Baseline Study of the FIOVANA Resilience 
Food Security Activity in Madagascar
October 2021|Volume I
IMPEL | Implementer-Led Evaluation & Learning Associate Award
ABOUT IMPEL
The Implementer-Led Evaluation & Learning (IMPEL) Associate Award works to improve the design and implementation of 
Bureau for Humanitarian Assistance (BHA)-funded resilience food security activities (RFSAs) through implementer-led 
evaluations and knowledge sharing. Funded by the United States Agency for International Development (USAID) BHA, IMPEL 
will gather information and knowledge in order to measure performance of RFSAs, strengthen accountability, and improve 
guidance and policy. This information will help the food security community of practice and USAID to design projects and 
modify existing projects in ways that bolster performance, efficiency, and effectiveness. IMPEL is a seven-year activity (2019-
2026) implemented by Save the Children (lead), TANGO International, Tulane University, Causal Design, and Innovations for 
Poverty Action.
RECOMMENDED CITATION
IMPEL. (2021). Baseline Study of the FIOVANA Resilience Food Security Activity in Madagascar (Vol. 1). Washington, DC: The 
Implementer-Led Evaluation & Learning Associate Award
PHOTO CREDITS
Save the Children / Charlie Forgham-Bailey.
DISCLAIMER
This report is made possible by the generous support of the American people through the United States Agency for 
International Development (USAID). The contents are the responsibility of the Implementer-Led Evaluation & Learning (IMPEL) 
award and do not necessarily reflect the views of USAID or the United States Government.
CONTACT INFORMATION
IDEAL Activity
c/o Save the Children
899 North Capitol Street NE, Suite #900
Washington, DC 20002
www.fsnnetwork.org 
IMPEL@savechildren.org
PREPARED BY: 
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Acknowledgments i
ACKNOWLEDGMENTS
The Causal Design research team would like to recognize several individual’s instrumental in creating 
and refining this document. This includes Emanuel da Costa, Tsarafidy Rasendraharison, Joseph 
Menakuntuala, Fahim Safi, and Jacques Zily with ADRA, and Mohamoud Ahmed, Emily Marshall, and 
Tom Spangler with IMPEL. Special thanks also to Mara Mordini, Justin Mupeyiwa, Benita O’Colmain, Arif 
Rashid, and Adam Trowbridge with USAID for their ongoing technical support.
IMPEL | Implementer-Led Evaluation and Learning
ii Table of Contents
TABLE OF CONTENTS
Acknowledgments ............................................................................................................................. i
List of Tables.....................................................................................................................................iii
List of Figures.....................................................................................................................................v
Acronyms .........................................................................................................................................vi
Executive Summary..........................................................................................................................vii
1. Introduction ................................................................................................................................1
2. Methodology and Limitations.......................................................................................................3
2.1 Evaluation Design...........................................................................................................................3
2.1.1 Randomization Strategy................................................................................................3
2.1.2 Sampling Frame.............................................................................................................4
2.1.3 Questionnaire Development.........................................................................................5
2.1.4 Field Preparation...........................................................................................................6
2.1.5 Data Collection..............................................................................................................6
2.2 Challenges and Limitations ............................................................................................................7
2.2.1 Survey Challenges.........................................................................................................7
2.2.2 Limitations.....................................................................................................................7
3. Findings.......................................................................................................................................8
3.1 Characteristics of the Study Population.........................................................................................8
3.2 Food Security................................................................................................................................10
3.3 Child Nutrition and Health ...........................................................................................................13
3.3.1 Nutrition......................................................................................................................13
3.3.2 Anthropometry ...........................................................................................................16
3.4 Women’s Health, Maternal Nutrition, and Reproductive Health................................................18
3.5 Water, Sanitation, and Hygiene Practices....................................................................................22
3.6 Agriculture....................................................................................................................................24
3.7 Poverty Measurement .................................................................................................................30
3.8 Gender Dynamics.........................................................................................................................32
3.8.1 Use of Financial Resources..........................................................................................32
3.8.2 Credit...........................................................................................................................34
3.9 Resilience......................................................................................................................................35
3.9.1 Ability to Recover from Shocks and Stresses Index ....................................................35
3.9.2 Social Capital Indices...................................................................................................37
3.9.3 Absorptive Capacity Index...........................................................................................38
3.9.4 Adaptive Capacity Index..............................................................................................39
3.9.5 Transformative Capacity Index ...................................................................................41
4. Comparison of Treatment and Control Groups............................................................................44
5. Conclusion.................................................................................................................................45
Annex A: Balance Tables ..................................................................................................................46
Annex B: Additional Tables...............................................................................................................53
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
List of Tables and Figures iii
VOLUME II
Annex C: Pre-Analysis Plan
Annex D: Data Collection Tool
LIST OF TABLES
Table 1. Number of fokontany and households surveyed............................................................................5
Table 2. Individual response rate..................................................................................................................5
Table 3. Household response rate ................................................................................................................6
Table 4. Basic household-level statistics.......................................................................................................8
Table 5. Baseline individual-level demographic information .......................................................................8
Table 6. Head of household statistics.........................................................................................................10
Table 7. Prevalence of moderate and severe food insecurity in the population .......................................10
Table 8. Percentage of households responding “yes” to “During the past 30 days, was there a time when 
you or others in your household…” disaggregated by household type*....................................................11
Table 9. Mean scores of households with poor, borderline, and acceptable Food Consumption Score...12
Table 10. Percentage of households with poor, borderline, and acceptable Food Consumption Score ...12
Table 11. Prevalence of children 6–23 months consuming a diet of minimum diversity...........................13
Table 12. Percentage of children ages 6–23 months receiving a minimum acceptable diet.....................14
Table 13. Minimum meal frequency, breast-fed and non-breastfed children 6–23 months.....................14
Table 14. Prevalence of exclusive breastfeeding of children under 6 months of age................................15
Table 15. Percentage of children under 5 (0–59 months) years old who had diarrhea in the prior 2 weeks
....................................................................................................................................................................15
Table 16. Percentage of children under 5 (0–59 months) years old with diarrhea treated with Oral 
Rehydration Therapy ..................................................................................................................................15
Table 17. Percentage of children under 5 years old who had fever, diarrhea, or acute respiratory 
infection in the past 15 days and received treatment within 24 hours from a health facility ...................16
Table 18. Prevalence of underweight children under 5 years old..............................................................17
Table 19. Prevalence of underweight children under 5 years old, disaggregated .....................................18
Table 20. Percentage of women of reproductive age consuming a diet of minimum diversity.................19
Table 21. Percentage of births receiving at least four antenatal care visits during pregnancy..................20
Table 22. Contraceptive prevalence rate....................................................................................................20
Table 23. Percentage of women in a union who have knowledge of modern family planning methods 
that can be used to delay or avoid pregnancy............................................................................................21
Table 24. Percent of women in a union who made decisions alone or with their partner jointly about 
modern family planning methods...............................................................................................................21
Table 25. Percent of non-pregnant adolescent girls and women of reproductive age practicing a modern 
method of birth control ..............................................................................................................................21
Table 26. Percentage of households using basic drinking water services..................................................22
Table 27. Water use per capita...................................................................................................................23
Table 28. Percentage of households with soap and water at a handwashing station on the premises....23
Table 29. Percentage of households in target areas practicing correct use of recommended household 
water treatment technologies....................................................................................................................23
IMPEL | Implementer-Led Evaluation and Learning
iv List of Tables and Figures
Table 30. Percentage of households in target areas practicing open defecation ......................................24
Table 31. Percentage of households with access to basic sanitation services...........................................24
Table 32. Crops grown ................................................................................................................................25
Table 33. Percentage of farmers who used financial services (savings, agricultural credit, and/or 
agricultural insurance) in the past 12 months & percentage of farmers who practiced the value chain 
interventions promoted by the activity in the past 12 months..................................................................25
Table 34. Improved management practices/technologies for cassava ......................................................26
Table 35. Improved management practices/technologies for rice ............................................................26
Table 36. Improved management practices/technologies for cloves ........................................................26
Table 37. Yield of targeted agricultural commodities within target areas.................................................28
Table 38. Poultry farming............................................................................................................................28
Table 39. Beekeeping..................................................................................................................................29
Table 40. Fishing .........................................................................................................................................30
Table 41. Poverty measures........................................................................................................................31
Table 42. Disaggregated poverty measures by household types...............................................................32
Table 43. Percentage of women and men in a union who earned cash in the past 12 months ................32
Table 44. Percentage of women in a union and earning cash who report participation in decisions about 
the use of self-earned cash.........................................................................................................................33
Table 45. Percentage of men in union and earning cash who report spouse/partner participation in 
decisions about the use of self-earned cash...............................................................................................33
Table 46. Percentage of women/men in a union with access to credit .....................................................34
Table 47. Percentage of women/men in a union who make decisions alone or jointly about credit........34
Table 48. Ability to recover from shocks and stresses index......................................................................35
Table 49. Social capital indices....................................................................................................................37
Table 50. Absorptive capacity index ...........................................................................................................38
Table 51. Adaptive capacity index ..............................................................................................................40
Table 52. Transformative capacity index ....................................................................................................42
Table 53. Household roster balance table..................................................................................................44
Table 54. Food security...............................................................................................................................46
Table 55. Child nutrition and health ...........................................................................................................46
Table 56. Anthropometry............................................................................................................................47
Table 57. Women’s health, maternal nutrition, and reproductive health .................................................48
Table 58. Water, sanitation, and hygiene...................................................................................................48
Table 59. Agriculture–cassava, rice, and cloves..........................................................................................49
Table 60. Agriculture–yield.........................................................................................................................49
Table 61. Poverty measurements...............................................................................................................50
Table 62. Use of financial resources...........................................................................................................50
Table 63. Credit...........................................................................................................................................50
Table 64. Resilience.....................................................................................................................................50
Table 65. Crops grown, gender...................................................................................................................53
Table 66. Agriculture value chain practices, gender...................................................................................53
Table 67. Modern practices for cassava, gender........................................................................................54
Table 68. Modern practices for rice, gender ..............................................................................................55
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
List of Tables and Figures v
Table 69. Modern practices for cloves, gender ..........................................................................................55
Table 70. Agricultural yield, gender............................................................................................................56
Table 71. Poultry farming, gender ..............................................................................................................56
Table 72. Beekeeping, gender.....................................................................................................................58
Table 73. Fishing, gender............................................................................................................................58
Table 74. Non-governmental organizations in the control areas...............................................................59
Table 75. Non-governmental organizations in the treatment areas..........................................................59
Table 76. Non-governmental organization activity ....................................................................................60
LIST OF FIGURES
Figure 1. FIOVANA regions and communes..................................................................................................1
Figure 2. Age and gender composition (unweighted) of the sample ...........................................................9
Figure 3. Distribution of weight-for-age z-score.........................................................................................17
Figure 4. Dietary diversity score for women of reproductive age ..............................................................19
Figure 5. Box plot of yield estimates for target crops.................................................................................27
Figure 6. Distribution of expenditure .........................................................................................................31
Figure 7. Most common shocks reported by households...........................................................................37
IMPEL | Implementer-Led Evaluation and Learning
vi Acronyms
ACRONYMS
ADRA Adventist Development Relief Agency
ANC Antenatal care
ARI Acute respiratory infection
BHA Bureau for Humanitarian Assistance
BL Baseline
CNA Child no adult
F&M Female and male adult
FAO Food and Agriculture Organization
FCS Food Consumption Score
FIES Food Insecurity Experience Scale
FNM Adult female no adult male
HAZ Height-for-Age Z-Score
ICC Intraclass Correlation Coefficient
IE Impact evaluation
IMPEL Implementer-Led Evaluation & Learning
MAD Minimum acceptable diet
MDD Minimum dietary diversity
MDD-C Minimum dietary diversity–children
MMF Minimum meal frequency
MNF Adult male no adult female
N Sample size
NGO Non-governmental organization
ORT Oral rehydration therapy
PPP Purchasing power parity
RFSA Resilience Food Security Activity
SE Standard error
USAID United States Agency for International Development
WASH Water, sanitation, and hygiene
WAZ Weight-for-Age Z-Score
WHZ Weight-for-Height Z-Score
WRA Women of reproductive age
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Executive Summary vii
EXECUTIVE SUMMARY
This report captures baseline round data and observations collected from May to June 2021 of the 
FIOVANA Resilience Food Security Activity (RFSA) that is being implemented by the Adventist 
Development Relief Agency (ADRA) in eastern Madagascar. This activity attempts to address and 
mitigate acute levels of food insecurity experienced by communities in this region of Madagascar. Apart 
from general demographics, study indicators include food security; child nutrition and health; women’s 
maternal nutrition and reproductive health; water, sanitation, and hygiene practice; agricultural practice 
and production; poverty measurement; gender dynamics; and resilience. A complimentary endline
survey is anticipated to be conducted from May to June of 2025.
Baseline Study Methodology
The impact evaluation (IE) of the FIOVANA RFSA relies on a clustered randomized controlled trial design 
to analyze differences between treatment and control groups. In the case of the FIOVANA RFSA IE, the 
intervention allowed for randomization at the commune level to create a set of treatment and control 
communes. As a result, the analysis will focus on the effect of the full RFSA on food security and other 
development outcomes across treatment households. Overall, analysis at baseline suggests that the IE is 
well placed to estimate these differences at endline, and that treatment and control groups are similar 
enough on key characteristics.
Study Limitations
Several factors posed potential challenges that the research team will adjust for or monitor throughout 
the course of the IE activity, such as the presence of several other large development organizations 
working in the region, which could present challenges in identifying the effects of the FIOVANA RFSA. 
This is complicated by the challenges imposed by the coronavirus that will potentially increase the 
involvement of outside aid groups. The research team will work closely with implementing partners to 
gauge this issue and take steps ahead of endline data collection to mitigate, if necessary. Finally, safety 
protocols that limited contact with beneficiaries and enforced social distancing measures ruled out the 
possibility of collecting some baseline indicators, specifically height measurements for women of 
reproductive age (WRA) and children.
Key Findings
Demographic Profiles
By the end of data collection efforts, 4,300 households had been surveyed. There were little to no 
observed trends of variance between the treatment and control groups on both the individual and 
household level, suggesting that the two groups are similar in terms of overall demographic 
characteristics. Self-reported household head characteristics follow a similar trend as other individual 
indicators in showing little difference between treatment and control households. 
Food Security
Reported severe drought conditions in the area suggest reduced food security across the entire region 
of the study. Food security was estimated using two standard measurement approaches, the Food 
Insecurity Experience Scale (FIES) and the Food Consumption Score (FCS) index. Based on the FIES, 18% 
of the population is facing severe food insecurity, and over 81% are at least moderately food insecure. 
The FCS, which calculates overall consumption levels across food groups while accounting for cultural 
IMPEL | Implementer-Led Evaluation and Learning
viii Executive Summary
and regional weights for food preference and importance, finds that fewer than 31% of all households 
have an acceptable FCS, and just under 7% of households are considered to have a poor FCS. 
Disaggregation by household adult type shows some differences as well, with 21% of households with 
an adult female and no adult male having an acceptable score, compared to 33% of households with 
both male and female adults present.
Child Nutrition and Health
Overall quality of diet for children 6 to 23 months of age appears to be poor across the survey 
population. Only 5% of all children aged 6 to 23 months met Minimum Dietary Diversity (MDD) criteria,
and only 3% met minimum acceptable diet standards. Approximately 11% of children under 5 are 
reported to have experienced diarrhea within the last 2 weeks. Of those that experienced diarrhea, only 
22% reported the use of oral rehydration therapy to treat symptoms. Of those that experienced fever, 
diarrhea, or acute respiratory infection, 58% report being treated at a health facility within 24 hours of 
symptoms. Given the ongoing COVID-19 pandemic, weight of children under 5 years of age was the only 
anthropometric indicator captured in the baseline round of surveys. More than a quarter of children are 
classified as underweight, and approximately 8% are severely underweight. Roughly 38% of children 
were within normal weight ranges.
Women’s Health, Maternal Nutrition, and Reproductive Health
Observations around the health and reproductive decisions of women of reproductive age among 
surveyed households suggest poor food consumption diversity but minimally acceptable levels of access 
to health personnel during pregnancy. Altogether, 9% of women consumed a diet that meets the MDD 
criteria. Nearly 65% of women consumed three or fewer food groups. On average, 75% of women 
received the recommended number of antenatal care visits (at least four) during their most recent 
pregnancies, with the majority of these (80%) being with midwives. Additionally, 33% of women report 
using a modern method of birth control.
Water, Sanitation, and Hygiene (WASH) Practices
Based on indicator criteria, 16% of all households have access to basic drinking water services. While 
most water sources are available year-round and within 30 minutes of a residence, only 19% of 
households have access to an improved source. Regarding treatment, sanitation, and hygiene practice,
only 4% of households were observed to have handwashing facilities available in the home, and half 
report treating water with filtering or disinfecting processes. Most households (66%) practice open 
defecation, and only 3% have household-level improved sanitation facilities. Among those not practicing 
open defecation, 84% utilize unimproved technology (uncovered pit latrine), and 48% share the facility 
with other households.
Agriculture
More than 94% of households are engaged in raising crops. Reported crops of focus include cassava, 
rice, and cloves. Of these, cassava and rice are grown by the large majority of farmers in the sample, 
84% and 92%, respectively. A little more than a quarter (28%) grow cloves as a cash crop. The average 
yield for cassava is slightly over 5,000 kg/ha. Rice yield is approximately 3,900 kg/ha. Clove yield is 
estimated to be 670 kg/ha. Regarding other support practices, the use of credit and savings is not 
common among farmers, and crop insurance is almost non-existent. Value chain participation activity 
suggests that 32% of farmers participate in at least one value chain activity. 
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Executive Summary ix
Livestock of focus in the baseline includes poultry, bees, and fish. In the baseline sample, 67% of farmers 
raised poultry, with only 27% reporting egg production in the past week. Of significant note is that 67% 
of poultry farmers report poultry dying in the last year, with most farmers (94%) reporting that they had 
not used any modern livestock raising practices such as vaccinations. Responses around beekeeping 
suggest that the practice is not widespread. Only 68 individuals in the survey reported raising bees, and 
a small number (15) utilized modern hives. Fishing was reported by 287 individuals, with the majority 
((97%) indicating that fish would be consumed in the home compared to 55% indicating that at least 
some of the fish would be brought to market for sale.
Poverty Measurement
Based on daily per capita expenditures of less than $1.90 per day (2011 purchasing power parity (PPP), 
the poverty rate among households surveyed is approximately 72%. The depth of poverty of the poor is 
33%, which means that the average poor person is 33% below the poverty line. In monetary terms, this 
means it would require an additional $0.63 per person per day to bring every poor person out of 
poverty.
Gender Dynamics
Gender dynamics are captured through the eight indicators in this section. Among women who are 
earning cash, a large majority (84%) reported that they participate in decisions about how to use the 
cash, whether solely or jointly with others. Women and men in a union report borrowing at similar rates 
across any source (36%). However, men in a union participate in decisions about credit at a much higher 
rate than women in a union (p-value = 0.00), with 87% of men reportedly making credit decisions, which 
is consistent across age groups.
Resilience
Resilience indicators were captured through several questions, including indices that were constructed 
to assess overall resilience capacities. Generally, households perceive their ability to meet their current 
needs as worse than the previous year and suspect that their future ability to meet these needs will 
deteriorate. A large share of households (80%) listed the drought as one of the shocks affecting them. 
