23.01.2014 Views

An Economic Assessment of Banana Genetic Improvement and ...

An Economic Assessment of Banana Genetic Improvement and ...

An Economic Assessment of Banana Genetic Improvement and ...

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

94 CHAPTER 7<br />

ity in the two technology dem<strong>and</strong> equations<br />

is also correlated. Estimating them jointly<br />

was not considered to be worthwhile, because<br />

the set <strong>of</strong> explanatory variables is<br />

identical in both equations (Greene 2000,<br />

616), implying no gains in statistical efficiency.<br />

Thus, each technology was treated<br />

separately.<br />

Dependent Variables<br />

We considered eight practices recommended<br />

to farmers for soil fertility management in<br />

banana groves: (1) mulch with grass; (2)<br />

mulch with crop residues; (3) mulch with<br />

kitchen residues; (4) addition <strong>of</strong> cattle manure,<br />

(5) goat manure, (6) pig manure, or (7)<br />

poultry manure; <strong>and</strong> (8) composting <strong>of</strong><br />

homestead refuse to make manure (compost<br />

manure). The eight practices were subjected<br />

to a factor analysis procedure using principal<br />

components to analyze correlation between<br />

them <strong>and</strong> determine whether they<br />

could be represented by a smaller number <strong>of</strong><br />

components. 3 Based on the criterion <strong>of</strong> an<br />

eigenvalue greater than unity, the eight soil<br />

management practices were grouped into<br />

four independent packages according to four<br />

unobserved factors. 4 The latent factors were<br />

interpreted as mulching, animal manure,<br />

composting, <strong>and</strong> traditional technologies.<br />

Based on the results from the factor<br />

analysis, the eight types <strong>of</strong> organic fertilizers<br />

were collapsed into two dependent variables:<br />

mulching practices <strong>and</strong> manure application.<br />

Use <strong>of</strong> mulching technology is<br />

defined as the practice <strong>of</strong> applying crop residues<br />

(nonbanana crop residues) or grass to<br />

mulch banana plantations. Use <strong>of</strong> this type<br />

<strong>of</strong> mulch materials in banana production<br />

involves costs for gathering, transportation,<br />

<strong>and</strong> application, which may not be affordable<br />

by some households. Use <strong>of</strong> manure<br />

technology is defined as the practice <strong>of</strong> applying<br />

animal waste (from cattle <strong>and</strong> pigs)<br />

or compost manure in the banana plantations.<br />

These organic fertilizers also involve<br />

a cost <strong>of</strong> access, transportation, <strong>and</strong> application<br />

that may limit their use.<br />

The omitted category (1 – δ) is the traditional<br />

method <strong>of</strong> recycling banana residues<br />

<strong>and</strong> applying household refuse, used by virtually<br />

every farmer on at least some mats.<br />

Organic household refuse (kitchen residues,<br />

goat <strong>and</strong> poultry waste) is collected in mixtures<br />

from the homestead as part <strong>of</strong> the<br />

cleaning activity <strong>and</strong> spread between mats<br />

to control weeds, though some farmers reported<br />

that use was not deliberate. Farmers<br />

are advised to compost the household refuse<br />

<strong>and</strong> other organic materials before applying<br />

it in banana plantations to facilitate quick<br />

decomposition (Tushemereirwe et al. 2003).<br />

Therefore, this method may not be adequate<br />

for soil fertility management in banana<br />

plantations. In addition, little additional information<br />

would be gained by investigating<br />

3<br />

A principal components analysis with Varimax rotation was applied (STATA 8.0). Principal components<br />

analysis partitions the sum <strong>of</strong> the variances for a set <strong>of</strong> variables by taking uncorrelated, linear combinations <strong>of</strong><br />

those that account for the latest eigenvalues <strong>of</strong> the covariance matrix. The components are not easily interpreted,<br />

however. The Varimax rotation procedure, one <strong>of</strong> several types <strong>of</strong> rotation, enables a more straightforward<br />

interpretation <strong>of</strong> components in terms <strong>of</strong> factors “loading” high on some variables <strong>and</strong> low on others (Stevens<br />

2002).<br />

4<br />

The four factors explain 73 percent <strong>of</strong> total variance in the adoption components. Only variables with factor<br />

loading equal to or greater than 0.5 were considered significantly correlated with the underlying latent variable<br />

<strong>and</strong> hence grouped together. Factor 1 explains about 28 percent <strong>of</strong> the variance in the eight management<br />

practices. Organic household refuse (kitchen residues, goat manure, <strong>and</strong> poultry manure) were highly correlated<br />

with the first latent factor. Mulching techniques using the organic materials that are gathered from sources<br />

other than the banana crop (crop residues <strong>and</strong> grass) loaded on factor 2 <strong>and</strong> explain 18 percent <strong>of</strong> the total variance<br />

in the use <strong>of</strong> soil fertility management practices. Organic manure from animals that are rarely kept in the<br />

homestead (cattle <strong>and</strong> pigs) also clustered together, while the technique <strong>of</strong> composting household refuse before<br />

applying it to banana plants instead <strong>of</strong> applying it directly seems to be used independently <strong>of</strong> other soil fertility<br />

management practices.

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!