Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti

Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti February 2022 | 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, Innovations for Poverty Action, and the International Food Policy Research Institute. RECOMMENDED CITATION IMPEL. (2022). Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I). Washington, DC: The Implementer- Led Evaluation & Learning Associate Award. PHOTO CREDITS Reginald Louissaint Jr. / Save the Children 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: Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) TABLE OF CONTENTS List of Tables ...................................................................................................... iii List of Figures ..................................................................................................... iii Acronyms ........................................................................................................... iv Summary of Key Findings .................................................................................... v 1. Introduction ................................................................................................... 1 1.1. Adapting an Impact Evaluation into a COVID-19 Impact Study (Overview) ................................. 1 1.2. Study Objectives and Research Questions .................................................................................... 2 2. Methodology ................................................................................................. 3 2.1. Round 1 Quantitative .................................................................................................................... 3 2.1.1. Tools and Indicators .............................................................................................................. 4 2.1.2. Sampling ................................................................................................................................ 4 2.1.3. Training and Fieldwork .......................................................................................................... 5 2.2. Phase 1 Qualitative ....................................................................................................................... 6 2.2.1. Objectives and Tools ............................................................................................................. 7 2.2.2. Sample Size and Selection ..................................................................................................... 7 2.2.3. Training and Fieldwork .......................................................................................................... 7 2.3. Round 2 Quantitative .................................................................................................................... 8 2.3.1. Tools and Indicators .............................................................................................................. 8 2.3.2. Sampling ................................................................................................................................ 9 2.3.3. Training and Fieldwork .......................................................................................................... 9 2.4. Phase 2 Qualitative ....................................................................................................................... 9 2.4.1. Objectives and Tools ........................................................................................................... 10 2.4.2. Sampling .............................................................................................................................. 10 2.4.3. Training and Fieldwork ........................................................................................................ 11 2.5. COVID-19 Protocols and Ethical Considerations ......................................................................... 12 2.6. Data Analysis (Quantitative) ....................................................................................................... 12 2.7. Data Analysis (Qualitative) .......................................................................................................... 13 2.8. Limitations and Delimitations ..................................................................................................... 14 3. Overview of Food Security, Shocks, and COVID-19 in Haiti (2018–Present) .. 16 3.1. COVID-19 ..................................................................................................................................... 17 3.2. Other Shocks ............................................................................................................................... 18 3.3. Economic Instability .................................................................................................................... 18 3.4. Food Security .............................................................................................................................. 19 4. Results ......................................................................................................... 23 4.1. Changes in Poverty and Food Security ........................................................................................ 23 Table of Contents i IMPEL | Implementer-Led Evaluation and Learning 4.1.1. Poverty ................................................................................................................................ 23 4.1.2. Food Security ...................................................................................................................... 25 4.2. Perceptions of the Impacts of COVID-19 .................................................................................... 31 4.3. COVID-19 Precautions and Access to Healthcare ....................................................................... 