Mapping the Ecosystem of Education Data for Internally Displaced Persons in the Middle East and Beyond: Issues, Challenges, and Recommendations January 2021 This publication was prepared independently by Daniel Shephard, Elizabeth Buckner, Anne Smiley, and Rachel Hatch of FHI 360 on behalf of Social Impact, Inc. at the request of the United States Agency for International Development. This publication is part of the Middle East Education Research, Training, and Support (MEERS) activity (AID-OAA-I-14-00075, AID-OAA-TO-17-00022). TABLE OF CONTENTS Acknowledgements...........................................................................................................................................................2 Executive Summary ...........................................................................................................................................................3 1. Data mapping ........................................................................................................................................................3 2. Factors Influencing Data Collection ................................................................................................................3 Conceptual.................................................................................................................................................................3 Technical/Methodological .......................................................................................................................................4 Institutional.................................................................................................................................................................4 Political........................................................................................................................................................................4 3. Recommendations...............................................................................................................................................4 Introduction ........................................................................................................................................................................6 Historical Context........................................................................................................................................................7 Methods...............................................................................................................................................................................8 Findings...............................................................................................................................................................................10 1. Descriptions of Literature and Participants.................................................................................................10 2. Mapping IDP Educational Data Sources........................................................................................................11 2.1 Global Data Aggregator: Internal Displacement Monitoring Centre (IDMC)..............................................11 2.2 Primary Data Collection by National Governments........................................................................................12 2.3 Primary Data Collection by International Organizations................................................................................13 2.4 Other Primary Data Collection Efforts.............................................................................................................22 2.5 IDP Education Data Snapshot for Iraq, Syria, and Yemen............................................................................23 3. Factors Influencing Data Collection ..............................................................................................................24 3.1 Conceptual Factors.............................................................................................................................................24 3.2 Technical/Methodological Factors.....................................................................................................................25 3.3 Institutional Factors............................................................................................................................................26 3.4 Political Factors...................................................................................................................................................32 Conclusion.........................................................................................................................................................................35 Recommendations...........................................................................................................................................................37 Governments, as appropriate for their contexts, should consider:...........................................................37 International organizations and civil society should consider:.....................................................................37 Donors should consider: ......................................................................................................................................38 Appendix 1: Interview Guide........................................................................................................................................39 References.........................................................................................................................................................................42 TABLES Table 1: Example educational data from IDMC primary data collection in Somalia .......................................11 Table 2: Government IDP Data Sources...................................................................................................................12 Table 3: Existing Education Relevant IDP Data Sources from International Partners....................................13 Table 4: DTM MSLA Summary Coverage, Strengths, and Limitations ...............................................................16 Table 4: REACH MSNA Summary Coverage, Strengths, and Limitations.........................................................18 Table 5: Education Cluster JENA & 5Ws Summary Coverage, Strengths, and Limitations...........................20 Table 6: JIPS Summary Coverage, Strengths, and Limitations...............................................................................22 Table 7: Other Data Sources with Potential IDP EiE Relevance..........................................................................22 Table 8: IDP Education Data Sources for Iraq, Syria, and Yemen .......................................................................23 Table 9: Similarities and Differences in IDP Conceptualization............................................................................24 FIGURES Figure 1: Respondent Country Coverage..................................................................................................................10 Figure 2: Example Educational Data from Iraq's 2018 Location Assessment (by Governorate)..................15 Figure 3: Example Educational Data from Syria's 2014 Multi-Sectoral Needs Assessment...........................17 Figure 3: IDP Education Data Collection Coverage (2015-2020), DTM & REACH........................................18 Figure 5: Example Educational Data from JENA in Aden, Yemen (2016)..........................................................19 Figure 5: Example Educational Data from Yemen's Education Cluster 5Ws....................................................20 Figure 7: Example Educational Data from the 2016 Iraq Urban Profiling by JIPS.............................................21 1 ACRONYMS DTM Displacement Tracking Matrix EiE Education in Emergencies EGRIS Expert Group on Refugee and IDP Statistics EMIS Educational Management Information System GEC Global Education Cluster GPE Global Partnership for Education HCT Humanitarian Country Team HDX Humanitarian Data Exchange HHS Household Survey HNO Humanitarian Needs Overview IASC Inter-Agency Standing Committee’s IDMC Internal Displacement Monitoring Centre IDPs Internally Displaced Persons IIEP International Institute for Education Planning INEE Inter-Agency Network for Education in Emergencies IOM International Organization for Migration IRIS International Recommendations on IDP Statistics JENA Joint Education Needs Assessment JIPS Joint IDP Profiling Service KII Key Informant Interview MEERS Middle East Education Research, Training, and Support MICS Multiple Indicator Cluster Survey MSLA Multi-Sector Location Assessment MSNA Multi-Sector Needs Assessment NRC Norwegian Refugee Council ODI Overseas Development Institute UNHCR United Nations High Commission on Refugees UNOSAT United Nations Operational Satellite Applications Programme UNRWA United Nations Relief and Works Agency for Palestine Refugees in the Near East USAID US Agency for International Development 2 ACKNOWLEDGEMENTS This resource is made possible through the US Agency for International Development (USAID) Middle East Education Research, Training, and Support (MEERS) program. The contents of this report were developed by FHI 360, a prime to Social Impact on the USAID MEERS program. The team at FHI 360 thanks Christine Capacci-Carneal, USAID and Andrew Epstein, Social Impact for their initial support of the concept and reviews of subsequent iterations of the study. We are also immensely grateful to our reviewers from a wide array of organizations, who provided invaluable feedback throughout the process, including: Anjuli Shivshanker, USAID; Nina Papadopoulos, USAID; Christelle Cazabat, IDMC; Sebastien Hine, Independent Consultant; Camille Lailheugue, UNESCO; Luke Pye, UNESCO; Matteo Valenza, UNICEF; Anna Seeger, UNESCO; Patrick Montjourides, NORRAG; Craig Williams, JIPS; Nicolas Servas, UNICEF; Chloe Goldthorpe, REACH; Lorenza Rossi, IOM-DTM; and Jean Claude Ndabananiye, UNESCO. 3 EXECUTIVE SUMMARY There were over 50 million people around the world experiencing internal displacement in 2019, with nearly 12 million of these in the Middle Eastern countries of Iraq, Syria, and Yemen. Unfortunately, a lack of systematic data on internally displaced persons (IDPs) can exacerbate the educational challenges of IDP children. This paper demonstrates that there are more educational data sources for IDPs than most practitioners are aware and make use of. However, the IDP educational data landscape—and the IDP data landscape more broadly—lack consistent systems which reduce the utility of existing data for local, national, and cross-national use. Through a review of the literature and interviews with sector experts, we have mapped existing data sources and identified factors that can help or hinder the creation of a more systematic educational data collection system for IDPs. We have organized these factors into four types: conceptual, technical/methodological, institutional, and political. DATA MAPPING At the global level, the most important data aggregator for tracking the number of IDPs globally is the Internal Displacement Monitoring Centre (IDMC). They rely on a range of primary data sources coming from national governments and international organizations. However, because there is limited data on education from IDMC the following data mapping focuses on primary data collection efforts that contain educational data. Since IDPs remain in their countries of origin, one of the most important potential sources of education data remains the country’s Educational Management Information System (EMIS) and other national surveys. However, such systems need to be adapted in important ways to be responsive to the needs of IDPs and often need to be supplemented by data collection activities of international partners and civil society organizations. In addition to government sources, there are several important primary sources of IDP data collected by international partners that can inform educational programing: • IOM-DTM’s Multi-Sector Location Assessment (MSLA) • REACH’s Multi-Sector Needs Assessment (MSNA) • The Education Cluster’s 5Ws & Joint Education Needs Assessment (JENA) • Joint IDP Profiling Service (JIPS) In the mapping, we note other pertinent data sources that are less frequent, less systematic, or less accessible than those mentioned above. FACTORS INFLUENCING DATA COLLECTION The factors that can influence the collection of education data for IDPs are conceptual, technical/methodological, institutional, and political. CONCEPTUAL How data collectors define IDP can affect data comparability. There are three main dimensions of the accepted normative definition: that the person left their habitual place of residence without crossing an international border; that this movement was not their choice, but rather that it was forced on them by man-made or natural disasters;1 and that they have not found a durable solution through sustainable 1 As the climate crisis continues to affect people’s movements, so-called “slow onset” natural disasters are likely to become an increasingly relevant type of displacement that will need to be addressed by data collectors. 4 (re-)settlement and integration in a new place or through return and reintegration in their habitual place of residence.2 Across contexts, there are variations in the location, reason, and timing dimensions of this definition: 1. Location: There are no standard within-country thresholds for a move to be captured as displacement. 2. Reason: There is no standard method for determining if a displacement is “forced.” 3. Timing: There are different temporal cut-off points to determine how long someone counts as displaced and how quickly they become classified as having a durable solution. TECHNICAL/METHODOLOGICAL The conceptual factors noted above influence two technical/methodological elements of educational data collection for IDPs. The third technical factor is unrelated to the conceptual factors. 1. The geographic precision of data collection instruments can alter the identification of IDPs and when they become returnees. 2. Different data collection methods of determining IDP status can result in different data including self-identification, location proxies, and screening questions. 3. Data collectors may lack a sampling frame which prevents the collection of representative data on educational indicators for IDPs. INSTITUTIONAL The institutional factors that can hinder or help educational data collection for IDPs relate to the institutional structure of national government ministries, international and civil society organizations, and how they interface with each other. POLITICAL The countries with the highest numbers of IDPs in need of education in the Middle East and globally are countries impacted by conflict. Political factors that always affect data collection will be even more salient in such contexts. However, such political factors are not limited to governments and intergroup conflicts, but also extend to inter-organizational politics in the international community. RECOMMENDATIONS We provide a series of recommendations based on our analysis of the literature and of the interviews with respondents. These are organized as recommendations for governments, the international community, funders, and all stakeholders. Here we highlight several key recommendations: Government should consider: 1. Creating an interoperability data standard between the Ministry of Education, the ministries in charge of emergency response and displacement tracking, and humanitarian partners. 