Socio-economic Research And Monitoring On Pemba And Unguja: Social Anthropologist's Inception Report

MINISTRY OF AGRICULTURE, LTVESTOCK AND NATURAL RESOURCES

ZANZTBAR CASH CROPSX'ARMTNGSYSTEMSPROJECT (ZCCFSP)

MARCH 1995

SOCIO-ECONOMICRESEARCH AND MONITORING ONPEMBAAND UNGUJA SocialAnthropologist'sInception Report

BY: Martin Walsh

zccFsP PO BOX 2283 Ztnzibal. Telephone: Fax :

(054) 33121 (054) 33121

CONTENTS Page

Preface

2

Acronyms

2

INTRODUCTION

J

INITIAL TASKS

J

Introductionsand Institutional Review

J

Contactwith Anthropologistsand Other Researchers

5

Library Researchand Literahre Review

5

Review of Socio-economicResearchand Monitoring in ZCCFSP

6

Socio-economicResearchand the Socio-economicsSection

6

ThePM Programme Farming Systemsand Fanning SystemsZones

8

Itealth Ranbing

11

Monitoring Systems

t4

WORKPI.AN

l5

Monitoring of Trials

l5

CashCropCaseStudies

l6

Researchon HouseholdResourceFlows andAllocations

l8

Individual CaseSndies

l8

CommunityCaseStudies

t8

PRAs and the FarmingSystemsZones

19

FarmerResearchGroups

l9

ZCCFSPResearchMethodoloeies

20

StaffTraining

20

SocialDevelopmentIssues

2l

References

22

Arnex l: Termsof Referencefor the ZCCFSPSocialAnthropologist

24

Arnex 2: First Workplan for the ZCCFSPSocialAnthropologist

25

Annex 3: Formal Meetingsand Contacts

26

Annex 4: Anthropologistsand Other Researchers

29

PREFACE It was agreedwith the ZCCFSPField Managerthat in view of their overlapping content the bnef Situation Summary Report being prepared by the Social Anthropologistfor the end of Novembershould be combinedwith lns Inception Report, due in March 1995 (seethe earlierworkplan in Annex 2). The following reportis the outcomeof that decision. Here I would like to take the opportunityto thank all of my colleaguesfor their warrn and friendlywelcometo Zamibu and ZCCFSPand for providingan atmospherein which work is enjoyableand neverlackingin fruitful debate. Let us hope that the peopleof Zaruibarwill benefitthereby. Martin Walsh ZCCFSPSocialAnthropologist Wete,Pemba March1995

ACROI\-YMS APOS

AssociateProfessional OfficerScheme

BDDEA

BritishDevelopment DMsion in EasternAfrica

GTZ

GermanAgencyfor TechnicalCo-operationLtd.

FINNIDA

FinnishlnternationalDevelopmentAgency

FRGs

FarmerResearchGroups

FS

FarmingSystems

IFAD

InternationalFundfor AgriculturalDevelopment

MALNR

Ministry of Agriculture,LivestockandNatural Resources

NCDP

Programme NationalCoconutDevelopment

NRI

NaturalResourcesInstitute

ODA

Development Administration Overseas

PRA

ParticipatoryRuralAppraisal

TCO

TechnicalCo-operationOfficer

ZCCFSP

Zamibu CashCropsFarmingSystems Project

ZEDO

ZanzibarEnterpriseDevelopment Organisation

INTRODUCTION l. The Zarnbar CashCropsFarmingSystemsProject(ZCCFSP)beganin November 1991with the primaryobjectiveof identifyinganddevelopingalternativeexport crops to cloves. Followingthe recommendations of a Mid-termReviewheld in NovemberDecember1993this objectivewasbroadened anda revisedProjectFrameworkfor the periodNovember1993to November1995was drawnup. This describedthe project asfollows. 'A multi-disciplinary team,within the Researchsub-commission of the MALNR [Ministry of Agriculture, Livestock and Natural Resources] will conductresearchinto: cashcropswithin the farmingsystemsof the two islands;and marketingsystems. The aim of the project is to develop:the contributionof cashcropsto householdincomeandto the wider economy;the capacityof the public andprivate sectorto conduct research;andthe capacltyofthe privatesectorto trade in agricultural produce." 2, The Mid-term Review also recommendedthe appointment of a social anthropologist(TCO) for the remainderof the project. Recruitmentby NRI took placein nrd-1994andled to the appointmentof the authorof this report. I arrivedin Zaruibw,Ungujaisland,on29 Augustandproceeded to post at Wete on Pembaisland on 2 September 1994. Sincethen,andfollowingmy termsof reference(reproducedin Annexl), I havebeenengagedin work on both islands. 3. This situationsummaryreport describesthe major initial tasksundertakento the end of November 1994 and outlines a workplan for future work togther with the socio-economics sectionof ZCCFSP. As suchit completesthe workplanfor October andNovemberdrawnup duringthe visit of the ProjectManagerin September1994 andaddsto andmodifiesthe projectionscontainedtherein(seeBarret 1994,paragraph 18,andAnnex2).

INITIAL TASKS Introductions and Institutional Review 4. A seriesof visitswasmadeto differentprojectsanddepartments within MALNR as well as somein other ministriesand non-govemment agencies.Most of thesevisits took the form of official appointments, thoughthis was not exclusivelythe case. On Pemba I was accompaniedby SuleimanSheheand occasionallyAweina Omar, in the socio-economics colleagues sectionof ZCCFSP.On Ungujavisitsweremadein the companyof the new ZCCFSPmarketingeconomist,Peter Oldham,sometimes togetherwith his counterpartin the Mnistry of Trade,Industriesand Marketing, Jihad Abdullahi.

5. Other generalmeetingsprovidedthe opportunityto make further contactswith ministry staff and have both formal and informal discussionswith them. These meetings included the Unguja Farming SystemsZones Workshop in Kizimbani (2019194), the ResearchandExtension2nd Co-ordinatingMeetingin Wete (2519194), the PembaFarming SystemsZonesWorkshopin Wawi (lllll94), the visit of the British High Commisionerto ZCCFSPPemba(I5llll94) andthe first meetingof the FarmingSystems ZonesWorkingGroupin Wete(24llll94). 6. Annex3 list the namesof the peoplemet,togetherwith their positions,institutions andthe datesdiscussions wereheld. 7. The purposeof these appointmentsand discussionswas manifold. First, to introducemyselfand my colleagues, the currentwork of ZCCFSPand our own work within it. Second,to gatherbasicinformationaboutother projectsand departments, especiallywithin MALN& and learn about their current objectivesand programmes, personneland institutionalstructuresand other relevantcharacteristics.Third, to review points of contact,both existingandpotential,betweentheir own work and that of ZCCFSPincludingits socio-economics section. Fourth, to benefitfrom the wide experienceof the indMduals met with and the general insights into social and inZamibar whichtheycouldprovide. economicdevelopment 8. The discussions held were thereforetypically wide-ranging. They providedan excellentinsightinto the institutionalstructureandcapacityof MALNR and the varied activities,past,presentandprojected,of its manyconstituentparts. Theyprovidedan opportunityto discussZCCFSP'sown work, to understandexternalperceptionsof it and to explain its current objectivesand programme. They provided many detailed observationsaboutagriculturalproductionandmarketingandthe social,economicand other factorswhich impingeupon these.However,they also revealeda generalfailure researchandmonitoring,including to incorporatesocio-economic andanthropological a concernwith genderissues,into projectactivities. 9. Someprojectsanddepartments wereableto providedataof relevanceto our work. The most usefulof thesewere recentsurveyand censusresultsnot in the ZCCFSP libraries. It is clearthat most of the project-orientedsocio-economic researchwhich hasbeendonein recentyearshasbeenbasedupon formal questionnaires and surveys. research,especially By contrasttherehasbeenvery little qualitativeor anthropological projects. its The for conducting this type of within MALNR and current capaclty researchis thereforealmostnon-existent. 10. To detailall of the subjectsdiscussed or evento summarise the mainpointswhich emergedwould constitutea reportin itself. Whererelevantthesewill be incorporated into futurereports. I l. Individualmeetingsanddiscussions of this kind outsidethe usualframeof ministry and project work are evidentlyof somevalue, as an ongoing activity and not just as part of a newcomer'sfamiliarisation process.A numberof individualsandinstitutions, especiallyon Unguja and outside MALN& could not be contacted or were not

availableduring this round of meetings,and I therefore anticipate making further in the future. appointments

Contact with Anthropologistsand Other Researchers 12. Giventhe generallack of qualitative(andhigh qualrty)researchreferredto above (paragraph9), a determinedeffort hasbeenmadeto locateother anthropologists and who areworking inZar.z:ibwor havedoneso in the recentpastacademicresearchers with a view to finding out what up-to-dateinformationis available,publishedor togetherwith a brief descriptionof the otherwise. A list of overseasresearchers, principalsubjectsof their research, is givenin Annex4. 13. Most of thoselisted have alreadybeencontacted. a few remainto be traced. Someof their researchis directlyrelevantto my own and vice versa,and it is hoped that a fruitful exchangeof ideasandinformationwill result. It is particularyheartening are interestedin genderissues,and this should to note that manyof theseresearchers providea usefulsourceof inspirationandcorrectiveto the currentlack of emphasis on women in developmentin project-orientedresearchwithin MALNR and other ministries.

