The problem still is, however, that it has been easier to develop technology that suits the
"progressive farmer" than it is to develop technology that fits within the narrow margins of low
access and resource-poor farmers (R61ing, 1988. p. 71). Thus, the extension message that goes
to the different categories of farmers most often is still one message, developed on experiment
stations and appropriate only for "progressive farmers," those with the best resource base,
Figure 4. Without a means to modify the technology to suit the resources and environments of
the "laggards," the message cannot result in improved technology for farmers whose resources
and environments do not reflect those found under experimental conditions on the stations.


MSA as an Alternative to the "Progressive Farmer" and an improvement to the "Target
Categories" Strategies

Modified Stability Analysis or MSA (Hildebrand, 1984) provides a means for solving the
lingering problems associated with developing and diffusing improved agricultural technologies
to all categories of farmers in any community. In MSA, a wide range of farmers and their fields
in a community or region can be part of a single research domain. To the extent that a
technology might be appropriate (desired and possible to use) on any of the farms or fields in
the research domain, any field on any farm in the domain can be used as an environment for on-
farm testing. This makes the research process more efficient. Research results from an on-farm
testing program can be used to identify recommendation domains based on the environment in
each field (both the biophysical base environment and the socioeconomic modifications to the
base) and on the different farmers' evaluation criteria.

A critical element in the design of an on-farm trial to make it amenable to analysis by MSA is
that each environment (field) have the same set of treatments, or at least a common set of two
or more treatments. In projects such as those in the Illinois Sustainable Agriculture Network,
if groups of farmers could agree on a common set of treatments, the resulting design would be
amenable to MSA. Individual farmers could still include their own specialized treatment, in
addition to the common set if they wanted additional information. For purposes of MSA, it is
not necessary to replicate treatments within an environment. If farmers want to replicate, then
treatment averages within an environment can be used for the MSA. However, it is more
important for purposes of MSA to have more environments than more replications within
environments. With a large number of cooperators involved in a single year, it may not be
necessary to wait two or more years before having definitive results (Stucker and Hicks, 1992;
Stroup et al., 1993).

A cooperative on-farm research program designed in this way results in multiple extension
messages suitable for well defined environments and different evaluation or selection criteria.
Each "target category" can receive a message appropriate to its conditions, Figure 5. An
example from Brazil will be used to illustrate the potential of MSA in this kind of a program.

Table 1 demonstrates the nature of recommendation domains specified by appropriately designed
and analyzed on-farm maize fertility research conducted in a series of communities, a research