2. Implementing an EMIS system that is more dynamically updated at the school-level to inform school-level practice for students experiencing high mobility. 3. Develop an educational sector plan to improve the resilience of the educational sector’s ability to handle displacements, including its data system. Yemen’s experience can provide an example. International organizations and civil society organizations should consider: 2 For more details on durable solutions for IDPs see the IASC framework: 2. IASC, Brookings, and Universitat Bern, IASC Framework on Durable Solutions for Internally Displaced Persons. 2010, Washington, DC: Brookings. 5 1. Empowering a single multi-lateral agency to lead efforts in coordinating and harmonizing IDP data collection efforts and system strengthening to support national authorities. 2. Improving the capacity of staff to collect and use educational data to inform their work with IDPs and host communities. 3. Supporting community-based data collection efforts that can inform the education of IDPs, such as the work of the PAL network and the community-based EMIS experience of Save the Children in Nepal and Sri Lanka. Donors should consider: 1. Incentivizing organizations to adhere to internationally established data standards. 2. Providing more proportionate funding for educational services and data during humanitarian crises. 3. The full set of recommendations can be found at the end of the report. 6 INTRODUCTION In this working paper, the authors outline existing large-scale data sources that can inform educational programming and policy for IDPs and explore how they might address challenges and find opportunities to improve the data landscape for the education of IDPs. There were over 50 million people around the world living in conditions of internal displacement in 2019, with nearly 12 million of these in the Middle Eastern countries of Iraq, Syria, and Yemen [3]. 3 Such large-scale displacements have disrupted the educational trajectories of millions of children [3-5]. Unfortunately, a lack of systematic data on IDPs can exacerbate the educational challenges of IDP children [6, 7]. In response, an increasing number of governments and the international community have devoted more efforts to collect data to support the education of IDPs. The United Nation’s Guiding Principles on Internal Displacement defines internally displaced persons as those “persons or groups of persons who have been forced or obliged to flee or to leave their homes or places of habitual residence, in particular as a result of or in order to avoid the effects of armed conflict, situations of generalized violence, violations of human rights or natural or human-made disasters, and who have not crossed an internationally recognized State border” [1]. Until the 1990s, IDPs received little attention from the international community compared to refugees. This situation began to change in the 1990s with publication of the United Nations’ Guiding Principles on Internal Displacement in 1998 which calls for the systematization of data collection on IDPs [1, 7]. While gaps in IDP educational data remain, calls for improved data for IDPs continued into the first two decades of the twenty-first century [8, 9]. In 2019, the Geneva Education in Emergencies Data Summit called for “further efforts […] to include those that are absent from the global statistical picture, such as IDPs” and agreed to “increased prioritization of IDP education data” [9]. This lack of systematic data and poor awareness of existing data on education for IDPs continue to hamper the realization of their right to education [6, 10-13]. This educational data gap is particularly pertinent in Iraq, Syria, and Yemen, which together account for nearly 25 percent of all IDPs globally. Fortunately, there are an increasing number of good practices that can be built upon to fill in some of these gaps. In the Middle East region and globally, there are multiple data collection efforts that could be better utilized to inform educational programming and policy for IDPs, but a lack of awareness about these sources of data presents a challenge. To that end, this study will address the following questions: 1. How are IDP education data collected and used in the Middle East and other crisis-affected contexts globally? 2. What are the factors that support or constrain collecting and using IDP education data in the Middle East and globally? 3. How can various stakeholders improve the collection and use of data to support the education of IDPs in the Middle East and globally? 3 IDMC data on the total number of IDPs includes those displaced due to conflict/violence and natural disasters. 7 HISTORICAL CONTEXT Improving the education and systematic data collection for IDPs have been integral to the guidance that has emerged since the 1990s. Principle 23 of the Guiding Principles on Internal Displacement calls for ensuring that IDPs receive education. Accompanying resolutions highlighted the need for a “more comprehensive and coherent system of collecting data” [1, 7]. That same year, the Internal Displacement Monitoring Centre (IDMC) was founded as part of the Norwegian Refugee Council (NRC) [14], with a mandate to aggregate and report on the number of IDPs globally. Since 2017, IDMC’s work has expanded to conducting research on the drivers and effects of internal displacement, including the role of education [14, 15]. International organizations and networks increased their focus on both Education in Emergencies (EiE) and IDP data collection in the 2000s. In 2000, the Inter-Agency Network for Education in Emergencies (INEE) was founded to create a single network for sharing global public goods related to EiE. In 2004, INEE developed the influential Minimum Standards for Education, which included the importance of recognizing IDP children and the use of disaggregated educational data to inform decisions [16]. In 2004, the International Organization for Migration (IOM) initiated the Displacement Tracking Matrix (DTM) to assess the situation of IDPs in Iraq and its methods, scope, and geographic reach have expanded since [17]. In 2007, the Global Education Cluster (GEC) was integrated into the Inter-Agency Standing Committee’s (IASC) cluster approach to humanitarian work. The Educational Clusters at national level coordinate implementation and data collection efforts in crisis contexts in which the cluster system is activated. The REACH initiative was created in 2010 as a joint initiative of IMPACT Initiatives, ACTED, and the United Nations Operational Satellite Applications Programme (UNOSAT) in order to provide evidence on needs for humanitarian decision making, in many cases this includes IDPs and educational needs [18]. In 2009, the Joint IDP Profiling Service (JIPS) was created as an inter-agency service to support governments, humanitarian and development organizations in improving locally owned data and analysis on displacement contexts [8, 19]. The importance of improving data for decision making in humanitarian settings, including for IDPs, was further emphasized in 2016 and 2017. In 2016, the Grand Bargain was launched calling for specific reforms to humanitarian aid, including support for IDPs. Among the reforms was a commitment to using “a shared open-data standard”, a “common digital platform,” and to “improve joint and impartial needs assessments” [20]. Resulting efforts have included investing in capacity building for data collection, including efforts by OCHA, JIPS, and IOM-DTM [8, 21]. Similarly, in 2017, the New Way of Working, published by OCHA included a specific call to address the needs of IDPs, address barriers to education, and jointly harmonize data collection and use [22]. In 2019, the Education in Emergencies Data Summit in Geneva generated a commitment by participating organizations to improve the prioritization of IDP education data [9]. In 2020, the Expert Group on Refugee and IDP Statistics (EGRIS) developed a background paper for the United Nations Statistical Commission on “International Recommendations on IDP Statistics” which mentions the need to collect IDP education data [23] but does not provide detailed guidance on education-related data collection. This year also saw the creation of the Alliance for Data on Children on the Move in March with a primary objective to “to improve statistics and data on migrant and forcibly displaced children.”4 4 A description of the alliance can be found here: https://bit.ly/idpdataalliance 8 METHODS This study combined a desk review with semi-structured interviews to map existing practices regarding education data collection for IDPs and identify good practices and challenges relevant for the Middle East and globally. Due to our regional and global focus, we reflect on the relevance of findings for the Middle East and globally throughout the paper. For the desk review, both academic and grey literature were reviewed. Academic literature was identified through a structured search of the ERIC and Sociological Abstracts databases using ProQuest and search terms to identify literature at the intersection of IDPs, education, and data.5 Grey literature was identified by reviewing the online publications of IDMC, UNHCR, GEC, IOM, REACH, UNICEF, and UNESCO. In addition to this desk review of the literature, interview participants also shared relevant documents that arose during the interviews. 6 In total we screened over 100 articles and identified over 50 of relevance. Semi-structured interviews lasting 30 to 60 minutes were conducted with 31 professionals and researchers with expertise related to IDP data collection or IDP educational program implementation.7 The interview was divided into questions for users of data and collectors of data. All respondents answered questions about what the priorities are for improving such data, what challenges exist for finding and using data, and finally asked them to share good practices. The questions for implementers also asked where they find data to inform educational programming for IDPs. For data collectors, the interview also covered questions about their methodology for data collection and sharing and other data sources they would suggest. Our sampling strategy began with the major IDP education data collectors and aggregators globally to learn about the challenges and opportunities associated with each set of methodologies, while also seeking out examples of innovation and good practices at local levels. Using a mixture of purposive and snowball sampling, we included respondents working at global and national levels; working with multilateral agencies (e.g., UNESCO), non-governmental organizations (e.g., Save the Children), and universities; with coverage across multiple regions. To ensure the maximum relevance for the Middle East, we sampled more participants with experience in the Middle East (n=10) than from each of the other regions. Meanwhile, to ensure our ability to draw from good practices and learn from challenges experienced globally, we included respondents from Africa, Asia, Latin America, and individuals working at a global level. We used semi-structured interviews to enable us to both capture data that directly addresses questions identified by us as researchers and to explore new topics of relevance that are brought up by our respondents. Similarly, our mixture of purposive and snowball sampling ensured we covered several pre-identified perspectives of importance while also allowing the research to expand iteratively based on participant input. The study followed the following analytical steps for the literature review and interviews. For the literature review, first, one researcher screened the results for potential relevance by investigating their title, abstract, and when necessary the full text. Key information regarding IDP educational data was extracted into an excel database based on the various resources. The organization of the database was 5 The search was conducted on February 21, 2020 using the following search string: (internal* NEAR/1 displac*) AND (MAINSUBJECT.EXACT("Education") OR ABS(educat* OR school* OR learn*)) AND (data OR MAINSUBJECT.EXACT("Data") OR evidence*). 6 One limitation of this latter method was that we did not ask every participant to share relevant documents but only did so when such documents were referenced during the interview. 7 We do not provide a detailed list of respondent affiliations in order to maintain respondent anonymity. Interview respondents were assured of such anonymity in order to increase their comfort in providing candid and complete responses and reduce any risk of adverse impacts on their professional lives. 9 organized primarily around data sources, challenges, and opportunities. From this information, a summary document was written about IDP education data broadly and each identified IDP educational data source specifically. These summary documents informed how interviews were conducted (e.g., which probes were used), analyzed, and presented in the final report. The bulk of the introductory content of the final report is derived from this desk review. For the interviews, memos containing both analytical and descriptive elements were produced after interviews. The research team met periodically to synthesize memos, document emerging themes, and identify extant gaps that could be filled by additional interviews. From this process, a coding framework that was iteratively adapted throughout the interviews. Finally, a final codebook as agreed to and this coding was applied to all transcripts by the first author. This final coding approach most closely resembled a mixture of pattern coding and focused coding [24]. This study contains several limitations that should be noted for the literature review and the interviews. The literature review search was limited to two academic databases and screening was done by a single interviewer. Third, the lack of peer-reviewed literature combined with primarily descriptive grey literature limited the insights that could be derived from the literature review. Several limitations of the interview component of the study should also be noted. First, we recruited and interviewed respondents in English which may have limited our access to specific types of respondents and perspectives.8 Second, our study prioritized respondents from the Middle East and thus this study could be complimented by deeper dives into other regions. Third, while all authors contributed to open coding, developing memos, and the creation of the code-book—final coding was completed by a single author. While this approach may enhance consistency, it may also have limited the emergence of different conceptualizations of the application of codes by different authors. Despite these limitations, this study constitutes the first rigorous investigation of the issues facing the collection and use of data to support IDP education as identified in the literature and across multiple actors and contexts around the globe. 8 The research team had access to Arabic, French, Spanish, and English speakers, but all interviews were done in English—likely because of respondent recruitment outreach was done through English-speaking networks. 