Race and sex on the Net 8
steps before utilizing the computer software: developing a coding scheme and dictionary
building. Using a coding scheme, which makes up the central component of a textual
analysis, “the large numbers of words contained in a piece of text are classified in to
content categories” (Bengston & Xu, 1995, p.6). The coding scheme for this particular
data set was theory driven, as discussed in the review of literature above, included themes
from past research (White, 2000), and looked for relevant emergent themes. The higher
order concepts of racialization, sexualization, and commodification guided the initial
steps.
During the data gathering process, a list of possible themes was noted as well as a
initial list of the words or phrases that denoted them. This list was expanded more
systematically through a close reading of a sample of the posts. Because of the size of the
population a smaller sample was chosen. There are two potential problems with a
completely random sample. First, some continents generated more posts than others and
some countries within continents had more posts than others. Asia accounted for 78% of
all posts and Thailand and the Philippines accounted for 43% of Asian posts. Second, this
unequal representation may result in some countries being disproportionately assessed for
local sex tourism discourse, or left out all together. The sample goal was approximately
forty percent of total posts (Bengston & Xu, 1995, p.10). The real number would end up
being higher due to the pre-coding during data collection. The posts were put into two
groups, smaller number of posts and larger number of posts (n>20). All posts in the
former group were included and every nth post in the latter, depending on the total
number.
Convention