Innovation and equity in rural development

Material Information

Innovation and equity in rural development
Series Title:
Preliminary papers. The Rockefeller Foundation. April 29-30, 1975
Conference on Social Science Research in Rural Development
Roling, Niels
Ascroft, Joseph
Chege, Fred Wa
Wageningen -- The Netherlands -- Department of Extension Education
University of Iowa -- School of Journalism
Michigan State University -- Department of Agricultural Economics
Place of Publication:
Gainesville, Fla.
Farming Systems Support Project, University of Florida
Publication Date:

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
The University of Florida George A. Smathers Libraries respect the intellectual property rights of others and do not claim any copyright interest in this item. This item may be protected by copyright but is made available here under a claim of fair use (17 U.S.C. §107) for non-profit research and educational purposes. Users of this work have responsibility for determining copyright status prior to reusing, publishing or reproducing this item for purposes other than what is allowed by fair use or other copyright exemptions. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder. The Smathers Libraries would like to learn more about this item and invite individuals or organizations to contact Digital Services ( with any additional information they can provide.

Full Text
Niels R61ing, Dep. of Extension Educ., Wageningen,
The Netherlands
Joseph Ascroft, School of Journalism, Univ. of
Fred Wa Chege, Dep. of Agricultural Econ., MSU,
East Lansing
Paper read for the Meeting of the ISA Research Committee on Innovative Processes in Social Change at the VIIIth World Congress of Sociology. Toronto, August 1974

The generalisations of diffusion theory are the result of observing current practice. They do not give recommendations for optimal practice. Yet, diffusion tenets have diffused widely and often assume a normative role for praxis, in that they justify and reinforce current practice. Current practice boils down to "progressive farmer extension strategies" whereby-scarce government resources are concentrated on a few progressive farmers in the hope that this effort will be multiplied widely through diffusion. The paper raises the question whether this is a strategy designed for equitable development or a case of "to those who have shall be given". In itself, the strategy seems practical and-efficient, especially from an agency-centered view-point.
Theoretically, diffusion processes can be seen as equalisers in the sense that they reflect a tendency toward entropy or sameness for as far as availability and utilisation of information are concerned. However, in practice, a number of imperfections of equalisation by diffusion can be mentioned which lead to the conclusion that diffusion processes do indeed tend to increase inequity and that diffusion generalisations can, therefore, only be used as guides for devising strategies to avoid what they predict. This state of affairs implies that those who seek equitable development must not carry out more surveys of current practice but must seek to experiment to proto-type test alternatives to current practice. The paper concludes with a brief report on one such experiment in Kenya which did show that one can aim one's efforts at laggardly farmers and not only obtain 100 % adoption among those reached directly, but also have an immediate diffusion effect of about three farmers for each farmer reached.

At a Farmers' Training Center is a rural district somewhere in Africa,
grass roots field workers receive systematic training for the first time. We witness the occasion at a moment when the trainer, himself a diploma-level extension worker, explains diffusion theory with the
help of an AID manual on extension. At issue is the presentation of the
bell-shaped diffusion curve: 2'% innovators, 121% early adopters,
34% early majority, 34% late majority ..... At this point, a question
is raised: "Sir, why are early and late majority both 34%?"
The Aid manual has dug a deep. hole for the trainer. He hesitates.
Finally he says: "I think they are neighbours".
The anecdote illustrates two things. First, the extent to which diffusion theory has diffused. And second, that diffusion processes have their weaknesses in disseminating ideas, and thereby, in generating change. We shall pay some attention to the first point, but it is especially the. second which is at issue in the present paper. For we would like to examine the influence of diffusion of innovation processes on equity in rural development. Diffusion is usually seen asa god-sent'autonomous process which assures the trickle down of income- and welfare generating ideas and thereby guarantees their distribution among all members of a population without effort of the change agent.
Thus, one looks at how well diffusion processes distribute the benefits of new technology. Such a viewpoint sufficed when the emphasis was on economic growth and on breaking down the barriers of traditionalism. However, times have changed. Instead of traditional tribesmen and warriors, we have masses of small-holders whose lack of opportunities rather than resistance to change seems to be the major bottleneck in development. Also, inequities are rapidly emerging in once egalitarian tribal societies. Classes of landless pleasants, rural unemployed, slum dwellers and seasonal laborers emerge, where formerly each man had a right to farm and an independent existence. Coupled to the fact; that, in most developing nations, alternative employment is so slow in forthcoming that more instead of fewer members of the rapidly growing population must find a living in rural areas, at least for the next several decades, these conditions have
x the authors wish to aknowledge the helpful comments of professor A.W. van
den Ban, Depavtment of Extension EducationsWageningen