10 FINDINGS This section provides the findings from the study organized into four parts. The first section describes the documents identified and participants interviewed. The second section provides an overview of the key identified educational data sources for IDPs. The third section highlights challenges facing the educational data ecosystem for IDPs. Finally, we end by providing an overview of several promising practices and opportunities that can shed light on improving data collection and use to improve the understanding of the educational needs and experiences of IDPs. DESCRIPTIONS OF LITERATURE AND PARTICIPANTS The academic literature on education data for IDPs is sparse. The structured search returned 119 academic papers of which only nine were deemed relevant after screening their titles and abstracts. From the grey literature, we identified 46 documents of relevance. The grey literature was published by various international and multi-lateral organizations (e.g., IDMC and UNICEF) and white papers generated from projects (e.g., the Brookings-Bern Project on Internal Displacement). Interview participants were well-balanced between males and females (41% female). They had substantial relevant experience, with 75 percent working in the education sector for more than four years and more than three years in EiE. Most respondents worked at a global level (58%) and those who worked nationally were predominately from the Middle East. The interviews were fairly balanced between those who generate data (n=16) and those who use data (n=21). 9 The respondents came from 19 countries or territories (see Figure 1). Figure 1: Respondent Country Coverage 9 The totals sum to more than the total number of interviews because five respondents answered that they are both data generators and users. 11 MAPPING IDP EDUCATIONAL DATA SOURCES This section provides a brief overview of the most prominent data collection efforts related to IDPs and education organized into four parts. First, an introduction to IDMC, which is the most important IDP data aggregator and has recently conducted some limited primary data collection, including on education. Second, we briefly discuss national government data collection systems, focusing on Educational Management Information Systems (EMIS) and Registries. Third, we introduce four important IDP education data collection efforts by international organizations. Finally, we provide a non-exhaustive list of other data sources that have less systematic coverage in the Middle East and globally. GLOBAL DATA AGGREGATOR: INTERNAL DISPLACEMENT MONITORING CENTRE (IDMC) The most important secondary data aggregation source is IDMC. As noted above, IDMC has the mandate to collect and aggregate IDP data at a global level. The data reported by IDMC is aggregated at the national level and published in a yearly report and a mid-year update in September although the online database is updated internally more often. 10 Their focus has been on aggregate population figures which are primarily used to track changes in the number of IDPs and to conduct national, regional, and international advocacy for IDPs. While IDMC’s role remains primarily that of aggregation, triangulation, and analysis; it has started conducting primary data collection on specific, policy-relevant topics. This has included studies on the education of IDPs. For example, IDMC published results from primary studies in four contexts that included data on the education of both IDPs and host communities [25], we have included two example figures from the Somalia case study below. Given this experience, IDMC plays an important role in supporting the aggregation of IDP education data and providing targeted insights. Table 1: Example educational data from IDMC primary data collection in Somalia Percentage of displaced children going to school before and after displacement Percentage of children in the host community going to school in 2017 and in 2019 Note: Data and graphics from IDMC (2020). Measuring the costs of internal displacement on IDPs and Hosts: Case studies in Eswatini, Ethiopia, Kenya and Somalia. Geneva: IDMC. 10 The data are available on their website at https://bit.ly/idpdatainternal-displacement 12 PRIMARY DATA COLLECTION BY NATIONAL GOVERNMENTS The primary national data source for education should be its National Education Management Information Systems (EMIS). However, crisis-affected countries may not have a functioning EMIS system or it may have ceased to function properly during the crisis. There are some examples of crisis-affected countries effectively using EMIS to track IDP educational access and performance. For example, UNESCO has supported Syria’s transition from a manually entered EMIS using an annual school survey to an operational EMIS using a school-based information management system that is dynamically updated and linked to individual student and teacher identification numbers. This system includes data on students’ movements which are used as a proxy of IDP status along with access, performance, and well-being data. The system is currently deployed in Damascus with efforts to expand it more broadly. South Sudan also provides lessons learned as they have developed an EMIS system that includes the collection and reporting of attendance using mobile phones. While not a national entity, the United Nations Relief and Works Agency for Palestine Refugees in the Near East (UNRWA) has its own EMIS for UNWRA-run schools, which provides helpful lessons for large-scale educational data systems in a context with parallel educational provision and periodic internal displacement. UNRWA’s lessons learned including the development of the system in stages which built local trust over time, the different types of data collected at different intervals with more standardized centralized data collected less frequently and needs based local data collection (e.g., more in-depth local data collected on dropouts when dropout rates exceed a threshold), and methods for tracking students as they move between schools using electronic student identification numbers internally (between UNRWA schools) and using paper-based transfer notices between UNRWA and national schools. These experiences can provide some lessons learned to others. Additionally, national registries of IDP status can be linked to educational data when they exist. Examples of such registries include Colombia’s Victims Unit and Iraq’s registration system. The Colombia example allowed individuals to self-identify as IDPs and the Iraq system links individual registrations to place of residence thus enabling the tracking of displacement. The Colombia case provides potential for learning from their experience linking the registry with other national data systems, including education. However, the system in Iraq is less operational, in part due to many people lacking official identification documents. Unfortunately, these national systems are often unable to respond to educational data needs during large displacement events and therefore there is often a need for international partners and civil society to provide support in educational data collection for IDPs. Table 2: Government IDP Data Sources Source Type Frequency Coverage Initiation IDP Relevance EMIS Facility￾based11 Yearly* National Government May contain migrant status and school moves which can be linked to forced displacement patterns. Registry Self￾registration Ongoing National Government/ Other IDP registers often include data useful for EiE such as: age and student status. Examples include Colombia, Iraq, and Sudan.12 11 Increasingly EMIS are using individual-level identification with dynamic updating instead of periodic facility-level surveys. 12 For example. the Victims Unit registration in Colombia , the Ministry of Migration and Displacement registration requirements , and the DTM Registration in Sudan 13 PRIMARY DATA COLLECTION BY INTERNATIONAL ORGANIZATIONS Table 3 outlines existing education-relevant data sources for IDPs. The two most prevalent and relevant sources for educational data on IDPs are IOM-DTM’s Multi-Sector Location Assessment and REACH’s Multi-Sector Needs Assessment. Each includes repeated data collection in over a dozen countries. Table 3: Existing Education Relevant IDP Data Sources from International Partners Source Type Frequency Coverage Initiation Sample questions / Indicators MSLA (IOM￾DTM) KII National* Annually 20+ countries HCT/OCHA Do children [in this location] have access to primary education? [Yes, No] If yes, which type of education facility? [TLC, FPE] Distance to the nearest education facility? [<1km, 1-2km, 3-5km, 6- 10km, >10km] % of children attending primary learning facility [0%, <25%, 25-50%, 51- 75%, >75%] MSNA (REACH) HHS Affected Areas Random Sample Annually13 12+ countries HCT/OCHA What are your households’ three priority needs? [Education is 1 of 15 options] Do you intend to move to another location? [<2wks, <1month, <3month] Are any individuals (6-17) in your household attending school? [Yes, No] # of children (6-17) in your household who do not attend any form of education is: Why are these children not attending any form of education? Are you able to afford the necessary school supplies for your children to attend school? [Yes, No] JENA (Ed Cluster) KII, FAC, etc. Local* Irregular 5-6 countries yearly Cluster leads or donors How many children [F/M], attended pre-crisis? How many children [F/M], are attending today? Is this school functional? 5Ws (Ed Cluster) IP Reports Local* Monthly (or more) 20+ countries Automatic Activities: Who is doing what? where? when? for whom? Joint-IDP Profiling Service Household survey (HHS) & KII Irregular 5 ongoing countries Government and partners Is the government of the local authority [or any other non-government actors] currently providing education services in the city? [Check all] (KII) Current number of teachers who actually work in this city (on duty) (KII) What is the capacity of the [school]? (OBS) 13 Annual data collection does not occur in all countries. 14 Local* Household members currently attending school (and their level) (HHS) Main reason for children not attending school or not attending regularly (HHS) Vocational training ever attended by household members (HHS) Note: KII = key informant interviews, HCT = Humanitarian Country Team, HHS = household survey, FAC = facility-based, TLC = temporary learning spaces, FPE = formal primary education, IP = implementing partners (humanitarian organizations). F = female, M = male. National and local refer to the approximate sample coverage of the data collection. The national data sources do not necessarily have full national coverage, but they approximate it. The REACH MSNA often aims to cover all the areas of the country with a substantial number of effected individuals. The local data sources usually sample from the areas in which educational implementing partners are active or a requested geographic area in the case of JIP 15 IOM-DTM’s Location Assessments (DTM-MSLA) IOM-DTM’s MSLA collects data from key informants at the area level (e.g., village, subdistrict) in various sectors including education, usually annually.14 The education questions are not standardized and DTM staff are advised to engage with sectoral experts to select relevant questions that meet their needs. Despite this, most MSLAs use almost identical sets of questions based on a standard data dictionary. 15 The questions focus on estimated educational access. This includes indicators on the supply (e.g., whether there are enough facilities and teachers) and on the access (e.g., the estimated proportion of a certain age-group attending school). Trained enumerators from the same area collect data from multiple key informants, such as district education officers or village leaders. The IOM-DTM manages a list of thousands of key informants compiled from government contacts, local leaders, and active civil society organizations. Data validation is done by comparing the estimates of different key informants, the enumerators’ own observations, and the enumerators’ assessment of the reliability of the key informants. The data provided are estimates by key informants who may or may not be sectoral experts. Examples of common questions include whether displaced children have access to primary education, the distance to the nearest educational facility, teacher and learner resource availability, and an approximate estimate of the percentage of children attending school (0%, <25%, 25-50%, 51- 75%, >75%). Data files are not always publicly available; however, data are available upon request and reports for most data collection waves can be found on the DTM website by searching for “location” or “site” assessment. Figure 2: Example Educational Data from Iraq's 2018 Location Assessment (by Governorate) Note: Data and graphic are taken directly from the Iraq Integrated Location Assessment III (p. 39) [26]. 14 The DTM collects other data, for example on the number of IDPs, more regularly. The other sectors are those from the cluster system. 15 The data dictionary can be found here: https://bit.ly/idpdatadictionary 16 In addition to the MSLA, the IOM-DTM tracks the estimated number of IDPs by country subdivision at regular intervals—often five times per year. The frequency of this population-level data can provide more dynamic information for educational planning and for designing education data collection efforts. The DTM-MSLA is useful for informing planning and advocacy. For planning, the DTM-MSLA identifies where there are insufficient educational facilities for IDPs and where large proportions of displaced children remain out of school (at the primary school level). These data are provided at sufficiently precise administrative levels to enable educational service providers to identify areas of greater need for primary education. The trends of change in terms of educational needs and provision by location can inform advocacy on unaddressed needs and to encourage the sharing of “good practices” from locations that have shown improvements in educational access over time. Finally, on a technical level, the DTM’s population-level data collection on the total number of IDPs16 by location can be used to inform the sampling frame of other data collection efforts seeking to provide representative data on IDPs. Table 4: DTM MSLA Summary Coverage, Strengths, and Limitations Strengths Limitations • Summarizes education access • Includes online guidance • Repeated data collection • Triangulates data • Lacks data on barriers/quality • Less precise key informant estimates • Estimates may include host and IDPs • Limited number of indicators on education • Difficult to navigate data/reports Geographic Coverage (Since 2015): Afghanistan, Argentina, Antigua and Barbuda, Burundi, Burkina Faso, Bangladesh, Bahamas, Bolivia (Plurinational State of), Central African Rep., Cameroon, Dem. Rep. of the Congo, Dominica, Ecuador, Ethiopia, Guatemala, Haiti, Indonesia, Iraq, Lao People's Dem. Rep., Libya, Sri Lanka, Mali, Myanmar, Mongolia, Mozambique, Malawi, Nigeria, Nepal, Peru, Philippines, Papua New Guinea, Sudan, Somalia, South Sudan, Chad, Ukraine, Vanuatu, Rep. of Yemen, Zimbabwe. REACH’s Multi-Sector Needs Assessment (MSNA) REACH provides support for a wide array of data collection needs in crisis-affected contexts. However, the most relevant data source on this topic is their Multi-Sector Needs Assessment (MSNA). This is done yearly at the request of UN OCHA and the cluster system by way of the Humanitarian Country Team (HCT) to inform the Humanitarian Needs Overviews (HNOs) and Humanitarian Response Plans (HRPs). In contexts where IDPs constitute a population of concern and are considered in the design of the sample, data from the MSNA can be disaggregated for IDPs. The data are collected at the household level, usually using a probability-based sample, typically through personal interviews using the mobile data collection platform KOBO collect. 