Agriculture Reference
In-Depth Information
1 p
Where: Ln (
represents the logit with
π
denoting the proportion of success that
)
p
the dependent variable y equals 1 (= tree growers or fields with trees) and 1-
the
proportion of success that y equals 0 (= non-tree growers or fields without trees);
α
π
β i is the partial regression coefficient; ETHNIC = Ethnicity of the
household head (0 = non-Ilocano, 1 = Ilocano); NFLABOR = Amount of time spent
by household in non-farming activities (days per month); EDUC = Highest education
attained by the household head (0 = until primary, 1 = high school and up);
EXTENSION = Whether or not the household received extension on tree growing or
agroforestry (0 = no, 1 = yes); LABORAV = Number of household members contrib-
uting to household income or farm labor (0 = one person, 1 = two or more persons);
MEMBER = Whether or not the household is a member of a village-based institution
(0 = no, 1 = yes); AGE = Age of the household head; TOTLAND = Total land area
owned (both formally and informally) by the household (ha); AGROECOZONE =
location of farm (ll = lowland, hl = hilly lowland and ul = upland); CAPTALAV =
Aggregate variable for the amount of monetary capital available to the household
(0 = none or low, 1 = moderate); PLOTSIZE = Size of individual fields cultivated by
a household (ha); SLOPE = Slope of individual fields cultivated by a household (0 =
flat, 1 = sloping); TENADAP = Tenure in individual fields cultivated by a household
(0 = insecure tenure, 1 = secure tenure); DISTANCE = travel distance between field
and house of respondent (minutes); and SOILTEXURE = soil texture in individual
fields (0 = fine, clay texture, 1 = medium loam texture, 2 = coarse sandy texture).
Based on analysis of correlations between explanatory variables, some variables
were eliminated from the regression analysis, e.g., non-farm labor (NFLABOR)
was included and preferred over non-farm income also because the latter resulted
in less reliable data. Some categorical variables have been transformed into binary
variables in order to better fit the logistic regression model. The variable
CAPTALAV has been transformed into ranks, meaning it is an aggregate variable
based on the values of several underlying measurements including household sav-
ings, income and assets. Each of these observations has been accorded a value (0 =
low, 1 = moderate) based on the researchers' assessment of relative importance of
each. The variable ETHNIC has been included because of indications that tree
growing is associated with certain ethnic groups, e.g., the Ifugao traditionally living
in upland areas and accustomed to rice and tree growing and the Ilocano tradition-
ally living in lowlands areas and mostly engaged in seasonal cropping and not so
much in tree growing. Most interviewed farmers are Ilocano, explaining the distinc-
tion in two ethnic groups: Ilocano and non-Ilocano.
The statistical package SPSS version 12.1 for windows is used for the data quan-
titative analyses. The means of the several variables describing the household and
field characteristics of tree and non-tree growers and their fields have been com-
pared using the
is a constant and
2 test for counted data and either the GLIM univariate analysis or
the T test with equal variances not assumed for numeric or measurement data. The
variables measured at the plot-level are nested within the households and therefore,
a general linear model with nested terms was specified.
χ
Search WWH ::




Custom Search