led to a greater interest in equity and distribution, next to growth. "We know in effect, that there is no rational alternative to policies of greater social equity" (McNamara, 1972), or, "There is no viable alternative to increasing the productivity of small-scale agriculture if any significant advance is to be made in solving the problems of absolute poverty in the rural areas ..... or of achie-ving long-tern stable economic growth" (McNamara, 1973). Clearly it is too early to breed a small elite of highly productive farmers who can provide the food for masses of workers employed in industry and services. And thus, we look not at how well diffusion processes distribute the benefits of new technology, but at how badly they do it.
Such a concern could lead to an exercise of old wine in new sacks, a reinterpretation of known facts. The present paper does aim to contribute to such a re-interpretation, to be sure. But is also attempts to offer new directions. For after examining the effect of the diffusion of innovations on equity, both theoretically and empirically, we would like to present some results of an effort to experimentally change current practice to achieve more equitable development in a rural area in Kenya.
At the onset, we would like to stress one point. We work with the "African model". That is, we assume societies with a rather egalitarian tribal past which are only beginning to learn the meaning of relative wealth and poverty. This point must be made, because those who work with the "Latin American model" and therefore assume a traditional society characterised by latifundistas, peasants and dependency relationships, will find it difficult to understand our concern with the diffusion of innovations as a contributing factor in inequity, over and above structural factors and especially the distribution of land ownership.

The diffusion of.innovations research tradition is probably unique among social sciences in the -extent of its empirical base (Havelock et al, 1969, p.11.11). Its body of generalisations has been disseminated with enthusiasm, clarity and care (e.g. Rogers, 1962 and Rogers and Shoemaker, 1971). Yet, this state of affairs may have its drawbacks. Diffusion research generalisations adequately draw donclusions about current practice, which is very different from offering recommendations for optimal practice. Yet, diffusion research generalisations often become normative for practice, precisely because they have diffused so widely.
1. The generalisations re-inforce an extension focus on progressive farmers by
showing that innovations do trickle from progressive farmers downwards. Of
course, the generalisations derive from the fact that most extension services
follow this strategy of least resistance, but that does not mean that it is
a strategy for optimum effect.
2. The generalisations re-inforce and systematise the use of adopter categorisations. There are few extension workers who do not classify their farmers in terms of progressiveness or innovativeness and make use of this classi*fication to concentrate on farmers who are quick to follow advice, of
sufficient means, knowledgeable and homophilous with the extension worker.
Diffusion tenets show that there are the farmers who have contact with
extension and thus justify current target group selection principles in a
situation where the extension worker has to make some choice because he
cannot cover all.
Also, generalisations on. adopter categorisation allow rei-fication. The
laggards, for instance, whose main pre-occupation is said to be the rearview mirror (Rogers, 1962, p. 71), are deemed incapable to change, frustrated, and fatalistic hardcores, even in societies where change is so recent that late or no adoption is often still more a question of inability than of the resistance to change bred by a long history of failure, oppression,
frustration and relative deprivation. The only real "hard cores" we
encountered in Kenya, for instance, were "local politicians", local exMauMau fighters and their supporters, who were so deeply frustrated in the
Post-Independence period that they now actively resist all change.
3. The basic tenet of diffusion research that innovations diffuse autonomously
from those in direct contact with external sources of information to other members of the community, insures a multiplier effect for the activites of
the field worker in a situation where he can only have with a

