beyond higher yields: the impact of sweetpotato …...cip program report 1999 – 2000 331 beyond...

CIP Program Report 1999 – 2000 331

Beyond Higher Yields: The Impact of SweetpotatoIntegrated Crop Management and Farmer FieldSchools in Indonesia

E. van de Fliert1, N. Johnson2, R. Asmunati3, and Wiyanto3

A pilot program of sweetpotato (Ipomoea batatas (L.) Lam) integrated cropmanagement (ICM)-farmer field schools (FFSs) was implemented in six com-munities in Indonesia, using protocols developed jointly by a team of farmers,researchers, and development workers. Monitoring and evaluation studiesshowed that participation in the FFS enhanced farmers’ crop managementknowledge and skills. Several of their changed cultivation practices led toincreased net income as a result of reduced cultivation costs and/or in-creased yields. Farmer participation in research was shown to havecontributed to the relevancy, appropriateness, and impact of the sweetpotatoICM-FFS protocols.

During 1994–97, CIP, with support fromUPWARD (The Users’ Perspectives forAgricultural Research and Development isa CIP-affiliated network of Asian research-ers conducting participatory R&D projectsin root crop systems.) and in collaborationwith public and private sector groups,implemented a project to develop aprotocol for a sweetpotato ICM-FFS inIndonesia. Collaborators were Mitra Tani,a local nongovernment organization(NGO); the National Research Institute forLegume and Tuber Crops; and DutaWacana Christian University (Table 1).Project activities were implemented inmajor sweetpotato growing areas in Eastand Central Java, where sweetpotato isgrown as an important cash crop through-out the year, mostly in rotation with rice.The project strategy relied on participatoryapproaches and methods at all stages:needs assessment and project design, ICM

1 CIP-ESEAP, Bogor, Indonesia.2 CGIAR-PRGA and CIAT, Cali, Colombia.3 Mitra Tani, Yogyakarta, Indonesia.

development and farmer learning protocolsapplying the FFS approach, pilot-scaleimplementation of the sweetpotato ICM-FFS, and monitoring and evaluation. (Fordetails of project strategy, activities, andoutputs, see Van de Fliert et al., 1996; Vande Fliert and Braun, 1997; Braun and Vande Fliert, 1997; Asmunati et al., 1999; Vande Fliert, 1999.)

To institutionalize the sweetpotato ICM-FFS model that was developed and toallow for large-scale farmer learning andimplementation, staff from the NationalIPM Program (NIPMP) and 30 local NGOsunderwent FFS facilitators’ training:NIPMP staff in June 1997, NGOs staff inApril 1998. Follow-up programs wereimplemented and funded by these localextension organizations, and a secondresearch project was initiated to evaluatetheir activities over a 2-yr period (1998–99) (Table 1). The work was carried out byMitra Tani, with methodological andfinancial support from CIP and UPWARD.

332 Research on Sweetpotato

Rigorous process and impact evaluationwere considered critical elements of athorough documentation of thesweetpotato ICM research, development,and dissemination experience.Sweetpotato ICM is knowledge-intensiveand site-specific. Its goal is to achievesustainable, collective agroecosystemmanagement by well-informed and skilledfarmers (Van de Fliert and Braun, 1999).The FFS learning process provides farmersnew knowledge and skills through discov-ery and experimentation (Van de Fliert,1993). Therefore, the monitoring andevaluation study not only measuredeconomic impact of ICM, but analyzedthe FFS implementation and disseminationprocess. In addition, it assessed impact inrelation to changed farmer capacity andcultivation practices, and farm-leveleffects using additional indicators forhuman, social, and environmental capital.

In 2000, the sweetpotato ICM-FFS projectwas selected as one of three cases to beincluded in a study on the costs andimpacts of participatory research andgender analysis conducted and funded by

the Consultative Group on InternationalAgricultural Research (CGIAR) systemwideprogram on Participatory Research andGender Analysis for technology develop-ment and institutional innovation (PRGA).The objective of the PRGA study was toassess the implications of incorporatingdifferent types of user participation intoresearch on farmers’ natural resourcemanagement, with specific attention onwomen and the poor. Four categories ofimpacts are evaluated: 1) adoption andimpact of the technologies developed,2) strengthening of human and socialcapital among participating individualsand communities, 3) establishment offeedback links to formal research, and4) costs of doing research.