Library Researchand Literature Review andrelated 14. Beforeembarkingfor Zannbu I begana reviewof the anthropological literatureon the islandsandtheir peopleusingthe Universityof SussexandInstituteof DevelopmentStudieslibraries,together with my own collection. In the liule spare time availableI have continuedwith this work in Zanabar, consultingunpublished literatureaswell. Smith(1992) archivesandreadingthroughthe 'grey' (development) published material,thoughhis focusis is a usefulintroductionto both andunpublished restrictedto agriculturalresearchand developmentnarrowly understood. In addition to ZCCFSP'sown librarieson Pembaand Unguja,a numberof librariesin Zannbar town have been used: the Natural History Museum library, the Zamibiu National Archives,the Zamibartown public library and the privatecollectionof Mzee M. A. Ghassany(ExecutiveChairmanof the CashCropsandFruit Authority and an authority on the historyof MALNR andagricultureinZamibu). 15. This work is ongoing. It is alreadyapparent,however,that a lot of the relevant literature is not availablelocally. In future it may thereforebe desirableto visit the librariesin the University of Dar es Salaam(thoughaccessis difficult), the University of Nairobi (the mainlibrary andthe Institute of African Studies),the British Institute in EasternAfrica (alsoNairobi)andthe libraryof Fort JesusMuseumin Mombasa.This will be done if and when opportunity arisesand preferablyin conjuction with other activities. Otherwisereadilyavailableitems(that canbe photocopiedin the UK) will be orderedthroughNRI. 16. Although the historiansand others have kept fairly busy, there has been a surprisinglack lack of professionalanthropologicalresearchon both Pembaand

Unguja,especiallyresearchwhich is action-oriented.The classicpiece of research remainsMiddleton's(1961)studyof landtenure,basedupon threemonths'work for the colonialgovernmentin 1958. This underlinesthe potentialwider importanceof researchundertakenwithin ZCCFSP,and the needto ensure appliedanthropological its dissemination throughoutMALNR andother institutionswhich might benefit.

Reviewof Socio-economic Researchand Monitoring in ZCCFSP Socio-economic Researchand the Socio-economics Section 17. A lot of valuablesocio-economic researchhas alreadybeenundertakenwithin ZCCFSP. Much of this hastakenthe form of one-offstudieswritten up into reports. Theseinclude,for example,studiesof rural incomeearningopportunitiesinZat:r;ibar and of marketingandtransporton Pemba(Fox and Packham1994; Fox et al. 1993). In somecasesthesestudiesraiseas many questionsas they solve,a good example beingthe whereabouts of the "missingmiddlemen"of Pembaisland(Fox el al. 1993. paragraph28). Nonetheless they providea wealth of basicbackgroundinformation and,asin the casejust mentioned,form an excellentstartingpoint for furtherresearch. 18. Researchof this kind has typically been led by expatriateproject stafl and sometimesby outside consultants. This is perhapsunavoidableunder current giventhe fact that almostall of the MALNR project counterpartscome circumstances, from a backgroundof training in agriculture,agronomyand relatedsubjects,there being a markedabsenceof economistsand otherstrainedin the social sciencesamong the local staff It will be difficult to provide all of the training that is needed,whether on-the-jobor offit, duringthe remaininglife of the project. The alternativeis to draft in Zamibais who alreadyhavethe appropriatequalificationsin terms of their training and/orwork experience.This would probablymeanrecruitingoutsideof MALNR: a backgroundin agronomyis not necessary at all, ratheran understanding of the basic methodsandprinciplesof socialscienceor economics, dependingon the positionto be filled. 19. At presentZCCFSPcomprisesa multi-disciplinary teamof expatriatesanda more or lessmono-disciplinary to differentdisciplinaryfunctions teamof local staffassigned (in the agronomy,post-harvestand socio-economics sections). Unlessa concerted effort is madeto changethis imbalance,it will be difficult for MALNR to sustainthe multi-disciplinaryemphasisof ZCCFSP's farming systems approach when the leave. expatriates 20. As currentlyconstitutedthe socio-economics sectioncomprisestwo counterparts on Pembaand four (threeactive) on Unguja. Early indicationsare that they may have difficulty in meetingthe increasedworkload resulting from the arrival of the TCO social anthropologrstand marketingeconomist,both of whom are detailedto work with the socio-economics section. It maybe relevantto mentionherethat they share with other MALNR staff the problemof low salariesand lack of incentives. Some membersof the socio-economics sectionhave shown themselveswining to work longerhoursthan usual for examplein caseswhereit is easierto interviewfarmersin 6

the late afternoonthan during office hours - but feel that they are offered insufficient incentivesto do so. In this caseit may be worth reviewingcurrent practiceand eliciting the views and suggestionsof staff more systematicallythan hashitherto been the case.

ThePM Programme a progranrme of ten village 21. Betweenmid-1992andmid-1993ZCCFSPorganised Participatory Rural Appraisals (PRAs, otherwise referred to as PlRRAs, Participatory/Rapid Rural Appraisals),five on eachisland. As describedin the final was a resoundingsuccessand achieveda report (Thomas1993),the PRA progranrme A numberof importantobjectives. wealth of informationwas gatheredin the ten informationanddescriptionsof communities studied,includinga lot of socio-economic the constraintsupon cash croppingin each of these areas. This informationhas provideda considerable of local farmingsystems, boost to ZCCFSP'sunderstanding and hasalsomadea crucialcontributionto the definitionand descriptionof farming systemszoneson the two islands. 22. The PRAsalsohad an importantinstitutionalimpact. They involvedthe training of staff from different sectionsin MALNR in the methodologyof PRAs and their subsequentcollaborationin multi-disciplinaryPRA teams working in the field, a considerabledeviationfrom traditional MALNR practice. As a result the PRA programmehad a direct impact upon the 36 individualparticipantsas well as an indirect impactupon MALNR as a whole in that it helpedto disseminatethe farming systemsapproachfosteredby ZCCFSP. 23. Theway in whichPRAswereusedby ZCCFSPwas somewhatdifferentfrom that prescribedin muchof the literature,andthis wasthe subjectof somediscussionat the time betweenproject staff and various external advisers. The primary objective of PRAs is often taken to be to provide a meansby which local communitiescan participatein preparingand implementingplans for action based upon a rapid of their problemsand prioritisationof their needs- the reasoningbehind assessment this being that locally plannedand managedinterventionsare more likely to succeed than solutionsand projectsimposedfrom outsidethe community. ZCCFSP'sPRA prograrnmedid not leadto the formationof actionplansor indeedactionsof anykind in mostof the villagesinvolved. Oneimportantexceptionhasbeenin Dayaon Pemba wherea farmer'sresearchgrouphasbeenformedwhichhasbeenableto act, with the helpof ZCCFSP,on someof the problemsidentifiedduringthe villagePRA conducted therein November1992. Otherwisethe communitieswhich were the subjectsof PRAs havenot benefiteddirectlyfrom the exercise,thoughthey may do so indirectlybecause of the long-termimpactsof the PRA programmeas a whole on MALNR policy and practice. 24. Arguablythe correctrole for PRAs in MALNR is as a tool to be usedby the extensionservice,working in collaborationwith the researchsections,to identify and prioritiselocal problemsand take action upon them in close collaborationwith the community. However, glven current institutional constraints, including limited 7

resources,it is unlikelythat anythinglike an action-orientedPRA progranrmecould be developedwithin MALNR andits ExtensionSub-commission at present. 25. Critics may argue that there was therefore a mismatchin ZCCFSP'sPRA programmebetweenthe methodused andthe intendedobjectives. To the extentthat this mismatchwas not evident - the method did, after all, attain various project objectives- then it did not matter. However,we haveno recordof what most of the participatingvillagersthink aboutthe outcome,or what may seemto them a lack of outcome. Accordingto recentreports"surveyfatigue"hasbecomeprevalentin some areasof Unguja,and it is not impossiblethat the PRA programmehascontributedto this. The mismatchalso meansthat ZCCFSPhas trained MALNR stafl including manyof its own, to think of PRAs primarilyas researchtools, when in fact they are not. During workshopsand meetingson the descriptionof farming systemszones researchers both within andoutsideof ZCCFSPhavebeenquickto suggestPRAsas a gathering moreinformationon thesezones,especiallythosezoneswhich have way of not alreadyhadPRAsconductedwithin them(howeverseeparagraph77 below). 26. VillagePRAsmaynot thereforebe the mostappropriateway eitherto do research or to train staff in future. This is not to saythat villagePRAs haveno placewithin MALNR or that different PRA methodscan have no role in research. As Robert pointedout in a discussion of this issuewith the FieldManagerPRAsdo not Chambers havenorms (though the literature hastendedto createthese)but can and shouldbe However,if PRA tools areto be usedin research adaptedto differentcircumstances. then they shouldbe carefullyselectedand applied(for further discussionof one of 36-47below). One of the advantages thesemethods,wealthranking,seeparagraphs groups is many of the pitfalls referredto above farmer research that of working with can be avoided: it is possibleto conductresearchusing a variety of methodologies while alsohelpingto find solutionsto groupmembers'problems.At the sametime it is possibleto fosteranddevelopstaffmembers'useof participatorymethodsaswell as to stimulatecollaborationbetweendifferentsectionsof MALNR.

