Best practice endline survey report: Uganda coffee farm college cohort 2018 Implemented by TechnoServe

FEBRUARY 2021

Best Practice Endline Survey Report

Uganda Coffee Farm College Cohort 2018

Implemented by TechnoServe

www.laterite.com

2

EXECUTIVESUMMARY..................................................................................................................4INTRODUCTION...............................................................................................................................7RESEARCHDESIGN.........................................................................................................................8HOUSEHOLDDEMOGRAPHICS.................................................................................................11AGRONOMYBESTPRACTICES..................................................................................................12REJUVENATION...........................................................................................................................................................16PRUNING........................................................................................................................................................................17INTEGRATEDPESTANDDISEASEMANAGEMENT(IPDM)....................................................................18SAFEUSEOFPESTICIDES.......................................................................................................................................19COFFEENUTRITION..................................................................................................................................................20WEEDING.......................................................................................................................................................................22MULCHING.....................................................................................................................................................................23SOILEROSIONCONTROL........................................................................................................................................24SHADEMANAGEMENT.............................................................................................................................................25RECORDKEEPING......................................................................................................................................................26

OTHERAGRONOMYPRACTICES..............................................................................................27COFFEEDRYING..........................................................................................................................................................27COFFEEPLANTING....................................................................................................................................................27INTERCROPPING.........................................................................................................................................................29

HOUSEHOLDCHARACTERISTICS............................................................................................30HOUSEHOLDCOMPOSITIONANDFARMERAGE.........................................................................................30EDUCATION...................................................................................................................................................................31EMPLOYMENTSTATUS–OUTSIDETHECOFFEEFARM..........................................................................32PAIDLABORANDCHILDLABOR.........................................................................................................................32LANDOWNERSHIP....................................................................................................................................................33COFFEEPRODUCTION..............................................................................................................................................34ASSETOWNERSHIP...................................................................................................................................................35

PERCEPTIONS&IMPACT...........................................................................................................37THEMEOFINTEREST:GENDER...............................................................................................39FINANCIALDECISIONS(INCOME&SAVINGS)..............................................................................................39FOODDECISIONS........................................................................................................................................................42ALLDECISIONS............................................................................................................................................................44

TABLEOFCONTENTS

3

COFFEEFARMRESPONSIBILITY.........................................................................................................................45INTRA-HOUSEHOLDLABOURDISTRIBUTION..............................................................................................46

FINANCIALPROFILE....................................................................................................................47INCOMEDISTRIBUTION..........................................................................................................................................47SAVINGS..........................................................................................................................................................................48

VULNERABILITYPROFILE.........................................................................................................49FINANCIALSHOCKS..................................................................................................................................................49FOODSHORTAGES.....................................................................................................................................................49HOUSEHOLDDIET......................................................................................................................................................50PROGRESSOUTOFPOVERTYINDEX.................................................................................................................51MULTI-DIMENSIONALPOVERTYINDEX.........................................................................................................52

APPENDICES..................................................................................................................................54

4

The TechnoServe Coffee Farm College Program, supported by Stichting Coffee Agronomy

Training,KeurigDrPepper,JacobsDouweEgberts(JDE),andEnveritas,isafour-yeartraining

program that aims to improve incomes for 30,000 coffee farming households in Central and

WesternUgandabyincreasingtheircoffeefarmproductivity.Thiswillbeachievedbypromoting

improvedcoffeeagronomicpracticesatthefarmlevelthroughatwo-yearstructuredandfarm-

basedtrainingprogramandtheimprovementofcoffeefarmers'accesstorecommendedinputs.

TheprogramworkedwithfarmersintheSembabuledistrictfromAugust2018toendofOctober

2020,deliveringatotalof20trainingsessions.Farmershavebeentrainedon14topicsdesigned

toincreaseyieldsandimprovecoffeequalitywhichinturnwillincreasefarmerincomeandlead

to improved socio-economic outcomes for their families. There was some disruption to the

trainingcausedbyCOVID-19,withgrouptrainingstoppedfromthemiddleofMarch,andoneto

onehouseholdon-farmtrainingdeliveredtotheendoftheprogram.Giventhaton-farmtraining

tooktwotothreemonthstoreachallhouseholds,despiteashortextensiontotheprogram,two

plannedreviewmoduleswerenotdelivered.A total of 8,262 households registered in this

cohortandattendedatleastonetraining.Ofthese,6,916farmersfrom6,161households(75%

ofregisteredhouseholds)canbeconsideredas“trained”havingattendedatleastseven(50%)of

the14topics,and39%oftrainedfarmersarewomen.Thetargetforthe2018Cohortis6,000

householdstrained.

Farmerswere assessed on ten agronomy best practices: record keeping, weeding, coffeenutrition,safeuseofpesticides,integratedpestanddiseasemanagement(IPDM),erosioncontrol,

shademanagement, rejuvenation, pruning, andmulching. Adoption of these best practices is

based on rules developed by TechnoServe based on their experience and expertise in coffee

farming.

Figure1.PercentageofHouseholdsAdoptingeachBestPractice

EXECUTIVESUMMARY

AGRONOMYBESTPRACTICEADOPTIONSUMMARY

5

AsshowninFigure1,baselineadoptionoftheagronomybestpracticeswaslow.Nohouseholdswere adopting safe use of pesticides and a negligible share of householdswere adopting the

nutritionandrecordkeepingbestpracticesatbaseline.

Theresultsoftheendlinesurveyshowanincreaseinadoptionofbestpracticeswitheightoutofthetenbestpracticeshavingastatisticallysignificantincrease.Mulchingwastheonlybestpracticethatshowedadecreasedadoptionratefrom10%offarmersmulchingduringthe

baselinetoonly4%duringtheendline,andsafeuseofpesticidesremainedat0%.Overall,the

shareofhouseholdsthatadoptedhalf(five)ormorebestpracticesincreasedfrom6%atbaseline

to33%atendline.Inaddition,70%ofhouseholdssawatleastonenewbestpracticeadopted,

andalmosthalf(48%)sawatleasttwonewbestpracticesadopted,sincebaseline.

Thekeydriversofyieldimprovementareexpectedtobepruning,rejuvenation,weedingandpestanddiseasemanagement.Althoughsoilhealthandnutritionareimportant,thesoilsinSembabulearefertileandhighinorganicmatter,unlikelytobelimitingyieldformostfarms.

Asoilandleafsurveyconductedbytheprogramshowedthatsoilswereabletosustainyieldsof

3kilogramsofcherrypertreewithouttheadditionoffertilizer.

Householdmembers.Householdshavesixmembersonaverage,fourofwhicharechildren(0to18years).Overall,thelevelofeducationismoderate,with84%ofthewomenand90%ofthe

menhavingreceivedsomelevelofformaleducation.Aquarter(25%)ofmalefarmerseitherhave

hademploymentinthepastorarecurrentlyemployedoutsideoftheircoffeefarmscomparedto

just16%ofthewomen.

Land Ownership. At endline, farmers own a self-reported mean of 2.1 hectares of totalagriculturalland,0.9hectaresofwhichisplantedwithcoffee,withanaverageof992coffeetrees.

HouseQuality.Thebuildingqualityofthefarmers’housesintheSembabuledistrictisgood,withthemajorityofhouseshaving improvedroofs,walls,and floors.Over four-fifths (85%)of the

householdshaveaccesstoelectricity,eitherfromthemainlineorsolarpanels.

WASH. Access to water, sanitation, and hygiene has increased, but remains limited. 12% ofhouseholdshaveaccesstoanimprovedtoilet,upfromjust7%atbaseline.Only26%ofthehouses

haveanimprovedwatersource,animprovementfrom17%atbaseline.

Assets.Manyhouseholdshavetheirownmeansoftransport–68%ofhouseholdsownabicycle,39%ownamotorcycle, and4%owna car.Mosthouseholds (95%)have at least onemobile

phone. Most households have some livestock, with 70% of households owning at least one

chicken.

CoffeeIncome.Atendline,coffeehouseholdswerehighlydependentoncoffee,with60%ofthehouseholdsreportingthatmorethanhalfoftheirtotalhouseholdincomecomesfromcoffee.This

isan increase fromthebaseline,wheremajorityofhouseholds(64%)reportedcoffee income

representedbetween25-50%ofincome.

HOUSEHOLDCHARACTERISTICSSUMMARY

FINANCIALPROFILESUMMARY

6

Savings.Mosthouseholds(86%)wereusingatleastonesavingsmethodatendline,comparedtoonly50%atbaseline.Thisincreasewasdrivenbyhouseholdsjoininginformalsavinggroups.

Shocks&Shortages.Amajorityofthehouseholdsatendline(70%)reportedtheyhavebeenaffectedbyatleastoneseriousfinancialshockinthepastyearandlessthanhalf(40%)reported

havingsufferedfromafoodshortageinthepastyear.Thisisanimprovementcomparedtothe

baseline,whenahighershareofhouseholdsreportedfinancialshocksandfoodshortages.

Poverty.UsingtheProgressoutofPoverty(PPI)andMulti-DimensionalPoverty(MPI)indices,householdswereassessedontheirpovertyprofilesatendline.Only17%ofthehouseholdswere

likelytofallbelowtheinternational$1.90/day2011PPPpovertylineusingthePPI,amuchlower

proportioncomparedtoruralhouseholdsinCentralUganda(22%).UsingtheMPI,two-thirds

(66%) of the households are considered multidimensionally poor, compared to 55% of the

householdsnation-wide.

Farmerperceptionsonyield,time,andincome.Coffeefarmersoverwhelminglyself-reportedincreasesincoffeeyields,timespentworkingoncoffee,andincomefromcoffee,sincejoiningthe

CoffeeFarmCollege inAugust2018.81%ofhouseholdsreportedexperiencingan increase in

yield sinceattending training, and40%acknowledged this increasewassubstantial.For time

spentworkingoncoffee,84%ofhouseholdsreportedspendingmoretimeoncoffee,with31%

statingthisincreasewashigh.Finally,57%ofhouseholdsreportedexperiencinganincreasein

income,with13%statingthisincreasewashigh.

Contactmodalities.Farmerswereaskedabouttheusefulnessofadditionalvisits,phonecalls,andSMSmessagesfromTechnoServefarmertrainers.Onascalefrom0to10,visitswereseenas

the most useful modality (mean score of 7.4), followed by phone calls (7.3), and then SMS

messages(6.8).

PERCEPTIONSSUMMARY

7

InOctober2020,thetwo-yearCoffeeFarmCollege,whichbeganinAugust2018,endedinthe

SembabuledistrictoftheCentralregionofUganda.Throughoutthecourseofthisprogram,Coffee

FarmCollegefarmersweretrainedon14topicstosustainablyimprovetheircoffeeyieldsand

supporttheirhouseholds.

Thisreportanalyzesdatafromanendlinesurveyof581householdsconductedinOctoberand

Novemberof2020tomeasuretherateofadoptionofagronomyBestPractices(BPs)aftertwo

years of the program and collect information on household characteristics. The survey was

conductedbothatthefarmandthehouseholdlevel.

This report hasbeen createdwith a strong emphasis on learning about theprofiles of coffee

farmersintheSembabuledistrictofUganda.Weexploretherelationbetweenattendanceinthe

trainingprogramandbestpracticeadoption,aswellashouseholdcharacteristicswhichcouldbe

influencing participation and adoption. We also explore intra-household gender relations

includingdecision-makingandlabourroleswithinhouseholds.Theseinsightswillhelpinform

thetrainingcontentandtheongoingimplementationoftheFarmCollegeinothercohorts.

Figure2.Mapofsurveyedhouseholds,colouredbyparish(Courtesy:OpenStreetMapcontributors&Tableau)

INTRODUCTION

8

LateriteLtd., adata, research, andadvisory firm, led thedevelopmentof the researchdesign,

supportedsurveydevelopmentandimplementation,andconductedtheanalysisforthisstudy.

AteamofeightenumeratorsweremanagedbyTechnoServetoconductthesurveyusingtablets.

DesignandSampleSelection

Overalldesign.Thisisatwotime-periodstudy(baselineandendline)trackingatotalof600households before and after the start of the Coffee Farm College trainings delivered by

TechnoServeinUganda’sSembabuleDistrict.Allhouseholdsincludedinthesamplearefarmers

that signed-up during sensitization to participate in the program. This study does not use a

counterfactual.

Samplingstrategy.Thedesignforthisstudyisbasedonaone-stageclusteredrandomsamplingmethod,withstratificationattheparishlevel:

• Stratification:thesamplewasstratifiedover15parishesacrossthreesub-countiesinthe

SembabuledistrictofUganda.

• Clustering:withineachparish,40householdswererandomlyselectedfromFarmCollege

sign-uplistscreatedduringtraininggroupformation.