Other common shocks listed included rising food prices, livestock disease, and crop pests. Out of the 
average of 2.3 shocks experienced across the sample, households perceived those shocks to be severe in 
nature, likely impacting perceived ability to recover. There does seem to be an overall trend that 
households with male adults both perceive a higher ability to recover from shocks and lower exposure 
and intensity of experienced shocks.
This page is intentionally left blank.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Introduction 1
1. INTRODUCTION
The primary objective of the impact evaluation is to measure the impact of the Resilience Food Security 
Activity (RFSA) package of interventions on improving health, nutrition, and income in the targeted 
communities and to determine possible attribution to changes in key indicators. The Adventist 
Development Relief Agency (ADRA) FIOVANA activities aim to achieve sustainable improvement of food 
and nutrition security and resilience of vulnerable populations in Atsimo Atsinanana and Vatovavy￾Fitovinany regions.
1 Figure 1 below shows the regions and communes where the intervention will be 
taking place. The activity serves a large population of households in southeastern Madagascar with 
interventions designed to improve maternal and infant health and nutrition, increase incomes, and 
improve resilience capacities. This report summarizes the results of the baseline study conducted in 
May–June 2021.
Figure 1. FIOVANA regions and communes 2
 1 https://adra.mg/our-projects/fiovana/
2 https://adra.mg/our-projects/fiovana/
IMPEL | Implementer-Led Evaluation and Learning
2 Introduction
The evaluation seeks to inform the larger conversation around the efficacy of RFSA interventions among 
vulnerable populations. The impact evaluation of the activity uses a randomized controlled trial design 
which randomized communities into both treatment and control groups.
The baseline study relies on quantitative methods to measure baseline indicators collected in the RFSA 
target areas, as well as control areas in the same region. The survey provides baseline estimates on the 
status of communities and households across the United States Agency for International Development’s 
(USAID) Bureau for Humanitarian Affairs (BHA) standard indicators. Causal Design has worked closely 
with BHA and relevant stakeholders to identify key learning objectives and to ensure that the baseline 
survey and study are able to contribute to this learning where possible.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Methodology and Limitations 3
2. METHODOLOGY AND LIMITATIONS
The impact evaluation of the FIOVANA RFSA uses a randomized controlled trial design at the commune 
level. A baseline survey of households was conducted in May–June 2021. The endline survey is planned 
for the second quarter of 2025. Below we describe an overview of the evaluation design. A more 
detailed description of the methodology can be found in Volume II, Annex C.
2.1 Evaluation Design
After discussion with the ADRA team, researchers deemed village or fokontany (village)-level 
randomization infeasible because (1) there are a number of planned commune-level interventions, and 
(2) there are logistical and political complications that arise when excluding some fokontany from 
activities within a commune. The communes are large, with an average population of over 11,000, 
which would generally be considered too large for randomization. In this case, however, numerous
communes meet ADRA’s inclusion criteria (124)—roughly double the number of communes in which 
ADRA had planned to work. Therefore, the best solution was to randomize at the commune level and 
later sample fokontany within the commune for the baseline survey. Based on our initial calculations, 
this commune-level randomization would be adequately powered to detect most impacts of interest.
2.1.1 Randomization Strategy
Selection Criteria
ADRA shared the criteria they used to determine if communes in the districts of Farafangana, Manakara, 
Mananjary, Vangaindrano, Vohipeno, and Vondrozo would be eligible for intervention. Based on these 
criteria, ADRA concluded that 124 communes of the 178 analyzed were eligible. We used these same 
criteria for our matching procedure. The selection criteria included population, level of poverty, length 
of the hungry season (“soudure”), remoteness, security, and access to a health clinic.
Matched Pairs
We chose a matched pair randomization approach to ensure better balance prior to baseline data 
collection compared to what could be achieved through stratified random sampling. In matched pair 
randomization, we first match units based on variables related to outcomes. In other words, we pair 
together communes that look “similar” based on available data. We then randomly assigned one unit 
from each pair to treatment and the other unit to control. Using ADRA’s selection criteria, we put 124 
communes into pairs or small groups of similar communes.
We made the groupings based on (1) being in the same district, (2) having similar population totals, (3) 
having similar distances to the district capital, and (4) having an ongoing United States Agency for 
International Development activity. We considered other criteria, but there was little variation in other 
factors such as poverty levels or soudure (hungry season) length. Three communes were unmatched 
due to outlier values on one or more of the characteristics. ADRA accidentally included the commune of 
Manakara in the list of selected communes and was not considered.
We shared the groupings with ADRA during our January 2020 trip. We asked the ADRA team to review 
our groupings and modify them as needed based on additional knowledge, including geographic location 
IMPEL | Implementer-Led Evaluation and Learning
4 Methodology and Limitations
(for example, coastal versus inland) or other characteristics deemed relevant for the types of 
interventions that might be carried out. We also suggested that they flag any communes they would not 
want to be considered for randomization for programmatic or logistical reasons. These would be taken 
out of consideration for pairing and assignment.
ADRA’s revised groupings resulted in minor adjustments to the list and produced groups of two to four 
communes. We made one additional change before randomization; we dropped the commune of 
Analavory in the district of Manakara, which was an outlier within its group in terms of population. This 
final list gave us a maximum of 54 pairs, meaning 54 communes each in treatment and control. After 
further investigation, ADRA found that at least two of the treatment communes were too remote to 
consider for implementation. The finalized list has 50 communes in treatment and 50 in control.
Sampling Calculation
We used the following assumptions to calculate the needed sample size for the baseline:
• We calculated sample size based on a cluster randomized design.
• Intra-class correlation. We used an intra-class correlation coefficient (ICC) of 0.12. The ICC 
estimate comes from height-for-age Z-scores (HAZ) in the Madagascar Enquête 
Anthropométrique et Développement de l'Enfant 2011 household data.
• Significance level of 5%.
• Power level of 80%.
• Expected reduction in stunting over the life of the activity of 8 percentage points.
• Inflation Factor for the number of households with children under 5 years old to sample. From the 
aforementioned 2011 data set, 24% of the population is under 5 years old in the provinces of 
Toliara and Fianarantsoa. This implies an inflation factor of 1.2.
• Non-response factor between baseline and endline. We used 5%.
To capture sufficient children under 5 years old and to account for attrition, we would need at least 
2 * 1,600 * 1.18 * 1.05 = 3,965 households total, or 20 households per cluster.
2.1.2 Sampling Frame
Fokontany within each commune were sampled using the 2018 INSTAT census, which was deemed to be 
the most consistent and updated data source available at the time. Researchers sampled Fokontany 
with probabilities proportional to size. Researchers sampled two fokontany in communes with under 
15,000 residents, three were sampled in communes with more than 15,000 residents, and one 
commune (Ambandrika) only had one fokontany. The number of communes in each category and the
resulting sample size are summarized below.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Methodology and Limitations 5
Table 1. Number of fokontany and households surveyed
Number of fokontany 
sampled in each 
commune
Number of communes
Total number of 
fokontany to be 
surveyed
Number of households to 
be surveyed (20 per 
fokontany)
1 1 1 20
2 83 166 3,320
3 16 48 960
Totals 100 215 4,300
Table 2. Individual response rate
Outcomes
Number in 
roster
Number 
surveyed
Response 
rate
Average age 
(years)
Women of reproductive age 5,214 4,885 93.7% 28.4
Children under 5 years old 4,916 4,211 85.7% 2.0
Farmers 4,192 4,135 98.6% 43.9
Women in a union 3,342 3,078 92.1% 36.4
Three of the sampled fokontany were replaced after teams discussed the local conditions with the 
commune mayor. Two fokontany were replaced because of security concerns, and one was replaced 
because it was very remote and would have required nearly a week’s travel on foot to reach all the 
hamlets. Replacement fokontany were randomly sampled from the remaining fokontany.
2.1.3 Questionnaire Development
The baseline survey was developed using previous baseline surveys used by BHA and refined in 
consultation with BHA and the implementing partners. 
The following survey modules were included:
• Module A: Household identification.
• Module B: Roster.
• Module C: Food access.
• Module D: Child nutrition and health.
• Module E: Women’s nutrition and health.
• Module F: Water, Sanitation, and Hygiene (WASH).
• Module G: Agriculture.
• Module H: Household expenditure.
• Module J: Gender and cash use.
• Module K: Gender and credit.
• Module R: Resilience.
In addition, we created a short commune-level survey to capture community-level variables such as 
public service availability, the activities of development or aid projects, and the presence of local 
community groups.
IMPEL | Implementer-Led Evaluation and Learning
6 Methodology and Limitations
2.1.4 Field Preparation
Travel restrictions during the COVID-19 pandemic prevented international travel. The team conducted 
training in the town of Manakara. Because most of the team had also worked on the nearly identical 
baseline survey for the Maharo RFSA, training was conducted over a shorter period. The team 
conducted full practice sessions over 3 days and conducted pilot surveys in two fokontany near 
Manakara. The training was interrupted by a 2-week quarantine after a possible COVID-19 case among 
the enumerators.
Dates of training:
6–13 April 2021, and 29 April–2 May (after quarantine).
Organization of teams:
There were 35 enumerators, 11 controllers, and 13 supervisors. There were 11 teams, 10 of which were 
assigned to specific communes, and one mobile team that joined other teams as needed. The 
controllers verified data quality, and one person from each team was charged with collecting weights of 
children.
2.1.5 Data Collection
Data collection began on May 4, 2021 and ended June 12, 2021. There were 4,300 households surveyed 
in the FIOVANA baseline. Table 3 depicts the household response rate.
Table 3. Household response rate3
Total households 
sampled Refused or absent Total households 
surveyed Response rate
Control 2,345 185 2,160 92.10%
Treatment 2,266 126 2,140 94.40%
All 4,611 311 4,300 93.30%
Household sampling
As household lists were not reliably available, field teams used the following method to sample 
households. First, working with fokontany leaders, the team would list each hamlet4 and the 
approximate number of households in each. The number of households to be surveyed in each hamlet 
was determined by a field method equivalent to sampling by probability proportional to size. Finally, at 
the hamlet level, random numbers were chosen, for example, 5, 7, and 19, and the 5th, 7th, and 19th 
homes were surveyed by following the same direction around the hamlet.
Data quality control
Each enumerator was introduced to a household by a supervisor or controller. The supervisor assisted at 
the beginning of the interview to ensure that the interview began well and that the interviewer 
recorded key household data (number of households, number of children, number of women of 
 3 Total sampled is the number of people who were contacted by the enumeration team (including replacements for refusals or 
absent households). Refused or absent account for people who were not available or were unable to be reached. Total 
surveyed are the people who were actually surveyed which aligns with our target sample.
4 Hamlets are groups of houses within the fokontany. This is not an additional level of sampling because hamlets are not 
sampled. Rather it is a way to randomly sample households in the absence of lists of names.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Methodology and Limitations 7
childbearing age, number of cash-earners in the household, etc.) in their notebook in addition to 
recording the data in the tablets. The notes confirmed that data was properly entered and that the 
correct numbers of people were included in the different modules. All data was collected via the mobile 
data collection software SurveyCTO. 
Two lead supervisors permanently monitored the field teams until the end of the surveys to verify the 
data and to solve problems promptly. Each team supervisor was asked to complete a fokontany 
information survey on their tablet to track the number of households sampled, number absent or 
refused, and number surveyed. In the case of households with unusual responses (case of 0 expenses, 
for example), the lead supervisors confirmed responses in person. Each team was visited by the lead 
supervisors at least twice during the field. This was necessary not only for the survey conduct but also to 
troubleshoot technical issues with the tablets, which was a collective problem for all teams. Five 
supervisors also listened to 5-minute recordings from the enumerators. These recordings covered the 
first 5 minutes of module H, household expenditures. However, most of the recordings were unusable 
due to various timing issues at the beginning of the module. 
2.2 Challenges and Limitations
2.2.1 Survey Challenges 
The start of the survey was initially delayed due to potential positive COVID-19 cases among the team 
members. The positive COVID-19 cases necessitated last-minute replacements of a supervisor and 
several enumerators. Ultimately, the COVID-19 protocols established for the baseline worked. We did 
not detect further illnesses or potential exposures once the team was in the field, and data quality 
should not have been affected. 
Because of problems with some of the tablets and issues with uploading survey data, we found some 
duplicate surveys in the raw data, and initially, some surveys we did not upload. We identified and 
dropped duplicates, and we identified and uploaded missing surveys. One tablet did not update the 
survey version successfully, so an older version was used for 38 surveys. However, we made only minor
updates after this version, which should not affect the validity of these surveys.
2.2.2 Limitations
• There are several other large development organizations working in the region. This could present 
challenges in identifying the effects of the FIOVANA RFSA. We are collecting information at the 
commune level at both the baseline and endline. This information will help us to control for other 
interventions.
• Yield calculations using recall data are necessarily rough estimates.
• The survey was conducted during the COVID-19 pandemic, and to maintain distance between 
enumerators and respondents and to minimize contact, only one anthropometric measure was 
collected at baseline. After discussions with BHA and the implementing partner, we determined 
that child weight could be safely measured if the caretaker was asked to weigh the child. Thus, we 
can calculate the weight-for-age z-score (WAZ) for children under 5 years old. We did not collect 
height for adults and children. However, because of the randomized design, collecting these 
measures at the endline will still be informative.
IMPEL | Implementer-Led Evaluation and Learning
8 Findings
3. FINDINGS
3.1 Characteristics of the Study Population
This section provides the basic demographic information for the baseline sample. As mentioned 
previously, we surveyed 4,300 households. Table 4 shows that most households have both a male and 
female adult present. For the whole sample, 16% of households have only an adult female present, and 
nearly 2% have only an adult male present. The average household has 5.8 people, including two 
children under the age of 5 years old. Results are not reported for cells with less than 30 observations 
and are denoted with an n/a for not available.
Table 4. Basic household-level statistics
Control Treatment All
Percent households with adult male and female 83.1% 82.7% 82.9%
Percent households with adult female only 15.8% 16.7% 16.3%
Percent households with adult male only 1.1% 0.6% 0.9%
Household size 5.77 5.92 5.84
Percent of households engaged in farming 94.6% 93.7% 94.1%
Percent responding as head of household* 48.5% 45.3% 47.0%
* This is based on who responded to module F, a section that asks to speak to an adult or the head of household or 
a responsible adult. All other sections asked specifically for the head of household or other members of the family 
(e.g. farmers, caretakers of children under 2)
Table 5 provides the basic individual demographic information of the sample. The household sample 
represents over 25,000 individuals. The average age is 21 years old, and approximately 51% of those in 
the sample are female. Figure 2 shows the population by age and gender.5 Among those 15 years old
and over, 33% engage in some type of farming or livestock activity. For those over 10 years old, 67% 
worked for cash in the previous year. Slightly more than half of those school-age or older have at least 
some schooling.
Table 5. Baseline individual-level demographic information
Control Treatment All
Outcome Mean
Standard 
Error (SE) N Mean SE N Mean SE N
Age 20.8 (0.19) 12,535 20.9 (0.20) 12,555 20.85 (0.14) 25,090
Percent of sample who are
female 51.1% (0.52) 12,535 51.4% (0.52) 12,557 51.2% (0.37) 25,092
Percent of adults 15 years and 
older who are farmers 34.6% (0.70) 6,324 31.5% (0.68) 6,415 33.0% (0.49) 12,739
Percent of school age persons
with at least some schooling 65.9% (0.54) 10,409 68.9% (0.52) 10,409 67.4% (0.38) 20,818
Percent of adults 15 years and 
older who worked for cash 63.4% (0.71) 6,253 61.7% (0.71) 6,361 62.5% (0.50) 12,614
 5 The figure shows the age distribution across the full sample 
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 9
Control Treatment All
Outcome Mean
Standard 
Error (SE) N Mean SE N Mean SE N
Percent of sample women of 
reproductive age 30.4% (0.48) 12,535 30.4% (0.48) 12,557 30.4% (0.34) 25,092
Percent of sample under 5
years old 16.5% (0.38) 12,535 17.1% (0.39) 12,557 16.8% (0.27) 25,092
Percent of sample 5–14 years 
old 33.2% (0.49) 12,535 31.9% (0.49) 12,557 32.6% (0.34) 25,092
Percent of sample 10-24 years 
old 34.1% (0.00) 12535 33.6% (0.00) 12,557 33.9% (0.00) 25,092
Figure 2. Age and gender composition (unweighted) of the sample
Table 6 lists the characteristics of the person self-identified as the head of household. The sample is 
comprised of 27% female-headed households. Roughly 32% of household heads have no formal 
schooling, and 35% completed primary school.
IMPEL | Implementer-Led Evaluation and Learning
10 Findings
Table 6. Head of household statistics
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Age of head of households 44.65 (0.38) 2,156 45.46 (0.39) 2,141 45.06 (0.27) 4,297
Percent of female head of 
households 25.3% (1.09) 2,157 27.9% (1.15) 2,142 26.6% (0.79) 4,299
Percent of head of households that 
did not attend school 33.0% (1.18) 2,157 31.9% (1.17) 2,142 32.4% (0.83) 4,299
Percent of head of households with 
some schooling, less than primary 34.2% (1.19) 2,157 30.2% (1.18) 2,142 32.2% (0.84) 4,299
Percent of heads of households that 
completed primary or more 32.8% (1.17) 2,156 37.9% (1.23) 2,141 35.4% (0.85) 4,297
3.2 Food Security
The first indicator in this section is the prevalence of food insecurity, which is measured using the Food 
Insecurity Experience Scale (FIES) developed by the Food and Agriculture Organization (FAO).
The responses to these questions are analyzed by estimating a Rasch model using tools developed by 
the FAO in R software.6 7 Only households responding to all questions are included in the analysis. 
Furthermore, extreme responses—those responding either “yes” or “no” to all questions—are excluded 
from the estimation but included in the final calculation of the prevalence rates. 
Table 7 summarizes the prevalence of moderate and severe food insecurity in the household based on 
the FIES. In the sample, approximately 18% of individuals are considered to be experiencing severe food 
insecurity, and over 81% are at least moderately food insecure. Severe food insecurity is higher in 
treatment areas (21%) compared to control (15%). The t-test comparing treatment and control based on 
the number of questions answered “yes” has a p-value of 0.00, confirming the higher food insecurity in 
treatment areas. The answers to the eight questions are disaggregated by household type in Table 8.
Table 7 shows the percent of respondents answering “yes” to each question. Overall, 16% of households 
responded “yes” to all eight questions for the period of the previous 30 days. Only 2.5% (107) 
households answered “no” to all questions. Consistent with the assumption that each question captures 
a progressively more severe experience, the proportion of “yes” answers generally declines with each 
question. The exception is between questions 4 and 5. More households (87%) reported eating less 
because of lack of money or other resources than reported skipping a meal (60%). 