Farming Systemsand Farming SystemsZones 27. ZCCFSP,as its title implies,uses and actively promotesa farming systems of this hasbeenthe dMsion of both islandsinto farming approach.Onedevelopment (currently five zones on Pembaand four on Unguja) and various efforts systems of theseas research,planningand extensiontools designedto securethe acceptance quarterly report (ZCCFSP1994)recognises that this will within MALNR. The latest be a long-termprocess,and I agree. However,I alsohave somereservationsabout our definition and use of farming systemszones,and this is probablya good placeto statethem. 28. To an experienced anthropologistfarmingsystemsare a statementand often a simplificationofthe obvious. Thevalueof farmingsystemsresearchis in introducinga more holistic perspectiveto agronomistsand in encouraginga definite move away from the traditionalprotocolsof agriculturalresearchtowardsa moreparticipatoryonpractice. This is alsothe major significanceof ZCCFSP's farm and farmer-sensitive 8

approachwithin MALNR, which has been dominatedhistorically by on-station and researcher-led research. 29. The definition of farming systemsper se, however,is a tricky business. I understandfarming systemsas a convenientshorthandfor refening to farmers and where they farm and what they farm and how they farm and what they do with their produce. This is perfectlyintelligible: the difficulty comesin tryrng to put this into practiceby distinguishingone farmingsystemfrom another. The problem,in other words,liesin convertingthe generaldefinitioninto guidelinesfor classification. give priority to 30. A quick glanceat the literatureindicatesthat differentresearchers differentcriteria. The NCDP-sponsored surveyconductedon both islandsin 1988 resultedin the choice of land tenure statusas the main criteria for defining farming systems.Variousareas,however,were excludedfrom the survey- Zamibartown, the coral rag, and irrigated rice- and sugarcane-growing areas- and it is not clear what criteria (other than lack of potential for coconut production)were employedin excludingthem or how they might be treatedin the classificationof farming systems (Wirth et al. 1988). A subsequent (1989)NCDP surveyof the coral rag areasled to the rejectionof landtenurestatusas the principalcriterionand an alternativeanalysis basedupon a ratherconfusingmixtureof variables:'typical coral rag farms" (with a very high proportion of coral rag land) versusothers,male- versusfemale-headed households,andfarmson Pembaversusfarmson Unguja(Krain et al. 1992). 31. Here it might be pointedout that ZCCFSPhasneverproducedits own coherent classificationof farming systems.An early report (Thomas 1992) emphasisedthe importanceof "householdfarming systems"over the "land use systems"used by NCDP and others. Following the PRA prognunme,however, a classificationof householdfarmingsystemswas suggested whichwasbasedon the combinationof two different setsof features: the identificationof four different householdwealth groups throughoutrural Zarmbarplus the division of the islandsinto seven'broad agroecologicalsystems",threeon Pembaandfour on Unguja. As statedin the final report of the PRA programmethis would gwe a total of 28 householdfarming systems, recognisingthat further researchwould be required to confirm and develop this analysis(Thomas1993,puagraphs29-31).The householdwealthgroupswere based upon the resultsof wealthrankingexercisesconductedaspart of the ten village PRAs. The methodologyand practiceof wealthrankingsuffersfrom a numberof problems, and I will discussthesebelow (see paragraphs36-47). In any event the use of householdwealth groupsin this way to developa classificationof farming systems proved impracticable,and the idea was quietly dropped. The "agro-ecological includingthe systems",however,remained,and,aftervariouschangesandrefinements, from ZCCFSP integration of economic information derived studies, they have metamorphosed into our nine farming systemszones. Although this is not made explicit,the imptcation is that thesezonesrepresentdifferentfarmingsystems. 32. This brings me to my first major difficulty with the conceptof farming system zones. It seemsto me that a "farming systemszone" is a contradictionin terms and zonesor land use systemshiding that the ZCCFSPzonesare really agro-economic behinda falsename. This shouldbe evidentfrom the historyjust outlinedand the 9

gradualchangein emphasis from householdfarmingsystemsto farmingsystemszones. Moreover the reasoningbehindmy argumentis anticipatedand set out quite clearly in the earlierreport: "In Zar:r;tbara distinctionmust be drawn between"land use systems" and"householdfarmingsystems".Thisis very important,asalthoughit is relativelyeasyto identify differentland use systemsrelatedto their geographical location, physical environment and land tenure arrangements,it is quite another thing to differentiate between households,as the majority of householdswill use severaldifferent "land use systems". ZCCFSPneedsto work with householdsrather makethe decisionsaboutfarmingactivities" than areas,as households (Thomas1992:l-2). 33. Agro-economicfactors have been given priority in drawing up the ZCCFSP variablesremain farmingsystemszones,not only becausemanyof the socio-economic zones fit into even if they them unknown, but becauseit would be very difficult to were. This doesnot meanthat we shouldabandonthe zonesor even,at this late stage, changetheir name;but we shouldrecognisethem for what they are and be aware of form the best their limitations. Oneof theselimitationsis that they do not necessarily framework for undertakingsocio-economicand anthropologicalresearchnor for makinginterventionswhich have socialdevelopmentissuesas one of their primary concerns. 34. I haveanother,related,reservationaboutthe promotionof zonalplanningwithin MALNR which also relatesback to my earlier observationsabout village PRAs (see paragraphs2l-26 above). Pemba and Unguja comprise a patchwork of local communitieswhich sharemany featureswith their nearestneighboursbut also differ from one anotherin numerousways,a fact underlinedby the village PRAs. Thereare also differentways of simpliSing this situationfor the purposesof research,planning and actualintervention.During the colonialperiod,for example,it was commonfor racial and ethnic criteria to be employed. The most widely acceptedand used classificationhasotherwisebeenthe distinctionbetweenurbanand rural Zamrbar and zones,the plantationareas the divisionof the latter into two principalagro-ecological andthe coral rag. ZCCFSP'sfarmingsystemzonesrepresenta further refinementof factorsadded. Thereis danger,however, with otheragro-economic this classification, that the useof thesezonesfor planningwill foster insensitivityto local differencesand needs. It is ironic that villagePRAs,a planningtool designedto maximisesensitivity have and are being usedby the project to help definemuch to local requirements, largerplanningunits. before seriousconsideration 35. I would suggestthat this is a questionwhich deserves recruitment zonal The of ZCCFSP attemptsto coax MALNR into a zonal future. planners,for example,eitherto supplantor supplementdistrict agriculturalofficers, to local needsand opporhrnitiesunless would do little to increaseresponsiveness accompanied by radicallynew working methodswhich did not stop at the level of the (in paragraphT4above)that villagePRAs(or slimmed zone. I havealreadysuggested down, more topicalversionsthereof)usedby the extensionservicemight be one of l0

thesemethods. It might thereforemakemore sensefor ZCCFSPto concentrateon encouragingExtensionand othersto implementa participatoryapproachwithin the existinginstitutionalframeworkthan to spenda lot of time in attemptingto createa newframeworkbaseduponfarmingsystems(readagro-economic) zones.