Variablesof interest.Thisstudyfocusesonthreesetsofvariables:(i) farmerandhouseholdcharacteristics, including data on demographics, diet, assets, and labour; (ii) coffee-related

characteristics, including farm-based observations of agronomy best practices; and (iii)

participation levels in theprogram.Households showedvaryingdegreesof attendance in the

trainingsessionsconductedoverthetwoyears.

Hypotheses.Themainhypothesisunderlyingtheresearchdesignisthathighattendancefarmers(i.e. the farmers thatattendhalformoreof the training topicsessions)willbemore likely to

experienceanincreaseintheadoptionofagronomybestpracticesbytheendoftheCoffeeFarm

College. In addition, our prior experience is that households with better socio-economic

indicatorsatbaselinewillbemorelikelytoattendthetrainingsessionsandthusmorelikelyto

adopt agronomy practices. Consequently, we will seek to measure the relative best practice

adoption amongst the different attendance groups and also examine socio-economic

characteristics of these farmers to understand some of the determinants of adoption. It is

importanttonotethatassociationsbetweenattendanceandadoptionofagronomybestpractices

willnotconstituteevidenceofimpact;rather,suchassociationsareexpectedsymptomsthatwe

would observe in the casewhere the program did have an effect on adoption of coffee best

practices.

BestPracticeBaselineandEndlineSurveys

SurveyCompletion.HouseholdswerevisitedinSeptemberandOctober2018forthebaselinesurvey,and inOctoberandNovember2020 for theendlinesurvey. Inbothsurveys,datawas

collected on demographics, diet, assets, and labour. In addition, farm-based agronomy best

practiceswereobserved.Theoverallrateofsurveycompletionatbaselinewas100%with600

households interviewed. The overall rate of survey completion at endlinewas 97%,with 19

householdseitherrefusingtobesurveyed,orunabletobesurveyed.Thedatawasweightedusing

inverse probability weights. These weights were determined by taking the inverse of the

probabilityofsamplingeachhouseholdsigned-upwithinaparish.

RESEARCHDESIGN

HHDESIGN

9

MatchingtoAttendance.HouseholdsthatparticipatedinthebaselinesurveyandattendedtheCoffee Farm College are identified via a baseline farmer tracking survey and via in-house

matching.Intotal,464(77%)ofthebaselinehouseholdsinthesamplewerematchedbeforethe

startoftheendlinesurvey.Duringtheendlinedatacollection,anadditional38householdswere

matchedtotheirattendancerecords,foratotalof502(86%)matchedhouseholds.Households

thatarenotmatchedareconsideredasnotattendingCoffeeFarmCollege.

Survey Sections. The surveys included two main sections. The first section covered thedemographic and socio-economic status of each household. With regard to demographics,

farmerswereaskedabouttheirmaritalstatus,age,education,employmentstatus,andhousehold

size.Therespondentswerealsoaskedtoprovidearecountofeachchildlivinginthehousehold,

theirage,andeducationstatus.Intermsofsocio-economiccharacteristics,farmerswereasked

aboutlandandassetownership,incomesources,nutritionaldiet,andfinancialshocks.Therewas

alsoastrongemphasisoncoffeetreeownershipandproduction;farmerswereaskedaboutthe

numberof coffee treesownedand themethodsandquantitiesof coffee soldduring themost

recentcoffeeharvestseason.

Thesecondsection focusedon theagriculturalpractices thatwere taughtduringCoffeeFarm

College. First, farmers were asked to show records on coffee sold and costs incurred in the

productionofcoffee.Farmerswerethenaskedaseriesofquestionsandobservationsregarding

pesticides, including products used on the coffee farm and howpesticidesweremanaged, to

determinewhether the farmerswere using these products safely. Finally, the data collectors

requestedwhethertheycouldvisitfarmers’maincoffeefieldtodeterminetheimplementation

ratesofthegoodagriculturalpracticesattheendoftheCoffeeFarmCollege.Thisappraisement

wasprimarilybasedondirectobservationofthecoffeefieldbythedatacollectors.

Additional Questions at Endline. In the endline survey, additional questions around intra-household decision-making and gender relations were added, as well as questions around

householdperceptionsoftheTechnoServeCoffeeFarmCollegeanditsimpactoncoffeeyields,

timespentoncoffeeproduction,andincomeearnedfromcoffee.

10BestPractices.Overall farmerswereassessedon their1) recordkeeping,2)weeding,3)coffeenutrition,4)integratedpestanddiseasemanagement(IPDM),5)safeuseofpesticides,6)

erosion control, 7) shade management, 8) rejuvenation, 9) pruning, and 10) mulching best

practices.

BestPracticeRules.Eachbestpracticewasassessedusingspecificrulesbasedonagronomicbestpracticesforcoffeefarming.Coffeehouseholdsmeetingtheseruleswereconsideredtohave

adoptedthebestpractice. In thisendlinereport,differentruleswereusedthanatbaseline in

ordertobeconsistentwithotherTNSreports.Assuch,baselineresultsreportedinthisendline

reportdiffertothoseinthebaselinereport,astheyhavebeenrecalculatedusingdifferentrules.

Appendix5documentsthesechanges.

Endline Sample. The analysis of best practices and household characteristics has beenstructuredaroundtheentireendlinesampleof581coffee-farminghouseholds.

10

Thereportcontinueswithasummaryofthemainhouseholdcharacteristicsandaprofileofthe

average Sembabule coffee farming household participating in the Coffee Farm College. This

profileisfollowedbyadetaileddescriptionoftheagriculturalpracticeendlinesurveyresultsand

adetaileddescriptionoftheselectedhouseholdcharacteristics.

Picture1.Coffeetreeloadedwithcoffee,SembabuleDistrict

11

HOUSEHOLDDEMOGRAPHICS

12

Farmerswereassessedon10agronomybestpractices (BP): rejuvenation,pruning,mulching,

coffeenutrition,intercropping,weeding,shademanagement,erosioncontrol,integratedpestand

diseasemanagement(IDPM),safeuseofpesticides,andrecordkeeping.

Table1.BestPracticeAdoptionRatesatBaselineandEndlineIncludesallsurveyedhouseholdsatendline,bothattendingandnon-attending Baseline Endline 2018 2020

No. of Households 581 581 Best Practices % Adoption Rate % Adoption Rate Shade Management*** 51% 67% Weeding*** 40% 64% IPDM*** 40% 61% Erosion Control*** 38% 57% Rejuvenation*** 29% 55% Pruning*** 16% 46% Coffee Nutrition*** 3% 17% Record Keeping*** 2% 4% Mulching*** 10% 4% Safe Use of Pesticides 0% 0% Average # of best practices adopted *** (out of 10) 2 BPs 4 BPs

Households adopting 5 or more best practices ***

6% 33%

Households adopting 2 or more additional best practices at endline compared to baseline

N/A 48%

Notesforinterpretingthetableofresultspresentedabove:

§ Thestars(*) indicatethesignificancelevelofbestpracticeadoptiondifferencesbetweenbaselineandendline,usingalogitregressionwithsamplingprobabilityweighing,clusteredattheparishlevel.

§ ***Significantatthe1%level,**significantatthe5%level,and*significantatthe10%level.§ SafeUseofPesticidesisdefinedforasamplesizeof421householdsatbaselineand419householdsat

endline,asallnon-usersofpesticideareconsideredas“NotApplicable”.§ Thepercentagesareweightedusingsamplingprobabilityweights.

Adoptionofbestpracticesvariedwidelyatbothbaselineandendline,asshowninTable1.Theendline survey results show a statistically significant increase in the average number of best

practicesadopted,from2atbaselineto4atendline,and48%ofhouseholdsadoptedtwoormore

additionalbestpracticeswhencompared tobaseline.Allbestpractices (excludingsafeuseof

pesticides) reported a statistically significant change in adoption rate, eight practices in an

increasingdirectionandonepractice inadecreasingdirection (mulching).Thepercentageof

householdsadoptingatleasthalf(5)ofthebestpracticesincreasedfrom6%atbaselineto34%

atendline.

Forthepracticeswherefarmersscoredrelativelywell:

AGRONOMYBESTPRACTICES

13

ShadeManagementiscommoninmixedcoffeefarmingsystems,suchasthoseinUganda,withhigh numbers of fruit trees, such as bananas in the coffee fields, resulting in relatively high

adoption of the shademanagement best practice (67%at endline). Almost half (48%) of the

householdshadgoodshadelevelsatendline(>20%shade)and40%ofhouseholdshadplanted

shadetreesinthepastthreeyears,whichwillprovideshadeintime.

Weeding canhaveadramatic impactoncoffeeyield.CoffeeFarmCollege taught farmers theimportance of weeding under the canopy by hand-pulling, using a panga, or mulching. The

trainingalsoexplainedwhyweedingbydiggingunderthecanopyisbadforcoffee.Adoptionof

theweedingbestpracticeincreasedfrom40%to64%frombaselinetoendline.Thiswasledby

asignificant increase in thepracticeofpullingweedsbyhandaswellasbyadecrease in the

practiceofweedingbydigging.

IPDMadoptionincreasedfrom40%adoptionatbaselineto61%atendline.CoffeeFarmCollegetaughthouseholdshowtouseculturalmethods,suchascrophygiene,traps,goodnutrition,etc.

toreducetheincidenceofpestsanddiseases,ratherthanreactivelymanagingthepestordisease

once the tree has been infected. At endline farmers reported knowing more methods for

preventingbothtwigborersandcoffeewiltdisease(CWD).

Figure3.PercentageofhouseholdsadoptingeachBPatBaselineandEndline

TherearenumerousBestPracticeswherefarmerscontinuetohavelowadoption:

SafeUseofPesticidescontinuestobeaconcern,withnohouseholdspassingthebestpractice,despite72%ofthefarmersusingatleastonetypeofpesticideinthepast12months.Almostall

households(99%)failedfornotusingtheminimumappropriatepersonalprotectiveequipment

(mask,goggles,andgloves).

CoffeeNutritionpracticescontinuetobepoor,with50%offarmsnotusinganyfertilizer,and17%usingthewrongfertilizers,suchasUreaandCANinsteadofNPK(25:05:05or22:06:12),

DAP,compost,ormanureandzincfoliarfeeds.

14

Mulchingwas theonlybestpractice that reportedsignificantdecrease inoverall adoptionatendline (4%) compared to baseline (10%). While the use of mulch increased from 16% of

households applyingmulch to19%at endline, thebest practice formulching requires a2cm

minimumthickness,whichthehouseholdsdidnotmeet.

Figure4showsthatonly6%ofthehouseholdswereadoptinghalf(five)ormoreofthe10BPsatbaseline.Atendline,thisfigureincreasedto33%.Atendlinemosthouseholdswereadopting

threetofivebestpractices.

Figure4.DistributionofnumberofBPsadoptedatBaselineandEndline

Figure5showsthat70%ofhouseholdssawanimprovementofatleastonenewbestpracticeadoptedsincebaseline,andalmosthalf(48%)sawanimprovementoftwoormorebestpractices.

Figure5.ChangeinnumberofBPsadoptedcomparedtoBaseline

Understanding the overall adoption rate of best practices of coffee-farming households at

baselineandendlineiscriticaltoobservetheeffectoftheprogram,butitisequallyinteresting

from a programmatic perspective to investigate the correlations between attendance to the

TechnoServeCoffeeFarmCollegeandbestpracticeadoption.

10%

21%

27%

23%

13%

5%

1% 0%1%

5%

15%

20%

25%22%

9%

2%

0 1 2 3 4 5 6 7

Perc

enta

ge o

f hou

seho

lds

Number of BPs Adopted

Distribution of Number of BPs Adopted: Baseline and EndlineAll Households (n = 581)

Baseline(2018)

Endline(2020)

15

Thissub-sectionanalyzesbestpracticeadoptionatendlineforthe502householdsthatcouldbe

matchedwith attendance records to themonthlyCoffee FarmCollege trainings.1 It compareshouseholdswhohaveattendedmorethanhalfofthetrainingtopics(7outof13)2,the“TrainedHouseholds”, to households who attended less than seven topics (1 to 6 topics) the LowAttendanceHouseholds.

Foracomparisonofthecharacteristicsoftrainedandlowattendancehouseholds,pleaserefertoAppendix1.Foracomparisonofbestpracticeadoptionbytrainingoneachspecifictopic,pleaserefertoAppendix2.Foradoptionratesfornon-attendinghouseholds,pleaserefertoAppendix3.Fortheparisheswithhighestadoptionrates,pleaserefertoAppendix4.Table2.LowAttendancevs.TrainedHouseholds:BestPracticeAdoptionRatesbyTrainingAttendanceLevelsatEndline Low Attendance

Households Trained

Households Overall Average Overall Average No. of Households 63 443 Shade 60% 66% Weeding 58% 65% IPDM 58% 64% Erosion Control* 72% 57% Rejuvenation 53% 54% Pruning 50% 46% Coffee Nutrition 16% 18% Record Keeping 4% 5% Mulching 5% 4% Safe Use of Pesticides 0% 0% Average # of best practices adopted (out of 10) 4 BPs 4 BPs

Households adopting more than half (5) of the best practices

26% 35%

Notesforinterpretingthetableofresultspresentedabove:

§ Thesignificancelevelofbestpracticeadoptiondifferencesbetweenbaselineandendlinewasestimatedusingalogitregressionwithsamplingprobabilityweighing,clusteredattheparishlevel.