Table 7. Prevalence of moderate and severe food insecurity in the population
Prevalence rate (not 
Moderate or Severe)
Prevalence rate 
(Moderate + Severe)
Prevalence rate 
(Severe) N
All 18.7% 81.3% 17.9% 4,293
Treatment 15.7% 84.3% 20.9% 2,137
Control 21.8% 78.2% 15.2% 2,156
 6 The model produced a reliability score of 0.86, which suggests a good model fit.
7 Cafiero et al., Methods for Estimating Comparable Prevalence Rates of Food Insecurity Experienced by Adults throughout the 
World, (Rome, Italy: FAO, 2016)
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 11
Table 8. Percentage of households responding “yes” to “During the past 30 days, was there a time 
when you or others in your household…” disaggregated by household type*
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
...were worried you would not have 
enough food to eat because of a 
lack of money or other resources 91.6% (1.22) 2,115 92.5% (1.17) 2,097 92.1% (0.61) 4,212
F&M Worried 91.3% (1.26) 1,771 92.5% (1.21) 1,731 91.9% (0.67) 3,502
FNM Worried 93.0% (2.12) 319 93.4% (1.79) 354 93.2% (1.25) 673
MNF Worried n/a n/a 25 n/a n/a 12 85.2% (8.26) 37
...were unable to eat healthy and 
nutritious food because of a lack of 
money or other resources 92.6% (1.29) 2,115 95.8% (0.86) 2,097 94.2% (0.74) 4,212
F&M Healthy 92.6% (1.14) 1,771 95.7% (0.85) 1,731 94.1% (0.66) 3,502
FNM Healthy 93.2% (2.39) 319 96.9% (1.18) 354 95.1% (1.36) 673
MNF Healthy n/a n/a 25 n/a n/a 12 80.9% (9.89) 37
...ate only a few kinds of foods 
because of a lack of money or other 
resources? 92.6% (1.28) 2,115 95.9% (0.77) 2,097 94.3% (0.73) 4,212
F&M Few foods 92.8% (1.16) 1,771 95.8% (0.78) 1,731 94.3% (0.65) 3,502
FNM Few foods 92.4% (2.35) 319 97.0% (1.06) 354 94.8% (1.41) 673
MNF Few foods n/a n/a 25 n/a n/a 12 n/a n/a 37
...had to skip a meal because there 
was not enough money or other 
resources to get food? 56.7% (2.43) 2,115 61.2% (2.55) 2,097 58.9% (1.59) 4,212
F&M Skipped 56.0% (2.58) 1,771 59.7% (2.55) 1,731 57.8% (1.64) 3,502
FNM Skipped 60.6% (3.68) 319 69.7% (3.92) 354 65.3% (2.71) 673
MNF Skipped n/a n/a 25 n/a n/a 12 40.6% (10.10) 37
...ate less than you thought you 
should because of a lack of money 
or other resources? 84.9% (1.89) 2,115 88.8% (1.29) 2,097 86.9% (1.17) 4,212
F&M Ate less 83.7% (2.01) 1,771 88.1% (1.45) 1,731 85.9% (1.29) 3,502
FNM Ate less 90.6% (2.01) 319 93.4% (1.47) 354 92.0% (1.25) 673
MNF Ate less n/a n/a 25 n/a n/a 12 80.8% (9.05) 37
...did not have food because of a 
lack of money or other resources? 23.7% (1.97) 2,115 29.8% (2.56) 2,097 26.7% (1.46) 4,212
F&M Runout 22.4% (2.11) 1,771 28.3% (2.66) 1,731 25.4% (1.50) 3,502
FNM Runout 30.2% (3.55) 319 37.1% (3.72) 354 33.7% (2.91) 673
MNF Runout n/a n/a 25 n/a n/a 12 n/a n/a 37
...were hungry but did not eat 
because there was not enough 
money or other resources? 23.8% (2.73) 2,115 31.4% (3.00) 2,097 27.6% (1.73) 4,212
F&M Hungry 23.2% (2.73) 1,771 30.2% (3.09) 1,731 26.7% (1.76) 3,502
FNM Hungry 27.8% (4.05) 319 37.4% (3.49) 354 32.7% (2.63) 673
IMPEL | Implementer-Led Evaluation and Learning
12 Findings
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
MNF Hungry n/a n/a 25 n/a n/a 12 18.4% (7.19) 37
...went without eating for a whole 
day because of a lack of money or 
other resources? 17.7% (1.95) 2,115 23.3% (2.84) 2,097 20.5% (1.36) 4,212
F&M Whole Day 17.1% (2.02) 1,771 22.4% (2.83) 1,731 19.7% (1.35) 3,502
FNM Whole Day 21.4% (3.01) 319 28.1% (3.78) 354 24.9% (2.25) 673
MNF Whole Day n/a n/a 25 n/a n/a 12 9.9% (5.46) 37
Not applicable (n/a)
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF).
The second indicator is the Food Consumption Score (FCS). This is a weighted sum of eight food groups 
consumed by the household in the previous 7 days. The weights are based on the food group’s 
importance in the diet. For example, meat and dairy have a weight of 4, staples have a weight of 2, and 
sugars have a weight of 0.5. The FCS ranges from 0 to 112. Scores below 22 are considered to have a 
poor consumption score. Scores 22–35 are considered borderline. Acceptable scores are above 35. As 
shown in Table 9, the mean FCS is 33, and approximately 7% of individuals are considered to have poor 
FCS. Fewer than 31% of households have an acceptable FCS. The majority of households have a 
borderline FCS. 
The FCSs disaggregated by household type are in Table 10. Households with both an adult male and female 
present have a higher FCS than female-only households (p-value = 0.00). There are too few male-only 
households to detect whether the FCSs for these households differ from those of other groups. 
Table 9. Mean scores of households with poor, borderline, and acceptable Food Consumption Score
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household FCS (0–112) 32.89 (0.50) 2115 32.3 (0.58) 2097 32.6 (0.32) 4212
F&M 33.25 (0.51) 1771 32.73 (0.57) 1731 32.99 (0.31) 3502
FNM 31.09 (0.62) 319 30.11 (0.85) 354 30.59 (0.51) 673
MNF n/a n/a 25 n/a n/a 12 32.9 (2.72) 37
Not applicable (n/a)
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
Table 10. Percentage of households with poor, borderline, and acceptable Food Consumption Score
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households with 
poor consumption score (<22) 5.9% (0.92) 2,115 8.0% (1.44) 2,097 6.9% (0.73) 4,212
F&M poor 5.3% (0.96) 1,771 6.9% (1.37) 1,731 6.1% (0.69) 3,502
FNM poor 7.8% (1.89) 319 12.8% (2.49) 354 10.4% (1.58) 673
MNF poor n/a n/a 25 n/a n/a 12 20.3% (9.20) 37
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 13
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households with 
borderline consumption score 
(22–35) 62.0% (2.13) 2,115 62.3% (2.10) 2,097 62.1% (1.21) 4,212
F&M border 60.8% (2.23) 1,771 61.1% (2.28) 1,731 61.0% (1.32) 3,502
FNM border 68.8% (3.48) 319 68.%2 (3.33) 354 68.5% (2.45) 673
MNF border n/a n/a 25 n/a n/a 12 55.0% (9.13) 37
Percent of households with an 
acceptable consumption score 
(>35) 32.1% (2.63) 2,115 29.8% (2.89) 2,097 30.9% (1.52) 4,212
F&M acceptable 33.9% (2.73) 1,771 32.0% (3.05) 1,731 32.9% (1.57) 3,502
FNM acceptable 23.4% (3.17) 319 19.0% (3.46) 354 21.1% (2.21) 673
MNF acceptable n/a n/a 25 n/a n/a 12 24.7% (8.14) 37
Not applicable (n/a)
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
3.3 Child Nutrition and Health
This section presents findings on child health and nutrition. This includes quality of diet, breastfeeding 
practices, healthcare treatment, and incidences of diarrhea. This section also includes child 
anthropometric data. Due to safety concerns brought on by COVID-19, however, we were only able to 
capture and calculate weight at baseline.
3.3.1 Nutrition
Prevalence of children 6–23 months consuming a diet of minimum dietary diversity (MDD) and percent 
of children 6–23 months receiving a minimum acceptable diet (MAD) are the indicators of nutrition
reported for children under 2 years old. The MDD-Children (MDD-C) uses the following eight food 
groups: (1) breastmilk, (2) grains, roots, and tubers, (3) legumes and nuts, (4) dairy products (milk, 
yogurt, and cheese), (5) flesh foods (meat, fish, poultry, and liver/organ meats), (6) eggs, (7) vitamin A￾rich fruits and vegetables, and (8) other fruits and vegetables. The criteria for achieving an MDD-C is 
consuming at least five of the eight food groups. 
Table 11 shows the percentage of children meeting these criteria. Of the children sampled, 5% met 
these criteria. Grains are the most consumed (80%), followed by vitamin-A-rich vegetables and fruits 
(61%), dairy (6%), other vegetables and fruits (5%), flesh foods (4%), and legumes (2%). Eggs were 
consumed the least by children (1%).
Table 11. Prevalence of children 6–23 months consuming a diet of minimum diversity 
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Children (ages 6–23 months) 6.8% (0.012) 639 3.9% (0.011) 639 5.4% (0.007) 1,278
Male children (ages 6–
23 months) 7.4% (0.016) 304 4.4% (0.013) 317 5.9% (0.008) 621
IMPEL | Implementer-Led Evaluation and Learning
14 Findings
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Female children (ages 6–
23 months) 6.3% (0.017) 335 3.4% (0.014) 322 4.9% (0.011) 657
MAD is defined as eating a certain number of times per day (minimum meal frequency) in addition to 
having an MDD. The MAD uses six or seven food groups depending on whether the child is breastfed or 
not to define MDD.8 Table 12 shows that only 3% of children are receiving a MAD. Roughly the same 
percentage of boys and girls are receiving a MAD (3%). 
Minimum meal frequency (MMF) is used to calculate the MAD. The MMF must be calculated for both 
breastfed and non-breastfed children. The criteria to meet this if the child is breastfeeding is that they 
have to have three or more feedings of solid, semi-solid, or soft foods and be between the ages of 9 and 
23 months. In total, 46.5% of breastfed children met these criteria, which is illustrated in Table 13. The 
criteria for non-breastfed children is that they have to be between the age of 6 and 23 months and have 
four or more feedings of solid, semi-solid, or soft foods in addition to two or more milk feedings. 
Breastfed children 6–8 months must receive at least two feedings. Table 13 shows that 4.4% of non￾breastfed children surveyed met this standard. 
Table 12. Percentage of children ages 6–23 months receiving a minimum acceptable diet
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Children (ages 6–23 months) 3.8% (0.009) 639 2.5% (0.007) 639 3.1% (0.005) 1,278
Male children (ages 6–23 
months) 4.0% (0.013) 304 2.6% (0.010) 317 3.3% (0.008) 621
Female children (ages 6–
23 months) 3.5% (0.010) 335 2.3% (0.011) 322 2.9% (0.008) 657
Table 13. Minimum meal frequency, breast-fed and non-breastfed children 6–23 months
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Minimum meal frequency for 
breastfed children ages 6–23 
months 48.7% (0.034) 556 44.3% (0.037) 538 46.5% (0.019) 1,094
Minimum meal frequency for 
non-breastfed children ages 6–
23 months 2.1% (0.021) 83 5.9% (0.027) 101 4.4% (0.019) 184
Table 14 illustrates the prevalence of exclusive breastfeeding among children under 6 months. This is 
defined as the children under 6 months who were exclusively breastfed during the day preceding the 
survey, excluding any oral rehydration solution. Overall, 64% of children under the age of 6 months were 
exclusively breastfed. 
 8 The MAD does not include breastmilk as a food group. It includes dairy products as a food group for breastfed children and 
excludes them for non-breastfed children
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 15
Table 14. Prevalence of exclusive breastfeeding of children under 6 months of age
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Children (under 6 months of 
age) 60.8% (0.047) 176 65.6% (0.042) 209 63.5% (0.030) 385
Male children (under 6 
months of age) 65.4% (0.047) 88 65.9% (0.060) 97 65.6% (0.036) 185
Female children 
(under 6 months of 
age) 55.9% (0.071) 88 65.4% (0.051) 112 61.4% (0.045) 200
The next two indicators focus on the percentage of children under 5 (0–59 months) who had diarrhea in 
the prior 2 weeks and the percentage of children under 5 (0–59 months) with diarrhea treated with oral 
rehydration therapy (ORT). A positive incidence is defined as a child experiencing an episode of diarrhea 
any time in the 2 weeks that preceded the survey, while ORT is defined as receiving an oral rehydration 
solution, recommended home fluids, or increased fluids. According to the survey results, 11% of children 
experienced diarrhea. Both boys and girls across the control and treatment areas had diarrhea at a 
similar rate. The prevalence of children who had diarrhea and were treated with ORTs is about 23% 
overall.
Table 15. Percentage of children under 5 (0–59 months) years old who had diarrhea in the prior 2
weeks
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Children (ages 0–59 months) 11.8% (0.016) 2,089 10.8% (0.011) 2,112 11.3% (0.008) 4,201
Male children (ages 0–59 
months) 11.9% (0.017) 1,044 10.3% (0.011) 1,016 11.1% (0.009) 2,060
Female children (ages 0–
59 months) 11.8% (0.017) 1,045 11.2% (0.015) 1,096 11.5% (0.010) 2,141
Table 16. Percentage of children under 5 (0–59 months) years old with diarrhea treated with Oral 
Rehydration Therapy
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Children (ages 0–59 months) 17.4% (0.034) 243 27.8% (0.038) 234 22.5% (0.025) 477
Male children (ages 0–
59 months) 16.6% (0.041) 122 33.1% (0.068) 114 24.5% (0.039) 236
Female children (ages 0–
59 months) 18.2% (0.052) 121 23.2% (0.050) 120 20.7% (0.037) 241
Table 17 displays a custom indicator that shows the percent of children under 5 years old who have had 
diarrhea or acute respiratory infection (ARI) in the past 15 days and received treatment within 24 hours 
from a health facility. Children received treatment for fever (64%) at a higher rate than for diarrhea 
(44%) and for ARI (43%). However, there were more children who had a fever and received treatment 
over the past 15 days than those who had diarrhea or ARI. 
IMPEL | Implementer-Led Evaluation and Learning
16 Findings
Table 17. Percentage of children under 5 years old who had fever, diarrhea, or acute respiratory 
infection in the past 15 days and received treatment within 24 hours from a health facility
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Children (0–59 months) who 
received treatment within 24 hours 
from a health facility 54.8% (0.030) 660 59.6% (0.029) 772 57.5% (0.020) 1,432
Children (0–59 months) 
who had a fever and 
received treatment within 
24 hours from a health 
facility 59.8% (0.035) 473 67.8% (0.034) 541 64.2% (0.023) 1,014
Children (0–59 months) 
who had diarrhea and 
received treatment within 
24 hours from a health 
facility 44.2% (0.041) 171 44.6% (0.059) 174 44.4% (0.033) 345
Children (0–59 months) 
who had ARI and received 
treatment within 24 hours 
from a health facility 40.1% (0.064) 114 45.3% (0.048) 141 43.2% (0.040) 255
3.3.2 Anthropometry
Anthropometric indicators traditionally include measures of prevalence rates of wasting (weight-for￾height z-score (WHZ)), stunting (HAZ), and being underweight (WAZ). Given limitations to data collection 
related to COVID-19 safety protocols, only the weight of children under 5 years old was captured in the 
baseline round of surveys. Children without a known birth month and year were excluded. Children with 
WAZ scores of less than -2 standard deviations are considered underweight, and those with scores of 
less than -3 standard deviations are considered severely underweight. A little over a quarter of children 
were underweight, and 8% were severely underweight. Roughly 38% of children are considered to have 
a normal weight. Figure 3 shows the distribution of WAZ scores.9
 9 The figure shows WAZ scores across the full sample of children under 5 years of age.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 17
Table 18. Prevalence of underweight children under 5 years old
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Weight-for-age z-score (WAZ) -1.3 (0.05) 2,049 -1.3 (0.06) 2,047 -1.3 (0.04) 4,096
Percent of children under 
5 years old underweight 26.3% (1.27) 2,049 25.5% (1.76) 2,047 25.9% (1.09) 4,096
Percent of children under 
5 years old severely 
underweight 9.1% (0.94) 2,049 6.7% (1.03) 2,047 7.9% (0.76) 4,096
Percent of children under 
5 years old with a normal 
weight* 39.3% (1.82) 2,049 37.6% (1.62) 2,047 38.4% (1.12) 4,096
* Children with WAZ scores between -1 and 2 are considered normal weight. 
Figure 3. Distribution of weight-for-age z-score
Table 19 illustrates the underweight statistics disaggregated by gender and age group. Boys have a 
lower WAZ on average (-1.41) than girls (-1.15) (p-value = 0.00). Older children (2–5 years old) also have 
a lower z-score than infants under 2 years old (p-value = 0.00). This is also reflected in both the 
underweight and severely underweight percentages. 
IMPEL | Implementer-Led Evaluation and Learning
18 Findings
Table 19. Prevalence of underweight children under 5 years old, disaggregated
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Female WAZ -1.15 (0.07) 1,033 -1.15 (0.06) 1,072 -1.15 (0.04) 2,105
Male WAZ -1.41 (0.06) 1,016 -1.4 (0.08) 975 -1.41 (0.05) 1,991
0–23 months WAZ -0.94 (0.08) 796 -1.07 (0.11) 810 -1.01 (0.06) 1,606
24–59 months WAZ -1.5 (0.05) 1,253 -1.4 (0.06) 1,237 -1.45 (0.04) 2,490
Percent of female children 
under 5 who are underweight 23.4% (1.46) 1,033 22.4% (1.95) 1,072 22.9% (1.25) 2,105
Percent of male children under 
5 who are underweight 29.3% (1.83) 1,016 28.8% (2.09) 975 29.0% (1.38) 1,991
Percent of children ages 0–23 
months underweight 21.9% (1.75) 796 25.4% (2.67) 810 23.7% (1.67) 1,606
Percent of children ages 24–59 
months underweight 29.1% (1.52) 1,253 25.5% (1.81) 1,237 27.3% (1.24) 2,490
Percent of female children 
under 5 years that are severely 
underweight 7.7% (1.30) 1,033 6.1% (1.50) 1,072 6.9% (0.91) 2,105
Percent of male children under 
5 years that are severely 
underweight 10.5% (1.13) 1,016 7.3% (1.32) 975 8.9% (0.90) 1,991
Percent of children 0–23 
months that are severely 
underweight 8.2% (1.31) 796 7.8% (1.59) 810 8.0% (1.08) 1,606
Percent of children 24–59 
months that are severely 
underweight 9.7% (1.14) 1,253 6.0% (0.91) 1,237 7.8% (0.81) 2,490
Female children under 5 years 
with a normal weight 43.4% (1.98) 1,033 41.2% (2.13) 1,072 42.3% (1.32) 2,105
Male children under 5 with a 
normal weight 35.1% (2.44) 1,016 33.8% (1.98) 975 34.4% (1.61) 1,991
Percent of children 0–23 
months with a normal weight 48.6% (2.66) 796 44.0% (3.05) 810 46.3% (1.93) 1,606
Percent of children 24–59 
months with a normal weight 33.4% (1.87) 1,253 33.5% (1.97) 1,237 33.5% (1.06) 2,490
* Children with WAZ scores between -1 and 2 are considered normal weight.