WealthRanking 36. Wealth ranking exerciseswere an important componentof the village PRAS undertakenby ZCCFSPin 1992-93.Theyhavesincebeenusedasan independent tool in helpingto monitorthe impactof the pilot trials on Pemba.The primaryobjectiveof wealthrankingin this contexthasbeento providea rapid assessment of the statusof farmersparticipatingin the trials vis-d-visone anotherand other communitymembers so that the role of relative wealth and poverty in determiningtheir participationand performance in the trials canbe appraised.Threesuchexercises havebeencompleted sinceJune1994 in Kengeja(the siteof a cinnamontrial), Msuka(alsocinnamon),and Mkanyageni(blackpepper) 37. Theseexercises haveprovidedinterestinginsightsinto local perceptionsof wealth. Their appropriateness as a monitoringtool, however,is doubtful, especiallyglven variousproblemsin their useandapplication.Thefollowingcommentsarebasedupon my own observationsof the Mkanyageniexerciseand an analysisof the data and written reportsfor all threePembarankings. I havechosento discusstheseat some lengthbecause they raiseimportantpointsaboutthe relationbetweenthe objectivesof monitoring,the methodswhichshouldbeused,andour capacrtyto usethem. 38. I will beginby focusrngupon the actualperformanceof the exercises.A first problemhingesupon the choiceof households which are to be ranked. Ideally this would compriseall of the households in the communityin which the trial participants live, andwould includeall of the participants'own households.On Pembaandin the context of the pilot trials, however, this is not very easyto achieve. Many rural communitieson the island do not form discreteor boundedentitiesbut comprise scatteredhamletsinhabitedby groups of close kin, with other, less cohesive, settlements typicallystrungout alongthe mainroads. The homesof trial participants are therefore likely to be scatteredover a wide area and a large number of local hamletsandothersettlements.Thetotal numberof households in the areain which all of the participantslive thus frequentlyexceedsthe numberwhich can be treated in a singlerankingexercise(not much more than 100 householdsin total). As a result ZCCFSPresearchers havehadto excludesometrial participants'households from the rankingexercises andonly a subsample of the subjectsof our monitoringareranked(in less two cases thana third of the total). This meansthat we canonly thenlook at the role of wealthrank within this subsample of trial participants.The othershaveto be excludedfrom anyanalysisbasedupontheseexercises. 39. This problem does not usually arise in wealth ranking exercisesbecausea scatteredtargetpopulation(in this caseour trial participants)is not normallystipulated in advance. This problemdid not thereforearisein the villagePRAs (althoughthe wealth ranking exercisesconducted in these suffered from their own sampling 11

problems).In the contextof the trialsthereis no way to solvethis problem: all of the participants'households mightbe coveredby a seriesof wealthrankingexercises, but this would only createanotherproblem,deriving from the fact that separatewealth rankingexercisescanbe comparedbut not combinedinto a singleset of figuresand thereforeranks. This is not to mentionthe additionaltime which more than one rankingexercisewould take. 40. In the Msukaexercisethis is what in fact happened.The residentialscatterof trial participantsled to the choiceof two locations,Sebleniand Kichakapumu,in which separatewealthrankingexercises were conducted(andevenso only 13 out of a total participants 46 of originaltrid were included).The two rankingswere then combined into one, an invalidprocedure.Moreoverthe numberof households in eachexercise (31 place) was too small in each while in Kichakapumutwo of the informantsonly rankedthe households into two groups,anotherinvalidprocedure.Theseprocedural elrors unfortunatelyinvalidatethe whole exercise,thoughthey shouldbe thought of as a consequence of the geographyof settlementin the Msuka areaas of anythingelse. The scatterof participantsled to the choiceof two settlements, while the smallnumber of householdsin eachof theseled to the smallsamplesize and the inability of two informantsto distinguishmorethantwo wealthranksin oneof them. 41. The choiceof informantscanalsobe a sourceof bias. All threeinformantsin the Mkanyageniexerciseweremenin the sameagegroup (late 20s,early30s),and might be expectedto have more similar perceptionsthan three or more informantsof differentagesand gender. A further biaswas introducedby the fact that the first two informantswere interviewedin the presenceof one another. Not surprisingly, therefore,they selectedthe samenumberof wealthgroups,employedthe samecriteria in differentiating them,andagreedin their allocationsof almostall the 108households they were askedto rank. The third informant was interviewedlater, but also in the presenceof the first two. Ideallyotherinformantsaswell asexternalobserversshould not be in attendance, might biasan informant'sresponses. because their very presence Other factors may also bias the results, including informants' perceptionsof researchersthe responsesthey are assumedto want or, indeed,may suggestby promptingor askingleadingquestions.The existenceof so manypotentialsourcesof bias naturally reducesconfidencein the validity of the rankingsproduced by these exercises. 42. Analysisof the dataalsohasits problems,somemore seriousthan others. One problemis the time takento completethe analysisandwrite the 2-3 pagereports. In the Mkanyagenicasethis took morethantwo weeks,althoughit mustbe saidthat this is consistentwith the generalpaceof work on Pemba,and so is not a problemfor the wealthrankingexercises alone. 43. Far more seriousthan this, however,is what I considerto be a flaw in the methodologyof wealth ranking. This concernsthe last and crucial stepin analysis,the final groupingof householdsinto wealth strataas an average(which is basedupon conversioninto numericalvalues)of the differentranksassignedto them by individual informants. The locus classicuson wealthrankingoffers the following adviceon how to proceedafterassigning averagescoresto the households beingranked: t2

"The next stepis to list the householdnumbersaccordingto their score from the smallestnumberto 100, which is from richest to poorest. needto be groupedinto a numberof wealthstrata. Next the households (To use the absolutescoresobtainedwould not be valid but would representa falsedegreeof accuracy.)As a rule of thumb,the number of final groupsshouldnot be more than the averagenumber[of] piles usedby the informants.It canbe fewer,dependingon the researcher's the interests,but not lessthanthree. Normally,for easeof comparison, size of the final groups shouldbe roughly equal. When the size of groupscannotbe equalbecauseof the clusteringof scores,it is better for the middle group or groups to be larger and the extremes (richest/poorest) to be slightlysmaller."(Grandin1988:30) 44. This is very impreciseadvice,and in effectleavestoo muchto the whim of the analyst(s). In the Mkanyagenicasethis manifesteditself as follows, When the they eachplacedonly two or three informantswere askedto rankthe 108households, in the wealthiestof four categories.Their descriptionsafterwardsleft no households doubt that these householdswere led by very wealthy individuals with significant resourcesat their disposal,andthat they stoodout businessinterestsandconsiderable in this way from most of their fellows. Subsequentanalysisof the Mkanyagenidata, however,produceda final wealthrankingin which the wealthieststratumcontained3l betweeninformants'perceptions households.Therewas thus a markeddiscrepancy canbe andthe analysiswhich resulted. It might be saidthat this kind of discrepancy avoidedby an experiencedresearcher.However,the fact that it can occur at all must the data be ascribedto the imprecisionof the method. As an experimentI reanalysed from this and other wealthrankingsandfound it quite difficult in a numberof casesto identify unequivocalcut-off points at which to separatedifferentstrata. I conclude that it would be relativelyeasyfor differentanalyststo producevery differentanalyses, restingupon quite arbitrarydecisionsaboutcut-off points or the other factorswhich oughtto be considered. Grandinsuggests 45. As a resultwealthrankingis liableto be very mucha hit-and-missaffair. It can also leadto a very falseimpressionof the significanceof wealth differencesin rural communities,suggestinga patternof permanentor incipientdifferentiationwhere in the methodof wealthrankingrestsupon fact theremaybe noneat all. This is because consideringdMsions within the households without classification of the synchronic household(for examplethe different accessof husbandsand wives to household resources), the patternof relationsbetweenthem(for examplethe poolingof resources andof thesedivisions within largerdomesticgroups),andthe evolutionof households and relationsover time. Thereforemany of the apparentdifferencesbetweenstrata might be a simplereflectionof the (assumingtheir identificationto be unproblematic) fact that the householdswithin themare at differentstagesof the developmentalcycle. 46. This certainlyseemsto havebeenthe casein Mkanyageni.It was noticeablethat amongthe principalcriteriausedby informantsto classifyhouseholdswere the age, sex and marital statusof the householdhead. The secondstratumwas dominatedby householdswith relativelyyoung and active male householdheads,the third by l3

households with olderandlessactivemalehouseholdheads,andthe fourth by widows people andold with a limitedcapacityfor labour. I alsoaskedaboutrelationsbetween householdsin different strata and informants observedthat one local family might include householdsin all three of these categories,while householdsin the upper stratumfrequentlyhelpedtheir kin in lower strata during times of need. Informants also confirmedthat householdswere apt to changetheir position in this nominal rankingover time. 47. To sumup, wealthrankingexercises sufferfrom a host of problemswhich render themof dubiousvaluein the monitoringof trials andtheir impacts. It would be more usefulto askparticipantsdirectlyabouttheir age,domesticstatus,accessto land and oFfarm enterprises.I thereforerecommend that wealthrankingbe discontinuedas a monitoringtool.