CIP/UPWARD and PRGA impact evalua-tions of the ICM-FFS activities arecomplementary. While both studies lookedat the impact of FFS implementation onsweetpotato farmers’ knowledge andproduction practices, the PRGA studyadditionally focused on the role of userparticipation in the development of theICM-FFS. Taken together, their results

Table 1. Overview of activities of the CIP/UPWARD-suppor ted sweetpotato ICM-FFS1 development and evaluation projects in Indonesia, 1994-99.Research phase Needs assessment

and baseline2Technology (ICM)

development3FFS

development andinstitutionalization4

PM&E ofSP ICM FFS

implementation5

Project I: Sweetpotato ICM and ICM-FFS development1994/95 wet season PRA, RK, FO1995 dry season RK, FO 8 Pilot FFS in SP1995/96 wet season RK, FO 17 (3) Write modules1996 dry season 14 (6) Pilot SP ICM FFS1996/97 wet season 12 (2) Revise Modules1997 dry season 6 ToT NIPMPProject II: PM&E5

1997/98 wet season Process M&E1998 dry season ToT NGOs Process and impact1998/99 wet season Impact evaluation1999 dry season Analysis1 ICM-FSS = integrated crop management-farmer field school.2 PRA = participatory rural appraisal methods, RK = season-long record keeping, FO = field observations.3 Number of farmer-managed trials; number of researcher-managed trials in parentheses.4 NIPMP = National Integrated Pest Management Program, Indonesia.5 PM&E = participatory monitoring and evaluation.


should provide insight into how userparticipation influenced the orientationand activities of the research project, andhow those changes in turn affected thefinal impacts of the project technologies.

Methods

The CIP/UPWARD monitoring and evalua-tion study applied an analytic frameworkdeveloped with all stakeholders during aworkshop at the beginning of the project.Indicators for impact evaluation had beendetermined with the stakeholders earlierduring the ICM research project. Theframework considered the various programimplementation levels for monitoring andevaluation: 1) training of facilitators,2) FFS implementation, 3) horizontaldissemination within the community,4) ICM implementation by trained farmers,5) farm level effects, and 6) implicationsfor research and development institutions.For each evaluation level, informationsources, variables, and indicators weredetermined, and data collection methodsdesigned as follows.

Pre-FFS:

• Observations and internal evaluationduring facilitator training.

• Individual qualitative baseline inter-views with FFS participants (who laterbecome ICM farmers).

During FFS:

• Process observation and recordingduring FFS sessions.

• Pre-tests/post-tests on ICM farmers’knowledge.

Post-FFS:

• Evaluation meetings with participantsand facilitators.

• Individual interviews with facilitators,village officials, and traders.

• Two-step individual interviews with ICMfarmers (FFS graduates) and randomlyselected non-ICM farmers.

• Season-long record keeping of post-FFSsweetpotato cultivation by ICM farmers.

• Observations in ICM and non-ICMfarmers’ fields.

• Annual evaluation and planning work-shops with all stakeholders.

Farmers were intensively involved inmethods design, data collection andanalysis, and presentation during a finalseminar with research and extensionpolicymakers.

The PRGA study was designed to investi-gate the benefits and costs of includingusers in research, particularly on howimpacts vary depending on the nature ofparticipation. The conceptual framework isbased on a typology of participation,which defines types of participation inrelation to who makes decisions aboutresearch priorities and activities (Lilja andAshby, 1999). For each of the stages of theresearch process (design, testing, anddissemination) hypotheses were developedrelating type of participation to each ofthe four categories of expected impacts ofdifferent types of participation (Johnsonand Lilja, 2001). These hypotheses helpedguide the design of the empirical methodsin each of the three cases. The hypothesesfor the ICM-FFS study, which used pre-dominantly collaborative participation, areshown in Table 2.

Both qualitative and quantitative data andanalyses were used to examine thehypotheses regarding participation andimpact. The study relied greatly onavailable reports and economic data fromthe project. Available implementation andproduction data were analyzed to assesseconomic impacts. PRGA and CIP staffcollected additional data on human andsocial capital impacts of participationduring a field trip in August 2000 byconducting semi-structured interviews withparticipants and key informants in theproject communities. Additional inter-views with scientists, policymakers, andextension and development workers were


conducted to assess the extent to whichthere was feedback from the project to thestakeholder groups. Finally, through theanalysis of the available data sets andinterviews with project staff, the PRGAstudy team assessed the overall impact ofuser participation on project goals, activi-ties, outputs, and costs.