3.4 Women’s Health, Maternal Nutrition, and Reproductive 
Health
This section focuses on the health and reproductive decisions of women of child-bearing age. The first 
indicator measures the percent of women of reproductive age (WRA) consuming a diet of minimum 
dietary diversity (MDD-W). WRA includes all women in the household 15–49 years old. MDD is 
measured by counting the number of food groups a woman consumed during the previous day and 
night. The food groups are grains, white roots and tubers, and plantains, pulses (beans, peas, and 
lentils), nuts and seeds, dairy, meat, poultry and fish, eggs, dark green leafy vegetables, other vitamin A-
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 19
rich fruits and vegetables, other vegetables, and other fruits. The criteria for MDD is met when a woman 
eats at least five of the 10 food groups specified.
Table 20 shows that about 3% of women consumed a diet that met the MDD requirements. Figure 4
illustrates the distribution of the MDD score.10 Almost 65% of women consumed three or fewer food 
groups, with almost 60% of women consuming three food groups. The most common food group 
consumed was grains (99%). The next most common food groups consumed are dark green leafy 
vegetables (88%), vitamin A-rich fruits and vegetables (27%), other fruit (21%), meat, poultry, and fish 
(17%), pulses (beans, peas, and lentils) (8%), eggs (1%), dairy (1%), and nuts (1%).
Table 20. Percentage of women of reproductive age consuming a diet of minimum diversity
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Women ages 15–49 years with MDD 3.2% (0.006) 2,416 3.4% (0.008) 2,467 3.3% (0.005) 4,883
Women (ages 15–18) 3.8% (0.012) 465 3.3% (0.011) 495 3.5% (0.008) 960
Women (age 19+) 3.1% (0.006) 1,951 3.4% (0.008) 1,972 3.3% (0.005) 3923
*Minimum dietary diversity (MDD)
Figure 4. Dietary diversity score for women of reproductive age
 10 The figure shows the distribution of the MDD score across the full sample of women of reproductive age.
0 10 20 30 40 50
Percent
-1 0 1 2 3 4 5 6 7 8 9 10
Dietary diversity score
IMPEL | Implementer-Led Evaluation and Learning
20 Findings
Table 21 shows the percent of women of reproductive age with a live birth in the past 5 years who 
received at least four antenatal care (ANC) visits from skilled health personnel during their most recent 
pregnancies. Skilled health personnel refers to doctors, nurses, midwives, skilled birth attendants, or 
clinical officers. Women qualify as having received ANC if they have had at least four ANC visits. Across 
both control and treatment groups, about 67% of these women met these criteria during their most 
recent pregnancy. More than 80% of these ANC visits were with midwives.
Table 21. Percentage of births receiving at least four antenatal care visits during pregnancy
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
ANC for WRA who had a live birth 
during the last 5 years*
66.1% (0.029) 1,347 68.2% (0.024) 1,372 67.1% (0.018) 2,719
*Women of reproductive ages 15–49
*Antenatal care (ANC)
The Contraceptive Prevalence Rate measures non-pregnant women of reproductive age (15–49) who 
are currently using or whose partners are currently using at least one contraceptive method. Table 22
displays this rate. Nearly 33% of non-pregnant women between the ages of 15–49 in a union are using 
birth control, with all of the women using modern birth control. Modern birth control includes female 
and male sterilization, injectables like Depo-Provera, intrauterine devices, contraceptive pills, implants, 
and female and male condoms. Traditional birth control methods include rhythm and withdrawal. 
Table 22. Contraceptive prevalence rate
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Non-pregnant women 15–
49 in a union using birth 
control
35.2% (0.027) 1,178 29.6% (0.025) 1,125 32.5% (0.014) 2,303
Non-pregnant 
women aged 15–
49 in a union using 
modern birth 
control
35.2% (0.027) 1,178 29.6% (0.025) 1,125 32.5% (0.014) 2,303
Non-pregnant
women aged 15–
49 in a union 
traditional birth 
control 0.0% n/a 0 0.0% n/a 0 0.0% n/a 0
Table 23 depicts the percent of women in a union who have knowledge of modern family planning 
methods that can be used to delay or avoid pregnancy. Of the women in a union between the ages of 15 
and 49, 78% reported having knowledge of modern family planning methods. The rate at which women
in a union reported having knowledge of modern family planning methods is similar across age groups.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 21
Table 23. Percentage of women in a union who have knowledge of modern family planning methods 
that can be used to delay or avoid pregnancy
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Women in a union (ages 15–49) 78.0% (0.031) 1,302 78.2% (0.034) 1,263 78.1% (0.018) 2,565
Women in a union (15–
19)
74.3% (0.048) 125 69.4% (0.059) 119 71.8% (0.039) 244
Women in a union
(ages 20–29)
83.7% (0.026) 457 78.3% (0.036) 445 81.0% (0.021) 902
Women in a union
(ages 30–49)
74.9% (0.042) 720 79.7% (0.035) 699 77.3% (0.020) 1,419
Table 24 presents findings on decision-making about family planning. The questions used to calculate 
this indicator are the following: “Did you or your partner use any of these modern contraceptive 
methods in the past 12 months?” and “Who usually makes the decision on whether or not you should 
use contraception, you, your (husband/partner), you and your (husband/partner) jointly, or someone 
else?” Of the women who have used modern family planning methods in the past 12 months, almost 
93% report making that decision either by themselves or with their partner jointly. 
Table 24. Percent of women in a union who made decisions alone or with their partner jointly about 
modern family planning methods 
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Women in a union (ages 15–49) 92.7% (0.021) 482 93.6% (0.019) 429 93.1% (0.008) 911
Women in a union (ages 
15–19)
90.8% (0.056) 34 n/a n/a 20 92.5% (0.038) 54
Women in a union (ages 
20–29)
94.2% (0.026) 181 96.6% (0.016) 158 95.3% (0.013) 339
Women in a union (ages 
30–49)
91.8% (0.024) 267 91.6% (0.027) 251 91.7% (0.012) 518
Table 25 is a custom indicator that shows women of reproductive age practicing at least one modern
method of birth control by age regardless of whether they are in a union or not. About 24% of women 
are practicing a modern method of birth control, with women between the ages of 20 and 49 using 
modern methods more (24%) than adolescent girls (9%) (p-value = 0.00). 
Table 25. Percent of non-pregnant adolescent girls and women of reproductive age practicing a
modern method of birth control
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Adolescent girls and women of 
reproductive age (ages 15–49) 26.1% (0.022) 2,191 21.9% (0.016) 2,235 24.0% (0.012) 4,426
Adolescent girls (ages 
15–19) 14.7% (0.027) 513 8.9% (0.014) 529 11.8% (0.016) 1,042
Women of 
reproductive age
(ages 20–49) 29.6% (0.022) 1,678 25.9% (0.020) 1,706 27.7% (0.013) 3,384
IMPEL | Implementer-Led Evaluation and Learning
22 Findings
3.5 Water, Sanitation, and Hygiene Practices
The percent of households using basic drinking water services indicator is measured by three criteria: (1) 
having access to an improved water source, such as a public tap or protected well, (2) having that source 
within 30 minutes round-trip of the home, and (3) having that source available year-round. Fewer than 
16% of all households meet all three of these criteria. While nearly all sources are available year-round, 
and 84% have water within 30 minutes, only 19% have access to an improved source.
Table 26. Percentage of households using basic drinking water services
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households with 
improved water source 16.9% (2.67) 2,159 20.7% (3.31) 2,140 18.8% (1.99) 4,299
F&M 15.9% (2.39) 1,807 19.6% (3.36) 1,769 17.7% (1.96) 3,576
FNM 22.3% (5.07) 325 25.5% (4.64) 359 24.0% (3.54) 684
MNF n/a n/a 27 n/a n/a 12 22.8% (8.35) 39
Percent of households with 
water source within 30 minutes 95.4% (0.85) 2,159 96.5% (0.58) 2,140 95.9% (0.59) 4,299
F&M 95.7% (0.79) 1,807 96.6% (0.60) 1,769 96.1% (0.55) 3,576
FNM 93.5% (2.18) 325 95.7% (1.24) 359 94.7% (1.34) 684
MNF n/a n/a 27 n/a n/a 12 100.0% (.) 39
Percent of households with 
water available year-round 86.0% (1.80) 2159 82.7% (2.34) 2140 84.3% (1.23) 4,299
F&M 85.3% (1.97) 1807 82.4% (2.44) 1769 83.9% (1.32) 3,576
FNM 89.8% (2.17) 325 83.6% (3.58) 359 86.6% (2.13) 684
MNF n/a n/a 27 n/a n/a 12 86.9% (8.37) 39
Percent of households with 
access to basic drinking water 
services 13.9% (2.23) 2159 16.1% (3.02) 2140 15.0% (1.82) 4,299
F&M 13.1% (1.94) 1807 15.3% (3.04) 1769 14.2% (1.80) 3,576
FNM 18.3% (4.74) 325 18.9% (3.72) 359 18.7% (3.02) 684
MNF n/a n/a 27 n/a n/a 12 22.8% (8.35) 39
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
Water use was estimated for only those treating water. This was done by asking what containers were 
used to collect water and how often water was collected. The volume of the containers was verified by 
the enumerator. On average, households report using 16 liters of water per person per day. Only 23% of 
households use the recommended 20 liters per day. Recalculating the BL16 indicator, the percentage of 
households using basic drinking water services, to include water use, recognizing the smaller sample, 
only 3.6% of households have access to basic drinking water services.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 23
Table 27. Water use per capita
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Water use per capita per 
day(liters) 16.6 (0.69) 990 14.8 (0.64) 1,128 15.6 (0.42) 2,118
F&M 15.53 (0.69) 829 14.02 (0.71) 929 14.72 (0.45) 1,758
FNM 21.65 (1.89) 149 18.71 (1.13) 193 20.00 (1.07) 342
MNF n/a n/a 12 n/a n/a 6 n/a n/a 18
Percent of household with 
access to basic drinking water 
services including water use 3.6% (0.93) 1002 3.5% (0.71) 1136 3.6% (0.62) 2,138
F&M 3.0% (0.87) 838 3.1% (0.69) 934 3.0% (0.54) 1,772
FNM 5.6% (2.22) 152 5.8% (1.80) 196 5.7% (1.64) 348
MNF n/a n/a 12 n/a n/a 6 n/a n/a 18
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
Regarding water treatment and sanitation, only 3.8% of households were observed to have 
handwashing facilities available on the premises (Table 28). Half of the households (50%) report treating 
water, and solar disinfection and flocculation are the most common methods. 
The majority of households (66%) practice open defecation, and less than 3% have household-level 
improved sanitation facilities. Among those not practicing open defecation, 84% use a pit latrine without 
a slab, which is considered unimproved, and 48% share the facility with other households. Table 28–
Table 31 summarize the water and sanitation variables by household type.
Table 28. Percentage of households with soap and water at a handwashing station on the premises
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Percent of households with 
handwashing available 3.5% (0.90) 1,929 4.1% (0.90) 1,863 3.8% (0.55) 3,792
F&M 3.3% (0.82) 1,605 4.2% (0.87) 1,536 3.7% (0.53) 3,141
FNM 4.5% (1.83) 301 3.3% (1.22) 316 3.9% (1.11) 617
MNF n/a n/a 23 n/a n/a 11 8.3% (6.06) 34
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
Table 29. Percentage of households in target areas practicing correct use of recommended household 
water treatment technologies
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households treating 
water 45.5% (4.21) 2159 53.4% (3.59) 2140 49.5% (2.37) 4,299
IMPEL | Implementer-Led Evaluation and Learning
24 Findings
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households 
with treated water by 
adding bleach or chlorine 
before drinking 42.1% (3.20) 1015 37.3% (3.14) 1142 39.5% (2.24) 2,157
Percent of households 
with treated water by 
filtration before drinking 10.8% (1.96) 2,159 13.7% (2.46) 2,140 12.2% (1.38) 4,299
Percent of households 
with treated water by 
solar disinfection 35.2% (3.59) 2,159 42.2% (3.30) 2,140 38.7% (1.92) 4,299
Percent of households 
with treated water by 
boiling before drinking 0.5% (0.22) 2,159 0.9% (0.27) 2,140 0.7 (0.16) 4,299
Table 30. Percentage of households in target areas practicing open defecation
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households practicing 
open defecation 63.3% (4.71) 2,159 68.0% (3.68) 2,140 65.7% (2.43) 4,299
F&M 62.8% (4.64) 1,807 68.4% (3.84) 1,769 65.6% (2.47) 3,576
FNM 65.5% (6.66) 325 66.5% (5.70) 359 66.0% (3.99) 684
MNF n/a n/a 27 n/a n/a 12 64.0% (8.64) 39
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
Table 31. Percentage of households with access to basic sanitation services 
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of households with access 
to basic sanitation facilities 2.8% (0.85) 2,159 2.4% (0.72) 2,140 2.6% (0.58) 4,299
F&M 3.1% (0.98) 1,807 2.8% (0.82) 1,769 2.9% (0.66) 3,576
FNM 0.9% (0.52) 325 0.3% (0.21) 359 0.6% (0.28) 684
MNF n/a n/a 27 n/a n/a 12 6.5% (4.70) 39
*Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
3.6 Agriculture
Crops
More than 94% of households farm, and 33% of household members 15 years old and over are classified 
as farmers. The targeted crops identified by the implementing partner were cassava, rice, and cloves. As 
shown in Table 32, cassava and rice are grown by 84% and 92% of farmers, respectively. Approximately 
one-quarter of farmers (28%) grow cash crop cloves. Other crops commonly grown in the area are 
coffee, citrus, sweet potatoes, and bananas.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 25
Table 32. Crops grown
Control Treatment All
Percentage of farmers growing.... Percent SE N Percent SE N Percent SE N
Cassava 84.8% (2.02) 2,008 83.0% (3.09) 1,934 83.9% (1.57) 3,942
Rice 92.6% (1.11) 2,008 91.6% (0.85) 1,934 92.1% (0.67) 3,942
Cloves 25.3% (3.47) 2,008 30.4% (3.96) 1,934 27.8% (2.04) 3,942
Coffee 39.1% (3.15) 2,008 36.1% (2.68) 1,934 37.6% (1.84) 3,942
Citrus 28.5% (3.71) 2,008 29.1% (3.51) 1,934 28.8% (2.12) 3,942
Mango 9.5% (1.74) 2,008 9.1% (2.13) 1,934 9.3% (1.31) 3,942
Sweet potatoes 40.5% (3.85) 2,008 37.0% (3.78) 1,934 38.8% (2.68) 3,942
Banana 38.8% (3.12) 2,008 35.4% (3.24) 1,934 37.1% (1.70) 3,942
Vanilla 13.8% (2.16) 2,008 14.3% (2.10) 1,934 14.1% (1.22) 3,942
The use of credit and savings is not common among farmers, and crop insurance is almost non-existent. 
Less than a third of farmers use at least one of the value chain participation activities. The most 
commonly used practices are purchase of agricultural inputs (14%), trade or sale of livestock/crops from 
your home/community with a buyer within your community (12%), and trade or sale of livestock/crops 
from primary regional markets (7%). Drying was practiced by fewer than 3% of farmers.11 None of the 
other practices listed, including the use of extension, contract farming, or processing produce, were 
used by more than 1% of farmers.12
Table 33. Percentage of farmers who used financial services (savings, agricultural credit, and/or 
agricultural insurance) in the past 12 months & percentage of farmers who practiced the value chain 
interventions promoted by the activity in the past 12 months
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of farmers using agricultural 
credit 5.4% (1.14) 2,103 5.1% (0.69) 2,003 5.2% (0.62) 4,106
Percent of farmers who saved 12.2% (1.51) 2,105 9.8% (0.98) 2,005 11.0% (0.90) 4,110
Percent of farmers using insurance 0.3% (0.15) 2,105 0.3% (0.16) 2,005 0.3% (0.09) 4,110
Percent of farmers reporting at least 
one value chain activity13 36.4% (6.40) 1,645 26.0% (4.58) 1,535 31.5% (3.09) 3,180
The agricultural module included a list of agricultural practices of interest to the implementing partner. 
Table 34–Table 36 summarize the most common practices used by farmers for cassava, rice, and cloves. 
 11 Drying refers to drying produce. Drying is a means of conservation, but this does not include processing the produce. 
12 The full list of value chain activities in the survey include: Purchase inputs for crops, purchase inputs for livestock, use of 
training and extension services, contract farming, drying produce, processing produce, trading or marketing produce through 
agro-dealers and/or community associations, use of formal marketing systems for livestock, use of formal marketing systems 
for crops, did not practice any of these activities in the past 12 months, trade or sale of livestock/crops from your 
home/community with a buyer from outside of your community, trade or sale of livestock/crops from your home/community 
with a buyer within your community, trade or sale of livestock/crops from a primary regional market. 13 The question on participation in value chain activities had the additional inclusion criteria that the farmer sell crops or 
livestock for income. This excluded 930 farmers.
IMPEL | Implementer-Led Evaluation and Learning
26 Findings
Only those practiced by at least 5% of farmers are shown.14 The most commonly used improved 
management practices/technologies for cassava were sowing after significant rain (5%) and inter￾planting (12%). For rice, the most common practices were sowing after significant rain (12%) and using 
organic manure (6%). For cloves, intercropping was practiced by 28% of farmers and the use of soil 
cover by 7% of farmers. However, 80%, 77%, and 62% of cassava, rice, and cloves farmers, respectively, 
used none of the listed practices. 
Table 34. Improved management practices/technologies for cassava
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of farmers using at least 
one practice for cassava 21.5% (2.06) 1,705 21.0% (2.56) 1,601 21.2% (1.45) 3,306
Percent of farmers interplanting for 
cassava 10.9% (1.55) 1,708 12.1% (1.89) 1,604 11.5% (0.97) 3,312
Percent of farmers sowing after 
significant rain for cassava 5.3% (1.37) 1,708 3.9% (1.27) 1,604 4.6% (0.81) 3,312
Percent of farmers that did not use 
modern practices one for cassava 79.7% (2.05) 1,708 79.4% (2.53) 1,604 79.6% (1.37) 3,312
Table 35. Improved management practices/technologies for rice
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of farmers using at least 
one practice for rice 24.0% (3.05) 1,847 22.9% (3.51) 1,770 23.5% (2.22) 3,617
Percent of farmers sowing after 
significant rain for rice 12.2% (2.96) 1,853 10.8% (3.35) 1,775 11.5% (2.03) 3,628
Percent of farmers that use 
organic manure for rice 5.4% (1.15) 1,853 6.8% (1.59) 1,775 6.1% (0.89) 3,628
Percent of farmers that did not 
use modern practices one for 
rice 76.9% (2.96) 1,853 77.5% (3.46) 1,775 77.2% (2.19) 3,628
Table 36. Improved management practices/technologies for cloves
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of farmers using at least one 
practice for cloves 42.5% (6.90) 528 36.0% (5.78) 576 39.0% (4.04) 1,104
Percent of farmers using inter￾planting for cloves 31% (6.99) 528 24.9% (4.30) 576 27.7% (3.83) 1,104
Percent of farmers using soil cover 
for cloves 5.9% (2.14) 528 7.3% (2.70) 576 6.7% (1.68) 1,104
Percent of farmers that did not use 
modern practices one for cloves 59.3% (7.08) 528 64.2% (5.76) 576 61.9% (4.09) 1,104
 14 Only those practiced by at least 5% of farmers are shown. The full list of practices included in the survey for all three crops is: 
Organic manure, compost, performing weeding, sowing after useful rain, crop association, crop rotation, use of improved 
seeds, use of climate information (rain forecast, disaster risks, etc.), windbreak, and soil cover.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 27
The yield estimates for the three crops are presented in Table 37. Because the yield calculation is based 
on farmers’ recall of production and farmers’ estimates of the area planted, these numbers should be 
taken as rough estimates. The survey first asks farmers about the number of plots, the area of each plot, 
and how much of each plot was devoted to each of the three crops. Next, farmers are asked to recall the
total production of each crop across all plots in the previous year. While most farmers measure land in 
Ares (1/100 of a hectare), many farmers measure production in volume, not weight, and this must be 
converted to kilograms.