small proportion of the farmers, es specially when he relies heavily on individual farm visits as his main method of extension. Diffusion research shows,
in effect, that there is no need to focus on more than a fraction of the
It must be added here that one related lesson which the practicioner has
usually not learnt is that next year's seed and fertiliser requirements can
not be based on linear prognoses but must take account of the snowballfashion acceleration of adoption.
4. "The most important single strategy of change advocated to change agents by
diffusion researchers is that of working through opinion leaders" (Sen and
Bhowmik, 1970, p. 1). Most change agents have learnt this lesson well,
albeit in a simple form. For opinion leaders are usually taken to be those
progressive farmers who are also leaders. That "interpersonal diffusion
is mostly homophilous" (Rogers and Shoemaker, 1971, p. 212) or that "the opinion leader is very much like the person he advises" (Robertson, 1971,
p. 184) is usually forgotten and also, that,.in a rapidly stratifying
society, each stratum may have its own opinion leaders), instead of a few
persons functioning as such for the whole community.
The following quote illustrates the usual short-cut which results (Republic of Indonesia, 1973, p. 61 a.f.). After having said that "ideal demonstrator
farmers" should be progressive, influential, sufficiently educated, represen
tative and of sufficient means, the author continues: "Sure enough, it will
not be easy to find the ideal demonstrator farmer,......... but it must be considered quite possible to find always a farmer who is willing to follow
advice and to play a leading part in a farmers' meeting without insisting
on being paid for that" (italics in original).
The net effect of.the diffusion of difussion research tenets, themselves based
on observations of current practics, has been to re-inforce, condone and systematise current practice. That practice is to provide intensive service:; to a handful of innovative, wealthy, large, educated and information seeking farmers and to. expect that the effect of these services will reach other farmers by autonomous diffusion processes.
2) personal communication Kees van Woerkum, dep. of Ext. Educ, Wageningen.
Van den Ban (1963, p. 182/83)also found that farmers choose, as advisors,
not very progressive farmers, but people-who are somewhat more progressive
than they themselves.

The question is: How realistic is this ex"'ectation? Is it a case of "to those who have shall be given", or is it a strategy designed for equitable development? In trying to answer this question we assume a responsibility. For if we were to show that the "progressive farmer strategy", as currently practiced, is detrimental to the achievement of equitable development, it will be upon us to' provide alternatives for the very practical "progressive farmer strategy".
1. Progressive farmers have large farms so that the extension worker's direct
effect on total production is greater than if he works with less progressive
2. Progressive farmers are those who can be expected to form the future core
of commercial. farmers who provide the nation with food and export earnings.
3. Progressive farmers have a high sense of efficacy. (Smith and Inkeles, 1966).
Thus they are eager for information. They follow advice. One does not waste
time convincing them. One gets quick results for annual reports and promotion.
4. Progressive farmers.demand services. One cannot by-pass them. The complain
if they are neglected. Some are powerful enough to threaten the career of
the extensionworker (Leonard, in prep.).
5.. Progressive farmers have means to try new ideas. They do not always complain
'Of lack of money. Others need credit, the provision of which is notoriously difficult in case of small-holders in developing nations. For these reasons also, demonstration plots are usually laid out on the farms of the progressives. Governments do not usually pay for the necessary inputs.
6. Progressive farmers are usually homophilous with the extension worker. It is
easy for them to communicate. In some cases, the progressive farmers are
more knowledgeable.and sophisticated than the field workers. They go directly
to high officials. Such situations are seen as an embarassment to'the
service. "our field workers cannot face the farmers".
7. Progressive farmers provide a challenge to the fieldworker. They keep him on
his toes and thus invigorate the service.
8. Extension workers learn from progressive farmers what to tell other farmers. In all, the progressive farmer strategy is clearly a very efficient and aittr n'tiv, L one, especially from an agency-centered viewpoint.
van den Ban, personal comm.