Results

Due to limitations of space and the factthat the PRGA analysis is still in process,only a limited number of key results areincluded here. The first set of resultspertains to measurable impacts of theICM-FFS on farmers’ knowledge, skills,production practices, and social relationsin the FFS villages. This analysis isfollowed by a discussion of the role offarmer participation in the design anddevelopment of the ICM-FFS, and itsimplications for impact.

Impact of ICM-FSS on farmer knowledge,skills, production practices, and socialrelations.FFS implementation and process evalua-tion. The impact assessments wereconducted in the six communities whereNIPMP conducted pilot-scale sweetpotatoICM-FFS in 1997/98. Four of the districtswere the same as where the earlier ICMresearch project had been conducted, i.e.,Mojokerto and Magetan in East JavaProvince and Karanganyar and Magelang

in Central Java. (In East Java, the FFSswere implemented in the same hamlets,although with different farmers than thoseinvolved in FFS development.) The twoadditional districts were Sleman inYogyakarta Province, and Kuningan inWest Java Province. Because thesweetpotato FFSs were organized for riceIPM-FFS alumni groups, the actual villagesselected in Magelang and Sleman werenot really major sweetpotato growingcommunities. Selection of ICM farmerswas based on the NIPMP criteria for FFSimplementation, and basically involvedthe rice IPM-FFS alumni.

An overview of sweetpotato ICM-FFSimplementation in the six NIPMP sites isgiven in Table 3. All FFSs conducted atleast 17 sessions. Although an initial 25–28participants were selected in each FFS,the average attendance per session rangedfrom 16–19 participants, which is compa-rable to IPM-FFS implementation in othercrops (Eveleens et al., 1996). Althoughonly 27% of FFS participants were women,they were substantially represented inthree communities: Magelang (49%),Sleman (56%), and Kuningan (60%). Thevariation in women’s participation reflectsthe regional variation of women’s involve-ment in sweetpotato cultivation. Inaddition to curriculum activities, all FFSgroups designed and conducted their ownexperiments to test and refine the ICMguidelines and to practice experimental

Table 2. Hypotheses for the ICM-FFS study.Stage Type of participation HypothesisDesign Consultative/Collaborative Farmer participation results in an increase in the proportion of the

targeted beneficiary group that could be reached by the projectbecause the priority topic chosen for research is more relevant tothe needs and priorities of targeted farmers (H1).

Testing Collaborative Farmer participation at the testing stage results in an increase inthe number of potential adopters within the target group since thespecific technology or technologies selected for recommendationare more appropriate, given farmers' criteria and constraints (H2).

Dissemination Collaborative There is an increase in the probability that potential adopters forwhom the technology or technologies are appropriate will beaware of them and be willing and able to adopt them and torecommend them to others (H3).


methods. The experiments dealt witheither fertilization or varietal evaluation.

Both training-of-facilitators and FFSmodels increased knowledge amongtrainees and demonstrated potentialimpact of ICM in sweetpotato cultivation.Knowledge increase of FFS participantsranged from 28% to 254% per group (Table3). Increased yields (Figure 1) and moreeffective or efficient use of external inputswere demonstrated on collective learningplots in most of the sweetpotato ICM-FFSs.The farmers generally credited the FFSprocess with adding to their knowledgeand skills. Knowledge areas singled outincluded pests and diseases, naturalenemies, seed health, soil health, andplant nutrient management. Skills areasincluded routine field observations,experimentation, fertilization, seedbedpreparation, pest management (none orreduced pesticide use), and vine lifting.