§ SafeUseofPesticides is defined fora sample sizeof419households, asall non-usersofpesticideareconsideredas“NotApplicable”.

§ Thepercentagesareweightedusingsamplingprobabilityweights.

Ofthehouseholdsthatcouldbematchedtoattendance,88%,(443outof506)hadattendedat

leastseventopicsoftheCoffeeFarmCollege,andwithinthese443households,75%attended

11–13topics–indicatingahighattendancerate.Ofthe63householdsinthe“LowAttendance”group,60%attended4orfewertopics.Householdsinthe“Trained”grouphavehigheradoptionratesin6ofthe10BPsandloweradoptionratesin4BPs.However,onlyoneofthesedifferences

wasfoundtobestatisticallysignificant-erosioncontrol,whichislowerfortrainedfarmers.Other

thanthis,householdsattendingmorethan7trainingtopicsisnotassociatedwithgreaterbest

1Thebaselinesamplewasrandomlydrawnfromthosethatsigned-upaspotentialprogramparticipants,ratherthanfromattendancesheets.Farmerswerethenmatchedtotheirattendancerecordsanditisassumedthatifnotmatchedtotheattendancerecordsthatthehouseholddidnotattendtraining.2Atthetimeofwriting,attendancedataisavailableuntilAugust2020anddoesnotincludethelasttrainingonsustainablefarming.

16

practiceadoptiononanyoftheother9bestpracticeswhencomparedtohouseholdsattending

fewer than7 training topics.Whencompared tobaseline,however,householdswhoattended

trainingshaveimprovedbestpracticeadoption.

Rejuvenationadoption increasedsignificantly from29%at

baselineto55%atendline–thesecondmostimprovedbest

practice after pruning. As seen in Figure 6 the maindeterminantofthisincreaseinadoptionwasthereductionin

thenumberofstemsfrom5ormore,downto3or4main

stems.Atbaseline,47%ofhouseholdsaveraged5ormore

mainstems.Atendline,thisfigurewasjust23%,indicatinga

significantdecrease.

FarmersinUgandatypicallyrejuvenatetheircoffee,butnot

systematically.Theywillallownewsuckerstogrowatany

timeandoftenhaveupto10mainstems,allofdifferentages,

someolderthaneightyears.TheCoffeeFarmCollegetaught

farmersabouttheimportanceofrejuvenatinginasystematic

way,maintaining youngmain stems (less than eight years

old) and keeping nomore than fourmain stems, although

twoorthreeareideal.

Whenaskedwhyfarmersdecidedtorejuvenatetheircoffee

farm, 70% mentioned that learning and practicing this

activity in the training session led to their decision to

rejuvenate.25%mentionedthatthetrainervisitingtheirfarmwaswhatledtotheirdecision,and

28%mentionedseeingtheimpactinthedemonstrationplotasthereasonforrejuvenating.

Figure6.NumberofMainStemsatBaselineandEndlineMaximumoffourmainstemsrecommended

REJUVENATION

Picture2.Rejuvenatedcoffeetree

17

Pruningwasthemostimprovedbestpracticewithanincreasefrom16%atbaselineto46%at

endline. This increase was due to an increase in the number of pruning methods used, as

householdsaveragedlessthan2methodsatbaseline,andmorethan3methodsatendline.

TheCoffeeFarmCollegetaughtfoursimplepruning

methods: remove the dead branches, the branches

touchingtheground,theunwantedsuckers,andthe

cross-overbranchesinordertoopenthetreecentres.

Pruningopensupthetreetolightandaircirculation,

bothprerequisitesforgoodfloweringandlowdisease

and pest incidence. Adoption of the pruning best

practice required the trees to be pruned using a

minimumofthreeofthefourmethodsthattheCoffee

FarmCollegetaught.

AsshowninFigure6,therewasincreaseintheusageof three of the four pruning methods between

baseline and endline. The opening tree centres

methodincreasedfrom21%adoptionatbaselineto

71%atendline.Twoothermethodsimprovedbyat

least18percentagepoints,andonemethod(removal

of unwanted suckers) decreased by 10 percentage

points.

Figure7.PruningMethodsUsed–Baselinev.Endline

21%

51% 50%

29%

71%

41%

68%55%

Centres opened Unwanted suckers removed Dead branches removed Branches touching the groundremoved

Perc

enta

ge o

f hou

seho

lds

Pruning Methods

Pruning Methods Used: Baseline and EndlineAll Households (581)Baseline

(2018)Endline(2020)

PRUNING

Picture3.Apartiallyprunedcoffeetree

18

IPDMadoptionincreasedfrom40%adoptionatbaselineto61%atendline.Thisimprovement

wasdrivenbyanincreaseintwigborerandcoffeewiltdisease(CWD)preventionmethods.At

baseline,farmersknew1.7methodsonaverageforpreventingtwigborers,and0.5methodson

averageforpreventingCWD.Atendline,theseincreasedto1.9methodsfortwigborerprevention

and0.7methodsforCWDprevention–amodestincrease.

Theuseofpesticides– insecticides, fungicides,andherbicides– ismoderatelyhighinUganda

(72%).Manyfarmersaresimplyunawareofalternativepestanddiseasecontrolmethods.The

CoffeeFarmCollegetaughthouseholdshowtouseculturalmethods,suchascrophygiene,traps,

goodnutrition,etc.toreducetheincidenceofpestsanddiseases,ratherthanreactivelymanaging

thepestordiseaseoncethetreehasbeeninfected.

As IPDMmethodscannotallbeeasilyobserved, thesurveyused twoknowledgequestions to

assesshouseholdknowledgeonmanagingacommonpest,TwigBorerandacommondisease,

CoffeeWilt.

Figure8andFigure9showthechangeinknowledgeofthevariousmanagementmethodsforCWDandTwigBorer infestations.After training, the percentage of farmers not knowing any

methodsdecreased,andfewerfarmersreportedincorrectmethods.Therewasalsoanincrease

inknowledgearoundthemostcrucialpractices,theuprooting,removal,andburningofinfested

twigsandtrees,whichmorethantwo-thirdsofhouseholdscorrectlyidentifiedasamethodat

endlineforbothCWDandTwigBorers.

Figure8.CoffeeWiltDiseaseManagementMethodsKnown

INTEGRATEDPESTANDDISEASEMANAGEMENT(IPDM)

19

Figure9.CoffeeTwigBorerManagementMethodsKnown

DuetothehighuseofpesticidesinUganda(72%reportedusingpesticidesatendline),safeuse

of pesticides is an important best practice. Unfortunately, as was the case at baseline, no

householdspassedthisbestpractice.Thisfailureisprimarilyduetooneofthefiveelementsof

safepesticideuse–theuseofpersonalprotectiveequipment(PPE)–specificallymasks,goggles,

andgloves.Figure10belowdisplaysthesefiveelementsandshowswherefarmersinSembabuledistricthaveimprovedandwheretheyarestillfallingshort. Figure10.ElementsofSafePesticideUse–Baselinevs.EndlineAllfiveelementsrequiredtobeconsideredsafe.

Safe pesticide usage requires adoption of awide array of elements. The Coffee FarmCollege

taughtfarmersaboutthehazardsofusingpesticides:howtoreadpesticidelabels,whichPPEto

use, the safety procedures for purchasing pesticides, transportation, storage, and disposal of

pesticidecontainers.

10%2%

79%70%

12%1% 1% 0%

7%1%

87%80%

10%4% 1% 0%

No methodsknown

Incorrectmethods

mentioned

Removeinfested twigs

Burn infestedtwigs

Removeunwantedsuckers

Spraypesticidescontaining

imidachloprid

Spray otherpesticides

Reduceshade

Perc

enta

ge o

f hou

seho

lds

Twig Borer Management Methods: Baseline and EndlineAll Households (n = 581)

Baseline(2018)

Endline(2020)

SAFEUSEOFPESTICIDES

20

Whenusingpesticides,farmersshouldwearPPEtostopanypossiblerouteofentryintothebody

bywearingamask,goggles,gloves,spraysuit,andboots.Glovesareespeciallyimportantashands

canbeexposedtotheundilutedproductwhenmixingpesticides.Farmersareconsideredtobe

usingpesticidessafelywhenthey:i)havebeentrainedonhowtousepesticidessafely,ii)donot

storepesticidesinsidetheirhomesandinalockedstoreoutsideoftheirhome,iii)weargoggles,

amask,andglovesataminimum, iv)discardpesticidecontainersbyburningthem,returning

them to a contractor, or returned to the source, andv)donotusebannedpesticides suchas

Parathion,Carbofuran,orDieldrin.

Despite significant improvements in safe storage, disposal, and training received, 99% of

householdsfailedthePPErequirement.Figure11showsthisinmoredetailbyhighlightingthespecificitemsofPPEthatarelacking.Masks,gloves,andgogglesareallrequiredtopass,yetall

threeoftheseitemshaveverylowlevelsofadoptionatendline(17%,13%,4%respectively).

Furtherinvestigationisrequiredinconsultationwithconcernedstakeholderstounderstandif

pooradoptionofPPEisanissueofavailabilityorcost,similarlowadoptionratesareseenin

othercountriesinEastAfrica.

Figure11.PersonalProtectiveEquipmentUse–Baselinev.Endline

AdoptionofthecoffeenutritionBPincreasedfrom3%atbaselineto17%atendline.50%offarms

surveyedatendlinearenotusinganyfertilizer(includingcompostormanure),and17%areusing

thewrongfertilizers,suchasUreaandCANinsteadofNPK(25:05:05or22:06:12),DAP,compost,

ormanureandzincfoliarfeeds.

Generally, the soils in Uganda are some of the best in East Africa; the soil and leaf survey

conducted at baseline indicated which nutrients were lacking and provided parish-specific

recommendations,bothtocorrectthesoilsandfeedthecoffeetrees.Usingjustnutrientsavailable

inthesoilcoffeetreescanproduceupto3kilogramsofcherrypertree.Forfarmerswhowishto

improveyieldsfurtherandarenotaimingfortheorganicmarket,theuseofadditionalnutritional

products can increase yields. Coffee Farm College farmers were taught the importance of

balancedcoffeenutrition,makingandapplyingorganicproductssuchascompostandmanureas

wellasapplicationofcertainformulationsofNPK,atthecorrectrate(usingameasure)andin

35%

3% 4%

38%

1% 1% 1%

37%

17%13%

61%

4% 6% 3%

No protectionused

Mask Gloves Boots Goggles Spray suit Hat

Perc

enta

ge o

f hou

seho

lds

Protective Equipment Used: Baseline and EndlineHouseholds applying pesticides (n = 419)

Baseline(2018)

Endline(2020)

COFFEENUTRITION

21

the correct place, under the tree canopy where the

feeder roots grow, in order to prevent fertilizer

wastage.

Thecoffeenutritionbestpracticerequiresthattrees

shownosignsofnutrientdeficienciesandfarmsapply

atleastonerecommendednutritionalproductwhich

canincludecompostormanure(countedasone),NPK,

DAP,limeorzincfoliarfeed.

Atendlinehalfofhouseholds(48%)didnotuseany

nutritionalproducts,includingcompostormanureon

their farms, and 5% of farms showed nutritional

deficiencies. Only 27% of farms used compost or

manure,aslight increase fromthebaseline. 21%of

farms used a recommended formulation of NPK, up

from7%atbaseline.Acombined13%offarmsused

NPK17:17:17(6%notrecommended),CAN(7%not

recommended) or Urea (4% not recommended).

Figure12displayssomeoftheimprovementsmadesince the baseline and also highlights the need for

increaseduseofappropriatenutrition.

Intermsofapplication,42%offarmsusinggranularfertilizersdidnotuseameasuretoapply

fertilizer, and thus do not know the amount given to each tree. All households applied the

fertilizerinthecorrectplace,whichisunderthetree,anddidnotbroadcastitoverthefield.This

isanimprovementfromthebaseline,where2%ofthehouseholdsapplyinggranularfertilizers

broadcastedthefertilizeralloverthefield.