That diffusion research, with its emphasis on survey methods, developed generalisationswhichpresuppose the progressive farmer strategy shows that the strategy is indeed widely practiced. Should diffusion researchers stop surveying and start proto-type testing alternatives to current practice? We shall examine the question from value judgement that equitable development is desirable. We define equitable development as social change processes which increase the extent to which members of a population are able to elicit similar outcomes, regarding their physical, security and social needs, from their environment.
In theory, diffusionprocesses are equalisers. Assume a tribal village some hundred years ago. It was largely closed to new information. As a result, the village was characterised by sameness or entropy. In fact, the antropologist could visit the place, describe one house, one farm and he had described them all. He had no need for statistics.
Some years later, some of the villagers who had been taken away as slaves to work in Brazilian cocoa plantations, return. They have done away with the bark cloth the villagers are wearing. Instead, they drape colorful cottons around their bodies. They alo plant cocoa to be able to buy corrugated iron roofs. These attractive outcomes for some create relative povertyamong the rest. And the normal reaction to relativepoverty is innovation. The rest thus also plants cocoa and also buys cotton dresses and iron sheets. This caricaturised version of early development in Yoruba land nevertheless demonstrates that diffusion processes can be seen as restorers of equity after a difference in outcomes has been introduced. One could even go so far as to say that the "natural tendency" for communities is to reach a stable state, i.e., a condition of sameness or equity, and that diffusion processes are the main mechanism for reaching that stable state. The S-curve describing the
diffusion of an innovation in a community describes the restoration of equity over time.
And thus, the anthropologist could go back to look at the. cocoa farm. Or could he? If diffusion processes tend to restore equity, why does UNRISD (Griffin, 1972) and others (Freebairn, 1973) worry so much about the social consequences of the "Green Revolution"? Why do such great inequities emerge.also in previously more egalitarian societies?

Diffusion processes are, in fact, imperfect equalisers. Several reasons for this state of affairs can be logically'deduced and empirically shown.
1. Innovations do not arrive in rural communities one by one, nor is there time
for equalizing processes to catch up on each innovation. Instead., innovations come in rapid succession. While some members of the community are still adopting an earlier innovation, others are already reaping benefits from more
recently introduced ones.
2. Innovations take time to diffuse. But even if it takes only ten years for a
new cash crop to diffuse, those who plant early receive an income over an
additional numberof years which may put them ahead so much that later adopters
may find it impossible to catch up.
3. Early adopters reap windfall profits (Rogers, 1962, p. 276). They start
producing at a time when the product is still relatively scarce and prices
are high or when subsidies or inputs are still offered as incentives. Later adopters may find prices low, while some may be frustrated by actual prohibitions to further adopt. An example are the coffee quota imposed as a
result of international commodity agreements.
4. Having money or more money relatively earlier than others allows acquisition
of additional resources when they are still cheap.

INNOVATION most upper lower least TOTAL
progr. middle middle progr.
1. grade cattle 93% 70% 49% 11% 60%
2. coffee 72 68 30 0 45
3. tea 22 12 3 0 10
4. pyrethrum 22 13 10 0 12
5. hybrid maize 63 32 22 0 31
6. certified potatoes 9 2 1 0 21
7. maccadamia nuts 30 31 16 0 21
8. pigs 57 35 21 0 30
TOTAL PERCENT 368% 263% 152% 11% 212%
BASE 92 97 102 63 354
Source: J. Ascroft, et al. (1973, p. 23)
In 1970, we carried out a survey (Ascroft et al, 1971) in Tetu Division of Central Province Kenya, among a random sample of 354 registered farmers. Using the 8 innovations in the table above, the farmers were given a progressiveness score based on the sum of the number of years a farmer had been using any of th4 eight innovations. The scores were then used to categorise the farmers in most progressives, upper middle, lower middle en least progressives. The two variables in the above table are thus not independent. Yet the figures show the magnitude of difference between most and least progressive farmers. "TOTAL PERCENT" shows that the most progressives had adopted an average of 3.7 innovations, while the least progressives had adopted an average of 0.1 innovations.

INCOME IN KSHS most upper lower least TOTAL
progr. middle middle progr.
299 and less 11% 13% 35% 81% 32%
300-999 19 43 .35 16 30
1000 and over 70 44 30 3 38
TOTAL PERCENT 100% 100% 100% 100% 100%
BASE 79 91 112 59 341
Source: F. Chege (in preparation).
In 1973, the same farmers interviewed in 1970 were interviewed again for a panel study (Chege, in prep). Income had not been studied in 1970 for fear of useless data. We tried in 1973, asking each respondent how much he had made from his beans, his tea, his milk, off-farm employment, etc. If one adds the total for each respondent, one obtains a figure which may not reflect actual income, but which seems comparable across respondents. The variable does discriminate between progressiveness categories, with 70% of the most progressives making 1000 shillings or more and 81% of the least progressives making 29.9 shillings or less.

most upper lower least TOTAL progr middle middle progr
1 plot 60% 76% .91% 96% 80%
2 plots 32 18 7 2 15
3 or more plots 8 6 2 2 5
TOTAL 100% 100% 100% 100% 100%
BASE 92 97 102 63 354
Source: J. Ascroft et al. (1973, p. 29)
When land registration and adjudication had been completed in the area of study in 1959, each farmer had one parcel of land. In 1970, 40% of the most progressives had acquired 1 or more additional plots, as against only 4% of the least progressives.

