Food	
  for	
  the	
  Hungry-­‐ Mozambique
P.	
  L.	
  480	
  TITLE	
  II
Multi-­‐Year	
  Assistance	
  Program
Final	
  Evaluation
Submitted	
  by	
  Marcus	
  L.	
  Catsam
On	
  behalf	
  of	
  Marcus	
  Catsam	
  Consulting	
  LLC
Final	
  Draft-­‐ July	
  9,	
  2013
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 1
Table	
  of	
  Contents
Abbreviations	
  and	
  Acronyms..........................................................................................................2
Executive	
  Summary………………………………………………………………………………………………………………………3
I.	
  Introduction……………………………………………….………………………………………………………………………………6
II.	
  Overview………………………………………………………………………………………………….………………..……………..6
A. Overview	
  of	
  Food	
  Security	
  Situation	
  in	
  Mozambique………………………………………..…………………..6
B. Overview	
  of	
  the	
  FH	
  MYAP	
  In	
  Cabo	
  Delgado…..………………………………………………..…………………….7
III.	
  Evaluation	
  Methodology…..………………………………………………………………………………………………………7
A. Sampling…………………………………………………………..…………………………………………………………………..7
B. Data	
  Collection	
  Instruments	
  and	
  Process……..……………………………………………………………………….8
C. Data	
  Quality,	
  Processing	
  and	
  Analysis……………………………………………………………………………………9
IV.	
  Constraints	
  and	
  Limitations…………………………………………………………………………………………………….10
V.	
  Findings….………………………………………………………………………………………………………………….……………12
A. Strategic	
  Objective	
  1	
  (Agriculture)……………………………………………………………………………………….12
B. Summary	
  of	
  Baseline	
  and	
  Final	
  Agricultural	
  Indicator	
  Values	
  for	
  SO1…………………….…………….21
A. Strategic	
  Objective	
  2	
  (Health)………………………………………………………………………………………………24
B. Summary	
  of	
  Baseline	
  and	
  Final	
  Health	
  Indicator	
  Values	
  for	
  SO2…………………………………………..31
A. Strategic	
  Objective	
  3	
  (Community	
  Capacity	
  Building)…………………………………………………………..34
B. Summary	
  of	
  Baseline	
  and	
  Final	
  CCB	
  Indicator	
  Values	
  for	
  SO3……………………………………………….38
VI.	
  Conclusions	
  and	
  Recommendations………………………………………………………………………………………..39
A. Program	
  Strengths……………………………………………………………………………………………………………….39
B. Program	
  Challenges……………………………………………………………………………………………………………..40
C. Recommendations……………………………………………………………………………………………………………….41
VII.	
  Resources	
  Consulted.…………………………………………………………………………………………………………….43
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 2
Abbreviations	
  and	
  Acronyms
AIDS Acquired	
  Immune	
  Deficiency	
  Syndrome
CCB Community	
  Capacity	
  Building
CDC Community	
  Development	
  Committee
CG Care	
  Groups
CSPro Census	
  and	
  Survey	
  Processing	
  System
DHS Demographic	
  and	
  Health	
  Survey
EHA Essential	
  Hygiene	
  Actions
ENA Essential	
  Nutrition	
  Actions
EWS Early	
  Warning	
  Systems
FANTA Food	
  and	
  Nutrition	
  Technical	
  Assistance
FFLG Farmer	
  Field	
  and	
  Life	
  Groups
FGD Focus	
  Group	
  Discussions
FH Food	
  for	
  the	
  Hungry
FFP USAID	
  Office	
  of	
  Food	
  for	
  Peace
FY Fiscal	
  Year
Ha Hectares
HDI Human	
  Development	
  Index
HH Households
HIV Human	
  immunodeficiency	
  virus
IPTT Indicator	
  Performance	
  Tracking	
  Table
IR Intermediate	
  Results
ITN Insecticide	
  treated	
  bed	
  net
JAM Joint	
  Aid	
  Management
Kg Kilograms
KPC Knowledge,	
  Practice	
  and	
  Coverage
KSI Key	
  Stakeholder	
  Interviews
LCL Lower	
  Confidence	
  Level
LQAS Lot	
  Quality	
  Assurance	
  Sampling
M&E Monitoring	
  and	
  Evaluation
MoA Ministry	
  of	
  Agriculture
MoH Ministry	
  of	
  Health
Mtz Mozambican	
  Meticais
MYAP Multi	
  Year	
  Assistance	
  Program
ORS Oral	
  rehydration	
  salts
PPS Probability	
  Proportional to	
  Size
PSO Program	
  Support	
  Officer
SO Strategic	
  Objective
SPSS IBM	
  SPSS	
  Statistics
ToR Terms	
  of	
  Reference
UCL Upper	
  Confidence	
  Level
USAID United	
  States	
  Agency	
  for	
  International	
  Development
VSLA Village	
  Savings	
  and	
  Loan	
  Associations
WHO World	
  Health	
  Organization
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 3
Executive	
  Summary
Background
Over	
   the	
  past	
   two	
  decades,	
  Mozambique	
  has	
  struggled	
   to	
  achieve	
  economic	
  and	
  social	
  stability.	
  Among	
  
the	
  poorest	
  provinces,	
  Cabo	
  Delgado	
   falls	
  behind	
  all	
  other	
  provinces	
  on	
  almost	
  every	
  indicator	
  of	
  social	
  
and	
   economic	
   development.	
   	
   In	
   addition	
   to	
   low	
   income	
   and	
   underdevelopment,	
   Cabo	
   Delgado	
   has	
  
extremely	
  high	
  rates	
  of	
  food	
  insecurity	
  that	
  impedes	
  individual	
  and	
  collective	
  human	
  development.	
  	
  
Infertile,	
  sandy	
  soils	
  in	
  the	
  province	
  lead	
  to	
  low	
  agricultural	
  productivity,	
  while	
  a	
  short	
  agricultural	
  season	
  
creates	
   an	
   extended	
   hungry	
   period.	
   	
   At	
   the	
   same	
   time	
   a	
   lack	
   of	
   crop	
   diversity	
   creates	
   significant	
  
vulnerability	
   to	
   pest	
   infestations,	
   drought,	
   and	
   other	
   shocks.	
   	
   Low	
   household	
   purchasing	
   power	
   results	
  
from	
  limited	
  production	
  of	
  cash	
  crops	
  and	
  poor	
  market	
  access	
  exacerbated	
  by	
  poor	
  infrastructure.	
  Low	
  
literacy	
   levels	
   and	
   language	
   and	
   cultural	
   marginalization	
   compound	
   these	
   factors,	
   impeding	
   vulnerable	
  
populations	
  from	
  engaging	
  in	
  income-­‐generating	
  activities.	
  	
  Cassava,	
  the	
  staple	
  food	
  crop,	
  lacks	
  the	
  caloric	
  
content	
   and	
   nutrient	
   density	
   to	
   meet	
   the	
   nutritional	
   needs	
   of	
   pregnant	
   and	
   lactating	
   women,	
   and	
  
children	
   under	
   five	
   years	
   of	
   age,	
   while	
   limited	
   knowledge	
   of	
   optimal	
   nutrition,	
   health	
   and	
   hygiene	
  
practices	
  undermine	
  maternal	
  and	
  child	
  health.	
  	
  As	
  a	
  result	
  of	
  these	
  factors,	
  more	
  than	
  half	
  of	
  all	
  children	
  
under	
  five	
  in	
  the	
  province	
  suffer	
  from	
  chronic	
  malnutrition	
  and	
  over	
  thirty	
  percent	
  are	
  underweight.
To	
  address	
  these	
  challenges	
  Food	
  for	
  Hungry	
  (FH)	
  in	
  Mozambique	
  has	
  implemented	
  a	
  five-­‐year	
  Multi	
  Year	
  
Assistance	
  Program	
  (MYAP)	
  with	
  funding	
  support	
  from	
  USAID’s	
  Office	
  of	
  Food	
  for	
  Peace.	
  	
  The	
  MYAP	
  was	
  
initially	
  proposed	
   for	
  a	
   three-­‐year	
  period	
  and	
  was	
   subsequently	
  extended	
   to	
   five	
  years	
   through	
  no-­‐cost	
  
extensions.	
  	
  Through	
  the	
  MYAP	
  FH	
  sought	
  to:	
  (1)	
  improve	
  the	
  health	
  and	
  nutritional	
  status	
  of	
  children	
  0-­‐5	
  
years	
  of	
  age;	
  (2)	
  to	
  increase	
  agricultural	
  productivity	
  and	
  strengthen	
  agricultural	
  value	
  chains;	
  and	
  (3)	
  to	
  
increase	
   community	
   resiliency	
   to	
   shocks	
   for	
   31,577	
   households	
   (HH)	
   across	
   the	
   districts	
   of	
   Nangade,	
  
Mocimboa	
  da	
  Praia,	
  and	
  Palma.
In	
  March	
  2013	
  FH	
  hired	
  the	
  services	
  of	
  a	
  consultant	
  to	
  lead	
  a	
  final	
  evaluation	
  of	
  the	
  MYAP	
  to	
  assess	
  the	
  
degree	
   of	
   achievement	
   of proposed	
   outcomes	
   and	
   impacts;	
   to	
   document	
   constraints,	
   lessons	
   learned,	
  
and	
  successes	
  derived	
   from	
   the	
  program;	
  and to	
  assess	
   the	
   relevance	
  of	
   the	
  program’s	
  implementation	
  
strategies	
   and	
   approach.	
   The	
   evaluation	
   did	
   not	
   include	
   an	
   anthropometric	
   study;	
   that	
   study	
   will	
   be	
  
completed	
  as	
  a	
  separate	
  exercise	
  between	
  October	
  and	
  November	
  2013.
Evaluation	
  methodology
The	
   final	
   evaluation	
   was	
   conducted	
   using	
   a	
  mixed	
  methodology.	
   	
   This	
   included	
   a	
   population-­‐based	
   HH	
  
survey	
   and	
   qualitative	
   research	
   consisting	
   of	
   focus	
   group	
   discussions	
   (FGD),	
   key	
   stakeholder	
   interviews	
  
(KSI)	
  and	
  systematic	
  observations	
  at	
   the	
  HH	
  level.	
  	
  For	
   the	
  HH	
  survey	
  a	
  multi-­‐stage,	
  30-­‐cluster	
  sampling	
  
design	
  was	
  used	
  in	
  line	
  with	
  the	
  methodology	
  used	
  in	
  the	
  baseline	
  survey.	
  Thirty	
  clusters	
  were selected	
  
for	
  data	
  collection	
  from	
  among	
  the	
  total	
  number	
  of	
  MYAP	
  intervention	
  communities	
  through	
  systematic	
  
sampling using	
  probability-­‐proportional-­‐to-­‐size	
  (PPS).	
  Parallel	
  sampling	
  was	
  used	
  to	
  conduct	
  both	
  health	
  
and	
  agriculture	
  interviews	
  at	
  the	
  HH	
  level.
Three	
  data	
  collection	
   teams	
  were	
   formed	
   from	
  FH	
   staff,	
  each	
  consisting	
  of	
  one	
  Team	
  Leader,	
   two	
  Data	
  
Collection	
   Supervisors	
   and	
   eight	
   Enumerators.	
   	
   Each	
   team	
   was	
   assigned	
   to	
   a	
   project	
   district	
   based	
   on	
  
language	
   and	
   cultural	
   considerations.	
   	
   Teams	
   were	
   trained	
   on	
   data	
   collection	
   processes	
   and	
   protocols	
  
over	
   four	
   days,	
   and	
   data	
   collection	
   was	
   subsequently	
   completed	
   over	
   nine	
   days	
   between	
   March	
   and	
  
March	
  27,	
  2013.	
  	
  A	
  total	
  of	
  407	
  health	
  surveys,	
  333	
  agriculture	
  surveys,	
  eight	
  Focus	
  Group	
  Discussions,	
  six	
  
Key	
  Stakeholder	
  Interviews	
  and	
  72	
  systematic	
  observations	
  were	
  completed.
A	
  team	
  of	
  data	
  entry	
  clerks	
  was	
  trained	
  to	
  enter	
  quantitative	
  data	
  into	
  the	
  US	
  government’s	
  open	
  source	
  
Census	
   and	
   Survey	
   Processing	
   System	
   (CSPro	
   5.0)	
   software	
   with	
   a	
   user-­‐friendly	
   data	
   template	
   that	
  
mirrored	
  the	
  questionnaire	
  to	
  reduce	
  error.	
  	
  Data	
  entry	
  was	
  completed	
  simultaneous	
  to	
  data	
  collection.	
  	
  
Initial	
  data	
   cleaning	
  and	
   verification	
  and	
   preliminary	
   quantitative	
  analysis	
  were	
   conducted	
   using	
   CSPro.	
  
Final	
   quantitative	
   analysis	
   of	
   baseline	
   and	
   final	
   evaluation	
   data	
   was	
   conducted	
   in	
   SPSS	
   21.0.	
   	
   Relevant	
  
proportions	
   and	
   means	
   were	
   generated	
   for	
   each	
   indicator	
   and	
   t-­‐tests	
   were	
   conducted	
   to	
   test	
   the	
  
significance	
  of	
  differences	
  in	
  indicator values	
  between	
  the	
  baseline	
  and	
  final	
  evaluations.	
  	
  The	
  p	
  values	
  of	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 4
each	
   of	
   the	
   findings	
   were	
   subsequently	
   verified	
   using	
   an	
   Indicator	
   Confidence	
   Interval	
   Calculation	
  
Worksheet	
  provided	
  by	
  FH.	
  	
  Qualitative	
  data	
  was	
  analyzed	
  using	
  simple	
   trend	
  analysis	
  and	
  observations	
  
from	
   the	
   infrastructure	
   checklists	
   were	
   tabulated	
   and	
   processed	
   in	
   Excel.	
   	
   Analysis	
   resulting	
   from	
   the	
  
quantitative	
  and	
   qualitative	
   data	
  was	
   used	
   to	
   develop	
  a	
   final	
  evaluation	
   report	
   draft	
   for	
  which	
  FH	
   staff	
  
provided	
  feedback.	
  	
  That	
  feedback	
  was	
  incorporated	
  into	
  this	
  final	
  report.	
  
Findings
Of	
   the	
   16	
   indicators	
   measured	
   under	
   the	
   agriculture component,	
   15	
   of	
   them	
   experienced	
   increases	
  
between	
   the	
  baseline	
  and	
   final	
  evaluations.	
  	
  Four	
  indicator	
  values	
   fell	
  below	
  program	
   targets,	
  while	
   the	
  
rest	
  met	
  or	
  exceeded	
  target	
  values.
Both	
   respondent	
   farmers	
   and	
   FH	
   staff	
   confirmed	
   through	
   qualitative	
   discussions	
   that	
   productivity	
   has	
  
increased	
   in	
   participant	
   fields	
   and	
   the	
   consistent	
   productivity	
   growth	
   demonstrated	
   through	
   the	
  
program’s	
   Annual	
   Agriculture	
   Surveys	
   supports	
   these	
   assertions.	
   	
   Farmers	
   from	
   the	
   program’s	
   Farmer	
  
Field	
   and	
   Life	
   Groups	
   (FFLG)	
   believe	
   that	
   the	
   technologies	
   that	
   FH	
   introduced	
   played	
   a	
   role	
   in	
   these	
  
productivity	
  increases.	
  	
  They	
  also	
  expressed	
  a	
  high	
  degree	
  of	
  satisfaction	
  and	
  confidence	
  with	
  the	
  FFLG	
  as	
  
a	
  sustainable	
  means	
  to	
  transmit	
  technical	
  assistance	
  messages	
  and	
  support	
  replication	
  of	
  best	
  practices.	
  	
  
FFLG	
   members	
   continue	
   to	
   meet	
   even	
   in	
   the	
   absence	
   of	
   FH	
   staff.	
   	
   Given	
   the	
   staffing	
   and	
   resource	
  
constraints	
   of	
   the	
   Ministry	
   of	
   Agriculture,	
   this	
   is	
   an	
   important	
   gain	
   that can	
   serve	
   as	
   a	
   foundation	
   for	
  
continued	
  farmer-­‐to-­‐farmer	
  based	
  agricultural	
  extension	
  support	
  in	
  the	
  future.	
  
Market	
  links	
  and	
  economic	
  opportunities	
  have	
  also	
  increased	
  as	
  a	
  result	
  of	
  the	
  program.	
  	
  While	
  there	
  has	
  
been	
  no	
  significant	
  change	
  in	
   the	
  percent	
  of	
  HH	
  producing	
  cash	
  crops,	
  60%	
  of	
   those	
  who	
  produce	
  cash	
  
crops	
  sell	
  a	
  portion	
  of	
  those	
  crops	
  either	
  collectively	
  or	
  individually	
  at	
  local	
  markets,	
  resulting	
  in	
  at	
  least	
  
$85,000	
   in	
   additional	
   income.	
   The	
   program’s	
   marketing	
   activities	
   could	
   have	
   been	
   strengthened	
   by	
  
providing	
  training	
  and	
  mentoring	
  to	
  farmers	
  in	
  market	
  analysis,	
  negotiation,	
  value	
  addition	
  or	
  processing,	
  
and	
  by	
  strengthening	
  linked	
  businesses	
  to	
  ensure	
  integral	
  and	
  strong	
  value	
  chains.	
  
Engagement	
  with	
  VSLAs	
  has	
  also	
  provided	
  an	
  opportunity	
   for	
  at	
  least	
  3,550	
  men	
  and	
  women	
  to	
  engage	
  
with	
   financial	
   services,	
   albeit	
   informal	
   and	
   likely	
   for	
   the	
   first	
   time.	
   	
   They	
   have	
   saved	
   over	
   $200,000	
  
collectively,	
  earned	
  $44,500	
  on	
  those	
  savings	
  through	
  interest	
  income	
  and	
  provided	
  much	
  needed	
  capital	
  
to local	
  entrepreneurs	
  and	
  individuals.	
  	
  It	
  might	
  be	
  worth	
  exploring	
  whether	
  the	
  flexibility	
  of	
  the	
  current	
  
VSLA	
   modality-­‐ allowing	
   participants	
   to	
   save	
   if	
   they	
   want	
   to-­‐ may	
   keep	
   VSLAs	
   from	
   reaching	
   their	
   full	
  
potential	
  as	
  a	
  consumption	
  smoothing	
  and	
  vulnerability	
  reduction	
  mechanism.	
  
Of	
   the	
   12	
   health	
   indicators	
   included	
   in	
   the	
   IPTT	
   9	
   of	
   them	
   (75%)	
   experienced	
   statistically	
   significant	
  
changes	
   in	
   values	
   between	
   the	
   baseline	
   and	
   final	
   evaluations.	
   	
   Although	
   indicator	
   values	
   fell	
   below	
  
program	
  targets,	
  most	
  of	
  the	
  behavior	
  change	
  targets	
  may	
  have	
  been	
  unrealistically	
  ambitious,	
  especially	
  
given	
  that	
  key	
  changes	
  were	
  expected	
  during	
  a	
  shortened	
  Care	
  Group	
  cycle.	
  
Among	
  essential	
  nutrition	
  actions,	
  breastfeeding	
  behaviors	
  showed	
  marked	
  improvement,	
  both	
  in	
  terms	
  
of	
   early	
   initiation	
   and	
   exclusive	
   breastfeeding	
   of	
   infants	
   0-­‐ 6	
   months	
   old.	
   	
   The	
   percent	
   of	
   children	
  
exclusively	
  breastfed	
  in	
  MYAP	
  project	
  areas	
  now	
  exceeds	
   the	
  national	
  average	
  by	
  almost	
  35	
  percentage	
  
points.	
  Likewise,	
  values	
   for	
  “adequate”	
  complementary	
   feeding	
   (minimum	
  dietary	
  diversity	
  and	
  number	
  
of	
  meals)	
   improved	
   by	
   25	
   percentage	
   points	
   during	
   the	
   project.	
   	
   These	
   values-­‐ previously	
   substantially	
  
lower	
  than	
  the	
  national	
  average-­‐ are	
  now	
  above	
  that	
  average	
  by	
  approximately	
  10	
  percentage	
  points.
The	
   percentage	
   of	
   caretakers	
   who	
   can	
   name	
   three	
   signs	
   of	
   childhood	
   illness	
   and	
   three	
   methods	
   to	
  
prevent	
  HIV	
  infection	
  have	
  both	
  increased	
  significantly	
  during	
  the	
  program.	
  Reported	
  malaria	
  prevention	
  
has	
   improved	
   thanks	
   to	
   high	
   adoption	
   of	
   ITNs,	
   at	
   approximately	
   90%	
   of	
   all	
   caregivers	
   interviewed,	
  
representing	
  a	
   rate	
  of	
  ITN	
  use	
   that	
  is	
  40	
  percentage	
  points	
  higher	
   than	
   the	
  national	
  average.	
  There	
  has	
  
also	
  been	
  a	
  statistically	
  significant	
  decrease	
  in	
  reported	
  cases	
  of	
  diarrhea.	
  
Most	
   indicator	
   values	
   related	
   to	
   water,	
   sanitation	
   and	
   hygiene	
   have	
   all	
   seen	
   statistically	
   significant	
  
increases	
  over	
  baseline	
  levels	
  as	
  well.	
  	
  Caretaker	
  access	
   to	
  improved	
  water	
  sources	
  has	
  nearly	
  doubled,	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 5
and	
   reported	
  personal	
  hygiene	
  behavior	
  has	
  improved	
  by	
  approximately	
   40	
  percentage	
  points	
  over	
   the	
  
baseline.	
  	
  More	
  families	
  have	
  soap	
  and	
  water	
  available	
  to	
  them	
  and	
  significantly	
  more	
  caretakers	
  report	
  
washing	
   hands	
  at	
  all	
   the	
   necessary	
   times.	
  There	
   has	
   been	
  an	
  increase	
   of	
  almost	
   40%	
  in	
   proper	
  excreta	
  
disposal,	
  although	
  observations	
  of	
  HH	
  latrines	
  suggest	
  that	
  hygiene	
  indicator	
  values	
  could	
  improve	
  further	
  
through	
   refresher	
   training	
  on	
  proper	
  HH	
  latrine	
  maintenance	
  and	
  care.	
  	
  The	
  vast	
  majority	
  of	
  caretakers	
  
properly	
  store	
  water	
  in	
  containers	
  with	
  lids,	
  although	
  only	
  approximately	
  34%	
  treat	
  their	
  water.	
  	
  This	
  may	
  
be	
  the	
  result	
  of	
  the	
  costs	
  associated	
  with	
  water	
  treatment.	
  	
  
All	
  the	
  indicator	
  values	
  relating	
  to	
  community	
  capacity	
  building increased	
  over	
  the	
  baseline.	
  	
  Almost	
  600	
  
community	
   leaders	
   were	
   trained	
   and	
   HCA	
   was	
   completed	
   in	
   each	
   of	
   the	
   communities.	
   	
   Individual	
  
community	
  leadership	
  capacity	
  is	
  somewhat	
  mixed,	
  although	
  given	
  the	
  scope	
  of	
  the	
  program’s	
  activities	
  
this	
   isn’t	
   surprising.	
   	
   It	
   is	
   clear,	
   however,	
   that	
   the	
   program	
   has	
   achieved	
   a	
   high	
   level	
   of	
   community	
  
engagement	
  and	
  ownership	
   through	
   the	
  use	
  of	
  local	
  promoters	
   for	
  health	
  and	
  agriculture,	
   training	
  and	
  
support	
  to	
  create	
  Mother	
  Leaders	
  for	
  Care	
  Groups,	
  development	
  and	
  training	
  of	
  Farmer	
  Field	
  Life	
  Groups	
  
and	
  Community	
  Development	
  Committees.	
  	
  	
  The	
  leadership	
  and	
  technical	
  skills	
  that	
  these	
  individuals	
  and	
  
groups	
  have	
  gained	
  will	
  provide	
  a	
  strong	
  foundation	
  for	
  continued	
  health	
  and	
  wellbeing	
  in	
  their	
  families	
  
and	
  for	
  community	
  development	
  across	
  all	
  MYAP	
  intervention	
  areas.	
  	
  
FH	
   support	
   has	
   also	
   resulted	
   in	
   significant	
   gains	
   in	
   community	
   infrastructure	
   and	
   assets.	
   	
   The	
   most	
  
important	
  of	
   these	
  are	
   the	
  contributions	
   to	
  water	
  and	
   sanitation	
  infrastructure.	
  	
  	
  In	
  community	
  visits	
  it	
  
was	
  clear	
  that	
  these	
  contributions	
  have	
  been	
  an	
  important	
  success	
  in	
  the	
  program	
  and	
  are	
  likely	
  to	
  have	
  
contributed	
   in	
   important	
   ways	
   to	
   the	
   health	
   and	
   hygiene	
   outcomes	
   under	
   the	
   program’s	
   health	
  
component.
Conclusions	
  and	
  recommendations
The	
   combination	
   of	
   increased	
   productivity	
  and	
   increased	
   incomes	
   derived	
   from	
   FH	
  agricultural	
   support	
  
has	
  contributed	
  to	
  statistically	
  significant	
  gains	
  in	
  both	
  Household	
  Food	
  Provisioning	
  and	
  overall	
  Dietary	
  
Diversity	
   Scores,	
   key	
  measures	
   of	
  HH	
  food	
  access.	
  	
  In	
   the	
   remainder	
   of	
   the	
   program	
   FH	
  might	
   validate	
  
these	
  conclusions	
   through	
  both	
   the	
  program’s	
   final	
  Annual	
  Agriculture	
  Survey	
  in	
  2013	
  and	
  by	
  compiling	
  
full	
  marketing	
  income	
  and	
  VSLA	
  data.	
  	
  
Under	
  the	
  program’s	
  health	
  component	
  FH	
  has	
  facilitated	
  significant	
  improvements	
  in	
  essential	
  nutrition	
  
actions,	
   essential	
   hygiene	
   actions,	
   and	
   understanding,	
   prevention	
   and	
   treatment	
   of	
   childhood	
   illness.	
  	
  
These	
  changes	
  are	
  likely	
   to	
  have	
  improved	
  food	
  utilization among	
   the	
   target	
  beneficiaries	
  by	
  increasing	
  
overall	
  nutrient	
  intake	
  and	
  use.	
   Some	
  of	
   the	
  program’s	
  health	
  gains	
  are	
  more	
  modest	
   than	
  might	
  have	
  
been	
   expected.	
   The	
   modifications	
   to	
   the	
   Care	
   Group	
   model-­‐ from	
   a	
   five-­‐year	
   cycle	
   accommodating	
   all	
  
children	
  under	
  5	
  to	
  a	
  two-­‐year	
  cycle	
  accommodating	
  only	
  children	
  under	
  two-­‐ may	
  have	
  played	
  a	
  role	
  in	
  
these	
  differences.	
  	
  Future	
  programs	
  should	
  either	
  consider	
  a	
  return	
  to	
  the	
  standard	
  five-­‐year	
  Care	
  Group	
  
cycle	
  or	
  modify	
  the	
  Care	
  Group	
  training	
  modules	
  to	
  match	
  the	
  reduced	
  time	
  period.	
  The	
  anthropometric	
  
study	
   in	
   late	
   2013	
   will	
   provide	
   an	
   opportunity	
   to	
   more	
   closely	
   examine	
   the	
   relationship	
   between	
   the	
  
behavior	
  and	
  knowledge	
  improvements	
  highlighted	
  here	
  and	
  actual	
  rates	
  of	
  malnutrition.
The	
   program’s	
   emphasis	
   on	
   community	
   engagement	
   has	
   provided	
   strong	
   mechanisms for	
   knowledge	
  
transfer	
  and	
  behavior	
  change	
  and	
  a	
  solid	
  foundation	
  for	
  continued	
  community	
  development.	
  During	
  the	
  
remainder	
  of	
  the	
  program	
  FH	
  should	
  prioritize	
  transition	
  and	
  sustainability	
  planning	
  as	
  well	
  as	
  refresher	
  
training	
   on	
   some	
   of	
   the	
   project’s	
   key	
   messages	
   (such	
   as	
   the	
   diagnosis,	
   prevention,	
   and	
   treatment	
   of	
  
childhood	
  illness;	
  HH	
  latrine	
  use	
  and	
  maintenance; infrastructure	
  maintenance)	
   to help	
  ensure	
   that	
   the	
  
program’s	
  benefits	
  can	
  be	
  sustained	
  as	
  long	
  as	
  possible.	
  	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 6
I.	
  Introduction
In	
  line	
  with	
  the	
  evaluation	
  Terms	
  of	
  Reference	
  (TOR),	
  approved	
  by	
  USAID’s	
  Office	
  of	
  Food	
  For	
  Peace	
  
(FFP),	
  the	
  MYAP	
  final	
  evaluation	
  has	
  three	
  key	
  objectives:
1) Assess	
  the	
  degree	
  of	
  achievement	
  of	
  the	
  MYAP’s	
  proposed	
  outcomes	
  and	
  impacts;
2) Document	
  constraints,	
  lessons	
  learned,	
  and	
  successes	
  derived	
  from	
  the	
  program;	
  and
3) Assess	
  the	
  relevance	
  of	
  the program’s	
  implementation	
  strategies	
  and	
  approach	
  
The	
  audience	
  for	
  the	
  evaluation	
  findings	
  is	
  Food	
  for	
  the	
  Hungry,	
  the	
  USAID	
  Mission	
  in	
  Mozambique	
  
and	
  FFP as	
  well	
  as	
  other	
  stakeholders	
  who	
  can	
  benefit	
  from	
  the	
  evaluation’s	
  results.	
  	
  	
  
II.	
  Overview	
  of	
  Program	
  Context	
  and	
  Implementation
A.	
  Overview	
  of	
  the	
  food	
  security	
  situation	
  in	
  Mozambique	
  at	
  program	
  start
Over	
   the	
   past	
   two	
   decades,	
   Mozambique	
   has	
   struggled	
   to	
   achieve	
   economic and	
   social	
   stability	
  
through	
  the	
  effects	
  of	
  civil	
  war	
  and	
  harsh	
  natural	
  disasters.	
  	
  Rich	
  natural	
  resources	
  and	
  international	
  
support	
   have	
   fueled	
  economic	
  growth, which	
  improved	
  Mozambique’s	
  Human	
  Development	
  Index	
  
(HDI)	
   from	
   171/177	
   countries	
   in	
   2002	
   to	
   168/
177	
   countries	
   in	
   2004.	
   	
   Although	
   there	
   has	
   been
positive	
  change	
  in	
  Mozambique’s	
  historically	
  poor	
  health	
  and	
  social	
  indicators,	
  many	
  indicators,	
  such	
  
as	
   infant	
  mortality,	
   illiteracy	
   and	
   life	
  expectancy,	
   continue	
   to	
   rank	
   among	
   the	
  world’s worst.	
  Cabo	
  
Delgado,	
  among	
  the	
  poorest	
  provinces,	
  falls	
  behind	
  all	
  other	
  provinces	
  on	
  almost	
  every	
  indicator	
  of	
  
social	
  and	
  economic	
  development.
In	
  addition	
  to	
  low	
  income	
  and	
  underdevelopment,	
  Cabo	
  Delgado	
  has	
  high	
  levels	
  of underlying	
  food	
  
insecurity.	
   Soaring	
  above	
  all	
  of	
   the	
  provinces	
  in	
  Mozambique,	
  more	
   than	
  half	
  of	
  all	
  children	
  under	
  
five	
  suffer	
  from	
  chronic	
  malnutrition	
  and	
  thirty-­‐four	
  percent	
  are	
  underweight1
,	
  an	
  indicator	
  of	
  both	
  
chronic	
   and	
   acute	
   malnutrition.	
   	
   Cabo	
   Delgado’s acute	
   malnutrition	
   rates	
   are	
   lower	
   than	
   some	
  
provinces	
  and	
  fall	
  below	
  emergency	
  levels;	
  however,	
  given	
  the	
  high	
  rates	
  of	
  chronic	
  malnutrition	
  in	
  
Mozambique-­‐ and	
  the	
  known	
  irreversible	
  effects	
  of	
  chronic	
  malnutrition	
  on	
  childhood	
  development,	
  
cognition	
   and	
   long-­‐term	
   productivity2
-­‐ there	
   is	
   an important	
   role	
   for	
   prevention	
   activities	
   through	
  
behavior	
  change-­‐based	
  interventions.
Food	
   availability	
   is	
   undermined	
   by infertile,	
   sandy	
   soils	
   leading	
   to	
   low	
   agricultural	
   productivity.	
   	
   A	
  
short	
   agricultural	
   season	
   creates	
   an	
   extended	
   hungry	
   season	
   with	
   attendant	
   food	
   provisioning	
  
difficulties,	
   while	
   a	
   lack	
   of	
   crop	
   diversity	
   creates	
   significant	
   vulnerability	
   to	
   shocks	
   related	
   to	
   pest	
  
infestations,	
   drought,	
   and	
   other	
   exogenous	
   shocks.	
   	
   At	
   the	
   same	
   time,	
   poor	
   post-­‐harvest	
   crop	
  
storage	
  leads	
  to	
  food	
  loss.
An	
  extremely	
  low	
  level	
  of	
  household	
  purchasing	
  power,	
   resulting	
   from	
  poor	
  market	
  access,	
  limited	
  
infrastructure,	
  as	
  well	
  as	
  lack	
  of	
  marketable	
  products	
  or	
  services	
  affect food	
  access.	
  	