Figure12.FertilizerUsedinthePrevious12Months–Baselinevs.Endline

Forcomposting,therewasanincreaseincompostingseenonthefarmfrom7%ofhouseholds

makingcompost(eithercompostpit,heap,ormanureheap)atbaselineto14%atendline.This

assessmentonlyaccounts forcompostormanureheapsseenduringtheendlinevisit, theuse

during the rest of the year shown in Figure 12 is self-reported, indicating the difference innumbers.

Picture4.Nitrogendeficiencyshowingasyellowleaves

22

Compostorcompostedmanureisaslowrelease,naturalfertilizerreturningkeynutrientstothe

soiltohelpmaintainsoilqualityandfertility.Inaddition,compostbringsorganicmattertothe

soilwhichhelpstoretainmoistureandfeedsmicroandmacro-organisms.CoffeeFarmCollege

teachesfarmerstomakecompostheaps,whencomparedtomakingcompostinapit,heapsare:

easier to turn,donotgetwaterloggedandwillproducecompost faster.Figure13 shows thedistributionofthethreecompostmethodsforsurveyedfarmers,anddepictsthechangesfrom

baselinetoendline.

Figure13.TypeofCompostSeenonFarm–Baselinevs.Endline

Adoptionoftheweedingbestpracticeincreased

from40%to64%frombaselinetoendline.This

wasledbyasignificantincreaseinthepracticeof

pullingweedsbyhand(fromjust3%atbaseline

to59%atendline)aswellasbyadecreaseinthe

practice of weeding by digging, which can

damage the feeder roots. Figure 14 highlightssomeoftheseimprovements.

Weedscanhaveadramatic impactonyield, so

theareaunderthetreecanopy,whichisthemain

feederrootzone,shouldbekeptasweed-freeas

possible throughout the season. This can be

difficultduringtherains,giventhatweedgrowth

israpidintropicalclimates.CoffeeFarmCollege

taughtfarmerstheimportanceofweedingunder

the canopy by hand-pulling using a panga or

mulchingwithslashing,and intercroppingwith

beans as options for between the rows. Coffee

Farm College also explained why weeding by

diggingunderthecanopyisbadforthecoffee,as

it damages the feeder roots and can introduce

CoffeeWiltDisease.

7%

2% 2% 3%

14%

0%

6% 7%

Yes, HH is making compost Layered compost heap Pile of boma manure Compost pitOverall Type of Compost Seen

Perc

enta

ge o

f hou

seho

lds

Compost Seen by Type: Baseline and EndlineAll Households (n = 581)

Baseline(2018)

Endline(2020)

WEEDING

Picture5.Coffeetreeweededbydigging.

23

Adoptionofweedingrequiresthati)thefarmerpullweedsbyhand,orusesaslasherorpangato

weedunderthetreecanopy,ii)thefarmerweedsatleasttwiceperyear,andiii)therearenoor

limitedweedsunderthetreecanopylessthan30cmtall.

Wherepossible,observationsweremadetoassesstheweedingstatus,andotherquestionssuch

asfrequencyofweedingwereaskedofthefarmers.Atendline,16%ofhouseholdsfailedbecause

theywereexclusivelyweedingbydiggingunderthecanopy.Another16%ofhouseholdsfailed

becausemanyweedswereobservedbytheenumeratorunderthecanopyand8%ofhouseholds

failedbecauseweedshadreachedheightsofmorethan30cmunderthecanopy.

Figure14.WeedingMethodsUsed–Baselinevs.Endline

The mulching best practice decreased in

adoption rate from baseline to endline.

However, this decrease was despite the fact

the percentage of farms applying mulch

actually increased, from 16% at baseline to

19%atendline.AsFigure15 illustrates, thereason for the failure is attributed to the

thicknessofthemulchapplied,whichmustbe

morethan2cmtobeconsideredfullyadopted.

Only4%offarmsadheredtothisstandard.

Mulching has many benefits including the

suppression ofweeds, reduction ofmoisture

loss from the soil, reduction of soil

temperatures, and increase of soil organic

matter. Mulching is especially important if

shadelevelsarelow.CoffeeFarmCollegetaughtfarmersaboutthebenefitsofmulchingandhow

tomulch.Adoptionofmulchingrequiresthatthereisatleast2cmthickmulchapplied.

Whileoverall adoptionof thisbestpracticedecreased for this cohort, future trainings should

focusontheimportanceofmulchthicknesstoreapthebenefitsofmulching.

MULCHING

Picture6.Awellmulchedtree.

24

Figure15.MulchApplicationandThickness–Baselinevs.Endline

Adoptionofsoilerosioncontrolbestpracticeincreasedfrom

38%atbaselineto57%atendline.Ahouseholdisconsidered

fully adopted if they employ one or more erosion control

methodsonthefarm.Thisincreaseinadoptionistherefore

attributed to the increased use of certain methods,

specifically trenches and water traps (counted as one),

mulch, and stabilizing grasses. Figure 16 illustrates theimprovementsincebaselineonthesevariousmethods.

Soil erosion results in the loss of top-soil, leaving roots

exposed and resulting in a loss of soil fertility, ultimately

impacting yield. Coffee Farm College taught farmers the

importanceoferosioncontrol insoilmanagement,evenon

fields without slopes, where wind erosion can take place.

Farmers learned various techniques to reduce soil erosion

including planting stabilizing grasses, applyingmulch, and

terracingonsteepslopes.

Afterreceivingtraining,thenumberoffarmersnotusingany

form of erosion control decreased by 19%. The preferred

methodforfarmersinSembabuleDistrictweretrenchesand

watertraps,followedbymulch.

Figure16.ErosionControlMethodsSeen–Baselinevs.Endline

62%

25%16%

1% 1%

43% 45%

19%

4% 0%

No erosion control seen Trenches and water traps Mulch Stabilizing grasses TerracesPerc

enta

ge o

f hou

seho

lds

Erosion Control Methods Seen: Baseline and EndlineAll households (n = 581); multiple select option

Baseline(2018)

Endline(2020)

SOILEROSIONCONTROL

Picture7.Watertrap

25

Adoption of shade management best

practiceincreasedfrom51%atbaseline

to 67% at endline. At baseline, 23% of

householdshadashade levelof20%or

more,atendline,thisincreasedto48%of

households.

Shadeiscommoninmixedcoffeefarming

systems, such as those in Uganda, with

high numbers of fruit trees, such as

bananas in the coffee fields, resulting in

relativelyhighadoptionoftheshadebest

practice. Almost half (48%) of the

households had good shade levels at

endline (>20% shade) and 40% of

households had planted shade trees in

thepast threeyears,whichwill provide

shadeintime.

ShadeisimportantforRobustacoffee,whichevolvedinthetropicalforestsofUgandaandCongo.Ideally,shadelevelsshouldbebetween20-40%.Below20%,coffeetreeswillbesubjectedtohigh

temperaturesandhighlevelsofmoisturelossfromboththeleavesandthesoil.However,very

high levelsofshadecanresult inreduced floweringandthereforeyields.CoffeeFarmCollege

taughtfarmersabouttheimportanceofshademanagement,theidealshadetreevarieties,and

plantingtechniques.Adoptionofshademanagementrequiresthatmorethan20%ofthefarmis

coveredbyshadeorthatshadetreeshavebeenplantedinthelastthreeyears,eveniftheyare

notcurrentlyprovidingtherequiredshadelevels.

Figure17illustratesthemostcommonshadetreesinthecoffeefieldswerefruittrees:69%offarmshadbananas, significant increasefrom47%atbaseline,and22%of farmshadplanted

otherfruittreessuchasmangooravocado.Althoughtheselargefruittreesgivehighshadelevels,

and have other socio-economic benefits such as enhancing household diet and providing

additional income, the farm can only accommodate a limited number of them. Only 6% of

households had indigenous trees in the farm,while just 1%had planted exotic varieties like

grevillea,whicharegenerallyconsideredunsuitable.

Figure17.ShadeTreeTypes–Baselinevs.Endline

2%

47%

32%

13%1%

69%

22%

6%

Exotic trees(e.g., Grevillea)

Banana Other fruit trees(e.g., mango, avocado)

Indigenous trees(e.g., Albizia)

Perc

enta

ge o

f hou

seho

lds

Shade Tree Types: Baseline and EndlineAll households (n = 581)

Baseline(2018)

Endline(2020)

SHADEMANAGEMENT

Picture8.Bananaplantsinacoffeefarm

26

Figure18illustratestheshadelevelsfoundoncoffeefarmsinSembabuleDistrictatbaselineandendline. It shows there has been an increase in shade levels since baseline, with 49% of

householdshavingatleasta20%shadelevel,andfewerhouseholdshavingnoshade.

Figure18.ShadeLevel–Baselinevs.Endline

Adoptionoftherecordkeepingbestpractice

increasedslightlyfrom2%atbaselineto4%

atendline.TheCoffeeFarmCollegeprogram

provided a three-year pictorial record card

(as illustrated inPicture9) at the February

2019trainingandtrainedfarmersonhowto

keep financial records and calculate profit.

However,duringdatacollection,only5%of

thehouseholdspresentedarecordbookand

only4%hadwrittenrecordsonit.

Duringdatacollection,enumeratorsreported

that farmers were often times reluctant to

show their record cards as they had not

started using them. Additionally, some data

collectorsmightnothavewaitedlongenough

forfarmerstobringtheirrecordcardsduring

the interview. This, combinedwith the fact

that 68% of the households attended the

February2019training,ledLateriteandTechnoServetofollowuponhouseholdsthatattended

anyCoffeeFarmCollegetrainingtounderstandwhethertheyhadreceivedarecordbook.

Theteammanagedtoreach420of500householdstargetedbyphone,andofthese,62%stated

tohavereceivedarecordbookfromTechnoServe,bringingtherateofrecordbookownershipto

50%forallhouseholds.TechnoServewasnotabletofollowupinpersonwiththe163households

whostatedtheyhadkeptfarmrecordsintheirbooks.Thesehouseholdsarenotincludedinthebestpracticeadoptioncalculationinthisreport,whichmaythereforeunderestimatetheadoption

ofrecordkeeping.

Atendline67%ofhouseholdsknewhowtocalculatefarmprofitandloss,comparedtoonly32%

atbaseline.

RECORDKEEPING

Picture9.CoffeeFarmCollegerecordcardandfarmer'srecordbook.

27

Thefollowingpracticesareimportantagronomicpracticesforcoffeefarming;however,theydo

notapplytoallcoffeehouseholdssurveyed.Therefore,theyarenotconsideredkeybestpractices,

butwerearestillmonitoredduringtheendlinesurvey.

87% of households met the criteria for coffee

drying;however,atthetimeoftheendlinesurvey,

only 55 households (9%) were actively drying

coffee. Of these, the vast majority were drying

coffeecorrectlyontarpaulinoraplasticsheeton

theground.Eightofthesehouseholdsweredrying

the coffee onbare ground,which could result in

contamination from the soil and isn’t

recommended..

Proper coffee drying is crucial to produce

improvedqualitycoffeeandminimizetheriskof

contamination. As harvesting season had nearlyfinished when the survey was conducted,

households in Sembabule self-reported their

coffee-dryingpractices.Correctcoffeedryingwas

assessed based on whether (i) the farmer is

currentlydryingcoffeeonaplastictarpaulinoron

an elevated bed or (ii) the farmer is not drying

coffee,butplanstodosoinatarpaulinorontables.

Ofthe526householdsnotcurrentlydrying,88%

plantodryonatarpaulinorplastic,and3%plan

to dry on the bare ground. Coffee Farm College

promotedtheuseofplastictarpaulinsorraisedbedsfordryingcoffeetopreventcontamination

from drying the cherry on the bare ground. The training also stressed the importance of

harvesting red cherry, turningduringdrying, andusinggood storage techniques to avoid the

developmentofmouldorfurthercontamination.

Goodcoffeeseedlingsandcorrectplantingareessentialforahealthycoffeetreeandfuturehigh

production. The Ugandan Government’s Operation Wealth Creation (OWC) program, in

collaborationwiththeUgandaCoffeeDevelopmentAuthority(UCDA),hasprovidedseedlingsto

farmerssince2015.Duringthebaselinesurvey,41%ofcoffeehouseholdsreportedplantingnew

coffeeseedlings in thepast12months.Atendline,58%ofhouseholds reportedplantingnew

seedlings.

OTHERAGRONOMYPRACTICES

COFFEEDRYING

COFFEEPLANTING

COFFEEDRYING

Picture10.Coffeedryingonatarpaulin

28

AsillustratedinFigure19,therearenumerouselements to consider for proper coffee

planting: (i) the variety of coffee is elite or

clonalwithCoffeeWiltDiseaseresistance;(ii)

the new seedlings come from certified

nurseries (UCDA or Government); (iii) the

farmerappliedanapprovednutritionproduct;

(iv) seedlings look healthy; (v) the seedlings

wereplantedinrowsthreemetersapart;and

(vi) for seedlings taller than60cm, the stems

are bent to initiate sucker production. If

suckers are taller than 30cm, the best four

suckershavebeenselected.

