East Timor formally obtained its independence in 2002 following a protracted period of occupation by
Indonesia which ended in 1999. It was initially faced with a series of issues such as a low level of
infrastructure, poor quality germplasm for the major staple crops, and being one of the world’s poorest
nations (Piggin and Palmer, 2003). Many East Timorese experience annual periods of food shortage,
sometimes exacerbated by droughts and pest damage (Piggin and Palmer, 2003). In response to a shortage
of suitably adapted varieties for the major staple crops of East Timor, a project called Seeds of Life was
developed in 2000 to locate and test local and international crop varieties with the aim of improving the
germplasm stock in the country (Piggin and Palmer, 2003). Seeds of Life recruited willing farmers to
participate in On-Farm Demonstration Trials (OFDTs) in 2006 which was hoped to result in independent
replanting and seed dissemination by these participants to neighbouring farmers. Two international maize
varieties were extended to participating farmers for trial – these were LYDMR (Late Yellow with Downy
Mildew Resistance) and Suwan 5 (a popular Thai variety with Downy Mildew Resistance). Given the
reported potential for the new varieties to increase farm maize yields, the self-selection of participants in
the Seeds of Life program, and that the adoption process was only in its first phase, a significant proportion
of non-adoption following OFDTs was observed (approximately 32% of participants).
A survey conducted in 2007 provided data for the estimation of a binary probit regression model to assess
the reasons for non-adoption. Results obtained corroborated the findings of Seeds of Life researchers prior
to variety extension; yet initial testing of varieties did not explicitly involve the inclusion of factors that
were considered likely to affect the utility of prospective adopters. Non-inclusion of factors relevant to
household utility when assessing new crop varieties may lead to the selection of less than optimal varieties.
Stochastic dominance methods are a potential solution to this issue allowing researchers to consider the
impact of new crop varieties on household utility and thus adoption decisions prior to their extension.
Stochastic dominance methods can be derived from the same utility maximisation framework as the probit
regression model and easily incorporate non-normal distributions of returns. Their capabilities in assessing
high numbers of potential innovations and their similarity in ease of application to existing methods such as
mean-variance dominance analysis are also advantages. In this paper tests for stochastic dominance are
retrospectively applied to the two introduced and the local maize varieties to demonstrate their application
as a competitive and relevant ex ante technology assessment tool in developing countries.