Development of Health and Sustainability Indicators for a Tropical Highlands Agroecosystem - Integrated Assessment of Health and Sustainability of Agroecosystems - page 167

The first dimension has a score that is correlated with measures of allocative
efficiency of cattle production (heads of cattle [0.65], head of cattle/available labor
[0.56], and heads of cattle per acre [0.54]). The score of the second dimension is
most correlated with the proportion of LUU income that is nonfarm ( r 2 = 0.30) and
with the school dropout rate ( r 2 = 0.29). The highest correlations with the score of
the third dimension are with measures of allocative efficiency of food crop produc-
tion (bean yield per acre [0.36], proportion of land under beans [0.35], maize yield
per acre [0.33], proportion of land under maize [0.32], potatoes yield per acre [0.25],
proportion of land under potatoes [0.22]).
Table 6.10 shows the principle inertia of the six dimensions accounting for 75.9%
of the variation in researcher-proposed, SSL (village-level) indicators of health and
sustainability, the indicators most correlated with the score of each dimension and
the coordinates for their categories along these dimensions. The first and second
dimensions accounted for over 16% of the variation each, while each dimension
from the third to the sixth accounted for between 12% and 8%. The principle iner-
tias ranged from 0.19 for dimension 1 to 0.08 for dimension 6. Only the first three
dimensions represented significant average correlations between the indicators and
the scores ( p < .1).
Five categories had a quality less than 0.6 (distance to Nairobi = L [0.40], all cat-
egories of nuclear family outside village [0.46], and all categories of occurrence of
animal diseases [0.50]). The category with the lowest mass was soil classification = H
(0.003). The ones with the highest (0.014) mass were coffee production = A, nuclear
family outside village = L, proportion of farms using agrochemicals = L, proportion
of LUUs with bank accounts = L.
6.3.4 C of m p A r i s of n of f i in D i C A t of r s u i t e s
Six of the attribute classifications were common to both suites of indicators: (1)
equity, (2) environmental quality, (3) soil fertility, (4) pest and disease dynamics,
(5) infrastructure, and (6) knowledge. However, the focus was on different catego-
ries of indicators within each of the attribute classes, resulting in differences in the
indicators chosen. For example, communities focused mostly on productivity and
physical characteristics in the soil fertility attribute, while the researcher-proposed
suite focused on chemical fertility and physical classification of the soils. The choice
of indicators within the same category of an attribute differed between the two
suites. Among the indicators common to both suites were distance to water source,
frequency of hospital visits, number of livestock, availability of extension services,
accessibility of infrastructure, morbidity and mortality, quantities of yields, and
presence or absence of various farm enterprises. The use of livestock numbers and
cash crops as indicators of capital, wealth, or savings was common to both suites. An
important difference between the two suites was the presence of value-based mea-
sures such as “proper hygiene,” “good behavior,” “good variety,” and “good habits”
in the community-based suite. In addition, many of the indicators in this suite were
mostly in ordinal scale. Researcher-proposed indicators were mostly numeric, non-
value-based measures generally on the continuous scale.