While households improved on four of the

elements above, they regressed in two

elements.Inparticular,thecriteriaonsourcing

seedlings from certified nurseries saw the

sharpestdecreasefrom78%atbaselinetojust

49% at endline. At endline, farmers opted to

source seeds from non-certified sources,

including 16% who grew the seedling

themselves,and9%whousedseedlingsgrown

by neighbours. In addition, only 39% of

householdsappliedanapprovedfertilizertotheseedlings.Additionally,therewasasignificant

increaseintheuseofEliteorCWDresistantvarietiescomparedtobaseline,whichisapositive

step.

Figure19.ElementsofProperCoffeePlanting–Baselinevs.Endline

Picture11.Ayoungcoffeetree

29

Intercroppinghastwokeybenefits:diversificationofincomeandusageasacovercropduring

therainstoreducetheincidenceofsoilerosion.WiththewiderspacingassociatedwithRobusta

coffee,itispossibletointercropbetweentherowsofmaturecoffeetrees.

The intercropping is practiced correctly if farmers are intercropping their coffee trees with

legumes,suchasbeans,orfruittrees,suchasbananas.Farmersshouldnotusemaize,cassava,

creepers–suchaspumpkin,rootsandtubers,orgroundnuts–orothercropsonthecoffeefield,

whichareeithergreedycrops(maize)orwouldrequirethefieldtobedugforharvest.Thesurvey

observedallcropsusedandnotthemaincropsointerpretationofthefindingsisdifficult, for

exampleonemaizeplantinafieldofbeanswouldbeacceptable.

Figure20illustratesthecropsseenasanintercrop.Therewasanincreasedobservationofmaizeasanintercrop,seenat44%offarmsatendlinecomparedtojust11%atbaseline.Intercropping

usingsuitablecropslikebananaandlegumesalsoincreasedsignificantlyandtheuseofcassava

(another unsuitable crop) decreased between baseline and endline. For the small percentage

(4%)ofcoffeefarminghouseholdsinSembabulenotgrowinganyintercrops,thispracticewas

consideredas“notapplicable”.

Figure20.Inter-CroppingVarieties–Baselinevs.Endline

INTERCROPPING

30

The following section covers demographic, socio-economic, and financial information on the

entireendlinesampleof581households.

HouseholdComposition:Householdshave,onaverage,sixmembers,fourofwhicharechildren(0to18years).Bothfemaleandmalecoffeefarmersarelikelytobeintheirforties,withamean

ageof42 forwomenand48 formen.Of themale farmers,9%are less than30yearsof age,

whereas21%offemalefarmersarelessthan30andcanbeconsidered“youth”.

Figure21.DistributionofHouseholdSizeatEndline

Figure22.DistributionofNumberofChildrenatEndline

HOUSEHOLDCHARACTERISTICS

HOUSEHOLDCOMPOSITIONANDFARMERAGE

31

Figure23.MedianandMeanAgeDistributionbyGender

Overall,thelevelofeducationismoderate,with84%ofthewomenand90%ofthemenhaving

receivedsomelevelofformaleducation.Menareslightlybettereducated,as22%ofthemale

farmers have education above the primary level (either Secondary or above Secondary),

comparedtoonly19%forthewomen.

Figure24.LevelofEducationbyGender

Forchildrenaged6–14,99%ofbothboysandgirlswereenrolledinformaleducationatendline.

Thisisaslightincreasefromthebaselineforboys,where95%wereenrolled.Forchildrenaged

15–18,therehasbeenasignificantincreaseinschoolenrolmentsincethebaselineforbothboys

and girls. At baseline, just 76% of girls 15–18 were enrolled in school. At endline this rate

increasedto92%.Similarly,forboys,atbaselineenrolmentlevelsfor15–18yearoldsweremuch

lower(70%)comparedtoendline(83%).

EDUCATION

32

Figure25.ChildEducationbyGenderandAge

Aquarter(25%)ofthemalefarmerseitherhavehademploymentinthepastorarecurrently

employedoutsidetheircoffeefarmscomparedtojust16%ofthewomen.Thereisasignificant

increaseinmenwhoarecurrentlyemployedatendline(9%),comparedtobaseline(3%).

Figure26.EmploymentStatusbyGenderatEndline

Atendline,overhalf(59%)ofthefarmershadhiredpaidlabourtoworkonthecoffeefarminthe

past 12months. The average reportedwage for an eight-hourwork daywas 8,525Ugandan

Shillings,equivalentto$2.30USD.Thisisanincreasefrombaseline,wheretheaveragewagewas

reportedtobe7,887UgandanShillings,equivalentto$2.00USD.Thenationalminimumwagein

Ugandais130,000UgandanShillingspermonth,approximately5,909UgandanShillingsperday

;therefore,thewagesearnedonthecoffeefarmsareslightlyabovethemonthlyminimumwage.

Ofthe476householdswithschool-agedchildren(6-14yearsold),75%reportthattheirchildren

workinthecoffeefields.Forthese476households,4%ofhouseholdsreportedthatfarmwork

sometimesinterferedwithaccesstoeducationfortheirownchildren,especiallyduringharvest.

Lessthan1%(4farms)ofhouseholdsreportedhiringlabourbelow14yearsofage.

EMPLOYMENTSTATUS–OUTSIDETHECOFFEEFARM

PAIDLABORANDCHILDLABOR

33

Farmersownaself-reportedmeanof2.1hectaresoftotalagriculturalland,0.9hectaresofwhich

isplantedwithcoffee,withanaverageof992coffeetrees.Toaccountforoutliersduetoself-

reporting,the1stand99thpercentilesofthestatisticshavebeentrimmed.Themedianfortotallandis1.4hectares,with0.8hectaresofcoffeeland,and750coffeetrees.Two-thirds(66%)of

coffeetreeswereplantedafter2000,and68%ofhouseholdsstatedthattheyplannedtoplant

morecoffeeoverthenext12months.

Figure27.DistributionofTotalFarmSizeinHectares

Figure28.DistributionofNumberofCoffeeTrees

LANDOWNERSHIP

34

Figure29.DistributionofYearofPlanting

Production data was collected in kilograms of cherry, kiboko (sun dried cherry), and FAQ

(parchment)producedinthe2020harvestseason.Thisdataisself-reported,andcanbeseenas

aguidelineonly,givenlimitedrecordkeepingandthetendencytoroundyieldstothenearest100

kg.AsreportedinFigure30,anegligibleproportionofthefarmers(1%)reportednotsellinganycoffeein2020,whilethemajoritysoldcherry(64%).One-third(34%)offarmerssoldFAQ,while

27%soldkiboko.Additionally,2%ofhouseholdspre-soldtheircoffeepriortoharvest.Compared

tobaseline,thepercentageofhouseholdssellingFAQhasnearlydoubled,from18%atbaseline

to34%atendline.FAQfetcheshigherpricesthantheotherforms,sothisincreaseisapositive

finding.

Figure30.TypeofCoffeeSold

Atendline,themeanself-reportedyieldpertreewas2.0kgcherryequivalentwithamedianyieldof1.6kgcherryequivalentpertree.Thisisupfrombaseline,wherethemeanself-reportedyieldpertreewas1.3kgcherryequivalent,withamedianof0.9kgcherryequivalent.

Giventhemedianfarmhas750coffeetreesatendline,anestimatedfarmproductionwouldbe

about1,200kgofcherryequivalent.Soldaskiboko(600kg),thiswouldearnthehouseholdabout

$360USD.Conductingthesamecalculationusingmeansinsteadofmedians,householdswouldearnabout$595USD.

COFFEEPRODUCTION

35

Assumptionsforyieldpertreecalculations:

• 1kgofkiboko=2kgofcherry

• 1kgofFAQ=4kgofcherry

• 1kgofkiboko=$0.60USD

• 1stand99thpercentilesweretrimmedtoaccountformistakesinself-reporteddata

Householdsown,onaverage,8ofthe23surveyedassets.Thisisanincreasefromanaverageof

6assetsatbaseline.Thetopfivelargestimprovementssincebaselinehavecomefromaccessto

electricity(+17percentagepointsfrombaseline),ownershipofpropershoes(+14),ownership

oftelevision(+11),ownershipofatleastonepig(+9),andaccesstoimprovedwatersource(+7).

Twotransportationassets,–bicyclesandmotorcycles–decreasedsincethebaselineby5and1

percentagepointsrespectively.However,thisispartiallyoffsetbyanincreasedownershipofcars

ortrucks(+2).

HouseQuality:Thebuildingqualityof the farmers’houses in theSembabuledistrict isgood.Almosteveryhousehold(99%)hasimprovedroofs(corrugatedironortinsheets),85%ofthe

houseshaveaccesstoelectricity,mostlyfromsolarpanels,whichisupfromjust68%atbaseline.

Morethanhalf(56%)ofthehouseshavecementsealedfloors.

WaterandSanitation:12%ofthefarmershaveaccesstoanimprovedtoilet3,upfromjust7%atbaseline.Only26%ofthehouseshaveanimprovedwatersource,4yetthisisanimprovementfromjust17%atbaseline.Childrenareresponsibleforcollectingwaterin66%ofhouseholdsand

the average time to complete the return journey to collectwaterwas 39minutes. For some

households,thistookaslongasthreehours.

Livestock:Mosthouseholdshavesomelivestock,with70%ofhouseholdsowningatleastonechicken,48%owningatleastonepig,19%owningatleastonecow,and35%owningatleastone

goatorsheep.

Communication:Ownershipof communicationdevices ishighamong farmers in Sembabule.Almost all households (95%) own at least onemobile phone,which ismost usually used for

calling(99%),mobilebanking(61%),sendingandreceivingmessages(51%).Mosthouseholds

(84%)alsoownradios,whileamuchlowerproportion(28%)ownatelevision.

Transport:Manycoffeehouseholdsowntheirownmeansoftransport,68%ofhouseholdsownabicycle,and39%ownamotorcycle,with4%owningacar.

3Improvedaccesstosanitationconsistsofventilatedimprovedpits,flushtoilets,orcompostingtoiletsthatarenotsharedwithotherhouseholds. 4Improvedwateraccessconsistsofhavingaccesstoaprotecteddugwellorspring,tubewell,watertank,orapublicorprivatetap.Ifthewatersourceisnotatthehouse,itshouldnotbemorethana30-minuteroundtrip.

ASSETOWNERSHIP

36

Figure31.HouseholdOwnershipofAssetsatEndline

37

TohelpTechnoServeimprovetheprogramandassessitsimpactonfarmhouseholds,questions

wereaskedaboutfarmerperceptionsoftheFarmCollege.

ContactModalities(Message,Call,Visit)Toassesstheusefulnessofthevariouswaysfarmersarecontactedandtrained,farmerswere

askedtoscoreeachmodality(messages,calls,visits)from0–10onusefulness,with10beingthe

mostusefuland0beingnotatalluseful.Figure32reportsthefindings,showingthatonaveragevisitswereseenasthemostusefulmodalitywithanaveragescoreof7.4,,followedbyphonecalls

(7.3),andthenmessages(6.8).Note thatallof thecallsconductedwereduringtheCOVID-19

lockdownperiod,whichmayimpactresponsesforperceivedusefulnessastheyweretheprimary

modalityusedinthatperiod.

Figure32.PerceivedUsefulnessofTNSContactModalities(Meanof0-10scale)

PerceptionsonYieldChangeFarmers at endlinewere asked to rate the change in coffee yield that they experienced since

joiningCoffeeFarmCollege.AsseeninFigure33,81%ofhouseholdsreportedexperiencinganincrease in yield since attending training, and 40% said this increasewas large. Only 8% of

households reported a decrease in yields since training, and 10% reported no change. Figure33.PerceptionofYieldChangesinceattendingTNStraining

PERCEPTIONS&IMPACT

38

PerceptionsonTimeSpentonCoffeeProductionDuringendline,farmerswereaskedtoratethechangeintimespentoncoffeeproductionsince

joiningCoffeeFarmCollege.AsseeninFigure34,84%ofhouseholdsreportedexperiencinganincreaseintimespent,with31%sayingthisincreasewashigh.Thisislikelyduetotheincreased

attentiontobestpracticeadoption,asmanywerenotadoptingbestpracticesatbaseline.Only

6%ofhouseholdsreportedadecreaseintimespentoncoffeesincethetraining.