17 Due to COVID-19 the data collection methods have been altered in some contexts by relying on a quota-based, non-probability sample including a combination of purposive and snowball sampling of mobile phone numbers and remote interviews. Publications and select data are at the REACH Resource Centre18 and at the Humanitarian Data Exchange (HDX).19 16 This is sometimes referred to using the demographic term of “stock” of IDPs in documentation and reporting. 17 Using KOBO as a mobile data collection platform, see 18 The MSNAs on the REACH Resource Centre can be found at https://bit.ly/idpdatamsnas 19 The Humanitarian Data Exchange enables humanitarian organizations to upload and share data, see https://bit.ly/idpdatahumdata 17 Figure 3: Example Educational Data from Syria's 2014 Multi-Sectoral Needs Assessment Note: Data and graphic are taken directly from the 2014 Syria MSNA (p. 45) [27]. REACH also conducts other types of data collection of relevance to EiE for IDPs when supported by donors, including in-depth education-focused needs assessments, Camp Profile Assessments and Humanitarian Situation Monitoring using key informants (“community representatives”). For example, in Syria, the 2018 in-depth education needs assessment included information on grade repetition, tuition support, and teacher training;20 the 2019 Syrian Camp Profiling Assessments included the priority needs by sector; 21 and the monthly Humanitarian Situation Monitoring provided data on population movements and community-level sectoral needs (including the education sector) according to key informants.22 REACH has additional data collection mechanisms that do not focus on education and thus are outside of the scope this paper. In some cases, there is a brief mention of facility-level data—such as schools being closed or converted into shelters–but education indicators are not the focus. 23 REACH targets information gaps that meet humanitarian needs, as such some data collection may not be undertaken because the data already exist or there is no demand from the humanitarian sector. REACH’s needs assessments tend to provide more precise data when household surveys are used and more detailed data with more educational indicators than are included in the DTM-MSLA, both of which can aid in more precise educational planning. The data often identifies barriers to attending school and breakdowns by type of school provision available (i.e., formal or non-formal). Because the data are typically collected at the household level (although some are conducted with key informants at the community level) the estimates of out-of-school children, intent to move again, barriers to schooling, and access to schooling are more precise. Data users should be aware that the MSNA is representative of only the areas in which it was conducted, and when access is limited it is often limited for all humanitarian actors. Furthermore, in some instances data users may need to request access to data to disaggregate by IDP status or to conduct more fine-grained analysis. 20 See for example the data snapshots by location from 2018 available from REACH < https://bit.ly/idpdataREACHsyria > 21 See for example a rapid Camp Profiling Assessment in Northwest Syria in 2019 22 See for example November 2019 situation monitoring in for Northwest Syria 23 See for example the Murzuq Rapid Situation Overview 18 Table 5: REACH MSNA Summary Coverage, Strengths, and Limitations Strengths Limitations • More precise household level data • Includes barriers to education • Includes IDPs in and out of camps • Annual data collection is aspired to, but not always available • Coverage may be limited by priority, access, and sampling • Difficult to navigate data/reports24 • Limited number of indicators for education Geographic Coverage (Since 2015): Afghanistan, Burkina Faso, Central African Rep., Dem. Rep. of the Congo, Haiti, Iraq, Libya, Mali, Niger, Nigeria, Somalia, South Sudan, Syria, Chad, Ukraine Figure 4: IDP Education Data Collection Coverage (2015-2020), DTM & REACH Note: Includes countries in which educational data on IDPs were collected in the past five years. Light blue = both initiatives collected data; Dark blue = only DTM collected data; Purple = only REACH collected data. Education Clusters’ 5Ws and Joint Education Needs Assessments (JENA) The Education Clusters, co-led by UNICEF, Save the Children and Ministries of Education, conduct periodic Joint Education Needs Assessments (JENA) and collect data on educational service provision through 5Ws and ActivityInfo, which are then reported in UNICEF SitReps and different humanitarian dashboards (ActivityInfo dashboards, OCHA Periodic Monitoring Reports, Humanitarian Dashboards and Humanitarian Insight). 25 JENAs are joint data collection and analysis exercises coordinated by the Education Cluster. The goal of JENAs is to generate comprehensive data in crisis affected areas to facilitate a shared understanding of the needs of schools, learners and their families in displaced settings and other humanitarian contexts. The toolkit for JENAs has recently been updated by GEC and is available on their website.26 However, JENAs are currently conducted in a limited number of countries with Education Clusters (approximately 15% in 2019, 25% in 2020) 27 and do not happen at sufficiently 24 For example, different naming conventions are used, the organization of data by report makes comparison difficult, data and reports are not always both available, and different datasets are available on the REACH website and HDX. It should be noted that this is a shared limitation of all of the data sources covered in this paper. 25 This includes the 5-Ws (Who is doing What, Where, When, and for Whom), the reach of implementing partners’ educational activities compared to targets and the need, as well as other data (see Footnote 28). 26 The updated 2019 guide can be found here: 27 These percentage estimates include JENAs done jointly with REACH as well as JENAs that focus on refugees and not IDPs. 19 consistent intervals to enable the analysis of trends. They are usually conducted upon demand from cluster leads or cluster members. Encouragingly, an increasing number of Education Clusters are conducting JENAs and this data source will hopefully become increasingly comprehensive. Figure 5: Example Educational Data from JENA in Aden, Yemen (2016) Note: The graph is made by the authors and the data are from the 2016 JENA in Aden, Yemen [28]. In addition to JENAs, the cluster members—especially the cluster leads (UNICEF and Save the Children)—often collect more detailed data to support their projects. The Education Cluster Coordinator and Information Management Officer support the harmonization and synthesis of comparable indicators from these individual organization data collection efforts to develop a broader picture of educational needs, provision, and outcomes. Reports from JENAs and other data collection efforts can be found at HumanitarianResponse.Info and on HDX. However, despite the utility of the JENAs, the most up-to-date, consistent, and accessible data from the Education Clusters is their reporting on education activities of implementing partners compared to needs and targets through the 5Ws dashboards, 28 which can be found by exploring their respective Dashboards by navigating to the relevant country page on the Global Education Cluster website. Such data are usually not disaggregated by IDP status; however, in contexts for which IDPs constitute the primary population of concern (such as Iraq, Syria, and Yemen) these data can serve as proxy indicators of educational programming for IDPs. 29 The Education Clusters’ 5Ws25 are useful for planning and monitoring of education service provision— and comparing provision to needs and commitments. Due to the familiarity of the 5Ws in the EiE sector [29], we will not go into depth about it here. 30 JENAs are useful for informing planning, coordination, 28 For Iraq, a brief overview of the dynamic ActivityInfo 5Ws Dashboards can be found here: . An example excel of the types of indicators of . An example excel of the types of indicators of ActivityInfo can be found here: https://bit.ly/idpdataActivityInfoIraq. The Iraq Education Cluster dashboard itself can be viewed here: https://bit.ly/idpdataGECIraq. It should be noted that most Education Clusters do not use ActivityInfo but rather report their 5Ws using PowerBI Dashboards.ActivityInfo can be found here: . It should be noted that most Education Clusters do not use ActivityInfo but rather report their 5Ws using PowerBI Dashboards. The Iraq Education Cluster dashboard itself can be viewed here: https://bit.ly/idpdataGECIraq. 29 In the case of Iraq, data can often be differentiated by IDP status in part because the Iraq cluster response was designed with IDPs as the primary target group. 30 A webinar on navigating humanitarian operations data can be found at INEE https://bit.ly/idpdatawebinar 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Student Attendance (on day of visit) Teacher Attendance (on day of visit) Total Female Male 20 and advocacy among education cluster partners. However, as noted above, JENAs may or may not include IDPs as a category of concern depending on the emergency context. In locations where IDPs constitute a substantial group of persons of concern, JENAs can inform the joint delivery of educational services for IDPs by identifying where needs remain and where service gaps persist. JENAs can also strengthen coordination among various educational service providers. The data collection process itself can help generate more comparable data collection methods among partners and conversations about what data exist and what data remain missing. The analysis of the JENAs can help cluster partners coordinate a joint response—for example by informing the HRPs. Finally, JENAs can inform a shared understanding of the situation and joint advocacy efforts among partners. However, extrapolation from JENAs should be done with care as they are cross-sectional and while some are representative (nationally or of the target areas) others are not. Figure 6: Example Educational Data from Yemen's Education Cluster 5Ws Note: These data were accessed on November 16, 2020 at Table 6: Education Cluster JENA & 5Ws Summary Coverage, Strengths, and Limitations Strengths Limitations JENA • Improve data coordination • Provide detailed data on education needs • Not regularly conducted • Limited geographic scope (depending on type of crisis) • Sometimes not representative (often facility or key￾informant based data) • Difficult to navigate data/reports 5Ws • Regularly updated dashboards • Useful for monitoring activities • Assist in coordinating actions • Restricted to activities • Limited primarily to monitoring • Depends on IP self-reports Geographic Coverage: JENAs (since 2015): Central African Rep., Ethiopia, Nigeria, South Sudan, Ukraine, Vanuatu, and Yemen (Aden) JENAs with REACH (Since 2015): Afghanistan, Indonesia (Central Sulawesi), Iraq, Libya, Myanmar (North Rakhine), Syria 5Ws (September 2020): Afghanistan, Burundi, Chad, Ethiopia, Mali, Nigeria, South Sudan, Syria, Bangladesh, Cameroon, Colombia, Iraq, Myanmar, Occupied Palestinian Territories, Rep. of Yemen, Central African Rep., Dem. Rep. of the Congo, Libya, Niger, Somalia, Ukraine Note: The two data sources have distinct objectives but are both under the purview of the Education Cluster and are therefore grouped together. We separately list JENAs conducted with and without REACH whenever such a distinction was possible. There may be multiple JENAs within a country during the five-year period (e.g., South Sudan). We exclude JENAs that do not include IDPs. The “Occupied Palestinian Territories” are included for completeness without intention to make a statement about their status. 21 The Joint IDP Profiling Service (JIPS) JIPS is a unique data service that designs and coordinates data collection exercises focused primarily on household-level data from IDPs and host communities, at the request of national governments and humanitarian and development actors[19]. Since 2015, JIPS is supporting data collection in a limited number of contexts (see Table 6). The frequency of data collection is country specific. For example, data collection in Syria was undertaken in 2015, 2016, and from 2018 to 2020; meanwhile, in Yemen, data collection among IDPs was only conducted in 2010. The data collection is preceded by intensive stakeholder and community engagement to ensure buy-in and to develop the data collection tools. This is then followed with brief household enumeration surveys to map the population and a more extensive sample-based household survey, complemented by qualitative methods when needed. The methodology is presented more extensively through their online toolkit. 31 Examples of the work of JIPS can be found covering Syria (2018 to 2020) and Iraq in 2016. 32 JIPS also provides capacity building support and tools including Urban Profiling Guidance and several of the JIPS datasets can be explored online. 33 Figure 7: Example Educational Data from the 2016 Iraq Urban Profiling by JIPS Note: Data and graphic are taken directly from the 2016 Iraq Urban Profiling in Erbil (p. 47) [30]. The JIPS data are particularly useful for planning, coordinating, and policymaking. For planning, JIPS generates high-quality, detailed data on the needs, barriers, and resources of IDPs. This can identify where additional educational services are needed. From a coordination perspective, the process JIPS facilitates ensures that governments, UN agencies, and NGOs participate and thus helps to generate coordinated data collection and more standardized indicators across stakeholders. From a policymaking perspective, the JIPS data collection process ensures stakeholder engagement with the data collection 31 The toolkit can be found here: https://bit.ly/idpdataJIPStoolkit 32 The Iraq example can be found here: https://bit.ly/idpdataJIPSiraq. The Syria example is here: https://bit.ly/JIPSSyria 33 Urban Profiling Guidance can be found here: https://bit.ly/idpdataJIPSurbanprofiling and their dynamic reporting tool here: https://bit.ly/idpdatareporting 22 exercise from the beginning and generates data that is representative of the area of concern. However, JIPS data are often location specific—depending on the coverage—local policy inferences may be more appropriate than national policy inferences. Limitations Table 7: JIPS Summary Coverage, Strengths, and Limitations Strengths Limitations • Covers needs, access, & barriers • Strong stakeholder engagement • Mixed methods (HHS, KII, FGDs, etc.) • Currently has limited country coverage • Often has limited geographic coverage • Most data are cross-sectional making trend analysis difficult • Difficult to navigate data/reports Geographic Coverage (Since 2015): Afghanistan, Central African Rep., Honduras, Iraq, Mali, Myanmar, Philippines, Sudan, El Salvador, Somalia, Serbia, Syria, Ukraine, Kosovo Note: HHS = household surveys, KII = key informant interviews. This only includes data collection activities related to IDPs. OTHER PRIMARY DATA COLLECTION EFFORTS UNICEF’s Multiple Indicator Cluster Survey (MICS) included a question to determine IDP status in Iraq in 2018. The MICS is a nationally representative survey that provides data to inform national education policies. MICS already includes a module on migration, which was adapted in Iraq to determine if internal migration had been forced—thus qualifying a respondent as an IDP. However, the survey’s sampling frame excluded IDP camps and thus could only generate data on IDPs living outside camps. In addition, many educational (inter)national NGOs also collect educational data on IDPs when they are designing or implementing a project for which such data are relevant. For example, many interview respondents referenced data collected by organizations such as Save the Children (the co-lead of the education cluster), IRC, NRC, Mercy Corps, and others. Such data are often project specific, bound to a specific location, and are often not easily accessible to other organizations and individuals.34 In addition, there are data collection exercises conducted by other clusters, such as protection and health, with useful data on the number, sex, and location of school age IDP children. Table 8: Other Data Sources with Potential IDP EiE Relevance Source Type Frequency Coverage Initiation IDP Relevance UNICEF￾MICS HHS Periodic National Various / Government The 2018 Iraq MICS can be disaggregated by IDP status. MICS contains a module on general migrant status. Other HHS / FAC Various Various Various Migrant status and place of origin questions could be mapped on to displacement patterns to identity likely IDPs. Education data collected in defined IDP camps can represent IDP needs. Note: HHS = household survey, FAC = facility-based. EMIS may be done yearly through a census or more dynamically through electronic data collection at the school-level. 