  Moreover,	
  low	
  
literacy	
  and	
  education	
  levels,	
  compounded	
  by	
  language	
  and	
  cultural	
  marginalization,	
  particularly	
  of	
  
women,	
  impede	
  vulnerable	
  populations	
  from	
  engaging	
  in	
  income-­‐generating	
  activities.	
  
Food	
  Utilization is	
  affected	
  by	
  poor	
  diet	
  composition	
  and	
  diversity,	
  with	
  seasonal	
  hungry	
  periods. The	
  
staple food	
   (cassava)	
   lacks	
   the	
   caloric	
   and	
   nutrient	
   density	
   to	
   meet	
   the	
   nutritional	
   needs	
   of	
  
vulnerable	
  populations, pregnant	
  and	
  lactating	
  women, and	
  children	
  less	
  than	
  five	
  years	
  of	
  age.	
  	
  	
  At	
  
the	
  same	
  time,	
  limited	
  knowledge	
  of	
  optimal	
  nutrition,	
  health	
  and	
  hygiene	
  practices	
  impede	
  families	
  
from	
  fully	
  using	
  the	
  resources	
  at	
  hand.
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   1 Mozambique	
  Demographic	
  and	
  Health	
  Survey,	
  2003 2 Tracking	
  Progress	
  and	
  Child	
  and	
  Maternal	
  Nutrition:	
  a	
  survival	
  and	
  development	
  priority,	
  UNICEF	
  2009
3 Provided	
  by	
  the	
  FH	
  Community	
  Capacity	
  Building	
  Manager 2 Tracking	
  Progress	
  and	
  Child	
  and	
  Maternal	
  Nutrition:	
  a	
  survival	
  and	
  development	
  priority,	
  UNICEF	
  2009
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 7
B.	
  Overview	
  of	
  the	
  FH	
  MYAP	
  in	
  Cabo	
  Delgado
Through	
   the MYAP	
   Food	
   for	
   the	
   Hungry	
   sought	
   to	
   improve	
   the	
   health	
   and	
   nutritional status	
   of	
  
children	
   0-­‐5 years	
   of	
   age;	
   to	
   increase	
   agricultural	
   productivity	
   and	
   strengthen	
   agricultural	
   value	
  
chains;	
   and	
   to	
   increase	
   community	
   resiliency	
   to	
   shocks	
   for	
   31,577	
   households	
   (HH)	
   across	
   the	
  
districts	
   of	
  Nangade,	
  Mocimboa	
   da	
   Praia,	
  and	
   Palma	
  in	
   Cabo	
  Delgado	
   province. Table	
   1	
   shows	
   the	
  
program	
  Results	
  Framework.
Table	
  1:	
  FH	
  Mozambique	
  MYAP	
  Results	
  Framework
SO1. Protect	
   and	
   enhance	
  
livelihood	
   capacities	
   of	
  
vulnerable	
   farming	
   households	
  
via	
   agricultural	
   production	
   and	
  
marketing
SO2. Protect	
   and	
   enhance	
  
human	
   capabilities	
   through	
  
improved	
  health	
  and	
  nutritional	
  
status	
   of	
   pregnant	
   women,	
  
preschool	
   children,	
   and	
   HIV+	
  
individuals	
  
SO3. Increased	
   community	
  
capacity	
   to	
   influence	
   factors	
  
that	
   affect	
   food	
   security	
   and	
  
resiliency	
  to	
  shocks
IR	
   1.1:	
  Agricultural	
   productivity	
  
and	
   production	
   diversified	
   and	
  
increased	
  
IR	
   1.2:	
   Natural	
   resource	
   base	
  
protected	
  and	
  enhanced
IR	
   1.3:	
   Market-­‐led	
   income	
  
sources	
   increased	
   and	
  
diversified	
  
IR2.1:Improved	
  use	
  of	
  Essential	
  
Nutrition	
   Actions	
   (ENA)	
   by	
  
pregnant	
  women	
  and	
  mothers	
  
of	
  young	
  children
IR2.2: Improve	
   mother’s	
  
ability	
   to	
   prevent,	
   diagnose	
  
and	
   manage	
   diseases	
   that	
  
exacerbate	
   malnutrition,	
  
including	
   dietary	
   management	
  
of	
  illness	
  
IR2.3: Improved	
   access	
   to	
  
clean	
   water,	
   sanitation	
  
facilities,	
  and	
  Essential	
  Hygiene	
  
Actions	
  (EHA)	
  
IR3.1: Increased	
   leadership	
  
capacity	
   of	
   existing	
   formal	
   and	
  
informal	
   community	
   leaders	
   to	
  
address	
   factors	
   that	
  affect	
   food	
  
security
IR	
   3.2:	
   Increased	
   community	
  
level	
   economic	
  
infrastructure/assets	
  
IR	
  3.3: Increased	
   ability	
   to	
  
predict	
  and	
  mitigate	
  shocks
III.	
  Evaluation	
  Methodology
A.	
  Sampling
Mixed	
  methodology
The	
  final	
  evaluation	
  was	
  conducted	
  using	
  a	
  mixed	
  methodology,	
  consisting	
  of	
  a	
  population-­‐based	
  HH	
  
survey,	
  focus	
  group	
  discussions	
  (FGD),	
  key	
  stakeholder	
  interviews	
  (KSI)	
  and	
  systematic	
  observations	
  
at	
  the	
  HH	
  level.	
  	
  
Quantitative	
  Sampling
For	
   the	
  HH	
   survey-­‐ the	
   primary	
  methodology	
   for	
   the	
   comparison	
   of	
   program	
  indicator	
  values	
   over	
  
time-­‐ a	
   multi-­‐stage,	
   30-­‐cluster	
  sampling	
   design	
   was	
   used	
   based on	
   the	
   KPC	
   survey	
   methodology	
  
outlined	
  in	
   the	
  KPC2000+	
  Field	
  Guide.	
   Under	
   these	
   standards,	
  separate	
   surveys	
  are	
  conducted	
   for	
  
each	
   program	
   target	
   group-­‐ in	
   this	
   case	
   a	
   health	
   survey	
   for	
   primary	
   caregivers of	
   children	
   0-­‐24	
  
months	
   and	
   an	
   agriculture	
   survey administered	
   to	
   identified	
   HH	
   heads-­‐ using	
   parallel	
   sampling.	
  
Under	
   the	
   KPC	
  methodology, 10	
   surveys	
   are	
   administered	
   at	
   random	
  within	
   each	
   cluster	
  for	
   each	
  
target	
   group,	
   for	
   a	
   total	
   (target)	
   sample	
   size	
   of	
   600	
   (300	
   for	
   health	
   and	
   300	
   for	
   agriculture).	
   This	
  
sample	
  size	
  generates a	
  confidence	
  level	
  of	
  95%,	
  with	
  a	
  confidence	
  interval	
  of	
  approximately	
  10%.
The	
   30	
   clusters	
   for	
   the	
   final	
   evaluation	
   were	
   selected	
   from	
   among	
   the	
   total	
   number	
   of	
   MYAP	
  
intervention	
   communities	
   (50)	
   through	
   systematic	
   sampling using	
   probability-­‐proportional-­‐to-­‐size
(PPS),	
   ensuring	
   that	
   the	
   total	
   sample	
   distribution	
   was	
   similar	
   to	
   the	
   population	
   distribution	
   across	
  
MYAP	
   districts.	
   The	
   final	
  list	
   of	
   selected	
   clusters	
   can	
   be	
   found	
  in	
  Annex	
  1. In	
  addition,	
  in	
  order	
   to	
  
ensure	
  random	
  selection	
  of	
  HH	
   for	
  survey	
  administration	
  within	
  each	
  cluster,	
  data	
  collection	
  teams	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 8
worked	
   with	
   community	
   leaders	
   to	
   identify	
   the	
   community	
   center.	
   	
   From	
   that	
   location	
   the	
   teams	
  
spun	
  a	
  bottle	
  and	
  began	
  data	
  collection	
  at	
  the	
  first	
  closest	
  house	
  in	
  the	
  direction	
  to	
  which	
  the	
  bottle	
  
pointed.	
  	
  Subsequent	
  surveys	
  were	
  administered	
  using	
  the	
  “nearest	
  door”	
  approach	
  until	
  the	
  target	
  
number	
   of	
   surveys	
   was	
   completed	
   per	
   cluster. The	
   data	
   collection	
   team	
   interviewed	
   primary	
  
caregivers	
   of	
   children	
   between	
   the	
   ages	
   of	
   0	
   and	
   24	
   months	
   for	
   the	
   health	
   survey,	
   while	
   the	
  
agriculture	
  survey	
  was	
  administered	
  to	
  identified	
  HH	
  heads	
  in	
  each	
  HH	
  unit.	
  	
  
Qualitative Sampling
The	
   qualitative	
   design consisted	
   of (1)	
   key	
   stakeholder	
   interviews	
   with	
   project	
   staff	
   and	
   MoH	
  
personnel,	
   (2)	
   focus	
   group	
   discussions	
   with	
   farmer	
   groups,	
   mother	
   leaders	
   and	
   community	
  
development	
  association	
  and	
  (3)	
  structured	
  observations	
   of	
   use/maintenance	
   of	
  key	
  infrastructure	
  
supported	
  by	
  the	
  project,	
  including	
  HH	
  latrines,	
  school/public	
  latrines	
  and	
  tube/shallow	
  wells. These	
  
methodologies	
  were	
  selected	
  to	
  provide	
  information	
  to	
  complement	
  or	
  expand	
  upon	
   findings	
   from	
  
the	
  survey.	
  	
  
Key	
   staff	
   interviews	
   were	
   planned	
   with	
   one health	
   and one agriculture	
   program	
   staff from	
   each	
  
project	
   district.	
   	
   Likewise	
   one	
   focus	
   group	
   discussion	
   was	
   planned	
   with	
   mother	
   leaders,	
   farmer	
  
groups	
  and	
  community	
  development	
  associations	
  in	
  each	
  project	
  district	
  (for	
  a	
   total	
  of	
  3	
  FGD	
  with	
  
each	
  group)	
  based	
  on	
  random	
  selection	
  from	
  among	
  communities	
  that	
  were	
  not	
  included	
  in	
  the	
  HH	
  
survey. At	
  least	
  one	
  structured	
  observation	
  for	
  each	
  infrastructure	
  type	
  was	
  also	
  planned	
  at	
  the	
  HH	
  
(latrines)	
   and	
   community	
   (public/school	
   latrines)	
   levels	
   in	
   each	
   of	
   the 30	
   clusters	
  selected	
   for	
   HH	
  
surveys	
  to	
  avoid	
  the	
  need	
  for	
  additional	
  logistics	
  support	
  and to	
  ensure	
  efficiency	
  in	
  data	
  collection.	
  
B.	
  Data	
  collection	
  instruments	
  and	
  process
Quantitative
The	
   team	
  used	
   the	
  same	
  surveys	
   for	
   the	
  quantitative	
  portion	
  of	
   the	
  exercise	
  as	
   those	
  used	
   for	
   the	
  
baseline	
  survey.	
  	
  Only	
  minimal,	
  largely	
  cosmetic	
  changes	
  were	
  made	
  on	
  a	
  small	
  number	
  of	
  questions	
  
to	
  increase	
  clarity	
  and	
  ensure	
  respondent	
  understanding.	
  	
  However,	
  in	
  order	
  to	
  ensure	
  comparability	
  
between	
   the	
   baseline	
   and	
   final	
   evaluations,	
   no	
   conceptual	
   or	
   content	
   changes	
   were	
   made	
   in	
   the	
  
tools.	
  	
  During	
  enumerator	
  training	
  the	
  surveys	
  were	
  translated	
  into	
  Makonde,	
  Kimwane	
  and	
  Makua,	
  
the	
   three	
   primary	
  local	
  languages	
   of	
   the	
  MYAP	
  intervention	
   districts.	
  	
  Copies	
   of	
   the	
   survey	
   can	
   be	
  
found	
  in	
  Annex	
  2.	
  
To	
   facilitate	
   data	
   collection	
   and	
   quality	
   control,	
   three	
   data	
   collection	
   teams	
   were	
   formed	
   from	
   FH	
  
staff,	
  each	
  consisting	
  of	
  one	
  Team	
  Leader,	
  two Data	
  Collection	
  Supervisors	
  and	
  eight Enumerators.	
  	
  
Each	
   team	
   was	
   assigned	
   to	
   a	
   project	
   district	
   based	
   on	
   language	
   and	
   cultural	
   considerations.	
  
Furthermore,	
   teams	
  were	
  assigned	
   to	
  districts	
  where	
   they	
  do	
  not	
  implement	
  programs	
  in	
  order	
   to	
  
reduce	
  potential	
  bias. 	
  	
  Team	
  Leaders	
  were	
  responsible	
  for	
  daily	
  team	
  planning,	
  ensuring	
  logistics	
  and	
  
overall	
   quality	
   control	
   as	
   well	
   as	
   daily	
   communication	
   with	
   the	
   Final	
   Evaluation	
   Consultant.	
   	
   Data	
  
Collection	
   Supervisors	
   oversaw	
   enumerators	
   and	
   supported	
   data	
   collection	
   efforts,	
   spot	
   checking	
  
quality	
  and	
  providing	
  daily	
  feedback.	
  	
  Enumerators	
  administered	
  the	
  survey	
  and	
  filled	
  out	
  the	
  survey	
  
forms.	
  	
  In	
  each	
  data	
  collection	
  team	
  the	
  four	
  enumerators	
  were	
  broken	
  into	
  two	
  pairs.	
  In	
  each	
  pair,	
  
one	
   person	
   administered	
   the	
   survey,	
   while	
   the	
   other	
   person	
   completed	
   the	
   form.	
   	
   To	
   further	
  
mitigate	
  against	
  bias	
  and	
  ensure	
  accurate	
   responses	
   to	
  culturally	
  sensitive	
  questions	
  each	
  pair	
  was	
  
composed	
   of	
   one	
   male	
   and	
   one	
   female	
   member. A	
   list	
   of	
   data	
   collection	
   teams	
   can	
   be	
   found	
   in	
  
Annex	
  3.
Data	
  collection	
  training	
  was	
  conducted	
  during four	
  days.	
  	
  One	
  full	
  day	
  was	
  provided	
  for	
  Team	
  Leaders	
  
and	
   Supervisors,	
   focused	
   on	
   ensuring	
   understanding	
   of	
   the	
   survey	
   tools	
   and	
   providing	
   a	
   detailed	
  
overview	
  of	
  HH	
  sampling	
  and	
  the	
  data	
  collection	
  process.	
  	
  Subsequently,	
  two	
  full	
  days	
  of	
  classroom-­‐
based	
  training	
  were	
  provided	
  for	
  the	
  full	
  data	
  collection	
  teams	
  to	
  ensure	
  standard	
  understanding	
  of	
  
survey	
  questions	
  and	
  key	
  concepts,	
  to	
  provide	
  an	
  overview	
  of	
  HH	
  selection	
  and	
  to	
  practice	
  using	
  the	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 9
tool	
  through	
  simulation	
  activities.	
  	
  A	
  third	
  day	
  of	
  training	
  was	
  used	
  to	
  field	
  test	
  the	
  full	
  data	
  collection	
  
process,	
  from	
  HH	
  selection	
  to	
  survey	
  tool	
  administration	
  and	
  quality	
  control.	
  
A	
   total	
   of	
   407	
   Health	
   surveys	
   and	
   333	
   Agriculture	
   surveys	
   were	
   administered	
   during	
   the	
   survey	
  
exercise.	
  	
  To	
  account for	
  non-­‐response and	
  potential	
  errors	
  in	
  data	
  collection,	
  teams	
  were	
  instructed	
  
to	
  complete	
  11	
  surveys	
  in	
  each	
  cluster.	
  	
  Data	
  
collection	
  was	
  completed	
  between	
  March	
  18	
  
and	
   March	
   27,	
   2013	
   (nine	
   days).	
   	
   It	
   is	
  
important	
   to	
   note	
   that	
   since	
   the	
   health	
  
survey	
   was	
   considered	
   the	
   “entryway”	
   to	
  
each	
   family,	
   agriculture	
   surveys	
   were	
   only	
   conducted	
   after	
   the	
   health	
   survey	
   was	
   completed	
   and	
  
where	
   the	
   household	
   head	
  was	
  available	
   for	
   interview.	
   	
  As	
  a	
   result,	
   the	
   number	
   of	
   health	
   surveys	
  
conducted	
  exceeds	
  the	
  number	
  of	
  agriculture	
  surveys.
Qualitative
FGD and	
   KSI were	
   conducted	
   directly	
   by	
   the	
   Evaluation	
   Consultant based	
   on	
   interview	
   guides	
  
developed	
   in	
   alignment	
   with	
   the	
   ToR	
   and	
   reviewed	
   /	
   approved by	
   FH	
   staff.	
   An	
   FH	
   translator	
  
supported	
   the	
   FGDs by	
   translating	
   questions	
   from	
   English	
   to	
   Portuguese.	
   Local	
   field	
   staff	
  
subsequently	
  translated	
   from	
  Portuguese	
  into	
   the	
  respective	
  local	
  language.	
   Responses	
   to	
  KSI	
  and	
  
FGDs were	
   recorded	
   into	
   a	
   notebook	
   and	
   subsequently	
   transferred	
   into	
   electronic	
   transcripts.	
  	
  
Qualitative	
  data	
  collection	
  tools	
  and	
  transcripts	
  can	
  be found	
  in	
  Annexes 4 and	
  5.
Structured	
   observation	
   checklists	
  were	
   developed	
   using	
  FH	
   technical	
   specifications	
   for	
  HH	
  latrines,	
  
school	
   latrines	
   and	
   tube/shallow	
   wells 3 to	
   support	
  
collection	
   of	
   systematic,	
   objective	
   and	
   quantifiable
observations	
   regarding	
   use	
  and	
  maintenance	
   of	
  key	
  water	
  
and	
   sanitation-­‐related	
   infrastructure.	
   	
   These	
   tools	
   were	
  
reviewed	
   and	
   modified	
   and	
   subsequently	
   translated	
   into	
  
Portuguese.	
   Copies	
  of	
  the	
  checklists	
  can	
  be	
  found	
  in	
  Annex	
  
6. Observations	
  were	
  conducted	
  by	
  Data	
  Collection	
  Supervisors	
  during	
  HH	
  surveys.	
  	
  Copies	
  of	
  original
observations	
  checklists	
  were	
  stored	
  in	
  the	
  FH	
  office	
  in	
  Pemba.	
  
C.	
  Data	
  Quality,	
  Processing	
  and	
  Analysis
Quality	
  Control
To	
  ensure	
  the	
  highest	
  quality	
  data	
  possible	
  quality	
  controls	
  were	
  instituted	
  at	
  five levels	
  during	
  data	
  
collection	
  and	
  three levels	
  during	
  data	
  processing.
Five	
  levels	
  of	
  data	
  collection	
  quality	
  control:
1)	
   Enumerators	
   were	
   split	
   into	
   pairs	
   to	
   ensure	
   complete	
   and	
   accurate	
   data	
   collection	
   /	
   recording	
  
based	
  on	
  clear	
  segregation	
  of	
  duties;	
  
2)	
  One	
  supervisor	
  was	
  assigned	
  to	
  two	
  enumerator	
  pairs.	
  	
  Each	
  supervisor	
  was	
  required	
  to	
  observe	
  
each	
  pair	
  at	
  least	
  one	
   time	
  per	
  day	
  and	
  provide	
  clear	
   feedback	
  for	
  improvement	
  directly	
  following	
  
the	
  observed	
  interviews based	
  on	
  a	
  systematic	
  quality	
  control	
  checklist;
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   3 Provided	
  by	
  the	
  FH	
  Community	
  Capacity	
  Building	
  Manager
Table	
  2:	
  Total	
  surveys	
  completed	
  by	
  target	
  group
Health Agriculture
407 333
Table	
  3:	
  Total	
  Qualitative	
  Interviews	
  Conducted
Focus	
  Group	
  Discussions Staff	
  Interviews Key	
  stakeholder	
  interviews
§ 3 Farmer	
   Field	
   Life	
  
Groups
§ 3	
  Mother	
  Leader	
  Groups	
  
§ 2	
   Community	
   Development	
  
Associations
§ 2	
  FH	
  Agriculture	
  staff
§ 3	
  FH	
  Health	
  staff § 1	
   MOH	
   Provincial	
  
Nutritionist
Table	
  4:	
  Infrastructure	
  Observations
§53	
  HH	
  latrine	
  observations
§10	
  school	
  latrine	
  observations
§9	
  tube	
  /	
  shallow	
  well	
  observations
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 10
3)	
  Supervisors	
  reviewed	
  each	
  of	
  the	
  forms of	
  their	
  respective	
  teams	
  for	
  accuracy	
  and	
  completeness	
  
prior	
  to	
  submission	
  to	
  the	
  Team	
  Leader;
4)	
   Team	
   Leaders	
  reviewed	
   all	
   forms for	
   accuracy	
   and	
   completeness	
   prior	
   to	
   submission	
  to	
   the	
   FH	
  
Mocimboa	
  de	
  Praia	
  office	
  for	
  data	
  entry;	
  and
5)	
  Forms	
  were reviewed	
  each	
  day prior	
  to	
  data	
  entry and	
  clarifications	
  or	
  corrections	
  were	
  requested	
  
as	
   needed.	
   	
   Overall	
   trends	
   from	
   this	
   review	
   were	
   also	
   fed	
   back	
   to	
   the	
   Team	
   Leaders	
   to	
   support	
  
improvement	
  and	
  standardization.
Three levels	
  of	
  data	
  processing	
  quality	
  control:
1)	
  Daily	
  on-­‐site	
  mentoring	
  and	
  oversight	
  to	
  address	
  arising	
  concerns /	
  challenges;
2)	
  The	
  data	
  entry	
   team	
  maintained a	
  daily	
  log	
  of	
   form	
  numbers	
  and	
  identified	
  challenges.	
  	
  The	
  log	
  
was	
   used	
   to	
   re-­‐code	
   individual	
   questions	
   or	
   re-­‐enter	
   entire	
   questionnaires	
   (depending	
   on	
   the	
  
magnitude	
   of	
   challenges	
   identified)	
  each	
   afternoon.	
  In	
  extreme	
   cases	
   (e.g.,	
  multiple	
   key	
   questions	
  
with	
  errors)	
  questionnaires	
  were	
  sent	
  back	
  to	
  the	
  field	
  for	
  additional	
  clarifications;	
  and
3)	
  A	
  sample	
  of	
  15%	
  of	
  all	
  questionnaires	
  was	
  re-­‐entered	
  (double	
  entry) after	
  the	
  initial	
  steps	
  of	
  data	
  
entry	
  and	
  correction	
  were	
  complete.	
  
Data	
  Processing	
  and	
  Analysis
Five	
   data	
   entry	
   clerks	
   (three	
   Peace	
   Corps	
   volunteers	
   and	
   two	
   local	
   FH	
   staff)	
   were	
   oriented	
   on	
  
quantitative	
   data entry	
  with	
   a	
   user-­‐friendly	
  data	
  template	
   in	
   CSPro5.0.	
   The	
   template	
  mirrored	
   the	
  
questionnaire	
  to	
  facilitate entry	
  and	
  to	
  reduce	
  error. Beginning	
  from	
  the	
  third	
  day	
  of	
  data	
  collection,	
  
data	
  entry	
  was	
  completed	
  simultaneous	
  to	
  data	
  collection.	
  	
  Data	
  entry was	
  completed	
  by	
  March	
  30th
2013.	
  Original	
  forms	
  were	
  stored	
  in	
  numerical	
  order	
  according	
  to	
  questionnaire	
  IDs	
  in	
  the	
  FH	
  office	
  in	
  
Pemba.	
  	
  Data	
  cleaning	
  and	
  verification	
  was	
  conducted	
  during	
  three	
  days	
  in	
  Pemba by	
  running	
  basic	
  
frequencies	
   and	
   crosstabs	
   and	
   crosschecking /	
   correcting	
   errors	
   by	
   referring	
   back	
   to	
   the	
   original	
  
forms.	
   Preliminary	
   quantitative	
   analysis	
   was	
   conducted	
   in	
   CSPro	
   to	
   generate	
   basic	
   descriptive	
  
statistics.	
  	
  
Qualitative	
   data	
   trends	
   were	
   identified	
   from	
   FGDs	
   and	
   KSI	
   by	
   organizing	
   the	
   qualitative	
   data	
   into	
  
themes.	
   	
   Observations	
   from	
   the	
   infrastructure	
   checklists	
   were	
   tabulated	
   and	
   processed	
   in	
   Excel.	
  	
  
Together	
   the	
   quantitative	
   and	
   qualitative data	
   was used	
   to	
   provide a	
   preliminary	
   debriefing	
   for	
  
Senior	
  FH	
   staff	
  and	
   stakeholders	
   from	
   the	
  MoH and	
  GoM	
  Office	
   for	
  Infrastructure	
  Development	
  in	
  
Pemba.	
  	
  Staff	
  from	
  the	
  USAID	
  Mission	
  in	
  Maputo	
  was	
  invited	
  to	
  the	
  debriefing;	
  however,	
  they	
  were	
  
unable	
   to	
  attend.	
   The	
  debriefing	
  in	
  Pemba	
  provided	
  an	
  opportunity	
   to	
   receive	
   feedback	
   regarding	
  
the	
  preliminary	
   findings	
   from	
  both	
  FH	
  staff	
  and	
   the	
  external	
  stakeholders	
  present.	
  	
  	
  This	
   feedback,	
  
especially	
  regarding	
  context	
  factors,	
  has	
  also	
  been	
  incorporated	
  into	
  this	
  report.	
  	
  
Final	
  quantitative	
  (comparative)	
  analysis	
  of	
  baseline	
  and	
  final	
  evaluation	
  data was	
  conducted	
  in	
  SPSS	
  
21.0.	
  Relevant	
  proportions	
  and	
  means	
  were	
  generated	
  for	
  each	
  indicator	
  and	
  t-­‐tests	
  for	
  independent	
  
samples	
   were	
   conducted	
   to	
   test	
   the	
   significance	
   of	
   differences	
   between	
   the	
   baseline	
   and	
   final	
  
evaluations.	
  	
  The	
  p	
  values	
  (statistical	
  significance)	
  of	
  each	
  of	
  the	
  findings	
  were	
  subsequently	
  verified	
  
using	
  the	
  Indicator	
  Confidence	
  Interval	
  Calculation	
  Worksheet	
  provided	
  by	
  FH	
  for	
  that	
  purpose.
IV.	
  Constraints	
  and	
  limitations
Every	
  effort	
  was	
  made	
  to reduce	
  bias,	
  ensure	
  data	
  fidelity	
  and	
  integrity of	
  the	
  information	
  presented	
  
in	
  this	
  report.	
  	
  Nonetheless-­‐ as	
  with	
  the	
  results	
  of	
  any	
  data	
  collection	
  and	
  analysis	
  process-­‐ some	
  key	
  
limitations	
  should	
  be	
  kept	
  in	
  mind	
  while	
  reading	
  this	
  report:
§ No	
   use	
   of	
   a	
   control	
   group.	
   The	
   data	
   that	
   is	
   presented	
   does	
   not	
   compare	
   the	
   results	
   of	
  
intervention	
   groups	
   with	
   control	
   groups,	
   thus	
   limiting	
   the	
   degree	
   of	
   inference	
   regarding	
  
causality.	
   What	
   the	
   analysis	
   does	
   measure	
   is	
   the	
   degree	
   to	
   which	
   changes	
   between	
   baseline	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 11
values	
   and	
   final	
   evaluation	
   values	
   are	
   statistically	
   significant.	
   	
   While	
   some	
   effort	
   was	
   made	
  
through	
   the	
  qualitative	
  work	
   to	
  identify	
  confounding	
   factors	
  outside	
  of	
  FH	
   support	
   that	
  might	
  
have	
  contributed	
   to	
   these	
  changes,	
   those	
  discussions	
   (and	
  by	
  proxy	
  discussions	
  of	
  attribution)	
  
are	
  limited.
§ Purposive	
   exclusion	
   of	
   intervention	
   areas	
   from	
   the	
   sample:	
   The	
   city	
   centers	
   of	
   Mocimboa	
   da	
  
Praia	
  and	
   Palma	
  were	
  excluded	
   from	
   the	
   sample based	
   on	
  a	
   consensus	
   decision	
   taken	
   by	
   the	
  
evaluation	
   management	
   team.	
   	
   This decision	
   obeyed	
   two	
   key	
   factors:	
   	
   (1)	
   the	
   centers	
   were	
  
excluded	
   in	
   the	
   baseline	
   sample;	
   and	
   (2)	
   although	
   FH	
   did	
   implement	
   MYAP	
   activities	
   in	
  
Mocimboa city	
   center,	
   only	
   health	
   interventions	
   were	
   undertaken	
   there.	
   Given	
   the	
   high	
  
population	
  density	
  in	
  Mocimboa	
  (representing	
  a	
  full	
  40%	
  of	
  all	
  of	
  the	
  MYAP	
  health	
  participants)	
  
it	
  would	
  have	
  rendered	
  the	
  parallel	
  sampling	
  technique	
  virtually	
  moot,	
  elevating	
  the	
  evaluation	
  
logistics	
   costs	
   and	
   increasing	
   the	
   time	
   required	
   for	
   the	
   exercise.	
   	
   Based	
   on	
   these	
   factors	
   the	
  
decision	
   was	
   taken	
   to	
   only	
   include	
   areas	
   with	
   both	
   health	
   and	
   agriculture	
   activities	
   in	
   the	
  
sample.	
   	
   Likewise, three	
   locations	
   with	
   extremely	
   difficult	
   access	
   during	
   the	
   rainy	
   season	
  
(Maculo	
   and	
   Ulo	
   in	
   Mocimboa	
   da	
   Praia	
   and	
   Lalane	
   in	
   Palma)	
   were	
   also	
   excluded	
   in	
   order	
   to	
  
facilitate	
  data	
  collection	
  and	
  efficiency	
  of	
  resource	
  use.	
  	
  Together	
  these	
  five locations	
  represent	
  
66,148	
   inhabitants	
   of	
   the	
   total	
   population	
   of	
   141,709	
   of	
   the	
   50	
   MYAP	
   intervention	
   areas,	
   or	
  
almost	
  47%	
  of	
  the	
  total	
  population.	
  In	
  other	
  words,	
  the	
  sample-­‐ and	
  the	
  statistics	
  derived	
  from	
  
it-­‐ represent	
  only	
  slightly	
  more	
  than	
  half	
  of	
  the	
  full	
  intervention	
  population.
§ Purposive	
   exclusion	
   of	
   a	
   large	
   proportion	
   of	
   health	
   participants from	
   the	
   sample.	
   	
   Similarly,	
  
Mocimboa	
   da	
   Praia	
   city	
   center	
   represents	
   a	
   full	
   40%	
   of	
   the	
   total	
   MYAP	
   health	
   participant	
  
population,	
  which	
  means	
  that	
  almost	
  half	
  of	
  the	
  health	
  participant	
  population	
  was	
  purposively	
  
excluded	
  from	
  the	
  sample.	
  	
  These	
  limitations	
  should	
  be	
  kept	
  in	
  mind	
  when	
  reading	
  the	
  findings	
  
and	
  conclusions	
  of	
  the	
  report.	
  	
  
§ Challenges	
  with	
  using	
  FH	
  staff	
  for	
  data	
  collection.	
  Survey	
  data	
  was	
  collected	
  by	
  FH	
  staff,	
  which	
  
does	
   potentially	
   introduce some	
   bias into	
   the	
   data.	
   	
   In	
   order	
   to	
   reduce	
   such	
   bias,	
   staff	
   was	
  
assigned	
   to	
   collect	
   data	
   in	
   locations	
   where	
   they	
   do	
   not	
   normally	
   work.	
   	
   Another	
   challenge	
   in	
  
using	
   FH	
   staff	
   was	
   that	
   almost	
   half	
   of	
   staff	
   assigned	
   to	
   the	
   data	
   collection	
   task	
   had	
   never	
  
participated	
  in	
  data	
  collection	
  previously.
4
	
  	
  Although	
   four	
   full	
  days	
  were	
  spent	
  on	
   training,	
   the	
  
lack	
  of	
  previous	
  experience	
  and	
  overall	
  limited	
  understanding	
  of	
  key	
  survey	
  concepts	
  (e.g.,	
  the	
  
need	
   for	
  precision,	
  an	
  orientation	
  to	
  detail)	
  led	
  to	
  a	
  large	
  learning	
  curve,	
  persistent	
  challenges	
  
and	
   the	
   need	
   for	
   consistent	
   and	
   very	
   close	
   follow-­‐up.	
   Most	
   challenges	
   were	
   addressed	
   by	
  
sending	
   forms	
   back	
   for	
  completion	
  or	
  clarification.	
  However	
   the	
  process	
  of	
  gathering	
  multiple	
  
corrections	
  and	
  clarifications	
  could	
  have	
  introduced	
  slight	
  bias	
  into	
  the	
  results.	
  	
  Approximately	
  
15-­‐20%	
   of	
   overall	
   survey	
   forms	
   required	
   clarifications	
   or corrections. Three	
   forms	
   had	
   to	
   be	
  
completely	
  discarded	
  as	
  a	
  result	
  of	
  these	
  factors.	
  