Figure34.PerceptionofTimeSpentonCoffeesinceattendingTNStraining

PerceptionsonCoffeeIncomeFarmersatendlinewereaskedtoratethechangeinincomefromcoffeesincejoiningCoffeeFarm

College.AsseeninFigure35,57%ofhouseholdsreportedexperiencinganincreaseinincome,with 13% saying this increasewas high. 17% reported having lost income since the start of

training,and24%reportednochange.GiventheendlinewasconductedduringtheCOVID-19

pandemic,when incomeswere already low formany farmers, it is possible that the farmers

reportingalossofincomesincethetrainingsareareflectionofthepandemic,andnotnecessarily

theimpactoftheCoffeeFarmCollege.Nevertheless,itappearsthatthemajorityofparticipating

coffee farmers in Sembabule District perceive the program to have generated a positive

contributiontoincome,whichisaprimarygoaloftheCoffeeFarmCollege.

Figure35.PerceptionofIncomeChangefromCoffeesinceattendingTNStraining

39

In the endline survey, additional questions relating to intra-household decision-making and

gender relationswere added to help understand howwomen’s participation in the program

translates to intrahousehold relations (e.g., gender roles andpreferences,decision-makingon

incomeandsavings,decision-makingonfoodpurchases).

Threequestionsguidedtheresearch:

• Howarewomenabletotranslatetheirparticipationintheinterventionintotheirability

tonegotiateintrahouseholdrelationstomakedecisionsovertheuseofincome?

• Howcanweunderstandthedifferentrolesandpreferencesofmenandwomentoplan

interventionsthatmeettheobjectivesandaimsoftheprogram?

• Howarewereachingandmeetingtheneedsofdifferentwomenthroughtheprogram?

InordertounderstandhowwomencantranslateparticipationintheCoffeeFarmCollegeinto

theirabilitytonegotiateintrahouseholdrelations,questionsondecision-makingonincomeand

savingswereasked.

Figure36reportstheprimarydecision-makeronincomedecisionsforbothcoffeeincomeandnon-coffeeincome.Itshowsthatmakingdecisionsjointly(bothmanandwomanequally)isthe

most commonway households in Sembabule District make decisions on both coffee income

(29%)andnon-coffeeincome(33%).However,whenhouseholdsdonotmakedecisionsjointly,

both coffee income and non-coffee income decisions are primarily made by men. Men are

involvedatleastequallyin80%ofcoffeeincomedecisions,comparedtojust48%ofwomen.For

non-coffeeincome,womenplayamoreprimaryroleastheyareinvolvedatleastequallyin57%

ofhouseholds.However,menarestillthemajoritydecision-makersonnon-coffeeincome,with

77%ofhouseholdsincludingmenatleastequallyinthesedecisions.

Itisimportanttonotethat5%ofhouseholdsatendlinearesingle-adultfemalehouseholds,and

13% of households are single-adult male households, which are indicated below. In these

households,decisionsaremadebyonlywomenoronlymen,asitistheonlyoption.

Figure36.Decision-makeronCoffeeandNon-CoffeeIncome

THEMEOFINTEREST:GENDER

FINANCIALDECISIONS(INCOME&SAVINGS)

40

Figure 37 reports the primary decision-maker on decisions for both major and minorexpenditures. Major expenditures include expenses including: 1) house construction and

renovation;2)payingschoolfees;3)buyingabicycleormotorbike;4)goingonareligioustrip;

or5)purchasinglargeappliancesforthehouse.Minorexpendituresincludeexpensessuchas:1)

foodpurchasesfordailyconsumption;2)rural-urbantransport;3)buyinghouseholditemslike

soapandcosmetics,charcoal,andfirewood;or4)kerosenepurchases.

Men tend to make sole decisions regarding major expenditures of the household. The most

commonwayforcoffeehouseholdstomakeminorexpendituredecisionsisjointly(bothmanand

womanequally) (23%). Formajor expenditures,men are involved at least equally in80%of

households,comparedtojust45%ofwomen.Forminorexpenditures,however,womenplaya

leadingrole,aswomenare involved in thesedecisionsat leastequally in65%ofhouseholds,

comparedtojust57%formen.

Again,itisimportanttonotethat5%ofhouseholdsatendlinearesingle-adultfemalehouseholds,

and13%ofhouseholdsaresingle-adultmalehouseholds,whichare indicatedbelow. Inthese

households,decisionsaremadebyonlywomenoronlymen,asitistheonlyoption.

Figure37.Decision-makeronMajorandMinorExpenditures

Figure38 reports theprimarydecision-makeronsavingsdecisions,andshowsthat themostcommondecisionmakersonsavingsaremenbythemselves(33%).Jointdecision-making(both

manandwomanequally)wasalsocommon(28%).Overall,menareinvolvedinsavingsdecisions

atleastequallyin74%ofhouseholds,comparedtojust54%forwomen.

Again,itisimportanttonotethat5%ofhouseholdsatendlinearesingle-adultfemalehouseholds,

and13%ofhouseholdsaresingle-adultmalehouseholds.Inthesehouseholds,decisionswould

bemadebyonlywomenoronlymen,asitistheonlyoption.

41

Figure38.Decision-makeronSavings

Enumeratorswerethenaskedtoobservewhoansweredthequestiontogaugeintra-household

genderrelations.Figure39disaggregatestheinformationbasedonwhoansweredthequestionandshowsastrongbiastowardsthegenderofthepersonresponsiblefordecision-making.For

example, in the 116 households where women are the sole decision-makers on savings, the

woman answered the question in 92%of the cases, suggesting thatwomen are speaking for

themselveswhenitrelatestosomethingtheyareresponsiblefor.Asimilarpatternwasobserved

with male participants, where the vast majority of men answered when they were solely

responsible fordecisionsonsavings.For jointdecisionhouseholds(eitherequallyor inpart),

bothmenandwomenparticipatedinansweringthequestion,butmenmoresothanwomen.

Figure39.Decision-makeronSavings,bywhoansweredthequestion

42

Inordertofurtherunderstandtherolesofmenandwomeninthehousehold,questionsregarding

foodpurchasedecisionsandcropproductionparticipationwerealsoasked.Figure40reportstheprimarydecision-makerforbothfoodpurchaseandcropproductionparticipationdecisions.

Forfoodpurchasedecisions,itismostcommonformentomakethesedecisionsbythemselves(30%).Menareinvolvedatleastequallyinfoodpurchasedecision-makingin69%ofhouseholds,

comparedtojust48%forwomen.

Forcropproductionparticipationdecisions,itismostcommonforhouseholdstomakethesedecisionsjointlyandequally(39%).Womenareinvolvedinthesedecisionsatleastequallyin

79%ofhouseholdsatendline,comparedto60%formen.

Figure40.Decision-makeronFoodPurchaseandCropProductionParticipation

Again,afteraskingthesequestionsonfoodproductiondecisions,theenumeratorswereaskedto

observe who answered the questions to gauge intra-household gender relations. Figure 41disaggregatestheinformationbasedonwhoansweredthequestionandagainshowsastrong

biastowardsthegenderofthepersonresponsiblefordecision-making.Forexample,inthe177

householdswheremenaretheonlydecision-makersonfoodpurchases,themanansweredthe

questionin70%ofthecases.Asimilarandevenmorepronouncedpatterninseenwithwomen,

wherethevastmajorityofwomenansweredwhentheyweresolelyresponsiblefordecisionson

foodpurchases.Forjointdecisionhouseholds(eitherequallyorinpart),bothmenandwomen

participatedinansweringthequestion,butmenmoresothanwomen.

FOODDECISIONS

43

Figure41.Decision-makeronFoodPurchaseDecisions,bywhoanswered

AsimilaranalysiswasconductedoncropproductionparticipationdecisionsinFigure42.Again,there isastrongbias towardsthegenderof thepersonresponsible fordecision-making,with

womenandmenspeaking for themselveswhentheyareresponsible,andanswering together

whenthereisjointresponsibility.

Figure42.Decision-makeroncropproductionparticipation,bywhoanswered

44

Figure43 providesa summaryof all of thedecisionsaskedabout in theendline survey, andhighlightsthemostcommonanswer.Itshowsthatjoint-decisionmakingisverycommonwith

menmakingslightlymoredecisionsthanwomenforsomedecisioncategories.However,there

arestillasignificantpercentageofhouseholdswherewomenare leadingthedecision-making

process.

Figure43.Summaryofdecision-makingacrosssevendecisions

Tofurtherassessoveralldecision-makingacrossalltypesofdecisions(income,savings,food),

Figure44analyzesthehouseholdswheremen,women,orbothareinvolvedatleastequallyinall decisions. In 41% of householdsmen are involved at least equally in all decisionswhere

questionswereasked(income,savings, food). In29%ofhouseholdswomenwere involvedat

leastequallyinalldecisions.In32%ofhouseholds,bothmenandwomenwereinvolvedatleast

insomepartinalldecisions.Thesefindingsreinforcethatmenareoveralltheprimarydecision-

makersinSembabuleDistrict,butthatjointdecision-makingandfemale-decisionhouseholdsare

verycommonaswell.

Figure44.Decision-makingbygenderacrossalldecisions(income,savings,food)

Thesefindingshighlightthatwomenandmenareresponsiblefordifferentdecisionswithinthe

household,andthatmanydecisionsaremadejointly.Bothwomenandmenfeelfreetoanswer

thequestionsabouttheirdecision-makingwhentheyareresponsible,andallowtheirpartnerto

answerwhen theyareresponsible.For jointdecision-makinghouseholds,men tend tobe the

primaryquestionrespondent,butwomenalsocontributeaconsiderableamount.

ALLDECISIONS

45

Inordertofurtherassessrolesandresponsibilitiesonthefarm,respondentswereaskedwhois

most responsible for thecoffee farm. Inaddition,at theendof thesurvey,enumeratorswere

asked to recordwhich respondent (man orwoman) answered themajority of the questions

throughoutthesurvey.Figure45 illustratesthatmenareprimarilyresponsibleforthecoffeefarm(44%)andwerealsotheprimaryrespondenttothemajorityofsurveyquestions(53%).

Figure45.CoffeeFarmResponsibilityandQuestionAnsweringatEndline

Toassessthisfurther,Figure46depictswhoansweredthemajorityofthequestionsthroughouttheentireendlinesurvey,dependingonwhoisresponsibleforthefarm.Similartothedecision-

makinganalysis,astrongbiastowardsthegenderofthemostresponsiblefarmerwasfound.That

is,whenwomenweremostresponsible,womenoverwhelmingly(93%)answeredthequestion.

Asimilarpatternwasfoundformen.Thisfurthersuggeststhatwomenareempoweredenough

to:a)beresponsibleforthecoffeefarm;b)answerquestionsontheirresponsibility.Italsoshows

thatmenandwomenallowtheirpartnerstoanswerfortheirownwork,asignofcooperative

intra-householdgenderrelations.

Figure46.CoffeeFarmResponsibility,bywhoansweredthequestions

COFFEEFARMRESPONSIBILITY

46

Toassessgenderroleswithrespecttocoffeeproduction,questionswereaskedonthegender

responsibilityforfourcoffeefarmingpractices:pruning,harvesting,weeding,andplanting.

Figure47reportsthefindings,andshowsthatmenandwomenhavedifferentresponsibilitieson the farm.Men are highly involved in all four practices, but are least involved in planting.

Womenareprimarilyinvolvedinharvestingandweeding,andmuchlessinvolvedinpruningand

planting.Harvestingandweeding,beinglesstechnicalskills,arealsotheactivitieswhereother

householdmembers(oftenchildrenaged6–14)aremostinvolved,andalsotheactivitieswhere

hiredlabourisused.

Figure47.GenderresponsibilityforPruning,Weeding,Planting,andHarvesting

Tofurtherassesstheintensityoflabourrequiredfromeachgender,Figure48reportsthemeannumberactivitieswhereeachmemberisinvolved,andshowsthatmenareinvolvedin2.8ofthe

4activitiesonaverage,comparedto2.2forwomen.However,thismetricdoesnotcapturethe

timespentoneachactivity,whichwouldbeatruertestofintensityoflabour.

Figure48.MeanNumberofResponsibilitiesbyGender

INTRA-HOUSEHOLDLABOURDISTRIBUTION

47

This section is dedicated to understanding the income profile of households. Examining the

financial profile of a household is useful to understand the wealth and vulnerability of the

household.

A major program goal of Farm College is to increase incomes and reduce poverty through

increasedcoffeeproductionandsales.Figure49showsthedependenceofhouseholdsoncoffeeforincomebasedonself-reporteddata.

Figure49.PortionofTotalHouseholdIncomefromCoffee

In 2020, coffee households were highly dependent on coffee, with 60% of the households

reportingthatmorethanhalfoftheirtotalhouseholdincomecomesfromcoffee.Thisisashift

frombaseline,where themajority of households (64%) recorded coffee income representing

between 25% and 50% of income. The COVID-19 pandemic may be a factor in this shift as

households may have lost other sources of income and are more dependent on coffee. As

reportedinFigure50,othermajoreconomicactivitiesintheareaare:1)thesaleofmaize,beans,andbananas(57%)2)runningpersonalbusinesses(16%);3)thesaleoflivestock(8%).

