Confidentiality and Copyright

1.  Guaranteeing anonymity

Background information sheets were collected from each participant (see Appendix 2). These sheets asked contributors to sign their name to a clause giving permission to use of the recording for linguistic research. While preservation of a contributor's anonymity is strictly observed for material collected in private settings, many public speakers and broadcasters and well-known figures such as the Prime Minister and sports personalities, are easily identifiable from the content of the extract. In these cases anonymity is impossible to achieve and was not attempted.

For the recordings collected in private settings, we promised people who contributed to the corpus that as far as possible their identity would be protected. Consequently, when transcribing material, names were changed to names of equivalent length and phonological structure to protect the identity of people referred to. Phone numbers, addresses and other names (of schools, businesses, etc) were also changed where their use would easily identify private individuals. Moreover, the tapes will not be released to anyone other than bona fide researchers who need them for phonetic or phonological analysis. Currently, use of the tapes is restricted to researchers working at Victoria University of Wellington.

Speakers were given a unique identification number and after relevant social information was entered on the database, their background information sheet was filed with the individual's identification number on it. Thereafter, there is no direct connection between the background information sheet and the relevant recording.

The identification number not only allows a speaker's pseudonym and demographic information to be linked to the other background information held about that contributor without revealing their identity on the transcript, but also permits rapid access to that information on the database. This process enables speakers who appear on multiple extracts (for example, radio interviewers and news readers) to be unambiguously identified as the same person, thus permitting monitoring of maximum word limits for individuals. It has also allowed the development of a relational database to select datasets on the basis of any combination of demographic variables or text category.

2. Copyright

Discussion with Gerry Knowles and Geoffrey Leech, corpus researchers at Lancaster University, indicated that in the light of their experience great care should be taken to obtain copyright clearance for all recorded broadcast material. Though this did not appear likely to be a problem in New Zealand at the time recording began, it turned out in retrospect to be very valuable advice as public radio was subsequently privatised.

Janet Holmes wrote to and subsequently talked to appropriate representatives of the television and radio corporations from whom we wished to record material. They were extremely cooperative and all agreed that we could use any broadcast material, on condition that it was recorded primarily for bona fide research purposes, and would not be used for commercial gain. Following the Lancaster University researchers' advice, this agreement was obtained in writing. A large increase in the privatisation of broadcasting took place during the period of data collection, and the copyright situation changed dramatically in New Zealand. At a late stage in the project we were required to produce our copyright permission by the director of a private arm of the broadcasting company which had been established in order to sell broadcast material. There is no doubt that the cost of obtaining copyright permission at the later stage would have been prohibitive.