§ Limitations	
  in	
  survey	
  instrument.	
   There	
  were	
  several	
  questions	
  on	
   the	
  agricultural	
  survey	
   (the	
  
entire	
   battery	
   of	
   questions	
   for	
   sections	
   C,	
   D	
   and	
   F)	
   that	
   sought	
   detailed	
   agricultural	
   yield,	
  
livelihoods	
  and	
  economic	
  data	
  from	
  participants	
  based	
  on	
  participant	
  recall.	
  	
   Participant	
  recall	
  
becomes	
   less	
   and	
   less	
   reliable	
   the	
   further	
   away	
   you	
   get	
   from	
   the	
   event	
   under	
   scrutiny	
   (e.g.,	
  
farmers	
  have	
  a	
  hard	
  time	
  remembering	
  precise	
  yields	
  the	
  further	
  they	
  are	
  from	
  the	
  last	
  harvest).	
  	
  
The	
  decision	
  was	
  made	
  not to	
  modify	
  the	
  questionnaire	
  significantly	
  for	
  these	
  questions	
  despite	
  
the	
   recognition	
   that	
   the	
   data	
   would	
   be	
   of	
   limited	
   use	
   for	
   anything	
   other	
   than	
   illustrative	
  
purposes.	
  
§ Simultaneous	
   qualitative	
   and	
   quantitative	
   data	
   collection.	
  In	
   line	
  with	
   the	
  ToR,	
   the	
  evaluation	
  
team	
  undertook	
  the	
  quantitative	
  and	
  qualitative	
  data	
  collection	
  activities	
  simultaneously.	
  	
  	
  This	
  
decision	
   obeyed	
   financial	
   and	
   time	
   constraints	
   and	
   represented	
   a practical	
   and	
   cost	
   efficient	
  
approach	
   to	
   the	
   evaluation	
   exercise.	
   Under	
   ideal	
   circumstances,	
   qualitative	
   and	
   quantitative	
  
data	
  collection	
  should	
  be	
  undertaken	
  as	
  separate	
  steps	
   to	
  ensure	
   that	
  any	
   findings	
  uncovered	
  
from	
   the	
  quantitative	
  exercise	
  can	
  be	
  “unpacked”	
  and	
  more	
   fully	
  explored	
   through	
  qualitative	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   4 Based	
  on	
  a	
  rapid	
  survey	
  of	
  data	
  collection	
  staff	
  conducted	
  during	
  enumerator	
  training.	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 12
inquiry.	
   	
   Because	
   in	
   this	
   case	
   quantitative/qualitative	
   data	
   collection	
   and	
   analysis	
   occurred	
  
simultaneously,	
  some	
  gaps	
  remain	
  that	
  could	
  not	
  be	
  adequately	
  addressed.	
  
§ Language	
   barriers	
   in	
   data	
   collection.	
   	
   Although	
   every	
   effort	
   was	
   made	
   to	
   translate	
   the	
   data	
  
collection	
  instruments	
  into	
  local	
  languages	
  (survey)	
  and	
  Portuguese	
  (qualitative	
  tools),	
  language	
  
barriers	
  did	
  arise	
  during	
   the	
  process.	
  	
  The	
  evaluation	
  leader	
  spoke	
  basic	
  Portuguese,	
  but	
  FGDs	
  
and	
  responses	
  required double	
  translation:	
  from	
  English	
  to	
  Portuguese	
  to	
  local	
  languages,	
  back	
  
into	
  Portuguese	
  and	
  English.	
  	
  Every	
  effort	
  was	
  made	
  to	
  clarify	
  concepts;	
  however, nuance	
  in	
  the	
  
FGDs	
  was	
  likely	
  lost	
  as	
  a	
  result	
  of	
  these	
  factors.	
  	
  At	
  the	
  same	
  time,	
  in	
  some	
  locations	
  participants	
  
spoke	
  none	
  of	
  the	
  three	
  foreseen	
  languages	
  for	
  which	
  the	
  survey	
  had	
  been	
  translated.	
  	
  In	
  these	
  
cases the	
  languages	
  could	
  be	
  spoken	
  and	
  understood,	
  but	
  translations	
  of	
  the	
  survey	
  form	
  were	
  
conducted	
  on	
   the	
  spot.	
  	
  This	
  could	
  have	
  led	
   to	
  misinterpretation	
  and	
  misunderstanding	
  of	
   the	
  
questions.	
   The	
   estimated	
   proportion	
   of	
   these	
   occurrences	
   was	
   small	
   (<10%	
   of	
   total); it	
   is	
  
nonetheless	
  important	
  to	
  keep	
  in	
  mind.
§ Lack of	
  original	
   (raw)	
  baseline	
  data.	
  It	
  is	
  important	
   to	
  note	
   that	
   the	
  original	
  baseline	
  data	
  was	
  
not	
  available at	
   the	
   time	
  of	
  analysis as	
  a	
   result	
  of	
  significant	
  staff	
  overturn	
  at	
  FH	
  between	
   the	
  
baseline	
   and	
   final	
  evaluations.	
  	
   Instead,	
   the	
   consultant	
   used	
   a	
   version	
   of	
   the	
  baseline	
  data	
   in	
  
Excel	
   that	
   included dummy	
   (composite)	
   variables	
   for	
   key	
   indicator	
   values	
   such	
   as exclusive	
  
breastfeeding,	
   hygiene	
   behaviors	
   and	
   signs	
   of	
   childhood	
   illness.	
   	
   The	
   original	
   data	
   used	
   to	
  
develop these	
  composite	
  variables	
  was	
  not	
  available,	
  limiting	
  the	
  possibility	
  for	
  verification.	
  	
  At	
  
the	
  same time,	
  no	
  data	
  dictionary	
  was	
  available	
  for	
  the consolidated	
  baseline	
  dataset leading	
  to	
  
questions	
  about	
   how	
   the	
  values	
  were	
  generated.	
   However,	
  the	
  FH	
   health	
   team	
  provided	
   the	
  
consultant	
  with	
  a	
  list	
  of	
  the	
  variables	
  from	
  which	
  each	
  composite value	
  was consolidated.	
  	
  This	
  
allowed	
   for	
   the reproduction	
   of	
   those	
   values	
   in	
   the	
   final	
   evaluation	
   dataset,	
   thus	
   facilitating	
  
comparative	
  analysis	
  in	
  SPSS.	
  	
  While	
  the	
  baseline	
  values	
  generated	
  in	
  SPSS	
  for	
  the	
  purposes	
  of	
  
this	
  analysis	
  correspond	
  in	
  the	
  large	
  majority	
  of	
  cases	
  to	
  those	
  included	
  in	
  the	
  original	
  baseline	
  
report	
   and	
   IPTT,	
   there	
   are	
   a	
   couple	
   of	
   indicator	
   values	
   that	
   do	
   not.	
   	
   In	
   these	
   cases	
   tests	
   of	
  
statistical	
  significance	
  were	
  conducted	
  in	
  two	
  ways:	
  (1)	
  by	
  conducting	
  t-­‐tests	
  of	
  the	
  differences	
  
between the new	
  baseline	
  values	
  with	
  the	
  final	
  evaluation values	
  in	
  SPSS; and	
  (2)	
  by	
  comparing	
  
the	
  confidence	
  intervals	
  of	
  the	
  original	
  baseline	
  values	
  with	
  the	
  values	
  of	
  the	
  final	
  survey.	
  
V.	
  Findings-­‐ Organized	
  by	
  Results	
  Framework	
  /	
  Indicators
A.	
  Strategic	
  Objective	
  1	
  (Agriculture):	
  Protect	
  and	
  enhance	
  livelihoods	
  of	
  vulnerable	
  farming	
  
households
Under	
   the	
   agricultural	
   component,	
   FH	
   proposed	
   to	
   protect	
   and	
   enhance	
   farming	
   livelihoods	
   by	
  
improving food	
   availability	
   and	
   access.	
   	
   FH	
   proposed	
   to	
   achieve	
   this by	
   (1)	
   increasing	
   agricultural	
  
productivity;	
   (2)	
   decreasing vulnerability	
   to	
   shocks	
   through	
   crop	
   diversification	
   and	
   improved	
  
storage;	
   and	
   (3)	
   increasing purchasing	
   power	
   through	
   cash	
   crop	
   production,	
   engagement	
   with	
  
markets	
  and	
  development	
  of	
  VSLAs.	
  	
  	
  Findings	
  for	
  each	
  of	
  the	
  indicators	
  relating	
  to	
  SO1	
  are	
  included	
  
below,	
  organized	
  by	
  Intermediate	
  Result.	
  	
  	
  	
  Overall	
  impact	
  indicators	
  for	
  SO	
  are	
  discussed	
  at	
  the	
  end	
  
of	
  this	
  section.
Intermediate	
  Result	
  1.1:	
  Agricultural	
  productivity	
  and	
  production	
  increased	
  and	
  diversified
FH	
  proposed	
  food	
  source	
  diversification	
  as	
  a	
  key	
  strategy	
  to	
  improve	
  food	
  access	
  and	
  availability in	
  
the	
  target	
  area.	
  To	
  that	
  end,	
  FH	
  proposed:	
   (1)	
  promotion	
  of	
  protein and	
  micronutrient-­‐rich	
  perennial	
  
dark	
  green	
  leaf	
  crop	
  production	
   (moringa,	
  cassava)	
   to	
  help	
  meet	
  demands	
   for	
  dietary	
  diversity;	
   (2)
the	
  promotion	
  of	
  several	
  locally	
  available	
  high-­‐protein,	
  high-­‐nitrogen	
  fixing	
  (leguminous)	
  crops;	
  and	
  
(3)	
  the	
  introduction	
  of	
  sustainable	
  agricultural	
  technologies	
  to	
  increase	
  production	
  and	
  improve	
  soil	
  
quality.	
   The	
  relevant	
  indicators	
  and	
  targets	
  for	
  these	
  practices	
  are	
  included	
  in	
  Table	
  5.
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 13
Table	
  5:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR1.1
Percent	
  of	
  HH	
  with	
  intensive	
  dark	
  green	
  leaf	
  production 95%
Percent	
   of	
   HH	
   adopting	
   and	
   diversifying	
   production	
   by	
   including	
   at	
   least	
   1	
   new	
  
leguminous	
  nutritious	
  food	
  crop	
  in	
  farming	
  system 90%
Percent	
   of	
   beneficiaries	
   (farmers)	
   using	
   at	
   least	
   3	
   sustainable	
   agricultural	
  
technologies	
  in the	
  past	
  year 70%
Percent	
   of	
   fruit	
   and	
   cashew	
   farming	
   training	
   beneficiaries	
   practicing	
   at	
   least	
   two	
  
improved	
  fruit	
  growing	
  techniques	
  in	
  the	
  past	
  year 60%
Percent	
  of	
  HH	
  adopting	
  sustainable	
  non	
  chemical	
  grain	
  and	
  seed	
  storage	
  protection	
  
techniques 90%
Productivity	
  (kg/ha)	
  for:	
  
(1)	
  Sesame	
  
(2)	
  Groundnut	
  
(3)	
  cow	
  peas	
  
(4)	
  rice	
  (Nerica)
Sesame:	
  500
Groundnut:	
  500
Cow	
  peas:	
  550
Rice:	
  700
FH	
   promoted protein	
   and	
   micronutrient-­‐rich	
  
perennial	
   dark	
   green	
   leaf	
   crop	
   production	
   to
help	
   meet	
   demands	
   for	
   dietary	
   diversity and	
  
support	
   the	
   MYAP’s	
   health	
   objective.	
   Dark	
  
green	
   leaves	
   from	
   moringa	
   and	
   cassava	
   are	
  
available	
   even	
   at	
   the	
   driest	
   time	
   of	
   the	
   year	
  
when	
   little	
   else	
   is	
   growing.	
   Moringa	
   in	
  
particular	
   is	
   high	
   in	
   protein,	
   Vitamin	
   “A”	
   and	
  
“C”,	
   potassium,	
   calcium	
   and	
   other	
   minerals.	
  	
  
Figure	
   1 shows	
   that production	
   of	
   dark	
   green	
  
leaf	
  increased	
  at	
  a	
  statistically	
  significant	
  level	
  
over	
   the	
   baseline,	
   from	
   71.7%	
   to	
   91%	
   (LCL	
  
89.7%	
   /	
   UCL	
   94.1%).	
  	
   Given	
   that	
   84%	
   of	
   HH	
  
produced	
  cassava (LCL	
  80.2%	
   /	
  UCL	
  88%),	
  it	
  is	
  
not	
   surprising	
   that	
   the	
  level	
  of	
  HH	
  with	
  green	
  
leaf	
  production	
  is	
  high.	
  
FH	
   also	
   promoted the	
   production	
   of	
   several	
  
leguminous	
  crops	
  that	
  are	
  both	
  high	
  in	
  protein (to	
  
support	
  health	
  objectives	
  under	
  SO2)	
  and	
  nitrogen	
  
content	
   (to	
   support	
   improved	
   soil).	
   	
   These	
   are	
  
drought	
   tolerant	
   crops,	
  appropriate	
   for	
  the	
   sandy
soils	
  of	
  the	
  intervention	
  districts,	
  and according	
  to	
  
FH	
  agriculture	
  staff	
  they have	
  low	
  labor	
  and	
  input	
  
requirements,	
   making	
   them	
   relatively	
   easy	
   to	
  
adopt.	
   	
   As	
   Figure	
   2 illustrates,	
   the	
   proportion	
   of	
  
HH	
   incorporating	
   at	
   least	
   one	
   leguminous	
   crop	
  
into	
   their	
   farming	
   systems	
   has	
   increased	
   by	
  
approximately	
   20 percentage	
   points over	
   the	
  
baseline,	
   from	
   37.3%	
   to	
   58%	
   (LCL	
   52.7%	
   / UCL	
  
63.3%).	
   This	
   change,	
   while	
   statistically	
   significant,	
  
is	
   approximately	
   30 percentage	
   points below	
   the	
  
program	
  target.
Baseline	
  N=	
  300;	
  Final	
  N=	
  333
p=	
  .000
Baseline	
  N=	
  300;	
  Final	
  N=	
  333
p=	
  .000
	
  	
  	
  	
  71.7%	
  
91%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  1:	
  Percent	
  of	
  HH	
  with	
  
intensive	
  dark	
  green	
  leaf	
  produc]on	
  
Target	
  95%	
  
37.3%	
  
58%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  2:	
  Percent	
  of	
  HH	
  plan]ng	
  at	
  
least	
  one	
  leguminous	
  crop	
  
Target	
  90%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 14
It	
  is	
  important	
  to	
  note	
  that,	
  despite	
  an	
  increase	
  
in	
   the	
   proportion	
   of	
   families	
   incorporating	
  
legumes	
  into	
  their	
  HH	
  production,	
  there	
  has	
  not	
  
been	
   a	
   statistically	
   significant	
   increase	
   in	
   the	
  
production	
   of	
   any	
   single	
   leguminous	
   crop since	
  
the	
  baseline.	
  	
  As	
  Table	
  6 (left)	
  demonstrates,	
  the	
  
overall	
  proportion	
  of	
  families	
  planting	
  any	
  single	
  
legume	
  has	
  remained	
  the	
  same	
  over	
  time.	
  	
  One	
  
potential	
   factor	
   contributing	
   to	
   the	
   lack	
   of	
  
increase	
   may	
   be	
   that	
   FH	
   stopped	
   distributing	
  
free	
   legume	
   seeds	
   in	
   2011.	
   	
   Seeds	
   were	
  
distributed	
  to	
  participants	
  in	
  the	
  first	
  three	
  years	
  of	
  the	
  program only.
5
To	
  support	
  increased	
  diversity	
  in	
  production	
  and	
  cash crop	
  potential	
  (covered	
  in	
  greater	
  detail	
  under	
  
IR1.3,	
   below),	
   FH	
   supported	
   improved	
   practices	
   for	
   fruit	
   tree	
   and	
   cashew	
   production.	
   	
   Fruit	
   tree	
  
production	
  was	
  not	
  a	
  viable	
  production	
  alternative	
  as	
  a	
  result	
  of	
  the	
  sandy	
  soil	
  and	
  coastal	
  climate	
  of	
  
the	
  intervention	
  areas.6
	
  	
  Support	
  was	
  provided	
   for	
  improved	
  cashew	
  production.	
  In	
  2010	
  and	
  2011	
  
FH	
   distributed	
   cashew	
   seedlings	
   to	
   1,300	
   farmers.	
  They	
  also	
   received	
  intensive training	
   on	
   cashew	
  
production	
   techniques,	
  including	
  land	
  preparation,	
  mulching,	
  spacing and	
  disease	
  and	
  pest	
  control.
According	
   to	
   the	
   2012	
   FH	
   Annual	
   Agricultural	
   Survey,	
   73%	
   of	
   FH	
   beneficiaries	
   involved	
   in cashew	
  
production	
   practiced at	
   least	
  two	
   improved	
   techniques	
   in	
   the	
   year	
   prior	
   to	
   the	
   survey.7 This	
   data	
  
could	
  not	
  be	
  verified,	
  as	
  the consultant	
  did	
  not	
  have	
  access	
  to	
  the	
  original	
  dataset	
  from	
  the	
  annual	
  
survey.	
  	
  
To	
  conserve	
  and	
  build	
  organic	
  matter	
  in	
  the	
  soil	
  
and	
   increase	
   water	
   infiltration	
   and	
   retention,	
  
FH	
   promoted	
   several	
   sustainable	
   technologies,	
  
including the	
   use	
   of	
   green	
   manures	
   and	
   cover	
  
crops,	
  compost,	
  crop	
   rotation,	
  intercropping	
  of	
  
grains	
   with nitrogen-­‐fixing	
   legumes as	
   well	
   as	
  
agroforestry.	
   As	
   Figure	
   3 demonstrates,	
   the	
  
proportion	
   of	
   HH	
   using	
   at	
   least	
   three	
  
sustainable	
   technologies	
   has	
   increased	
  
significantly	
   over	
   the	
   baseline,	
   from	
   32.7%	
   to	
  
67%	
  (LCL	
  61.9%	
  /	
  UCL	
  72%).	
  	
  This	
  is	
  only	
  slightly	
  
lower	
  than	
  the	
  program	
  target.
We	
  can	
  see	
  in	
  Table	
  7 (below)	
  that	
  technology
use	
   has	
   increased	
   across	
   four technology	
  
categories.	
  Adoption	
  of	
  composting,	
  intercropping,	
  
crop	
   rotation	
   and	
   reduced	
   burning	
   have	
   all	
  
increased	
  at	
  a	
  statistically	
  significant	
  level.	
  	
  	
  Most	
  of	
  
these	
   increases	
   have	
   been	
   by	
   at	
   least	
   20
percentage	
   points.	
   	
   Reduced	
   burning	
   has	
   seen	
   the	
  
greatest	
   change,	
   at	
   almost	
   50 percentage	
   points.	
  	
  
There	
  was	
  no	
  change	
  in	
  adoption	
  of	
  green	
  manure	
  
or	
  agroforestry. Farmers	
  in	
  FGD	
  also	
  mentioned	
  line	
  
sowing	
   and	
   crop	
   spacing	
   as	
   key	
   technologies	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   5 Food	
  for	
  the	
  Hungry	
  Mozambique	
  2012	
  Annual	
  Agriculture	
  Survey	
  report 6 Interview	
  with	
  FH	
  Agriculture	
  Manager
7 Unlike	
   the	
   other	
   data	
   in	
   this	
   report,	
   this	
   figure	
   comes	
   from	
   the	
   FH	
   Annual	
   Agriculture	
   Survey	
  and	
   reflects	
   only	
   project	
  
participants.	
  	
  No	
  specific	
  question	
  was	
  included	
  in	
  the	
  survey	
  regarding	
  cashew	
  production	
  techniques
Table	
  6:	
  Legume	
  production
Leguminous	
  
variety BL Final Sig.
Cow	
  pea 32% 39.3% N
Jugo	
  bean 10% 11.7% N
Soloco	
  bean 0% 0.3% N
Boer	
  bean 1% 2.4% N
Other	
  beans 1% 0.9% N
Groundnut 15% 12.3% N
Cashew 25% 29.7% N
Baseline	
  N=	
  300;	
  Final	
  N=	
  333
p=	
  .000
Table	
  7:	
  Technology	
  adoption
Technology BL Final Sig.
Green	
  Manure 39.3% 48.6% N
Composting 32.0% 61.3% Y
Intercropping 42.7% 64.6% Y
Agroforestry 16.7% 23.7% N
Crop	
  rotation 17.7% 52.0% Y
Reduced	
  
burning 30.3% 79.0%
Y
32.7%	
  
67%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  3:	
  Percent	
  of	
  HH	
  using	
  at	
  least	
  
3	
  sustainable	
  technologies	
  
Target	
  70%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 15
introduced	
  by	
  the	
  project	
  that	
  have	
  helped	
  them	
  increase	
  yields.
FH	
   introduced	
   the	
   use	
   of	
   neem	
   leaves	
   and	
   mango	
   ash	
  to	
   protect	
   seed	
   and	
   grain	
   during storage.	
  
According	
   to	
   the	
  2012	
  Annual	
  Agricultural	
  Survey,	
  63%	
  of	
  project	
  participants	
  used	
  neem	
  and	
  63%
used	
  mango	
  ash.	
  	
  In	
   two	
  FGDs with	
  members	
   of	
   the	
  Farmer	
  Field	
  and	
   Life	
  Groups	
   (FFLG),	
   farmers	
  
expressed	
   frustration	
  with	
  organic	
  pesticides,	
  which	
  they	
  perceived	
  as	
  working	
  poorly.	
  	
  In	
  one	
  case	
  
the	
   farmers	
  explained	
  that	
  FH	
  had	
  provided	
  misinformation	
  regarding	
  a	
  mistaken	
  variety	
  of	
  tree	
   (it	
  
looked	
   like	
   neem,	
   but	
   was	
   actually	
   another-­‐ unrelated-­‐ tree	
   variety).	
   	
   Logically	
   it	
   did	
   not	
   work	
  
because	
  it	
  did	
  not	
  have	
  the	
  properties	
  of	
  an	
  organic	
  pesticide.	
  	
  In	
  the	
  other	
  case	
  farmers	
  said	
  that	
  the	
  
neem mix	
  did	
  not	
  work	
  for	
  root	
  plants,	
  explaining that	
  pests	
  had	
  eaten	
  both	
  cassava	
  and	
  groundnut
production	
  treated	
  with	
   the	
   remedy.	
  	
  Although	
   the	
   farmers	
  in	
   the	
  second	
  group could	
  not	
  provide	
  
exact	
  figures,	
  they	
  suggested	
  that	
  this	
  had	
  led	
  to	
  large	
  losses.	
  	
  
According	
  to	
  FH	
  staff,	
  the	
  error	
  of	
  the	
  first	
  case	
  was	
  resolved	
  as	
  soon	
  as	
  the	
  error	
  was	
  discovered.	
  	
  In	
  
the	
  second	
  case	
  FH	
  staff	
  explained	
   that	
  neem	
  takes	
  as	
  many	
  as	
   three	
  years	
  to	
  produce	
  seeds	
  with	
  
strong	
  insecticidal	
  properties.	
   Although	
  the	
  leaves	
  can	
  also	
  be	
  used	
  as	
  an	
  insecticide,	
  they	
  are	
  not	
  as	
  
potent	
  as	
  an	
  insecticide	
  as	
  the	
  seeds.	
  In	
  either	
  form	
  neem	
  is	
  an	
  effective	
  preventive	
  measure when	
  
used	
  in	
  conjunction	
  with	
  other	
  methods,	
  but	
  alone	
  it	
  does	
  not	
  kills	
  all insects. 8
	
  	
  Given	
  the	
  limitations	
  
of	
  neem	
  and	
  the	
  challenges	
  with	
  pest	
  infestation in	
  the	
  MYAP	
  program	
  areas	
  it	
  is	
  unclear	
  why	
  FH did	
  
not	
  promote other pesticide	
  alternatives in the	
  program.	
  When	
  asked,	
  FH	
  staff	
  explained	
  that	
  USAID	
  
does	
   not	
  allow	
   the	
   purchase	
   of	
   chemical	
   pesticides.	
  However,	
   this	
  alone	
   seems	
  like	
  an	
  insufficient	
  
explanation.	
  	
  It	
  is	
  important	
  to	
  note	
  that	
  farmers’	
  willingness	
  to	
  adopt	
  new	
  technologies	
  will	
  depend	
  
largely	
   on	
   whether	
   their	
   own	
   experience with	
   such	
   technologies is	
   positive.	
   	
   A	
   disappointing	
  
experience	
   with	
   one	
   technology	
   could	
   make	
   farmers	
   skeptical	
   of	
   other	
   new	
   technologies	
   in	
   the	
  
future.	
  
One	
  of	
  the	
  key	
  objectives	
  of	
  the	
  support	
  to	
  agriculture	
  was	
  to	
  increase	
  production	
  and	
  productivity
for	
  key	
  crops.	
  	
  As	
  mentioned	
  in	
  the	
  limitations	
  section,	
  the	
  original	
  questionnaire	
  asked	
  respondents	
  
to	
   estimate	
   yield	
   and	
   sales	
   data	
   based	
   on	
   recall.	
   	
   Although	
   recall	
   can	
   be	
   an	
   effective	
   method	
   for	
  
estimating	
   yields	
   within	
   a	
   few	
   days	
   (or	
   even	
   weeks)	
   of	
   the	
   harvest,	
   recall	
   becomes	
   less	
   and	
   less	
  
reliable	
  with	
  the	
  passage	
  of	
  time.9
	
  	
  Given	
  that	
  the	
  harvest	
  to	
  which	
  the	
  survey	
  referred	
  took	
  place	
  in	
  
June	
   of	
   2012-­‐ a	
   full	
   8	
   months	
   prior	
   to	
   the	
   survey	
   exercise-­‐ the	
   specific	
   data	
   regarding	
   yields	
   is	
  
unreliable.	
  	
  A	
  more	
  accurate	
  way	
  to	
  look	
  at	
  performance	
  is to	
  look	
  at	
  overall	
  trends	
  from	
  FH’s	
  Annual
Agricultural	
   Surveys	
   which	
   used	
   LQAS	
   sampling	
   to	
   collect	
   crop	
   cuttings	
   and	
   extrapolate	
   mean	
  
productivity	
  among	
  sampled	
  FH	
  participants from	
  across	
  all	
  program	
  locations.	
  Data	
   from	
  both	
  the	
  
evaluation	
  survey	
  AND	
  the	
  FH	
  Annual	
  Agricultural	
  Surveys	
  are presented	
  in	
  Figures	
  4a	
  and	
  4b below.
We	
   can	
   see	
  in	
  Figure	
   4a	
  (below,	
   on	
   page	
   13)	
  that	
   the final	
  evaluation	
  survey	
  data	
   shows	
  negative	
  
productivity	
   growth for	
   all	
   crops,	
   except rice,	
   which	
  showed	
   slight	
   improvements.	
   At	
   least	
   four
factors	
  may be	
  influencing	
  these	
  results:
(1)	
  The	
  first,	
  mentioned	
  above,	
  has	
  to	
  do	
  with	
  the	
  unreliable	
  nature	
  of	
  farmer	
  recall	
  to	
  describe	
  
yield	
  data	
  so	
  many	
  months	
  after	
  the	
  harvest;	
  	
  
(2)	
   The	
   second	
   is	
   that	
   the	
   agricultural	
   component	
   might	
   be	
   better	
   served	
   by	
   direct	
  
measurement	
  of	
  agricultural	
  participants	
  rather	
  than	
  full	
  populations.	
  By	
  including	
  data	
  for	
  the	
  
full	
  population,	
  the	
  results	
  may	
  not	
  fully	
  capture	
  the	
  agricultural	
  growth	
  of	
  farmers;
(3) The	
   third	
   factor	
   has	
   to	
   do	
   with	
   the	
   challenges	
   in	
   effectively	
   standardizing	
   local	
  
measurements	
   into	
   kilos.	
   	
   At	
   least four	
   different	
   local	
   agricultural	
   weight	
   measurements	
   are	
  
used	
  in	
  the	
  communities	
  where	
  the	
  evaluation	
  survey	
  was	
  conducted	
  (“cans”,	
  “bags”,	
  “baskets”	
  
and	
  “buckets”), each	
  of	
  which	
  also	
  has	
  a	
  range	
  of	
  measures	
  (small,	
  medium,	
  large,	
  etc.).	
  	
  	
  While	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   8 Email	
  communication	
  with	
  the	
  FH	
  Agricultural	
  Program	
  Officer,	
  May	
  24,	
  2013
9 Diskin,	
   Patrick,	
   Agricultural	
   Productivity	
   Indicators	
   Measurement	
   Guide, Food	
   and	
   Agriculture	
   Technical	
   Assistance	
  
(FANTA),	
  December	
  1997
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 16
an	
   agricultural conversion	
   chart	
   was	
   developed	
   for	
   use	
   by	
   enumerators	
   during	
   the	
   survey	
  
exercise,	
  the	
  range	
  of	
  options	
  and	
  measures	
  (in	
  addition	
  to	
  new	
  ones	
  that	
  showed	
  up	
  during	
  the	
  
field	
   exercise)	
   led	
   to	
   significant	
   extrapolation	
   in	
   the	
   best-­‐case scenario	
   and	
   downright	
  
guesswork	
  in	
  the	
  worst	
  case;	
  
(4)	
  The	
   fourth factor	
  may	
  be	
  related	
   to	
   the	
  cultural	
  context.	
  Discussions	
  of	
  how	
  much	
   farmers	
  
produce	
  are	
  sensitive.	
  	
  This	
  is	
  owed	
  to	
  the	
  fact	
  that	
  (a)	
  farmers	
  understand	
  that	
  decisions	
  may	
  
be	
  made	
  about	
   future	
   support based	
   on	
   their	
   responses	
  and	
   (b)	
  much	
   in	
   the	
   way	
   that	
   in	
   the	
  
West	
   it	
   would	
   be	
   considered	
   inappropriate	
   to	
   openly	
   discuss	
   salaries,	
   some	
   farmers	
   may	
  
consider	
  it	
  inappropriate/uncomfortable to	
  discuss	
  agricultural	
  yields	
  and	
  salaries.	
  	
  These	
  factors	
  
may	
  have	
  resulted	
  in	
  underreporting.	
  	
  
The data	
  from	
  Figure	
  4b comes	
  from	
  the	
  FH Annual	
  Agricultural	
  Surveys,	
  with	
  the	
  first	
  survey	
  (2010)	
  
used	
  as	
  a	
  proxy	
  for	
  the	
  baseline	
  and	
  2012	
  used	
  as	
  a	
  proxy	
  for	
  the	
  final	
  evaluation.	
   Assuming	
  that	
  the	
  
LQAS	
   and	
   crop-­‐cutting	
   methodologies	
   were	
   followed	
   correctly	
   and	
   that	
   data	
   was	
   managed	
   and	
  
analyzed	
   properly the	
  annual	
   survey	
   data	
  might	
   be	
   a	
  more	
  accurate	
   reflection	
   of	
   overall	
   trends	
   in	
  
productivity for	
  project	
  participants.	
  FGDs	
  with	
  two	
  farmer	
  groups	
  and	
  conversations	
  with	
  three	
  FH	
  
agriculture	
   staff	
  also	
   suggest	
   that	
   yields	
   have	
   indeed	
   increased.	
   	
   Respondent	
   farmers	
  emphatically	
  
described	
  gains	
  in	
  productivity	
  and	
  they	
  attribute	
  those	
  gains	
  to	
  use	
  of	
  technologies	
  promoted	
  by	
  FH	
  
and	
   introduced	
   through	
   FFLGs.	
   	
   Coincidentally, according	
   to	
   the	
   FH	
   Agricultural	
   Program	
  Manager	
  
the data	
  from	
  the	
  Annual	
  Agricultural	
  Surveys also	
  aligns	
  more	
  closely	
  with	
  average	
  productivity	
  data	
  
for	
  the	
  Cabo	
  Delgado	
  province,	
  further	
  suggesting	
  that	
  the	
  data	
  may	
  be	
  a	
  more	
  accurate	
  reflection	
  of	
  
productivity. Figure	
  4b	
  clearly shows yearly	
  productivity	
  increases	
   for	
  sesame,	
  cowpeas,	
  groundnut
and	
  rice.
10
Given	
  that	
  multiple	
  sources	
  of	
  data	
  suggest	
  that	
  productivity	
  has	
  increased	
  during	
  the	
  program-­‐ and	
  
taking	
   into	
   account	
   the	
   known	
   methodological	
   limitations	
   with	
   the	
   baseline/final	
   evaluation	
   data	
  
collection	
   instrument (described	
   above)-­‐ it	
   is	
   the	
   consultant’s	
   opinion	
  that	
   the	
  results	
   of	
   the	
   final	
  
evaluation	
  likely	
  do	
  not accurately	
  reflect	
  the	
  program’s	
  contribution	
  to	
  agriculture.	
  	