Monitoring Systems 48. To date ZCCFSPhas not developeda coherentand ef[ective approachto monitoringthe socio-economic impactsof its on-farmtrials and other field activities. Thereare a numberof reasonsfor this, one of them beingthat the field progranrme itself hasbeenin stateof constantflux aswell asbeing somewhatdifferent on the two islands(whereboth the methodologyof trials and the cropsinvolvedhavediffered). This situationstill prevailsthoughis likely to changeasthe emphasisof the progranrme shifts away from conductingon-farm trials in the traditional way to working instead with farmers'researchgroups. 49. Socio-economicmonitoring on both islandshas revolved around the use of differentmonitoringforms, eachdesignedwith a differentapplicationand field activity in mind. In somecasesthesecombinesocio-economic and agronomicinformation, while someforms havebeendesignedfor specificcrops and the activitiesassociated with these.The overallresultis very confusing. 50. On Ungujaan attemptwasmadeto improvethis situationby producingandfieldtestinga'Tarmer ProfileData CollectionForm". Theseformsarefour pageslong and are designedto elicit information on the farmer, householdcomposition,plot ownershipand use, householdassets,extensioncontactsand accessto financial resources.So far about37 formshavebeenfilled in for farmersinvolvedin the chilli campaign. Someeffort was madeto introducetheseforms to the socio-economics sectionon Pembabut they were not adopted. On Pembaa different set of forms has been sporadicallyin use. The most recentlyproducedof these is a '?ilot Trial Monitoring Form, Socio-economics", designedfor use with farmersinvolvedin the pilot cinnamontrials. This form is only two pageslong and containssectionson the farmer,householdlabour,land ownershipanduse,incomeand agriculturalconstraints, includingconstraints to plantingcinnamon. 51. Both of the forms describedcan and havebeenusedto collect interestingand usefulinformation. Theycould alsono doubtbe improved,perhapsby combiningthe best featuresof eachof them - althoughit is difficult to know what questionsto ask t4

andhow bestto askthembeforeundertakingqualitativeresearch.Staff could alsobe bettertrainedin completingthe forms. A more seriousproblem,however,lies in our capacityto makeuseof the datawhichtheycouldsupply. Completedformstendto lie idle in files,andno attemptis madeto analyseor do anythingfurtherwith them. One reasonfor this is that the formsdo little morethanprovidebaselineinformation. This is all very well - the possession of baselineinformationon programmeparticipantsis a necessary prerequisitefor manymonitoringsystems- but it doesnot in itself perform the taskof monitoringimpactsor elicitingthe reasonsfor these. 52. Although it is tempting to treat these forms as the starting point for the developmentof a comprehensive monitoring system,a further note of caution is necessary.Previousattemptsto establishdatabases of differentkinds on both Pemba andUngujahavefounderedon the failureto keepthemupdated. This may partly be due to lack of time and inclinationon behalfof the staff involved. This is an activity whichZCCFSP(andMALNR) in its currentinstitutionalstatecannotsustain,andit is hardto envisageit beingpossiblewithout the recruitmentor assignment of staff solely to the task. Anotherobstaclein the way of creatingandmaintaininglargeor complex databasesstemsfrom the difficulty of working with moderninformationtechnologyin Zanzrbar, wherecomputerequipmentsuffersfrom the humidityandcomputerusersare hampered by frequentshortages of electricity. 53. Needlessto saythe currentprocessof windingup on-farmtrials andthe shift of focus to farmers'researchgroupsmakesthe developmentof monitoringsystemsas originally envisagedredundant. Farmers'researchgroups require a quite different approach,in a context where intensivecontact allows more qualitative and detailed work. This andotherfutureplansarediscussed in the sectionswhichfollow.

WORKPLAII 54. This sectionof the report outlinesa workplan for the year beginningJanuary 1 9 95 . Monitoring of Trials 55. Followingthis reviewanddiscussion with the AgronomySectiona new andmore pilot directapproachto trial monitoringwasinitiatedin December1994. This is based upon open-ended interviewswith the trial participantsboth as a group and, in some cases,asindMduals. 56. The purposeof theseinterviews,in additionto gatheringbackgroundinformation on the communityand participants,is to explainthe performanceof the trials and differentaspectsof the farmers'response.Interviewersare guidedby a checklistof questions.Themeetingswith farmersareheldtogetherwith projectagronomists, who also ask their own questionsto the group as a whole. The resultinginformation, togetherwith any additionsfrom project stafl is written up into brief (2-3 page) narrativereports. 15

57 Interviewsbeganin December,in Makombeniand Shangafu(both sitesof black peppertrials) and continuedin the New Year in Mkanyageni(black pepper),Jomvu andMsuka(bothcinnamonsites). Giventhe largenumberof trial sites- originallyla the aim was to completethe monitoringinterviewsin a singlevisit, dealingwith one site per week. This schedulehasproveda little ambitious,andthe currentgoal is to finishthis work in April. It is thenplannedto compileall of the reportstogetherandto write a shortpapersummarising the salientpoints. 58. It has alreadybecomeclearthat manyof the problemsexperienced in the trials were a resultof eitherimperfectplanningandcontrolon the part of ZCCFSP(e.g.the provisionof defectiveseedsand seedlings)or unrealisticexpectationson the part of the farmersinvolved. A lot of farmersjoined the trials in the beliefthat the project evidently led by wealthy Europeans- would distribute free materials and provide variousbenefitsto themotherthanjust cheapseedsand/orseedlings,and whenthese presumedbenefitswere not forthcoming,their interestin the trials often waned, and management of the trial cropssufferedaccordingly.While this attitudeon the part of farmers no doubt reflects their past experienceof agricultural extensionand other projectinterventions, it mustalsoowe somethingto the naivetyof our own approach. 59. Consideration of thesepreliminaryresultshasalreadyplayeda part in the decision to wind down the pilot trials on Pemba. The ongoingattemptto persuadeindMdual trial participantsto establishtheir own seedlingnurserieswith initial ZCCFSP is alreadyrunninginto relateddifficulties(i.e. in manyvillagesthe farmers assistance declarethemselvesunwilling or unableto managenurseriesunlessadditionalinputs are providedby ZCCFSP),andwill haveto be rethought. 60. A similar approachto trial monitoringhas been recommended to the Socioyet put practice. economicsSectionon Unguja,but not into 61. As well as providingimportantinformationto the project about reasonsfor the successand failure of trials, this methodis designedto improve the interviewing and reportingskillsof sectionmembers,encouraging themto focusupon the analysisof a problem(e.g.why did a largenumberof farmersin sitex not follow up on their initial ordersfor seedlings?) andits description.Projectstaffalreadydo this in everydaylife andinformaldiscussion(e.9.aboutthe trial sitesandtheir visits to them)but tend not to think of recordingthis asimportantknowledgefor the project. It is hopedthat the open-endedinterview and narrativereport approachwill give them greaterconfidence to do so. In this respectthe methodis alreadybeginningto proveeffective.

CashCrop CaseStudies 62. Giventhat the pilot and othertrials on PembaandUngujawere mainlyproviding informationon cropswhich had not yet beenharvestedor marketed,it was felt that it would be of considerable valueto undertakedetailedstudiesof indMdualcropswhich had 'takenoff ascashcropsin particularlocalareas.