31 4.4. Impacts of Drought, Poor Water Access, and Infrastructure ...................................................... 32 4.5. Agriculture and Other Livelihoods .............................................................................................. 33 4.6. Psychosocial ................................................................................................................................ 36 5. Conclusions .................................................................................................. 39 Annex A: Qualitative Tools ................................................................................ 41 Phase 1 Qualitative KII Guide English .................................................................................................. 41 Phase 2 Qualitative FGD Guide English ............................................................................................... 45 Phase 2 Qualitative Research Questions Guidance for FGD Leaders French ...................................... 49 Annex B: Quantitative Tools ............................................................................. 52 Round 1 English Questionnaire ........................................................................................................... 52 Round 1 Creole Questionnaire ............................................................................................................ 74 Round 2 Questionnaire ....................................................................................................................... 98 VOLUME II Annex C: CPVOD-19 Research Safety Protocol and Guidance Annex D: GHQ-12 Additional Analysis Annex E: Protocol ii Table of Contents Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) LIST OF TABLES Table 1. Quantitative Round 1 & 2 data collection dates and sample sizes ................................................. 9 Table 2. Phase 2 qualitative sample ............................................................................................................ 11 Table 3. Total verified cases of COVID-19 by department, October 2021 .................................................. 17 LIST OF FIGURES Figure 1. Study timeline ................................................................................................................................ 3 Figure 2. Timeline of shocks in Haiti 2018–present .................................................................................... 16 Figure 3. Consumer Price Indices (food, general) 2018–2021 .................................................................... 19 Figure 4. IPC analysis 2020–2021 (with projections for 2021–2022) ......................................................... 20 Figure 5. IPC analysis 2018–2019 (with projections for 2019-2020) .......................................................... 22 Figure 6. Prevalence of poverty by commune ............................................................................................ 23 Figure 7. Depth of poverty by commune .................................................................................................... 24 Figure 8. Food Consumption Score by commune ....................................................................................... 25 Figure 9. Consumption days by food group ................................................................................................ 26 Figure 10. Changes in per capita food expenditures and quantities .......................................................... 27 Figure 11. Household Dietary Diversity Score by commune ...................................................................... 28 Figure 12. Change in Household Hunger Scale by commune ..................................................................... 29 Figure 13. Household Hunger Scale question responses by commune ...................................................... 29 Figure 14. Reported livelihood activities .................................................................................................... 34 Figure 15. Asset ownership by commune ................................................................................................... 35 Figure 16. Change in assets between rounds ............................................................................................. 36 Figure 17. GHQ-12 and the Household Hunger Scale ................................................................................. 38 List of Tables and Figures iii IMPEL | Implementer-Led Evaluation and Learning ACRONYMS BAC Communal Agriculture Extension Office BHA Bureau for Humanitarian Assistance CPI Consumer Price Index DHS Demographic and Health Survey FCS Food Consumption Score FFP USAID’s Office of Food for Peace FGD Focus Group Discussion GHI Global Hunger Index GHQ-12 General Health Questionnaire- 12 HDDS Household Dietary Diversity Score HDVI Haitian Deprivation and Vulnerability Index HH Household HHS Household Hunger Scale IMPEL Implementer Led Evaluation and Learning IP Implementing-Partner IPC Integrated Phase Classification KI Key Informant KII Key Informant Interview KL Kore Lavi M&E Monitoring and Evaluation ODK Open Data Kit PPP Purchasing Power Parity RFSA Resilience Food Security Activity USAID United States Agency for International Development USD United States Dollar VSLA Village Savings and Loan Association iv Acronyms Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) SUMMARY OF KEY FINDINGS In early 2020, the Kore Lavi Strategic Objective 2 (SO2) impact evaluation was discontinued due to the COVID-19 pandemic and heightened insecurity. However, the USAID Bureau for Humanitarian Assistance (BHA) seized this unique opportunity to conduct a mixed-methods study in three Kore Lavi communes in the Northwest Department and upper Artibonite Department (Bombardopolis, Mole Saint Nicolas, and Anse Rouge) where data was collected prior