5. Ideas diffuse easily. They are not subject to scarcity. One can pass on
an idea and keep it. But implementation of ideas takes resources which are subject to scarity. Differences in resource endowment, such as the
power to command traditional land rights or input supplies, differences
in health, available family labor to clear virgin land for cashcrops,
intelligence, etc., may imply great differences between farm households
in the extent to which they can benefit from innovations. There are some
practicioners who believe that late adopters have only themselves to
blame, in that they are stupid, lack entrepreneurship, and so on.
As said, very many reasons, most of which cannot be blamed on the farmer, can be mentioned. Another reason may be, for instance, that an important
innovation "strikes" at a time when the family circumstances of the
life cycle of the individual make early adoption impossible
6. As we saw earlier, extension services tend to focus on progressive
farmers. These farmers tend to become a fixed clientele of the
extension worker, so that new information is always channeled to the same farmers, strengthening their advantages through early adoption.
verbal communication of W. .Meynen, ISS, the Hague.

most upper lower least TOTAL
progr. middle middle progr. AV. FARMSIZE (ACRES) 9,7 6,7 4,2 2,6 6,0
Source: J. Ascroft et al (1973, p. 29)
Whether as a result of recent acquisition, or as a result of having more land to start with, the most progressives had a farm which was, on average, nearly four times the size of the farm of the least progressives.

FARMVISIT BY EXTENSION most upper lower least TOTAL
AT LEAST ONCE LAST YEAR progr. middle middle progr.
crop extension worker 100% 96% 85% 41% 84%
livestock extension w. 93 77 63 22 68
home economics 10 7 3 2 6
TOTAL PERCENTx 203% 180% 151% 65% 158%
BASE 92 97 102 63 354
TYPES OF CONTACT AT most upper lower least TOTAL
LEAST ONCE LAST YEAR progr. middle middle progr.
Farmers' Training Center 48% 26% .13% 5% 24%
Crop Demonstration 92 76 76 37 70
Livestock Demonstration 91 72 61 32 67
Home Ec. Demonstration 74 60 50 32 56
TOTAL PERCENTx) 305% 234% 200% 156% 217%
BASE 92 97 102 63. 354
Source: J. Ascroft et al., (1973, p. 32).
Totals add up to more than 100% because farmers may have received different types of workers or visited more than one type of group extension activity,
at last once since "thesame time last year".
All figures demonstrate a highly skewed distribution of the provision of vxteiision services. It should be noted, however, that a very high proportion of the farmers had been visited. Kenya has a very favourable extension/farmer ratio. Interesting is also that the four home-economics extension workers had reached 56% (+ 5%) of the 12.000 registered farm families in'the area througireliance on group extension methods instead of the individual farm visit.
m, ,,

It is indeed likely that an extension worker gradually develops a fixed clientele. When an extension worker starts in a traditional village,-he will at first spend most of his time on promoting innovations. The more he succeeds, however, the more time he will have to devote to follow-up and maintenance of adopted innovations, until he reaches a point where he cannot handle more farmers. If he has further innovations to promote, he will do so on farms where he comes for maintenance. We have some tentative empirical support for this contention.
7. Credit is given to those who can give collateral. Thus costly, and therefore
often profitable, innovations can easiest be adopted by those who are
relatively well off.
8. Research stations usually concentrate on developing recommendations which
benefit especially the larger farmers and the high-potential areas, and
couch their recommendations in terms which only allow their application by
large, educated farmers (Mbithi, 1972).
9. Although ideas are not subject to scarcity, they are subject to loss of
fidelity. Those who receive information second, third, or nth mouth, will have less, less specific and less reliable information. One experiment inIndia (Sinha and Mehta, 1972) showed that, of information broadcast to a
primary audience, only 14% reached a secondary audience via informal channels. Lowdermilk (1972) observed in Pakistan that, of a whole package of
practices required for the successful production of HYV grains, only the
new seed diffused to the smaller farmers. In our experiment we observed that,
of those who'adopted as a result of hearing from those we had trained,
all adopted the recommended maize variety, but only 40% used the compound fertilisers recommended, while 69% did not apply dust against stalkborers.