  In	
  other	
  words,	
  
despite	
   the	
   results	
   of	
   the	
   evaluation	
   survey,	
   which	
   suggest	
   limited	
   productivity	
   gains,	
   the	
   other	
  
sources	
  of	
  data	
  in	
  the	
  program	
  indicate	
  strongly	
  that	
  productivity	
  is	
  likely	
  to	
  have	
  increased between	
  
the	
  baseline	
  and	
  final	
  evaluations.
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   10 Because	
  the	
  consultant	
  did	
  not	
  have	
  access	
  to	
  the	
  Annual	
  Agricultural	
  Survey	
  raw	
  data	
  files the	
  validity	
  of	
  these	
  findings	
  
could	
  not	
  be	
  verified.
87	
  
286	
  
161	
   293	
  
	
  	
  	
  	
  51.8	
  
163.2	
  
	
  	
  94.3	
  
334.6	
  
0	
  
200	
  
400	
  
600	
  
800	
  
1000	
  
1200	
  
Sesame	
   Ground	
  
Nut	
  
Cow	
  peas	
   Rice	
  
Figure	
  4a:	
  Agricultural	
  Produc]vity-­‐	
  
Final	
  Evalua]on	
  Survey	
  
Baseline	
   Final	
  
428	
   426	
   383	
  
661	
   677	
   617	
  
745	
  
1028	
  
0	
  
200	
  
400	
  
600	
  
800	
  
1000	
  
1200	
  
Sesame	
   Ground	
  
Nut	
  
Cow	
  peas	
   Rice	
  
Figure	
  4b:	
  Agricultural	
  Produc]vity-­‐	
  
FH	
  Annual	
  LQAS	
  Surveys	
  
Baseline	
  (2010)	
   Final	
  (2012)	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 17
For	
  the	
  final	
  Annual	
  Agricultural	
  survey	
  in	
  2013	
  it	
  might be	
  worth	
  investing	
  in	
  an	
  external	
  consultant	
  
who	
   can	
   oversee	
   the	
   LQAS	
   methodology, sampling,	
   data	
   collection,	
   management	
   and	
   processing.	
  
This	
  will	
  support	
  objective	
  validation	
  of the productivity	
  figures presented	
  above.	
  
Overall	
   diversification	
   of	
   production does not	
   seem	
   to	
   have	
   increased	
   during	
   the	
   program.	
   The	
  
percent	
   of	
  HH	
   planting	
   3	
   or	
  more	
   crops	
   has	
   not	
  increased	
   significantly	
   during	
   the	
   program.	
  At	
   the	
  
baseline	
   55.3%	
   of	
   HH	
   planted	
   3	
   or	
   more	
   crops,	
   while	
   at	
   the	
   final	
   evaluation	
   the	
   number	
   had	
  
remained	
  statistically	
  constant	
  at	
  58.3%	
  (LCL	
  53%	
  /	
  UCL	
  63.6%).
Intermediate	
  Result	
  1.2:	
  Natural	
  resource	
  base	
  protected	
  and	
  enhanced
To	
   address	
   challenges	
   with	
   soil	
   fertility,	
   FH	
   proposed to	
   increase cover	
   cropping	
   in	
   fields	
   with	
  
leguminous	
  crops to fix	
  nitrogen	
  and	
  add	
  organic	
  matter	
  in	
  the	
  soil.	
  Other	
  strategies	
  to	
  protect	
  and	
  
improve	
   the	
   soil	
  and/or	
  increase	
  nutrients	
  and	
  included reduced	
   field	
  burning,	
   reduced	
   soil	
   tillage,	
  
and	
  incorporation	
  of	
  mulch.	
  	
  
Table	
  8:	
  IPTT	
  Indicator	
  and	
  targets	
  for	
  IR1.2
Percent	
  of	
  NRM	
  training	
  beneficiaries	
  that	
  implemented	
  3	
  OR	
  MORE	
  improved	
  NRM	
  
practice in	
  their	
  fields	
  in	
  the	
  past	
  year 50%
See	
  the	
  indicators regarding	
  technologies	
  and	
  legume	
  production	
  under	
  IR	
  1.1.
Intermediate	
   Result	
   1.3:	
   Market	
   links	
   established	
   and	
   households	
   making	
   use	
   of	
   economic	
  
opportunities
To	
   address	
   deficits	
   in food	
   access and	
   increase	
   community	
   resilience,	
   FH	
   proposed	
   the	
  
implementation	
   of	
   interventions	
   to	
   increase	
   and	
   diversify rural	
   income,	
   including:	
   1)	
  support for	
  
production	
  and	
  sales	
  of high	
  value	
  cash	
  crops-­‐ sesame, cashew and groundnut;	
  and	
  2)	
  establishment	
  
of Savings Groups to	
  mobilize	
  community	
  saving	
  and	
  encourage	
  entrepreneurial	
  activity.	
  
Table	
  9:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR1.3
Percent	
  of	
  HH	
  producing	
  at	
  least	
  1	
  high	
  value	
  improved	
  cash	
  crop 80%
Percent	
  of	
  women	
  participating	
  in	
  cash	
  crop	
  production 50%
Average	
  savings	
  group	
  distribution	
  value	
  per	
  group	
  at	
  end	
  of	
  cycle $200
Total	
  funds	
  generated	
  for	
  all	
  savings	
  groups	
  per	
  cycle	
  (US$) $52,800
Number	
  of	
  legalized	
  associations	
  (10	
  of	
  Savings	
  Groups) 120
The	
   percent	
   of	
   HH	
   involved	
   in	
   cash	
   crop	
   production	
   has	
   not	
   increased.	
   As	
   Figure	
   5 shows,	
   at	
   the	
  
baseline 34%	
  of	
  respondent	
  HH	
  planted	
  at	
  least	
  
one	
  cash	
  crop	
  (cashew,	
  groundnut,	
  sesame).	
  	
  At	
  
the	
  final	
  evaluation	
  that	
  figure	
  is	
  38.1%,	
  with	
  no	
  
statistically	
   significant	
   difference between	
  
baseline	
  and	
  final	
  evaluation.	
  	
  
Table	
   10 (below)	
  further	
   shows	
   that	
  there	
   has	
  
been	
   no	
   significant	
   increase	
   in	
   production	
   of	
  
any	
   of	
   three	
   key	
   cash	
   crops. According	
   to	
   the	
  
FH	
  Agriculture	
  Program	
  Manager,	
  these	
   figures
may	
  have	
  remained	
  constant	
  because-­‐ with	
  the	
  
exception	
   of	
   sesame-­‐ FH	
   promoted	
   the	
   use	
   of	
  
improved	
   varieties.	
   	
   In	
   most	
   cases	
   these	
  
varieties	
   were	
   promoted	
   among	
   participants	
  
who	
   were	
   already	
   producing	
   cash	
   crops and	
  
Baseline	
  N=	
  300;	
  Final	
  N=	
  333
p=	
  .459:	
  NO	
  difference
34%	
   38.1%	
  
0%	
  
20%	
  
40%	
  
60%	
  
80%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  5:	
  Percent	
  of	
  HH	
  producing	
  at	
  
least	
  one	
  high	
  value	
  cash	
  crop	
  
Target	
  80%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 18
were	
  not	
  introduced	
  to	
  new	
  farmers.	
  	
  It	
  is	
  also	
  important	
  to	
  bear	
  in	
  mind	
  that	
  while	
  the	
  percentage	
  
of	
   HH	
   producing	
   the	
   crops	
   has	
   not	
   increased	
   overall	
   productivity	
   of	
   cash	
   crops	
   has	
   increased,	
  
according	
  the	
  Annual	
  Agricultural	
  Surveys.	
  	
  
Using	
  the	
  gender	
  of	
  HH	
  heads	
  as	
  a	
  proxy	
  for	
  the	
  full	
  
population	
   of	
   female	
   agriculture	
   participants,	
   the	
  
evaluation	
   survey	
   showed	
   that	
   63%	
   of	
   female	
   HH	
  
heads	
   have	
   participated	
   in	
   cash	
   crop production,	
  
approximately	
  13%	
  above	
  the	
  target	
  of	
  50%.	
  	
  This	
  increase	
  is	
  not	
  surprising	
  given	
  that	
  this	
  indicator	
  
value	
  was	
  listed as	
  less	
   than	
   5	
   percent	
  in	
   the	
   baseline-­‐ suggesting	
   that	
   the	
  value	
  was	
  an	
  estimate.	
  	
  
Calculating	
  this	
  value	
  based	
  on female	
  HH	
  heads	
  at	
  the	
  baseline	
  would	
  have	
  been	
  likely	
  to	
  produce	
  
similar	
  proportions	
   to	
   those	
   from	
   the	
   final	
  evaluation.	
  The	
  original	
  baseline	
  data	
  was	
  not	
  available,	
  
making	
  it	
  impossible	
  to	
  verify	
  this	
  assumption.	
  
Producers	
   of	
   cash	
   crops	
   have	
   increased	
   their	
   engagement	
   with	
   local	
   markets.	
   	
   The	
   MYAP	
   has	
  
supported	
  the	
  development	
  of	
  15	
  marketing	
  forums	
  through	
  which	
  associated	
  farmers	
  sell	
  produce-­‐
mostly	
   sesame-­‐ to	
  a	
   single	
  identified	
   buyer	
   out	
   of	
  Maputo.	
  	
  According	
   to	
   data	
   provided	
   by	
   the	
  FH	
  
Agriculture	
  team	
   (included	
  in	
  Table	
  11), participants	
  in	
   forums	
  have	
  collectively	
  sold	
  a	
  total	
  of	
  over	
  
95,000kgs	
  of	
  sesame	
  worth	
  approximately	
  $87,900.	
  	
  Unfortunately,	
  no	
  data	
  was	
  available	
  regarding	
  
the	
   number	
   of	
   farmers	
  
who	
   participate	
   in	
   those	
  
sales,	
   despite	
   multiple	
  
requests	
   for	
   those	
  
figures.	
   	
   If	
   we	
   know,	
  
however,	
   that	
   9%	
   of	
   all	
  
agricultural	
   beneficiaries	
  
(total	
  beneficiaries:	
  7887)	
  
produce	
   sesame	
   we	
   can	
  
calculate	
   that	
   these	
  
collective	
   earnings	
   have	
  
been distributed	
  among	
  roughly	
  709	
  farmers	
  (7887	
  x	
  .09).	
  	
  This	
  would	
  work	
  out	
  to	
  approximate	
  sales	
  
per	
   capita	
   of	
   $123.97	
   ($87,900	
   /	
   709).	
   	
   This	
  back-­‐of-­‐the-­‐envelope	
  figure	
   does	
   not	
   account	
   for any	
  
relevant	
   expenses	
   incurred.	
   	
   Once	
   labor	
   costs	
   are	
   figured	
   in	
   it	
   seems	
   like	
   a	
   relatively	
   low	
   level	
   of	
  
earnings.	
  	
  
Data	
  management has	
  been	
  a weakness	
  in	
   the	
  
agricultural	
   component	
   of	
   the	
   program.	
  
Although	
  there	
  have	
  clearly	
  been	
  earnings	
  as	
  a	
  
result	
  of	
  collective	
  marketing,	
  without	
  the	
  data	
  
regarding	
   the	
   number	
   of	
   people	
   who	
  
participated	
   in	
   those activities	
   it	
   is	
   difficult	
   to	
  
understand	
   the	
   impact	
   of	
   these	
   activities	
   on	
  
mean	
  incomes.	
  The	
  overall	
  data	
  point	
  ($87,000)	
  
is	
  a valuable	
  metric	
  only	
  if	
  we	
  know	
  how	
  many	
  
people	
   are	
   sharing	
   those	
   profits	
   and	
   at	
   what	
  
rate. It	
   is	
   a	
   weakness	
   in	
   program’s design	
   that	
  
no	
  indicators	
  were	
  included	
   the	
  IPTT	
   regarding	
  
mean	
   earnings	
   for	
   participants	
   in	
   collective	
  
marketing	
  activities.	
  	
  	
  
In	
   terms	
   of	
   cash	
   crop	
   sales,	
   however,	
   Figure	
   6
shows that	
   of	
   those	
   HH	
   that	
   produced cash	
   crops,	
   61%	
  sold	
   at	
   least	
   part	
   of	
   what	
   they	
   produced.	
  
Table	
  10:	
  HH	
  participation	
  in	
  cash	
  crops	
  
Cash	
  crop Baseline Final Sig
Groundnut 15% 12.30% N
Cashew 25% 29.70% N
Sesame 7% 9% N
Table	
  11:	
  Sesame	
  sales	
  through	
  collective	
  marketing
DISTRICTS
2010/2011 2011/2012
Qty	
  	
  (kg) Cash	
  (USD) Qty	
  (kg) Cash	
  (USD)
Moçimboa	
   6,000 $5,357 22,629 $21,821
Palma 2,715 $2,424 6,042 $5,826
Nangade 7,500 $4018 50,268 $48,473
TOTALS 16215	
  kg $11,799 78939kg $76,120
Final	
  Evaluation	
  data	
  only:	
  N=	
  127
Sold	
  	
  
61%	
  
Did	
  not	
  
sell	
  
39%	
  
Figure	
  6:	
  Percent	
  of	
  cash	
  crop	
  
producers	
  who	
  sold	
  /	
  did	
  not	
  sell	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 19
While	
   no income	
   figures	
   were	
   available	
   for	
   those	
   sales,	
   the	
   survey	
   shows	
   that 61%	
   of	
   cash	
   crop	
  
producers	
  earned	
  at	
  least	
  some	
  income	
  from	
  cash	
  crop	
  sales.11 Furthermore,	
  according	
  to	
  the	
  most	
  
recent	
   FH	
   Annual	
   Agricultural	
   Survey	
   many	
   of	
   the	
   MYAP’s	
   agricultural	
   participants	
   sell	
   goods	
  
individually	
   in	
   local	
   markets.	
   	
   Although	
   sales	
   volume	
   and	
   earnings	
   data	
   are	
   unavailable	
   for	
   those	
  
sales,	
  we	
   do	
  know	
   from	
   the	
  evaluation	
   that	
   roughly	
   23%	
   of	
  all	
   respondents	
  in	
   the	
   final	
  evaluation	
  
(N=333)	
  sold	
  at	
  least	
  one	
  crop	
  in	
  the	
  year	
  previous	
  to	
  the	
  survey	
  exercise.	
  	
  Together	
  this	
  data	
  tells	
  us	
  
that	
  HH	
  in	
  the	
  MYAP	
  intervention	
  areas	
  have	
  generated	
  income	
  from	
  their	
  productive	
  activities.	
  	
  
Even	
  with	
   the	
  modest success	
   in	
   linking	
   farmers	
   to	
   collective	
  markets,	
  a	
   few	
   challenges	
   should	
   be	
  
highlighted.	
   	
   First,	
   farmers	
   in	
   collective	
   marketing	
   depend	
   entirely	
   on	
   FH	
   for	
   market	
   linkages.	
  	
  
Farmers	
   themselves	
   and	
   FH	
   staff	
   confirmed	
   this.	
   	
   Farmers	
   have	
   not	
   received	
   training	
   in	
   basic	
  
marketing	
  concepts	
  such	
  market	
  studies,	
  value	
  addition	
  or	
  price	
  negotiation.	
  	
  Furthermore,	
  across	
  all	
  
intervention	
  districts	
  all	
  farmers	
  who	
  sell	
  through	
  forums	
  sell	
  to	
  a	
  single buyer	
  who	
  determines	
  the	
  
price;	
   farmers	
   do	
   not	
   actively	
   negotiate	
   in	
   price	
   or	
   contract	
   negotiations.	
   	
   At	
   the	
   same	
   time,	
   this	
  
buyer	
  travels	
  to	
  the	
  farmers	
  to	
  collect	
  and	
  pay	
  for	
  the	
  produce,	
  which	
  means	
  that	
  the	
  farmers	
  have	
  
not	
   built	
   the	
   capacity	
   to	
   plan	
   and	
   manage	
   the	
   logistics	
   normally	
   involved	
   in	
   collective	
   marketing.	
  
These	
   findings	
  run counter	
   to	
   the	
  understood	
  purpose	
  of	
  collective	
  marketing, which	
  is	
   to	
  increase	
  
the	
  negotiating	
  and	
  earning	
  power	
  of	
   farmers.	
  	
  	
  Farmers	
  are	
  entirely	
   subject	
   to	
   the	
  buyer.	
  	
  In	
  one	
  
community	
   we	
   visited	
   during	
   FGDs,	
   for	
   example,	
   respondents	
   said	
   that	
   one	
   year	
   the	
   buyer	
   had	
  
reached	
   his	
   quota	
   of	
   purchases,	
  and	
  as	
  a	
   result	
  he	
  was	
   no	
  longer	
  interested	
  in	
   buying	
   from	
   them,	
  
leaving	
  them	
  with	
  a	
  glut	
  of	
  product.	
  	
  
Conversations	
   with	
   FH	
   staff	
   suggested	
   that	
   these	
  factors have	
   been	
   significant	
   weaknesses	
   in	
   the	
  
program’s marketing	
  component.	
  	
  Although	
  there	
  have	
  been	
  short-­‐term	
  gains	
  in	
  sales	
  and	
  incomes,	
  
these	
  gains	
  may be	
  unsustainable	
  past	
  the	
  project’s	
  end.	
  
FH	
  supported	
  the	
   formation	
  of	
  savings	
  groups	
   (VSLAs) in	
  order	
  to	
  smooth	
  consumption	
  and	
  reduce	
  
vulnerability	
  during	
  lean	
  periods.12	
  	
  According	
  to	
  data	
  provided	
  by	
  the	
  FH	
  M&E	
  Manager,	
  196 savings	
  
groups	
  have	
  been	
  formed since	
  the	
  start	
  of	
  the	
  project.	
  	
  Of	
  these,	
  32	
  have	
  been	
  legalized.	
  	
  Together	
  
3552	
   participants	
   (1693	
   women	
   /	
   1859	
   men)	
   in	
   VSLAs	
   have	
   generated a	
   cumulative	
   total	
   of	
  
approximately	
   $201,15013 in	
   savings,	
   approximately	
   $44,50014 in	
   interest	
   income and	
   $13,915	
   in	
  
social	
  investment	
   funds	
  during	
   the	
  life	
  of	
   the	
  program.	
  	
  The	
  average	
   savings	
  group	
  distribution	
   for	
  
each	
  cycle	
  is	
  approximately	
  $776.50,	
  more	
  than $500	
  above	
  the	
  target	
  of	
  $200.	
  	
  The	
  total	
  of	
  funds	
  
generated	
  for	
  all	
  savings	
  groups	
  per	
  cycle	
  (mean	
  of	
  total	
  savings	
  +	
  total	
  interest	
  +	
  total	
  social	
  funds/	
  5	
  
years)	
  is	
  $51,913,	
  slightly	
  below	
  the	
  target	
  of	
  $52,800.
The	
  savings	
  and	
  loan	
  methodology	
  used	
  by	
  FH	
  follows	
  the	
  standard	
  VSLA	
  model	
  introduced	
  by	
  local	
  
NGO	
  Opavela,	
  with	
  slight	
  adjustments	
  in	
  its	
  implementation	
  to	
  respond	
  to	
  perceived	
  local	
  needs.	
  	
  In	
  
typical	
   VSLAs	
   members	
   establish	
   standard	
   contribution	
   amounts and	
   timelines	
   (e.g.,	
   weekly	
  
contributions) and	
   impose	
   a	
   penalty	
   on	
   those	
  members	
  who	
   do	
   not	
   pay	
   into	
   the	
   savings	
   account.	
  	
  
According	
   to	
   the	
   FH	
   Agricultural	
   Program	
   Manager	
   VSLAs	
   supported	
   by	
   the	
   MYAP	
   have	
   not	
   been	
  
required	
   to	
   establish	
   standard	
   contribution	
   amounts	
   or	
   penalties.	
   FH	
   made	
   this	
   decision	
   to	
  
accommodate	
   the	
   extreme poverty	
   within	
   MYAP	
   intervention	
   communities and	
   to	
   support	
   a	
   high	
  
level	
   of	
   VSLA	
   participation.	
   This	
   adaptation	
   shows	
  strong responsiveness	
   to	
   community	
   concerns;	
  
however,	
   it may	
   also	
   decrease	
   the	
   effectiveness	
   of	
   the	
   intervention	
   as	
   a	
   means	
   to	
   smooth	
  
consumption	
  and	
   reduce	
  vulnerability	
   to	
  shocks.	
  	
  Those	
  people	
  who	
  are	
   “too	
  poor”	
   to consistently
contribute	
   to	
   VSLAs	
   may be	
   only	
  slightly	
   less	
   vulnerable	
   to	
   shocks	
   than they	
   were	
   prior	
   to	
   their	
  
participation and	
  may	
  be	
  far	
  less	
  prepared	
  for	
  shocks	
  relative	
  to	
  their	
  peers	
  who	
  contribute	
  regularly.	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   11 Baseline	
  data	
  was	
  not	
  available	
  for	
  cash crop	
  sales,	
  making	
  comparison	
  impossible 12 Telephone	
  discussion	
  with	
  FH	
  Food	
  Security	
  Director,	
  April	
  19,	
  2013 13 Mtz	
  6,023,527	
  converted	
  to	
  US	
  $,	
  based	
  on	
  exchange	
  rate	
  of	
  1USD	
  =	
  29.	
  9449Mtz,	
  based	
  on	
  rates	
  found	
  on	
  www.xe.com 14 Mtz 1,333,757	
  converted	
  to	
  US	
  $,	
  based	
  on	
  exchange	
  rate	
  of	
  1USD	
  =	
  29.	
  9449Mtz,	
  based	
  on	
  rates	
  found	
  on	
  www.xe.com
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 20
Likewise,	
  under	
  a flexible	
  implementation	
  model	
  the	
  risk	
  associated	
  with	
  loans	
  dispersed	
  through	
  the	
  
VSLAs	
  may	
   also	
   be	
   disproportionately	
   distributed	
   among	
   the	
   members	
   who	
   have	
   contributed	
   the	
  
most,	
   effectively	
   increasing	
   their	
   vulnerability	
   as	
   well.15 It	
   is	
   important	
   to	
   note	
   that	
   none	
   of	
   these	
  
assumptions	
  was	
  verified	
  due	
  to	
  limited	
  access	
  to	
  the	
  program’s	
  VSLA	
  data.	
  
Nonetheless	
  it is	
   clear	
  is	
   that	
  FH’s support	
   for	
  VSLAs	
   has	
  led	
   to increased	
  income	
  through	
  interest	
  
earnings	
  while providing 3,500	
  men	
  and	
  women	
  access	
  to	
  more	
  than $200,000	
  in	
  available	
  capital	
  for	
  
livelihoods	
  investment	
  or	
  other	
  household	
  consumption	
  needs.	
  
Impact	
  indicators for	
  SO1	
  are	
  included	
  in	
  Table	
  12
Table	
  12:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  SO1
Average	
  HH	
  Dietary	
  Diversity	
  Score 4
Average	
  number	
  of	
  months	
  of	
  adequate	
  food	
  provisioning >10.5
According	
   to	
   FANTA,	
   household	
   food	
   access	
   is	
  
the	
   ability	
   to	
   acquire	
   sufficient	
   quality	
   and	
  
quantity	
   of	
   food	
   to	
   meet	
   all	
   household	
  
members’	
   nutritional	
   requirements	
   for	
  
productive	
   lives.	
  HH	
  access	
   to	
   food depends	
   on	
  
the	
   ability	
   of	
   households	
   to	
   obtain	
   food	
   from	
  
their	
   own	
   production,	
   stocks,	
   purchases,	
  
gathering,	
  or	
  through	
  food	
  transfers	
  of	
  one	
  kind	
  
or	
  another.	
   Access	
  to	
  food	
  also	
  depends	
  on	
  the	
  
resources	
   available	
   to	
   individual	
   HH members.	
  
The	
  Household	
  Dietary	
  Diversity Score	
  measures	
  
the	
   mean	
   number	
   of	
   different	
   food	
   groups	
  
consumed	
  during	
  a	
  defined	
  time	
  period	
  and	
  is	
  a	
  
strong	
   proxy	
   indicator	
   of	
   the	
   socio-­‐economic	
  
status	
  of	
   the	
  household,	
   reflecting	
   their	
   relative	
  
ability	
  to	
  access	
  food	
  in	
  appropriate	
  quantities	
  and	
  quality.	
  	
  
As	
  Figure	
  7 (above)	
  shows,	
  there	
  has	
  been	
  a	
  statistically	
  significant	
  change	
  in	
  the	
  HH	
  Dietary	
  Diversity	
  
Score from 3.25 at	
  the	
  baseline	
  to	
  5.17	
  at	
  the	
  final	
  
evaluation,	
   a	
   full	
   point	
   above	
   the	
   target.	
   On	
  
average,	
   HH consume	
   almost	
   two	
   more	
   food	
  
groups	
   now	
   than	
   they	
   did	
  at	
   the	
   baseline.	
  	
  	
  	
  We	
  
can	
  see	
  in	
  Table	
  13	
  that	
  food	
  group	
  consumption	
  
patterns	
  have	
  changed	
  significantly	
  across	
  almost	
  
all	
   food	
   groups,	
   with	
   the	
   exception	
   of	
   grains,	
  
fish/seafood	
   and	
   dairy	
   products	
   which	
   have	
   all	
  
remained	
   statistically	
   unchanged.	
   	
   Seafood	
   is	
   a	
  
staple	
   food,	
   available	
   with	
   relative	
   ease	
   in	
  
program	
   communities	
   along	
   the	
   coast.	
  	
  
Consumption	
  was	
  reasonably	
  high	
  at	
  the	
  baseline	
  
and	
   remained	
   statistically	
   constant	
   at	
   the	
   final	
  
evaluation.	
   Those	
   communities	
   that	
   do	
   not	
  
consume	
   seafood	
   are	
   likely	
   to	
   be	
   further	
   inland,	
  
where	
   seafood	
   consumption	
  would	
   represent	
  an	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   15 The	
   flexible	
   nature	
   of	
   VSLA	
   contributions	
   may	
   also	
   help	
   explain	
   why	
   aggregate	
   savings	
   seems	
   lower	
   than	
   one	
   might	
  
expect	
  after	
  five	
  years	
  of	
  programming:	
  since	
  savings	
  were	
  optional,	
  overall	
  savings	
  rates	
  might	
  be	
  lower	
  than	
  they	
  would	
  
have	
  been	
  under	
  a	
  model	
  with	
  minimal	
  required	
  contributions.	
  	
  	
  
Table	
  13:	
  Consumption	
  by	
  Food	
  Group
Food	
  
Group BL Final Sig.
Cereals 95.3% 97.9% N
Roots	
  /	
  tubers 13.3% 30.6% Y
Vegetables 42.3% 61.3% Y
Fruits 25.3% 41.7% Y
Meat,	
  poultry 7.0% 21.0% Y
Eggs 4.0% 12.3% Y
Fish	
  /	
  seafood 60.3% 63.7% N
Legumes 9.7% 44.7% Y
Dairy 2.7% 3.0% N
Oils	
  or	
  fats 22.0% 68.5% Y
Sugar,	
  honey 18.0% 33.6% Y
Miscellaneous 25.3% 36.0% Y
Baseline	
  N=	
  300;	
  Final	
  N=	
  333
p=	
  .000
3.25	
  
5.17	
  
0	
  
1	
  
2	
  
3	
  
4	
  
5	
  
6	
  
7	
  
8	
  
9	
  
10	
  
11	
  
12	
  
Baseline	
   Final	
  
Figure	
  7:	
  Household	
  Dietary	
  Diversity	
  
Score	
  
Target	
  4	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 21
additional	
   cost	
   for	
   families.	
   	
   This	
   would	
   have	
   been	
   true	
   both	
   before	
   and	
   after	
   the	
   intervention,	
  
explaining	
  the	
  lack	
  of	
  change.	
  	
  Grains	
  are	
  the	
  staples	
  of	
  the	
  local	
  diet.	
  Unsurprisingly,	
  more	
  than	
  95%	
  
of	
  respondents	
  at	
  both	
  the	
  baseline	
  and	
  final	
  evaluations	
  consumed	
  these	
  items. Dairy	
  consumption	
  
has	
  remained	
  consistently	
  low	
  at	
  both	
  the	
  baseline	
  and	
  final	
  evaluations.	
  	
  This	
  is	
  because	
  there	
  is	
  no	
  
local	
  production	
  of	
  dairy	
  in	
  Cabo	
  Delgado.	
  Because	
  of	
   the	
  remote	
  location	
  of	
   the	
  province	
   the	
  vast	
  
majority	
   of	
   dairy	
   products	
   (such	
   condensed,	
   evaporated	
   or	
   powdered	
   milk)	
   are	
   imported,	
   making	
  
them	
  prohibitively	
  expensive	
  for	
  most	
  poor	
  families.	
  As	
  a	
  result	
  of	
  their	
  cost,	
  most	
  families	
  choose	
  to	
  
consume	
  other	
  more	
  accessible	
  items.	
  
A	
   HH’s ability	
   to	
   meet	
   its	
   food	
   needs	
   often	
  
varies due	
   to	
   factors	
   like the	
   quality	
   of crop	
  
production (for	
   any	
   number	
   of	
   reasons,	
  
including	
   climate	
   factors,	
   soil	
   quality,	
   input	
  
quality,	
   etc.)	
   and	
   an	
   increase or	
   decrease	
   in	
  
income	
   sources.	
  	
  If	
   we	
   look	
   at	
   Figure	
   8	
   we	
   can	
  
see	
   that	
   the	
   mean	
   number	
   of	
   Months	
   of	
  
Adequate	
   Household	
   Food	
   Provisioning	
   has	
  
experienced	
   a	
   statistically	
   significant	
   increase,	
  
from	
  9.87	
  to	
  11.16,	
  achieving	
   the	
   target	
   for	
   the	
  
program.
Figure	
  9	
  (below)	
  shows	
  that peak	
  hunger	
  occurs
between	
   November	
   and	
   March	
   in	
   both	
   data	
  
sets;	
   however, the	
   magnitude	
   of	
   hunger	
  
experienced	
  between	
  January	
  and	
  March	
  reduced	
  dramatically	
  at	
  the	
  final	
  evaluation.
Overview	
  and	
  Discussion	
  of	
  Agriculture	
  Findings
Table	
  14 Comparative	
  Summary	
  of	
  Baseline	
  and	
  Final	
  Agricultural Indicator Values for	
  SO1
Indicator BL FE
%	
  Point	
  /	
  
Numerical	
  
Difference
Sig.
IMPACT:	
  Livelihoods protected	
  and	
  enhanced
Average	
  number	
  of	
  months	
  of	
  adequate	
  food	
  provisioning 9.87 11.16 +1.29 Y
Average	
  Household	
  Diversity	
  Score 3.25 5.17 +1.92 Y
Production	
  diversified	
  and	
  productivity	
  increased
Percent	
  of	
  HH	
  with	
  intensive	
  dark	
  green	
  leaf	
  production 71.7% 91% +19.3% Y
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
Jan	
   Feb	
   Mar	
   April	
   May	
   June	
   july	
   Aug	
   Sept	
   Oct	
   Nov	
   Dec	
  
%	
  with	
  Inadequate	
  Food	
  
Figure	
  9:	
  Months	
  of	
  Adequate	
  HH	
  Food	
  Provisioning	
  in	
  Compara]ve	
  Perspec]ve	
  
Baseline	
  
Final	
  
Baseline	
  N=	
  300;	
  Final	
  N=	
  333
p=	
  .000
9.87	
  
11.16	
  
0	
  
1	
  
2	
  
3	
  
4	
  
5	
  
6	
  
7	
  
8	
  
9	
  
10	
  
11	
  
12	
  
Baseline	
   Final	
  
Figure	
  8:	
  Average	
  number	
  of	
  months	
  
of	
  Adequate	
  Food	
  Provisioning	
  
Target	
  >10.5	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 22
Percent	
   of	
   HH	
   adopting	
   and	
   diversifying	
   production	
   by	
  
including	
  at	
  least	
  1	
  new	
  leguminous	
  nutritious	
   food	
  crop	
  in	
  
farming	
  system
37.3% 58% +20.7% Y
Percent	
   of	
   beneficiaries	
   (farmers)	
   using	
   at	
   least	
   3	
  
sustainable	
  agricultural	
  technologies	
  in	
  the	
  past	
  year 32.7% 67% +34.3% Y
Percent	
   of	
   fruit	
   and	
   cashew	
   farming	
   training	
   beneficiaries	
  
practicing	
  at	
  least	
  two	
  improved	
  fruit	
  growing	
  techniques	
  in	
  
the	
  past	
  year
0 73% + NA
Percent	
  of	
  HH	
  adopting	
  sustainable	
  non	
  chemical	
  grain	
  and	
  
seed storage	
  protection	
  techniques 0 63% + NA
Productivity	
  (kg/ha)*
(1)	
  Sesame	
  
(2)	
  Groundnut	
  
(3)	
  Cow peas	
  
(4)	
  Rice (Nerica)
*Data	
  from	
  Annual	
  Agricultural	
  Survey)
428
426
383
677
661
617
745
1028
+233
+191
+362
+351
NA
Natural	
  resource	
  base	
  enhanced	
  and	
  protected
Percent	
   of	
   NRM	
   training	
   beneficiaries	
   that	
   implemented	
   3	
  
OR	
  MORE	
  improved	
  NRM	
  practice	
  in	
  their	
  fields	
  in	
  the	
  past	
  
year
32.7% 67% +34.3% Y
Market	
  links	
  and	
  economic	
  opportunities
Percent	
  of	
  HH	
  producing	
  at	
  least	
  1	
  high	
  value	
  improved	
  cash	
  
crop
34% 38.1% +4.1% N
Percent	
  of	
  women	
  participating	
  in	
  cash	
  crop	
  production <5% 63% + NA
Average	
  savings	
  group	
  distribution	
  value	
  per	
  group	
  at	
  end	
  of	
  
cycle 0 $776.5
0
+ NA
Total	
  funds	
  generated	
  for	
  all	
  savings	
  groups	
  per	
  cycle	
  (US$) 0 $51,91
3
+ NA
Number	
  of	
  legalized	
  associations	
  (10	
  of	
  Savings	
  Groups) 0 32 + NA
Of	
   the	
  16 indicators	
  included	
  in	
   the	
  IPTT	
   for	
  agriculture16,	
  15 of	
   them	
  experienced	
  positive	
  change.	
  	
