Environmental Engineering Reference
In-Depth Information
Key points in the
experimental design
included:
●
The sample was divided into 115 Bubis and 81 Fang. The respondents were
located in both urban and rural areas. The sampling strategy was opportunis-
tic, with people chosen according to availability.
●
No prior notice was given to the interviewees, although the village chief (or
household head, in urban areas) was consulted about the study and asked for
permission to carry out interviews. No respondent was given any monetary
reward for participating.
●
People were asked about their personal preferences and consumption of bush-
meat species. They were asked to rank their top three most preferred species
and their top three most consumed species from all species together first, and
then with the species separated by taxonomic group. Basic information about
the respondents (age, sex, location, ethnic group, profession, marital status)
were also obtained. There was a broad range covered for each of these variables.
●
Preference and consumption scores were ranked as 0
Not mentioned,
First place.
●
Data on market prices and availability were taken from the published
literature.
1
Third place, 2
Second place, 3
Data analysis
included:
●
A preliminary exploration of the data. This involved drawing histograms of
the preference and consumption scores, and creating a cross-tabulation to
show the relationship between the most preferred and the most consumed
species for each ethnic group.
●
They used principal components analysis (PCA) to see how the Bubi and Fang
differed in their preference and consumption scores. PCA can be useful in
reducing datasets with many explanatory variables down to a few composite
variables, although interpretation of these composite variables can be difficult.
●
They used ordinal logistic regression to find out which variables best
explained consumption scores (see Section 3.4 for a weblink). The variables
they tested were preference, age, sex and ethnic group.
●
They had data on market availability and prices for 11 species, which they
used in a multiple linear regression of the effect of a species' availability, price
and mean preference score on its mean consumption score.
The main
results
were:
●
The Fang named a much wider variety of species than the Bubi, and several of
the species they named were not found on the island. The Fang were main-
taining their traditional bushmeat preferences, and importing bushmeat from
the mainland to satisfy these preferences.
●
Generally for the most widely consumed species, consumption was signifi-
cantly related to preference, as well as to ethnic group. Age and sex were not
significant in the regression models.