T6

63. A rough modelfor this type of studyis providedby a paperwritten by Gerlach (1963) on the developmentof the fresh fish trade in the hinterlandof the southern Kenyacoast. While this studydoesnot dealwith cashcropsas such,it doesgive a clearaccountof how tradein a particularcommoditywas developed,andhow various obstaclesto this developmentwere overcome(so clear that it has been used in a participatorybusinesstraining programmefor women's groups in Kenya). It was proposedthat the Zamibarcasestudiesbe begunwith a similarformat in mind. 64. ln additionto providinginformationon the cashcropsconcerned, thesecashcrop casestudiesare designedto producegeneralinsightsinto the development of cropsas widely-marketedcommodities,and especiallyto indicatehow and why variouslocal constraintswere overcomein the courseof this development, what problemsmight still remain,and what other impactsit may have had. It is hopedthat the lessons learnedwill then feedinto the formulationof ZCCFSP'sown strategyfor stimulating similardevelopments. 65. Two casestudieswere initiatedon Pemba,one on sweetpotatoesin Makangale andthe other on turmericin Muwambe. The first of theseconcernsa crop normally thoughtof primarilyasa food crop (andthereforeneverbeforethe subjectof ZCCFSP interest,also becauseit has no obvious export potential)in an area which is of additionalinterestfor a numberof reasons. theseincludea patternof land tenure problemsand disputes,invoMng immigrant and indigenouscommunities,and a reputationfor sharplydecliningsoilfertility, It is alsoin PembaFarmingSystemsZone l, which was not coveredby any of ZCCFSP'svillage PRAs. The secondcase concernsa crop which has alreadybeenconsideredin somedetail by ZCCFSP,on Ungujaaswell ason Pemba. 66. Togetherwith the MarketingEconomistand membersof the Socio-economics Sectionon Ungujatwo studieswere also selectedandbegunon that island. Both of these concern fruit crops: orangesin Ndijani and pineapplesin Machui and thereabouts. These Unguja case studiesare also being combinedwith cost of productionstudiesinitiatedby the MarketingEconomist. 67. The Pembacasestudiesbeganin Novemberandthe Ungujastudiesin December 1994. Theyarebeingundertakenby sectionmembersfollowingthe initial formulation of researchquestions(in the form of an openchecklist)in conjunctionwith the Social Anthropologist.Againthe methodis oneof open-ended interviewsandthe production of narrativereports. It was originallyhopedthat two of the four studieswould be completedin February. Fieldworkand working notesfor three of the studieswere finished by the end of the month but the full writing up of the results has been postponedto April, whenwork on the fourth studywill alsobe finalised. Fieldwork for an additionalcasestudy on Boribo Muyuni mangoesby the APO Agronomist on Ungujahasalsobeencompletedandis readyto be writtenup.

t7

Researchon HouseholdResourceFlowsand Allocations 68. Intensiveanthropologicalresearchinto householddecision-making,resource flows, and the constraints,opportunitesand impactsof cash crop productionand marketing at local level has begun using two different methods and sources of information. Theseare individualcasestudiesandcommunitycasestudies,discussed separately below.

Individual CaseStudies 69. The individualcase studiesfocus upon project field staff and their families. Followingpreliminaryvisits with one Pembacolleagueto his natal and other homes, this idea was suggested to project staff on Pembaand Unguja and met with general approval. It was emphasisedthat, as is the norm in anthropologicalresearch, informants'confidentialitywould be fully respected,and that the resultswould be generalisedand combinedwith those emergingfrom the communitycasestudiesand othersources. 70. In mid-December a participatorytrainingsessionwasheldfor thePembafield staff on how to draw kinship diagrams. Participantswere encouragedto follow up by recordinginformationon their own familiesusingthe methodspresented.This served as an entrdeto the casestudiesthemselveswhich beganin January1995 andhave so far involvedfive Pembastaffmembers.Similarwork will beginon Ungujaduringthe nextextendedvisit there(not yet scheduled). anthropologrst's 71. The adoptionof this methodin additionto the communitycasestudiespresentsa number of advantages. Project staff and their kin represent a wide spread of communitieson the two islands,including town dwellers, and becauseof their familiarity with the researchand the researcherit relatively easy to pursue some questionsin considerable depth. It is hopedthat directinvolvementof the field staffin this researchwill alsoincreasetheir understanding of it, the informalresearchmethods issuesin general. used,andof socialdevelopment 72. It is anticipatedthat the bulk of this researchwill be completeby June and incorporatedin theInterimReportto be producedthen(seeparagraphT5below).

CommunityCaseStudies 73. The communitycasestudieswill focus upon ZCCFSP'sfarmerresearchgroups andtheir members'local communities.This will involvefieldworkundertakenby the aswell asfollow-upwork by othersectionmembers.A complementary anthropologist progranrmeof agronomicresearchis beingplannedon Pemba,and hopefullycan also be undertakenon Unguja. 74. Fieldworkwill begin on Pembain April with Daya Farmer'sResearchGroup. Two monthsof the anthropologist'stime hasbeenprovisionallyassignedto eachof 18

four researchgroups, Daya and Kangaganion Pembaand Gamba and Ndijani (or perhapsa new researchgroup) on Unguja. This meansthat April and May are assignedto Daya, June and July provisionally to Gamba, and so on through to November. Field researchwill not take up all of this time, but as much as possible (and/ornecessary) givenotherprojectcommitments. 75. Thepreliminaryresultsof this researchaswell asof the individualcasestudieswill be presentedin an Interim Report scheduledfor June 1995. A final report will be preparedat the endof theyear. Interestingandpotentiallyimportantfindingswill also, asandwhentheyarise. with colleagues of course,be discussed

PRAs and the Farming SystemsZones (seeAnnex2 below) no furthervillagePRAs are 76. Contraryto earlierassumptions at presentplannedfor either island. Researchon the Farming SystemsZones, however,will continue,largelyaspart andparcelof the work alreadydescribed. 77. The Pemba FS Zones Working Group, hosted by ZCCFSP and with representatives from differentsectionsandprojectswithin MALN& hasrecommended that data on each of the zones should be consolidated,each section/project contributingasit canandusingthe variousmethodsat its disposal,without recourseto multi-disciplinaryPRASat present. The researchbeingundertakenby the agronomic and socio-economicsectionsof ZCCFSP on Pemba,which uses selectedPRA of is alreadyaddingto the descriptionofthe zonesand an understanding techniques, the problemsandopportunitieswithin eachof them,andwill continueto do so. 78. A similar situationprevailson Unguja, althoughthere has been no further of the FS Zoneswithin theMnistrv there. discussion

Farmer ResearchGroups 79. Thereis no doubtthat the FarmerResearchGroups(FRG$ compriseone of the most promisinglines of further developmentfor ZCCFSP. Their full potentialfor participatoryresearchhas yet to be realisedand the current workplan has been designedwith a view to makingmuchfuller useof them. discussionhas 80. Needlessto say,variousissuesremainundecided. Considerable and continuesto take place over the best ways to form and work with FRGS. Following an exploratoryworkshopwith GambaFRG on Unguja it was decidedto adopt a more systematicapproachto working with the FRGs on Pemba. It was proposedthat one of the first stepsin working with a group shouldbe to hold one or more one-dayparticipatoryworkshopsto elucidatetheir perceptionof the main constraintsto agriculturaland cashcrop productionand to discusspossiblesolutions with them. It was also suggestedthat this should be followed by a more regular patternof meetingswith the FRGsthanhadhithertobeenthe case. 19

81. In Decembera workshopwas held with DayaFRG at which their constraintsto cashcrop productionwere discussed andregularmonthlymeetingswere agreedupon. In the same month backgroundinformation was collected on the formation of KangaganiFRG and the other group activities of its members(all of them women) prior to holdinga workshopwith themin late January1995. Becauserelativelywork had been done before in Kangagani,this workshop focusedon the constraintsto agriculturalproductionin general.Again,regularmonthlymeetingswere agreedupon. 82. ZCCFSPPembais currentlyplanningits work programmesfor both Daya and Kangagani: this includesthe anthropologicalresearchalreadyreferred to above (paragraphT4).Agronomicresearchon individualfarmersandtheir plots which will be linkedto the anthropological researchhasalreadybegunin Daya. 83. Giventhe project's decisionto limit the numberof FRGs (perhapsaddingone moreon eachisland)the issueof selectionapproaches seemsto be lesscriticalthan it oncewas. It would seemsensible to reviewthis issueat a ratherlater stage,whenthe outcomeof work with existingFRGscanbe properlyassessed.

ZCCFSPResearchMethodologies 84. Giventhe processual natureof ZCCFSPandthe varietyof researchmethodologies which it hasemployedover the years,project staff agreedthat it would be usefulto preparea paper summarisingthese and their various pros and cons. The social anthropologisthasagreedto do this in readinessfor the March 1995review mission from BDDE,\ Nairobi.

StaffTraining 85. The anthropologisthasalsobeenand will continueto providean ongoinginput into stafftraining. This focusesin particularuponparticipatoryresearchmethodsand field staff on Pembahave alreadyprogressedconsiderablyin their use of various techniques,developingopen-endedchecklists,conductingsemi-structured interviews andwriting narrativereportson the basisof these.Thiswork is alsobeingextendedto Pembawith the helpof the projectMarketingEconomist. 86. Constructiveuse is also being madeof the anthropologist'sexperiencein and familiarity with participatory training methods, especially in the planning and organisation of trainingandotherworkshopson both PembaandUngujafor ZCCFSP and other MALNR staff and othersoutsideof the Ministry. Although ZCCFSPstaff havealreadyattendedsometrainingof trainersworkshopstheir use of participatory training techniqueshasnot beenreinforcedthrough constantpracticeand it is evident that they still have someway to go before they can becomefully effective trainers. The anthropologist's input to this work on both islandswill thereforecontinueas and whenadviceandassistance arerequired.