34 This lack of accessibility may be due to a lack of a technical system, an organization’s rules regarding data protection do not allow any sharing, a lack of personnel with a data focus, and a lack of prioritizing such sharing. 23 IDP EDUCATION DATA SNAPSHOT FOR IRAQ, SYRIA, AND YEMEN Table 9 provides an overview of these various data sources that are available for the focal countries for MEERS; namely, Iraq, Syria, and Yemen. Iraq provides a case to learn from as it has the greatest data coverage in terms of data source and years included. In Syria, REACH and JIPS provide the widest range of data while the MSLA data collection exercise are managed by OCHA instead of DTM. Yemen has the least data availability; however, there is an ongoing JENA. Table 9: IDP Education Data Sources for Iraq, Syria, and Yemen Country DTM REACH GEC JIPS Other Iraq 2015*, 2016, 2017, 2018, 2019 2015, 2016, 2017, 2018, 2019, 2020* Dashboard HDX 2015-2016 MICS 2018 Syria 2014, 2018*, 2020 Dashboard HDX 2015, 2016, 2018-2020 EMIS (Damascus*) OCHA MSLA Yemen 2016, 2018 Dashboard HDX 2010, 24 FACTORS INFLUENCING DATA COLLECTION Drawing on the interviews, the following section highlights factors that may hinder or facilitate IDP educational data collection and use. While some factors are unique to IDPs, others are more broadly applicable to crisis-affected contexts and EiE. We use the following typology which emerged from our analysis through a synthesis of memos on all interviews and informed our final codebook: 1. Conceptual factors about how an internally displaced person is defined and operationalized. 2. Technical/methodological factors regarding how data are collected. 3. Institutional factors regarding how systems are structured. 4. Political factors regarding competition for power or status. Some of the insights are unique to IDP educational data. Other insights are applicable more broadly to data collection among IDPs and other displaced groups. The latter insights are equally pertinent for educational data collection and should be considered by actors involved in educational data collection and are therefore included. CONCEPTUAL FACTORS The international community lacks an operational definition of IDPs, even though it shares a normative definition. There are three main dimensions of the accepted normative definition. First, that the person left their habitual place of residence without crossing an international border. Second, that this movement was not their choice, but rather that it was forced on them by man-made or natural disasters. Finally, that they have not found a durable solution through long-term resettlement or return to their habitual place of residence (for a further discussion on durable solutions for IDPs see the IASC framework [2]). While there is general agreement about these three dimensions, the specifics are inconsistent among governments, data collectors, and EiE implementers. This results in different conceptualizations that have real consequences for people in crisis. These differences can be organized into three questions linked to location, reason for migration, and timing. The structure presented combines our own analysis with other frameworks, most notably the work of EGRIS [23]. Table 10: Similarities and Differences in IDP Conceptualization “We have very different methodologies and very different things that we're counting because IDP data—I mean that the founding documents on IDP is basically two and none of them are a very operational. All of them are aspirational, more human rights-based, […] And so in the end, they don't even tell you what to count. […] So we will end up with different figures.” (010) Location (where) Reason (why) Timing (when) Normative Similarity Internal move (within country) Forced move (not volitional) Recent and temporary move (intend to return or move again) 25 Operational Differences What distance is sufficient to be counted as an IDP? For example, the household, neighborhood, municipality, county, state, or region. What type of force is sufficient for the move to be considered forced? For example, potential vs. actual violence, slow vs. rapid onset natural disasters, interpersonal vs. intergroup violence, destruction of homes vs. livelihoods. Is this determined by self￾declaration, by data collectors, or by legislation? What is the cut-off date for displacement status? For example, does IDP status only apply to the first￾generation or also later generations? Is resettlement considered a durable solution based on intent to stay, ability to return, quality of life, or time displaced?35 TECHNICAL/METHODOLOGICAL FACTORS Three technical factors impacting IDP education data collection were highlighted by the literature and participants: geographic precision, status determination, and sampling frames. The first two of these— geographic precision and status determination—are related to operationalizing the conceptual definition of IDPs for data collection. The third relates to developing a representative survey when there is no census or other traditional sampling frame from which a random sample can be drawn. Several of these issues affect IDP data beyond the education sector; however, respondents noted the importance of these issues in the context of education data. The geographic precision of data collection instruments affects the identification of IDPs. The use of larger geographical units to define displacement causes less precise detection of displacement changes. When displacement is defined as leaving a small geographic unit, such a method will detect more IDPs and fewer returnees. For illustration, if a family is forced to leave their home but remain in their village data collection methods using a smaller geographical unit are more likely to identify them as IDPs while larger geographic units may be less likely to (e.g., district, or county level key informant methods). The same problem applies to identifying a family as having returned to their habitual residence and therefore being subtracted from the total number of IDPs. If a family fled their village and county but later returned to a different village in the same county some methods will identify them as having returned while others will still identify them as being displaced. The precision of identification is related to both the question design (what is asked) and the respondent level (who is asked). This challenge is also discussed in the International Recommendations on IDP statistics [see 23]. Different methods of determining IDP status can result in different data. In some cases, individuals can self-identify as IDPs. This is the case with the registry of victims in Colombia, which includes over 35 For a normative discussion of durable solutions see the IASC Framework on Durable Solutions for Internally Displaced Persons [2]. “…The system is geography based, […] the data [are] collected at the lowest administrative level . Now the lowest administrative level is the lowest available, but not all of these countries— especially emergency countries—[…] have geo￾coordinates […] for the sub-district, for example . So then you do it at the district level. […] In another country you do it at sub-district level. But maybe another agency does it differently. So somebody can be an IDP that has moved to another place or a returnee depending on what the administrative level you choose -- not that you choose, but that you have available.” (010) “We try as much as possible to ask at village level. So it's possible you are displaced from your village in the same state and local governments of origin.” (028) 26 seven million IDPs. This empowers the individual, but likely allows individual’s subjective differences in IDP definitions to influence the status determination. In other cases, the location of the respondent is used as a proxy for IDP status. For example, individuals from areas affected acutely by the emergency, those living in IDP camps, or students attending schools designated as being for IDPs are assumed to be IDPs. Although this provides an efficient proxy, there are often host community members in IDP camps and schools as well as many IDPs who are not located in (or originating from) such designated areas. Other data systems use a series of questions to determine if an individual meets the criteria of being an IDP. Such question-series address whether a person moved (location) and if the move was forced (reason). However, such questions are not standardized. Some contexts and organizations specify certain spans of years to determine IDP status (i.e., they must have been displaced before and/or after a specific date). Regardless of the method used, the data are usually an undercount because some IDPs will deliberately avoid detection, with some even destroying their identifying documents that might reveal their association with the places affected by displacement. This is particularly true of conflict affected countries such as Iraq, Syria, and Yemen. Such individuals will remain invisible. It is also true that non-IDPs may identify as IDPs if they believe they will receive additional services. However, most respondents felt this latter problem could be mitigated by providing and funding services for both IDPs and the host community. Data collectors may lack a sampling frame which prevents the collection of representative data on educational indicators for IDPs. Traditionally, sampling frames are based on a census. However, some countries do not have a recent census, and even if it does exist, it is unlikely to capture the distribution of IDPs [6, 31]. IDPs have often been displaced since the most recent census, or the census methodology may not capture IDPs due to their mobility and the fact that a non-trivial number of IDPs live in informal settlements which may be systematically excluded from a household-based census enumeration. A factor that has facilitated the conducting of representative surveys in some contexts is the use of DTM data to create a sampling frame for IDPs. For example, the IOM-DTM population data has been used to create the sampling frame for household level surveys in Iraq. One of the few examples of a nationally representative survey of IDPs is the 2018 MICS in Iraq. Other respondents highlighted ways in which census systems could be adapted to better capture IDPs, such revisions are discussed in some detail in the International Recommendations on IDP Statistics [23] and are therefore not replicated here. JIPS has developed a helpful guide on sampling in displacement settings including technical guidance and illustrative cases [32]. This consideration is important for informing the collection of representative IDP education data. INSTITUTIONAL FACTORS Institutional factors play a role at the national and international level, and at their intersection in terms of helping and/or hindering EiE data collection and use to support IDPs and their education. National governments have the primary responsibility for educational data collection and service provision for IDPs. However, during rapid-onset and protracted emergencies, their data systems—especially EMIS— may falter and will need to be more resilient, dynamic, and interoperable. “They're trying to come up with a nationally representative sample . But no one really knows what share IDP scores constitute” (012) “We use DTM data as the sampling frame, because DTM data is the best sampling frame you can find.” (010) “We have put you in the census so that we'll have some idea […] we need to make [IDPs] first visible. Only then can we think about education.” (022) 27 National level IDP educational data collection occurs in contexts that require supplements and adjustments to governments’ educational data systems to make them more responsive. EMIS can provide useful data on IDP education; however, it is not primarily designed to do so. Furthermore, in displacement contexts, EMIS is often disrupted and annual data collection cycles are not sufficiently responsive. Disruption may occur because authorities and data collectors can no longer access specific geographic areas thus creating data gaps where the emergency is most acute, those tasked with collecting and using data are themselves displaced, or the destruction of infrastructure used for data collection and reporting. When the Ministry of Education or National Statistical Office is unable to collect data, educational data can be collected by humanitarian organizations and government ministries tasked with handling emergencies and/or displacement. However, whenever possible this data should be made safe but also interoperable with EMIS, for example using harmonized methods, statistical standards, and indicators. One example is the use of the same unique identifiers for schools (and individuals) that can enable data merges—while ensuring individual safety and data protection. In Iraq, the Ministry of Migration and Displacement maintains a registry of IDPs, but it is unclear if this is interoperable with the data systems of the Ministry of Education. Meanwhile, humanitarian partners sometimes have better data on IDPs than the government, but again these data are not designed to be interoperable. In Yemen, various authorities also maintain registries of IDPs but are often unwilling to share them36 with international organizations providing support with educational data collection, service provision, and funding. This is unfortunate, given the fact that humanitarian data have been used in Yemen to inform Educational Sector planning [33, 34]. The challenge is not unique to the Middle East, as respondents also noted the need for improved inter-agency data coordination and interoperability in several other contexts. Colombia provides an example of a good practice in this regard. The Colombia Victims Unit maintains a registry of victims of the conflict which includes IDP status as a category and is linked to an individual’s national ID. This national ID enables data merging with the Ministry of National Education (and other ministries). Given the sensitivity of the data, strict data sharing protocols are put in place in Colombia, primarily between government agencies. Furthermore, the literature review noted the importance of avoiding inadvertently revealing personally identifiable information through merging of datasets [35]. Several respondents and articles highlighted examples of how the use of the same school-level IDs can help improve interoperability between EMIS and civil society organizations without disclosure of personal data. Examples of this practice include the Ethiopian Education Cluster and the community based EMIS work of Save the Children in Nepal and Sri Lanka. “Access [to specific geographic areas] and [the ability to collect] data are enemies” (006) “[…] in those government controlled areas […] government monitors […] are going on collecting some data that is being used […] But [the government’s] capacity, again, remains very weak . […] The coverage of that data is often very questionable, sometimes or limited.” (015) 36 Although some have been willing to allow organizations to look at the data within the confines of the authorities‘ offices. This may suggest that creating data access points (in person or remote) may be useful. 