to the start of the pandemic. The study looks at the impacts of recent shocks (COVID-19, climatic, social-political) on poor households,1 and the capacities and strategies these households have used to mitigate these events. This study builds on the first round of data collected from 276 households surveyed in January of 2020 by conducting a second round of data collection with the same households in January and February 2021, with the addition of questions related to COVID-19, shocks, and resilience. These two studies provide a panel database of 255 households (21 households lost at follow-up). Additionally, complementary qualitative research was conducted in two phases. Phase 1 was intended to inform hypotheses and modifications to the second round of the quantitative data collection tool. It took place in January and February of 2021 prior to the second round of quantitative data collection. Phase 2 of the qualitative research was conducted in June 2021 after the second round of quantitative data collection to further explore the impacts of COVID-19, other shocks, and resilience of poor households in the study areas. Phase 2 was also meant to better illuminate the quantitative findings. The study has identified several findings of particular importance to understanding the current food/nutrition security and poverty situation of poor households in the study areas. Key Finding 1: Changes in Poverty and Food Security According to the Integrated Phase Classification (IPC), food security in Haiti has worsened in the past 3 years, with the number of Haitians facing high levels of acute food insecurity nearly doubling between 2018 and 2021. The September 2021–February 2022 IPC indicates, in the communes focused on in this study, Bombardopolis and Mole Saint Nicolas are in Phase 4 (food security emergency), and Anse rouge is in Phase 3 (food security crisis). The cited drivers of poor food security are economic (inflation, reduction in remittances), poor harvests (poor rainfall), hurricane Laura (August 2020, mainly in the southern parts of the country), and the secondary impacts of COVID-19 (lockdowns, restrictions negatively affecting livelihoods). Between early 2020 (Round 1) and early 2021 (Round 2), the breadth and depth of poverty has increased, with a highly significant shift observed in Anse Rouge. Food security-related outcomes show mixed directions of change between rounds, but in all cases, a very large proportion of households experiencing food insecurity and hunger. The overall prevalence of households living on less than the equivalent of $1.90 per day per capita increased from 81% to 87%. The greatest increase was observed in Mole Saint Nicholas (from 68% of 1 Kore Lavi (KL) identified the poorest 11% of households in each targeted commune through a full census of all households. KL gathered basic information to estimate their poverty level using the Haitian Deprivation and Vulnerability Index (HDVI), a poverty proxy KL developed. Only these poorest households were included in the SO2 Impact Evaluation baseline survey. Summary of Key Findings v IMPEL | Implementer-Led Evaluation and Learning households in Round 1 to 80% in Round 2). This increase in poverty is largely due to a significant decrease in the purchasing power of households. The Purchasing Parity Index for Haiti increased by 24% between 2019 and 2020, resulting in a 34% increase in the poverty line after accounting for additional inflation in early 2021. Thus, while overall reported expenditures remained largely stable between the two rounds, reported quantities purchased indicated a significant decline in the estimated overall quantity of food consumed per person. These findings are triangulated by the qualitative data, where participants in both phases indicated price increases, with several Phase 2 Focus Group Discussion (FGD) participants reporting the prices of some basic commodities had as much as doubled. Food security was measured in the quantitative surveys using two different diet-related indicators and one experiential indicator. The Household Dietary Diversity Score (HDDS) measures the diversity of foods consumed within a household during the previous day and is reported as a mean or median value. The Food Consumption Score (FCS) is a measure of diet diversity and the frequency of consumption in the household during the previous 7 days. The resulting FCS is then reported as a prevalence, categorized into three food consumption groups (poor, borderline, and acceptable). The Household Hunger Scale (HHS) measures the perceived experience of hunger in the household over the past month and is reported as a prevalence. The scale is categorized into severe hunger, moderate hunger, and little/no hunger. Both the HDDS and the FCS showed that diets were poor in both Round 1 and Round 2, with infrequent consumption of fruits, vegetables, and animal proteins. The mean HDDS remained similar between rounds. The prevalence of poor/borderline FCS decreased from 81% of households in Round 1 to 71% in Round 2, mainly due to slight increases in consumption frequencies of staples, legumes, and fats. However, the FCS does not account for the actual quantities consumed. Further analysis of the estimates of 7-day food expenditures and quantities purchased revealed that households purchased smaller quantities of vegetables and fats in Round 2. There were no significant differences in the quantities of staples and legumes purchased despite increases in the FCS score for these categories. This suggests that households reporting more frequent consumption are likely consuming smaller quantities per day rather than actually increasing consumption. According to the HHS, the prevalence of households reporting moderate or severe hunger increased from 34% in Round 1 to 44% in Round 2. Key Finding 2: Perceptions of Impacts of COVID-19 The COVID-19 pandemic was not generally perceived