New Zealand voice. This is very likely due to his monotonous reading style.
* The practice voice was used to familiarise listeners with filling in the questionnaire; data was coded, but not analysed. A NAm-accented female voice was used in administering the questionnaire in America, but the more "neutral" NZE practice voice was used in the European and Asian experiments to avoid biasing listeners toward NAm.