This chapter describes the kinds of data analysed in this book, how they are generated, and how they are to be interpreted. A number of important methods and fundamental methodological issues in phonological and phonetic research are discussed, including some new methods.
The phonological methods used in Chapter 3, Phonological Preliminaries, are briefly discussed. Then, the procedure by which each idiolect was phonetically analyzed is described, step by step. A method of impressionistically classifying vowels for their level of phrasal stress is given in detail. The clitic/non-clitic distinction is defined. This enables the exclusion of forms that may be phonologically reduced, in cases where this makes the pattern more clear. I discuss the criteria and rationale behind the choice of a particular time slice within the acoustic vowel as maximally representative of the phonetic quality of the vowel nucleus. Numbers of measured, excluded, deleted, and total tokens are given. A method of locating and accounting for outliers is described, along with some typical results of applying this method: most cases of apparent outliers are legitimate cases of extreme phonetic variation. The final section discusses statistical methods, including the bootstrap technique, which is used to create easily-interpreted displays of large amounts of data, as well as the statistical tests which are used to show that sets of F1-F2 measurements are significantly different.