Figure50.OtherIncomeSourcesatEndline

FINANCIALPROFILE

INCOMEDISTRIBUTION

48

For any household, savings are an important shield against financial shocks and coffee price

volatility.14%ofthehouseholdsreportedthattheydonotuseanymethodofsavings.Thisis

downdramaticallyfrom51%atbaseline,whichappearstobedrivenbyariseinmembershipto

informalsavingsgroups.

Farmerswereaskedabouttheirmainsavingmethods,whichisreportedinFigure51.47%ofthe farmersmentioned informalsavinggroupsas theirmainsavingsmethod,updramatically

fromjust18%atbaseline.13%offarmersuseformalsavingsgroups(e.g.,SACCOs),andonly7%

ofthehouseholdsuseformalsavingbankaccountsastheirsavingmethod.Thesameproportion

of households (7%) report fattening of livestock as their main saving mechanism. A lower

proportionoffarmers(5%)useamobilemoneyaccountforsavings.

Figure51.SavingMethodsUsedatEndline

SAVINGS

49

This section explores the vulnerability of households by studying the financial shocks and

incidenceoffoodshortagesfacedbyfarmersinSembabule.

Ashockorcrisisisdefinedasanyeventthatleadstoaseriousreductioninthefarmer’sasset

holdings,causingthehouseholdincometofallsubstantiallyorresultinasignificantreductionin

food consumption. A majority of the households at endline (70%) reported they have been

affectedbyatleastoneseriousfinancialshockorcrisisduringthelastyear.Atbaseline,thisfigure

was80%.

Householdsintheendlinesurveyreportedanaverageof1.4financialshocksinthelastyear,with

themostcommonbeingaserioushealthproblemordeath(51%),drought(30%),andlowcoffee

yields(27%).Figure52reportstheremainingmostcommonshocks.Figure52.TopFinancialShocksFacedbyHouseholds

Foodshortagescanbedebilitatingforhouseholds,adverselyaffectingtheirhealthandabilityto

workontheirfarms.Atendline,235households(40%)reportedfacingfoodshortagesoverthe

lastyear,eachforanaverageofonemonth.Thisisdownconsiderablyfrombaseline,when65%

experiencedfoodshortagesforanaverageoftwomonths.Figure53belowdepictsthemonthsof the year in which food shortages were most common, and shows August to October

experiencing the most food shortages. The peak came in September, when 65% of the 235

householdswhoexperiencedshortage,wenthungry.

VULNERABILITYPROFILE

FINANCIALSHOCKS

FOODSHORTAGES

50

Figure53.Mainmonthsoffoodshortagefor235householdsfacingfoodshortage

In order to combine data from a large number of individual food groups, a single indexwas

constructedusingamethodpromotedbytheFoodandAgricultureOrganizationoftheUnited

Nations(FAO2011).Thehouseholddietarydiversityscore(HDDS)reflects, inasnapshot,the

economicabilityofahouseholdtoaccessavarietyoffoods.Studieshaveshownthatanincrease

indietarydiversityisassociatedwithsocio-economicstatusandhouseholdfoodsecurity.

TheHDDStakes integervaluesbetween0and12; thehigherthescore, thebetter thedietary

diversity. The average household in the endline survey has a HDDS score of 8, indicating a

moderatedietarydiversityformostfarminghouseholds.

Themostfrequentfoodgroupsconsumedbysampledfarmerswerepulsesorlegumessuchas

beansorpeas(97%),fruits(93%),rootsandtubers(92%),andspicesorcondiments(99%).A

largemajorityofthehouseholdsalsoconsumedvegetables(86%),oilsorfats(82%),sugaror

honey(73%)andmilkproducts(73%).Consumptionofcerealsfellfrom77%atbaselinetojust

60% at endline. A lower percent of households (13%) reported having consumed meat

(predominantlybeef), fish (24%),andeggs (15%). Matooke,a stapledish inmostofUganda

consistingofmashedstarchybanana,wasoneofthemostwidelyconsumedfooditemsatendline,

withmorethanhalf(77%)ofthehouseholdsreportinghavingeatenthisdishduringthepast24

hours.Otherpopulardishesincludeboiledormashedcassava(38%)andugali(33%).

HOUSEHOLDDIET

51

Figure54.MainTypesofFoodConsumed

This sectionprovidesabriefoverviewof twopovertyassessment tools– theProgressoutof

Poverty Index (PPI) and the Multi-Dimensional Poverty Index (MPI). The purpose of these

povertyassessmenttoolsistounderstandtherelativepovertyprofileofthefarmersinvolvedin

theFarmCollegeprogram.

ThePPIisapovertymeasurementtoolcomprisedofacountry-specificsurveywith10simple,

easy-to-answermultiple-choicequestions.5ItwasdevelopedbyGrameenBankandmanagedbyInnovationsforPovertyAction(IPA).ItisprimarilyusedbyNGOs,socialenterprises,andafew

foundations and has been customized for 61 developing countries to date, including Uganda

(Schreiner2011).

Thetotalscorethatahouseholdcanachieverangesbetween0(mostlikelytobebelowapoverty

line)and100(leastlikelytobebelowapovertyline).ThePPIscoringcanbeusedtoestimate

threebasicquantities.First,thepovertylikelihoodofahousehold,whichistheprobabilitythata

householdhasper-adultorequivalentper-capitaconsumptionbelowapovertyline.Second,the

PPI scorecanestimate thepoverty rateofagroupofhouseholdsatapoint in timeusing the

averagescoresofeachhousehold.Third, it canestimate thechanges in thepovertyratesofa

groupofhouseholdsbetweentwopointsintime.

TheSimplePovertyScorecard(Schreiner2011),usedtomeasurethepovertylikelihoodof

farmersinCentralUganda,isbasedonindicatorsrelatedtothehouseholdcomposition(suchas

thenumberofhouseholdmembers),education(suchasthelevelofeducationofa

woman/wife),housingquality(suchasthewallandroofmaterialsofthehouse),andthe

ownershipofdurableassets(suchasmobilephoneorshoeownership).Foradetailed

descriptionofthequestionsusedtobuildthepovertyprofileofFarmCollegecoffeefarmers,

pleasereferto

Appendix6.

5FormoreinformationonthePovertyProbabilityIndex,visithttps://www.povertyindex.org/about-us.

POVERTYANALYSIS

PROGRESSOUTOFPOVERTYINDEX

52

TheresultsofthissectioncomparethePPP-basedpovertyrateofcoffeefarmersinSembabuleto

theUgandaNationalHouseholdSurvey2016((UBOS)2018),anationallyrepresentativesurvey

covering7,500householdscountry-wide.

Figure55belowshowstheresultsofthePPIanalysisforSembabuleDistrictfrombothbaselineandendline.Theregionalvaluesdonotchangebetweenbaselineandendlineasthesamesurvey

isused.ThePPIvalueforaregionrepresentsthepercentageofagivenpopulationthatlivesbelow

thepovertylineindicated.

Figure55.PPI-BasedPovertyRate

Asillustratedabove,PPI-basedpovertyhas fallensincebaseline inallcategoriescomparedto

nationalandregionalaverages.Thatis,thehouseholdsbelongingtothe2018cohortoftheFarm

Collegeareinabetterpositionintermsofpovertylevels,whenusingPPIcalculations.Only17%

ofthehouseholdsatendlinearelikelytofallbelowthe$1.90/day2011PPPpovertyline,which

isdownfrom26%atbaseline,andnowlessthanCentralRegionaverages(22%).Atthenational

povertylinelevel,only1%ofSembabulehouseholdssurveyedfallbelowthepovertyline.

The globalMPI is an internationalmeasure of poverty that complements traditional income-

basedpovertymeasuresbycapturing thedeprivations thateachperson faceswithrespect to

education,health,andlivingstandards. ItwasdevelopedandmanagedbytheOxfordPoverty&HumanDevelopmentInitiativeandsupportedbyUNDP.ItisprimarilyusedbyUNDP(Human

DevelopmentReport),WorldBank,andotherlargeinstitutions,andcoversover100countries.6

Understandingthemultiplefacetsofpovertylikelackofeducation,healthdeprivation,andpoor

standardsoflivingiscriticaltotrulyassessingtherelativepovertyofthebeneficiaries.TheMPI

tool provides a multi-dimensional perspective of the relative poverty of farmers. Note that

incidence refers to percentage of people that are multi-dimensionally poor/deprived, andintensity refers to proportion of indicators in which these poor people are deprived. The

6Formoreinformation,visit:http://hdr.undp.org/en/content/multidimensional-poverty-index-mpi

MULTI-DIMENSIONALPOVERTYINDEX

53

TechnoServeMPIwasadapted7 to theendline surveyandcompares thedataof the surveyedhouseholds to that of the Uganda 2016 Demographic and Health Survey (Uganda Bureau of

Statistics,2018).

TheresultsinFigure56showthat,onaverage,thecoffeehouseholdsinthe2018cohortofCoffeeFarmCollegeareworseoffthanthenationalandregionalaverages.66%ofthehouseholdsare

consideredmulti-dimensionallypoorcomparedto55%ofthehouseholdsnation-wide,and31%

of the households in the Central Region of Uganda. However, this is slightly improved since

baseline,where69%ofhouseholdswereconsideredpoor.

Figure56.MultidimensionalPoverty–IncidenceandIntensity

Figure 57 shows the contribution of the different indicators to overall poverty. The maincontributorforthefarmersinthe2018CohortoftheFarmCollegeisthenutritioncomponent,as

more than a quarter of the households (28%) have an HDDS score below or equal to 8, the

thresholdfordeprivationinnutrition.

7ThemainchangeistheuseofaHouseholdDietaryDiversityIndex(HDDS)insteadofusingBodyMassIndex(BMI)basedonscoresfor calculating nutritional deprivation. Numerous studies using cross-sectional data for sub-Saharan African and South Asiancountriesshowadirectlinkbetweendietarydiversityandthenutritionaladequacyofadiet(HoddinottandYohannes,2002;Ogle,2001;Bhargava,2015;Hatloyetal.,1998).Also,ArimondandRuel(2004)showthatdietarydiversitydoespredictheight-for-weightz-scores(HAZ),weight-for-age(WAZ)z-scores,andundernutrition–theverymeasuresusedintheglobalMPI.WhilethisMPImaynotbecompletelycomparablewiththeglobalone,theHDDSweemployedprovidessimilarinformationatmuchlowerdatacollectioncost,henceitsuseisjustified.

54

Figure57.ContributionofeachIndicatortoOverallPoverty

Appendix1.CharacteristicsofAdoptingHouseholds

Acomparisonofthecharacteristicsof(i)householdsadoptingfourormorebestpractices–High

Adopters,tothose(ii)householdswhohaveimplementedthreeorfewerbestpractices–Low

Adopters,revealsthatthetwogroupsareslightlydifferent.Differencesarestatisticallysignificant

for women’s age, women’s education, men’s literacy rates, land planted with coffee, asset

ownership,andfinancialshocksfaced.Thatis,highadoptionhouseholdsarestatisticallydifferent

thanlowadoptionhouseholdsinthefollowingways:

1%significancedifferences• Womeninhighadoptionhouseholdsareyoungeronaverage(age40)comparedtolow

adoptionhouseholds(age44).

• Women inhighadoptionhouseholdshavemore formaleducation(88%),comparedtolowadoptionhouseholds(82%).

• Menhavehigherratesofliteracyinhighadoptionhouseholds(91%),comparedtolowadoptionhouseholds(82%).

• Low adoption households have less land planted with coffee trees (0.8 hectares)comparedtohighadoptionhouseholds(1.2hectares)

• Highadoptionhouseholdsownmoreassets(9)comparedtolowadoptionhouseholds(8)

5%significancedifferences• Highadoptionhouseholdsweremorelikelytohavefacedafinancialshockinthelastyear.

Table3.AverageCharacteristicsofLowandHighBestPracticeAdoptingHHsStatisticallysignificantdifferencesarehighlightedingreen.

Low Adoption Households

High Adoption Households

Adopting less than 4 BPs

Adopting 4 or more BPs

No. of Households 384 197 Number of household members 6 7 Number of children 4 4

APPENDICES

55

Age of woman*** 44 40 Age of man 48 47 Formal education rate – women*** 82% 88% Formal education rate – men 89% 92% Literacy rate – women 75% 80% Literacy rate – men*** 82% 91% Total agricultural land (hectares) 2.0 2.2 Land planted with coffee (hectares)*** 0.8 1.2 % Faced financial shocks in last year** 67% 77% % Faced food shortages in last year 38% 43% Asset Ownership (number of assets)*** 8 9 Coffee majority of income 59% 62% HH Dietary Diversity Score 8.0 8.2 Thestars(*)representthesignificancelevelofthedifferencebetweenthehighadoptionandlowadoptiongroups.***Significantatthe1%level,**Significantatthe5%level,*Significantatthe10%level.