The average yield for cassava is slightly over 5,000 kg/ha. Rice yield is approximately 3,900 kg/ha. Clove 
yield is estimated to be 670 kg/ha, although clove yield is often measured per bush and not per hectare,
and the area covered by cloves may have been difficult to estimate. Figure 5 shows the distribution of 
yields. Despite trimming extreme values, some of the yield values, particularly for rice, are likely 
overestimated.
Figure 5. Box plot of yield estimates for target crops
0 5,000 10,000 15,000 20,000
Yield (kg per ha)
Cassava Cloves
Rice
IMPEL | Implementer-Led Evaluation and Learning
28 Findings
Table 37. Yield of targeted agricultural commodities within target areas
Control Treatment All
Outcome Mean Median SE N Mean Median SE N Mean Median SE N
Producers 
of cassava
reporting
yield in kg 
per ha* 5333.1 3000 -365.08 1389 4805.7 3000 -421.92 1315 5074.29 3000 -272.12 2704
Producers 
of cloves 
reporting
yield in kg 
per ha 680.78 231.11 -134.96 160 665.04 220 -110.15 165 672.17 222.22 -74.69 325
Producers 
of rice 
reporting
yield in kg 
per ha* 4315.18 2400 -358.79 1606 3459.56 2333.33 -282.73 1554 3897.38 2359.13 -247.35 3160
*Kilogram (kg)
*Hectare (ha)
Livestock
The targeted livestock are poultry, bees, and fish. The summary statistics for poultry farmers are 
presented in Table 38. In our sample, 67% of farmers raised poultry, and the average number of birds 
raised per farmer in the last year was 28. Only 27% of farmers reported any egg production from 
chickens in the previous week. Farmers tend to both sell poultry (61%) and consume it (80%). Over two￾thirds of farmers (67%) reported poultry dying in the last year. Vaccinations were the only practice used 
by more than 3% of poultry farmers, and most farmers (94%) did not use any of the listed practices. 
Poultry yield, defined as the total kilograms consumed or sold divided by the number of birds kept, is 
estimated to be 2.4 kg/bird.
Table 38. Poultry farming
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of farmers raising poultry 67.5% (2.69) 2,141 65.7% (2.52) 2,046 66.6% (1.64) 4,187
Percent of farmers using 
vaccinations 3.0% (0.77) 1,433 5.1% (1.40) 1,340 4.0% (0.72) 2,773
Percent of farmers not using any 
of the poultry practices 93.9% (1.13) 1,433 92.4% (1.56) 1,340 93.2% (0.73) 2,773
Percent of farmers reporting eggs 
in previous week 28.0% (2.66) 1,420 25.5% (2.08) 1,322 26.8% (1.69) 2,742
Percent of farmers selling poultry 
in the last year 59.8% (2.42) 1,427 63.2% (2.30) 1,327 61.5% (1.38) 2,754
Percent of farmers consuming 
own poultry in the last year 82.8% (1.82) 1,426 77.9% (2.03) 1,327 80.4% (1.33) 2,753
Percent of farmers experiencing 
poultry dying in the last year 67.3% (2.43) 1,427 66.3% (2.51) 1,327 66.8% (1.81) 2,754
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 29
Only 68 individuals in the survey raised bees. Half raised them for both food and the market. As found in
Table 39, the average beekeeper had four hives. Only 15 farmers reported having modern hives;
therefore, the yield was excluded in this report. However, traditional hives, which are more common, 
yielded less than 2 liters of honey. 
Table 39. Beekeeping15
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Total number of hives owned by 
beekeepers 4.66 (1.35) 32 4.03 (0.42) 37 4.25 (0.54) 69
Number of modern hives owned 
by beekeepers 2.39 (1.55) 31 0.31 (0.22) 38 1.02 (0.73) 69
Number of traditional hives
owned by beekeepers 1.93 (0.35) 31 3.68 (0.27) 38 3.08 (0.37) 69
Liters of honey produced per year-
-traditional hives 1.78 (0.49) 27 1.99 (0.73) 33 1.92 (0.53) 60
Percentage using beekeeping equipment (listed below)
Percent of beekeepers using 
beehive 87.2% (9.81) 32 92.8% (4.32) 37 90.8% (4.73) 69
Percent of beekeepers using 
smokers 49.0% (13.11) 32 16.7% (6.49) 37 28.1% (8.41) 69
Percent of beekeepers using 
frames and foundations 18.4% (9.03) 32 7.3% (5.19) 37 11.2% (5.22) 69
Percent of beekeepers using veil 
and gloves 13.6% (5.43) 32 9.8% (6.94) 37 11.1% (4.85) 69
Percent of beekeepers using 
feeders 3.8% (3.53) 32 0.0% (.) 37 1.4% (1.56) 69
Percent of beekeepers using 
heated knifes 29.8% (11.54) 32 22.4% (4.15) 37 25.0% (4.68) 69
Percent of beekeepers using 
uncapping forks 6.9% (4.40) 32 7.3% (5.19) 37 7.1% (3.47) 69
Percent of beekeepers using tub 
for wax/honey 3.8% (3.53) 32 1.5% (1.56) 37 2.3% (1.95) 69
Percent of beekeepers using 
extractors 0.0% (.) 32 1.5% (1.56) 37 1.0% (0.96) 69
Percent of beekeepers using food￾grade buckets 6.2% (4.00) 32 9.9% (6.93) 37 8.6% (4.38) 69
Percent of beekeepers using 
double sieves 9.2% (5.13) 32 3.0% (3.13) 37 5.2% (3.26) 69
Percent of beekeepers using 
containers 3.8% (3.53) 32 3.6% (3.03) 37 3.7% (2.60) 69
Percent of beekeepers using 
queen excluders 2.1% (2.11) 32 3.0% (3.13) 37 2.7% (2.10) 69
Percent of beekeepers using 
queen cages 5.9% (4.02) 32 6.9% (4.27) 37 6.6% (2.72) 69
Percent of beekeepers using 
beekeeper's suits 3.8% (3.53) 32 3.0% (3.13) 37 3.3% (2.64) 69
 15 The sample size is small so these results may not be representative and should be interpreted with caution
IMPEL | Implementer-Led Evaluation and Learning
30 Findings
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Percent of beekeepers using nylon 
brush 5.9% (4.02) 32 1.5% (1.56) 37 3.1% (2.20) 69
Percent of beekeepers using 
attracts swarms 8.2% (4.53) 32 5.8% (3.88) 37 6.6% (2.77) 69
Percent of beekeepers using filter 
cloths 11.3% (5.20) 32 22.5% (6.25) 37 18.6% (5.53) 69
Percent of beekeepers using 
knives 32.7% (13.20) 32 43.2% (10.76) 37 39.5% (8.87) 69
Percent of beekeepers using 
buckets 12.0% (6.29) 32 24.8% (6.04) 37 20.2% (5.53) 69
In the sample, 287 individuals report fishing. Roughly half of fishermen fish for both home consumption 
and market, 46% of those fishing fish for home consumption only, and fishing for the market only is not 
common.
Table 40. Fishing
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Percent of fishers fishing for 
food only 56.5% (8.85) 160 33.9% (8.27) 127 45.5% (5.93) 287
Percent of fishers fishing for 
market only 1.2% (0.72) 160 5.0% (2.58) 127 3.03% (1.37) 287
Percent or fishers fishing for 
both food and market 42.4% (8.89) 160 61.1% (7.78) 127 51.5% (5.53) 287
Percent of fishers using a 
pirogue 29.7% (7.36) 159 27.3% (7.56) 127 28.6% (5.01) 286
Percent of fishers using nets 43.3% (7.87) 159 45.5% (6.79) 127 44.4% (5.17) 286
Percent of fishers using 
containers 33.6% (8.12) 159 30.6% (7.05) 127 32.1% (4.42) 286
3.7 Poverty Measurement
This section presents three measures of poverty, all based on household expenditure. The measures are 
per capita daily expenditure, the percentage living on less than $1.90 per day (2011 purchasing power 
parity (PPP)), and the depth of poverty of the poor. The equivalent of $1.90 in current Malagasy Ariary 
was determined to be 2,443.16 17
The poverty measures are summarized in Table 41. The mean per capita expenditure is 2,070 Ariary, or 
about $1.61 per day. The poverty rate is approximately 72%. The depth of poverty of the poor is 33%, 
which means that the average poor person is 33% below the poverty line. In monetary terms, this 
 16 The 2011 PPP used was 700.228 Ariary/dollar. The Consumer Price Index used for 2011 was 72.11 and 132.5 for 2020. This 
gives us the poverty line = 1.9 * 700.228 * 132.5 / 72.18 = 2243 17 The expenditure module of the survey followed standard practices for expenditure calculation. Frequent items, primarily 
foods, used 7-day recall. Less frequent purchases used a 30-day or 12-month recall. A rental equivalent is used to value housing 
and durable goods. Enumerators reported a few extreme cases of people with essentially zero expenditures who lived on what 
they could collect, forage, or receive from begging. 
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 31
means it would require $0.63 per person per day to bring every poor person out of poverty. Figure 6
shows the distribution of expenditure values across the full sample. The poverty line is shown in dark 
blue.
Table 41. Poverty measures
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Prevalence of Poverty: 
Percentage of people living on 
less than $1.90/day 69.5% (1.19) 2,160 74.3% (1.18) 2,140 71.9% (0.84) 4,300
Daily per capita expenditure 
(Ariary) 2,099.07 (23.37) 2,160 2040.82 (22.50) 2,140 2,069.51 (16.20) 4,300
Daily per capita expenditure 
(Dollars) 1.63 (0.02) 2,160 1.59 (0.02) 2,140 1.61 (0.01) 4,300
Depth of Poverty of the Poor: 
Mean percent shortfall of the 
poor relative to the $1.90/day 
2011 PPP poverty line 32.81 (0.00) 1,404 32.41 (0.00) 1,512 32.6 (0.00) 2,916
Figure 6. Distribution of expenditure
These measures are further broken down by household type (Table 42). The types are F&M, FNM, MNF,
and child no adult (CNA). In the present case, there were no CNA households. There are no large 
IMPEL | Implementer-Led Evaluation and Learning
32 Findings
differences between F&M and FNM households. There are very few MNF households, and therefore it is 
difficult to draw conclusions about this group.
Table 42. Disaggregated poverty measures by household types
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Prevalence of 
Poverty: Percentage
of people living on 
less than $1.90/day 69.5% (1.19) 2,160 74.3% (1.18) 2,140 71.9% (0.84) 4,300
F&M 69.7% (1.29) 1,808 74.1% (1.28) 1,769 71.9% (0.91) 3,577
FNM 68.6% (3.14) 325 75.3% (2.77) 359 72.1% (2.10) 684
MNF n/a n/a 27 n/a n/a 12 68.3% (8.69) 39
Daily per capita 
expenditure (Ariary) 2,099.07 (23.37) 2,160 2,040.82 (22.50) 2,140 2,069.51 (16.20) 4,300
F&M 2,087.56 (24.82) 1,808 2,025.98 (24.03) 1,769 2,056.39 (17.26) 3,577
FNM 2,184.62 (71.94) 325 2,151.65 (64.46) 359 2,167.27 (48.04) 684
MNF n/a n/a 27 n/a n/a 12 2,387.16 (193.13) 39
Daily per capita 
expenditure 
(Dollars) 1.63 (0.02) 2,160 1.59 (0.02) 2,140 1.61 (0.01) 4,300
F&M 1.62 (0.02) 1,808 1.58 (0.02) 1,769 1.6 (0.01) 3,577
FNM 1.7 (0.06) 325 1.67 (0.05) 359 1.69 (0.04) 684
MNF n/a n/a 27 n/a n/a 12 1.86 (0.15) 39
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
3.8 Gender Dynamics
Gender dynamics are captured through the six indicators in this section. Because the survey targeted 
only men and women in a union, they are the only ones included in this section. These indicators 
explore male and female financial resources and access to credit.
3.8.1 Use of Financial Resources
This section presents findings on participation in cash earning activities across men and women in a 
union. Table 43 illustrates that, of the number of women and men in a union, 74% participated in cash 
earning activities in the past year. According to respondents, it was more common for men in a union to 
be cash earners (89%) than for women in a union (59%) (p-value = 0.00).
Table 43. Percentage of women and men in a union who earned cash in the past 12 months
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Cash earners in a union 74.5% (0.024) 3,220 73.9% (0.016) 3,164 74.2% (0.011) 6,384
Cash earning women in a 
union 58.6% (0.045) 1,560 58.7% (0.034) 1,562 58.6% (0.019) 3,122
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 33
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Women in a union 
(ages 15–19) 51.6% (0.065) 120 55.9% (0.051) 110 53.7% (0.036) 230
Women in a union 
(ages 20–29) 62.2% (0.050) 455 55.5% (0.044) 429 59.0% (0.027) 884
Women in a union 
(ages 30–49) 62.1% (0.041) 682 62.9% (0.034) 696 62.5% (0.019) 1,378
Cash earning men in a union 89.6% (0.012) 1,660 88.7% (0.014) 1,602 89.2% (0.008) 3,262
Men in a union (ages 
15–19) n/a n/a 20 n/a n/a 29 74.1% (0.074) 49
Men in a union (ages 
20–29) 84.2% (0.022) 329 88.8% (0.021) 270 86.3% (0.016) 599
Men in a union (ages
30+) 91.2% (0.013) 1,311 89.0% (0.016) 1,303 90.1% (0.009) 2,614
Table 44 shows the percentage of cash-earning women in a union who report participation in decisions 
about the use of self-earned cash. The large majority of women in a union (84%) reported that they 
participate in decisions about how to use their cash, whether solely or jointly, with others.
Table 44. Percentage of women in a union and earning cash who report participation in decisions 
about the use of self-earned cash
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Cash earning women in a union 86.6% (0.031) 934 82.1% (0.041) 914 84.4% (0.016) 1,848
Women in a union (ages 
15–19) 86.7% (0.064) 67 63.1% (0.111) 57 74.8% (0.063) 124
Women in a union (ages 
20–29) 87.1% (0.034) 286 83.8% (0.043) 242 85.6% (0.024) 528
Women in a union (ages 
30–49) 84.0% (0.039) 432 82.0% (0.044) 440 83.0% (0.018) 872
Table 45 highlights the percentage of cash-earning men in a union who report that their spouse 
participates in decision-making about the use of self-earned cash. Approximately 82% of men perceive 
their wives to be involved in decision-making. There is little variation across age groups.18
Table 45. Percentage of men in union and earning cash who report spouse/partner participation in 
decisions about the use of self-earned cash
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Cash earning men in union 82.9% (0.024) 1,479 80.8% (0.025) 1,426 81.9% (0.015) 2,905
Men in a union (ages 
15–19) n/a n/a 15 n/a n/a 20 82.7% (0.078) 35
 18 The sample for baseline 34 indicator (Percentage of women in a union and earning cash who report spouse/partner 
participation in decisions about the use of self-earned cash) was not going to be reflective of all cash earning women in a union,
so it was left out of the reporting.
IMPEL | Implementer-Led Evaluation and Learning
34 Findings
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Men in a union (ages 
20–29) 81.7% (0.034) 280 81.3% (0.036) 240 81.5% (0.021) 520
Men in a union (ages
30+) 83.1% (0.024) 1,184 80.8% (0.027) 1,166 82.0% (0.016) 2,350
3.8.2 Credit
This section presents findings on the use of credit among men and women in a union. Table 46 shows 
that over a third of men and women in a union (36%) borrowed in the previous 12 months. Women and 
men in a union report borrowing at similar rates (36% and 35%). 
Table 46. Percentage of women/men in a union with access to credit
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Men and women in a union 35.6% (0.030) 3,220 35.9% (0.029) 3,164 35.8% (0.017) 6,384
Men in a union 36.0% (0.030) 1,660 36.5% (0.027) 1,602 36.2% (0.016) 3,262
Men in a union (ages 
15–19) n/a n/a 20 n/a n/a 29 24.1% (0.075) 49
Men in a union (ages 
20–29) 39.1% (0.041) 329 39.8% (0.043) 270 39.4% (0.026) 599
Men in a union (ages
30+) 35.4% (0.030) 1,311 36.0% (0.027) 1,303 35.7% (0.017) 2,614
Women in a union 35.2% (0.030) 1,560 35.4% (0.032) 1,562 35.3% (0.018) 3,122
Women in a union 
(ages 15–19) 26.2% (0.062) 120 33.1% (0.065) 110 29.5% (0.045) 230
Women in a union
(ages 20–29) 41.4% (0.043) 455 36.7% (0.039) 429 39.1% (0.025) 884
Women in a union
(ages 30–49) 34.0% (0.031) 682 37.3% (0.036) 696 35.7% (0.022) 1,378
Of men and women in a union who report borrowing, a majority of them (69%) participate in decisions 
about credit. These are decisions made by the respondents alone and with a spouse/partner or another 
individual. As illustrated in Table 47, men in a union participate in decisions about credit at a much 
higher rate than women in a union (p-value = 0.00), with 87% of men in a union reportedly making 
credit decisions. 
Table 47. Percentage of women/men in a union who make decisions alone or jointly about credit
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Men and women in a union 69.9% (0.024) 1,165 67.3% (0.020) 1,115 68.6% (0.014) 2,280
Men in a union 88.9% (0.021) 615 86.7% (0.029) 585 87.8% (0.016) 1,200
Men in a union (ages 15–
19) n/a n/a 5 n/a n/a 8 n/a n/a 13
Men in a union (ages 20–
29) 89.8% (0.029) 138 91.3% (0.032) 109 90.5% (0.024) 247
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 35
Control Treatment All
Outcome Percent SE N Percent SE N Percent SE N
Men in a union (ages 
30+) 88.7% (0.022) 472 85.6% (0.037) 468 87.1% (0.019) 940
Women in a union 49.4% (0.046) 550 46.7% (0.065) 530 48.0% (0.035) 1,080
Women in a union (ages 
15–19) 40.0% (0.106) 36 38.4% (0.116) 32 39.1% (0.082) 68
Women in a union (ages 
20–29) 47.4% (0.051) 189 34.6% (0.057) 162 41.6% (0.037) 351
Women in a union (ages 
30–49) 53.5% (0.058) 239 52.8% (0.062) 243 53.1% (0.037) 482
3.9 Resilience
3.9.1 Ability to Recover from Shocks and Stresses Index
The ability to recover from shocks and stresses index reflects the ability to recover from negative events 
that have impacted the household. On average, households score 2.6 on this index. This index is 
composed of indices that reflect how households perceived their recovery as well as the total number 
and severity of shocks the household experienced over the past year. Table 48 below illustrates that 
there was little to no observable difference across treatment and control household responses (see 
Annex A for tests of significance). There is also little variation in the perceived ability to recover across 
the different household types (p-value = 0.30).
