Cart 0


Multiple Overlapping Deprivation Analysis (MODA) is a new tool designed to capture the deprivations that each child faces at the same time. MODA builds upon the Bristol approach to multidimensional child poverty applied in the Global Study on Child Poverty and Disparities, as well as the Oxford Poverty and Human Development Initiative’s (OPHI) Multidimensional Poverty Index. MODA adopts a whole-child oriented view, which can be used to create a comprehensive picture of children experiencing deprivations of basic needs, and permits comparison across countries and within countries. MODA identifies overlapping deprivations at both the household and individual child level across child relevant dimensions, such as (in CC-MODA) water, sanitation, housing, protection against domestic violence (analyzed for all children), and nutrition and health (analyzed for children under the age of five), and education and information (analyzed for children aged 5-17 years). MODA captures how many children experience overlapping deprivations (incidence) and how many deprivations they face on average (intensity). MODA can be broken down by dimension and children’s characteristics to show how the composition of multidimensional poverty changes for different regions, ethnic groups and other individual or household features – with useful implications for policy and programming.

MODA does not imply causal relationships between deprivation/poverty on the one hand and profiling variables on the other. MODA does, however, show statistical associations between the child’s personal and household characteristics and the dimensional deprivation, deprivation level and the deprivation overlap, and helps to identify “who” is deprived and/or poor. As such MODA can help scholars to generate hypotheses about causal relationships. Real causal analysis (or answering the question “why” some children are more deprived than others) can be inspired by MODA but in itself can only be based on a theoretical framework and analytical statistical techniques specifying all types of interactions between the dependent and explanatory variables (and mutually between the explanatory variables). MODA is a descriptive compilation of information describing children’s status of well-being. It describes the deprivation status of children, identifies the deprived children and the deprivations they suffer from; it shows how the different deprivations overlap, how the deprivations are distributed across the child population, and how these results change depending on children’s personal and household characteristics. These characteristics are used as profiling variables to draw-up a picture of multiply deprived children. This information is meant to inspire and guide further in-depth research on the causal relationship of the different determinants of child deprivation, and possibly assist policy-makers in designing effective policy interventions. As an example, when the profile shows that children without a birth certificate are more likely to be deprived in health than children with a birth certificate, the observed health issues are not solved by providing children with a birth certificate. The reason why children cannot access health care facilities should be investigated through further research on the supply and demand problems around their access to health care.

There are two applications of MODA:

  1. Cross-country MODA study (CC-MODA), identifies deprived children in low and middle-income countries applying the MODA-approach, using the most recent globally comparable data from the Demographic and Health Survey (DHS) and Multiple Indicator Cluster Survey of UNICEF (MICS). CC-MODA uses standardized definitions of age groups, dimensions, indicators, profiling variables and thresholds. CC-MODA allows international comparisons.
  2. National MODA (N-MODA): identifies deprived children in applying the MODA-approach using a suitable dataset for each country (either an international dataset or a national specific dataset). N-MODA adopts country-specific definitions of age groups, dimensions, indicators, profiling variables and thresholds. While the results are not comparable with other countries, they can reveal more detailed and richer information on the extent and characteristics of child deprivations.

The MODA web portal functions as an assembly point for all studies and works related to MODA; the portal consists of various sections which comprise information on N-MODA, CC-MODA, background materials, guidelines and publications. Currently, one of the main parts of the MODA webportal is the CC-MODA analysis. This interactive application shows the results of CC-MODA by country and in a country comparative setting. Results can be adjusted by country, age group, dimension, profiling variable and/or deprivation threshold according to the user’s preference. A short description and example is provided for each part of the multidimensional deprivation analysis to assist with the interpretation of the results. All of the graphs and figures can be downloaded through the “export”, “add to cart” or “CSV” function (as figure or table) to enable the further use of MODA results for UNICEF country offices or other users.

MODA (both cross-country and national) makes five important contributions to existing knowledge on child well-being.

  • First, MODA concentrates on the child as the unit of analysis, rather than the household. Children experience poverty differently from adults especially with regards to developmental needs, which can have lasting effects (UNICEF, 2000). For this reason, the survey data used to measure child well-being reflects child-specific information.
  • Second, MODA acknowledges that children’s needs are not homogenous across their childhood. For this reason MODA adopts a life-cycle approach analyzing separate age groups to reflect the different needs of early childhood, primary childhood and adolescence.
  • Third, MODA enhances knowledge of compartmentalized or sector-by-sector approaches (e.g. consumption, health, and education) with an overlapping deprivation analysis. This analysis indicates which of the multiple facets of child poverty are experienced simultaneously. The information found when analyzing the overlapping and non-overlapping groups points towards mechanisms needed for effective policy design that will address children’s needs as accurately as possible.
  • Fourth, the whole-child oriented view of MODA supports the focus on equity, because it allows to concentrate on highly deprived groups in the society and to create profiles which assist in determining their geographical and social position. The ‘whole-child’ view acknowledges that child deprivation is multidimensional in nature and therefore centers on the child rather than dimensions by looking at whether a child experiences several deprivations simultaneously, if any.
  • Fifth, the results of MODA are presented in an interactive web portal allowing the user to explore many interactions and cross-tabulations between the deprivations and the underlying variables. The web portal gives easy and user-friendly access to a large quantity of data and results, allowing the user to fully exploit the richness of the data and estimations provided.