Effects at farmer level (human capital). Toassess the impact of the FFS on partici-pants’ knowledge and skills, post-FFScomparisons were made between farmers

who had attended the ICM-FFS (ICMfarmers) and those who had not (non-ICMfarmers). Regarding production-relatedknowledge and practices, results show thatICM farmers are more likely to understandthe concept of natural enemies and do amore thorough job on field observationthan non-ICM farmers (Table 4). Enhancedknowledge and skills resulted in severalchanges in farmers’ sweetpotato cultiva-tion practices. More ICM than non-ICMfarmers practiced seed selection and goodwater management. Fertilization practiceschanged in that more ICM farmers appliedmore balanced fertilization. ICM farmerswere significantly more likely to usepotassium (K) fertilizer than non-ICMfarmers (P < 0.05), and they applied higheraverage rates. More ICM than non-ICMfarmers used organic fertilizer, albeit atlower application rates. Several seasonslater, however, farmer researchers reportedthe use of organic manure as one of themajor benefits of ICM. Over time stillmore farmers are beginning to use it andunderstand its benefits. Surprisingly,pesticide application frequency was not

Table 3. Overview of implementation and process evaluation of six NIPMP-conducted sweetpotato ICM farmer field schools, 1997-98.DistrictVariable

MojokertoEast Java

MagetanEast Java

KaranganyarCentral Java

MagelangCentral Java

SlemanYogyakarta

KuninganWest Java

Major sweetpotato growing area

yes yes yes no no yes

FFS implementation season

Jan–Jun 98(wet/dry)

Aug–Dec 97(dry/wet)

Sep 97– Feb98 (wet)

Sep–Dec 97(dry/wet)

Aug–Nov 97(dry/wet)

Nov 97–Apr98 (wet)

Meetings (no.) 17 17 17 18 17 17Participants (range) 10–25 11–27 14–27 11–27 11–27 13–28 Meetings (avg. no.) 16 18 19 18 16 17Women participation (% of trainees)

0% 0% 7% 40% 56% 60%

Experiments conducted1

Fertilization(application

rate; organic)

Varieties; Napplication

rates

Fertilization(potassium,N-K rates)

Varieties Varieties Organicfertilizer

Avg. knowledge increase between pre- and post-test2

56%(13)

28%(20)

48%(24)

126%(25)

40%(10)

254%(19)

1 In addition to standard FFS trials.2 Pre- and post-tests consisted of 10–15 questions each. Scores correspond with the number of correctly answered questions. Percentage knowledge increase is calculated by ((score post-test – score pre-test) / score pre- test)* 100% The figures in the table represent the average knowledge increase of all participants per FFS who did both pre- and post-test (number of respondents given in parentheses).


Figure 1. Sweetpotato storage root yield, market price, gross and net income, and cultivation costs at six ICM-FFSsites, by baseline farmers (1994/95 wet season), on the sweetpotato ICM-FFS plot (for seasons, see Table 3), andby ICM and non-ICM farmers during a post-FFS season. No post-FFS data were available for non-ICM farmers inMagelang. Note: The high yields obtained on the FFS learning plots likely result not only from application of the ICMpractices, but also from the fact FFS plots are often relatively fertile and intensely managed. Moreover, they are quitesmall and extrapolating kg/plot yield to t/ha would be expected to overestimate the yield gains.

0 10 20 30 40 50market price of sweetpotato storage roots

(US$/t)

0 10 20 30 40 50Sweetpotato cultivation cost

(US$/ha)

0 250 500 750 1000Gross income from sweetpotato

(US$/ha)

0 250 500 750Net income from sweetpotato

(US$/ha)

Mojokerto

Magetan

Karanganyar

Magelang

Sleman

Kuningan

Sweetpotato storage root yield (t/ha)

Mojokerto

Magetan

Karanganyar

Magelang

Sleman

Kuningan

Mojokerto

Magetan

Karanganyar

Magelang

Sleman

Kuningan

0 10 20 30 40 50 60

Non-ICM farmers, post-FFS

ICM farmers, post-FFS

FFS plot


influenced by ICM training. Thirty percentof ICM farmers applied 1 or 2 applicationsof pesticides on the sweetpotato cropduring the post-FFS season, which isexactly the same as before training, andslightly more than among non-ICM farmers(21%), who, however, give only 1 or 2applications as well. One reason may bethe already low levels of pesticide use inthese areas. More ICM than non-ICMfarmers, however, began to implementfield sanitation to manage sweetpotatoweevil.

In addition to cultural practices, the FFSalso included yield estimation of the cropin the field. Sweetpotato farmers in Javaoften sell their standing crop, so the abilityto accurately assess yields is an advantagewhen negotiating a selling price. Duringthe needs assessment, farmers had identi-fied their weak bargaining power as amajor constraint to achieving a goodincome from sweetpotato.