  
Of	
  the	
  6	
  for	
  which	
  statistical	
  tests	
  of	
  significance	
  were	
  conducted,	
  5	
  of	
  them	
  experienced	
  statistically	
  
significant	
  change.	
  Four	
  indicator	
  values	
  fell	
  below	
  program	
  targets,	
  while	
  the	
  rest	
  met	
  or	
  exceeded	
  
target	
  values.17
For	
   production	
   and	
   diversification,	
   overall	
   outcomes	
   were	
   somewhat	
   mixed,	
   with	
   a	
   significant	
  
increase	
   in	
   the	
   percent	
   of HH	
   planting	
   dark	
   green	
   leaf	
   vegetables	
   and	
   incorporating	
   at	
   least	
   one	
  
leguminous	
  plant	
  into	
  their	
  production.	
   At	
  the	
  same	
  time,	
  however, no	
  single	
  leguminous	
  crop	
  has	
  
seen	
  statistically	
  significant increases	
   in	
   production	
   nor	
   has	
   there	
   been any	
  statistically	
  significant	
  
increase	
  in	
  the	
  percent	
  of	
   farmers	
  planting	
  three or	
  more	
  crops,	
  suggesting	
  that	
  overall	
  production	
  
diversification	
  objectives	
  may	
  have	
  fallen	
  short.	
  If	
  we	
  use	
  the	
  yield/production	
  data	
  derived	
  from	
  the	
  
FH	
  Annual	
  Agricultural	
  Survey	
  (which, for	
  reasons	
  described	
  above	
  may	
  be	
  somewhat more	
  reliable	
  
than	
  the	
  yield	
  data	
  from	
  the	
  final	
  evaluation)	
  then	
  it	
  appears	
  as	
  if	
  productivity	
  has	
  increased	
  across	
  
all	
  four	
  of	
  the	
  program’s	
  target	
  crops.	
  	
  These	
  figures	
  should	
  be	
  viewed	
  cautiously,	
  however,	
  since	
  the	
  
data	
  could	
  not	
  be	
  verified.	
   Nonetheless,	
  it	
  is	
  important	
  to	
  note	
  that	
  both farmers	
  and	
  FH	
  staff	
  also	
  
confirmed	
  through	
  qualitative	
  discussions	
  that	
  they	
  have	
  seen	
  increases	
  in	
  productivity.	
   If	
  possible	
  it	
  
is	
  advisable	
  that	
  FH	
  work	
  with	
  an	
  external	
  consultant	
  to	
  facilitate	
  the	
  final	
  Annual	
  Agriculture	
  Survey	
  
to	
  validate	
  the	
  results	
  presented	
  in	
  this	
  report.	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   16 The	
  indicators	
  regarding	
  technology	
  adoption	
  under	
  IR1.1	
  and	
  NRM	
  under	
  IR1.2	
  are	
  the	
  same.	
  	
  They	
  are	
  counted	
  once. 17 Productivity	
  values	
  from	
  the	
  FH	
  Annual	
  Agricultural	
  Survey	
  (and	
  NOT	
  the	
  Final	
  Evaluation	
  survey)	
  are	
  being	
  used	
  here
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 23
There	
  was	
  a	
  high	
  degree	
  of	
  adoption	
  of	
  sustainable	
  agriculture	
  practices	
  to	
  enhance	
  and	
  protect	
  the	
  
natural	
  resource	
  base.	
  	
  With	
  the	
  exception	
  of	
  green	
  manuring	
  and	
  agroforestry	
  all	
  other	
  technologies	
  
promoted	
  by	
  the	
  program	
  saw	
  statistically	
  significant	
  increases.	
  	
  While	
  the	
  impacts	
  of	
  these	
  practices	
  
were not	
   measured	
   or	
   verified	
   during	
   the	
   evaluation	
   exercise,	
   farmers	
   from	
   the	
   program’s	
   FFLGs
expressed	
  the	
  belief	
  that	
  the	
  technologies	
  were	
  at	
  least	
  partially	
  responsible	
  for	
  gains	
  in	
  production.	
  
The	
  one	
  exception	
  was	
  with	
  organic	
  pesticides.	
  	
  These	
  were	
  viewed	
  as	
  ineffective	
  by	
  farmers	
  in	
  FGD,	
  
and	
   FH	
   staff	
   confirmed	
   that	
   the	
   use	
   of neem may have	
   undermined	
   that	
   component	
   altogether.	
  	
  
Farmers	
  are	
   satisfied with	
  FFLGs	
  as	
  a	
   sustainable	
  means	
   to	
   transmit	
  technical	
  assistance messages	
  
and	
  support	
   replication	
   of	
   best	
   practices	
   among	
   farmers,	
   and	
   they	
   confirmed	
   that	
   they	
   meet	
   as	
  
groups	
  even	
  in	
  the	
  absence	
  of FH	
  staff.	
  	
  Given	
  the	
  staffing	
  and	
  resource	
  constraints	
  of	
  the	
  Ministry	
  of	
  
Agriculture,	
   this	
   is	
  an	
   important	
  benefit that	
   can	
   serve	
  as	
   the	
   foundation	
   for	
   continued	
   farmer-­‐to-­‐
farmer	
  based	
  agricultural	
  extension	
  support	
  in	
  the	
  future.
Market	
  links	
  and	
  economic	
  opportunities have	
  increased	
  as	
  a	
  result	
  of	
  the	
  program.	
  	
  While	
  there	
  has	
  
been	
   no	
   significant	
   change	
   in	
   the	
   percent	
   of	
   HH	
   producing	
   cash	
   crops,	
   at	
   least	
   60%	
   of	
   those	
   who	
  
produce	
   cash	
   crops	
   are	
   selling	
   a	
   portion	
   (or	
   all)	
   of	
   them	
   either	
   collectively	
   or	
   individually	
   at	
   local	
  
markets.	
   Data	
  provided	
  by	
  FH	
  staff	
  regarding	
  income	
  from	
  collective	
  marketing	
  and	
  VSLAs	
  was	
  not	
  
verifiable, but taken	
   at	
   face	
   value	
   at	
   least	
   $130,000	
   in	
   new income	
  ($87,000	
   from	
   sesame	
   sales	
   +	
  
$44,500	
  in	
  interest	
  income	
   from	
  VSLAs)	
  has	
  been	
  generated	
   through	
   those	
  activities,	
  not	
  including	
  
any	
   income	
   generated	
   through	
   additional	
   individual	
   marketing	
   activity.	
   	
   Furthermore,	
   the	
   total	
  
collective	
  savings	
  of	
  VSLAs	
  represents	
  at	
  least	
  $200,000 of	
  available	
  capital	
  through	
  loans from VSLA	
  
funds.
Farmers	
   have	
   not	
   received training	
   to	
   support	
   increased	
   market-­‐based production.	
   Farmers	
   have	
  
not	
  gained	
  skills	
  in	
  market	
  analysis,	
  negotiation,	
  value	
  addition	
  or	
  processing,	
  and	
   the	
  program	
  has	
  
not	
   strengthened	
   other	
   linked	
   business	
   to	
   ensure	
   integral	
   and	
   strong	
   value	
   chains. As	
   a	
   result	
  
marketing	
  gains	
  may	
  be	
  short-­‐lived.	
  Farmers	
  are selling	
  collectively	
  through	
  markets,	
  but	
  that	
  is	
  the	
  
result	
   of	
   good	
   timing,	
  the	
   persistent	
   support	
   of	
   FH and	
   the	
   unusual circumstances	
   that	
   bring	
   the	
  
buyer	
  to	
  them,	
  rather	
  than	
  expecting	
  them	
  to	
  transport	
  to	
  the	
  buyer.	
  	
  Farmers	
  are	
  dependent	
  on	
  FH	
  
for	
  the	
  existing	
  market	
  linkages in	
  the	
  program;	
  without	
  the ongoing support of	
  FH	
  and	
  the	
  existing	
  
interest	
  on	
  the	
  part	
  of	
  the	
  buyer	
  that	
  linkage	
  will	
  not	
  be	
  sustained.	
  	
  	
  
VSLAs	
   are working:	
  HH	
   are	
   saving	
  money,	
  earning	
   interest	
   and	
   giving/receiving loans.	
   	
  It	
  might	
   be	
  
worth	
   exploring	
   whether	
   the	
   flexibility	
   of	
   the	
   groups may	
   keep	
   VSLAs	
   from	
   reaching	
   their	
   full	
  
potential	
   as	
   a	
   consumption	
   smoothing	
   and	
   vulnerability	
   reduction	
   mechanism.	
   Since	
   the	
   most	
  
vulnerable	
   can	
   always	
   find	
   reasons	
   not	
   to	
   save,	
   they	
   be	
   perennially	
   at	
   risk.	
   	
   Furthermore,	
   this	
  
arrangement	
   could increase	
   the	
   vulnerability	
   of	
   members	
   who	
   do	
   contribute	
   regularly	
   by	
  
disproportionately	
  placing	
  the	
  risks	
  for	
  VSLA	
  loans	
  on	
  their	
  shoulders.
Overall	
   sustainability	
  of	
   the	
  agriculture	
  program	
  could	
  be	
  at	
   risk	
  when	
   the	
  program	
  ends	
   since	
   the	
  
Ministry	
  of	
  Agriculture	
  does	
  not	
  have	
  the material	
  or	
  human	
  resources	
  capacity	
  to	
  provide	
  ongoing	
  
follow-­‐up	
  or	
  extension	
  support.	
  The	
  MoA has	
  a	
  list	
  of	
  formally	
  “recognized”	
  volunteers,	
  including	
  the	
  
149	
   community-­‐based	
   agricultural	
   promoters	
   who	
   have	
   worked	
   with	
   FH	
   agricultural	
   and	
   VSLA	
  
participants.	
   	
   However,	
   the	
   Ministry	
   has	
   logistics	
   challenges	
   and	
   resource	
   constraints	
   that	
   do	
   not	
  
allow	
   them	
   to	
   provide	
   any	
   kind	
   of	
   incentives	
   for	
   the	
   volunteers,	
   which	
   means	
   ongoing	
   support	
   is	
  
unlikely	
  to	
  happen	
  in	
  any	
  meaningful	
  way.	
  
Despite	
   these	
  challenges,	
   the	
  program’s	
  agricultural	
  component	
  has	
  achieved	
   results	
  in	
  production	
  
levels	
  and	
  income	
  generation	
  that	
  are	
  likely	
  to	
  have	
  contributed	
  to	
  increased	
  food	
  access.
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 24
B.	
  Strategic	
  Objective	
  2	
  (Health): Protect	
  and	
  enhance	
  human	
  capabilities	
  through	
  improved	
  
health	
  and	
  nutritional	
  status	
  of	
  pregnant	
  women,	
  preschool	
  children	
  and	
  HIV+	
  individuals
To	
  address	
  food	
  utilization,	
  FH	
  promoted positive	
  health	
  and	
  nutrition	
  behaviors	
  for children	
  0-­‐ 24
months of	
   age.	
   	
   The	
   program	
   used the Care	
   Group	
   model to: 1)	
   improve	
   nutrition	
   behaviors	
   and	
  
practices;	
  2)	
  strengthen	
  home	
  management	
  of	
  illness;	
  and	
  3)	
  promote	
  Essential	
  Hygiene	
  Behaviors.	
  	
  	
  
At	
  the	
  same	
  time,	
  FH	
  supported	
  the	
  reduction	
  of	
  fecal	
  and	
  water	
  borne	
  diseases	
  through	
  improved
access	
   to	
   clean	
  water	
  and sanitation	
  facilities.	
   Findings for	
  each	
   of	
  the	
   indicators	
   relating	
   to these	
  
interventions	
  are	
  included	
  below,	
  organized	
  by	
  Intermediate	
  Result.	
  	
  
Intermediate	
  Result	
  2.1:	
  Improved	
  use	
  of	
  Essential	
  Nutrition	
  Actions	
  (ENA)	
  by	
  pregnant	
  women	
  and	
  
mothers	
  of	
  young	
  children.
Appropriate	
  breastfeeding	
  and	
  complementary	
  feeding	
  practices	
  play	
  a	
  significant	
  role	
  in	
  improving	
  
the	
   health	
   and	
   nutrition	
   of	
   young	
   children	
   and	
   confer	
   significant	
   long-­‐term	
   benefits	
   during	
  
adolescence	
  and	
  adulthood. The	
  relevant	
  indicators	
  and	
  targets	
  for	
  FH	
  activities	
  related	
  to	
  ENA	
  are	
  
included	
  in	
  Table	
  15.
Table	
  15:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR2.1
Percent	
  of	
  children	
  0-­‐6 months	
  of	
  age	
  who	
  started	
  breastfeeding	
  within	
  one	
  hour	
  after	
  
birth 80%
Percent	
  of	
  children	
  0-­‐6m	
  who	
  were	
  exclusively	
  breastfed	
  in	
  the	
  last	
  24h 90%
Percent	
  of	
  children	
  12-­‐15	
  months	
  of	
  age	
  who	
   received	
  breast	
  milk	
  during	
   the	
  previous	
  
day 98%
Percent	
  of	
  infants	
  6-­‐8	
  months	
  of	
  age	
  who	
  received	
  solid,	
  semi-­‐solid	
  or	
  soft	
  foods	
  during	
  
the	
  previous	
  day 98%
Percent	
  of	
  breastfed	
  and	
  non-­‐breastfed	
  children	
  6-­‐24 months	
  of	
  age	
  who	
  received solid,	
  
semi-­‐solid	
   or	
   soft	
   foods,	
   with	
   the	
   minimum	
   required	
   dietary	
   diversity	
   and	
   minimum	
  
number	
  of	
  times	
  a	
  day
98%
Percent	
   of	
   caregivers	
  who	
   took	
   their	
   child	
  aged	
   0-­‐24m	
   for	
  an	
  evaluation	
   of	
   nutritional
status	
  in	
  the	
  last	
  four	
  months 95%
Figure	
  10 shows	
  that	
  a	
  total	
  of	
  65.8%	
  (LCL	
  61.2%	
  /	
  UCL	
  70.5%) of	
  children aged	
  0-­‐6	
  months across	
  the	
  
program	
   districts	
   breastfed	
   within	
   one	
   hour	
  
of	
  birth,	
  representing	
  a	
  statistically	
  significant	
  
increase	
   over	
   the	
   baseline.	
   	
   According	
   to	
  
WHO,	
  infants	
  who	
  breastfeed	
  within	
  the	
  first	
  
hour	
   interact	
   more	
   with	
   their	
   mothers,	
   stay	
  
warmer	
  and	
  cry	
  less.	
  Early	
  breastfeeding	
  has	
  
also	
   been	
   linked	
   with	
   decreased	
   neonatal	
  
mortality.	
   Additionally,	
   a	
   causal	
   association	
  
between	
   early	
   breastfeeding	
   and	
   reduced	
  
infection-­‐specific	
   neonatal	
   mortality	
   has	
  
been	
   reported,	
   supporting	
   the	
  
recommendation	
   of	
   early	
   initiation.18 Breast	
  
milk has	
  anti-­‐infective	
  properties	
  that	
  help	
  to	
  
protect	
   the	
   infant	
   against	
   infections.	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   18 Essential	
  Nutrition	
  Actions	
  Improving	
  Maternal-­‐Newborn-­‐Infant	
  and	
  Young Child	
  Health	
  and	
  Nutrition,	
  World	
  Health	
  
Organization,	
  2011.	
  	
  Found	
  on	
  May	
  13,	
  2013	
  at:	
  
http://www.who.int/nutrition/EB128_18_backgroundpaper2_A_reviewofhealthinterventionswithaneffectonnutrition.pdf
Baseline	
  N=	
  299;	
  Final	
  404
p=	
  .019
	
  	
  	
  	
  25.4%	
  
	
  65.8%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
  Evaluason	
   Final	
  Evaluason	
  
Figure	
  10:	
  Percent	
  of	
  children	
  who	
  
breasced	
  within	
  1	
  hour	
  ader	
  birth	
  
Target	
  80%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 25
Colostrum,	
   the milk	
   produced	
   in	
   the	
   first	
   few	
   days,	
   is	
   the	
   first	
   "essential	
   immunization".19 The
increase in	
  this	
  indicator	
  is approximately	
  14 percentage	
  points lower	
  than	
  the	
  target.	
  	
  However,	
  it	
  is	
  
important	
   to	
   highlight	
   that	
   in	
   the	
   Cabo	
   Delgado	
   context	
   the	
   colostrum	
   is	
   traditionally	
   considered	
  
“dirty”	
  milk.	
  	
  In	
  the	
  context	
  of	
  such	
  strong	
  negative	
  cultural	
  connotations	
  for	
  early	
  breastfeeding	
  the	
  
initial	
  target	
  might	
  have	
  been	
  unrealistic.
As	
  we	
   can	
   see	
   in	
   Figure	
   11,	
   exclusive	
   breastfeeding	
  among	
   children	
   between	
   0-­‐ 6 months has also	
  
increased	
   since	
   the	
   baseline,	
   increasing	
   approximately	
   70 percentage	
   points since	
   the	
   start	
   of	
   the	
  
program.	
  	
  Current	
  EBF	
  has	
  reached	
  a	
  level	
  of	
  
79.4% (LCL	
  72.3%	
  /	
  UCL	
  86.4%),	
  a	
  level	
  slightly
lower	
   than	
   the	
   program	
   target.	
   	
   Exclusive	
  
breastfeeding	
   for	
   the	
   first	
   six	
   months	
   is	
   an	
  
unequalled	
   way	
   of	
   providing	
   ideal	
   food	
   for	
  
the	
   healthy	
   growth	
   and	
   development	
   of	
  
infants.	
   	
   Breast	
   milk	
   is	
   the	
   natural	
   first	
   food	
  
for	
   babies,	
   it	
   provides	
   all	
   the	
   energy	
   and	
  
nutrients	
   that	
   the	
   infant	
   needs	
   for	
   the	
   first	
  
months	
   of	
   life,	
   promoting	
   sensory	
   and	
  
cognitive	
   development,	
   and	
   protecting
against infectious	
   and	
   chronic	
   disease.	
  
Exclusive	
   breastfeeding	
   reduces	
   infant	
  
mortality	
  due	
   to	
  common	
  childhood	
  illnesses	
  
such	
  as diarrhea	
  or	
  pneumonia,	
  and	
  helps	
  for	
  
a	
  quicker	
   recovery	
  during	
  illness.	
  Thereafter	
  infants	
  should	
  continue	
   to	
  breastfeed up	
   to	
  2	
  years	
  of	
  
age	
  (or	
  beyond).
20 	
  	
  
In	
   terms	
   of	
   breastfeeding	
   for	
   older	
   children,	
   the	
   percent	
   of	
   children	
   12-­‐ 15	
   months	
   of	
   age	
   who	
  
received	
  breast	
  milk in	
   the	
  24	
  hours	
  prior	
   to	
   the	
  survey remained	
  unchanged	
  between	
   the	
  baseline	
  
and	
  final	
  evaluations (96%	
  at	
  baseline,	
  94%	
  at	
  final	
  evaluation).	
  	
  This	
  indicator	
  was	
  high	
  at	
  baseline,	
  
with	
  targets	
  set	
  at	
  only	
  4 percentage	
  points above	
  initial	
  levels.	
  	
  	
  	
  
All	
  infants	
  should	
  start	
  receiving	
  foods	
  in	
  addition	
  to	
  breast	
  milk	
  from	
  6	
  months	
  onwards.	
  	
  As	
  Table	
  
16 demonstrates,	
  introduction	
  of	
  solid,	
  semi-­‐solid	
  or	
  soft	
  foods	
  for children	
  between	
  6	
  and	
  8	
  months
old	
   showed	
   no	
   significant	
  increase	
  over	
   the	
   baseline,	
  evidenced	
   by	
   the slight	
   overlap	
   between	
   the	
  
confidence	
   intervals	
   of	
   the	
   baseline	
   and	
   final	
   evaluation	
   proportions;	
   the	
   proportion	
   of	
   children	
  
between	
  6	
  and	
  8	
  months	
  who	
  receive	
  complementary	
  feeding	
  is	
  likely	
  to	
  lie	
  somewhere	
  between	
  the	
  
baseline	
  and	
  final	
  values.
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   19 Infant	
  and	
  Young	
  Child	
  Feeding	
  Summary	
  Sheet,	
  Save	
  the	
  Children;	
  found	
  on	
  May	
  29,	
  2013	
  at:	
  
http://www.savethechildren.org/atf/cf/%7B9def2ebe-­‐10ae-­‐432c-­‐9bd0-­‐df91d2eba74a%7D/Summary-­‐Sheets-­‐All.pdf 20 http://www.who.int/nutrition/topics/exclusive_breastfeeding/en/;	
  found	
  on	
  May	
  10,	
  2013
Baseline	
  N=	
  67;	
  Final	
  N=	
  126
p=	
  .000
Table	
  16:	
  Complementary	
  Feeding
Percent	
   of	
   children	
   between	
   6-­‐ 8	
   months	
  
who	
   received	
   any	
   solid,	
   semi-­‐solid	
   or	
   soft	
  
foods	
  during	
  the	
  previous	
  day
B.L.	
  % LCL UCL F.	
  E.	
  % LCL UCL
78% 65.4% 90.7% 95.9% 90.4% 100%
6%	
  
	
  	
  	
  79.40%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  11:	
  Percent	
  0-­‐	
  6	
  mos.	
  BF	
  
exclusively	
  in	
  the	
  last	
  24	
  hours	
  
Target	
  90%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 26
When	
   we	
   look	
   at	
   the	
   adequacy	
   of	
   complementary	
   feeding we	
   can	
   see	
   in	
   Figure	
   12 that,	
   although	
  
there	
   has	
   been	
   a	
   statistically	
   significant	
   increase	
   over	
   the	
   baseline,	
   the	
   frequency	
   and	
   variety	
   of	
  
foods	
   to	
   cover	
   the	
   nutritional	
   needs	
   of	
  
children	
  between	
  6-­‐ 24 months falls below	
  the	
  
program	
  target of	
  98%.	
  	
  At	
  the	
  final	
  evaluation	
  
27.6%	
   (LCL	
   22.3%	
   /	
   UCL	
   33%) of	
   children	
  
between	
   6-­‐ 24 months	
   receive	
   adequate	
  
diversity	
   and	
   sufficient	
   meals	
   per	
   day,	
  
approximately	
  70 percentage	
  points below	
  the	
  
target	
   for	
   the	
   program.	
   This	
   indicator	
   is	
   a	
  
complex	
   composite	
   of	
   minimum	
   number	
   of	
  
meals	
   and	
   minimum	
   dietary	
   diversity	
   for	
   the	
  
age	
  groups	
  of	
   6-­‐ 8	
  months	
  and	
   9-­‐ 24 months.	
  	
  
When	
   we	
   look	
   at	
   the	
   indicator	
   by	
   age	
   group,	
  
we	
   see	
   a	
   similar	
   pattern across	
   both	
   groups.	
  	
  
For	
  children	
  between	
  6-­‐ 8	
  months,	
  only	
  26.5%	
  
receive	
  adequate	
  dietary	
  diversity	
  and	
  number	
  
of	
  meals	
  (defined	
  as	
  4	
  food	
  groups	
  and	
  at	
  least	
  
2	
  meals	
  per	
  day).	
  	
  For	
  9-­‐ 24 months,	
  27.9%	
  receive	
  adequate	
  dietary	
  diversity	
  and	
  number	
  of	
  meals	
  
(defined	
  as	
  4	
  food	
  groups	
  and	
  at	
  least	
  3	
  meals).	
  	
  Given	
  the	
  complexity	
  of	
  this	
  indicator	
  the	
  program	
  
target	
  of	
  98%	
  was	
  highly	
  unrealistic	
  especially	
  since	
  the	
  baseline	
  value	
  was	
  close	
  to	
  0.	
  
Regular	
  Growth	
  Monitoring	
  and	
  Promotion	
   (GMP),	
  is	
  a key	
  part	
  of	
  identifying	
  children	
  with	
  growth	
  
deficiencies	
   and	
   preventing	
   further	
   nutritional	
   and	
   health	
   deterioration.	
   Health	
   personnel	
   have	
  
weighed	
   the	
   majority	
   of	
   children	
   regularly both	
   before	
   and	
   after	
   the	
   intervention.	
   	
   There	
   was	
   a	
  
statistically	
  significant	
  change	
  from	
  83.3%	
  to	
  93.7%	
  (LCL	
  91.2	
  /	
  UCL	
  96.1%).	
  	
  It	
  is	
  important	
  to	
  note,	
  
however, that	
  the	
  interpretation	
  of	
  the	
  MoH	
  health	
  card	
  posed	
  a	
  challenge for	
  survey	
  enumerators.	
  	
  
The	
  MOH	
  card	
  includes	
  both	
  a	
  weight-­‐for-­‐age	
  graph	
  and	
  an	
  area	
  to	
  note	
  the	
  weight-­‐for-­‐age and	
  the	
  
next	
  appointment	
  date.	
  	
  The	
  cards	
  were	
  filled	
  out	
  inconsistently	
  within	
  and	
  across	
  project	
  areas,	
  and	
  
in	
   many	
   cases the	
   cards	
   did	
   not	
   note	
   the	
   date of	
   the	
   previous	
   weighing	
   session.	
   Interviewers	
  
estimated dates based	
   on	
   their	
   individual	
   understanding	
   and	
   interpretation	
   of	
   the	
   cards,	
  which	
   is	
  
likely	
  to	
  have	
  introduced	
  bias	
  in	
  the	
  data.	
   Nevertheless,	
  it	
  was	
  clear	
  that	
  the	
  majority	
  of	
  children	
  are	
  
being	
  weighed	
  regularly.	
  
Intermediate	
   Result	
   2.2:	
   	
   Improve	
   mothers’ ability	
   to	
   prevent,	
   diagnose,	
   and manage	
   common	
  
childhood	
  illnesses
Under	
   IR2.2	
   FH proposed	
   to	
   address:	
   increased	
   knowledge	
   about	
   the	
   signs	
   of	
   childhood	
   illness,	
  
prevention	
   of	
   diarrhea	
   (through	
   promotion	
   of exclusive	
   breastfeeding;	
   hand-­‐washing	
   with	
   sand	
   or	
  
ash	
   using	
   household	
   hand-­‐washing	
   stations,	
   care	
   of	
   household	
   water	
   and	
   food,	
   promotion	
   of	
  
household	
   latrine	
   construction); prevention	
   and	
   treatment	
   of	
   malaria;	
   and	
   increased knowledge	
  
about	
  HIV	
  prevention.	
  	
  The	
  indicators	
  and	
  targets	
  under IR	
  2.2	
  are	
  included	
  in	
  Table	
  17.
Table	
  17:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR2.2
Percentage	
   of	
   participant	
   caregivers	
   of	
   children	
   age	
   0–24 months	
   who	
   know	
   at	
   least	
  
three	
  signs	
  of	
  childhood	
  illness	
  that	
  indicate	
  the	
  need	
  for	
  treatment	
   90%
Percentage	
   of	
   pregnant	
   women	
   or	
   caregivers	
   of	
   children	
   0-­‐24m	
   who	
   can	
   name	
  three
methods	
  of	
  preventing	
  HIV 70%
Proportion	
   of	
   caregivers	
   of	
   children	
   6-­‐24m	
   of	
   age	
   who	
   gave	
   the	
   same	
   or	
   more	
   food	
  
during	
  a	
  childhood	
  illness	
  in	
  the	
  past	
  two	
  weeks 85%
Baseline N=	
  231;	
  Final	
  N=	
  268
p=	
  .000
3%	
  
27.6%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  12:	
  Percent	
  of	
  6-­‐	
  24	
  month	
  olds	
  
with	
  adequate	
  dietary	
  diversity	
  and	
  
number	
  of	
  meals	
  
Target	
  98%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 27
The	
   percent	
   of	
   caregivers	
   of	
   children	
   0-­‐ 24
months	
  who	
  can	
  identify	
   three	
  or	
  more	
  signs	
  of	
  
childhood	
  illness has	
  increased	
  over	
  the	
  baseline	
  
at	
   a	
   statistically	
   significant	
   level.	
   	
   As	
   Figure	
   13
shows,	
   49%	
   (LCL	
   44.1%	
   /	
   UCL	
   53.9%) know	
   at	
  
least	
  three	
  signs	
  of	
  illness.	
  	
  While	
  this	
  represents	
  
a	
   gain	
   of	
   more	
   than	
   20 percentage	
   points,	
   the	
  
achievement	
   falls	
   approximately	
   30 percentage	
  
points short	
   of	
   the	
   program	
   target	
   of	
   80%.	
  	
  
Given	
   that	
   this	
   is	
   an	
   indicator	
   measuring	
  
knowledge and	
   not	
   practices	
   / behavior	
   we	
  
would	
  expect	
  to	
  see	
  higher	
  gains.	
  	
  Knowledge	
  is	
  
usually	
  the	
  easiest	
  level	
  of	
  result	
  to	
  influence,	
  as	
  
it	
   is	
   typically	
   the	
   most	
   linear	
   input-­‐output	
  
relationship	
  in	
  a	
  project	
  intervention.
Table	
  18 provides	
  a	
  break	
  down	
  of	
  knowledge	
  levels	
  by	
  typical	
  danger	
  sign	
  between	
  the	
  baseline	
  and	
  
final	
   evaluations.	
   	
   	
   As	
   the	
   table	
   shows,	
   there	
   has	
   been	
   virtually	
   no change	
   in	
   knowledge	
   levels	
  
regarding	
  danger	
  signs	
  on	
  6	
  out	
  of	
  7	
  key	
  signs.	
  	
  
The	
   only	
   one	
   that	
   showed	
   statistically	
  
significant	
  change	
  since	
  the baseline	
  is	
  “appears	
  
unwell,	
   won’t	
   play”.	
   	
   None	
   of	
   the	
   others	
   have	
  
seen	
   any	
   improvement	
   over	
   the	
   time	
   of	
   the	
  
MYAP.
A	
   total	
   of	
   26.2%	
   of	
   caregivers	
   reported	
   that	
  
their	
  child	
  between	
  0-­‐ 24 months	
  had	
  diarrhea	
  
in	
   the	
   two	
   weeks	
   previous	
   to	
   the	
  survey. This	
  
represents	
   a	
   statistically	
   significant	
   decrease	
  
compared	
   to	
   the baseline	
  of	
  36.8%,	
   suggesting	
  
that	
  behaviors	
  related	
  to	
  prevention	
  have	
  begun	
  to	
  change,	
  even	
  if	
  slightly. More	
  details	
  regarding	
  
hygiene	
  factors	
  that	
  influence	
  diarrhea	
  are	
  included	
  under	
  the	
  section	
  for	
  IR2.3,	
  below.
Malaria	
   prevention	
   has	
   also	
   improved	
   since	
   the	
   baseline.	
   Bed net	
   ownership	
   (all	
   types)	
   saw	
   a
statistically	
   significant	
   increase	
   from	
   73.6%	
   at	
  
the	
  baseline	
  to	
  86.4%	
  (LCL	
  83%	
  /	
  UCL	
  89.7%) at
the	
   final	
   evaluation.	
   	
   The	
   percent	
   of	
   children	
  
sleeping	
   under	
   bed	
   nets	
   did	
   not	
   increase	
  
significantly	
   over	
   the	
   baseline;	
   however,	
   the	
  
percent	
  of	
  HH	
  with	
  treated	
  bed	
  nets	
  increased	
  
dramatically,	
   from	
   36.8%	
   at	
   the	
   baseline	
   to	
  
90.5%	
   (LCL	
   87.5%	
   /	
   UCL	
   93.6%) at	
   the	
   final	
  
evaluation.	
  	
  Even	
  if	
   the	
  proportion	
  of	
  children	
  
sleeping	
   under	
   bed	
   nets	
   has	
   remained	
   the	
  
same,	
   almost	
   three	
   times	
   more	
   children	
   are	
  
sleeping	
  under	
  ITNs.	
  
The	
  percent	
  of	
  respondents	
  who	
  have	
  heard	
  of	
  
HIV/AIDS	
   has	
   increased	
   over	
   the	
   baseline,	
  
from	
  89.9%	
  (LCL	
  86.4%	
  /	
  UCL	
  to	
  93.3%) to	
  98.5	
  
(LCL	
  97.3%	
  /	
  UCL	
  99.7%).	
  	