"What do you usually Score of 6 Score of 5 Score or 4 TOTAL
spend more time on:" and over or less
1. helping farmers solve
problems with crops or 50% 40% 33% 43%
livestock they already have
2. Making farmers know crops
or livestocks they.did not 50% 60% 67% 57%
know before ("innovation
TOTAL 100% 100% 100% 100%
BASE 10 5 6 21
Total N is very small. These figures provide, therefore, no-more than
tentative support for the relationship. One of the problems in studying
the relationship is that one needs large numbers of farmers to get development scores for only a few areas of work of extension workers.
The 23 sublocations making up the arta of study were rankordered with the aid of scalogram analysis, using as items the aggregated data of the 1970 survey(R1ing. 1972). The sublocations were then divided into three categories based on the number of dichotomised items on which they scored. The scale reflects a measure of agricultural development. Extension workers responsible for the agriculturally most advanced sublocations tend to claim they spend more time on "MAINTENANCE" than their colleagues responsible for less developed sublocations.

most upper lower! least TOTAL
progr. middle middle progr.
no loans 66% 73% 85% 95% 79%
1 or more loans 34% 27% 15% 5% 21%
TOTAL 100% 100% 100% 100% 100%
BASE 73 91 111 59 339
Source: F.E. Chege (in prep.)
Nearly one-third of the most progressive farmers have obtained one or more loans, as against one-twentieth of the least progressives.

And that goes for adopters with second mouth information who adopted the
same year. The experience with clinics shows what to expect for
those further down the trickle down process.
It is probably distortion of information and possibly complete blockage
of further information transmission because of the absence of channels
between subsystems, that can account for such situations as we encountered
in Kenya: In the area where we worked, hybrid maize had been introduced
in 1963. Maize is the local staple food, There were no complaints about the taste
of the new varieties, while farmers growing local maize.had very low-yields,
forcing them to buy maize every year, and making the area- a net importer.
Farmers who adopted hybrid were enthusiastic about the results. In our
experimental training we had no trouble achieving 100% adoption rates
among farmers who had not yet adopted hybrid. The Kikuyu's who inhabit.
the'area are known as one of the most change-prone peoples of Africa.
Yet, when we did our survey in 1970, only 31% had-adopted hybrid maize.
Autonomous diffusion processes seem indeed very unreliable when it
comes to solving such problems as the threatening food crisis.
.. There is another, more psychological reason, for imperfections of
diffusion processes in equalising outcomes. As inequities begin to
emerge, farmers start to experience feelings of relative success and
failure. Those who experience success begin to learn that they can
determine their own fate ("efficacy" (Smith and Inkeles, 1966),
"expectancy of internal control" (Rotter, 1966)).
Thus, success. breeds success, while those who experience failure lose
confidence. The highly efficacious start seeking information more eagerly
and get more of it, a reason for Havelock et al (1969, p. 11, 25) to
remark that knowledge tends to go where there is most of it already 'and also a reason why the model of the problem solving, 'information seeking
client, as advocated by Westley (1970), may be a dangerous one to follow'
for a change agent ..... if he is concerned with equity, that is.
Other reasons for the imperfections of diffusion processes as equaliserscould probably be mentioned. However, the reasons given are sufficient to draw a few conclusions.