However, yield assessment skills of ICMfarmers did not visibly improve (Table 4).Nevertheless, attention to marketing in theICM-FFS seemed to have helped ICMfarmers in all except one district tonegotiate for a somewhat better price(Figure 1), although the data are onlysignificant for Mojokerto (P = 0.063). InMagetan, ICM farmers effectively re-ceived a lower per kg price than non-ICMfarmers (P = 0.056), because they wereunable to negotiate a higher per area unitprice with the traders, even though theyobtained higher yields.

Another important indicator of humancapital is the ability of farmers to experi-ment in their own fields. We found nodifference in experimentation betweenICM and non-ICM farmers. However, thatmay be due to a bias in the data. Nearlyall the non-ICM farmers who reportedexperimenting were from Mojokerto,where the farmer researchers who workedon the development of the ICM-FFS

Table 4. Differences in knowledge and skills, practices, and inputs and outputs of sweetpotato cultivation during a post-FFS season by ICM-FFS participation. Magelang is excluded from the analysis because there are no non-ICM cases in this community.Variable ICM (N=73) Non-ICM (N=50)Knowledge and skills Understand the concept natural enemy (%) 86 52 Practice routine/thorough field observation (%) 75/47 75/29 Conduct experiments (%) 18 16 Can assess yield within reasonable range (%) 48 53Cultural practices Practice seed selection (%) 45 37 Practice good water management (%) 68 58 Use N-fertilizer (%/kg N/ha) 96/223 92/242 Use P-fer tilizer (%/kg N/ha) 68/78 60/58 Use K-fertilizer (%/kg N/ha) 28/20 13/7 Use organic fertilizer (%/t/ha) 35/1.3 29/1.6 Pesticides applications (avg. no./season) 0.4 0.3 Practice field sanitation after experiencing weevil attack (%) 49 30Inputs and outputs1

Value of hired labor (US$/ha) 107 112 Value of inorganic fer tilizer use (US$/ha) 29 27 Value of total cultivation cost (US$/ha)2 201 237 Yield (t/ha) 20.5 19.5 Price received (US$/t) 32 30 Net income (US$/ha)2 391 3161 Exchange rate of Rp. 8,500/US$1.00.2 Significant at P < 0.05.


curriculum during 1994/95 had a particu-larly strong influence, resulting in anunrepresentative global average for non-ICM farmers. In Karanganyar andKuningan, the FFS groups continued withcollective rather than individual experi-mentation on ICM components in thesweetpotato crop after the FFS season.Observations of the study team confirmthat ICM farmers do have better skills andmore interest in experimentation.

Effects at the farm level (financialcapital). Changes in knowledge, skills,and practices appear to have translatedinto a 24% higher average net income forICM farmers compared with non-ICMfarmers. ICM farmers obtained slightlyhigher (5%) sweetpotato yields than non-ICM farmers. That was not, however,significant over the whole data set (Table4). By district, yields differed significantly(by t test) in Magetan and Sleman(P < 0.05) and Mojokerto (P < 0.10). Onaverage cultivation costs of ICM farmerswere 15% less (P < 0.05) than those ofnon-ICM farmers. But on a district basisthe difference was significant only inKaranganyar (P < 0.05) and Magetan(P < 0.10) (Figure 1). ICM farmers tendedto spend slightly, but not significantly,more on inorganic fertilizer inputs, whichwas mainly a result of their higher use ofthe more expensive fertilizers such as K.The cost of land rental and use of hiredlabor was the same for both groups.

Slightly higher yields and market priceswere responsible for ICM farmers receivinga 10% higher gross income fromsweetpotato than non-ICM farmers (Figure1). But the difference was significant onlyin Kuningan (P < 0.10). Due in part to theirlower cultivation costs, ICM farmers didachieve a significantly greater net in-come/ha than non-ICM farmers (P < 0.05),particularly those in Mojokerto (P < 0.10)and Karanganyar (P < 0.05).