by the key informants (Kis) and FGD participants as a specific shock in itself but rather an aggravating factor that worsens the impacts of other chronic and acute challenges. COVID-19 also lessens the ability of the poor to employ their usual survival strategies. In both phases of qualitative data collection, all KIs and FGD participants indicated that they were aware of few or no cases of COVID-19 in their areas. Secondary data indicate relatively few confirmed cases in Haiti, and informal sources suggest that cases have been concentrated in urban areas. Qualitative data indicate that remittances coming from elsewhere in Haiti and from abroad have decreased or stopped due to COVID-19. Social support within the communities, in general, has lessened because of social distancing, the closing of churches and schools, and general isolation due to COVID-19 restrictions. In addition, domestic and cross-border travel difficulty has made certain livelihoods more challenging. vi Summary of Key Findings Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) Key Finding 3: COVID-19 Precautions and Access to Healthcare Qualitative data indicate that the population is generally aware of the cause and preventive measures of COVID-19. However, the data also show that the pandemic has had a negative impact on accessing health care. The qualitative data indicate that people are aware of using masks and social distancing in public to reduce transmission, but few people use masks—generally because of a lack of vigilance or a lack of money to purchase masks. In Round 2, 20% of those interviewed indicated that they or someone in their household had delayed/skipped/been unable to complete health care visits in the previous year because of COVID-19. The specific reasons cited included fear of contracting COVID-19 (45%), lack of money to pay for healthcare (29%), lack of availability of healthcare (18%), and fear of being quarantined (4%). Qualitative data confirms these findings. FGDs and key informant interviews (KIIs) frequently indicate that people have been afraid to seek health care for fear of contracting COVID-19 and the stigma that suspected COVID-19 infection causes. Additionally, in Phase 1 of the qualitative data collection, the fear of being quarantined if seeking care for any COVID-19-like symptoms was a frequently cited concern. This was mentioned less often in the qualitative Phase 2 interviews and discussions. Key Finding 4: Impacts of Drought and Poor Water Access/Infrastructure The impacts of drought (acute and chronic) on agriculture and livestock livelihoods, coupled with the severe lack of water and agriculture infrastructure, such as irrigation, is one of the biggest underlying factors of household food insecurity, poverty, and low resilience in the areas studied. Drought and drinking/irrigation water access were mentioned as one of/or the most significant shocks/challenges in every qualitative KII and FGD. People are planting less than before (beans, corn, peanuts, potatoes) because of the lack of rain. Disease and lack of forage due to drought have caused significant loss of livestock. Round 2 added specific questions on drinking and irrigation water access. These data indicate that 76% of households must travel for more than 30 minutes to access drinking water, and 40% of households must travel for more than 2 hours to access drinking water. Over 40% of households report getting their drinking water from an unsafe source (mainly surface water and unprotected springs). Approximately 65% of households reported practicing agriculture in the past year (even if not cited as one of the main livelihoods). Of those, only 6% reported having access to irrigation (mainly a canal or ditch). The remaining 94% reported relying only on rainfall. Key Finding 5: Agriculture and Other Livelihoods The use of agriculture as part of the main livelihood strategies is fluid. Although commonly practiced, the evidence indicates that it is not considered a viable or reliable primary source of income. Investments in agriculture in these areas may not improve resilience or reduce poverty unless very large and long-term investments are made in water and irrigation infrastructure. Those relying solely on agriculture have among the worst food security outcomes, while households with two or more livelihoods were significantly more likely to report adequate food consumption levels. Households are selling off their productive agriculture assets in an attempt to cope with the worsening economic and food security situation. The overall percentage of households owning agricultural equipment declined from 48% to 10%, with similar significant trends in all three regions. Repeated Summary of Key Findings vii IMPEL | Implementer-Led Evaluation and Learning drought combined with a lack of agriculture infrastructure, such as irrigation, has further aggravated the challenges in relying on agriculture as a source of food and money. Young people are increasingly looking away from agriculture for other sources of income, such as moto-taxi driving. Key Finding 6: Psychological Distress Psychological distress is high in the surveyed populations, stemming from the challenges related to poverty, COVID-19, and other recent shocks. Round 2 survey data revealed high rates of psychological stress, with 28% of respondents scoring 7 or higher on the General Health Questionnaire-12 (GHQ-12).2 The prevalence varied significantly by commune, from 15% of respondents in Anse Rouge to 23% in Mole Saint Nicholas to 45% in Bombardopolis. This contrasts with 19% of respondents experiencing psychological stress in a 2011 national food security survey. Elevated psychological stress was triangulated by the qualitative data, where KIs and FGD participants reported experiencing greater general fear and stress than usual because of the pandemic. There was also clear evidence that psychological distress was associated with poverty. Among those with per capita daily spending above the estimated 2021 poverty line, none showed evidence of severe psychological stress, in contrast to 33% among the respondents with expenditures below the poverty line. Greater depth of poverty was most closely associated with feelings of depression, low self-worth, low confidence, and an inability to concentrate. 