As with the Welsh-English participants, six Spanish-English partici-
pants (from three recordings) were chosen from the main corpus. Five of the
participants were female and one was male. As Table 6.3 shows, two of the
female speakers were between the ages of 20 and 29, two were between
30 and 39, one was in her forties, and the male was in his fifties. All of the
participants had acquired Spanish at two years of age, and the language
spoken in the home was also Spanish. With regards to the age of acquisition
of English, however, the participants showed more variability. Two of the
speakers had learned English at age two, one speaker at age four, and the
others had not learned English until primary or secondary school. Two of
the six speakers had been to primary school in Spanish, while the remaining
participants had attended school in English. Five of the participants had had
English as the main language in secondary school, with one participant
reporting a bilingual secondary education. Three of the speakers had com-
pleted secondary school and three had completed a university degree. Two
of the participants were born in Cuba then later moved to Miami (one after
11 years and the other after 25 years), two had spent their entire lives in
Miami, one was born in Venezuela then moved to Miami after 20 years, and
the remaining participant was born in Miami but had spent time in several
different cities in the United States, mostly located in Florida and
Massachusetts.
Materials and data collection
As described in Chapter 5 digital audio recordings were collected by a
research team in Wales for the Welsh-English data set and in Miami, Florida
for the Spanish-English data set. Participants were recorded having informal
conversations in pairs for approximately half an hour. In order to minimise
the effect of the Observer’s Paradox (i.e. the notion that participants often
behave different from the norm when being observed (Labov, 1972)), the
investigator was not involved in the conversation. The participants were able
to choose their own bilingual recording partner and often chose a friend or
family member. Subsequently, the recordings were transcribed according to
the CHAT format (MacWhinney, 2000) and the Language Interaction Data
Exchange System (Barnett et al., 2000) in order to make them available to the
public via Talkbank (http://www.talkbank.org). The Welsh-English corpus
consists of approximately forty hours of natural speech data, whereas the
Spanish-English corpus is about 30 hours. (For more details of the two cor-
pora, see Chapter 5.) However for the purpose of this study three recordings
from each corpus were analysed. This is because only three conversations
(i.e. 90 minutes total) were transcribed in the Spanish-English corpus at the
time this research was conducted. For the sake of comparison, a similar
amount of data was chosen from the Welsh-English corpus. The six conver-
sations analysed are available via the Bilingualism Centre website (http://
http://www.bangortalk.org.uk/)..)
Factors Influencing Code-Switching 123