Agriculture Reference
In-Depth Information
tree growing and agroforestry covered the area. On the one hand, two out of five
communities were covered by a government reforestation project (Community-
based Forest Management or CBFM project). In this project, community members
cultivating or claiming public lands were encouraged to plant timber trees (mostly
Gmelina arborea
). The project's direct impact was mainly through workshops and
trainings and through the dispersal of tree seedlings. In addition, PLAN interna-
tional Philippines implemented the Sustainable Agriculture Program, which
entailed a tree growing component in all five communities. In this program selected
participants were given trainings on fruit-tree management and fruit-tree seedlings
for planting in their fields. The direct impact of this program seems limited since
only few members per village were included, but for both programs indirect impact
is estimated to be much larger.
The respondents in this study were selected by stratified random sampling based
on village records of households (n = 151, 76 tree planters and 75 non-planters).
Respondents were classified as tree planters when they planted or retained at least
seven trees on one of their plots, with a minimum density of 10 trees/ha. Survey
data were supplemented with field visits and informal interviews. Respondents
were household heads as defined by themselves (96 percent male). All data presented
are farmer-based and are derived from the same survey. Sample sizes however, vary
throughout the paper as not all respondents were able to answer every question.
3.2.3 Statistical Analysis
Several factors are hypothesized to affect tree growing in the study area based on
the exploratory survey and existing evidence in the literature. In order to identify
those factors that best explain tree growing in farm fields, binominal (or binary)
logistic regression analyses have been applied to the data of the 151 households and
associated farm fields. This type of regression is generally used when the dependent
relates to a dichotomy (in this study, the dependent at household level is composed
of the classes “tree growers” and “non-tree growers” and the one at field level is
composed of the classes “fields with trees” and “fields without trees”) whereas the
independents can be of any type (continuous, categorical or dichotomous; Garson
2006; Hosmer and Lemeshow 1989). Hence, the following models were applied:
Equation for
household
level regression:
p
Ln
(
1
p
ab
)
=+
ETHNIC
+
b
NFLABOR
+
b b
EDUC
+
EXTENSION
+
b
LABORAV
1
2
3
4
5
−
+
b
MEMBER
+
b
AGE
+
b
TOTLAND
+
b
CAPTALAV
6
7
8
9
Equation for the
field
level regression:
π
−π
Ln
(
1
)
=
ab
PLOTSIZE
+
b
S
LOP
E
+
b
TENADAP
+
b
DISTANCE
2
3
4
1
+
b
S
OILT
EXTURE
5
Search WWH ::
Custom Search