2 Using the binary method to calculate the GHQ-12 score gives a range of scores from 0–12. It is an internationally used tool to measure psychological stress. It is not meant as a screening of particular psychiatric disorders. viii Summary of Key Findings Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) 1. INTRODUCTION This section gives an overview of how this study evolved from a previously planned impact evaluation into the current study. 1.1. ADAPTING AN IMPACT EVALUATION INTO A COVID-19 IMPACT STUDY (OVERVIEW) In 2019, under the Implementer-Led Evaluation and Learning (IMPEL) Associate Award, Tulane University School of Public Health and Tropical Medicine (Tulane) was subcontracted to conduct a final impact evaluation of Kore Lavi (KL). Kore Lavi is a United States Agency for International Development (USAID) Bureau for Humanitarian Assistance (BHA)-funded Resilience Food Security Activity (RFSA) implemented in Haiti by a consortium of partners, led by CARE as prime. The objective of the quantitative final impact evaluation was to measure the impact of the Kore Lavi activities on food security and poverty of the poorest households targeted by Kore Lavi. The evaluation was designed to determine whether changes in food security and poverty indicators for direct participant households over time are attributable to the Kore Lavi food voucher distributions. The approach for this final impact evaluation was a quasi-experimental design, in which study areas were non-randomly assigned to an intervention or comparison group. The baseline survey3 data collection was conducted between November 2014 and December 2015. The target population and sampling frame for the baseline intervention group was the lowest 11% of vulnerable households in each commune identified through Kore Lavi’s vulnerability analysis. These households later benefited from Kore Lavi’s food vouchers for the duration of the project. The endline survey was planned to take place during the same period (November 2019 to December 2020) to avoid any possible seasonal bias. Rather than take a random sample at the endline from the same sampling frame used at baseline, the sampling frame was limited to only households that were included in the baseline survey, allowing paired pre-post analysis at the household level. The start of the endline survey data collection was delayed to early 2020 due to ongoing insecurity in Haiti. Data was collected from the first three communes (Bombardopolis and Mole Saint Nicholas in the Northwest Department and Anse Rouge in Artibonite Department) in January and February 2020. All three communes were from the KL intervention areas. A total of 276 households were interviewed in these three areas.4 The next planned period of data collection was to take place from May to July of 2020 in a different set of communes. However, due to the developing COVID-19 pandemic at this time, as well as heightened 3 “Food for Peace Impact Evaluation of the Kore Lavi Program’s Strategic Objective 2 in Haiti. Baseline Household Survey. Final Report” June 2016. 4 The planned sample size planned for the endline survey was approximately 1,200 households (600 in the intervention areas and 600 in the comparison areas). Introduction 1 IMPEL | Implementer-Led Evaluation and Learning insecurity, BHA decided to first postpone and then later cancel the remainder of the data collection and no longer conduct the final impact evaluation. Once it became evident that data collection in Haiti could be conducted safely if proper COVID-19 protocols were put into place, Tulane and BHA worked together to seize this unique opportunity to build on the data collected from these 276 households in early 2020. They proposed adapting it into a panel study of these households. They would re-interview them again one year later in early 2021 and add complementary qualitative research. Their goal would be to examine the impacts recent shocks (COVID- 19, climatic, social-political) had on poor households and better understand the capacities and strategies poor households have used to mitigate these events. These adaptations resulted in the study presented in this report. 1.2. STUDY OBJECTIVES AND RESEARCH QUESTIONS The objectives of the study are to: ● Assess how recent shocks (COVID-19, climatic, social-political) have affected poor households in three communes in the Northwest and Artibonite departments. ● Understand what capacities and strategies they have used to mitigate these events. ● Generate hypotheses about potential resilience capacities and their relationship to the various shocks, including COVID-19. The study research questions are: 1. How has household food/economic security status changed during the past year as a result of COVID-19 and its sequelae? 2. What shocks have been experienced by activity participant households (socio-political, climatic, COVID-19), and how have they coped with them? 3. What resilient characteristics are present in participant households? 