28 When EMIS is more dynamic, it is better able to capture the IDP population. Many national data collection systems lack the necessary frequency to inform the rapid-onset phase of a displacement emergency or continued volatility. These institutional limitations of most national educational data systems often prevent the timely and responsive collection of EiE relevant IDP data. A common challenge of IDP data collection is the population’s level of mobility with many families moving multiple times. While an annual data collection system may be sufficient under some circumstances, this is usually not the case. Respondents noted that one factor that can help increase the utility of EMIS for IDPs is the development of a more operational EMIS37 that could facilitate the ability of school administrators and teachers to coordinate educational service provision for mobile students. Such a system would capture when a student leaves a particular school and if they then enroll in a different school or if they remain out of school. This has been done in Syria with the support of UNESCO, starting in Damascus with ongoing efforts to expand to the rest of the country (see the second quote in the box below). A more dynamic, decentralized EMIS system has also been used in South Sudan, in which each state had its own Education Planning and Budgeting unit and each school reports data using a customized mobile phone platform with data collection including attendance. 38 Countries that have updated their EMIS to include student level data include Kenya, South Africa, and others [36, 37]. “The best solution—but that cannot be applied in all contexts—is a dynamic EMIS system that will collect data frequently. But those systems are not applicable to all contexts because in some countries you don't have the necessary capacities in the ministry to operate such a system.” (004) “About 2016, we identified a company in Syria who could customize software that could respond to the educational structure and educational needs inside the country. […] It's like when I go to school, when I enroll in school A, immediately the teacher captures me. My data [is linked], from my date of birth, to parents names, to where I was born , to my school, academic history, where I was before to if I have any health issues and so forth . That's all registered at the school level […] into this platform. So immediately that generates a unique [student] I.D. number. And so [if] I moved to another school B, […]the teacher A can take me off the out-of-school list. And then other school B receiving me can put me into the system. So in a way, really using technology tracking not only movement of students, but also the performance and also well-being health situation and of all learners. […] But this one is fully owned by the Ministry of Education. [But] it doesn't cover the whole country at all. […] So for now, with [our] support they expanded throughout Damascus […] then over time, the next three years or so, they'll be able to cover the whole country. […] It is to be a real time tracking and management system and not really focused on bulk figures […], but really focused on children and how they are doing in terms of academic performance, how they are in terms of their wellbeing and health and nutrition situation and also the teachers, how well they're supported, what sort of training needs they have.” (025) International community There is a lack of clarity in the international community about which organization has the primary responsibility for providing leadership to harmonize and coordinate educational IDP data collection. Currently there are initiatives by multiple organizations with different foci, different geographic coverage, different methods, and different mandates. This makes coordination and systematization difficult and does not provide a supportive enabling environment for national data 37 Some respondents spoke about “School-Based EMIS” to mean a dynamic EMIS that included student-level unique identifiers. Therefore, those interacting with stakeholders should beware that when actors speak about “school-based” EMIS that does not always mean data are only at the facility level. 38 For a summary of the decentralized school-based EMIS in South Sudan see: https://bit.ly/idpdataEMISfhi360 29 systems to improve their capacity and interoperability with the support of international and civil society organizations. The following examples illustrate the lack of clarity in responsibilities. • The Joint IDP Profiling Service (JIPS) coordinates data collection exercises on IDPs, based upon request from government, humanitarian and development actors. • The United Nations High Commission on Refugees (UNHCR) does not collect data on IDPs routinely, but only does so when they provide protection or assistance to IDPs [38]. 39 • IOM-DTM and the Education Cluster jointly developed guidance on data collection; meanwhile, they, REACH, and JIPS all maintain separate data guidelines and capacity building materials. This lack of clarity was noted as a challenge by respondents at both the global and national level. Furthermore, improved clarity of responsibility could enable the creation of a shared definition and operationalization of IDP data collection to create a more supportive enabling environment for national governments. Without such clarity, respondents noted that it is difficult to develop a systematic, approach when there are joint, parallel, and overlapping efforts by Education Clusters, IOM, UNESCO, UNICEF, REACH, UNHCR, UNOCHA, UNSTATS, JIPS and others without clear coordination. While each organization has important strengths to contribute to efforts to improve IDP educational data collection, there is a need for a single organization to provide strong leadership to harmonize and coordinate efforts effectively. Standard data collection procedures and guidance on their adaption (or the lack therefore) affects the quality, utility, and efficiency of IDP education data. Most respondents called for improved systematization of data collection processes for IDP education. Although some respondents were reticent to create a template for fear that unique situations would copy-and-paste inappropriate needs assessment questions, such reticence was a minority position that could be addressed by creating an adaptable module as part of a standard question set. Such systematization can draw from the guidance developed by IOM-DTM and the Global Education Cluster [39, 40], the ongoing discussions on indicator standardization at INEE’s Data and Evidence Collaborative [41], and the work of the Expert Group on Refugee and IDP Statistics which plans to release guidance in 2022 [23]. It should be noted that multiple organizations (Education Clusters, REACH, DTM, EGRIS) have released separate guidance documents in the past two years which creates a confusing environment for those looking for support. However, until a standardized set of questions and/or indicators is developed, respondents noted that systematization can be improved along other dimensions of data collection. For example, respondents “Collaborating with the NSO, the National Statistical Office, maybe we can include IDPs. But before we do that we must first and foremost decide among ourselves. We should come to a conclusion about what is an IDP and what is not an IDP.” (022) “There is no systematic approach across these different countries. I mean, it is often ad hoc” (015) “So what they’re trying to do now is to have kind of a uniform, or somewhat kind of that is standard data collection and tool and maybe a questionnaire […] that has input from various clusters” (019) “The systematization for me is one of the key issues. And I don’t know how that could be solved because it might be too much to ask people to enter into these systematic ways. (018) 39 The UNHCR experience with refugee data collection is certainly relevant; however, the point made here—and by many respondents—is the lack of clarity of who among international partners is responsible for data on IDPs in general and IDP education in particular. 30 shared how helpful it is when implementing partners use the same EMIS school identification numbers as the Ministry of Education. Such a change can enable data interoperability at the school level among implementing partners themselves and with the national EMIS. For example, the Ethiopia Education Cluster has recommended all partners use the EMIS school ID for data interoperability at that level of (dis)aggregation [42]. The National and International Intersection The national and international factors noted above can interact to create an environment that is conducive to improved IDP education data collection and use. A resilient national EMIS or even an EMIS with a relatively complete list of unique school identifiers can facilitate the actions of international organizations (as well as local civil society organizations). Likewise, clarity about the multilateral agency with an IDP data mandate and a systematic approach to educational data collection for IDPs will create an enabling environment for national governments. Such clarity will make it easier to identify lessons learned and obtain capacity building support which is often needed by the government educational data systems. Furthermore, more systematic data systems at both levels can facilitate that transition from humanitarian support during rapid onset disasters to sustainable national systems if displacement becomes protracted. “First of all, we need to really see how we can collect data and clean it up. […]. We have so much information out there [in South Sudan]. But how would it be delivered and how it is used for decision making is something else. This data needs to come out clearly. We need to have a channel of communication whereby we know that the data that we are going for is harmonized. […] We tried to harmonize it through the Bureau of Statistics […] but they need a lot of capacity building to have data sources, where to store the data, and how people can access the data.” (014) “We made it very clear […] to begin with, the idea of having that principle of having a system. We said, so that all partners who want to support the education system can align themselves to one single source of data involving the humanitarian situation. There are a lot of organizations working around the education [system] all over the country, not only in Syria, in Yemen as well. For their own purpose, they collect data, but that cannot be it. That cannot be the one to decide whether funding should go to one place or another. It should be one official source.” (025 “To expect governments to then do it in what are often fragile settings where they don't have capacity anyways to do the regular kind of mass monitoring , which is another tool we often use in countries where education monitoring information systems . You know, it's a very big challenge just because of those capacity issues as well in those member states that are often very weak as well in terms of their systems” (015) The new transitional educational sector planning tools can help facilitate linkages between international and national institutions to make use of IDP education data to inform national planning. The Global Partnership for Education (GPE) and the International Institute for Education Planning (IIEP) developed guidelines for creating transitional educational sector plans in fragile, conflict￾affected contexts in 2016 [43] with appraisal guidance in 2017 [44]. The guidance calls for consideration of forward planning to create an educational system that can handle displacement and the consideration of IDPs when conducting an educational sector assessment [43]. For example, Yemen developed a transitional education sector plan in 2019, which included “school-aged IDPs/returnees as a percentage 31 of the same age group in the resident community” drawn from UN OCHA and other IDP educational figures from the Yemen Education Cluster [34] and reporting on other relevant educational data [33]. 40 The motivation and profile of national and international staff affects data collection and quality. At the national level, several respondents spoke about how data flowed from the local to the national but rarely came back down in a way that was useful for staff at schools and local education offices. The regular extraction of data combined with either no feedback or slow feedback affects the motivation and accuracy of the data reporting and collection at the local level that feeds into EMIS and other data collection systems. Meanwhile, when there is a more dynamic EMIS with local access to a dashboard (and capacity to use it), staff were more motivated about the utility of the data and thus to continue accurately collecting it. At the international level, several respondents noted that the personal interest, motivation, and capacity of cluster coordinators and information managers often determined whether high quality data collection was prioritized. In one case it was explicitly mentioned that funding was not a problem for educational data collection in most contexts, but personnel profiles were—with most personnel having backgrounds (and training) focusing on implementation instead of data and date use. This suggests there is a need to train cluster staff on data collection as well as data use to create a culture that values data for its usefulness and not data as an accountability exercise that interferes with implementation. In addition, each cluster should ensure they do have an information manager, as not all do. Training and support materials for coordinators and information managers can play a critical role in empowering them to “unlock the capability of the community to collect and report data on education” (018). “And this was one of the things that they would get frustrated and they were like: ‘We don't feel motivated to collect more data that we just put in a form and we just send to [the capital] and then nothing happens.’ […] And then that affects data quality. Because they're not checking the data properly, that they're not verifying it because it is almost objectively a waste of their time. And these are people who are quite time constrained. They're very resource limited. So the way that you use them is important and is often neglected.” (011) “For me one of the key issue is the actually low levels of data literacy that you find in many institutions and organizations . And it's not a criticism or a thing is the nature of the job. So if you have someone who is a social worker or something , they might not be aware of the data needs often may not have been exposed to that. And I think it ought to integrate some of the basics in a basic toolbox for […] I think the basic toolbox in terms of collecting data and reporting [is] one of the priority items . What are some of the key dimensions of education ? How are they measured with different types of indicators ? So what kind of identity would you have ? And maybe some basic templates to collect data , too , to report some channels or so to report in a more centralized way.” (018) 40 Note, Iraq and Syria have not developed such plans although GPE lists Syria as an eligible country https://bit.ly/idpdataGPEpartners 32 POLITICAL FACTORS When internal displacement is related to conflicts, there are several political factors that influence the collection of educational data on IDPs. Different areas of the country may be under the control of different parties to the conflict making access difficult. This may make the continued use of national data systems (e.g., EMIS) impossible in areas that are not controlled by the government. For example, Yemen has three different authorities and in Syria opposition-controlled areas have provided a parallel education system and curriculum [45] although this is less common now. In such instances, international partners can help bring different educational data sources together. However, such a role for international partners is still predicated on their access and on the use of and acceptance of such data by the government. In Syria, the government reportedly has not accepted humanitarian data sources for informing its humanitarian response plans [46, 47]. Identifying IDPs may put them at risk since education can be used as a tool in a conflict [48, 49]. Identifying IDPs may enable the targeting of those students and schools either physically (targeted attacks, kidnapping, or imprisonment), fiscally (preferential resource allocations), or educationally (deploying re-education programs) [50, 51]. As a result, many IDPs may not want to be identified. This also highlights the importance of organizations prioritizing safety when they consider linking data sets that may inadvertently reveal personally identifiable information. “You have the government of Yemen. You have Ansar Allah. You have more recently the STC [Southern Transitional Council]. And then are as you drill down and you may have been a different patchwork of governance of district and tribal authorities” (001) “You don’t want to be identified as an IDP because then you’re a suspected rebel [in Sri Lanka].” (020) “…one of the reasons a lot of IDPs will not often be very willing to provide information because then it becomes easier to target them [..] Same for refugees.” (015) “So [where there is new violence] you have people that are scared again to approach the state, not as the way they were before. […] But maybe they're scared that other actors can do something bad to them. Retaliation or something.” (031) “When you look at the places like Iraq [or] Syria , […] when you look at individual minorities, there's a certain kind of our challenge with [identifying IDP status with Ministry of Education data collection]. But there is no alternative. And I think what you have to re-look at is what are the safeguards that can be built around that. […]These are issues that, you know, there's very hard conversations ahead.” (001) However, such safety and political concerns can sometimes result in data censorship that reduce the data’s usefulness. Respondents reflected on how a careful balance needs to be struck and that this can only be done through (often difficult) conversations with various stakeholders and communities. ““That's where everything is so highly politicized and there are so many triggers to kind of the conflict. Tensions are always […] simmering just below. So we had to anonymize data to a level where we're writing the report, it almost felt like... It was unclear what we were writing about. Like our findings were just sort of like: “And then in this one area, we found that…”, you know? But look, we couldn't name the area.” (017) 33 Government officials may serve as the gatekeepers for data collection activities in such a way that can skew the data collection or its subsequent interpretation and dissemination. This may result in entire areas not being represented in data collection. It may also limit which countries are collecting data if the country’s government resists data collection. The selection of key informants from/by the government may also bias the data if they are trying to represent a more positive picture for political reasons or a more negative one to get more funding or services. This challenge is related to the fact that IDPs remain in their country of residence and thus under the purview of the government and not of international agencies. Intergroup politics can affect the ethics and utility of IDP specific educational data collection and use—especially in conflict-affected locations. Among respondents with a stronger focus on the educational system and the governments roles and responsibilities there was a feeling that the benefits of adding a displacement marker to educational data systems would outweigh risks and costs, especially when a sufficient number of IDPs are in the context. However, among respondents who work more on protection-related and humanitarian-focused work, there was a feeling that when there are sufficient IDPs to justify an individual marker, there is also sufficient vulnerability in the host community to argue for focusing on shared vulnerabilities and avoiding the individual identification of IDPs which could create inter-communal tensions. The resolution of these two perspectives seems to come from two ideas shared by respondents. First, respondents shared a strong belief that the data collected must be linked to data use. Therefore, when there is a use for IDP-specific educational data then such disaggregation should be included. The second—and related—resolution to these different perspectives is the targeting issue alluded to above: when there is “uniform vulnerability” the utility of targeting IDPs for specific educational services is probably very small. However, when IDPs have unique educational needs that require changes to educational planning, programming, and policies then such data should be collected. When such unique needs are related to characteristics that are indirectly (but not directly) linked to IDP status—such as differences in language or curricular exposure—it may be possible to directly ask about those characteristics rather than IDP status. In short, respondents noted the importance of being conflict-sensitive when planning to collect or use education data that distinguishes IDPs. “And so the […] assessment isn't done or hasn't been done in Yemen. […] Because authorities give you access or not .” (023) “Look carefully, because sometimes the jumps in numbers can be that we've just had access to a new area or we identify more key informants” (010) “You would also have a question: the ethics and utility of trying to identify IDPs in classrooms. Because you're then introducing questions of, why are you asking? […] Why are you trying to identify children from other parts of the country who have potentially been displaced […] particularly issues around inequality due to ethnicity?” (011) “This kind of uniform vulnerability. You know, everyone is vulnerable [in South Sudan]. That you cannot really say this is this. This is the community.” (019) “In Yemen […] you have to maneuver around things. And so I think if you […] consider the host community and IDPs and people who are living in a community rather than just focusing on IDPs, you’ll have more of a chance in gaining acceptance and access” (023) “But it's quite important [to know IDP status in Syria] because for one minister to be able to plan” (025) 34 Interorganizational politics also influences the fragmentation of the IDP education data landscape. Different organizations may compete for funding and status to conduct data collection, with each organization using a different methodology. Different funders may also require different types of reporting and thus incentivize different data collection systems. Such fragmentation goes against the commitments in the New Way of Working and Grand Bargain and many respondents argued that there needs to be improved systematization and standardization among organizations. While such standardization was called for by many respondents, several also noted the need to be responsive to the context. Finally, this interorganizational political challenge overlaps with other challenges noted above: the conceptual challenge of a lack of a consistent definition of an IDP and the institutional challenge of a lack of a single entity with the mandate to collect and coordinate educational data for IDPs. “They're our big competitors […] Between donors and [organizations] there are politics. And so they come in even when we're present they come in, when we are established they come.” (010) “So you get these multiplicity of organizations competing over that space--for lack of a better term.” (015) 35 CONCLUSION National governments and the international community have increased their efforts in the past decade to collect the type of data to improve educational provision to IDPs. As a result of these efforts, there is often more relevant data available than stakeholders are aware. However, these data sources are fragmented, use different methods, and are not easy to find. Answers to the three research questions are shared below. Research Question 1: How are IDP education data collected and used in the Middle East and other crisis￾affected contexts globally? In addressing our first research question, we have mapped the most prominent data collection exercises that include educational data on IDPs. This includes the global aggregator for IDP data, IDMC. Because IDPs remain in their country of residence, the primary duty-bearer for education provision and related data collection is the government. As such, we discuss the role of EMIS and IDP registries as primary education data sources for IDPs. However, often these systems are not functioning or unable to respond to the crisis that is causing displacement. We therefore turn next to the most important and commonly cited data collection initiatives by international organizations with education data for IDPs, namely: IOM-DTM’s Multi-Sector Location Assessments, REACH’s Multi-Sector Needs Assessments, Education Clusters’ Joint Education Needs Assessments and 5Ws, and the Joint IDP Profiling Service. We also mention other data sources that are less widely available across geographies. Research Question 2: What are the factors that support or constrain collecting and using IDP education data in the Middle East and globally? The remainder of the paper addressed the second research question by exploring the factors that help or hinder the collection and use of educational data for IDPs. We found four broad types of factors: (1) conceptual, (2) technical/methodological, (3) institutional, and (4) political. There are efforts underway to address the conceptual and technical/methodological challenges presented by IDP data collection generally in the International Recommendation on IDP Statistics [23] and the work of JIPS [19]. These efforts, if successful, will help move these conceptual and technical factors from hindrances to helpers. However, this study identified more institutional factors that help and hinder the collection and use of education data for IDPs. Many of these are not being addressed systematically and jointly; different actors are moving in different directions even as they try to systematize. An example of this is the proliferation of guidance in just the past two years from the Education Cluster & IOM-DTM, JIPS, REACH, GPE, and EGRIS. While each of these pieces of guidance is helpful, taken together, the efforts at standardization risk further fragmenting a space that lacks a clear central organizing force. The political factors tend to be intensified versions of those that also affect refugee education data, because IDPs remain in their country of residence, which is where the emergency causing their displacement originated and where the duty-bearer (the national government) is affected by that emergency. Furthermore, international law has fewer mechanisms to protect IDPs and ensure their right to education than in the case of refugees—at least in theory. Lastly, intergroup tensions may be more salient and differentiating IDPs from host communities more difficult and more dangerous. 36 Research Question 3: How can various stakeholders improve the collection and use of data to support the education of IDPs in the Middle East and globally? The recommendation section below addresses the third research question regarding what stakeholders can do to improve the IDP educational data ecosystem. The recommendations are drawn from the literature, interviews, and our own perspectives We hope this paper will assist those working in the sector to navigate the existing data sources prior and to consider the factors and recommendations outlined herein before undertaking new data collection efforts. We further hope this effort and the efforts of others will lead to a more systematic, more useful data ecosystem to inform the education of the millions of internally displaced children in the Middle East and around the world. 