Page 673 - 2015-37

The utility Uti is random and the ith grower practices improved
crop production technology, if U1i>U2i. Thus, for farmer i, the prob-
ability of practicing improved crop production technology is given
as follows (Kishore et al, 2012):

P (A* > 0) = P (U1i > U2i)

= P{β1Fi(SE1i; G1i;)+ε1i}>{β2Fi(SE2I;G2i;) + ε2i} (2)

= P{( ε1i − ε2i) > Fi(SEti; Gti;) ( β1−β2)}

= P{Vi > Fi(SEti; Gti;) β = φ(Xiβ)

Where

P a probability function,

A* latent variable,

Vi random disturbance term,

β β1−β2 is a vector of parameters to be estimated,

Xiβ vector of explanatory variables,

φ(Xiβ) is the cumulative distribution function for Vi evaluated
at Xiβ.

However, A* is not observable. A is equal to one, if a farmer
practices

Improved crop production technology, which is observable i.e.,

A ={1 if A* > 0, 0 if A* ≤ 0} (3)

Empirical Model and Data Analysis

Although there are numerous studies using different models such
as linear regression models, nonlinear probabilistic models, partici-
patory rural appraisal, etc., for analysing the inluence of different
factors on decision variables, the use of logit and probit models are
more frequent (Pattanayak and Mercer, 1998; Subejo, 2000; Neu-
pane et al., 2002; Johnson, 2005; Adeogun et al., 2008; Muneer,

-8-