2 Introduction Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) 2. METHODOLOGY This study applied mixed methods, employing both qualitative and quantitative data collection as well as a review of existing data and studies related to food security and recent shocks in Haiti. The study was specifically designed to follow both rounds of quantitative data collection with qualitative data collection, each designed to inform the next data collection exercise and to further explore questions that came up in the initial analysis of the previously collected data. A timeline of the study data collection activities and analysis is shown in Figure 1 below. Figure 1. Study timeline Quant. Round 1 Jan/Feb 2020 Qual. Phase 1 Jan. 2021 Quant. Round 2 Jan/Feb 2021 Quant. Round 2 Jan/Feb 2021 Qual. Phase 2 June 2021 Qual. Phase 2 June 2021 Qual. Phase 2 June 2021 •Preliminary analysis •Identification of potential additional data needs •Analysis to draw hypotheses about impacts of COVID-19, other shocks •Used to inform design of quant. Round 2 •Preliminary analysis to identify areas for further qual. query in Phase 2 •Preliminary analysis to identify areas for further qual. query in Phase 2 •Preliminary analysis to identify areas for further qual. query in Phase 2 •Analysis of data •Analysis of data •Analysis of data •Re-analysis of previous qual. and quant. data as needed. •Re-analysis of previous qual. and quant. data as needed. •Re-analysis of previous qual. and quant. data as needed. •Re-analysis of previous qual. and quant. data as needed. •Reporting •Reporting •Reporting The qualitative portion of the study included Focus Group Discussions (FGDs), Key Informant Interviews (KIIs), and a small number of in-depth semi-structured interviews (see selection in the sampling sections). These were used to gather information about perceptions of shock exposure, coping capacities/resilience capacities, and this small sample’s perceptions on any resilience capacities that enabled them to cope with shocks, including those related to the COVID-19 pandemic. The qualitative data collection was split into two phases (Phase 1 qualitative and Phase 2 qualitative). The quantitative component focuses on the 2020/21 panel data, referred to in this report as Round 1 and Round 2. The quantitative portion of the study focused primarily on the changes in household food security and consumption expenditure indicators, coping behaviors and resilience capacities (assessed retrospectively) between early 2020 and early 2021. These changes will be further complemented by data collected at follow-up Round 2 on shocks, coping, resilience, and psychosocial well-being. These additional modules in Round 2 were informed by the Phase 1 qualitative data. 2.1. ROUND 1 QUANTITATIVE As noted above, Round 1 of the quantitative data collection was conducted before the study focus was shifted from an endline survey for the KL impact evaluation to a stand-alone study on the impacts of COVID-19 and other shocks. Methodology 3 Secondary data review, monitoring IMPEL | Implementer-Led Evaluation and Learning 2.1.1. TOOLS AND INDICATORS The Round 1 household survey questionnaire was adapted from the Kore Lavi Strategic Objective 2 (SO2) baseline study survey instrument. The questionnaire consists of separate modules for the following topics: ● Module A: Household identification and informed consent. ● Module B: Household roster. ● Module C: Household food diversity and hunger. ● Module H: Household expenditures. Additional questions on livelihoods and voucher program exposure/other food-security program exposure were also included in the tool Questions for Modules A through C were adapted using questions from the Food for Peace (FFP) Standard Indicators Handbook and the Demographic and Health Survey (DHS) questionnaire; and questions for Module H from the 2012 Living Standards Measurement Survey. The key indicators are presented in the Round 2 quantitative data collection section further below. 2.1.2. SAMPLING The 2014 KL baseline survey sample was designed to be representative of the poorest 11% of households in the surveyed communes, based on the Haitian Deprivation and Vulnerability Index (HDVI), in both KL intervention communes as well as non-KL intervention communes to serve as the comparison group). To improve comparability between the intervention and comparison area households in the endline, a nearest-neighbor analysis was done using the 2014 baseline data. In the intervention areas, the 10% of households with the most distant nearest neighbor in the comparison areas were identified. The endline sampling frame then consisted of the remaining 90% of households interviewed at baseline. A subset of households in each commune was randomly selected from this list to be interviewed in the endline study. The Round 1 sample of this study consists of those households sampled in three of the nine communes planned to be included for the Kore Lavi endline survey that were able to be interviewed before the pandemic began. To locate the households for Round 1, the surveyors were given the list of the 2014/15 Kore Lavi baseline survey household information provided to Tulane by BHA. This data included the following information: ● Geographic. o Commune (sub-department administrative unit). o Seccion Communal (sub-commune level administrative unit). o Localite (village/community name). o GPS coordinates (for some communes). ● Household. o Names of the household members. o If possible, the name of the person interviewed in the baseline survey. 4 Methodology Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) The survey team spoke with local residents to identify who in the communities could help identify the selected households based on the above information. If a household was not found, a substitute household from the same localite was randomly selected. If the household was found but nobody was home/available for interview, two to three re-visits were made in an attempt to interview the household. If nobody was found available for an interview, a substitute household within the same localite would be selected. However, in Round 1, about 45% of the selected households were not able to be located. In the time between the Kore Lavi S02 baseline survey and Round 1 data collection in early 2020, they may have moved away, dissolved, or the survey teams simply could not locate them with the available information. These households