Appendix2.BestPracticeAdoptionbyTrainingonEachSpecificTopic

Acomparisonofbestpracticeadoptionrateswasconductedforhouseholdsthatwereconsidered

trainedonaspecifictopic(sometimesthisrequiredattendanceatmultipletrainings)–AttendedTrainingonTopic–andthosewhowerenottopictrained–DidNotAttendTrainingonTopic.Noneofthedifferenceswerefoundtobestatisticallysignificant.

Table4.Bestpracticeadoptionbytrainingoneachspecifictopic Not Trained on

Specific Topic Trained on

Specific Topic

Overall Average Overall Average Shade 64% 66% Weeding 59% 65% IPDM 63% 63% Erosion Control 60% 59% Rejuvenation 47% 56% Pruning 44% 47% Coffee Nutrition 19% 17% Record Keeping 2% 5% Mulching 6% 4% Safe Use of Pesticides 0% 0% Average # of best practices adopted (out of 10) 4 BPs 4 BPs

Thestars(*)representthesignificancelevelofthedifferencebetweenthehighadoptionandlowadoptiongroups.***Significantatthe1%level,**Significantatthe5%level,*Significantatthe10%level.

56

Appendix3.BestPracticeAdoptionbyNon-AttendingHouseholds

Acomparisonofbestpracticeadoptionrateswasalsoconductedforthe79householdswhodid

notattendCoffeeFarmCollege.Table5reportstheseratesaswellastheadoptionratesandalsoreports the rates for Low Attending Households (less than 7 topics attended) and TrainedHouseholds(morethan7topicsattended).

Table5.Bestpracticeadoptionfornon-attendinghouseholdsatendline

Non-Attending Households

Low Attend. Households

Trained Households

Overall Average Overall Average Overall Average No. of Households 79 63 443 Shade 78% 60% 66% Weeding 62% 58% 65% IPDM 42% 58% 64% Erosion Control 44% 72% 57% Rejuvenation 61% 53% 54% Pruning 42% 50% 46% Coffee Nutrition 16% 16% 18% Record Keeping 2% 4% 5% Mulching 2% 5% 4% Safe Use of Pesticides 0% 0% 0% Average # of best practices adopted (out of 10)

3 BPs 4 BPs 4 BPs

Households adopting more than half (5) of the best practices

29% 26% 35%

Appendix4.BestPracticeAdoptionbyParishComparisonofhouseholdsfromthe15parishesshowsthatthereisavariationinbestpracticeadoptionbetweentheparishes–seeTable6.HighestAdoptingParishes

• Nsoga–Nsogaaveragedthemostamountofbestpracticesadopted(4.3),andthesecondmosthouseholdsadopting5+bestpractices(45%)• Mateete–38%offarmersinMateeteareadoptingmorethan5bestpractices,andhouseholdsaverage4.2bestpractices• Nakagongo–46%offarmersareadoptingmorethan5bestpractices(mostofallparishes);householdsaverage3.9bestpractices

LowAdoptingParishes

• Kasambya–only22%offarmersareadopting5ormorebestpractices,andtheyscorethelowestaveragebestpracticesatendline(3.4).• Lugusulu–farmersatendlineaveragedjust3.3bestpractices,thelowestofanyparish.

Table6.BestPracticeAdoption,byParishForeachbestpractice,thetwohighestadoptingparishesarehighlightedingreen,andthetwolowestadoptingparishishighlightedinred.(n=581)

Kabale Kasam- bya

Kasambya Lwebitakuli Kayunga Kido-

kolo Kinywa-

mazzi Lugus

ulu Lwebita-

kuli Mabindo Man- yama Mateete Mitete Naka-

gongo Naka- senyi Nsoga

No. of households Record Keeping 0% 6% 5% 7% 6% 0% 3% 6% 5% 8% 5% 3% 8% 5% 5% Safe Use of Pesticides 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% Weeding 72% 64% 54% 73% 64% 65% 60% 71% 67% 82% 58% 61% 59% 54% 61% IPDM 46% 47% 56% 56% 69% 51% 55% 65% 62% 53% 68% 74% 64% 74% 79% Rejuvenation 49% 58% 56% 68% 69% 43% 53% 53% 49% 45% 53% 45% 77% 59% 61% Pruning 41% 44% 38% 46% 61% 43% 35% 49% 49% 50% 30% 65% 62% 44% 58% Nutrition 8% 8% 10% 5% 6% 27% 20% 24% 15% 21% 40% 26% 13% 10% 16% Mulching 8% 3% 5% 5% 6% 8% 3% 0% 5% 3% 5% 0% 3% 0% 0% Erosion 51% 47% 59% 51% 56% 68% 40% 47% 64% 63% 73% 61% 54% 51% 76% Shade 82% 61% 64% 54% 50% 76% 63% 57% 74% 55% 88% 71% 54% 64% 71% Adopted 5+ BPs 21% 22% 21% 32% 42% 35% 30% 27% 41% 34% 38% 42% 46% 38% 45% Sum BPs (Endline) 3.6 3.4 3.5 3.7 3.9 3.8 3.3 3.7 3.9 3.8 4.2 4.1 3.9 3.6 4.3 Sum BPs (Baseline) 1.8 1.5 1.9 2.3 1.9 2.6 2.1 2.3 2.5 2.0 3.0 2.6 2.6 2.2 2.8

58

Appendix5.BestPracticeRuleChangessinceBaselineIn thebaselinereport,bestpracticeswerescoredusinganew,morecomplexrulemethodology,combining feedback fromHRNS.Atendline, thedecisionwasmadetorevertbacktothestandardTNSruleswhichareconsistentwithallothercountriesaswellastheprogramproposal.ThisendlinereportusesthesestandardTNSrules,andhasconvertedallbaselineresultstotheseusualrules.Table7displaystheoriginalbaselinerulesaswellastheusualTNSrulesusedatendline,anddescribesthedifferencebetweenthetwomethodologies.AdiscussionwillbeheldwiththewiderUCATteamontherulesinlightofthefindings.

Table7.RuleschangessinceBaselinetoreverttousualTNSrules

Original Baseline Rule Updated Rule at Endline (Usual TNS Rules) Comment

10 Best Practices Adopted if: Adopted if:

Record Keeping

1. Farmer has a Record Book AND 2. This has records of either coffee sold OR hired labour costs or other costs

1. Farmer has a Record Book AND 2. This has records of either coffee sold OR hired labour costs or other costs

No change. Already consistent with usual TNS rules.

Safe Use of Pesticides

N/A if farmer has not sprayed pesticide. If farmer has sprayed pesticide: 1. Farmer has been trained on how to safely use pesticide AND 2. Famer is not storing pesticide OR farmer stores pesticide outside the house and the store is locked AND 3. Farmer used masks googles and gloves at a minimum OR contractor used masks goggles and gloves items AND 4. Containers not washed and re-used for other purposes AND not thrown into the fields or compound

N/A if farmer has not sprayed pesticide. If farmer has sprayed pesticide: 1. Farmer has been trained on how to safely use pesticide AND 2. Famer is not storing pesticide OR farmer stores pesticide outside the house and the store is locked AND 3. Farmer used masks googles and gloves at a minimum OR contractor used masks goggles and gloves items AND 4. Containers not washed and re-used for other purposes AND not thrown into the fields or compound

No change. Already consistent with usual TNS rules.

59

AND not thrown into the rubbish pit AND not buried AND not thrown into pit latrine/toilet 5. No banned pesticides used

AND not thrown into the rubbish pit AND not buried AND not thrown into pit latrine/toilet 5. No banned pesticides used

Weeding

1. Farmer has pulled weeds by hand, used a slasher under the tree canopy, used a panga to weed under the tree canopy, dug under the tree canopy, or applied herbicides AND 1.1. Farmer has NOT exclusively dug under the tree canopy exclusively AND 2. Farmer weeds twice or more per year AND 3. There are few OR no weeds AND 4. If there are few weeds, they are less than 30cm tall AND 5. Weeds have not reached the flowering stage

1. Farmer has pulled weeds by hand, used a slasher under the tree canopy, used a panga to weed under the tree canopy, dug under the tree canopy, or applied herbicides AND 1.1. Farmer has NOT exclusively dug under the tree canopy exclusively AND 2. Farmer weeds twice or more per year AND 3. There are few OR no weeds AND 4. If there are few weeds, they are less than 30cm tall

Weed flowering criteria was removed in order to be consistent with usual TNS rules in other countries and cohorts.

IPDM

1. Farmer can name at least 4 out of 9 IPDM methods AND 2. Farmer mentions uprooting and burning infested trees for CWD AND 3. Farmer mentions removing and burning infested twigs for CTB

1. Farmer can name at least 3 out of 9 IPDM methods

Reverted to usual TNS rules which require 3 out of 9 methods, and do not specify conditions on specific methods.

Rejuvenation

1. Most trees have 4 main stems or less AND 2. The oldest trees are 8 years old or younger OR

1. Most trees have 4 main stems or less AND 2. The oldest main stems are 8 years or younger

Reverted to usual TNS rule for rejuvenation, which requires 4 or fewer main stems and main stems 8 years or younger.

60

1. At least one tree is older than 8 years AND 2. The oldest main stems are 8 years or younger AND 3. Most trees have 4 main stems or less AND 4. At least 25% of the farm has been rejuvenated AND 5. Suckers are less than 30cm tall OR Suckers are more than 30cm tall AND there are 4 or less suckers

Pruning

1. Trees have been pruned using 4 of the 5 methods listed

1. Trees have been pruned using 3 of the 4 methods listed

Reverted to usual TNS rule for pruning, which requires 3 methods to be seen. In addition, the responses choosing the option “Broken / unproductive stems and/or branches removed” were removed, as this option is not typically given in any other countries or cohorts. As such, the usual TNS rule of 3 of 4 methods is used.

Coffee Nutrition

1. Nearly all leaves are dark green AND 2. At least two of the following are used: compost or manure (counted as 1), NPK, Foliar Feeds, Lime and DAP. If foliar feed is used only count if this is zinc/boron based AND 3. IF fertilizer is applied to the soil, the fertilizer has been applied using a measure, and is not broadcast

1. Nearly all leaves are dark green AND 2. At least one of the following are used: compost or manure (counted as 1), NPK, Foliar Feeds, Lime and DAP. If foliar feed is used only count if this is zinc/boron based AND 3. IF fertilizer is applied to the soil, the fertilizer has been applied using a measure, and is not broadcast

Reverted to usual TNS rule for good soils (based in soil survey results since the baseline) where organic production is common, which requires one product be used, instead of two.

61

Mulching

1. Farmer has applied mulch AND 2. Mulch applied on up to 25% of the farm AND 3. Mulch is more than 2cm thick

1. Farmer has applied mulch AND 2. Mulch is more than 2cm thick

Reverted to usual TNS rules which do not include the 25% farm proportion criteria.

Erosion Control

1. At least two erosion control methods seen 1. At least one erosion control methods seen Reverted back to usual TNS rules which requires just one erosion control method seen. In addition, the options of trenches and water traps were combined into one single option, as trenches are not typically an option under TNS rules.

Shade

1. There is 20% shade or more OR 2. There is less than 20% shade but shade trees have been planted in the last 3 years

1. There is 20% shade or more OR 2. There is less than 20% shade but shade trees have been planted in the last 3 years

No change. Already consistent with usual TNS rules.

Other Agronomy Practices

Composting

Originally considered a BP No longer considered a BP Not usually considered a standalone BP as it is included under Coffee Nutrition. This is consistent with TNS rules in other countries and cohorts.

Intercropping

Originally considered a BP No longer considered a BP Not usually considered a standalone BP under usual TNS rules. Question was changed.

Coffee Planting Not considered a BP Not considered a BP No change Coffee Drying Not considered a BP Not considered a BP No change

Appendix6.MPIandPPIScorecardsFigure58.MPIScorecard

63

Figure59.PPIScorecard

64

(UBOS),UgandaBureauofStatistics.2018.“UgandaNationalHouseholdSurvey2016.”Kampala.FAO.2011.“GuidelinesforMeasuringHouseholdandIndividualDietaryDiversity.”

KenyaNationalBureauofStatistics.2015.KenyaDemographicandHealthSurvey2014.Rockville,MD,USA:KenyaNationalBureauofStatistics.

KenyaNationalBureauofStatistics.2013.KenyaIntegratedHouseholdBudgetSurvey2005-06.Nairobi:KenyaNationalBureauofStatistics.

Schreiner,Mark.2011.“SimplePovertyScorecard®Poverty-AssessmentToolKenya.”SimplePovertyScorecard,11March.

BIBLIOGRAPHY