The ICM-FFS teaches a broad range oftechnologies and practices relating notonly to crop production but also to farm

management and marketing. One way toassess the combined impact of the newknowledge, skills, and practices is to lookat how ICM-FFS attendance affects theoverall profitability of sweetpotato produc-tion. The dependent variable in the profitfunction is net income/ha fromsweetpotato production. Independentvariables that are expected to influenceprofitability include expenditure onfertilizer (mean value = US$43/ha) andhired labor mean value = US$100/ha), anddummy variables for land tenure status(26% farmed rented land), for whether thefarmer attended ICM-FFS (60% attended),and for water management practices used(59% used routine rather than sporadicirrigation). Community dummy variableswere also included to control for theinfluence of local conditions. A Cobb-Douglas functional form was assumed.

The results of the analysis show thatparticipation in ICM-FFS is significantlyand positively associated with net incomefrom sweetpotato production (Table 5).(One explanation for the fact that partici-pation in ICM-FFS is positively andsignificantly associated with net income

Table 5. Results of estimation of sweetpotato profit function coefficients; dependent variable is log of net income per hectare (N = 81, adjusted R2 = 0.50). Magelang is excluded from the analysis because there are no non-ICM cases in this community and Sleman is excluded because it is not an impor tant sweetpotato growing area.

Standardizedregressioncoefficient

P

(Constant) 0.000Log of hired labor costs/ha 0.089 0.335Log of fertilizer costs/ha 0.102 0.375Water management dummy (= 1 if irrigation is routine as opposed to sporadic)

0.290 0.002

ICM participation dummy (= 1 if attended ICM-FFS)

0.271 0.002

Land rental dummy (= 1 if rents)

-0.382 0.000

Karanganyar dummy 0.387 0.001Kuningan dummy 0.452 0.000Magetan dummy -0.059 0.553


from sweetpotato production could be thatthe ICM-FFS participants were alreadybetter farmers than nonparticipatingfarmers. To rule out this possibility,baseline production data collected beforethe development of the ICM technologiesand practices were analyzed, and did notshow that farmers who would subsequentlyparticipate in ICM-FFS had higher yieldsor net incomes than those who did not. Infact, the opposite was true, suggesting thatthe results presented in this paper mayunderestimate the economic benefit of theICM-FFS.)

Participants in ICM-FFS have higher netincomes than non-participants, aftercontrolling for other factors such as landtenure and water availability. The commu-nity dummy variables are significant forKuningan and Karanganyar, the communi-ties with very favorable ecologicalconditions for sweetpotato production, andwhere the response to the ICM-FFS wasmost enthusiastic. Neither fertilizerexpenditure nor hired labor costs weresignificantly associated with profitabilityof sweetpotato production. As mentionedearlier, ICM farmers use more costly K-fertilizer but less N-fertilizer than non-ICMfarmers (Table 4). In an earlier version ofthe analysis, gender was included as adummy variable, however, it was notsignificant and therefore results are notpresented here.

Effects at the community level (socialcapital). The collective activities of FFSsare carried out in a group and are intendedto strengthen the capacity of farmers towork together, to share information, and tolearn from each other (Van de Fliert,1993). To assess impacts of the ICM-FFSon these social capital variables, datawere collected on farmers’ practicesregarding sharing information about theFFS and about agriculture in general, andabout their participation in group activitieswith other FFS members. The majority ofICM farmers (68%) reported disseminatingsweetpotato ICM information to otherfarmers or neighbors. There was, however,

wide variation between the communities.In five of the six communities, over 90%of participants talked about the FFS withothers, but in Magetan only 41% did so.Magetan, it should be noted, was wherethe research had taken place. Most of thevillagers were aware of at least thegeneral issues of the sweetpotato ICM-FFS.

On average, each participant talked with3 people. Nearly all the informationexchanges on sweetpotato ICM occurredin the field (53%) or in home/groupmeetings (42%). In Karanganyar andSleman, dissemination also occurredbeyond the village where the ICM-FFSwas conducted. Men and women differedsignificantly in whom they talked to aboutFFS. Men talked much more with neigh-bors and almost exclusively with othermen. Women talked much more withrelatives and similarly almost exclusivelywith other women. ICM informationtransmitted to other nonparticipatingfarmers mainly dealt with variety andhealthy seed selection (62%), plantnutrient management (54%), sweetpotatopests and diseases (35%), soil preparationand health (10%), and yield assessment(10%). Spontaneous diffusion of informa-tion between farmers, however, cannot beconsidered an impact of the project.Interviews with farmers and other keyinformants suggest that the communitieshave a long history of sharing agriculturalinformation. The high levels of spontane-ous dissemination are related to highexisting levels of social capital in thecommunities. In several communities, FFSgroups have continued to engage incollective experimentation, somethingthey did not do before participating in thesweetpotato ICM-FFS.