were replaced with the nearest households that had taken part in the Kore Lavi SO2 baseline survey. Ultimately, 276 households were interviewed across the three communes in Round 1. An analysis was conducted to look for bias in the R1 sample. Using baseline data on food security outcomes and basic household demographics, no evidence of meaningful bias was observed in the sub- set sampled in R1 and those not sampled.5 These modifications to the sample (exclusion through nearest neighbor analysis, and large proportion of replacement households) render it technically non-probable, and so not statistically representative of the poorest 11% of households in the communes. However, the lack of evidence of meaningful bias in food security outcomes and basic household demographics make the data still amenable to statistical tests. However, the results should not be considered representative of the poorest 11% of households in the three communes, but rather indicative of these households. 2.1.3. TRAINING AND FIELDWORK The training for the first round of quantitative data collection took place in January 2020. Tulane’s local research partner, IFOS, was responsible for the coordination and fieldwork of the data collection. Tulane supported IFOS in the enumerator training, pre-testing before the start of fieldwork, and conducted regular data quality checks throughout the data collection period. Training took place at the IFOS office conference room in Port-au-Prince. Enumerators traveled to Port- au-Prince to take part in the training. The training was co-led by the Tulane survey expert, IFOS supervisor, and an IFOS hired consultant that had technical knowledge and experience on the collection and analysis of consumption/expenditure survey data in Haiti. The training lasted 6 days, including the pre-test. 5 Using the baseline data to compare households interviewed vs. those not interviewed at Round 1, a small but significant difference in the mean FCS was observed (40 among those not interviewed in Round 1 vs. 43 in those interviewed in Round 1), and a small borderline significant difference in the proportion of households headed by a man and woman (80% in the non- Round 1-interviewed vs. 86% among those interviewed in Round 1). No significant differences were observed in HDDS, HHS, or HH size. Methodology 5 IMPEL | Implementer-Led Evaluation and Learning For training and field guidance, manuals and guidelines were adapted from the baseline survey: 1. Supervisor and interviewer manuals include topics such as objectives of study, survey organization, supervisor roles and responsibilities, rules and regulations, ethics, and quality control. Guidelines for implementing survey and fieldwork procedures, including interviewing techniques and procedures for completing the questionnaires. 2. Question by question detailed explanations includes detailed explanations and instructions for each question in the questionnaire. The Phase 1 field data collection was conducted by one team, made up of seven members: one team leader and six enumerators. The team leader and the survey coordinator at IFOS conducted field supervision. Additionally, IFOS hired an expert in consumption-expenditure survey data collection and analysis in Haiti (Haitian national) to assist in the training and serve as an independent monitor of the fieldwork. Tulane staff conducted overall supervision remotely, including regular checks of the data quality. Data was collected using tablets and smartphones, programmed using Open Data Kit (ODK), which allowed skip patterns and data range checks to be automatically embedded in the application, as well as real-time monitoring of the data quality. Data checks were included. ● Checking the sample: validating interviewed households against the sample (comparison of household member names, locations) and verifying the correct unique household ID numbers were applied to ensure data from baseline and Round 1 were aligned. ● Checking data integrity: verifying unique IDs, comparing data-to-data dictionary of expected values, verifying labels, and checking values to ensure they fall within the valid range. ● Checks of data validity: verifying that skip patterns were correctly observed, no unexpected missing data points, excessive use of “don’t know” or “other,” and distributions of continuous measures to identify outliers. In Round 1, only four households required corrections to their unique IDs. These were all in replacement households, where the enumerator recorded the originally sampled household ID rather than the replacement household ID. Field supervisors kept records of all households that were replaced, so the IDs were also verified against this list. Additionally, enumerators validated that the household interviewed had received Kore Lavi vouchers. Only one household reported they had not received KL vouchers. This household was replaced following field protocol. 2.2. PHASE 1 QUALITATIVE The goal of the first round of qualitative data collection was to gain a basic understanding of the perceptions and main effects of COVID-19 and other recent shocks on the poor households in the communities being studied. This study also informed adaptations to the quantitative survey tool used in Round 2 (early 2021). 