1. Diffusion processes do indeed lead to inequitable development unless.
correcting measures are undertaken. At present, the tendency ofdiffusion processes to enhance inequit is only reinforced by the services,
which follow progressive farmer strategies.
2. The normative role which diffusion theory de facto plays, is detrimental
for equitable development. Diffusion researchers, including some of the
present authors, can even be blamed for the error of ommision of not having
highlighted the inequity-producing aspects of innovation diffusion, although
these aspects have always been very clear from their main findings.
The researchers might have focussed on how to avoid what they predict. 3. The third conclusion somewhat tentative and is meant to be of
heuristic value. For a long time, students of modernization were interested in studying individual modernity with a view to determine the factors which change traditional into modern people, i.e., to discover the key for making
the world modern. The assumption thus was that the following sequence exists:
However, the data collected in-this pursuit could very well be interpreted
differently. The fact that one finds laggards and progressives with very much
the same characteristics as traditional and modern people also in developed nations may mean that research on modernity has used an incorrect paradigm.
It may well be that "modern" and "traditional" are dyadic concepts, i.e.,
the one is there because of the other.
A "traditionalist", "laggard", farmer may have few resources, is not reached
by extension and may have lost confidence in his ability to affect his future
precisely because "progressive ", "modern", farmers have resources, are reached by Government services and have confidence. The paradigm may thus read as follows:
That the traditionalist may, once inequities have increased, become severely frustrated in his effort to soothe his -relative poverty, has been explained

elsewhere, as well as the various forms his reactions to frustration maytake (Roling, 1970). Suffice it to say herethat modernity may not be a desirable state to be reached by everyone. It may be a state of mind which can only be maintained by virtue of the presence of traditionalists. Modernity may be the product of an inequitable development.
If one takes seriously the need for a rapid spread of new technology and the need for a more equitable development in the rural areas of developing nations, it seems unavoidable to seek alternatives to the progressive farmer strategy and its concomitant reliance on autonomous diffusion to others. Consequently. it behooves diffusion researchers to experiment with such alternatives. In Kenya, efforts have been going on to do just that. The aim was to find replicable strategies for accelerating the flow of income-generating inno.vations to less progressive farmers. We shall not describe the whole experiment here. The design, research method and main results haven been published elsewhere (Ascroft et al (1973,), Roling (1974 a), Roling (1974 b)). Suffice it here to describe our method of farmer selection and the main results we obtained.
We asked extension workers to select, in their sublocation, 25 farmers for a special farmer training course on hybrid maize, by taking only those who did not have hybrid maize, coffee and grade cattle on their farm. The extension workers did, at first, not believe we could succeed, so most of them picked fairly progressive farmers or people, such as city workers who recently returned home, whom they felt could easily "make it". As a result, the participants in our first set of experimental courses were, on average, only slightly below average compared to the random-sample, and included many most progressive farmers.
We changed our strategy for the second set of courses, making it less replicable in the process. We explained the selection procedure more clearly to the extension workers and discussed their doubts. To please them, we allowed them to select up to three progressive farmers as "examples" for each group of 25. To select the rest, the extension workers filled out cards for 40 less progressive farmers lacking common innovations on their farms. On these cards, we asked for such information as farm size, innovations adopted, etc.

From the 40 cards, we ourselves picked the 25 smallest, least progressive. farmers. Table 1 shows the result of this selection procedure.
participants random sample
most progressive 5% 26%
upper middle 15 25
lower middle 38 25
least progr. 42 24
TOTAL 100% 100%
BASE 308 253
Only lower areas of Tetu from which also the participants were drawn.
The table shows that we succeeded in recruting less progressive farmers. In fact, 80% were below average. There was no trouble in getting these farmers to come to the courses. As a result of them, nearly all of the 308 farmers (97%) accepted the credit offered, bought the inputs and planted the, hybrid maize promoted during the course. What's more, for every farmer trained, about three other farmers also adopted in the same year, be it that some aspects of the innovation package did not get perfectly diffused. If "interpersonal diffusion" is mostly homophilous, one can expect that the second-mouth adopters were also less progressive. Unfortunately, we only had time to check this homophily for the diffusion effect of the first set of courses. For these courses, the second-mouth adopters were slightly less progressive than those trained, a result which conforms to expectations based on other research (Rogers and Shoemaker, 1971, p. 211). We cannot claim that one field experiment (with all the shortcomings. field experiments are bound to have, considering such factors as ongoing government activities, the difficulty to find controls (we used the sample survey results), etc.), carried out among a very gung-ho people (the Kikuyu's), on an innovation which had been introduced about nine years earlier, can prove our-case. Yet, we feel to have shown that it seems possible to find alternatives to the progressive farmer strategy which work and have better consequences for the equity of the ensuing development.