The contribution of farmer participatoryresearch to the impact of the ICM-FFSAs mentioned earlier, the research processthat led to the development of the ICMtechnologies and practices and the ICM-FFS curriculum were developed in aparticipatory mode. Farmer input was


obtained from needs assessment data andfrom an intensive process of collaborativeresearch with eight farmers in four commu-nities over a period of 3 years. Bothfarmers and researchers conducted experi-ments, and regular workshops were held topresent findings, analyze results, and planfuture work.

The importance of farmers’ contributions tothe research process was made clear earlywhen it was discovered that farmers didnot believe that pests were the mainconstraint to sweetpotato production. Theoriginal CIP/UPWARD project had calledfor the development of an integrated pestmanagement (IPM) FFS curriculum.Farmer input resulted in a shift from IPM toICM. Within the ICM framework, topicsthat farmers identified as important and onwhich they carried out experimentsincluded plant nutrient management,varietal selection, and cultural practicessuch as planting methods, vine lifting, andintercropping. In four of these five areas,the farmers’ experiments generated resultsfor practices and guidelines that wereincluded in the final curriculum. Theexperience working with farmers alsocontributed to the ICM-FFS focus onhealth, and its emphasis on teachingprinciples and experimental methods.

These impact results show that the majorareas in which the FFS enhanced farmersknowledge, productivity, and profitabilitywere in plant nutrient management, watermanagement, varietal and plantingmaterial selection, and overall betterobservation skills. An FFS focused entirelyon pest management would likely havehad much less impact. Fertilizer and seedselection were by far the two topics mostcommonly mentioned by ICM-FFS partici-pants when they talked to other farmers.That suggests that they are the elementsthat will be most widely disseminatedinformally. The results regarding individualand group experimentation indicate thatan ongoing process of productivity im-provement may have been set in motionby the ICM-FFS process.

Conclusions

Sweetpotato ICM-FFS has contributed toincreased knowledge and skills amongfarmers who participated. These humancapital increases have resulted in concretechanges in production and marketingpractices that contributed to higher yieldsand higher income from sweetpotatoproduction. These results were essentiallyachieved by improving the efficiency ofexisting practices rather than by introduc-ing totally new technologies. Farmerparticipation in the research process,particularly in the identification of pos-sible areas for improving efficiency,appears to have contributed to both therelevance and the impact of the ICM-FFS.

Whereas most of the impacts shown in thisstudy are significant statistically, they mayseem to be less so in practical terms. Themagnitude of the differences of individualparameters between ICM and non-ICMfarmers are often relatively small, buttaken as a whole they indicate the initia-tion of change in farmers’ cropmanagement behavior. As mentionedearlier the data may underestimate thetrue differences between ICM and non-ICM farmers, mainly as a result of theinability to control for spillover effectsfrom participants to nonparticipants. Giventhat ICM is complex and that increasedknowledge does not immediately result inchanged practices, it is likely that overtime the benefits from ICM-FFS may grow.Subsequent field visits as part of the PRGAstudy gave that impression. A final factorthat could have affected the magnitude ofthe observed impacts is the Asian financialcrisis that hit Indonesia in 1997. Theprofitability of sweetpotato was drasticallyreduced, hence the reduced incentives forfarmers to invest in it.

Nonetheless, the question of whether thebenefits justify the costs of ICM-FFS isvalid. This issue will be examined ingreater detail in the forthcoming PRGAstudy. Participatory development of thesweetpotato ICM-FFS took several years


and was intensive due to the amount ofcomponent research required to fill thegap of technological guidelines for thecrop. In addition, a broader and morewidely applicable FFS model was devel-oped and documented. That facilitatedsimilar CIP projects in other countries inand beyond the region. To assess cost-effectiveness of the project, it would haveto be compared with a conventional CIPresearch project with similar scope aimedat both productivity enhancement andnatural resource management, rather thanwith a conventional extension-orientedFFS.