20

SocialDevelopmentIssues 87. It goeswithout saylngthat one of the anthropologist'stasksis to ensurethat socialdevelopment issuesare given full consideration within the project at all levels from project planning to the actual conduct of research,its outputs and the recommendations and developmentactivitiesbasedupon these. Socialdevelopment concernsare an integralaspectof all the work describedin precedingparagraphs. HereI will only highlightsomeof the majorissueswhichI will pursue. 88. Genderissueshavehithertoreceivedvery little attentionin Zarzibarand in the variousprojectswithin MALNR as well as in other ministries. Notable exceptions includethe work of the VillageForestryDepartmentwith women'sgroups(described by Khatibu and Suleiman 1992), a report produced by the Zanzibar Enterprise DevelopmentOrganisation(ZEDO) on women's ownershipof land in Zanzibw @onkerloandAboud 1994),andthe ongoingwomen'sgroup projectslocatedwithin the Ministry of State,President'sOffice, (and) Womenand Children'sAffairs. The potential role of women's groups in ZCCFSP activities is currently under consideration, and one of the FRGson Pemba,Kangangani,is composedentirelyof women (who are membersof differentwomen'sgroups in the village). Particular researchas attentionwill be paidto genderissuesin the programmeof socio-economic well asin otheraspectsof ZCCFSP'swork, includingits trainingactivities. 89. ZCCFSPalsohasa clearlystatedinterestin povertyalleviation,thoughthe most effectivewaysto achievethis objectivearestill the subjectof discussion.Researchwill thereforebe directed to providing a deeperunderstandingof this issue so that appropriaterecommendations can be made. The availableliterature containsonly superficialanalysesof socialstratificationand classin Zalzibar, and detailedvillage andurbanstudiesof this issuehaveyet to appear. Referencehasalreadybeenmade (see paragraphs36-47 above) to ZCCFSP's wealth ranking exercisesand the suggestionthat the rankingthat theseproducedmay reflect different stagesin the developmental cycleof rural householdsas much as permanentwealth differencesin relationsand their the community. A clearerpicture of intra- and inter-household development over time in both rural andurbancontextsis clearlyneeded,as this may have important consequencesfor decisions about targeting households and communities in ZCCFSPaswell asin otherprojectsandministryactivities. 90. Socialdevelopment concernsalsoextendto the institutionalstructureof MALNR of institutional andZCCFSP'spresentandfuturepositionwithin it. The development capacityis alreadythe subject of considerableeffort by ZCCFSP and intensive discussions within the projectabouthow this canbestbe achievedin future. Onearea of discussionis the possiblefuture role of NGOs and here it is anticipatedthat the project anthropologist,who hasa wide experienceof working in the NGO sectoras well asin other institutionalenvironments, will havea potentiallyusefulcontributionto make. 91. Last but not least,the anthropologist is contributingto ongoingdiscussions about the institutional structure of ZCCFSP itself, including the need to improve 2I

communication within the project,especially betweenZatuibanandnon-Zatuibaristaff andbetweenthe two islands.

References Barrett, J. C. 1994 Report of a Management Visit to the Zqnzibar Cash Crop Farming SystemsProject: 19-23 September1994, NR[, Chatham (File: FSC

e2e4ssztsl8l046c). Cameron,Greg 1992 'The StateandCo-operatives \n Zamlbar', paperpresented at the Internntional Conferenceon the History and Cuhure of Zanzibar, 14-16 December. Donkerlo, Jannekeand Asha Aboud 1994 lV'omenand Land in Zanzibar. Zannbar: Zamibar EnterpriseD evelopmentOrganisation. Fox, Diana,SalumShaaliAli, AbdullaJumaKhamisand AweinaOmar Issa 1993 Marketingand Transportin PembaIsland,WorkingPaperNo. 93112,ZCCFSP, MALN& Zallu;ibar. Fox, Diana and JamesPackham 1994 Rural Income Earning Opportunitiesin Zanzibarwith Regiorul Analysis(2 volumes),TechnicalReportNo. TR 93110, ZCCFSP,MALN& Zannbar. Gerlach,Luther P. 1963 'Traderson Bicycle: A Studyof Entrepreneurship and CultureChangeamongtheDigo andDurumaof Kenya',Sociologus,13 (l),3249. Grandin,Barbar4 E. 1988 WealthRankingin SmallholderCommunities:A Field Manual. Rugby: Intermediate TechnologyPublications Ltd. Khatibu, MwantangaL andRahikaH. Suleiman 1992 Women'sParticipation in Forestry Activities in Zanzibar (Zanzibar Forestry Development Project TechnicalPaperNo.6). Helsinki: Commissionfor NaturalResources,Zarzibw; FINNIDA; FinnishNationalBoardof Forestry. Krain, E. et al. 1992 Farming Systemsof the Coral RagArea of Zanzibar, National CoconutDevelopmentProgramme, Zanzibw. Middleton, John 1961 StationeryOffice.

ODA

Land Tenurein Zanzibar. London: Her Majesty's

1993 Report on A Visit to Zanzibar: Mid-term Reviewof Zanzibar Cash CropsFarming SystemsProject, 29 November- 3 December1993,BDDEA' Nairobi.

Rose, Kimberely,D. E. 1994 Sicknessand Malnutrition on Pemba: An AnthropologicalStudyof the SocialRelationsof Food, Perceptionsof Sichtess, and Help-seekingBehaviour,reportfor Savethe ChildrenFund,Wete. 1992 Agricultural Researchand Developmentin Zanzibar: An Smith, P. D. Analysisof the Literature,Centrefor Arid Zone Studies,UniversityCollegeof North Wales,Bangor. Thomas,Mark 1992 Farming Systems(FS) Within ZCCFSP,Working PaperNo, WP 9215,ZCCFSP,MALNR, Zaru;,tbar. Thomas,Mark 1993 Final Report: P/RRAProgramme,Ten Village P/kRAs and the Findings of A WorkshopHeld at the EACROTANAL Centre, 3ll5/93 416193, ReportNo. P/RRA93lll, ZCCFSP,MALN& Zannbar. 1988 A BaselineSumeyfor the ldentification of Farming Wirth, Figga et al. Systemsin Zanzibar. Berlin: Centre for AdvancedTraining in Agricultural Development, TechnicalUniversityof Berlin. ZCCFSP 1994 Zanzibar CashCropsFarming SystemsProject Quarterly Report, I July - 30 September1994,ReportNo. ZNZ 94/95 Q2, ZCCFSP,MALNR, Zarulbar.

23

Annex l.: Terms of Referencefor the ZCCFSP SocialAnthronolosist "The dutiesof the appointment will include: a.

To work as part of the ZCCFSPinter-disciplinary team,with farmers andtradersin pilot villagesto establishthe impactsand opportunities for the productionandmarketingof candidatecrops. To contributeto the interpretationof thesecropsmadeby ruralhouseholds.

b.

To contributeto the developmentof methodologyfor the design, implementation and analysisof trials to assesscashcroppingimpacts andopportunitieswithin the farmingsystem.

c.

To conductresearchon individualand householdinvestmentsboth within and outsideagriculture. As far as possibleand where relevant, to placetheseindividualand householdstrategieswithin the wider contextof whole family investmentpatterns.

d.

from differentareaswith access To coverindividualsandhouseholds to and control of differentresourcelevels. To includeone or more categoriesof personsinvestingin agriculturebut not residentwithin the rural areas.Thisresearchshouldcoverboth islands. To usea casestudyapproachfor c. and d., stayrngin villagesduring field work. To plan with the rest of the team follow up research required,to testthe findingsfrom the casestudymaterial. To assess the potentiallynegativesocialimpactof candidatecashcrop developmentespeciallywith respectto women, children,and food crop cultivation. To usethis informationto give guidancein further selectionbetweencandidatecashcrops.

g.

To supervisethe APOS socialscientistand ensurea complementary Ungujaprogranrme specificallyin the monitoringof on-farmtrials.

h.

planning,researchand To contributeto the overallinter-disciplinary project. progr.mlme activities of the extension To train counterpartstaffin surveywork andanalysis."

(from ODA letterof appointment dated17August1994) It shouldbe pointedout that giventhe "process"natureof the project,someof these termshavealready,in practice,beenmodified. Item (g), for example,doesnot apply in full becauseanAPOSsocialscientistwasnot recruited.