6 Methodology Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) 2.2.1. OBJECTIVES AND TOOLS Objectives of Phase 1 qualitative: ● To better understand how COVID-19 is perceived by the population (as a shock, as an issue not impacting their communities, etc.), how they understand the effects of COVID-19, the impacts on their lives/livelihoods, and how they have responded to these impacts. ● To understand the changes in household resilience over the past 5 years, including farming practices, spending habits and other characteristics, and what projects may have helped build household resilience. ● To help inform the next steps of the data collection as part of the assessment on the impacts of the Coronavirus pandemic and other recent shocks on the poor households in the areas being studied. The interview guide for Phase 1 Qualitative is found in Annex A. 2.2.2. SAMPLE SIZE AND SELECTION In Phase 1 of the qualitative data collection, 15 key informant interviews (KIIs) were conducted, with a total of 16 key informants (KIs) (one interview was done with two informants together). Five interviews were conducted in each of the three communes. KIs included staff from the Mayor’s office, local leaders or “notab,” health workers, community development council members, community organization members, and small merchants. These key informants were identified by IFOS based on their contacts, drawing from the contacts used in the mid-term evaluation qualitative data collection. 2.2.3. TRAINING AND FIELDWORK The training for the first phase of qualitative data collection took place January 9–10, 2021. Tulane participated virtually (via Zoom), and IFOS led the in-country training and the practice sessions with the field interviewers. Tulane also held calls with the IFOS staff and the survey coordinator before, during, and after the training to ensure they had a full understanding of the methodology, tools, and purposes of the data collection. Data collection took place from January 11 to January 15, 2021. There were four field team members. They worked in pairs, with one note taker and one interviewer. The interviews were audio-recorded after securing a verbal informed consent from the participant(s), and the teams took extensive notes during and after each interview which were used to prepare a separate report for each interview. IFOS supervised fieldwork remotely (from Port-au-Prince). The survey coordinator also served in the field as one of the interviewers. Tulane remained in contact with the teams on a regular basis via WhatsApp. After the first day of interviews, Tulane held a phone call with the team leader to get initial insights on findings so that Tulane could provide feedback, such as additional lines of inquiry, clarifications of any of the tools, and discussion of the KI selection to ensure various information sources and viewpoints were accounted for. Methodology 7 IMPEL | Implementer-Led Evaluation and Learning 2.3. ROUND 2 QUANTITATIVE 2.3.1. TOOLS AND INDICATORS The second round of follow-up household data collection used the Round 1 questionnaire with only small modifications designed to address lines of query identified in Phase 1 of the qualitative and to assess the impacts of COVID-19 and other shocks. Modules added to the Round 2 tool include: ● General shock exposure and coping. ● COVID-19 specific shock exposure and coping. ● Perceptions of COVID-19 impacts (general, on livelihoods, on food access). ● Access and use of credit. ● COVID-19 knowledge, attitude, and practice. ● Psychosocial well-being. The household roster module was abbreviated in Round 2. The food security and consumption/ expenditure modules remained unchanged from Round 1. The quantitative survey instruments are found in Annex B. The quantitative survey instruments gathered a wide range of data. Key indicators focused on in the analysis include, but are not limited to, those in the list below. Indicators added to the Round 2 quantitative data collection are noted. Food Security Outcomes: ● Household Hunger Score (HHS). ● Food Consumption Score (FCS) and Food Consumption Score Groups. ● Household Dietary Diversity between Jan/Feb 2020 and the date of survey implementation. Economic Security Outcomes: ● Prevalence of poverty (below $1.90 per capita per day). ● Depth of poverty. Other high-level outcomes: ● Psychosocial well-being status–General Health Questionnaire 12 (GHQ-12). (Round 2 only). Other key indicators: ● Access to healthcare (Round 2 only). ● Livelihoods. ● Agricultural assets (Additional data in Round 2 only). ● Drinking and irrigation water access (Round 2 only). Coronavirus Specific measures: ● Knowledge, Attitude, Practice (KAP) related to Coronavirus (Round 2 only). ● Coronavirus specific shocks/coping (Round 2 only). Additionally, data from other modules was analyzed, and used to describe household characteristics and practices to further describe resilient characteristics where possible. 8 Methodology Study of the Impacts of COVID-19 and Other Recent Shocks in Haiti (Vol. I) 2.3.2. SAMPLING To conserve the panel design, the same 276 households surveyed in Round 1 (January/February 2020) were targeted to be re-interviewed in Round 2 (February 2021). Initial discussions with the local research partner (IFOS) indicated a strong likelihood that they would be able to locate these households for interview. As seen in the table below, of the 276 households interviewed in Round 1, 92% (255 households) were located and consented to be interviewed in Round 2. Table 1. Quantitative Round 1 & 2 data collection dates and sample sizes Dept. Commune Quant. Data collection dates- Round 1 Quant. Data collection dates- Round 2 Follow-up sample – Round 1 Follow-up sample – Round 1 Follow-up sample – Round 1 Follow-up sample – Round 2 Follow-up sample – Round 2 Follow-up sample – Round 2 Follow-up sample – Round 2 Methodology 9 Northwest Bombardopolis Jan/Feb 2020 Feb/March 2021 90 83 Mole Saint Nicolas Jan/Feb 2020 Feb/March 2021 90 86 Artibonite Anse Rouge Jan/Feb 2020 Feb/March 2021 90 86 Total 276 255 2.3.3. TRAINING AND FIELDWORK Because of the restrictions due to COVID-19, the team size for the Round 2 quantitative data collection was reduced to four enumerators and one supervisor. Training took place in Port-au-Prince, with Tulane supporting remotely. Because IFOS’s staff had the experience from Round 1, the training was considered a “refresher,” with a focus on the new or modified indicators as well as the additional COVID-19 safety protocols. As in Round 1, data was collected using table