  Likewise,	
  the	
  percent	
  of	
  caregivers	
  who	
  can	
  name	
  at	
  least	
  three	
  methods	
  
for	
  preventing	
  HIV has	
  increased.	
  As	
  Figure	
  14 (above) demonstrates,	
  there	
  has	
  been	
  an	
  increase	
  of	
  
Baseline	
  N=	
  295;	
  Final	
  N=	
  404
p=	
  .000
Table	
  18:	
  Recognition	
  of	
  signs	
  of	
  illness,	
  by	
  sign
Sign	
  of	
  illness BL Final Sig.
Appears	
   unwell,	
   won’t	
  
play 60% 70% Y
Won’t	
  eat	
  or	
  drink 33% 41% N
Lethargic,	
   difficult	
   to	
  
wake	
   6% 5% N
High	
  fever	
   88% 87% N
Rapid	
   or	
   difficult	
  
breathing 7% 13% N
Vomits	
  everything 8% 14% N
Convulsions 3% 1% N
Baseline	
  N=	
  266;	
  Final	
  N=	
  394
p=	
  .000
25.8%	
  
	
  	
  	
  49%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  13:	
  Percent	
  who	
  know	
  at	
  
least	
  3	
  signs	
  of	
  childhood	
  illness	
  
Target	
  80%	
  
9%	
  
27.7%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  14:	
  Percent	
  who	
  can	
  name	
  3	
  or	
  
more	
  HIV	
  preven]on	
  methods	
  
Target	
  70%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 28
approximately	
   20 percentage	
   points.	
   Nonetheless,	
   the	
   achievement	
   is	
   also	
   approximately	
   40
percentage	
   points below	
   the	
   program	
   target.	
   	
   Like	
   with	
   the	
   indicator	
   regarding	
   signs	
   of	
   childhood	
  
illness,	
  this	
  indicator	
  measures	
  knowledge	
  and	
  not	
  practices	
  or	
  behavior;	
  as	
  a	
  result	
  we	
  would	
  expect	
  
achievement	
  to	
  have	
  been	
  higher.
The	
  percent	
  of	
  children	
  experiencing	
   fever	
  in	
  the	
  two	
  weeks	
  prior	
  to	
  the	
  survey	
  remained	
  constant	
  
between	
  the	
  baseline	
  and	
  final	
  evaluations	
  (52%,	
  58%,	
  no	
  significant	
  difference).	
  	
  	
  	
  There	
  was	
  also	
  no	
  
statistically	
  significant difference	
  between	
   those	
  who	
   took	
   their	
  child	
   to	
   the	
  hospital when	
  he/	
  she	
  
had	
   fever between	
  the	
  baseline	
  and	
   final.	
  	
  There	
  is,	
  however,	
  a	
  statistically	
  significant	
  difference	
  in	
  
the	
   time	
  between	
   the	
  onset	
  of	
   fever	
  and	
   the	
  visit	
   to	
  hospital.	
  	
  At	
   the	
  baseline 53.6%	
  of	
  caregivers	
  
took	
  their	
  child	
  to	
  the	
  hospital	
  at the	
  onset	
  of	
  fever	
  (within	
  one	
  day),	
  while	
  this	
  figure	
  increased	
  to	
  
76.5%	
   (LCL	
   69.3%	
   /	
   UCL	
   83.6%) at	
   the	
   final	
   evaluation,	
   suggesting	
   that	
   there	
   has	
   been	
   an	
  
improvement	
  in	
  malaria	
  care	
  seeking	
  behavior.	
  	
  
Table	
  19	
  (below)	
  shows	
  key	
  values	
  of	
  the	
  indicators	
  relating	
  to	
  home-­‐based	
  care.	
  There	
  has	
  been no	
  
significant	
   change	
   over	
   the	
   baseline	
   in	
   the	
   proportion	
   of	
   caregivers	
   providing	
   a	
   sick	
   child	
   with	
  the	
  
same	
   or	
  more	
   food	
   during	
   illness.	
   	
   There	
  has	
   also	
   been no	
   significant	
   change	
   in	
   the	
   proportion	
   of	
  
caregivers	
  who	
  breastfed	
  a	
  sick	
  child	
  the	
  same	
  or	
  more	
  during	
  illness,	
  suggesting	
  that	
  household	
  care
is	
   still	
   lacking.	
  	
   Likewise,	
   there	
   has	
   also	
   been	
   no	
   significant	
   change	
   in	
   the	
   percent	
   of	
   children	
   who	
  
received	
   oral	
   rehydration	
   salts	
   (ORS) during	
   diarrhea,	
   although	
   at both	
   the	
   baseline	
   and	
   final	
  
evaluation	
  it	
  is	
  important	
  to	
  note	
  that	
  the	
  percent	
  of	
  children	
  receiving	
  ORS	
  was	
  relatively	
  high.
Table	
  19:	
  Key	
  treatment	
  indicators
Indicator B.L.	
  % LCL UCL F.	
  E.	
  % LCL UCL
Percent	
   of	
   caregivers	
   who	
   gave	
   children	
   0-­‐
24 months	
   the	
   same	
   or	
   more	
   food	
   during	
  
illness
37.5% 26.3% 48.7% 51.2% 40.6% 61.7%
Percent	
  of	
  caregivers	
  who	
  breastfed	
  children	
  
0-­‐ 24 months	
   the	
   same	
   or	
   more	
   during	
  
illness
51.1% 47.3% 66.9% 69.2% 60.4% 78.1%
Percent	
   of	
   children	
   0-­‐ 24 months	
   who	
  
received	
  ORS	
  during	
  diarrhea 85.2% 77.4% 92.9% 62.4% 52.5% 72.2%
Intermediate	
  Result	
  2.3:	
  	
  Improved	
  access	
  to	
  clean	
  water,	
  sanitation	
  facilities	
  and	
  essential	
  hygiene	
  
behaviors	
  (EHB)
According	
  to	
  the	
  original	
  project	
  design,	
  many	
  households	
  in	
  the	
  target	
  had	
  a low	
  level	
  of	
  access	
  to	
  
and	
  use	
  of	
  potable	
  water	
  and	
  proper	
   sanitation	
   facilities at	
   the	
   start	
  of	
   the	
  program.	
   To	
  decrease	
  
household	
   vulnerability	
   to	
   unclean	
   water	
   and	
   poor	
   sanitation,	
   FH	
   constructed wells	
   and	
   public	
  
latrines,	
   promoted	
   construction	
   of	
   simple	
   HH	
   latrines and	
   provided	
   training for	
   improved	
   hygiene	
  
behaviors through	
  Care	
  Groups.	
  	
  The	
  IPTT	
  indicators	
  and	
  targets	
  under	
  IR2.3 are	
  included	
  in	
  Table	
  19.
Table	
  20:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR2.3
Percentage	
  of	
  beneficiary	
  caregivers	
  demonstrating	
  proper	
  personal	
  hygiene	
  behaviors 95%
Percentage	
   of	
   beneficiary	
   caregivers	
   demonstrating	
   proper	
   environmental	
   hygiene	
  
behaviors 95%
Proportion	
  of	
  caregivers	
  demonstrating	
  proper	
  food/water	
  hygiene	
  behaviors 40%
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 29
FH	
   and	
   partner	
   JAM	
   constructed a	
   total	
   of	
   52 wells,	
   including	
   17 tube	
   wells	
   and	
   35 shallow	
   wells
across	
  30	
  communities	
  (approximately	
  60%	
  of	
  all	
  communities	
  in	
  the	
  MYAP).21 As	
  Figure	
  15 shows,	
  
access	
   to	
   improved	
   water	
   sources	
   has	
  
approximately	
   doubled	
   during	
   the	
   life	
   of	
  
the	
  program (LCL	
  58.4%	
  /	
  UCL	
  68.7%).	
  This	
  
is	
  a	
   statistically	
   significant	
   change	
   over	
   the	
  
baseline.	
   	
   Furthermore, of	
   those	
   with	
  
access	
   to	
   an	
   improved	
   water	
   source,	
   95%	
  
percent have access	
   year-­‐round	
   and	
   can	
  
access	
  the	
  water	
  source	
  within	
  a	
  10-­‐minute	
  
(median)	
   walk	
   from	
   their	
   home.	
   	
   It	
   is	
  
important	
   note,	
   albeit	
   anecdotally,	
   that	
  
during	
   several	
   field	
   visits	
   for	
   qualitative	
  
discussions	
   the	
   team	
  observed	
  a	
   high	
  level	
  
of	
  use	
  of	
  wells	
  constructed	
  with	
  FH	
  support.	
  	
  	
  
In	
  each	
  location	
  the	
  wells	
  were	
  in	
  constant	
  
use	
  by	
  community	
  members.	
  	
  
Hand	
  washing	
  is	
  one	
  of	
   the	
  most	
  effective	
  means	
  of	
  preventing	
  diarrheal	
  diseases,	
  along	
  with	
  safe	
  
stool	
  disposal	
  and	
  safe	
  and	
  adequate	
  household	
  water	
  supply.	
  Evidence	
  suggests	
  that	
  improved	
  hand	
  
washing	
   can	
   have	
   a	
  major	
   impact	
   on	
   public	
   health	
   and	
   significantly	
   reduces diarrheal	
   disease	
   and	
  
acute	
   respiratory	
   infection-­‐ the	
   two	
   leading	
   causes	
   of	
   childhood	
   mortality.	
   Because	
   hand	
   washing	
  
with	
  soap	
  can	
  prevent	
  the	
  transmission	
  of	
  a	
  variety	
  of	
  pathogens,	
  it	
  may	
  be	
  more	
  effective	
  than	
  any	
  
single	
   vaccine	
   or	
   hygiene	
   behavior.	
   During	
   the	
   course	
   of	
   the	
   MYAP	
   there	
   has	
   been	
   a	
   statistically	
  
significant	
   increase	
   in	
   the	
   percent	
   of	
  
caregivers	
   demonstrating	
   proper	
   personal	
  
hygiene behaviors. This	
   is	
   a	
   composite	
  
indicator	
   that	
   measures	
   (1)	
   whether	
  
caregivers	
  report	
  washing their	
  hands	
  at	
  two	
  
or	
   more	
   appropriate	
   times	
   and	
   (2)	
   whether	
  
they	
   have	
   available	
   water	
   and	
   soap/ash	
   for	
  
hand	
   washing.	
   Figure	
   16 illustrates	
   that	
  
composite	
   personal	
   hygiene	
   behaviors	
   have	
  
increased	
   approximately	
   40 percentage	
  
points over	
  the	
  baseline to	
  42.3%	
  (LCL	
  37.5%	
  
/	
   UCL	
   47.1%).	
  	
  	
   Despite	
   the	
   increase,	
  
achievement	
   on	
   this	
   indicator	
   is	
   still	
  
approximately	
   50 percentage	
   points below	
  
the	
   program	
   target.	
   	
   If	
   we	
   look	
   at	
   hand	
  
washing	
   behaviors	
   alone,	
   the	
   percent	
   of	
  
caregivers	
  who report washing their	
  hands	
  at	
  two	
  or	
  more	
  appropriate	
  times	
  has	
  virtually doubled,	
  
from	
   42.5%	
   at	
   the	
   baseline	
   to	
   83.4%	
  
(LCL 79.8%	
  /	
  UCL	
  88%).	
  	
  This	
  change	
  is	
  
statistically	
  significant.
Table	
   21 illustrates	
   the	
   detail	
  
regarding hand	
   washing.	
   	
   The	
   only	
  
response	
   that	
   did	
   not	
   change	
   since	
  
the	
   baseline	
   is	
   “never.”	
  
Approximately	
  5%	
  of	
   respondents	
  do	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   21 FH	
  Mozambique	
  program	
  data	
  provided	
  on	
  April	
  5,	
  2013
Baseline	
  N=	
  299;	
  Final	
  N=	
  404
p=	
  .000
Baseline	
  N=	
  299;	
  Final	
  N=	
  404
p=	
  .000
Table	
  21:	
  Hand	
  washing	
  practices
Hand	
  washing BL Final Sig.
Never 9% 5% N
Before	
  Food	
  preparation 44% 88.5% Y
Before	
  giving	
  food	
  to	
  children 22% 62.8% Y
After	
  defecating 33% 70.3% Y
After	
  caring	
  for	
  a	
  child	
  who	
  
defecated 27% 55.3% Y
34.1%	
  
63.1%	
  
0%	
  
20%	
  
40%	
  
60%	
  
80%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  15:	
  Percent	
  of	
  caregivers	
  with	
  
access	
  to	
  improved	
  water	
  sources	
  
2.7%	
  
42.3%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  16:	
  Percent	
  of	
  caregivers	
  
prac]cing	
  proper	
  personal	
  hygiene	
  
Target	
  95%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 30
not	
  wash	
  their	
  hands	
  at	
  any	
  of	
  the	
  appropriate	
  /	
  necessary	
  times.	
  	
  One	
  category	
  that	
  still	
  seems	
  low	
  
in	
   comparative	
   perspective	
   is	
   the	
   percent	
   of	
   respondents	
   who	
   wash	
   their	
   hands	
   after	
   caring	
   for	
   a	
  
child	
  who	
  defecated.	
  	
  It	
  increased	
  by	
  approximately	
  30 percentage	
  points but	
  is	
  still	
  far	
  less	
  common	
  
than	
  the	
  other	
  responses.	
  
Proper	
   environmental	
   hygiene	
   is	
   a	
   key	
   component	
   to	
   ensuring	
   the	
   health	
   of	
   young	
   children.	
   	
   As	
  
Figure	
  17	
  shows there	
  has	
  been	
  a	
  statistically	
  
significant	
   increase	
   in	
   the	
   percent	
   of HH	
  
practicing	
   proper	
   environmental	
   hygiene,
from	
  16.4%	
  at	
  baseline	
  to	
  59.9%	
  (LCL	
  55.1%	
  /	
  
UCL	
   64.7%) at	
   the	
   final	
   evaluation.	
  	
  	
  
Achievement	
  is	
  approximately	
  35 percentage	
  
points below	
   the	
   target.	
  	
   It	
   is	
   important	
   to	
  
point	
   out	
   that	
   in	
   the	
   baseline questionnaire
the	
   question	
   referring	
   to	
   disposal	
   of	
  
children’s	
   feces	
   was	
   phrased	
   as	
   “where	
   did	
  
children	
  go	
  to	
  the	
  bathroom?”	
  as	
  opposed	
  to	
  
“where	
  did	
  you	
  dispose	
  of	
  children’s	
   feces?”	
  	
  
The	
   question	
   was	
   not	
   modified	
   in	
   order	
   to	
  
ensure	
   comparability	
   of	
   responses between	
  
the	
   baseline	
   and	
   final	
   evaluations;	
   however,	
  
the	
   phrasing	
   of	
   the	
   question	
   may	
   have	
   also
affected	
  its	
  interpretation	
  by	
  respondents.	
  
When	
   we	
   think	
   about	
   environmental	
   hygiene	
   we	
   must	
   also	
   consider the	
   overall	
   practices	
   of	
   the	
  
family.	
  	
  To	
  reduce	
  open	
  defecation	
  and	
  support	
  the	
  program’s	
  hygiene	
  component	
  FH promoted HH	
  
latrines.	
  	
  While	
  FH	
  did	
  not	
  provide	
  materials,	
  
they	
  did	
  provide	
  training	
  and	
  orientation	
  on	
  a	
  
simple	
   model	
   for	
   HH	
   latrines	
   using	
   locally	
  
available	
   materials.	
   	
   The	
   latrine	
   is	
   a	
   simple	
  
dug	
   pit	
   (unlined),	
   covered	
   with	
   a	
   crosshatch	
  
of	
   sticks	
   and	
   mud	
   with	
   a	
   small	
   opening	
   left	
  
for	
   defecation	
   and	
   urination.	
   	
   During	
  
fieldwork	
  it was	
  clear	
  that	
  these	
  latrines	
  were	
  
widely	
   replicated	
   across	
   MYAP	
   intervention	
  
communities.	
   	
   	
   Based	
   on	
   53	
   individual	
   case	
  
observations,	
   the latrines	
   are	
   used	
   regularly
at	
   the	
   HH	
   level.	
  	
   However,	
   as	
   Figure	
   18	
  
shows,	
   only	
   40%	
   of	
   observed	
   latrines	
   had	
   a	
  
cover	
   for	
   the	
   latrine’s	
   opening,	
   creating	
   a	
  
potentially	
   significant	
   disease	
   vector	
   at	
   the	
  
household	
   level.	
  	
   Given	
   that	
   the	
   purpose	
   of	
  
the	
   latrines	
   was	
   to	
   improve	
   environmental	
   hygiene,	
   this	
   finding	
   is	
   important	
   as	
   it	
   potentially	
  
undermines	
  that	
  purpose.	
  	
  
Baseline	
  N=	
  299;	
  Final	
  N=	
  404
p=	
  .000
N=	
  53
NO	
  test	
  of	
  statistical	
  significance	
  performed
16.4%	
  
59.9%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  
Figure	
  17:	
  Percent	
  demonstra]ng	
  
proper	
  environmental	
  hygiene	
  
Target	
  95%	
  
Opening	
  
covered	
  
40%	
  
Opening	
  
Not	
  
covered	
  
60%	
  
Figure	
  18:	
  Percent	
  of	
  HH	
  latrines	
  
with	
  a	
  "lid"	
  for	
  the	
  opening	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 31
The	
   indicator	
   regarding water hygiene	
   behaviors measures the	
   degree	
   to	
   which	
   caregivers	
   take	
  
proper	
   measures	
   to	
   both	
   treat	
   and	
   store	
   water as	
   a	
   means	
   to	
   preventing	
   diarrhea.	
   As	
   Figure	
   19
shows,	
   this	
   indicator	
   value	
   has	
   seen	
   a	
  
statistically	
   significant	
   changed	
   from	
  4.7%	
  at	
  
the	
   baseline	
   to	
   29.2%	
   (LCL	
   24.8%	
   /	
   UCL	
  
33.6%) at	
   the	
   final	
   evaluation.	
  	
   This	
   falls	
  
approximately	
   10 percentage	
   points below	
  
the	
  program	
  target.	
  	
  If	
  we	
  look	
  individually	
  at	
  
the	
   component	
   indicators,	
   approximately	
  
34%	
   of	
   caregivers	
   practice	
   any	
   single	
   water	
  
treatment	
  method,	
  while	
  almost	
  all	
  (98%,	
  N=	
  
392) store	
  their water	
  in	
  containers	
  with	
  lids.	
  	
  
This	
   suggests	
   that	
   water	
   treatment	
   is	
   the	
  
major	
   barrier	
   to	
   overall	
   improved	
   food	
   and	
  
water	
   hygiene.	
  	
   It	
   is	
   unclear	
   why	
   this	
   is	
   the	
  
case,	
   although	
   the	
   increased	
   costs	
   of	
   water	
  
treatment	
  may	
   be	
   a	
   partial	
   explanation.	
   	
   To	
  
boil	
   water,	
   firewood	
   is	
   required.	
   	
   Firewood	
  
takes	
   time	
   to	
   collect, and	
   boiling	
   enough	
   water	
   for	
   a	
   family	
   would	
   dramatically	
   increase	
   both	
   the	
  
amount	
  of	
   firewood	
  and	
  the	
   time	
   required	
   to	
  collect	
  it.	
  	
  Additionally,	
  boiling	
  water	
   takes	
   time	
  and	
  
leaving	
  it	
  out	
  to	
  cool	
  is	
  sometimes	
  a	
  challenge	
  if	
  a	
  family	
  only	
  has	
  a	
  limited	
  number	
  of	
  pots	
  and	
  they	
  
are	
  needed	
  for	
  the	
  next	
  meal.22
Overview	
  and	
  Discussion	
  of	
  Health	
  Findings
Table	
  22 Comparative	
  Summary	
  of	
  Baseline	
  and	
  Final	
  Health	
  Indicator	
  Values	
  for	
  SO2
Indicator BL FE %	
  Point	
  
Difference Sig.
Essential	
  Nutrition	
  Actions
%	
  0-­‐5.9	
  who	
  started	
  breastfeeding	
  one	
  hour	
  after	
  birth 25.4% 65.8% +40.4% Y
%	
  0-­‐5.9m	
  who	
  were	
  exclusively	
  breastfed	
  in	
  the	
  last	
  24h 6% 79.4% +73.4% Y
%	
   12-­‐15	
   months	
   of	
   age	
   who	
   received	
   breast	
   milk	
   in	
   the	
  
previous	
  day 96% 94% -­‐2% N
% 6-­‐8	
  months	
  of	
  age	
  who	
  received	
  solid,	
  semi-­‐solid	
  or	
  soft	
  
foods	
  during	
  the	
  previous	
  day 78% 95.9% +17.9% N
% 6-­‐23m	
   who	
   received	
   solid,	
   semi-­‐solid	
   or	
   soft	
   foods,	
  
minimum	
  dietary	
  diversity	
  and	
  #	
  of	
  meals 3% 27.6% +24.6% Y
%	
  0-­‐23.9m	
  who	
   received nutritional evaluation	
  in	
   the	
  last	
  
4 months 83.3% 93.7% +10.4% Y
Diagnosis,	
  prevention	
  and	
  treatment	
  of	
  illness
%	
  Caregivers of	
  children 0–23	
  who	
  know	
  at	
  least	
  3	
  signs	
  of	
  
childhood	
  illness	
  that	
  indicate	
  the	
  need	
  for	
  treatment	
   25.8% 49% +23.2% Y
%	
   Of pregnant	
   women	
   or	
   caregivers	
   of	
   children	
   0-­‐23m	
  
who	
  can	
  name	
  3	
  methods	
  of	
  preventing	
  HIV 9% 27.7% +18.7% Y
%	
   Caregivers of	
   children	
   6-­‐23.9m	
   who	
   gave	
   the	
   same	
   or	
  
more	
  food	
  during	
  a	
  childhood	
  illness	
  in	
  the	
  past	
  two	
  weeks 37.5% 51.2% +13.7% N
Water,	
  sanitation	
  and	
  hygiene
%	
   Caregivers demonstrating	
   proper	
   personal	
   hygiene	
  
behaviors 2.7% 42.3% +39.6% Y
%	
   Caregivers demonstrating	
   proper	
   environmental	
   16.4% 59.9% +43.5% Y
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   22 Email	
  correspondence	
  with	
  FH	
  Health	
  Advisor,	
  May	
  24,	
  2013
Baseline	
  N=	
  299;	
  Final	
  N=	
  404
p=	
  .000
4.7%	
  
	
  29.2%	
  
0%	
  
10%	
  
20%	
  
30%	
  
40%	
  
50%	
  
60%	
  
70%	
  
80%	
  
90%	
  
100%	
  
Baseline	
   Final	
  	
  
Figure	
  19:	
  Propor]on	
  demonstra]ng	
  
proper	
  water	
  hygiene	
  
Target	
  40%	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 32
hygiene	
  behaviors
%	
   Caregivers demonstrating	
   proper	
   food/water	
   hygiene	
  
behaviors 4.7% 29.2% +24.5% Y
Of	
  the	
  12	
  health	
  indicators	
  included	
  in the	
  IPTT	
  9	
  of	
  them (75%)	
  experienced	
  statistically	
  significant	
  
changes	
   in	
   values	
   between	
   the	
   baseline	
   and	
   final	
   evaluations.	
   	
   Three	
   (25%)	
   did	
   not	
   experience	
  
statistically	
  significant	
  change	
  over	
  the	
  program	
  period.	
   All	
  indicator	
  values fell	
  below	
  the	
  expected	
  
targets	
   for	
   the	
   program,	
   although	
   the behavior	
   change targets	
   may	
   have	
   been	
   unrealistically	
  
ambitious,	
  especially	
  given	
  the	
  two-­‐year	
  window	
  within	
  the	
  modified	
  Care	
  Group model.	
  
Among essential nutrition	
   actions,	
   breastfeeding	
   behaviors	
   showed	
   marked	
   improvement,	
   both	
   in	
  
term	
   of	
   early	
   initiation and	
   exclusive	
   breastfeeding	
   of	
   infants 0-­‐ 6m.	
   The	
   percent	
   of	
   children	
  
breastfed	
  within	
  one	
  hour	
  of	
  birth	
  still	
  falls	
  behind	
  the	
  national	
  average	
  of	
  76%	
  (DHS	
  2011),	
  but	
  the	
  
percent	
   of	
   children	
  exclusively	
   breastfed	
   in	
  MYAP	
   project	
  areas	
   now	
  exceeds	
   the	
   national	
  average
(42.8%) by	
  almost	
  35 percentage	
  points at	
  79.4%.	
   Continued	
  breastfeeding of	
  children	
  12-­‐15m	
  was	
  
already	
  very	
  high	
  at	
   the	
  baseline,	
  essentially	
   reaching	
   the	
   target	
  before	
   the	
  project	
  began, and	
   the	
  
survey	
  results	
  show	
  that	
  those	
  behaviors	
  continued	
  through	
  the	
  project.	
  	
  
Rates	
  of complementary	
   feeding	
  for	
  children	
  6-­‐ 8	
  months	
  did	
  not	
  change	
  during	
  the	
  project. Those	
  
values	
  were	
  already	
  higher	
  than	
   the	
  national	
  average	
  of	
  68.5%	
  (DHS)	
  at	
  baseline.	
  	
  However,	
  values	
  
for	
  “adequate”	
  complementary	
  feeding	
  (minimum	
  dietary	
  diversity	
  and	
  number	
  of	
  meals) improved
by	
   25 percentage	
   points during	
   the	
   project.	
   	
   Before	
   the	
   MYAP those	
   values	
   were	
   12 percentage	
  
points below	
   the	
   national	
   average	
   of	
   15%; they	
   are	
   now	
   above	
   that	
   average	
   by	
   approximately	
   10
percentage	
  points.
The	
  indicator values regarding	
  diagnosis,	
  prevention	
  and	
   treatment	
  of	
  illness are	
  the	
  most	
  variable.	
  	
  
Although	
   the	
   percentage of	
   caretakers	
   who	
   can	
   name	
   at	
   least	
  three signs	
   of	
   illness	
   has	
   increased	
  
from	
  26%	
  to	
  49%,	
  none	
  of	
  the	
  individual	
  measures	
  of	
  knowledge	
  for	
  any	
  of	
  the individual	
  signs	
  has	
  
changed	
   at	
   a	
   statistically	
   significant	
   level	
   since	
   the	
   baseline.	
   	
   This	
   was	
   particularly	
   surprising	
   since	
  
these	
  indicators	
  gauge	
  knowledge	
  rather	
  than	
  behaviors.	
  	
  
There	
  has	
  been	
  a	
  statistically	
  significant	
   change	
  in	
  cases	
  of diarrhea.	
  	
  These	
  are presumably	
  linked	
  
with	
  improved	
  feeding	
  habits	
  as	
  well	
  as improved	
  hygiene	
  and	
  sanitation (under	
  IR2.3),	
  thus	
  helping	
  
prevent	
  diarrheal	
  disease.	
   Malaria	
  prevention	
  has also	
  improved	
  thanks	
  to	
  high	
  adoption	
  of	
  ITNs,	
  at	
  
approximately	
   90%	
   of	
   all caregivers	
   interviewed.	
   	
   This	
   represents a	
   rate	
   of	
   ITN	
   use	
   that	
   is	
   40
percentage	
  points	
  higher	
  than the	
  national	
  (rural)	
  average	
  of	
  49%	
  (DHS).	
  	
  Reports	
  of	
  fever	
  at	
  the	
  final	
  
evaluation	
   remained	
   at	
   par	
   with	
   level	
   of	
   the	
   baseline, which	
   may	
   be	
   due	
   to	
   the	
   fact	
   that	
   the	
  
proportion	
  of	
  children	
  sleeping under	
  nets	
  has	
  not	
  changed	
  significantly.	
  	
  One	
  might	
  also	
  expect that	
  
some	
   of	
   these	
   cases	
   of	
   fever	
  are due	
   to	
   other	
   causes,	
  although	
   the	
   survey	
   did	
   not	
  request	
   details	
  
regarding	
  the	
  causes	
  of	
  fever. The	
  percent	
  of	
  caregivers	
  with	
  knowledge	
  of	
  three	
  or	
  more	
  methods	
  to
prevent	
  HIV	
  has	
  also	
  increased.
Achievement	
   for	
   treatment	
   of	
   Illness	
   is	
   somewhat	
   mixed.	
   	
   On	
   the	
   one	
   hand	
   there	
   has	
   been	
   no	
  
increase	
   in	
   the	
   proportion of	
   caregivers	
   who	
   bring	
   their	
   child	
   to	
   the	
   hospital	
   when	
   he/she	
   has	
   a	
  
fever,	
  but	
  on	
   the	
  other	
  hand	
  there	
  has	
  been	
  a	
   statistically	
   significant	
  increase	
  in	
   the	
  proportion	
  of	
  
caregivers	
  who	
  do	
  so	
  within	
  a	
  day	
  of	
   the	
  outbreak	
  of	
   fever.	
  	
  This	
  suggests	
  both	
  increased	
  vigilance	
  
and	
   understanding	
   of	
   the	
   potential	
   gravity	
   of	
   fever	
   among	
   young	
   children. At	
   the	
   same	
   time,	
  
however,	
  there	
  has	
  been	
  no	
  statistically	
  significant	
  change	
  in	
  the	
  percent	
  of	
  caretakers	
  who	
  provide	
  
the	
  same	
  or	
  more	
  food	
  OR	
  the	
  same	
  or	
  more	
  breast	
  milk	
  during	
  illness.	
  
Indicator	
   values related	
   to	
   water,	
   sanitation	
   and	
   hygiene have	
   all	
   seen	
   statistically	
   significant	
  
increases	
  over	
  baseline	
  levels.	
  	
  Caretaker	
  access	
  to	
  improved	
  water	
  sources	
  has	
  nearly	
  doubled,	
  from	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 33
34%	
   to	
   63%.	
   	
   Personal	
   hygiene	
   behaviors have also	
   increased	
   significantly-­‐ by	
   approximately	
   40
percentage	
   points over	
   the	
   baseline.	
   	
   More	
   families	
   have	
   soap	
   and	
   water	
   available	
   to	
   them	
   and	
  
significantly	
  more	
  caretakers	
  report	
  washing	
  hands	
  at	
  all	
  the	
  necessary	
  times.	
  	
  Strangely,	
  a	
  relatively	
  
low	
   proportion	
   of	
   caretakers	
   (55%)	
   report	
   washing their hands	
   after	
   caring	
   for	
   a	
   child	
   who	
   has	
  
defecated.	
  	
  
Environmental	
   hygiene	
   outcomes	
   are mixed.	
   	
   There	
   has	
   been	
   statistically	
   significant	
   increase	
   in	
  
proper excreta	
   disposal-­‐ from	
   16%	
   to	
   60%.	
   	
   At	
   the	
   same	
   time,	
   observations	
   of	
   more	
   than	
   50	
   HH	
  
latrines	
  showed	
  potential	
  disease	
  vectors	
  resulting	
  from	
  improper	
  maintenance	
  and	
  care.	
  
Indicator	
   values	
   for	
   water	
   treatment	
   and	
   water	
   storage	
  are	
   also	
   variable.	
   Ninety-­‐eight	
   percent of	
  
caretakers	
  store	
  water	
  in	
  containers	
  with	
  lids,	
  but	
  only	
  approximately	
  34%	
  use	
  any	
  water	
  treatment	
  
method.	
  When	
  taken	
  as	
  a	
  composite value	
  only	
  approximately	
  29%	
  of	
  caretakers	
  show both	
  proper	
  
storage	
  and	
  treatment	
  of	
  water.	
   Nonetheless, this	
  represents	
  a	
  statistically	
  significant	
  increase	
  of	
  25
percentage	
  points over	
  the	
  baseline	
  value	
  of	
  4%.
Some	
  of	
  the	
  health	
  gains	
  have	
  been	
  more	
  modest	
  than	
  might	
  have	
  been	
  expected.	
  	
  There	
  are	
  three	
  
related factors that	
  may explain	
  this,	
  all	
  of	
  which	
  are	
  linked	
  to	
  the	
  way	
  the	
  Care	
  Group	
  methodology	
  
was	
  implemented in	
  the	
  MYAP:
(1)	
   The	
   time	
   frame	
   of	
   Care Group	
   support-­‐ by	
   design	
  the	
   MYAP envisioned	
   support	
   through	
   Care	
  
Groups	
   only	
   during	
   the	
   0-­‐24 month	
   interval	
   of	
   children’s	
   lives rather	
   than	
   the	
   0-­‐ 60-­‐month	
   time	
  
period	
  envisioned	
  in	
  the	
  original	
  Care	
  Group	
  model.	
   Sustained	
  behavior	
  change	
  in	
  most	
  contexts is	
  
likely	
  to	
  require	
  more	
  than	
  two	
  years of	
  training	
  and	
  follow-­‐up.	
   And	
  without	
  sustained	
  follow	
  up	
  for	
  
both	
   children	
   and	
   mothers	
   after	
   children	
   turn	
   two,	
   the	
   gains	
   in	
   behaviors	
   and	
   knowledge	
   could	
  
reverse.	
  In	
  short, the	
  modifications	
  of	
  the	
  CG	
  model	
  from	
  a	
  five-­‐year	
  to	
  a	
  two-­‐year process	
  may	
  have	
  
limited both	
  short-­‐term	
  achievement	
  and	
  long-­‐term	
  sustainability. Future	
  programs	
  might consider	
  
including	
  mothers	
   of	
   all	
   children	
   under	
   five	
   and	
   keeping	
   them	
   participating	
   in	
   the	
   CGs	
   for	
  the	
   full
length	
  of	
  the	
  project.	
  	