Table 48. Ability to recover from shocks and stresses index
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household ability 
to recover from 
shocks & stresses 
index
2.586 (0.062) 1,849 2.606 (0.045) 1,862 2.596 (0.033) 3,711
F&M 2.6 (0.062) 1,553 2.604 (0.047) 1,542 2.602 (0.035) 3,095
FNM 2.52 (0.099) 273 2.615 (0.068) 310 2.571 (0.050) 583
MNF n/a n/a 23 n/a n/a 10 2.441 (0.103) 33
Household ability 
to recover index 
(2–6)
2.594 (0.065) 1,884 2.612 (0.052) 1,881 2.603 (0.039) 3,765
F&M 2.598 (0.068) 1,576 2.601 (0.056) 1,560 2.599 (0.042) 3,136
FNM 2.586 (0.092) 283 2.671 (0.066) 311 2.63 (0.051) 594
MNF n/a n/a 25 n/a n/a 10 2.395 (0.111) 35
IMPEL | Implementer-Led Evaluation and Learning
36 Findings
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household total 
shocks 
experiences (0-
22)
2.292 (0.085) 2,160 2.283 (0.091) 2,140 2.287 (0.053) 4,300
F&M 2.357 (0.087) 1,808 2.355 (0.098) 1,769 2.356 (0.057) 3,577
FNM 1.959 (0.111) 325 1.937 (0.080) 359 1.948 (0.065) 684
MNF n/a n/a 27 n/a n/a 12 2.074 (0.385) 39
Household shock 
exposure index 
(0-176)
13.664 (0.586) 2,074 13.217 (0.571) 2,077 13.439 (0.364) 4,151
F&M 14.041 (0.584) 1,745 13.616 (0.597) 1,717 13.828 (0.377) 3,462
FNM 11.442 (0.821) 305 11.191 (0.631) 350 11.311 (0.479) 655
MNF n/a n/a 24 n/a n/a 10 16.829 (2.620) 34
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
The perceived ability to recover index reflects a household’s perspective on their ability to meet food 
needs now, relative to the previous year, as well as their ability to meet their future food needs. 
Households score an average of 2.6 (range of 2–6). In other words, households perceive their ability to 
meet their current needs as worse than the previous year and suspect that their ability to meet these 
needs will deteriorate in the future. There is variation in how different household types perceive their 
ability to recover (p-value = 0.00), with households with no female adult present reporting a lower 
ability to recover and households with no male adult present reporting a higher ability.
Households experienced an average of 2.3 out of 22 possible shocks in the previous year. The most 
commonly reported shocks are illustrated below in Figure 7, which shows the shocks experienced by at 
least 10% of households. A large majority of households (80%) listed the drought as one of these shocks. 
Other common shocks listed included rising food prices, livestock diseases, and crop pests.
In terms of the number and intensity of shocks experienced, households scored an average of 13.4 out 
of 176, which accounts for 22 shocks and four different levels of severity (regarding both the impact on 
the household economic situation and impact on household consumption). This suggests that out of the 
average of 2.3 shocks experienced, households perceived those shocks to be severe. The shock exposure 
index varies across household types (p-value = 0.00), with households with no female adult present 
reporting a greater number and intensity of shocks and households with no male adult present 
reporting the lowest.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 37
Figure 7. Most common shocks reported by households
3.9.2 Social Capital Indices
These indices convey the ability of households to draw on social networks to get support to reduce the 
impact of shocks and stresses on their households.19 They measure both the degree of bonding—social 
capital among households within their own communities—and the degree of bridging—social capital 
between households in the area and households outside their own community. Findings suggest that 
households are able to draw on other households within their community slightly more (average score 
of 54.4) than they are able to draw on other households outside of their community (average score of 
48.2). There is variation across household types in both the bonding subindex (p-value = 0.00) and the 
bridging subindex (p-value = 0.05). In both cases, households with both male and female adults present 
are more able to draw on community members for support than households with one or the other adult 
not present.
Table 49. Social capital indices
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household bonding 
subindex (0–100) 52.773 (1.892) 2,160 56.001 (2.187) 2,140 54.391 (1.093) 4,300
F&M 53.442 (2.028) 1,808 56.605 (2.166) 1,769 55.023 (1.095) 3,577
FNM 49.724 (2.451) 325 53.627 (2.810) 359 51.732 (1.805) 684
 19 Two questions that are included in the social capital index were accidentally deleted from the FIOVANA survey. Specifically, 
these questions were “who could households turn to inside (and outside) of the village if they needed help urgently.” Due to 
this, we do not create the full social capital index but only report two of the subindices that form the full index. Here the 
bonding index focuses only on whom households would give help to within their communities, and the bridging index focuses 
only on whom households would give help to outside of their communities. 
IMPEL | Implementer-Led Evaluation and Learning
38 Findings
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
MNF n/a n/a 27 n/a n/a 12 43.644 (7.412) 39
Household bridging 
subindex (0–100) 46.348 (2.113) 2,160 49.975 (2.371) 2,140 48.165 (1.271) 4,300
F&M 46.682 (2.270) 1,808 50.392 (2.348) 1,769 48.537 (1.334) 3,577
FNM 45.488 (2.554) 325 47.952 (3.319) 359 46.756 (1.887) 684
MNF n/a n/a 27 n/a n/a 12 38.943 (7.014) 39
* Gendered Household Type: Female and Male Adults (F&M), Adult Female no Adult Male (FNM), Adult Male no 
Adult Female (MNF)
3.9.3 Absorptive Capacity Index
The absorptive capacity index reflects the ability of households to prepare for, deal with, and mitigate 
the impact of shocks and stressors on well-being outcomes through preventive measures and positive 
coping strategies. Overall, households score 30.3 out of 100 on this index, reflecting a low ability to 
absorb shocks.20 As discussed below, there are few resources that households have access to that would 
enable them to better mitigate shocks. 
A key aspect of the absorptive capacity index captures financial resources that households have access 
to in order to absorb shocks. Overall, very few households have access to financial resources for 
absorbing shocks, lowering their ability to mitigate the impact of shock on well-being outcomes. 
Specifically, only 12% of households have cash savings, 1% have reported receiving remittances, and less
than 1% have access to insurance. Households on average own 6.4 different types of household and 
productive assets (out of 43), although this does not necessarily mean that households have a large 
asset stock.
Table 50. Absorptive capacity index
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household absorptive capacity 
index (0–100) 27.103 (1.456) 1,904 33.376 (1.834) 1,952 30.264 (1.108) 3,856
Household access to cash savings 
index (0–1) 0.118 (0.017) 2,159 0.115 (0.015) 2,140 0.117 (0.009) 4,299
Household asset ownership index 
- total type (0–31) 6.416 (0.187) 2,102 6.398 (0.147) 2,089 6.407 (0.109) 4,191
Household remittances index 
(0–1) 0.01 (0.003) 2,160 0.012 (0.003) 2,140 0.011 (0.002) 4,300
Household access to insurance 
index (0–1) 0.003 (0.002) 1,990 0.003 (0.002) 1,966 0.003 (0.001) 3,956
Household bonding social capital 
index (0–6) 1.112 (0.044) 2,160 1.199 (0.055) 2,140 1.156 (0.026) 4,300
 20 The three resilience indices were calculated by taking the first principal component of all of the sub-indices included in those 
tables. We then predicted the score for each household based on the weighted combination of the first component and the 
subindices. This score was rescaled to be between 0 and 100.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 39
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household access to informal 
safety nets index (0–6) 2.396 (0.169) 2,080 2.779 (0.136) 2,140 2.59 (0.100) 4,220
Household shock preparedness & 
responsiveness index (0–3) 0.922 (0.055) 2,160 1.098 (0.071) 2,140 1.01 (0.044) 4,300
Household access to humanitarian 
assistance index (0–1) 0.089 (0.026) 2,160 0.277 (0.037) 2,140 0.183 (0.025) 4,300
The absorptive capacity index also captures the level of social capital that households have access to in 
order to help them absorb shocks. Overall, findings suggest that most households have low social 
capital. The bonding social capital index reflects the number of types of individuals that households 
could draw on inside of their communities (out of three groups). On average, households feel able to 
draw on 1.2 of these types of individuals. Moreover, households reported that they have moderate 
access to and have been active in community organizations that typically serve as informal safety nets. 
On average, households have access to 2.5 of six types of safety nets.
Another element of absorptive capacity is how well a household is prepared to mitigate shocks21
through the availability of disaster preparedness groups in the community, as well as other household 
shock mitigation strategies. On average, households score a 1.01 out of 3 on this index, suggesting a 
lower ability to mitigate shocks.
The last dimension of absorptive capacity is the availability of humanitarian assistance in the 
community. Less than one-fifth of households (18%) reported that they had received emergency food or 
cash assistance from the government or from a non-governmental organization (NGO).
22
3.9.4 Adaptive Capacity Index
The adaptive capacity index measures the ability of households to manage resources and make 
proactive and informed choices to better prepare for and adapt to future shocks. The index is composed 
of several components that reflect different resources or adaptive abilities. On average, households 
score 31.8 out of 100 on this index,
23 which suggests that households have a limited ability to manage 
resources and adapt to future shocks. Households across the treatment and control groups perform 
similarly on this score. 
 21 This index does not include whether the household reports participating in any of the following activities: soil conservation
activities, flood diversion structures (i.e., protection of land/infrastructure from flooding), planting trees on communal land, or 
improving access to health services given available data.
22 Note that this index does not capture whether non-governmental organization/government assistance is available in their 
community, but they have not received it.
23 We followed precedent with TANGO and did not exclude factors that loaded negatively on the first component in the 
construction of this index. This is because we want the indices to be comparable across baseline and endline (where at endline,
the factors that load negatively may be different). However, the two scores are comparable on average.
IMPEL | Implementer-Led Evaluation and Learning
40 Findings
Table 51. Adaptive capacity index
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household adaptive 
capacity index (0–100) 29.73 (1.218) 1,886 33.966 (1.072) 1,863 31.846 (0.691) 3,749
Household bridging 
social capital index (0–6) 0.978 (0.049) 2,160 1.063 (0.060) 2,140 1.02 (0.030) 4,300
Household linking social 
capital (0–4) 0.041 (0.010) 2,160 0.029 (0.005) 2,140 0.035 (0.006) 4,300
Household social 
network index (0–6) 1.462 (0.094) 2,150 1.959 (0.092) 2,140 1.711 (0.061) 4,290
Household 
education/training index 
(0–8) 1.112 (0.053) 2,159 1.245 (0.056) 2,135 1.179 (0.038) 4,294
Household asset 
ownership index - total 
type (0–31) 6.416 (0.187) 2,102 6.398 (0.147) 2,089 6.407 (0.109) 4,191
Household access to 
financial resources (0–2) 0.403 (0.114) 2,160 0.787 (0.063) 2,140 0.595 (0.062) 4,300
Household livelihood 
diversification index (0–
20) 2.716 (0.099) 2,160 2.799 (0.105) 2,140 2.757 (0.059) 4,300
Household adoption of 
improved practices index 
(0–1) 0.07 (0.013) 1,990 0.059 (0.011) 1,966 0.064 (0.008) 3,956
Household exposure to 
information index (0–19) 4.215 (0.343) 2,160 4.248 (0.291) 2,140 4.232 (0.176) 4,300
Household 
aspirations/confidence 
to adapt index (0–16) 9.205 (0.100) 2,056 9.198 (0.122) 2,039 9.202 (0.059) 4,095
One aspect of the adaptive capacity index captures households’ social capital and networks, given that 
households that are able to leverage these networks more effectively may better prepare for and adapt 
to future shocks. Across these indicators, households score poorly, suggesting that households are not 
able to effectively leverage social capital and networks to adapt to shocks. In particular, the bridging 
social capital index reflects the number of types of individuals that households could draw on outside of 
their communities (out of three groups). On average, households feel able to draw on 1.0 of these types 
of individuals. The linking social capital index reflects how well-connected households are to 
government or NGO leaders and whether they can draw on them for help. Households score very low 
(0.04 out of 4) on this, suggesting that the majority of households neither know leaders nor are they 
able to ask leaders for help. Finally, the social network index captures household access to and 
participation in various support groups. Households score a 1.7 out of 6, suggesting that only a minority 
of households have access to and/or participate in these groups.
Another aspect of the adaptive capacity index captures the human resources, assets, and financial 
resources available to households to mitigate shocks. Overall, households have low levels of human 
capital and asset resources, suggesting constraints on the overall resource pool they are able to draw on 
in the face of shocks. The education/training index reflects the level of human capital in the household, 
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 41
specifically adult literacy, whether any adult has surpassed primary school, and the number of trainings 
in which household adults have participated. Households score low (1.8 out of 8) on this indicator, 
reflecting that overall household human capital is low. The asset ownership index illustrates the number 
of different types of assets a household owns (out of 43 types). On average, a household owns 6.4 
different types of assets. This could mean that overall household asset stock is low, although this does 
not reflect the value of each asset. Finally, the access to financial resources index reflects the financial 
resources available in the village through credit and savings institutions. Households have, on average, 
0.6 out of 2 of these institutions available to them.
A third aspect of the adaptive capacity index reflects how diversified and improved household livelihood 
activities are. In summary, household activities are not well-diversified, and few households have 
adopted improved practices. The livelihood diversification index reflects the number of different 
livelihood activities that households were engaged in over the past year. Overall, households were 
engaged in an average of 2.6 out of 20 activities, indicating that activities are not well-diversified. The 
adoption of improved practices index24 reflects whether households adopted improved crop or livestock 
practices, natural resource management practices, or improved storage practices. Overall, households 
score 0.06 out of 1, indicating that only a minority of households have adopted improved practices. The 
exposure to information index captures the number of topics that households have received 
information on in the past year, which relates directly to a household’s ability to make informed choices 
in order to better prepare for shocks. On average, households have received information on 4.2 out of 
19 available topics, highlighting that households have had limited exposure to information to help 
inform shock mitigation strategies.
Finally, the aspirations/confidence to adapt index reflects a household adult’s aspirations, confidence to 
adapt, and a sense of control over one’s life. On average, adults score 9.2 out of 16 on this index, 
reflecting a moderate sense of confidence to adapt.
3.9.5 Transformative Capacity Index
The transformative capacity index25 captures system-level resources, governance, and institutions that 
make up the enabling environment that promotes or limits a household’s capacity to respond to shocks 
and stressors. On average, households score 35.7 out of 100 on this index,26 indicating that there are not 
very strong institutions available to enhance household capacity to respond to shocks. Households 
across the treatment and control groups perform similarly on this score. 
 24 This index does not include a measure of quality for each of the service types. Instead, it only captures whether or not the 
service exists. Health services reflect whether NGOs are currently conducting health activities and not whether local health 
institutions are available.
25 Note that this index does not include the following sub-indices given data availability: access to livestock services and 
collective action.
26 We followed precedent with TANGO and did not exclude factors that loaded negatively on the first component in the 
construction of this index. This is because we want the indices to be comparable across baseline and endline (where at endline
the factors that load negatively may be different). For the case of the transformative index, the index that adjusts for negative 
factors is 35.11 on average across the sample. The main driver between these scores is that the infrastructure factor gets more 
weight in the score we show here.
IMPEL | Implementer-Led Evaluation and Learning
42 Findings
Table 52. Transformative capacity index
Control Treatment All
Outcome Mean SE N Mean SE N Mean SE N
Household transformative 
capacity index (0–100) 35.305 (2.474) 1,333 36.007 (2.655) 1,782 35.734 (1.747) 3,115
Household access to 
infrastructure index (0–4) 1.004 (0.104) 2,160 0.949 (0.107) 2,140 0.976 (0.068) 4,300
Household basic services 
index (0–4) 1.375 (0.082) 2,160 1.92 (0.074) 2,140 1.648 (0.061) 4,300
Household access to markets 
index (0–1) 1 (.) 2,160 1 (.) 2,140 1 (.) 4,300
Household access to 
communal natural resources 
index (0–4) 0.237 (0.070) 2,160 0.156 (0.065) 2,140 0.197 (0.036) 4,300
Household access to 
agricultural services index 
(0–1) 0.044 (0.032) 2,160 0.146 (0.065) 2,140 0.095 (0.038) 4,300
Household bridging social 
capital index (0–6) 0.978 (0.049) 2,160 1.063 (0.060) 2,140 1.02 (0.030) 4,300
Household linking social 
capital index (0-4) 0.041 (0.010) 2,160 0.029 (0.005) 2,140 0.035 (0.006) 4,300
Household social cohesion 
index (0–3) 0.169 (0.059) 2,160 0.16 (0.066) 2,140 0.165 (0.033) 4,300
Household local decision￾making index (0–1) 0.34 (0.027) 2,148 0.359 (0.025) 2,122 0.349 (0.015) 4,270
Household gender index (0–
3) 2.368 (0.080) 2,160 2.118 (0.047) 2,140 2.243 (0.053) 4,300
Household gender equitable 
decision-making index (0–2) 0.704 (0.036) 2,160 0.665 (0.048) 2,140 0.684 (0.018) 4,300
Household access to formal 
safety nets index (0–13) 1.787 (0.165) 1,339 1.562 (0.136) 1,800 1.649 (0.097) 3,139
Household local government 
responsiveness index (0–2) 0.871 (0.050) 2,160 0.951 (0.042) 2,140 0.911 (0.025) 4,300
One dimension of the transformative capacity index is the accessibility of infrastructure and services in 
the community. Communities have access to some basic services but even fewer key infrastructure 
types. The basic services index27 illustrates the number of services (police force, primary schools, health, 
and financial services) that are available in the community. Households have access to 1.6 out of four of 
these services on average. The access to infrastructure index reflects how many types of key 
infrastructure (electricity grid, piped water, mobile phone service, and roads) are available in the 
community. On average, households have access to 0.98 out of four of these types of key infrastructure. 
The next dimension of the transformative capacity index is the availability of economic institutions to 
support livelihoods. Access to these economic institutions is varied. All households report having access 
 27 This index does not include a measure of quality for each of the service types. Instead, it only captures whether or not the 
service exists. Health services reflect whether NGOs are currently conducting health activities and not whether local health 
institutions are available.
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Findings 43
to markets.28 Only 10%, however, report having access to agricultural extension services.
29 Few 
households have access to natural communal resources: on average, households have access to only 
0.20 out of four natural communal resources (communal grazing land, water source, firewood, and 
irrigation source). Specifically, 15% of communities have communal grazing land, 3% of communities 
have an irrigation source, less than 1% of communities have a communal water source, and no 
communities report a communal firewood source.
Another aspect of the transformative capacity index reflects the strength of households to support 
themselves through their networks. Overall, the ability of households to draw on their networks is low. 
In particular, the bridging social capital index reflects the number of types of individuals that households 
could draw on outside of their communities (out of three groups). On average, households feel able to 
draw on 1.0 of these types of individuals. The linking social capital index reflects how well-connected 
households are to government or NGO leaders and whether they can draw on them for help. 
Households score very low (0.04 out of 4) on this, suggesting that the majority of households either do 
not know leaders or they are not able to ask leaders for help. The social cohesion index30 illustrates how 
active households have been in various support groups in the community. On average, households 
report engaging in 0.17 out of three support groups, reflecting that participation in support groups is not 
common. Finally, the local decision-making index reflects how actively households participate in groups 
in their communities. About one-third of households (35%) report active participation.