The six pilot field schools were conductedin the selected districts to trigger interestfrom local authorities so that they mightallocate funds for larger-scale implementa-tion. But the economic crisis that hitIndonesia in 1997 onwards, and simulta-neous failures in some major rice bowls inIndonesia during 1998, caused all agricul-tural development to be allocated to riceproduction. Expansion of sweetpotato FFSimplementation within the governmentextension system has not yet beenachieved. The newly established Director-ate of Root Crops Production hascommitted itself to organizing a nation-wide sweetpotato ICM-FFS programwhenever the economic situation allows.It is hoped that the results of this study willcontribute to the implementation of theprogram. Additionally, NGO programshave continued to apply ICM-FFSapporaches in sweetpotato and othercrops, and these efforts will be evaluatedduring 2001-2002.

References

Asmunati, R., Wiyanto, and E. van deFliert. 1999. Integrating sweetpotatoICM and FFS approaches in governmentand NGO initiatives in Indonesia, In:Users’ Perspectives for AgriculturalResearch and Development (UPWARD),Learning to manage livelihoods: Newperspectives in rootcrop R&D.

UPWARD, Los Baños, Laguna,Philippines. p. 129–139.

Braun, A.R. and E. van de Fliert. 1997.The farmer field school approach to IPMand ICM in Indonesia: Userparticipation. In: UPWARD. Local R&D:Institutionalizing innovations in rootcropagriculture. (Users’ Perspectives forAgricultural Research and Development(UPWARD), Los Baños, Laguna,Philippines. p. 44–64.

Eveleens, K.G., R. Chisholm, E. van deFliert, M. Kato, P.T. Nhat, andP. Schmidt. 1996. Mid-term review ofPhase III report. FAO Inter-countryProgram for the Development andApplication of Integrated Pest Controlin Rice in South and Southeast Asia,Makati, Metro Manila, Philippines.

Johnson N. and N. Lilja. 2001.Characterizing and measuring theeffects of incorporating stakeholderparticipation in natural resourcemanagement research: Analysis ofresearch benefits and costs in three casestudies. Participatory Research andGender Analysis (PGRA) WorkingDocument. PRGA/Centro Internacionalde Agricultural Tropical (CIAT), Cali,Colombia.

Lilja, N. and J.A. Ashby. 1999. Types ofparticipatory research based on locus ofdecision-making. Participatory Researchand Gender Analysis WorkingDocument (PGRA) No. 6. PRGA/CIAT,Cali, Colombia.

Van de Fliert, E. 1993. Integrated pestmanagement: Farmer field schoolsgenerate sustainable practices. A casestudy in Central Java evaluating IPMtraining. Wageningen AgriculturalUniversity Papers (93)3 PUDOC.Wageningen, Netherlands. 304 p.

Van de Fliert, Elske. 1999. Integrative,farmer-participatory methodology forpoverty-sensitive research: Sweetpotatoand potato integrated crop managementin Southeast Asia. Paper presented atthe Centro Internacional de AgriculturalTropical International Workshop


Assessing the impact of agriculturalresearch on poverty alleviation held14–16 September 1999, San José, CostaRica, 14 p. Available at: http://www.ciat.cgiar.org/poverty/bar_papers.htm.

Van de Fliert, E. and A.R. Braun. 1997.One step back, two steps forward:Sweetpotato ICM development inIndonesia, In: UPWARD. Local R&D:Institutionalizing innovations in rootcropagriculture. UPWARD, Los Baños,Laguna, Philippines. p. 153–167.

Van de Fliert, E. and A.R. Braun. 1999.Farmer field school for integrated crop

management of sweetpotato: Fieldguides and technical manual.International Potato Center, Bogor,Indonesia. 286 p.

Van de Fliert, E., R. Asmunati, Wiyanto,Y. Widodo, and A.R. Braun. 1996. Frombasic approach to tailored curriculum:Participatory development of a farmerfield school model for sweetpotato. In:UPWARD. Into action research:Partnerships in Asian rootcrop researchand development. UPWARD, Los Baños,Laguna, Philippines. p. 59–74.

beyond higher yields: the impact of sweetpotato …...cip program report 1999 – 2000 331 beyond...

Documents