  This	
  would	
  be	
  more	
  consistent	
  with	
  the intention	
  of	
  the	
  Care	
  Group model.
(2)	
  Incentive	
  plans	
  for	
  Mother	
  Leaders	
  reduced	
  participation	
  in	
  Care	
  Groups-­‐ while	
  it	
  is	
  clear	
  that	
  Care	
  
Groups	
   support	
   widespread	
   replication	
   of	
   appropriate	
   knowledge,	
   practices	
   and	
   behavior,	
  Mother	
  
Leaders	
  in	
  all	
  FGD	
  pointed	
  to	
  high	
  attrition	
  among	
  participant	
  mothers	
  as	
  a	
  key	
  factor	
  undermining	
  
results.	
   When	
   asked	
   why	
   this	
   happened	
   mothers	
   across	
   all	
   FGDs	
   unanimously	
   echoed that	
   the	
  
distribution	
  of capulanas (colorful	
   traditional	
  wraps)	
   to	
  Mother	
  Leaders	
  had	
  created	
  large	
   tensions	
  
between	
  Mother	
  Leaders	
  and	
  many	
  participant	
  mothers.	
  	
  The	
  capulanas	
  were	
  part	
  of	
  FH’s	
  strategy	
  to	
  
provide	
  a	
  non-­‐monetary	
  incentive	
  to	
  Mother	
  Leaders	
  for	
  their	
  work	
  in	
  Care Groups.	
  	
  The	
  capulanas-­‐
printed	
   with	
   the	
   program’s	
   key	
   messages regarding	
   nutrition	
   and	
   health-­‐ resulted	
   in	
   significant	
  
jealousy.	
   	
   According	
   to	
   Mother	
   Leaders	
   across	
   all FGDs	
   as	
   many	
   as	
   half	
   of	
   all	
   mothers	
   of	
   0-­‐ 23.9	
  
stopped	
  participating	
  in	
  Care	
  Group	
  sessions	
  as	
  a	
   result.	
  	
  This	
  would	
  help	
  explain	
  why	
  some	
  of	
   the	
  
key	
  indicators-­‐ especially	
   regarding	
  knowledge-­‐ were	
  lower	
   than	
  might	
   have	
   been	
  expected.	
  	
  	
  This	
  
finding	
  is	
  supported	
  by	
  the	
  quantitative	
  data,	
  which	
  shows	
  that	
  only	
  70%	
  of	
  mothers	
  had	
  received	
  a
health	
  promotional	
  visit	
  in	
  the	
  previous	
  month.	
  	
  Fewer	
  than	
  half	
  had	
  received	
  more	
  than	
  one	
  visit.
(3)	
   The	
   current	
   course	
   of	
   modules	
   is	
   not	
   appropriate	
   for	
   a	
   two-­‐year	
   Care Group	
   cycle-­‐ there	
   are	
  
currently	
  nine	
  modules	
  in	
  the	
  program,	
  each	
  of	
  which	
  is	
  taught	
  over	
  the	
  course	
  of	
  approximately	
  6	
  
months.	
  	
  This	
  means	
  that	
  it	
  would	
  take	
  almost	
  5	
  full	
  years to	
  get	
  through	
  all	
  9	
  modules.	
  This	
  suggests	
  
that	
  NO	
  SINGLE	
  mother	
  in	
  the	
  program	
  has	
  received	
  all	
  nine	
  modules.	
  	
  In	
  fact,	
  under	
  this	
  model,	
  any	
  
single	
  mother	
  would	
  only	
  receive	
  approximately	
  half	
  of	
  the	
  full	
  modules	
  during	
  her	
  time	
  in	
  the	
  Care	
  
Group.	
  	
  This	
  may	
  help	
  explain	
  the	
  relatively	
  low	
  performance	
  on	
  the	
  key	
  knowledge	
  indicators.	
  	
  For	
  
future	
  programs	
  where	
  USAID	
  requests the	
  implementation	
  of	
  Care	
  Groups	
  for	
  only	
  the	
  0-­‐ 24	
  month	
  
age	
  group it	
  might	
  make	
  sense	
  to	
  consider	
  reducing	
  the	
  length	
  of	
  the	
  modules	
  (and	
  the	
  schedule	
  of	
  
trainings) to	
  ensure	
  that	
  all	
  participant	
  mothers	
  receive	
  all	
  relevant	
  messages	
  during	
  their	
  time	
  in	
  the	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 34
Care	
  Groups.	
   Refresher	
   training	
   could	
   be	
   provided	
  on	
  an	
   ongoing	
   basis	
   throughout	
   the	
  life	
   of	
   the	
  
program.	
  	
  
Despite	
  these	
  challenges,	
  the	
  changes	
  in	
  indicator	
  values	
  suggest	
  that	
  MYAP’s	
  health	
  activities	
  have	
  
led	
   to	
   changes	
   in	
   key	
   behaviors	
   that	
   serve	
   as	
   the	
   foundation	
   for	
   better	
   food	
   utilization.	
   Through	
  
improved	
   feeding	
   practices,	
   improved	
   hygiene,	
   decreased	
   diarrhea	
   and	
   improved	
  
prevention/treatment	
   of	
   other	
   illnesses	
   such	
   as	
  malaria	
   children	
   receive-­‐ and	
   can	
   properly	
   utilize-­‐
food.	
   	
   Changes	
   in	
   the	
   status	
   of	
   nutrition	
  among	
   participant	
   populations	
  will	
   be	
   verified	
   during	
   the	
  
anthropometric	
  exercise in	
  October-­‐November	
  2013.
C.	
   Strategic	
  Objective	
   3	
   (Community	
   Capacity	
   Building): Increased	
  community	
  capacity	
   to	
  
influence	
  factors	
  that	
  affect	
  food	
  security	
  and	
  resiliency	
  to	
  shocks
The	
  Community	
  Capacity	
  Building	
  component	
  of	
  the MYAP sought to	
  develop	
  community	
  capacity	
  to	
  
promote	
  and	
  sustain	
  local	
  development	
  initiatives	
  by	
  (1)	
  building	
  the	
  capacity	
  of	
  existing	
  formal	
  and	
  
informal	
  community	
  leaders; (2)	
  supporting	
  the	
  development	
  of	
  local	
  infrastructure	
  and	
  capacity	
  or	
  
their	
   maintenance;	
   and	
   (3)	
   increasing	
   community	
   capacity	
   to	
   mitigate	
   shocks.	
   	
   In	
   short,	
   this	
   SO	
  
underpins the	
  health	
  and	
  agriculture	
  findings	
  by	
  ensuring	
  that	
  local	
  communities	
  have	
  the	
  skills	
  and	
  
experience	
  to	
  manage	
  and	
  prioritize	
  community	
  development.	
  	
  Findings	
  for	
  the	
  indicators	
  relating	
  to	
  
SO3	
  are	
  included	
  below,	
  organized	
  by	
  Intermediate	
  Result.	
  	
  
Intermediate	
   Result	
   3.1:	
   Increased	
   leadership	
   capacity	
   of	
  existing	
   formal	
  and informal	
   community	
  
leaders	
  to	
  address	
  factors	
  that	
  affect	
  food	
  security
Under	
  IR3.1,	
  FH	
  built	
  the	
  capacity	
  of	
  Community	
  Development	
  Committees to	
  play	
  an	
  active	
  role	
  in	
  
community	
  development	
  efforts.
Table	
  23:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR3.1
Number	
  of	
  HCA/PRA	
  conducted	
  and	
  reports	
  prepared 48
Number	
  of	
  CDC	
  leaders	
  participating	
  in	
  training 1162
Percent of	
  CDC	
  members	
  who	
  are	
  women 38%
FH	
   formed	
   48	
   CDCs,	
   consisting	
   of	
   between	
   10	
   and	
   12	
   members	
   selected	
   through	
   a	
   process	
   of	
  
community	
  consensus.	
  	
  CDC	
  members	
  were	
  trained	
  in	
  internal	
  governance,	
  roles	
  and	
  responsibilities,	
  
approaches,	
   community	
   development	
   and	
   conflict	
   resolution.	
  	
   A	
   total	
   of	
   570	
   CDC	
   members	
   were	
  
trained	
  across	
  all	
  target	
  communities.	
  	
   FH	
  implemented	
  a	
  total	
  of	
  768	
  trainings.	
  	
  In	
  total	
  33%	
  of	
  CDC	
  
members	
  across	
  all	
  communities	
  are	
  women.	
  	
  Given	
  that	
  many	
  of	
  the	
  intervention	
  communities	
  are	
  
Muslim-­‐ where	
   community	
   participation	
   among	
   women	
   is	
   limited	
   by	
   culturally-­‐prescribed	
   gender	
  
norms-­‐ this	
  represents	
  an	
  important	
  achievement.23
FGD	
  were	
  planned	
  with	
  three	
  CDCs	
  in	
  three	
  different	
  communities	
  (one in	
  each	
  district)	
  in	
  order	
  to	
  
gauge	
   leadership	
   capacity	
   and	
   the overall	
   capacity	
   of	
   CDC	
   members	
   as	
   community	
   leaders.	
   	
   As	
   a	
  
result	
   of	
   scheduling	
   conflicts,	
   only	
   two	
   FGDs	
   were completed.	
   Of	
   the FGDs	
   conducted only one	
  
showed	
  a	
  basic	
  understanding	
  of	
  CDC roles	
  and	
  responsibilities,	
  development	
  plans	
  (past	
  and	
  future)	
  
and	
  sustainability	
  plans	
  for	
  community	
  infrastructure projects	
  (solar	
  powered	
  improved	
  water	
  pump	
  
and	
   market	
   stall).	
   	
   	
   The	
   other	
   group	
   could	
   not	
   describe	
   the	
   purpose	
   or	
   function	
   of	
   their	
   CDC or	
  
provide	
  any	
  details	
   regarding	
   the	
  community	
  development	
  activities	
  or	
  infrastructure	
  projects	
   that	
  
had	
  been	
  implemented with	
  FH	
  support.	
   The	
  FGDs gave	
  a	
  mixed	
  impression	
  of	
  the	
  CDCs.	
  	
  Although	
  
clearly	
  two	
  FGDs are	
  not representative	
  of	
  all	
  CDC	
  capacity,	
  they	
  do	
  illustrate	
  the	
  likely	
  variation	
  that	
  
exists in	
  CDC	
  capacity.	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   23 FY09	
  Annual	
  Results	
  Report
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 35
The	
   weaker	
   of	
   the	
   two	
   CDCs	
   said	
   that	
   FH	
   CCB	
   staff/volunteers	
   did	
   not	
   visit	
   their	
   community	
   very	
  
regularly,	
  which	
  might provide	
  partial	
  insight	
  into	
  the	
  observed	
  skills	
  deficits.	
  	
  Community	
  capacity	
  
building	
   requires	
   significant	
   time	
   and human	
   resource	
   accompaniment	
   to	
   be	
   successful.	
   	
   It	
   does	
  
seem	
   that	
   consistent	
   follow-­‐up	
   might	
   have	
   been	
   a	
   challenge	
   given	
   the	
   geographic	
   scope	
   of	
   the	
  
program	
   and	
   the	
   limited	
   staff	
   compared	
   to	
   the	
   program’s	
   other	
   components.	
  	
   The	
   differences	
   in	
  
staffing	
   levels	
   are	
   clear:	
   	
   (1)	
   the	
   Health	
   team	
   has	
   a	
   Manager,	
   two Coordinators,	
   seven Program	
  
Support	
  Officers	
  (PSOs)	
  and	
  45	
  community-­‐based	
  health	
  and	
  nutrition	
  facilitators;	
  (2)	
  the	
  Agriculture	
  
team	
  has	
  a	
  Manager,	
  an	
  Agriculture	
  Coordinator,	
  a	
  Savings	
  Coordinator,	
  six Agricultural	
  PSOs,	
  three
savings	
   PSOs	
   and	
   149	
   community-­‐based	
   agriculture	
   and	
   savings	
   promoters;	
   (3)	
   the	
   community	
  
capacity-­‐building	
   team	
   has	
   a	
   Manager,	
   a	
   coordinator	
   and	
   nine PSOs.24	
  	
   Although	
   the	
   CCB	
   team	
  
coordinates	
   with	
   the	
   other	
   programs’	
   community-­‐based	
   staff,	
   it	
   is	
   easy	
   to	
   imagine	
   that	
   the	
  
significantly	
  smaller	
  CCB	
  staff	
  would	
  face	
  difficulties	
  in	
  ensuring	
  consistent	
  follow-­‐up.	
  
According	
   to	
   the	
  M&E	
  Manager,	
  while	
  CDC	
   training	
  is	
   tracked,	
  overall	
  capacity	
  of	
   the	
  groups	
  is	
  not	
  
systematically	
   measured.	
   	
   Such	
   efforts	
   could	
   help	
   to	
   identify	
   specific	
   skills	
   gaps	
   and	
   support	
  
appropriate	
  planning	
  based	
  on	
  identified	
  problem	
  areas.	
  	
  
Intermediate	
  Result	
  3.2:	
  Increased	
  community	
  level	
  economic	
  infrastructure/assets
FH	
  supported	
  CDCs	
  across	
  MYAP	
  communities	
  in	
  the	
  development	
  of	
  community-­‐level	
  infrastructure,	
  
including	
  (1)	
  Access	
  roads	
  and	
  small	
  bridges;	
  (2)	
  improved	
  wells	
  to	
  protect	
  against	
  contamination	
  at	
  
source,	
   (3)	
   market	
   stalls	
   and	
   (4)	
   first	
   aid	
   facilities.	
   	
   Indicators	
   and	
   targets	
   for	
   this	
   IR	
   are	
   included	
  
below.
Table	
  24:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR3.2
%	
  of	
  CDCs	
  that	
  complete	
  community	
  development	
  projects 92%
#	
  of	
  community	
  development	
  projects	
  completed 66
Amount	
  contributed	
  by	
  community	
  for	
  small-­‐scale	
  community	
  projects	
  (US$) $97,000
As	
   of	
   April	
   5,	
   2013	
   a	
   total	
   of	
   87.5%	
   (42	
   of	
   48)	
   of	
   all	
   CDCs	
   had	
   completed	
   infrastructure	
   projects,	
  
approximately	
   4.5 percentage	
   points	
   (two CDCs)	
   below	
  
the	
   target.	
   In	
   total,	
   CDCs	
   have	
   completed	
   68	
   community	
  
infrastructure	
   projects. 25 These	
   projects refer	
   to	
  
infrastructure	
   activities	
   undertaken	
   in	
   partnership	
  
between	
   the	
   CDC	
   and	
   FH,	
   in	
   which	
   the	
   CDC	
   prioritized	
  
projects	
   and	
   secured	
   labor,	
   while	
   FH	
   provided	
   basic	
  
materials	
   for	
   their	
   construction,	
   such	
   as	
   tin roofing	
   and	
  
cement.	
   	
   The	
   market	
   stalls,	
   meeting	
   spaces	
   and	
  
schoolrooms	
   are	
   very	
   simple	
   structures,	
   in	
   most	
   cases	
  
essentially	
  four	
  pillars	
  and	
  a	
  roof.	
  	
  In	
  addition	
  to	
  planning,	
  
coordination	
   and	
   construction,	
   part	
   of	
   the	
   CDC	
   role	
   is	
   to	
  
ensure	
   that	
   the	
   community	
   infrastructure	
   is	
   adequately	
   used	
   and	
   maintained.	
   	
  	
   Although	
   the	
  
timeframe	
   of	
   the	
   evaluation	
   did	
   not	
   permit	
   systematic	
   observations	
   all	
   of	
   the	
   CDC	
   community	
  
projects,	
   during	
   the	
   two	
   CDC	
   FGDs,	
   observations	
   of	
   two	
   market	
   stalls	
   were conducted	
   (in	
   two	
  
communities	
  across	
   two	
  districts).	
  	
  Construction	
  on	
  one	
  had	
  been	
  halted	
   since	
  October	
  due	
   to	
   the	
  
illness	
   of	
   the	
   mason	
   who	
   had	
   agreed	
   to	
   complete	
   the	
   structure	
   free	
   of	
   charge	
   (community	
  
contribution),	
  while	
  the	
  other	
  although	
  complete,	
  was	
  not	
  being	
  used.	
  	
  	
  FH	
  staff	
  also	
  mentioned	
  at	
  
least	
  one	
  other	
  market	
  stall	
  that	
  had	
  been	
  completed	
  but	
  was	
  not	
  utilized.	
  	
  
When	
  asked	
  about	
  this,	
  FH	
  staff	
  explained	
  that	
  the	
  stalls	
  were	
  not	
  being	
  used	
  as	
  a	
  result	
  of	
  delays	
  on	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   24 FH	
  MYAP	
  Organizational	
  and	
  Staffing	
  Structure	
  Chart,	
  updated	
  August	
  2012 25 This	
  data	
  is	
  derived	
  from	
  a	
  table	
  provided	
  by	
  the	
  FH	
  M&E	
  Manager	
  on	
  April	
  5,	
  2013
Table	
  25:	
  CDC	
  Projects
Project	
  type Number
Schoolrooms 28
Tube	
  wells/shallow	
  wells 5
Market	
  stalls 14
Health	
  posts 14
Bridges/road repair 3
Maternal	
  care	
  center 1
Meeting	
  spaces 3
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 36
the	
  part	
  of	
  the	
  district	
  governments,	
  who	
  must	
  approve	
  and	
  then	
  formally	
  “inaugurate”	
  community	
  
infrastructure	
  before	
  it	
  can	
  be	
  used	
  by	
  the	
  communities.	
   It	
  is	
  unfortunate	
  that	
  items	
  that	
  have	
  been	
  
prioritized	
   by	
   communities	
   have	
   not	
   received	
   the	
   same	
   degree	
   of	
   priority	
   on	
   the	
   part	
   of	
   local	
  
government	
  officials.	
  	
  This	
  might	
  be	
  an	
  area	
  where	
  FH	
  could	
  exert	
  a	
  bit	
  more	
  influence.	
  	
  
In	
  addition to	
  this	
  basic	
  infrastructure,	
  the	
  FH	
  Community	
  Capacity Building	
  team	
  oversaw	
  water	
  and	
  
sanitation	
   infrastructure	
   development	
   and	
   training	
   in	
   support	
   of	
   the project’s health	
   component	
  
(SO2). For	
  this	
  purpose,	
  the	
  MYAP	
  subcontracted	
  with	
  Joint	
  Aid	
  Management,	
  a	
  South	
  African	
  NGO,
to	
  construct	
  school/public	
  latrines as	
  well	
  as	
  tube	
  wells and	
  shallow	
  wells.	
  	
   As	
  mentioned	
  under	
  SO2,
a	
  total	
  17	
  tube	
  wells	
  and	
  35	
  shallow	
  wells	
  (52	
  wells	
  in	
  total)	
  were	
  constructed	
  across	
  30	
  communities	
  
(approximately	
   60%	
   of	
   all	
   communities	
   in	
   the	
   MYAP).	
  	
  	
   This	
   is	
   likely	
   to	
   have	
   contributed	
   to	
   the	
  
increased	
  access	
  to	
  improved	
  water	
  sources	
  cited	
  under	
  the	
  health	
  component.	
  	
  At	
  the	
  same	
  time,	
  
JAM	
   supported	
   the	
   construction	
   of	
   45 school	
   latrines.	
   	
   Public	
   latrines	
   were	
   not	
   constructed.	
  	
  	
   As	
  
mentioned	
  above,	
  the	
  community	
  capacity	
  building	
  team	
  also	
  provided	
  orientation	
  and	
  training	
   for	
  
the	
  construction	
  of	
   simple	
  pit	
  latrines.	
  	
  The	
  evaluation	
   team	
   completed	
   structured observations	
  of	
  
nine	
   wells,	
   ten	
   school	
   latrines	
   and	
   53	
   HH	
   latrines	
   built	
   under	
   this	
   component	
   of	
   the	
   program to	
  
gauge	
  quality	
  of	
  use	
  and	
  maintenance.	
  	
  
Observations	
  of	
  the	
  tube	
  wells and	
  shallow	
  wells	
  (nine total,	
  5%	
  of	
  those	
  constructed)	
  showed	
  that	
  
all	
   wells	
   were-­‐ unsurprisingly-­‐ regularly	
   used.	
   	
   According	
   to	
   conversations	
   with	
   a	
   community	
  
member,	
  water	
   from	
  one	
  of	
  the	
  wells	
   (in	
  Mitumbate,	
  Mocimboa)	
  is	
  constantly	
  red,	
  rusty	
  and	
  salty,	
  
which	
   means	
   it	
   is	
   unsuitable	
   for	
   consumption.	
   	
   Six	
   of	
   the	
   wells	
   (66%	
   of	
   all	
   observations)	
   showed	
  
pooled	
  water	
  around	
  the	
  water	
  collection	
  chamber.	
  	
  Although	
  pooled	
  water	
  will	
  be	
  difficult	
  to	
  avoid	
  
altogether at	
   an	
   open	
   well,	
   if	
   unchecked	
   over	
   time	
   it	
   could	
   become	
   a	
   vector	
   for	
   mosquito-­‐borne	
  
illness	
   such as	
   malaria	
   or	
   even	
   elephantiasis,	
   both	
   of	
   which are	
   endemic.	
   	
  	
   Three	
   of	
   the	
   wells	
  
presented	
  constant	
  leaks	
  around	
   the	
  pump.	
  	
   Although	
   five	
  of	
   the	
  communities	
  visited	
  have	
  water	
  
committees,
26 maintenance	
  seems	
  to	
  be	
  a	
  nascent	
  issue with the	
  wells.	
  	
  Given	
  that	
  the	
  program	
  has	
  
not	
  yet	
  ended,	
  one	
  might	
  expect	
  these	
  small	
  challenges	
  to	
  grow	
  once	
  the	
  program	
  ends.
Observations	
  of	
  the	
  10	
  school	
  latrines	
  (approximately	
  20%	
  of	
  the	
  total	
  constructed)	
  showed	
  that	
  four
of	
   them	
   (40%)	
   were	
   not	
   used	
   regularly,	
   eight	
  
(80%	
   of	
   observations,	
   18%	
   of	
   total	
   constructed)	
  
offered	
   no	
   area	
   for	
   hand	
   washing and	
   three	
  
(30%)	
  were	
  not	
  maintained,	
  exhibiting	
  high	
  grass	
  
around	
   the	
  structure.	
  	
  Again,	
  when	
  asked	
  about	
  
these	
   findings	
   FH	
   staff	
   explained	
   that	
   the	
  
protocol	
   for	
   latrines	
   is	
   similar	
   to	
   that	
   described
above	
   regarding	
  market	
   stalls:	
  upon	
   completion	
  
of	
  infrastructure	
  the	
  MoH	
  assumes	
  responsibility	
  
for	
  maintenance	
  through	
  a system	
  in	
  which	
  both	
  
a	
   teacher	
   and	
   a	
   student	
   are	
   selected	
   to	
   be	
  
responsible	
  for	
  the	
  hygiene	
  and	
  sanitation	
  of	
  the	
  
latrine.	
  	
  The	
  latrines	
  were	
  formally	
  “delivered”	
  to	
  
the	
   MoH by	
   FH.	
   	
   Based	
   on	
   MoH	
   policy,	
   it	
   was	
  
subsequently	
  the	
  responsibility	
  of	
  the	
  school	
  to	
  ensure	
  proper	
  use	
  and	
  maintenance.	
  	
  Unfortunately	
  
this	
  has	
  not	
  happened.	
  
It	
  is	
  important	
  to	
  mention	
  that	
  there	
  were	
  challenges	
  in	
  the	
  relationship	
  between	
  JAM	
  and	
  FH.	
  	
  FH	
  
staff	
   offer	
   two	
   explanations	
   for	
   these	
   challenges:	
   (1) JAM’s	
   faith-­‐based	
   identity and	
   perceived	
  
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   26 The	
  structured	
  observations	
  included	
  a	
  small	
  section	
  with	
  questions,	
  asked	
  to	
  passersby/community	
  members/well	
  users
2	
  
8	
  
Figure	
  20:	
  Availability	
  of	
  handwashing	
  
facili]es	
  in	
  school	
  latrines	
  
Handwashing	
  
available	
  
Handwashing	
  
not	
  available	
  
Final	
  Evaluason	
  ,	
  N=	
  10	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 37
compatibility	
  of	
  vision	
  with	
  FH at	
   the	
   time	
  of	
  MYAP	
  proposal	
  development27 may	
  have	
  outweighed	
  
technical	
  considerations,	
  (2)	
  the	
  Northern	
  Mozambican	
  context	
  proved	
  to	
  be	
  difficult	
  for	
  JAM from	
  a
technical	
   perspective.	
   	
   Activities	
   were	
   not	
   properly	
   budgeted for	
   the	
   unique	
   drilling	
   and	
   technical	
  
requirements	
   of	
   coastal	
   communities, and	
   they	
   did	
   not	
   plan	
   accordingly.	
   	
   JAM left	
   Mozambique
without	
  completing	
  all	
  the	
  activities	
  under	
  their	
  contract.
As mentioned	
  under	
  the	
  health	
  section	
  of	
  this	
  report,	
  HH	
  have	
  shown	
  a	
  high	
  degree	
  of	
  acceptance	
  of	
  
HH	
  latrines.	
  	
  Based	
  on	
  53	
  observations,	
  100%	
  of	
   the latrines	
  show	
  regular	
  use.	
   Seventy	
  percent	
  of	
  
the	
   latrines	
   are	
   kept	
   clean (free	
   of	
   grass,	
   other	
   natural	
   growth	
   or	
   human	
   debris).	
   As	
   mentioned	
  
above,	
  in	
  60%	
  of	
  cases	
  the	
  latrine	
  opening	
  does	
  not	
  have	
  a	
  cover,	
  creating	
  a	
  potential	
  disease	
  vector.	
  	
  
One of	
  the	
  primary	
  concerns	
  with	
  the	
  latrines	
  is	
  related	
  to	
  their	
  design.	
  	
  The	
  pit	
  is	
  unlined, and	
  the	
  
remaining	
  structural	
  elements	
  are	
  made	
  of	
  mud	
  and	
  sticks.	
  	
  Most	
  families	
  cannot	
  afford	
  even	
  simple	
  
roofing	
  materials.	
  	
  As	
  a	
  result	
  80%	
  of	
  observed	
  latrines	
  did not	
  have roofs, making	
  them	
  precarious,	
  
especially during	
  the	
  rainy	
  season.	
  	
  Several	
  HH	
  mentioned	
  that	
  their	
  latrines	
  had	
  already	
  collapsed	
  at	
  
least	
  once.	
  	
  	
  On the	
  one	
  hand	
  it	
  makes	
  sense	
  to	
  use	
  a simple	
  design	
  requiring	
  few	
  costly	
  inputs.	
  	
  On	
  
the	
  other	
  hand,	
  encouraging	
  the	
  use	
  of	
  unlined	
  pit	
  latrines	
  with	
  a	
  stick	
  and	
  mud	
  platform	
  in	
  a	
  place	
  
with	
  six	
  months	
  of	
   rain	
  presents	
  potential	
  health	
  and	
  safety	
   risks,	
  as	
  well	
  as	
  presenting	
  a	
   threat	
   to	
  
sustainability.	
  	
  
Mother	
   Leaders	
   in	
   FDGs	
   mentioned	
   that	
   although	
   the	
   participation	
   of	
   husbands/partners	
   had	
  
improved	
   since	
   the	
   program	
   began,	
   male	
   participation continues to	
   be	
   an	
   important	
   factor	
   in	
  
facilitating	
   or	
   blocking	
   HH-­‐level	
   behavior	
   change.	
   One	
   of	
   the	
   biggest	
   challenges was	
   in	
   getting	
  
husbands	
  to	
  support	
  latrine	
  construction	
  (latrine	
  use has	
  not	
  been	
  a	
  challenge.	
  	
  Once	
  the	
  latrines	
  are	
  
built,	
  husbands	
  use	
  them).	
  As	
  one	
  mother	
  described,	
  many	
  men	
  would	
  tell	
  their	
  wives,	
  “you’re	
  the	
  
one	
  participating	
  in	
  the	
  training;	
  you	
  build	
  the	
  latrine.”	
  	
  Although	
  FGDs	
  with	
  Mother	
  Leaders	
  suggest	
  
that	
   this	
   barrier	
   has	
   reduced,	
  the	
   latrine	
   design	
   itself	
   should	
   help	
   facilitate	
   the	
   transition	
   and	
   not	
  
work	
   as	
   a	
   barrier	
   to	
   change.	
   	
   Latrines	
   that	
   require	
   re-­‐building	
   (digging)	
   every	
   year	
  may	
   become	
   a	
  
barrier	
  to	
  behavior	
  change.	
  	
  The	
  harder	
  behavior	
  change	
  is	
  (e.g.,	
  the	
  more	
  work	
  it	
  requires)	
  the	
  less	
  
likely	
  it	
  is	
  to	
  take	
  hold and	
  be	
  sustained.	
  	
  
Given	
   the	
  challenges	
  with	
  school	
  latrines	
   (e.g.,	
   their	
  limited	
  use	
  and	
  maintenance-­‐ and	
   the	
  unclear
link	
   to	
   the	
   target	
  age	
  group	
  of	
   children	
  0-­‐24	
  months)	
  and	
   the	
  cost	
  of	
   building	
   them,	
  it	
  might	
   have	
  
made	
  more	
  sense	
  to	
  invest	
  those	
  resources	
  in	
  a	
  slightly	
  more	
  durable	
  HH	
  latrine	
  design.	
  
Intermediate	
  Result	
  3.3:	
  Increased	
  ability	
  to	
  predict	
  and	
  mitigate	
  shocks
The	
   purpose	
   of	
   this	
   component	
   was	
   to	
   provide	
   communities	
   with	
   basic	
   understanding	
   of	
   early	
  
warning	
   systems	
   (EWS)	
   and	
   mitigation	
   processes	
   to	
   increase	
   resiliency	
   against	
   shocks	
   such	
   as	
  
destruction	
  of	
  crops	
  by	
  animals,	
  drought	
  or	
  illness.	
  	
  The	
  indicators	
  and	
  targets	
  for	
  IR3.3	
  are	
  included	
  
below:	
  
Table	
  26:	
  IPTT	
  Indicators	
  and	
  targets	
  for	
  IR3.3
Number	
  of	
  communities	
  with	
  improved	
  infrastructure	
  to	
  mitigate	
  the	
  impact	
  of	
  shocks 44
Number	
  of	
  communities	
  with	
  disaster	
  early	
  warning	
  and	
  response	
  systems	
  in	
  place 46
Number	
  of	
  months	
  of	
  increase	
  of	
  adequate	
  food	
  provisioning .5
Number	
   of	
   assisted	
   communities	
   with	
   safety	
   nets	
  to	
   address	
   the	
   needs	
   of	
   their	
   most	
  
vulnerable	
  members 46
In	
   the	
   second	
   year	
   of	
   the	
   program	
   risk	
   management	
   committees	
   were	
   formed	
   and	
   trained	
   in	
   all	
  
program	
  intervention	
  communities.	
  	
  With	
  support	
  from	
  FH, 44	
  risk	
  management	
  committees	
  formed
	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
  	
   27 According	
   to FH	
   Program	
  Director,	
   JAM	
  has	
  undergone	
  a	
  change	
  in	
  organizational	
  identity	
  over	
   the	
  last	
   few	
  years and	
  
they	
  no	
  longer	
  publicly	
  identify	
  as	
  a	
  faith-­‐based	
  organization.
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 38
EWS,	
   including	
   the	
   use	
   of	
   radios,	
   flags	
   and	
   drums	
   to	
   share	
   information	
   within	
   the	
   community.	
  	
  
According	
   to	
  FH	
   staff,	
   the	
  early	
  warning systems	
  are	
  essentially	
  defunct.	
  	
  Most	
  of	
   the	
  intervention	
  
communities	
   do	
   not	
   experience	
   disaster	
   or	
   emergency	
   situations	
   very	
   commonly,	
   making	
   them	
  
impractical	
  and	
  unsustainable.
A total	
  of 37 communities	
  have	
  completed	
  infrastructure	
  to	
  mitigate	
  shocks (under	
  IR3.2).	
  	
  The	
  most	
  
relevant	
  of	
  these	
  are	
  wells	
  (five),	
  market	
  stalls	
  (14),	
  health	
  posts	
  (14),	
  bridge	
  and	
  road	
  repair	
  (three)	
  
and	
  the	
  construction	
  a	
  small	
  maternal	
  care	
  center.
A	
  total	
  of	
  196	
  VSLAs	
  across	
  50	
  communities	
  have	
  accumulated	
  a	
  total	
  of	
  $18,195	
  in	
  social	
  investment	
  
funds	
  to	
  act	
  as	
  safety	
  nets	
  in	
  case	
  of	
  emergencies	
  or	
  unexpected	
  occurrences	
  at	
  the	
  community	
  level.	
  	