Another dimension captured by the transformative capacity index is the extent to which there are 
gender-related barriers in the community. Overall, there seem to be a moderate number of gender￾related barriers in the community. The gender index reflects constraints to gender-neutral behavior at 
the community level. On average, communities report 2.2 out of three gender-neutral behaviors are 
norms. The gender equitable decision-making index31 reflects how equitable decision-making is across 
male and female adults within the same household. On average, households score 0.68 out of 2 on this 
index, reflecting that out of two key household decisions, on average, 0.68 involve both male and 
female household members.
A final dimension of the transformative capacity index measures how available and reliable external 
sources of support are. Overall, households have access to a low number of these external resources. 
The formal safety nets index reflects the number of external safety nets (e.g., emergency food or cash 
assistance, agricultural inputs) available in the community. Overall, households have access to 1.6 out of 
13 formal safety nets. The government responsiveness index reflects whether households have access 
to a reliable police force and a peace committee.32 On average, households have access to 0.91 of two
of these resources.
 28 We assumed that any community in which a household reported selling crops to a local market had access to a market.
29 This index was calculated based on the percentage using agricultural extension services versus those with access, given 
available data.
30 This index ranges from 0–3 instead of 0–4 as we do not have data on whether community members came together for social 
events.
31 This index does not include measures of equitable decision-making around nutrition and child health as well as around 
savings. Thus, the index ranges from 0–2 instead of 0–4.
32 This indicator also corresponds to the baseline 24% of households that believe local government will respond effectively to 
future shocks and stresses.
IMPEL | Implementer-Led Evaluation and Learning
44 Comparison of Treatment and Control Groups
4. COMPARISON OF TREATMENT AND CONTROL GROUPS
Based on the data collected in the baseline, the research can evaluate the overall comparability of the 
treatment and control groups. This is done by comparing the mean values of a range of demographic 
and household-level characteristics and identifying any trends of statistically significant differences 
between the two groups. The table below illustrates the results of the exercise and confirms that the 
treatment and control group are, overall, balanced. There are no statistically significant differences in 
means between the two household groups at baseline. A joint test of orthogonality demonstrates that 
the balance characteristics are balanced (p-value = 0.976). Additional balance tables between treatment 
and control groups are in Annex A of this report. 
Table 53. Household roster balance table
Control Treatment
Outcome Mean N Mean N Difference P-value
Average age of people in the household 
roster 20.795 12,535 20.904 12,555 0.109 0.823
Percent of females in the household 
roster 51.047% 12,535 51.358% 12,557 0.311 0.641
Percent of farmers in the household 
roster 34.626% 6,324 31.508% 6,415 -3.118 0.078
Percent of people with at least some 
schooling in the household roster 65.895% 10,409 68.879% 10,409 2.983 0.31
Percent of people who worked for cash in 
the roster 52.306% 7,689 51.369% 7,746 -0.936 0.728
Percent of households with adult male 
and female present in the roster 83.058% 2,160 82.701% 2,140 -0.357 0.852
Percent of households with adult female 
only present in the roster 15.825% 2,160 16.709% 2,140 0.883 0.637
Percent of households with adult male 
only present in the roster 1.116% 2,160 0.59% 2,140 -0.527 0.075
Household Head
Average age of heads of households in 
the roster 44.651 2,156 45.462 2,141 0.811 0.39
Percent of female head of households in 
the roster 25.317 2,157 27.897 2,142 2.58 0.26
Percent that did not attend school in the 
roster 32.966 2,157 31.898 2,142 -1.068 0.719
Percent of people in the roster with some 
schooling, less than primary 34.228 2,157 30.171 2,142 -4.057 0.14
Percent of people in the roster that 
completed Primary or more 32.788 2,156 37.924 2,141 5.136 0.053
*Denotes significance at 10 pct., ** at 5 pct., and *** at 1 pct. The p-value for a joint test of orthogonality is 0.976
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Conclusion 45
5. CONCLUSION
The FIOVANA baseline survey was conducted in southeastern Madagascar in the districts of 
Farafangana, Manakara Atsimo, Mananjary, Vangaindrano, Vohipeno, and Vondrozo. This area is 
normally characterized by abundant rainfall and favorable growing conditions for a variety of crops, but 
this region is also remote with rugged terrain and poor road access. The baseline survey of 4,300 
households reflects an area dominated by agriculture. Over 94% of households farm and 92% of farmers 
grow rice. While farmers do grow cash crops, including cloves, coffee, and bananas, agriculture is limited 
by the lack of access to input and output markets, contract farming, credit, and extension.
Poverty rates in the region are high. The percentage living on less than $1.90 per day (2011 PPP) is 72%. 
Furthermore, 18% of individuals are considered to be facing severe food insecurity, and over 81% are at 
least moderately food insecure. Only 31% of households have an adequate Food Consumption Score. 
Very few children are consuming a diet of minimum diversity (5%) and minimum acceptability (3%). A 
quarter of children under 5 years old are underweight in the region. While women have a higher 
minimum dietary diversity than children, the percentage (9%) is still low, with most women consuming 
only three diverse food groups. 
Households have access to few resources that would enable them to better mitigate shocks and 
perceive that they have a low ability to recover from future shocks. Moreover, institutions available to 
enable households' capacity to respond to shocks seem to be weak. These findings underscore the 
importance of the resilience-focused activities that FIOVANA aims to deliver.
The baseline data shows a good balance between treatment and control communes. Looking ahead to 
the endline survey in 2025, there are several potential challenges. First, the length of time between 
baseline and endline implies that attrition may be an issue. Second, there are many organizations 
working in this region, some of which will likely have health, nutrition, or agriculture interventions in 
control as well as treatment communes. This may make it more difficult to identify the effects of the 
FIOVANA activity. To address this, the baseline survey collected information on current activities
operating in each commune, and the same will be done at endline. Furthermore, the outcome 
monitoring study planned for 2023 will provide an opportunity to assess changes in the practices and 
activities of households linked to FIOVANA activities.
IMPEL | Implementer-Led Evaluation and Learning
46 Annex A: Balance Tables
ANNEX A: BALANCE TABLES
Table 54. Food security
Outcome Control Treatment
Mean N Mean N Difference P-value
Household FCS (0–112) 32.891 2,115 32.302 2,097 -0.589 0.507
Percent of households with poor consumption 
score (<22) 5.88 2,115 7.967 2,097 2.087 0.294
Percent of households with borderline 
consumption score (22–35) 62.01 2,115 62.269 2,097 0.258 0.941
Percent of households with acceptable 
consumption score (>35) 32.109 2,115 29.764 2,097 -2.345 0.615
Percent of households worried about not 
having enough food to eat because of a lack of 
money or other resources 91.566 2,115 92.54 2,097 0.974 0.639
Percent of households unable to eat healthy 
and nutritious food because of a lack of money 
or other resources 92.566 2,115 95.798 2,097 3.231 0.058
Percent of households that ate only a few kinds 
of foods because of a lack of money or other 
resources 92.589 2,115 95.914 2,097 3.325 0.045
Percent of households that skipped a meal 
because there was not enough money or other 
resources 56.686 2,115 61.165 2,097 4.479 0.254
Percent of households that ate less than you 
thought you should because of a lack of money 
or other resource 84.907 2,115 88.783 2,097 3.876 0.104
Percent of households that did not have food 
because of a lack of money or other resources 23.676 2,115 29.752 2,097 6.076 0.1
Percent of households that are hungry but did 
not eat because there was not enough money 
or other resource 23.786 2,115 31.39 2,097 7.604 0.115
Percent of households that went without eating 
for a whole day because of a lack of money or 
other resource 17.74 2,115 23.329 2,097 5.59 0.18
*Denotes significance at 10 pct., ** at 5 pct., and *** at 1 pct. The p-value for a joint test of orthogonality is 0.000
Table 55. Child nutrition and health
Outcome Control Treatment
Percent N Percent N Difference P-value
MDD children ages 6–23 months 0.068 639 0.039 639 -0.029 0.127
MAD children ages 6–23 months 0.038 639 0.025 639 -0.013 0.309
Exclusively breastfed children under 6 
months of age
0.608 176 0.656 209 0.048 0.469
Children under 5 who had diarrhea 0.118 2,089 0.108 2,112 -0.011 0.601
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex A: Balance Tables 47
Outcome Control Treatment
Percent N Percent N Difference P-value
Children under age 5 who had fever, 
diarrhea, or acute raspatory infection (ARI) 
in the past 15 days and received treatment 
within 2 hours from a health facility or 
health service 
0.548 660 0.596 772 0.048 0.299
Children who had diarrhea and were given 
ORT 0.174 243 0.278 234 0.104* 0.07
*The results in this table are expressed as percentages
Table 56. Anthropometry
Outcome Control Treatment
Mean N Mean N Difference P-value
Weight-for-age Z-score (WAZ) children 0–59 
months -1.28 2,049 -1.271 2,047 0.01 0.918
Female children 0–59 months WAZ -1.153 1,033 -1.15 1,072 0.003 0.976
Male children 0–59 months WAZ -1.411 1,016 -1.401 975 0.01 0.924
Children 0–23 months WAZ -0.944 796 -1.071 810 -0.127 0.407
Children 24–59 months WAZ 0.238
Percent of children 0–59 months underweight 26.301 2,049 25.462 2,047 -0.84 0.702
Percent of female children 0–59 months that are 
underweight 23.402 1,033 22.385 1,072 -1.017 0.673
Percent of male children 0–59 months that are 
underweight 29.264 1,016 28.774 975 -0.49 0.862
Percent of children 0–23 months that are 
underweight 21.927 796 25.407 810 3.48 0.271
Percent of children 24–59 months that are 
underweight 29.094 1,253 25.496 1,237 -3.598 0.123
Percent of children 0–59 months that are 
severely underweight 9.106 2,049 6.679 2,047 -2.427 0.07
Percent of female children 0–59 months that are 
severely underweight 7.733 1,033 6.131 1,072 -1.602 0.46
Percent of male children 0–59 months that are 
severely underweight 10.509 1,016 7.27 975 -3.239 0.077
Percent of children 0–23 months that are 
severely underweight 8.194 796 7.781 810 -0.414 0.834
Percent of children 24–59 months that are 
severely underweight 9.688 1,253 5.98 1,237 -3.708 0.012
Percent of female children 0–59 months with a 
normal weight 43.444 1,033 41.147 1,072 -2.298 0.477
Percent of male children 0–59 months with a 
normal weight 35.101 1,016 33.771 975 -1.33 0.663
Percent of children 0–23 months with a normal
weight 48.622 796 44.005 810 -4.616 0.29
Percent of children 24–59 months with normal
weight 33.376 1,253 33.526 1,237 0.15 0.963
IMPEL | Implementer-Led Evaluation and Learning
48 Annex A: Balance Tables
Table 57. Women’s health, maternal nutrition, and reproductive health
Outcome Control Treatment
Percent N Percent N Difference P-value
Women of reproductive age (15–49) with MDD 0.032 2416 0.034 2467 0.002 0.867
Women in a union using birth control 0.352 1176 0.297 1121 -0.055 0.233
Women of reproductive age who had a live birth 
during the last 5 years that received ANC during 
last pregnancy 0.661 1347 0.682 1372 0.021 0.61
Women of reproductive age in a union who have 
knowledge of modern family planning methods 0.78 1302 0.782 1263 0.002 0.964
Women of reproductive age in a union who use a 
modern family planning method in the last 12 
months who made decisions about modern family 
planning methods in the past 12 months 0.927 482 0.936 429 0.009 0.801
Women of reproductive age who take at least one 
method of birth control 0.032 2416 0.034 2467 0.002 0.867
*Denotes significance at 10 pct., ** at 5 pct., and *** at 1 pct. The p-value for a joint test of orthogonality is 0.681
*The results in this table are expressed as percentages
Table 58. Water, sanitation, and hygiene 
Outcome Control Treatment
Mean N Mean N Difference P-value
Percent of households with improved water source 16.859 2,159 20.701 2,140 3.842 0.404
Water source within 30 minutes per household 95.364 2,159 96.453 2,140 1.089 0.237
Water available year-round per household 87.594 2,159 85.044 2,140 -2.55 0.394
Households with access to basic drinking water 
services 14.188 2,159 16.72 2,140 2.531 0.536
Handwashing available per household 3.503 1,929 4.097 1,863 0.593 0.681
Households that treat their water 46.528 2,159 54.22 2,140 7.692 0.256
Households that treat water by adding bleach or 
chlorine before drinking 2.118 2,159 2.224 2,140 0.106 0.936
Households that treat water by flocculation before 
drinking 19.884 2,159 20.295 2,140 0.41 0.903
Households that treat water by filtration before 
drinking 10.765 2,159 13.656 2,140 2.891 0.415
Households that treat water by solar disinfection 35.22 2,159 42.193 2,140 6.974 0.238
Households that treat water by boiling before 
drinking 0.521 2,159 0.898 2,140 0.377 0.306
Households practicing open defecation 63.33 2,159 67.99 2,140 4.66 0.505
Households using improved sanitation facilities 2.764 2,159 2.407 2,140 -0.357 0.738
Household water use per capita (liters) 16.58 990 14.839 1,128 -1.741 0.108
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex A: Balance Tables 49
Table 59. Agriculture–cassava, rice, and cloves
Outcome Control Treatment
Mean N Mean N Difference P-value
Cassava
Percent of farmers using at least one practice for 
cassava 21.453 1,705 20.998 1,601 -0.455 0.901
Percent of farmers using interplanting for cassava 10.915 1,708 12.093 1,604 1.179 0.681
Percent of farmers sowing after significant rain 
for cassava 5.281 1,708 3.92 1,604 -1.361 0.524
Percent of farmers that did not use any of the 
practices for cassava 79.684 1,708 79.44 1,604 -0.243 0.947
Rice
Percent of farmers using at least one practice for 
rice 24.036 1,847 22.917 1,770 -1.119 0.818
Percent of farmers using organic manure for rice 5.348 1,853 6.868 1,775 1.52 0.476
Percent of farmers sowing after significant rain 
for rice 12.238 1,853 10.746 1,775 -1.492 0.759
Percent of farmers that did not use any of the 
practices for rice 76.891 1,853 77.519 1,775 0.629 0.894
Cloves
Percent of farmers using at least one practice for 
cloves 42.449 528 35.975 576 -6.474 0.513
Percent of farmers interplanting for cloves 30.982 528 24.869 576 -6.113 0.486
Percent of farmers using soil cover for cloves 5.929 528 7.303 576 1.374 0.71
Percent of farmers that did not use any of the 
practices for cloves 59.322 528 64.207 576 4.885 0.626
Other Practices
Percent of farmers using agricultural credit 5.411 2,103 5.061 2,003 -0.35 0.808
Percent of farmers who saved 12.167 2,105 9.773 2,005 -2.394 0.215
Percent of farmers using insurance 0.312 2,105 0.251 2,005 -0.062 0.808
Percent of farmers reporting at least one value 
chain activity 36.423 1,645 26.035 1,535 -10.388 0.283
Table 60. Agriculture–yield
Outcome Control Treatment
Mean N Mean N Difference P-value
Farmers reporting cassava yield in kg per ha 5,340.62 1,387 4,805.695 1,315 -534.924 0.352
Farmers reporting cloves yield in kg per ha 680.779 160 665.038 165 -15.74 0.935
Farmers reporting rice yield in kg per ha 4,317.811 1,605 3,459.563 1,554 -858.247 0.068
Farmers reporting zero rice production 1.318 1,852 1.333 1,780 0.015 0.977
IMPEL | Implementer-Led Evaluation and Learning
50 Annex A: Balance Tables
Table 61. Poverty measurements
Outcome Control Treatment
Mean N Mean N Difference P-value
Percent poor per household 69.495 2,160 74.284 2,140 4.789 0.095
Daily per capita expenditure (Ariary) 2,099.067 2,160 2,040.822 2,140 -58.245 0.435
Daily per capita expenditure (Dollars) 1.633 2,160 1.588 2,140 -0.045 0.435
*Denotes significance at 10 pct., ** at 5 pct., and *** at 1 pct. The p-value for a joint test of orthogonality is 0.028
Table 62. Use of financial resources
Outcome Control Treatment
Percent N Percent N Difference P-value