24

Annex 2: First Workolan for the ZCCFSP SocialAnthropoloeist

a

l. To visit differentsectionsandprojectswithin the Ministry of Agricultureand other relevantministries/ agencies,for the purposesof introduction(to personneland programmes)and to review points of contact (past, presentand future) between programmeandrelevantaspectsof their own work. ZCCFSPandits socio-economic 2. To review the existing progranrmeof socio-economicdata collection and monitoringin ZCCFSPandthe capabilities of the staffinvolved. 3. To plansocio-economic involvementandinputsin the ongoingprogrammeof pilot andothertrialson PembaandUnguja. 4. To prepare,at the endof November,a situationsummaryreport of 5-10 pageson the resultsof tasksl-3 above. InterimWorkplan-December1994- March 1995 While a more detailedplanwill be presentedin the Novemberreport, it is anticipated that the periodDecemberthroughto Marchwill includethe followingtasks:l. Active involvementin pilot andothertrials. 2. Involvementin villagePRA exercises(on FarmingSystemszonesnot previously coveredby villagePRAs). 3. Initiationof a programmeof basicresearchon Pemba,lookingin particularat interand intra-household resourceallocationsandflows, genderrelationsand other factors which imprngeupon opportunitiesfor and constraintsupon expandedor otherwise modifiedcashcrop productionandmarketing. This researchwill alsoinvolveMinistry counterparts andincludeanimportanttrainingcomponent. 4. To preparean inceptionreport for the March 1995reviewmissionand to discuss Adviser,Dr PhilipEvans." this in advancewith the BDDEA SocialDevelopment (workplandatedI October 1994,agreedby the Field Managerandforwardedto NRI)

25

Annex 3: Formal Meetingsand Contacts Pemba Shariff Maalim Hamadi (Head), Plant Protection Division, Research Subcommission, MALNR (719194, 2819194) Matthis Spittel, Strengtheningthe Plant Protection Division of Zarulbar, Plant ProtectionDivision,ResearchSub-commission, MALNR (719I 94) Merja Mflkelli, FINNIDA Forestry Project, Forestry Sub-commission, MALNR (visitingfrom Unguja, 819194) Rashid Said Nassoor(Administrator),Savethe ChildrenFund,Wete(1319194). Juma Omar Hamad (Farming SystemsResearchOfficer), National Coconut Development Programme, ResearchSub-commission, MALNR (28I 9I 94, 6I | 0I 94) Masood SuleimanHamad, VillageForestryDepartment,ForestrySub-commission, MALNR (28lele4) SuleimanHamisi Ali (RegionalForestryOfficer),ForestrySub-commission, MALNR (28lele4) Khatid Zahor Muhiddin (Acting Head), Extension Sub-commission, MALNR

(28tete4) Rex M. Kenneh,ExtensionSub-commission, MALNR (3110194) Khatib Juma Khatib, IFAD, Food Crops Adaptive Research,ResearchSubcommission, MALNR (4110194 andvarious) Dr KassimJuma (Head),LivestockSub-commission, MALNR (6110194) Rashid H. Said (Head),AgriculturalSub-commission, MALNR (6110194) Rashid Juma Qlead),ForestrySub-commission, MALNR (6110194) Said Abdulrahman Abdi (RegionalCoconutExtensionOfficer),National Coconut Development Programme, ResearchSub-commission, MALNR (6I l0 I 94) Ati Saidi Hamadi (FisheriesDevelopmentOfficer and Acting Head),FisheriesSubMALNR (6110194) commission, Omar Majid Salim (Extension Officer), Fisheries Sub-commission,MALNR

(6tr0te4) 26

Ali Salim Chenga (PlanningOfficer),Ministry of State,President'sOffice, Women andChildren'sAffairs(l9ll0l94) Shaibu Abdallah Mohammed (Personneland AdministrationOfficer), Ministry of State,President'sOffice,WomenandChildren'sAffairs(l9ll0l94) Omar SaleheAli (AssistantPlanningand ProjectsOfficer), ), Ministry of State, President'sOffice,WomenandChildren'sAffairs(19110194) Halima Ali (ResearchOfficer, Sociologist),), Ministry of State,President'sOffice, WomenandChildren'sAffairs(26110194) Mwalimu Anas (OfficerIn Charge),MALNR (l5llll94)

Unsuia Amina Mohammed (AssistantCommissioner), ResearchSub-commission, MALNR (3118194 andvarious) Mohammed Abdallah Ghassany (Executive Chairman),Cash Crops and Fruit Authority,MALNR (l 19/94, 9I | | 194) SalehSadiq Osman(Permanent Secretary), MALNR (l7ll0l94) Mwadini Haji Kutenga (Acting Commissioner),Research and Extension Commission, MALNR (17I 1019 4) Abu Bakr (Assistant Commissioner),Agriculture Sub-commission,MALNR

(8true4)

Dr Mahmoud E. Mansour (Adviser),IFAD SmallholderSupportProject,Research MALNR (8I | | 194) Sub-commission, Rashid Said Masoud (AssistantCommissioner), ForestrySub-commission, MALNR

(8trve4) Majid Muhammed (Director),ZamibwRainfedRiceDevelopment Project,MALNR

(8l1ve4) Dr Gerard C. A. Bruin (PlantPathologist/ ChiefTechnicalAdvisor), Strengthening the Plant ProtectionDivision of Zanabw, Plant ProtectionDivision, ResearchSubcommission, MALNR (81| | 194) Maulid Musa Takrima (Director of Trade and Marketing), Ministry of Trade, IndustriesandMarketng (9I I U9a) 27

Omar Salim Ali (StatisticalOfficer), StatisticalUnit, Planning Sub-commission, MALNR (elrve4) HassanUssy Ameir (StatisticalOfficer), StatisticalUnit, PlanningSub-commission, MALNR (elrve4) I. H. Makame (Statistician), DemographicSection,Departmentof Statistics,Ministry ofPlanning(9llll94) (l0llll94) A. S. Bajaria (Manager),NationalBankof Commerce,Shangani,Zanzibar Dr Mohamed A. El-Lakany (Adaptive ResearchSpecialist),IFAD Smallholder MALNR (l I I | | I94) SupportProject,ResearchSub-commission, Professor Sayed H. Nassar (VegetableResearchConsultant),IFAD Smallholder MALNR (l I I | | I94) SupportProject,ResearchSub-commission, Haji Suleiman Hamad (Head),Food Crops Research,ResearchSub-commission,

(rvrve+) MALNR

28

Annex 4: Anthroooloqistsand Other Researchers

HegeAasbo,currentlydoing short-termanthopologicalresearchon socialandurban spaceinZandbu town for anM.A. at the Universityof Troms6,Norway. Glen Cameron, hasworked for someyearswith women'sco-operativeson Pemba islandand is writing a political sciencePh.D for the Schoolof Orientaland African Studies,London,on this subject.He hasalreadyproduceda shorthistoricaloverview of the co-operativemovementinZamrbu (Cameron1992). Helle Goldman, currentlywriting up a Ph.D at New York Universitydealingwith issuesof identity, stratificationand ideologicalhegemonyamongthe Swahili,Arabs and otherson Pembaisland. This is basedupon anthropological fieldwork conducted in the villagesof Mzambarauni andFurahabetween1992and1994. Peter Green, startingPh.D researchfor StockholmUniversity,Sweden,on the spatial conceptualisation of naturalresourcesand environmentin fishingvillagesin Unguja. Alreadyundertakenpreliminaryresearchin Nungwi. Bethan Jones, has lived and worked in Zamibar town for four years and is consideringstarting an M.A. and researchingthe changingstatus of women in the town. Karina Kielmann, a doctoralcandidateat JohnHopkinsUniversitySchoolof Public Health, USA, and currentlyworking on a GTZ internshipand a four-month study of women'sperceptionsof fertility andinfertilityon Pemba(not connectedto her Ph.D). Due to leavePembain January1995. Eberhard Krain, worked for the NationalCoconutDevelopmentProgrammefrom 1986andin 1990beganresearchfor a Ph.D on agriculturaleconomyand institutions in Zamibar (working on both islands)for the Friedrich-Wilhelms University,Bonn, Germany. He has recentlypresentedthe preliminaryconclusionsof this work to papers. MALNR in a seriesof discussion Kjersti Lrrsen, completeda Ph.D at the Universityof Oslo, Norway, and recently publisheda book (in Norwegian)on inititationandmarriageritualsnZaruibw town. Kimberley Rose,recentlycompletedresearchon health,nutritionandrelatedissuesin Kukuu andnearbyvillagesof southPemba.The preliminaryresultsof this researchare availablein a reportwrittenfor the Savethe ChildrenFund,Pemba(Kimberley1994). Ulrike Seibel, recently completedfield researchin fishing villages on Unguja, includingUroa andNungwi, looking at nutritionand householdresources.Currently BurkinaFaso. basedin Ouagadougou,

29

Chris Walley, starting anthropologicalfield researchfor a Ph.D at New York University,US.t looking at the issuesof tourism and environmenton Unguja and Mafiaislands.Currentlyfocusinguponthe latter.

30