37 RECOMMENDATIONS This section provides a set of suggestive recommendations for national governments, international and civil society organizations, and funders to consider improving the educational data ecosystem for IDPs. The recommendations are drawn from the discussion of facilitating and hindering factors above. Recommendations should be considered as a point of departure to be adapted to the unique contexts and capacities of each stakeholder. GOVERNMENTS, AS APPROPRIATE FOR THEIR CONTEXTS, SHOULD CONSIDER: • Creating an interoperability data standard between the Ministry of Education, the ministries in charge of emergency response and displacement tracking, and humanitarian partners. • Implementing an EMIS system that is more dynamically updated at the school-level to inform school-level practice for students experiencing high mobility. • Evaluating the risks and benefits of the inclusion of unique student identifiers. • Including a marker for student movement as a proxy for IDP status. • Providing timely access to data systems for schools and local authorities so that data can inform decisions at the level of implementation in a responsive fashion. • Including IDP screening questions in existing national surveys, especially in contexts of prolonged internal displacement such as Iraq, Syria, and Yemen. • Develop an educational sector plan to improve the resilience of the educational sector’s ability to handle displacements, including its data system. Yemen’s experience can provide an example [33, 34]. INTERNATIONAL ORGANIZATIONS AND CIVIL SOCIETY SHOULD CONSIDER: • Empowering a single multi-lateral agency to lead efforts in coordinating and harmonizing IDP data collection efforts and system strengthening to support national authorities. • Developing a standard set of educational indicators and data collection procedures for IDPs with the inclusion of explicit guidance on how to contextualize the standardized tool. • Encourage Educational Clusters and their implementing partners to make their data interoperable with EMIS, especially by linking all data collection to accepted EMIS-based unique school IDs. • Improving the capacity of staff to collect and use educational data to inform their work with IDPs and host communities • Continuing to ensure that both displaced children and host community children are made visible through data collection and are supported through educational service provision to reduce perverse incentives to self-identify as IDPs (or not). • Ensuring a balanced level of time and resources for both data collection and data use. • Supporting community-based data collection efforts that can inform the education of IDPs, such as the work of the PAL network41 and the community-based EMIS experience of Save the Children in Nepal and Sri Lanka. • Adapting existing surveys with educational data to include screening questions for IDP status, in particular the MICS42 should be adapted to include IDP status in contexts where there are large 41 For more information about the PAL network see: https://bit.ly/idpdatapalnetwork 42 The DHS EdData Surveys could also be adapted in this way although they were not mentioned by our respondents. The DHS has questions about family moves and related school changes, but does not ask questions to determine if the moves were forced. 38 numbers of displaced children such as Iraq, Syria, and Yemen. The 2018 Iraq MICS can serve as a template. DONORS SHOULD CONSIDER: • Standardizing data collection requests. • Incentivizing organizations to adhere to internationally established data standards. • Providing more proportionate funding for educational services and data during humanitarian crises. ALL STAKEHOLDERS SHOULD CONSIDER: • Ensuring data is shared back with schools and local education officials in a timely manner and a useful and useable format. There continues to be a concern that data systems often extract data but never deliver any utility back to those who need it to make decisions at the school-, district-, and even national level. 39 APPENDIX 1: INTERVIEW GUIDE FHI 360 MEERS - IDP Education Data - Interview Guide INTRODUCTION: Welcome. Thank you again for agreeing to speak with us. Let's start by introducing ourselves. My name is ____________. I work for ___________________ through the Middle East Education Research, Training, and Support Initiative (MEERS). We are interested in learning more about how EiE organizations and professionals find and use data related to IDPs – specifically, we are looking at the intersection of IDPs, education, and data. For the purposes of our conversation, when we say “Education in Emergencies” we are referring to education for those affected by disasters, conflict, and displacement. CONSENT: [Read consent form. If a participant chooses not to give consent, he / she cannot participate in the interview and the interviewer must stop the interview.] Did you get oral consent from each participant? ___ Yes ___ No INTERVIEW: Demographics First, we would like to ask some information about your background and your work. 1. What is your current position within your organization? What are your major responsibilities? 2. How long have you been working in the field of education? 3. How long have you been involved with work related to education in emergencies? 4. Would you consider your work to focus on a local level, national level, or global level? EiE Organizations and IDP Data Now we’d like to ask you about your organization’s work related to IDPs. 1. Does your organization collect or aggregate data related to IDPs? (If Yes, Skip Down to Q10 - Below) For Organizations that do IDP Education Programming but not Data Collection 2. Can you describe the types of programming or organizational activities that you and your organization carry out as they relate to IDPs and specifically, education for IDPs? a. Does your organization’s programming vary or otherwise distinguish between local learners, IDPs, and refugees? If so, how do you do so? b. What are the major challenges you face in designing EiE programming for IDPs? 3. Now we want to focus specifically on data availability. What types of data exist for understanding who, how many, and where internally-displaced children and youth are? a. What are the major challenges you face in accessing data on internally-displaced children and youth? b. What types of data exist for understanding the educational needs of internally-displaced children and youth? 40 c. What types of data exist for understanding educational access and experiences of internally-displaced children and youth? d. Does a lack of data limit or otherwise make your work more difficult? e. (If relevant) In your experience, how does data availability vary for IDPs and refugee education programming? 4. What are the key sources of data you consult when you need data on IDPs? a. Where do you look and who do you ask for such data? b. What types of data are publicly available on IDPs? Are there other forms of data you have found that are not publicly available? 5. Who are the partner organizations that your organization shares data on IDPs with, or that share these data with you? a. What are the existing processes or protocols for sharing data related to education for IDPs, particularly in conflict regions? 6. Based on your experience and expertise, what are the priorities for improving data for education for IDPs, particularly those in conflict settings? Relatedly, do you have any ideas or thoughts on novel approaches that could improve data collection for internally-displaced children and youth? 7. Based on your professional experience and expertise, what are the most pressing challenges for finding and using information on IDPs, and specifically on the educational needs and experiences of IDPs? 8. What country or organization provides an example of a good practice on the collection of data on IDPs and/or the use of data to inform IDP education? 9. Is there any other information that you would like to share with us? For Organizations that Collect Data on IDPs 10. Could you tell us more about the mandate of your organization and where IDPs fit in specifically? 11. We’d really like to understand more about how data on IDPs, and specifically on internally￾displaced children, is currently collected and made public or accessible. Could you tell us more about what types of data on IDPs you currently collect? a. How frequent is data collection? b. At what level is data aggregated? c. What forms of data are publicly available? (Are some data not made publicly available) d. What forms of disaggregation are currently carried out/ possible? 12. Do you, or any organization you know of, have specific initiatives to collect data specifically on internally-displaced children and youth? (Why or why not?) a. Do you or any organization you know of have specific initiatives to collect data on the educational needs and experiences of IDPs? 13. Could you tell us more about the processes and methods you use to collect and validate data on IDPs? 41 a. What are tools and processes are currently used to identify how many and where IDPs are? b. What tools or processes are in place to validate data? 14. (If relevant) Based on your professional experience, do data collection and aggregation procedures vary for local learners, IDPs, and refugees? If so, how? 15. Outside of your organization, are their other key sources of data you consult when you need data on IDPs? 16. Who are your partner organizations that your organization shares data on IDPs with, or that share these data with you? a. What are the existing processes or protocols for sharing data related to education for IDPs, particularly in conflict regions? 17. Based on your professional experience and expertise, what are the most pressing challenges in collecting information on IDPs, and specifically on the educational needs and experiences of IDPs? 18. Are there any promising developments in collecting information on IDPs, and specifically internally-displaced children and youth? Or, do you have any ideas or thoughts on novel approaches that could improve data collection for internally-displaced children and youth? 19. What country or organization provides an example of a good practice on the collection of data on IDPs and/or the use of data to inform IDP education? 20. Is there any other information that you would like to share with us? 42 REFERENCES 1. UN Economic and Social Council, Guiding Principles on Internal Displacement (E/CN.4/1998/53/Add.2). 1998, New York, NY: United Nations Economic and Social Council. 2. IASC, Brookings, and Universitat Bern, IASC Framework on Durable Solutions for Internally Displaced Persons. 2010, Washington, DC: Brookings. 3. IDMC, Global Report on Internal Displacement. 2020, Geneva: Norwegian Refugee Council & IDMC. 4. UNICEF and IDMC, Equitable access to quality education for internally displaced children 2019, New York & Geneva: UNICEF & IDMC. 5. UNHCR, Guiding Principles on Internal Displacement. 1998, Geneva: UNHCR. 6. IDMC, Twice invisible: Accounting for internally displaced children 2019, Geneva: IDMC. 7. UNGA, Protection of and assistance to internally displaced persons (A/RES/52/130). 1998 New York, NY: United Nations General Assembly. 8. UNGA, Human rights of internally displaced persons (A/HRC/RES/20/9). 2012, New York, NY: United Nations General Assembly. 9. USAID, NORRAG, and INEE, Education in Emergencies Data: A Long-term Vision and Action Agenda. 2019, Washington, DC: USAID, NORRAG, INEE. 10. Miranda, Á.S., A.C. Valderrama, and M. Kissenkoetter, Painting the Full Picture: Persistent Data Gaps on Internal Displacement Associated with Violence in El Salvador, Guatemala and Honduras. 2019, Geneva: Internal Displacement Monitoring Centre (IDMC). 11. Sarzin, Z., Stocktaking of Global Forced Displacement Data. 2017, Washington, DC: World Bank. 12. Sarzin, Z., Forced Displacement Literature Review. Joint Data Center Literature Review. 2019, Washington, DC: World Bank. 13. GAO, Internally Displaced Persons Lack Effective Protection. 2001, Washington, DC: United States General Accounting Office. 14. IDMC. About IDMC. 2020 [cited 2020 July 17]. 15. IDMC. An Institutional History of Internal Displacement. 2020 [cited 2020 July 17]. 16. INEE, Minimum standards for education in emergencies, chronic crisis and early reconstruction. 2004/2010, INEE: Paris. 17. IOM. About DTM. 2020 [cited 2020 August 20]; Available from: https://dtm.iom.int/about. 18. REACH. Who we are. 2020 [cited 2020 August 20]; Available from: https://www.reach￾initiative.org/who-we-are/. 19. JIPS. About JIPS. 2020 [cited 2020 September 24]; Available from: https://www.jips.org/about￾jips/. 20. Grand Bargain Secretariat, The Grand Bargain. 2016, Geneva: Grand Bargain Secretariat. 21. Metcalfe-Hough, V., W. Fenton, and L. Poole, Grand Bargain annual independent report 2019. 2019, London: Humanitarian Policy Group at ODI. 22. OCHA, New Way of Working, Development and Studies Branch, Editor. 2017, OCHA: New York. 23. EGRIS, International Recommendations on IDP Statistics (IRIS). 2020, Geneva: United Nations Statistical Commission Expert Group on Refugee and IDP Statistics. 24. Saldana, J., The Coding Manual for Qualitative Researchers. Third ed. 2016, Thousand Oaks, CA: Sage Publications. 25. IDMC, Measuring the costs of internal displacement on IDPs and Hosts: Case studies in Eswatini, Ethiopia, Kenya and Somalia. 2020, Geneva: IDMC. 26. IOM, Iraq Integrated Location Assessment III. 2018, Baghdad, Iraq: International Organization for Migration (IOM). 27. REACH, Syria Multi-Sectoral Needs Assessment. 2014, Geneva: REACH. 43 28. Save the Children and Yemen Education Cluster, Joint Education Needs Assessment. 2016, Aden: Save the Children & Yemen Education Cluster. 29. Buckner, E., A. Smiley, and S. Cremin, A New Way of Thinking About Education in Emergencies Data. Norrag 2019. Special Issue (02): p. 19-21. 30. UNHCR and JIPS, Displacement as challenge and opportunity. Urban profile: Refugees, internally displaced persons and host community. . 2016, Erbil Governorate, Kurdistan Region of Iraq: UNHCR. 31. Shafiq, M.N., School enrollment in Iraq during the U.S.-led invasion: A statistical analysis. International Journal of Educational Development, 2013. 33(2): p. 130-138. 32. JIPS, Sampling Guide for Displacement Situations & Practical Examples. 2020, Geneva: Joint IDP Profiling Service (JIPS). 33. GPE, Results Framework Data for Yemen. 2019, Washington, DC: Global Partnership for Education (GPE). 34. Yemen Ministry of Education, Transitional Education Plan 2019/20 - 2021/22. 2019, Washington, DC: Global Partnership for Education. 35. Morales, L.G. and T. Orrell, Data Interoperability: A Practitioner’s Guide to Joining Up Data in the Development Sector. 2018, Dubai: Global Partnership for Effective Development Co-operation. 36. Hure, M. and J. Taylor, Education Data: A Key to Inclusion of Refugee Learners in National Systems. Norrag, 2019. 2: p. 77-79. 37. GPE, Meeting the data challenge in education. 2019, Washington, DC: Global Partnership for Education (GPE). 38. UNHCR. Measuring forced displacement and statelessness. 2020 [cited 2020 September 20]; Available from: https://www.unhcr.org/refugee-statistics/methodology/. 39. Global Education Cluster, Guide to Education in Emergencies Needs Assessments and Analysis. 2019, Geneva: Global Education Cluster. 40. UNICEF and IOM, Global Education Cluster Guidance: Obtaining useful data from IOM's Dispalcement Tracking Matrix (DTM) to inform emergency education planning and response. 2019, Geneva: UNICEF. 41. INEE. Data and Evidence. 2020 [cited 2020 September 20]. 42. Ethiopia Education Cluster, Education in Emergency 2016/17 Response Strategy. 2016, Addis Ababa: Ethiopia Education Cluster. 43. GPE and IIEP, Guidelines for transitional education plan preparation. 2016, Paris: International Institute for Education Planning. 44. GPE, Guidelines fro Transitional Education Plan Appraisal. 2017, Washington, DC: Global Partnership for Education (GPE). 45. UNICEF, Curriculum, accreditation and certification for Syrian children in Syria, Turkey, Lebanon, Jordan, Iraq and Egypt. 2015, UNICEF: Amman. 46. Anderson, A., A. Magee, and S. Nicolai, Strengthening coordinated education planning and response in crises. 2020, London: Overseas Development Institute (ODI). 47. UNICEF, Syria Education Sector Analysis: The effects of the crisis on education in areas controlled by the Governmetn of Syria, 2010-2015. 2016, Demascus: UNICEF Syria. 48. Davies, L., Education and Conflict: Complexity and Chaos. 2004, London: Routledge Falmer. 49. Davies, L., Schools and war: urgent agendas for comparative and international education. Compare: A Journal of Comparative and International Education, 2005. 35(4): p. 357-371. 50. Wessel, M.v. and R.v. Hirtum, Schools as Tactical Targets in Conflict: What the Case of Nepal Can Teach Us. Comparative Education Review, 2013. 57(1): p. 1-21. 51. Pherali, T.J., Education and conflict in Nepal: possibilities for reconstruction. Globalisation, Societies and Education, 2011. 9(1): p. 135-154.