Women and men in a union who earned cash in the past 
12 months 0.745 3,220 0.739 3,164 -0.006 0.858
Men in a union and earning cash who report 
spouse/partner participation in decisions about the use 
of self-earned cash 0.829 1,479 0.808 1,426 -0.021 0.604
Women in a union and earning cash who report 
participation in decisions about the use of self-earned 
cash 0.866 934 0.821 914 -0.045 0.494
*Denotes significance at 10 pct., ** at 5 pct., and *** at 1 pct. The p-value for a joint test of orthogonality is 0.4865
*The results in this table are expressed as percentages
Table 63. Credit
Outcome Control Treatment
Percent N Percent N Difference P-value
People in a union who have access to credit 0.356 3,220 0.359 3,164 0.003 0.945
Men and women in a union who report 
making the borrowing decision
0.699 1,165 0.673 1,115 -0.026 0.456
* denotes significance at 10 pct., ** at 5 pct., and *** at 1 pct. The p-value for a joint test of orthogonality is 0.458
*The results in this table are expressed as percentages
Table 64. Resilience
Outcome Control Treatment
Mean N Mean N Difference P-value
Household adaptive capacity index (0–100) 29.73 1,886 33.966 1,863 4.236 0.034
Household bridging Social Capital index (0–6) 0.978 2,160 1.063 2,140 0.085 0.363
Household linking social capital (0–4) 0.041 2,160 0.029 2,140 -0.012 0.312
Household social network index (0–6) 1.462 2,150 1.959 2,140 0.497 0.003
Household education/training index (0–7) 1.112 2,159 1.245 2,135 0.133 0.11
Household asset ownership index—total type 
(0–31)
6.416 2,102 6.398 2,089 -0.018 0.944
Household access to financial resources (0–2) 0.403 2,160 0.787 2,140 0.384 0.011
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex A: Balance Tables 51
Outcome Control Treatment
Mean N Mean N Difference P-value
Household livelihood diversification index (0–
20)
2.716 2,160 2.799 2,140 0.084 0.619
Household adoption of improved practices 
index (0–1)
0.07 1,990 0.059 1,966 -0.011 0.559
Household exposure to information index (0–
19)
4.215 2,160 4.248 2,140 0.033 0.95
Household aspirations/confidence to adapt 
index (0–16)
9.205 2,056 9.198 2,039 -0.007 0.97
Household absorptive capacity index (0–100) 27.103 1,904 33.376 1,952 6.273 0.02
Household access to cash savings index (0–1) 0.118 2,159 0.115 2,140 -0.003 0.925
Household asset ownership index—total type 
(0–31)
6.416 2,102 6.398 2,089 -0.018 0.944
Household remittances index (0–1) 0.01 2,160 0.012 2,140 0.002 0.641
Household access to insurance index (0–1) 0.003 1,990 0.003 1,966 -0.001 0.769
Household bonding Social Capital index (0–6) 1.112 2,160 1.199 2,140 0.086 0.319
Household access to informal safety nets 
index (0–6)
2.396 2,080 2.779 2,140 0.383 0.111
Household shock preparedness & 
responsiveness index (0–3)
0.922 2,160 1.098 2,140 0.175 0.071
Household access to humanitarian assistance 
index (0–1)
0.089 2,160 0.277 2,140 0.188 0
Household transformative capacity index (0–
100)
35.296 1,333 36.04 1,782 0.744 0.851
Household access to formal safety nets index 
(0–13)
1.787 1,339 1.562 1,800 -0.225 0.335
Household access to markets index (0–1) 1 2,160 1 2,140 0.000
Household access to communal natural 
resources index (0–4)
0.237 2,160 0.156 2,140 -0.081 0.481
Household basic services index (0–4) 1.375 2,160 1.92 2,140 0.545 0
Household access to infrastructure index (0–3) 1.004 2,160 0.949 2,140 -0.056 0.731
Household access to agricultural services 
index (0–1)
0.044 2,160 0.146 2,140 0.102 0.154
IMPEL | Implementer-Led Evaluation and Learning
52 Annex A: Balance Tables
Outcome Control Treatment
Mean N Mean N Difference P-value
Household bridging Social Capital index (0–6) 0.978 2,160 1.063 2140 0.085 0.363
Household linking social capital (0–4) 0.041 2,160 0.029 2140 -0.012 0.312
Household social cohesion index (0–3) 0.169 2,160 0.16 2140 -0.009 0.934
Household gender equitable decision-making 
index (0–2)
0.704 2,160 0.664 2140 -0.039 0.617
Household local decision-making index (0–1) 0.34 2,148 0.359 2122 0.019 0.654
Household local government responsiveness 
index (0–2)
0.871 2,160 0.951 2140 0.080 0.312
Household gender index (0–3) 2.368 2,160 2.118 2140 -0.250 0.01
Household ability to recover from shocks & 
stresses index
2.586 1,849 2.606 1862 0.020 0.813
Household index of social capital at household 
level (0-100)
49.56 2,160 52.988 2140 3.428 0.339
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex B: Additional Tables 53
ANNEX B: ADDITIONAL TABLES
Table 65. Crops grown, gender
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Percent farmers growing 
cassava
89.3% (1.32) 3,028 66.4% (3.14) 914 83.9% (1.57) 3,942
Percent farmers growing 
rice
94.6% (0.63) 3,028 84.3% (2.32) 914 92.1% (0.67) 3,942
Percent farmers growing 
cloves
32.7% (2.38) 3,028 12.3% (1.52) 914 27.8% (2.04) 3,942
Percent farmers growing 
coffee
43.4% (2.08) 3,028 19.0% (1.79) 914 37.6% (1.84) 3,942
Percent farmers growing 
citrus
31.8% (2.43) 3,028 19.2% (2.05) 914 28.8% (2.12) 3,942
Percent farmers growing 
mango
10.9% (1.63) 3,028 4.2% (0.62) 914 9.3% (1.31) 3,942
Percent farmers growing 
cassava
89.3% (1.32) 3,028 66.4% (3.14) 914 83.9% (1.57) 3,942
Percent farmers growing 
sw_pot
40.8% (2.79) 3,028 32.6% (3.28) 914 38.8% (2.68) 3,942
Percent farmers growing 
cloves
32.7% (2.38) 3,028 12.3% (1.52) 914 27.8% (2.04) 3,942
Percent farmers growing 
banana
42.0% (2.12) 3,028 21.5% (1.46) 914 37.1% (1.70) 3,942
Percent farmers growing 
vanilla
17.0% (1.52) 3,028 4.6% (0.73) 914 14.1% (1.22) 3,942
Table 66. Agriculture value chain practices, gender
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Percent using agricultural 
credit
5.4% (0.65) 3,049 4.9% (0.81) 1,057 5.2% (0.62) 4,106
Percent of farmers who 
saved
11.6% (1.00) 3,051 9.2% (1.26) 1,059 11.0% (0.90) 4,110
Percent of farmers using 
insurance
0.3% (0.15) 3,051 0.1% (0.13) 1,059 0.3% (0.09) 4,110
Percent of farmers 
reporting at least one 
value chain activity
30.5% (2.72) 2,405 34.5% (4.67) 775 31.5% (3.09) 3,180
IMPEL | Implementer-Led Evaluation and Learning
54 Annex B: Additional Tables
Table 67. Modern practices for cassava, gender
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Percent using at 
least one 
practice for 
cassava
21.5% (1.52) 2,705 20.1% (2.53) 601 21.2% (1.45) 3,306
Organic manure 
for cassava
1.2% (0.20) 2,709 1.7% (0.65) 603 1.3% (0.24) 3,312
Compost for 
cassava
3.0% (0.59) 2,709 1.7% (0.73) 603 2.8% (0.51) 3,312
Performing 
weedings for 
cassava
0.03% (0.03) 2,709 0.3% (0.27) 603 0.1% (0.05) 3,312
Sowing after 
significant rain 
for cassava
4.9% (0.86) 2,709 3.6% (1.05) 603 4.6% (0.81) 3,312
Interplanting for 
cassava
11.5% (0.94) 2,709 11.4% (2.12) 603 11.5% (0.97) 3,312
Crop rotation for 
cassava
2.4% (0.54) 2,709 2.1% (0.59) 603 2.3% (0.49) 3,312
Use of improved 
seeds for cassava
0.2% (0.12) 2,709 0.3% (0.21) 603 0.2% (0.10) 3,312
Use of climate 
information for 
cassava
0.2% (0.15) 2,709 0.0% (.) 603 0.2% (0.12) 3,312
Wind break for 
cassava
0.0% (.) 2,709 0.0% (.) 603 0.0% (.) 3,312
Soil cover for 
cassava
0.3% (0.09) 2,709 0.0% (.) 603 0.2% (0.07) 3,312
Micro doses of 
fertilizer for 
cassava 
0.0% (.) 2,709 0.7% (0.50) 603 0.1% (0.10) 3,312
Crop rotation for 
cassava
0.4% (0.15) 2,709 0.3% (0.30) 603 0.4% (0.13) 3,312
Intercropping 
and Agroforestry 
for cassava
0.4% (0.17) 2,709 0.3% (0.18) 603 0.4% (0.15) 3,312
land leveling for 
cassava
0.1% (0.07) 2,709 0.0% (.) 603 0.1% (0.05) 3,312
Did not use any 
of the practices 
for cassava
79.2% (1.42) 2,709 81.0% (2.38) 603 79.6% (1.37) 3,312
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex B: Additional Tables 55
Table 68. Modern practices for rice, gender
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Percent using at least one 
practice for rice
23.8% (2.33) 2,853 22.4% (2.69) 764 23.5% (2.22) 3,617
Use of improved seeds for 
rice
0.1% (0.06) 2,860 0.4% (0.31) 768 0.2% (0.08) 3,628
Use of climate information 
for rice
0.2% (0.08) ,2860 0.0% (.) 768 0.2% (0.06) 3,628
Wind break for rice 0.0% (0.03) 2,860 0.0% (.) 768 0.0% (0.02) 3,628
Soil cover for rice 0.1% (0.06) 2,860 0.2% (0.20) 768 0.1% (0.06) 3,628
Micro doses of fertilizer for 
rice 
1.3% (0.27) 2,860 0.8% (0.24) 768 1.2% (0.22) 3,628
Crop rotation for rice 0.2% (0.07) 2,860 0.2% (0.17) 768 0.2% (0.06) 3,628
Intercropping and 
Agroforestry for rice
0.0% (0.04) 2,860 0.1% (0.14) 768 0.1% (0.04) 3,628
land leveling for rice 0.0% (.) 2,860 0.1% (0.14) 768 0.0% (0.03) 3,628
Did not use any of the 
practices for rice
76.9% (2.30) 2,860 78.4% (2.66) 768 77.2% (2.19) 3,628
Table 69. Modern practices for cloves, gender
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Percent using at least one 
practice for cloves
39.7% (4.02) 994 32.8% (6.03) 110 39.0% (4.04) 1,104
Compost for cloves 3.4% (0.94) 994 3.3% (2.03) 110 3.4% (0.88) 1,104
Performing weedings for 
cloves
0.0% (.) 994 0.0% (.) 110 0.0% (.) 1,104
Sowing after significant rain 
for cloves
0.5% (0.16) 994 0.5% (0.52) 110 0.5% (0.19) 1,104
Interplanting for cloves 28.2% (3.73) 994 23.7% (5.87) 110 27.7% (3.83) 1,104
Crop rotation for cloves 0.7% (0.35) 994 0.0% (.) 110 0.7% (0.31) 1,104
Use of improved seeds for 
cloves
0.6% (0.34) 994 0.5% (0.52) 110 0.6% (0.32) 1,104
Use of climate information 
for cloves
0.5% (0.34) 994 0.0% (.) 110 0.4% (0.30) 1,104
Wind break for cloves 0.1% (0.08) 994 0.0% (.) 110 0.1% (0.07) 1,104
Soil cover for cloves 6.7% (1.70) 994 6.5% (2.70) 110 6.7% (1.68) 1,104
Intercropping and 
Agroforestry for cloves
4.2% (1.43) 994 0.9% (0.76) 110 3.8% (1.22) 1,104
land leveling for cloves 0.1% (0.13) 994 0.0% (.) 110 0.1% (0.12) 1,104
Use of improved seeds for 
cloves
0.6% (0.34) 994 0.5% (0.52) 110 0.6% (0.32) 1,104
IMPEL | Implementer-Led Evaluation and Learning
56 Annex B: Additional Tables
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Did not use any of the 
practices for cloves
61.3% (4.07) 994 67.2% (6.03) 110 61.9% (4.09) 1,104
Table 70. Agricultural yield, gender
Male Female All
Outcome Mean SE N Mean SE N Mean SE N
Cassava yield in 
kg per ha
4,912.53 (262.89) 2,214 5,768.37 (403.90) 490 5,074.29 (272.12) 2,704
Cloves yield in 
kg per ha
616.48 (65.97) 276 919.8 (158.23) 49 672.17 (74.69) 325
Rice yield in kg 
per ha
3,858.06 (238.32) 2,503 4,046.44 (326.14) 657 3,897.38 (247.35) 3,160
Table 71. Poultry farming, gender
Male Female All
Outcome Percent SE N Percent SE N
Percen
t
SE N
Percent of farmers raising 
poultry
69.4% (1.66) 3,100 58.8% (2.30) 1,087 66.6% (1.64) 4,187
Use of improved poultry 
variety/breed
0.6% (0.18) 2,142 0.6% (0.41) 631 0.6% (0.17) 2,773
Use of improved feed 0.0% (0.01) 2,142 0.0% (.) 631 0.0% (0.01) 2,773
Use of improved shelters 0.4% (0.16) 2,142 0.0% (.) 631 0.3% (0.12) 2,773
Use of improved fodder 
production
0.1% (0.05) 2,142 0.0% (.) 631 0.1% (0.04) 2,773
Vaccinations 3.7% (0.74) 2,142 5.1% (1.31) 631 4.0% (0.72) 2,773
Use of para-veterinary 
services for poultry
0.2% (0.13) 2,124 0.1% (0.06) 624 0.2% (0.10) 2,748
Food preservation 0.1% (0.06) 2,124 0.2% (0.15) 624 0.1% (0.06) 2,748
Selective breeding 0.2% (0.11) 2,124 0.0% (.) 624 0.1% (0.08) 2,748
Habitat 3.2% (0.56) 2,124 1.5% (0.48) 624 2.8% (0.39) 2,748
Percent of farmers 
reporting eggs in 
previous week
27.6% (1.90) 2,124 23.4% (2.11) 630 26.6% (1.72) 2,754
Poultry yield in kg per 
bird
2.4% (0.02) 2,139 2.4% (0.03) 630 2.4% (0.02) 2,769
Percent selling poultry in 
the last year
61.6% (1.59) 2,124 60.9% (2.95) 630 61.5% (1.38) 2,754
Percent consuming own 
poultry in the last year
82.4% (1.42) 2,123 74.0% (2.75) 630 80.4% (1.33) 2,753
Percent of farmers 
experiencing poultry 
dying in the last year
64.9% (2.07) 2,124 73.2% (2.93) 630 66.8% (1.81) 2,754
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex B: Additional Tables 57
Male Female All
Outcome Percent SE N Percent SE N
Percen
t
SE N
Did not use any of these 
practices in the past 12 
months
93.8% (0.76) 2,142 94.2% (1.24) 631 93.9% (0.67) 2,773
IMPEL | Implementer-Led Evaluation and Learning
58 Annex B: Additional Tables
Table 72. Beekeeping, gender
Male Female All
Outcome Mean SE N Mean SE N Mean SE N
Total number of hives 4.15 (0.41) 65 5.92 (3.40) 4 4.25 (0.54) 69
Number of modern hives n/a n/a 13 n/a n/a 2 n/a n/a 15
Liters of honey per modern hive n/a n/a 13 n/a n/a 2 n/a n/a 15
Number of traditional hives 3.51 (0.39) 58 n/a n/a 2 3.44 (0.38) 60
Liters of honey per traditional hive 1.91 (0.55) 58 n/a n/a 2 1.92 (0.53) 60
Percentage using beekeeping equipment (listed below)
Beehive 90.2% (4.93) 65 n/a n/a 4 90.8% (4.73) 69
Smoker 29.8% (8.72) 65 n/a n/a 4 28.1% (8.41) 69
Frames and Foundation 11.9% (5.58) 65 n/a n/a 4 11.2% (5.22) 69
Veil and Gloves 11.8% (5.09) 65 n/a n/a 4 11.1% (4.85) 69
Feeder 1.4% (1.69) 65 n/a n/a 4 1.4% (1.56) 69
Heated Knife 25.7% (4.90) 65 n/a n/a 4 25.0% (4.68) 69
Uncapping Fork 7.6% (3.67) 65 n/a n/a 4 7.1% (3.47) 69
Tub For Wax/Honey 2.5% (2.10) 65 n/a n/a 4 2.3% (1.95) 69
Extractor 1.0% (1.02) 65 n/a n/a 4 1.0% (0.96) 69
Food-Grade Bucket 8.2% (4.46) 65 n/a n/a 4 8.6% (4.38) 69
Double Sieve 5.5% (3.51) 65 n/a n/a 4 5.2% (3.26) 69
Containers 3.9% (2.80) 65 n/a n/a 4 3.7% (2.60) 69
Queen excluder 2.8% (2.22) 65 n/a n/a 4 2.7% (2.10) 69
Queen cage 7.0% (2.88) 65 n/a n/a 4 6.6% (2.72) 69
Beekeeper's suit 3.5% (2.84) 65 n/a n/a 4 3.3% (2.64) 69
Nylon brush 3.2% (2.38) 65 n/a n/a 4 3.1% (2.20) 69
Attracts swarm 7.0% (2.92) 65 n/a n/a 4 6.6% (2.77) 69
Filter cloth 19.7% (6.09) 65 n/a n/a 4 18.6% (5.53) 69
Knife 38.7% (8.86) 65 n/a n/a 4 39.5% (8.87) 69
Bucket 20.6% (6.14) 65 n/a n/a 4 20.2% (5.53) 69
Table 73. Fishing, gender
Male Female All
Outcome Percent SE N Percent SE N Percent SE N
Percent fishing for food
only
46.5% (6.40) 204 43.3% (10.73) 83 45.5% (5.93) 287
Percent fishing for market 
only
1.7% (0.86) 204 6.0% (3.64) 83 3.0% (1.37) 287
Percent fishing for both 
food and market
51.8% (6.24) 204 50.7% (9.27) 83 51.5% (5.53) 287
Percent of fishers using a
pirogue
37.7% (6.14) 204 8.1% (5.94) 82 28.6% (5.01) 286
Percent of fishers using 
nets
55.6% (6.40) 204 19.2% (6.07) 82 44.4% (5.17) 286
Percent of fishers using 
containers
27.2% (4.88) 204 43.3% (7.20) 82 32.1% (4.42) 286
Baseline Study of the FIOVANA RFSA in Madagascar (Vol. I)
Annex B: Additional Tables 59
Table 74. Non-governmental organizations in the control areas
Organization/leader of the project Frequency Percent Cum.
ADRA 3 2.3% 2.26
AFAFI SUD 3 2.3% 4.51
DEFIS 23 17.3% 21.8
FAO 1 0.8% 22.56
FID 8 6.0% 28.57
PAM 6 4.5% 33.08
ONN 1 0.8% 33.83
USAID ACCESS 20 15.0% 48.87
WHH (WELT HUNGER LIFE) 2 1.5% 50.38
Church 7 5.3% 55.64
Malagasy Government 3 2.3% 57.89
Individual 1 0.8% 58.65
International NGO 27 20.3% 78.95
Malagasy NGO 12 9.0% 87.97
Other bilateral organization 8 6.0% 93.98
Other international organization 8 6.0% 100
Total 133 100
Table 75. Non-governmental organizations in the treatment areas
Organization/leader of the project Freq. Percent Cum.
ADRA 48 26.1% 26.09
AFAFI SUD 4 2.2% 28.26
CARE 1 0.5% 28.8
CRS 4 2.2% 30.98
DEFIS 25 13.6% 44.57
FID 7 3.8% 48.37
PAM 1 0.5% 48.91
ONN 2 1.1% 50
USAID ACCESS 20 10.9% 60.87
WHH (WELT HUNGER LIFE) 5 2.7% 63.59
Church 6 3.3% 66.85
Malagasy Government 4 2.2% 69.02
International NGO 28 15.2% 84.24
Malagasy NGO 12 6.5% 90.76
Other bilateral organization 5 2.7% 93.48
Other international organization 12 6.5% 100
Total 184 100
IMPEL | Implementer-Led Evaluation and Learning
60 Annex B: Additional Tables
Table 76. Non-governmental organization activity
Control Treatment All
Outcome
Percent 
reporting the 
activity
N
Percent 
reporting the 
activity
N
Percent 
reporting the 
activity
N
Administrative support (conseil, 
finance, personnel) 
0.0% 133 0.0% 184 0.0% 317
Construction, public buildings 0.0% 133 0.0% 184 0.0% 317
Construction, roads 2.3% 133 1.6% 184 1.9% 317
Construction, other (including 
water and irrigation services) 
18.8% 133 10.3% 184 13.9% 317
Health 30.8% 133 41.9% 184 37.2% 317
Education 6.8% 133 3.8% 184 5.1% 317
Professional training 1.5% 133 1.1% 184 1.3% 317
Agriculture 43.6% 133 51.1% 184 48.0% 317
Livestock 13.5% 133 21.7% 184 18.3% 317
Fishing/fish farming 3.0% 133 7.6% 184 5.7% 317
Reforestation/environment 6.0% 133 10.9% 184 8.8% 317
Industry/crafts 0.0% 133 0.0% 184 0.0% 317
Commerce/transport 0.8% 133 1.1% 184 1.0% 317
Savings and credit 0.0% 133 1.6% 184 1.0% 317
Conservation (fauna and flora) 2.3% 133 0.0% 184 1.0% 317
Cultural 0.0% 133 0.5% 184 0.3% 317
Food aid 6.0% 133 16.9% 184 12.3% 317
Women and child services 2.3% 133 1.6% 184 1.9% 317
Disability services 0.0% 133 0.0% 184 0.0% 317
IEC support 1.5% 133 1.6% 184 1.6% 317
Support of farmers groups 15.0% 133 16.9% 184 16.1% 317
Good governance 0.0% 133 1.6% 184 1.0% 317
Market access 0.0% 133 0.5% 184 0.3% 317
*N is an organization, not commune