  
Overview	
  and	
  Discussion	
  of	
  Community	
  Capacity	
  Building	
  findings
Table	
  27 Comparative	
  Summary	
  of	
  Baseline	
  and	
  Final	
  CCB	
  Indicator	
  Values	
  for	
  SO3
Indicator
BL FE
%	
  Point	
  /	
  
Numerical	
  
Difference
Sig.
Increased	
  community	
  leadership	
  capacity
Number	
  of	
  HCA/PRA	
  conducted	
  and	
  reports	
  prepared 0 48 + NA
Number	
  of	
  CDC	
  leaders	
  participating	
  in	
  training 0 570 + NA
Percent of	
  CDC	
  members	
  who	
  are	
  women 0 33% + NA
Increased	
  community	
  infrastructure	
  and	
  assets
% of	
  CDCs	
  that	
  complete	
  community	
  development	
  projects 0 87.5% + NA
Number	
  of	
  community	
  development	
  projects	
  completed 0 68 + NA
Amount	
   contributed	
   by	
   community	
   for	
   small-­‐scale	
  
community	
  projects	
  (US$) 0 + NA
Prediction	
  and	
  mitigation	
  of	
  shocks
Number	
   of	
   communities	
   with	
   improved	
   infrastructure	
   to	
  
mitigate	
  the	
  impact	
  of	
  shocks 0 37 + NA
Number	
   of	
   communities	
   with	
   disaster	
   early	
   warning	
   and	
  
response	
  systems	
  in	
  place 0 44 + NA
Number	
  of	
  months	
  of	
  increase	
  of	
  adequate	
  food	
  provisioning 9.87 11.16 +	
  1.29 Y
Number	
  of	
  assisted	
  communities	
  with	
  safety	
  nets	
  to	
  address	
  
the	
  needs	
  of	
  their	
  most	
  vulnerable	
  members 0 50 + NA
Given	
  that	
  the	
  baseline	
  for	
  the	
  majority	
  of	
  these	
  indicators	
  was	
  zero	
  (assuming	
  no	
  previous	
  support	
  
or	
   achievement	
   prior	
   to	
   the	
   program),	
   all	
   indicator	
   values	
   under	
   SO3	
   increased	
   over	
   the	
   baseline.	
  	
  
The	
  majority	
  of	
  the	
  indicator	
  values	
  are	
  just	
  under	
  the	
  cumulative	
  program	
  targets.
Almost	
   600	
   community	
  leaders	
  were	
   trained	
  and	
  HCA	
  was	
   completed	
  in	
  each	
   of	
   the	
   communities,	
  
but	
   community	
   leadership	
   capacity seems	
   to	
   be	
   mixed.	
   	
   FGD	
   conducted	
   with	
   two	
   CDCs	
   showed	
  
substantial	
  variation	
  in	
  understanding of	
  their	
  basic	
  functions.	
   This	
  variation	
  of	
  capacities	
  is	
  likely	
  to	
  
reflect	
   the	
   reality	
   of all	
   program	
   areas.	
   	
   Since	
   no	
   systematic	
   process	
   or	
   tool	
   has	
   been	
   used	
   to	
  
measure	
   CDC	
   capacity,	
   it	
   might	
   be	
   advisable	
   to	
   develop	
   and	
   administer	
   such	
   a	
   tool	
   prior	
   to	
   the	
  
completion	
   of	
   the	
   program	
   to	
   establish	
   a	
   more	
   accurate	
   understanding	
   of	
   CDC	
   skills,	
   deficits	
   and	
  
ongoing	
  capacity	
  needs.	
  	
   The	
  indicators	
  included	
  for	
  this	
  component	
  represent	
  a	
  flaw	
  in	
  the	
  design	
  
of	
   the	
   program	
   as,	
   on	
   their	
   own,	
   they	
   tell	
   very	
   little	
   about	
   CDC	
   capacity	
   to	
   identify,	
   prioritize,	
  
coordinate	
  and	
  manage	
  ongoing	
  development	
  efforts.
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 39
The	
  project	
  has	
  clearly	
  resulted	
  in	
  significant	
  gains	
  in	
  community	
  infrastructure	
  and	
  assets.	
  	
  The	
  most	
  
important	
   of	
   these	
  are the	
   contributions	
   to	
  water	
  and	
   sanitation	
  infrastructure.	
  	
  Given	
   the	
  limited	
  
time	
  of	
  the	
  final	
  evaluation	
  exercise	
  it	
  was	
  not	
  possible	
  to	
  visit	
  all	
  (or	
  even	
  a	
  representative sample)	
  
of	
   the	
   various	
   types	
   of	
   infrastructure	
   developed.	
   	
   However,	
   based	
   on	
   conversations	
   with	
   the	
   CDC	
  
members	
  and	
  observations	
  of	
  school	
  latrines,	
  wells	
  and	
  HH	
  latrines,	
  it	
  may	
  be	
  advisable	
  to	
  work	
  with	
  
community	
  leaders	
  and	
  local	
  government	
  to	
  develop	
  clear	
  sustainability	
  plans	
  for	
  this	
  infrastructure	
  
moving	
   forward.	
  	
  Given	
   the	
  hopes	
  of	
  communities-­‐ as	
  well	
  as	
   the	
   time	
  and	
  monetary	
  investments-­‐
that	
  this	
  infrastructure	
  represents,	
  the	
  last	
  months of	
  the	
  program	
  should	
  be	
  spent	
  trying	
  to	
  ensure	
  
that	
  communities can	
  use	
  and	
  maintain	
  their	
  new	
  assets	
  as	
  long	
  as	
  possible.
It	
   is	
   frankly	
   difficult	
   to	
   see	
   the	
   practical	
   value	
   added	
   of	
   the	
   component	
   regarding	
   prediction	
   and	
  
mitigation	
   of	
   shocks as	
   nothing	
   concrete	
   seems	
   to	
   have	
   happened	
   under	
   this	
  IR	
   since	
  early	
   in	
   the	
  
program.	
  	
  It	
  isn’t	
  clear	
  what	
  need	
  these	
  activities	
  address or	
  what	
  their practical	
  impacts	
  have	
  been.
The	
   theory	
   underpinning	
   SO3	
   makes	
   a	
   great	
   deal	
   of	
   sense:	
   empowering	
   communities	
   to	
   take	
  
ownership	
   over	
   their	
   development	
   processes	
   should	
   be	
   the	
   fundamental	
   task	
   of	
   al	
   development	
  
actors.	
  	
  The	
  initial	
  efforts	
   to	
  identify	
  and	
  strengthen	
  community	
  leaders	
  provide a	
  solid	
   foundation	
  
for	
  continued	
  work	
  in	
  the	
  zone.	
  	
  Hopefully	
  the	
  opportunity	
  for	
  such	
  work	
  presents	
  itself	
  so	
  that	
  these	
  
efforts	
  are	
  not	
  lost.
From	
   a	
   project	
   design	
   perspective	
   it	
   is	
   not	
   clear	
   why	
   a	
   separate	
   SO	
   was	
   created	
   for	
   community	
  
strengthening,	
  since	
   community	
   engagement	
   and	
   capacity	
   strengthening	
   underpin	
  success	
   in	
   both	
  
the	
   health	
   and	
   agricultural	
   objectives.	
   	
  A disproportionate	
   amount	
   of	
   the	
   CCB	
   team’s	
   energy	
  may	
  
have	
  been	
  spent	
  on	
  infrastructure	
  development,	
  perhaps	
  to	
  the	
  detriment	
  of	
  capacity	
  building.	
  	
  The	
  
focus	
  on	
  infrastructure	
  created	
  a	
  “pull”	
  away	
  from	
  most	
  of	
  the	
  other	
  capacity	
  development	
  activities	
  
included	
  under	
  the	
  SO,	
  including	
  training	
  and	
  mentoring	
  to	
  ensure	
  effective	
  use	
  and	
  maintenance	
  of	
  
the	
  infrastructure	
  itself.	
  It	
  is	
  worth	
  pondering	
  whether	
  is preferable	
   to	
  have	
  30	
  good	
  infrastructure	
  
projects	
   with	
   communities	
   capable	
   of	
   managing	
   them	
   sustainably	
   or to	
   have	
  many	
   more projects	
  
that	
  may	
  not	
  provide lasting	
  contributions.
VI.	
  Conclusions	
  and	
  Recommendations
A.	
  Program	
  Strengths
Farmers	
  in	
  FGD	
  clearly	
  confirm	
  that	
  productivity	
  has	
  increased since	
  they	
  began	
  participating	
  in	
  the	
  
MYAP.	
   	
   The consistent	
   productivity	
   growth	
   demonstrated	
   through the	
   Annual	
   Agriculture surveys	
  
supports	
   these	
   assertions.	
   For	
   the	
   final	
   Annual	
   Agricultural	
   Survey	
   in	
   2013, FH	
   should	
   devote	
  
resources	
  to	
  a	
  strong	
  final	
  methodological	
  and	
  analytical	
  exercise	
  to	
  validate	
  the	
  results	
  presented	
  in	
  
this	
  report.	
  	
  Increased	
  farmer	
  engagement	
  in	
  collective	
  markets has	
  provided	
  additional	
  income for	
  
communities.	
   Although	
   the	
   degree	
   of	
   sustainability	
   of	
   those	
   activities	
   is	
   unclear,	
   they	
   provide	
   a	
  
strong	
  foundation	
  of	
  learning	
  upon	
  which	
  farmers	
  can	
  continue	
  to	
  build.	
  	
  It	
  would	
  be	
  advisable	
  for	
  FH	
  
to	
  ensure	
   that	
  clear	
  data	
   regarding	
  incomes	
  and	
  distribution	
  of	
  benefits	
  are gathered,	
  collated	
  and	
  
stored	
   to	
  gain	
  a	
   fuller	
  understanding	
  of	
   the	
  magnitude	
  of	
  benefits	
  across	
  MYAP	
  intervention	
  areas	
  
and	
   to	
   facilitate	
   learning	
   at	
   an	
   agency	
   level.	
   	
   Engagement	
   with	
   VSLAs has	
   also	
   provided	
   an	
  
opportunity	
  for	
  at	
  least	
  3,550	
  men	
  and	
  women	
  to	
  engage	
  with	
  financial	
  services,	
  albeit	
  informal and
likely	
   for	
   the	
   first	
   time.	
   	
   They	
   have	
   saved	
   over	
   $200,000	
   collectively,	
   earned	
   $44,500	
   on	
   those	
  
savings	
   through	
   interest	
   income	
   and	
   provided capital	
   to	
   local	
   entrepreneurs	
   and	
   individuals.	
  
Together	
  the	
  increased	
  productivity	
  and	
  incomes	
  derived	
  from	
  the	
  program	
  are	
  likely	
  contributors to	
  
the	
  gains	
  we	
  have	
  seen	
  in	
  both	
  Household	
  Food	
  Provisioning	
  and	
  overall	
  Dietary	
  Diversity	
  Scores.
There	
  have	
  been	
  significant	
  improvements	
  in	
  many	
  of	
  the	
  essential	
  nutrition	
  actions including	
  early	
  
breastfeeding,	
   exclusive	
   breastfeeding	
   and	
   adequacy	
   of	
   complementary	
   feeding.	
   	
   	
   Although	
   the	
  
knowledge	
   indicators	
   for	
   the	
   recognition	
   of	
   signs	
   of	
   illness	
   and	
   methods	
   to	
   prevent	
   HIV	
   are	
  
surprisingly	
  low,	
  some	
  key	
  prevention	
  indicators	
  show	
  improvements,	
  including	
  significant	
  decreases	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 40
in	
  cases	
  of	
  diarrhea and	
  increases	
  in	
  the	
  use	
  of	
  treated	
  bednets to	
  combat	
  malaria.	
  	
  There	
  have	
  also	
  
been	
  increases	
  in	
  hospital	
  visits	
  within	
  the	
  first	
  day	
  of	
  fever.	
  	
   Access	
  to	
  improved	
  water	
  sources has	
  
almost	
   doubled	
   thanks	
   to	
   the	
   construction	
   of	
   water	
   infrastructure.	
   	
   Likewise,	
   knowledge	
   of	
   hand	
  
washing	
   behaviors,	
   proper	
   excreta	
   disposal and	
   food	
   and	
   water	
   hygiene have	
   all	
   improved	
  
significantly	
  over	
  the	
  baseline.	
  	
  Together	
  these	
  factors	
  are	
  likely	
  to	
  have	
  improved	
  food	
  utilization	
  by	
  
improving	
   overall	
   nutrient	
  intake	
  and	
   use.	
  	
  The	
  anthropometric	
   survey	
  in	
  October-­‐November	
   2013
will	
  confirm	
   the	
  impact	
  of	
  improved	
  knowledge	
  and	
  practices	
  on	
   the	
  nutritional	
   status	
  of	
  children,	
  
although	
  some	
  gains-­‐ even	
  if	
  modest-­‐ should	
  be	
  expected.	
  
The	
  program	
  has	
  achieved	
  a	
  high	
  level	
  of	
  community	
  engagement	
  and	
  ownership through:	
  the	
  use	
  
of	
   local	
   promoters	
   for	
   health	
   and	
   agriculture,	
   training	
   and	
   support	
   to	
   create	
  Mother	
   Leaders	
   for	
  
Care	
  Groups,	
  development	
  and	
   training	
  of	
  Farmer	
   Field	
   Life	
  Groups and	
  Community	
  Development	
  
Committees.	
  	
  	
  The	
  leadership	
  and	
  technical	
  skills	
  that	
  these	
  individuals	
  and	
  groups	
  have	
  gained	
  will	
  
provide	
  a	
  strong	
  foundation	
  for	
  continued	
  health	
  and	
  wellbeing	
  in	
  their	
  families	
  and	
  for community	
  
development	
  across	
  all	
  MYAP	
  intervention	
  areas.	
  	
  	
  According	
   to	
  staff	
  at	
   the	
  Ministry	
  of	
  Health,	
   the	
  
involvement	
  of	
  community	
  leaders-­‐ especially	
  the	
  use	
  of	
  community-­‐based	
  promoters-­‐ has	
  been	
  the	
  
cornerstone	
  of	
  the	
  program’s	
  success.	
  	
  
B.	
  Program	
  Challenges
Specific	
   challenges	
   within	
   each	
   technical	
   sector	
   have	
   already	
   been	
   addressed	
   in	
   each	
   respective	
  
section.	
  	
  This	
  section	
  will	
  briefly	
  highlight	
  identified	
  challenges	
  that	
  affect	
  the	
  whole	
  program.
The	
  geographic	
  spread of	
  the	
  program	
  may	
  have	
  been	
  an	
  impediment	
  to	
  timely	
  and	
  effective	
  follow-­‐
up.	
  	
  This	
  issue	
  was	
  raised	
  anecdotally	
  during	
  three	
  FGD,	
  but	
  given	
  the	
  terrain	
  and	
  distance	
  between	
  
locations,	
  program	
  technical	
  staff	
  are	
  unable	
  to	
  devote intensive time	
  in	
  each	
  location,	
  which	
  means	
  
they	
   rely	
  heavily on	
  local	
  promoters.	
  	
  While	
   the	
  FH	
  model	
  of	
  using	
  community-­‐based	
  promoters	
  is	
  
one	
   of	
   the	
   project’s	
   greatest	
   strengths,	
   reliance	
   on	
   local	
   promoters	
   may	
   translate	
   into	
   gaps	
   in	
  
technical	
  expertise	
  and	
  follow-­‐up.	
  	
  
Data	
   management is	
   a	
   challenge	
   across	
   the	
   program.	
   	
   Health	
   has	
   the	
   fewest	
   challenges,	
   likely	
  
because	
  the	
  sector	
  has	
  clear	
  data	
  standards	
  and	
  protocols.	
  	
  In	
  the	
  case	
  of	
  agriculture	
  and	
  CCB	
  there	
  
are	
  gaps.	
  	
  Some	
  of	
  these	
  are	
  owed	
  to	
  design	
   flaws:	
  	
   (1)	
   the	
  CCB	
  indicators	
   tell	
  very	
  little	
  about	
   the	
  
impact	
  of	
  CCB	
  activities;	
  (2)	
  the	
  agricultural	
  component	
  did	
  not	
  include	
  any	
  marketing indicators.	
  	
  But	
  
there	
  are	
  also	
  data	
  management	
  challenges:	
  data	
  that	
  should	
  exist	
  readily	
  is	
  difficult	
  to	
  access.	
  	
  On	
  
multiple	
  occasions	
  during	
  the	
  final	
  evaluation	
  basic	
  information-­‐ such	
  as	
  the	
  number	
  of	
  participants	
  
in	
   marketing	
   forums-­‐ was	
   requested	
   but	
   was	
   never	
   received.	
   	
   	
   Some	
   of	
   the	
   challenge	
   is	
   owed	
   to	
  
where	
  data	
  is	
  managed	
  and	
  who	
  “owns” it.	
  	
  Logically, technical	
  and	
  management	
  staff	
  delegate	
  tasks	
  
to	
  other	
  members	
  of	
  their	
  teams,	
  but	
  some	
  managers	
  do	
  not	
  seem	
  to	
  be	
  requesting	
   (or	
  reviewing)	
  
the	
  data	
  that	
  is	
  available,	
  which	
  does	
  raise	
  questions	
  regarding	
  how	
  program	
  management	
  decisions	
  
are	
  made.	
   	
   Likewise,	
  according	
   to	
   the	
  M&E	
  Manager,	
   the	
   guidance	
   for	
   calculating	
   indicator	
   values	
  
was	
  developed	
  only	
  a	
  year	
  ago and	
  some	
  of	
  the	
  indicators	
  within	
  the	
  current	
  PMP	
  still	
  do	
  not	
  have	
  
clear	
  definitions.	
  
Sustainability	
  will	
  be	
  a	
  challenge	
  across	
  all	
  components	
  of	
  the	
  program.	
  	
  The	
  agriculture	
  program	
  has	
  
begun	
   to	
   plan	
   for	
   the	
   post-­‐program	
   phase	
   by	
   meeting	
   with	
   the	
   MoA	
   and	
   supporting	
   the	
   official	
  
“recognition”	
   of	
   the	
   program’s	
   community-­‐based	
   promoters	
   as	
   official	
   MoA	
   volunteers.	
   	
   No	
  such	
  
action	
  planning	
  has	
  occurred	
  yet	
  for	
  the	
  health	
  or	
  CCB	
  teams.	
   The	
  MOH,	
  for	
  example, currently	
  has	
  
only	
   25	
   trained	
   health	
   promoters	
   across	
   the	
   50	
   project	
   districts.	
   Across	
   all	
   sectors	
   the	
   resource	
  
challenges	
  of	
  district	
  governments	
  and	
  Ministries	
  will	
  create	
  serious	
  challenges	
   for	
  effective	
   follow-­‐
up.	
   	
   Even	
   in	
   the	
   case	
   of	
   the	
   MoA	
   the	
   recognition	
   of	
   community	
   volunteers	
   means	
   very	
   little	
   in	
  
practical	
  terms	
  because	
  they	
  do	
  not	
  have	
  the	
  budget	
  to	
  provide	
  incentives or	
  transportation.	
  	
  	
  	
  
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 41
There	
   are	
   nine communities	
   with	
   no	
   overlap	
   between	
   health	
   and	
   agriculture	
   activities,	
   including
Mocimboa	
   town,	
   which represents	
   more	
   than	
   40%	
   of	
   the	
   total	
   number	
   of	
   program	
   health	
   and
nutrition	
  participants.	
  	
  Those	
  beneficiaries	
  have	
  not	
  benefited	
  from	
  the	
  agriculture	
  activities,	
  which-­‐
based	
  on the	
  program’s	
  logic	
  model-­‐ means	
  they	
  are	
  not	
  benefitting	
  from	
  increased	
  food	
  access.	
   In	
  
large	
  part	
  this	
  was	
  based	
  on	
  the	
  need	
  to	
  meet	
  the	
  target number	
  of program	
  participants	
  included	
  in	
  
thr approved	
  program	
  proposal;	
   the	
  health	
  program	
  expanded	
   to	
   the	
  city	
  center	
  in order	
   to	
   reach	
  
the	
  numbers.	
  	
  Nonetheless,	
  this	
  presents challenges	
  to	
  the	
  underlying	
  theory-­‐ and	
  presumably	
  to	
  the	
  
results-­‐ of	
  the	
  program.
C.	
  Recommendations
Agriculture
• Given	
  the	
  challenges	
  with	
  neem,	
  there	
  is	
  a	
  need	
  to	
  identify	
  other	
  viable	
  organic	
  crop,	
  seed	
  and	
  
storage	
  protection	
  mechanisms	
  that	
  do	
  not	
  require	
  such	
  a	
  long	
  period	
  for	
  maturation.
• Likewise,	
  there	
  is	
  a	
  need	
  to	
  mentor	
  staff	
  to	
  ensure	
  that	
  when	
  confronted	
  with	
  difficult	
  questions	
  
from	
  program	
  participants,	
  field	
  staff	
  can	
  respond	
  in	
  constructive	
  ways	
  to	
  help	
  solve	
  problems.
• It	
   might	
   be	
   worth	
   conducting	
   post-­‐facto	
   data	
   cleaning	
   and	
   analysis	
   for	
   the	
   previous	
   Annual	
  
Agricultural	
  surveys	
  and	
  ensure	
  that	
  the	
  final	
  survey	
  follows	
  sampling,	
  data	
  collection and	
  data	
  
analysis	
  protocols	
  precisely	
  in	
  order	
  to	
  provide	
  additional	
  clarity	
  regarding	
  productivity.	
  
• Data	
   regarding	
   collective	
   individual	
   and	
   collective	
   marketing	
   benefits	
   should	
   be	
   collected,	
  
compiled	
  and	
  analyzed	
  to	
  gain	
  a	
  greater	
  understanding	
  of	
  the magnitude	
  of	
  income	
  generation.
• Select	
   farmers	
   from	
   collective	
   marketing	
   should	
   be	
   linked	
   to	
   ongoing	
   training	
   opportunities	
  
regarding	
   key	
   marketing	
   skills	
   such	
   as	
   market	
   assessment,	
   value	
   addition	
   and	
   processing	
   in	
  
order	
   to	
   build	
   on	
   the	
   skills	
   they	
   have	
   gained	
   and	
   help	
   support	
   future	
   sustainability.	
   	
   Future	
  
program	
  should	
  include	
  an	
  intensive	
  training	
  component	
  for	
  market-­‐based	
  production	
  working	
  
with	
  specialized	
  experts	
  in	
  the	
  subject	
  matter.
• It	
   might	
   be	
   worth	
   exploring	
   the	
   possibility	
   of	
   a	
   long(er)-­‐term	
   contract	
   between	
   the	
   current	
  
buyer	
  and	
  the	
  marketing	
  forums	
  prior	
  to	
  the	
  close	
  of	
  the	
  program.	
  	
  This	
  may	
  be	
  one	
  way	
  to	
  help	
  
ensure	
  ongoing	
  benefits	
  from	
  those	
  activities.
• Before	
  the	
  program	
  ends	
  it	
  might	
  be	
  useful	
  to	
  analyze	
  mean	
  differences	
  in	
  available	
  savings	
  for	
  
groups	
   operating	
   based	
   on	
   flexible	
   standards	
   versus	
   those	
   that	
   have	
   implemented	
   the	
  model	
  
based	
   on	
   fixed	
   contribution	
   amounts	
   and	
   timelines.	
   	
   Focus	
   group	
   or	
   other	
   qualitative	
   work	
  
might	
   also	
   help	
   determine	
   whether	
   the	
   current	
   flexible	
   arrangement	
   provides	
   the	
   best	
  
opportunity	
   for	
   consumption	
   smoothing	
   and	
   reducing	
   risk	
   among	
   participants. This	
   could	
  
provide	
  useful	
  lessons	
  for	
  future	
  VSLA	
  work	
  moving	
  forward.	
  	
  	
  
• Future	
   programs	
   of	
   this	
   nature	
   might	
   consider	
   the	
   inclusion	
   of	
   a	
   small	
   quasi-­‐experimental	
  
design	
   comparing	
   cash	
   crop	
   participant	
   wellbeing	
   with	
   that	
   of	
   participants	
   who	
   do	
   not	
  
participate	
   in	
   cash	
   crops.	
   	
   This	
   would	
   provide	
   evidence of	
   impact	
   of	
   cash	
   crop	
   production	
   on	
  
food	
  access,	
  dietary	
  diversity	
  and	
  household	
  food	
  provisioning	
  
• Agricultural	
  indicators	
  and	
  measurement	
  tools	
  in	
  future	
  programs	
  should	
  be	
  clearly	
  linked	
  to	
  the	
  
expected	
   outcomes	
   of	
   the	
   program	
   to	
  ensure	
   that	
  measurement	
   provides	
   useful	
   insights	
   into	
  
program	
  progress	
  and	
  impacts
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 42
Health
• Future	
   programs	
  requiring	
   the	
   implementation	
   of	
   a	
  modified	
   version	
   of Care	
  Groups	
   focusing	
  
only on pregnant	
  women	
  and	
  children 0-­‐24 months	
  should	
  use	
   training	
  modules	
   that	
  allow	
  all	
  
topics	
   to	
   be	
   covered	
   either	
   during	
   a	
   two	
   year	
   period or	
   during	
   the	
   length	
   of	
   a	
   mother’s	
  
enrollment.	
  	
  This	
  may	
  require	
  shortening	
  some	
  of	
  the	
  modules	
  based	
  on	
  the	
  identification	
  of	
  the	
  
“minimum	
  required	
  package”	
  of	
  messages	
  per	
  topic	
  for	
  greatest	
  impact.	
  	
  Then	
  refresher	
  training	
  
can	
   be	
   provided	
   periodically	
   to	
   strengthen	
   certain	
   topics	
   or	
   provide	
   greater	
   detail	
   on	
   others.	
  	
  	
  
Clearly	
   a	
   “condensed”	
   series	
   of	
   modules	
   for	
   a	
   two-­‐year	
   time	
   period	
   should	
   be	
   tested	
   for	
  
effectiveness	
  prior	
  to	
  use.
• Future	
  programs	
  using	
  the	
  Care	
  Group	
  model	
  should	
  identify	
  incentive	
  plans	
  for	
  Mother	
  Leaders	
  
that	
  not	
  create	
  a	
  disincentive	
  for	
  participants.
Community	
  Capacity	
  Building
• Prior	
   to	
   the	
   end	
   of	
   the	
   program,	
   a	
   simple,	
   standard	
   tool	
   should	
   be	
   adapted	
   /	
   developed	
   and	
  
used	
   to	
  measure	
   the	
   capacity	
   of	
   CDCs.	
   	
  This	
  will	
   provide	
  a	
   clear	
   sense	
   of	
   the	
   skills	
   developed	
  
during	
  the	
  program	
  and	
  will	
  facilitate	
  follow-­‐on	
  skills	
  building	
  for	
  those	
  groups	
  that	
  show	
  large	
  
skill	
  deficits.	
  
• Sustainability	
  plans	
   should	
  be	
  developed	
   for	
  all	
  infrastructure	
  built	
  under	
   the	
  program.	
  	
  CDCs,	
  
district	
  and	
  local	
  governments	
  and	
  relevant	
  Ministries	
  should	
  be	
  included	
  in	
  this	
  activity.	
  
• Refresher	
  training	
  should	
  be	
  provided	
  for	
  HH	
  on	
  the	
  importance	
  of	
  covering	
  latrine	
  openings	
  to	
  
limit	
  disease.	
  	
  Local	
  alternatives	
  for	
  covers	
  should	
  be	
  developed	
  and	
  disseminated	
  to	
  ensure	
  that	
  
this	
  does	
  not	
  incur	
  additional	
  costs	
  for	
  families.
• Likewise,	
  local	
   roofing	
  alternatives	
   should	
  be	
  disseminated	
   to	
   families	
   to	
  protect	
  latrines	
   from	
  
the	
  elements	
  and	
  maintain	
  them	
  intact	
  for	
  longer	
  periods	
  of	
  time
	
  	
  
All	
  programs
• For	
  future	
  programs,	
  key	
  program	
  data	
  should	
  be	
  gathered/stored/housed	
  in	
  a	
  central	
  location	
  
to	
  ensure	
  appropriate	
  decision-­‐making,	
  learning	
  and	
  institutional	
  memory	
  
• Sustainability	
   planning	
   needs	
   to	
   be	
   incorporated	
   into	
   program	
   design	
   in	
   future	
   programs	
   and	
  
might	
  include	
  staggered	
  transitioning	
  of	
  program	
  staff	
  into	
  relevant	
  Ministries.	
  	
  This	
  clearly	
  will	
  
require	
   strong	
   working	
   relationships	
   with	
   relevant	
   government	
   offices	
   and	
   long-­‐term	
  
commitment	
  (and	
  resources)	
  on	
  the	
  part	
  of	
  FH.
• Future	
  programs	
  should	
  try	
  to	
  ensure	
  overlap	
  between	
  program	
  sectors	
  to	
  ensure	
  that	
  program	
  
participants	
  can	
  take	
  advantage	
  of	
  all	
  the	
  complementary	
  benefits	
  from	
  each	
  sector.
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 43
VII.	
  Resources	
  Consulted
“Essential	
   Nutrition	
   Actions-­‐ Improving	
   Maternal-­‐Newborn-­‐Infant	
   and	
   Young	
   Child	
   Health	
   and	
  
Nutrition,	
  Evidence	
  for	
  Essential	
  Nutrition	
  Actions,”	
  World	
  Health	
  Organization,	
  2011.
“Indicators	
   for	
   assessing	
   infant	
   and	
   young	
   child	
   feeding	
   practices:	
   conclusions	
   of	
   a	
   consensus	
  
meeting	
  held	
  6–8	
  November	
  2007	
  in	
  Washington	
  D.C.,	
  USA. WHO
The	
   Handwashing	
   Handbook:	
   A	
   guide	
   for	
   developing	
   a	
   hygiene	
   promotion	
   program	
   to	
   increase	
  
handwashing	
  with	
  soap,	
  World	
  Bank
Food	
  Security	
  Information	
  for	
  Action,	
  “Practical	
  Guidelines:	
  An	
  Introduction	
  to	
  the	
  Basic	
  Concepts	
  of	
  
Food	
  Security,” United	
  Nations	
  Food	
  and	
  Agriculture	
  Organization	
  
2003	
   Mozambique	
   Demographic	
   and	
   Health	
   Survey (DHS),	
   Measure	
   DHS/ICF	
   International	
   and	
  
USAID.
2011	
   Mozambique	
   Demographic	
   and	
   Health	
   Survey (DHS),	
   Measure	
   DHS/ICF	
   International	
   and	
  
USAID,	
  March	
  2013
Swindale,	
   Anne	
   and	
   Bilinsky,	
   Paula,	
  Household	
   Dietary	
   Diversity	
   Score	
   (HDDS)	
   for	
  Measurement	
   of	
  
Household	
   Food	
   Access:	
   Indicator	
   Guide VERSION	
   2,	
   Food	
   and	
   Agriculture	
   Technical	
   Assistance	
  
(FANTA),	
  September	
  2006.	
  	
  
Swindale,	
  Anne	
  and	
  Bilinsky,	
   Paula,	
  Months	
   of	
  Adequate	
  Household	
  Food	
   Provisioning	
   (MAHFP)	
   for	
  
Measurement	
  of	
  Household	
  Food	
  Access:	
  Indicator	
  Guide	
  VERSION	
  4,	
  Food	
  and	
  Agriculture	
  Technical	
  
Assistance	
  (FANTA),June	
  2010
Indicators	
   For	
   Assessing	
   Infant	
   And	
   Young	
   Child	
   Feeding	
   Practices,	
   Part	
   2:	
   Measurement,	
   World	
  
Health	
  Organization
Espeut,	
   Donna,	
   Knowledge,	
   Practices	
   and	
   Coverage	
   Survey	
   2000+	
   Field	
   Guide,	
   The	
   Child	
   Survival	
  
Technical	
  Support	
  Project	
  (USAID),	
  August	
  2001.
Diskin,	
   Patrick,	
   Agricultural	
   Productivity	
   Indicators	
   Measurement	
   Guide,	
   Food	
   and	
   Agriculture	
  
Technical	
  Assistance	
  (FANTA),	
  December	
  1997
Tracking	
  Progress	
  on	
  Child	
  and	
  Maternal	
  Nutrition,	
  UNICEF,	
  2009.
JP	
  Habicht,	
  CG	
  Victora	
  and	
  JP	
  Vaughn,	
  “Evaluation	
  Designs	
  for	
  Adequacy,	
  Plausibility	
  and	
  Probability	
  
of	
  Public	
  Health	
  Programme	
  Performance	
  and	
  Impact,”	
  International	
  Journal	
  of	
  Epidemiology,	
  1999;	
  
28:	
  10-­‐18
Whitson,	
  Donald	
  T.	
  	
   Food	
   for	
   the	
  Hungry	
  Multi	
  Year	
  Assistance	
   Program	
   Baseline	
   Report,	
  February	
  
2009.
Food	
  for	
  the	
  Hungry	
  MYAP	
  Organizational	
  Structure	
  and	
  Staff	
  Chart,	
  August	
  2012
Food	
  for	
  the	
  Hungry	
  Multi Year	
  Assistance	
  Program	
  Proposal	
  Revised	
  Narrative,	
  August	
  2008.
Food	
  for	
  the	
  Hungry	
  Annual	
  Agricultural	
  Surveys: August	
  2010,	
  September	
  2012
Food	
  for	
  the	
  Hungry	
  Mozambique	
  Annual	
  Results	
  Reports:	
  FY09,	
  FY10,	
  FY11,	
  FY12
Food	
  for	
  the	
  Hungry	
  MYAP	
  Final	
  Evaluation	
  Report
Final	
  Version	
  July, 2013 44
VIII.	
  Annexes