As'a result of our field experiment, we came to believe that it is not the characteristics of the farmers, such as their progressiveness, which are the prime determinants of the outcomes of development efforts. We believe rather, that it is the characteristics and deployment of Government services which matter, and especially their ability to provide and structure opportunities. Of course, the Government services are the more manipulable factor in rural development processes.

Ascroft, J., et al (1971) "The Tetu Extension Pilot Project", in Strategies for Improving Rural Welfare, Nairobi (Kenya): Institute for Development Studies, Occasional Paper 4, pp 37-80. Ascroft, J., et al (1973) Extension and the Forgotten Farmer,
Wageningen: Afdelingen Sociale Wetenschappen van de Landbouwhogeschool, no. 37. Ban, A.W. van den (1963) Boer en Landbouwvoorlichting, Assen: van Gorcum Chege,.F. (in prep.) Thesis, East Lansing (Mich): MSU, Dep. of Ag. Econ. Freebairn, D.K: (1973) "Income disparitiesin the agricultural sector regional and institutional stresses", chapter 6 in: T.T. Poleman and D.K. Freebairn (eds). Food, population and employment: the impact of the Green Revolution.
New York: Praeger, pp 97-119. Griffin, K. (1972) The green revolution: an economic analysis
Geneva: UNRISD;
Havelock, R.G. et al (1969) Planning for Innovation through Dissemination and Utilization of Knowledge, Ann Arbor (Mich.): Center for Research on the Utilization of Scientific Knowledge.
Leonard, D.K. (in prep.) Thesis, Davis (Calif.): USC Lowdermilk, M. (1972) Diffusion of Dwarf Wheat Production Technology in Pakistan's Punjaab, Ithaca: Cornell, Dep of Ext. Edac. Dissertation

Mbithi, P.H. (1972)
"Innovation in Rural Development", Nairobi: Institute for Devel~pment studies. Discussion paper no. 158.
.McNamara, R.S. (1972)
Address to the Board of Governors of the World Bank, IMF, Washington (DC): IBRD McNamara, R.S. (1973)
Address to the Board of Governors of World Bank, IMF, Narobi (Kenya).
Republic of Indonesia (1973) "Sempor Dam and Irrigation Project: Quarterly Progress Report, Appendix II b, Irrigation and Drainage, Agronomy and Agric. Extension, January through March
Djakarta (Indonesia): Ministry of Public Works and Electric Power, Directorate General of Water Resources Development, Directorate of Irrigation. Robertson, T.S. (1971) Innovative Behavior and Communication, New York: Holt, Rinehart and Winston. Rogers, E.M. (1962)
Diffusion of Innovations, New York: Free Press .Rogers, E.M. and F.E. Shoemaker, (1971) Communication of Innovations: A cross-cultural Approach. New York: Free Press. Ro1ing, N.G. (1970)
"Adaptations in development: a conceptual guide for the study of non-innovative responses of peasant farmers", Economic Development and Cultural Change, 19 (1), pp 71-85.
Roling, N.G. (1972)
"Determining the level of agricultural development in Tetu sublocations: towards measuring extension impact".
Nairobi (Kenya): Institute for Development Studies, Working paper 33.

Rl61ing, N.G. (1974 a) "From theory to action", Ceres, 7 (3), May-June, 1974, pp 22-25.
Roling, N.G. (1974.b) "Forgotten farmers in Kenya", Agricultural Progress, (in press).
Rotter, J.B. (1966) "Generalised expectancy of external versus internal control of reinforcement", Psych. Monographs, 80 (1), whole 609.
Sen,. L.K. and D.K..Bhowmik (1970), "Opinion leadership and interpersonal diffusion" chapter 5 in:
Rogers, E.M., J. Ascroft and N. Roling, Diffusion and Innovations in Brazil, Nigeria and India, East Lansing (Mich.): MSU, Dep. of Communication.
Sinha, B.P. and P. Mehta (1972), "Farmers' need for achievement and change pronesess in acquisition of information from a farm-telecast", Rural Sociology, 37 (3): pp 417-427. Smith, D.H. and A. Inkeles (1966), "The OM scale, a comparative socio-psychological measure of individual modernity", Sociometry 29 (4), pp 253-377.
Westley, B.H. (1970), "Communication theory and general systems theory: implications for plannedchange", Paper for the Annual Conference of the American Assoc. for Public Opinion Research.