Cross Country MODA (CC-MODA) uses international standards as guiding principles for choosing the most relevant dimensions of child well-being. The Convention on the Rights of the Child (CRC) (1989), in combination with the World Summit for Children (1990), the World Summit on Social Development (1995), and the Millennium Development Goals (2000), have served as a basis for the construction of a core set of dimensions that are essential to any child’s development irrespective of their country of residence, socio-economic status, or culture. For national MODA (N-MODA) international conventions on the well-being of children remain important, but can be complemented with the use of e.g. national norms, standards, or legislation; regional agreements; or academic theories, which may better serve the national context.

The MODA methodology can be applied using various datasets: internationally administered datasets, such as MICS or DHS; regional ones, such as the European Union Statistics on Income and Living Conditions (EU-SILC); national household surveys; or local surveys. The quality of the dataset to be used is important. A dataset used for MODA should fulfill the following requirements:

  • The sampling method and sample size should be adequate to be representative for the geographical area that MODA is intended to cover, ranging from a local entity (city, urban or rural area, district, or region) to a national state or a group of countries;
  • The data are preferably available not only on household level, but also on the individual – and especially child - level. Child level data are the only data that can make child-specific needs visible and that allow for the identification of age- and gender-driven differences (i.e. intra-household differences). Household level data can be used when indicators are equally applicable to all household members.
  • The data should comprise all the necessary variables for measuring all the dimensions seen to be essential for any child’s development;
  • The necessary information should be available for all the observations (i.e., each child to be included in the analysis), with as few missing values as possible;
  • The data should be available for all the age groups that the study aims to analyze.

When considering datasets for inclusion in a MODA type of analysis, it is important to consider the specific dimensions and indicators that are most relevant in a particular country context. The domains as specified in the CRC and discussed in the Step-by-step guidelines to MODA can guide the considerations. It is advisable to review the potential list of dimensions and the related potential indicators in order to determine whether information on the indicators is available in the datasets under consideration. Equally important is local knowledge about the assumed variance in each of the indicators; for example in countries where all or nearly all children/households have access to safe drinking water, the inclusion of the variable in MODA, and thus in the dataset used for MODA, is not important since it will not reveal any deprivation.

Cross-Country MODA (CC-MODA) relies on two main datasets that are publicly available and comparable for most low income countries: the Demographic and Health Survey (DHS) and Multiple Indicator Cluster Survey (MICS). These datasets have been chosen because they provide the most recent available data that captures child deprivation in low and middle income countries. The single requirement for country inclusion in CC-MODA is that countries need to have a recent DHS or MICS available with a reference period between 2007 and 2012. For National MODA (N-MODA), national datasets can be used as the indicators do not need to be internationally comparable. National data can be seen as a way to fill data gaps and to improve the deprivation analysis by including more dimensions on age-specific needs. For N-MODA, datasets are selected on the basis of their relevance to child well-being, possible unit of analysis, applicability to the country context, and data quality.

Step-by-step guidelines to MODA are available to assist in carrying out the entire process of the multidimensional deprivation analysis. The purpose of the guidelines is to serve as a reference for countries interested in adopting MODA. It discusses decisions that are to be made throughout the analysis and it points out the various risks of each of the choices. The guidelines to MODA refer to the general MODA methodology and leave room for creativity, adjustment, and tailoring to the country context when carrying out national MODA (N-MODA), in order to capture issues that are most relevant to the country context.

The guidelines are complemented by the CC-MODA Technical Note, which describes the methodology and technicalities of the Cross-Country MODA. CC-MODA is a special application of the MODA methodology utilizing internationally comparable datasets to analyze child deprivation using standardized definitions of age groups, dimensions, indicators, profiling variables and thresholds.

The Global Study on Child Poverty and Disparities was launched in late 2007, with 54 countries from all regions of the world joining it. The Study has been instrumental in advocating for the recognition that child poverty is multidimensional, and cannot be measured through income/consumption alone. The Study adopted an approach to child poverty based on the ground-breaking deprivation of basic needs methodology, developed in the early 2000’s by Gordon, Townsend, Nandy, Pantazis and Pemberton at the University of Bristol. Over the last decade, there have been major advances in survey collection in low and middle income countries, as well as progress in multidimensional deprivation theory and data. It was therefore timely to take advantage of these survey improvements and reflect upon the conceptual foundations of the Study. MODA, a collaborative effort between UNICEF Office of Research and UNICEF Division for Policy and Strategy, was borne out of these reflections, taking into account how to utilize the latest household survey improvements to the fullest, as well as building on recent advances in multidimensional poverty theories.

MODA builds upon the Global Study on Child Poverty and Disparities (Bristol) methodology; both follow a rights-based approach to child well-being and account for the number of deprivations experienced by each child. However, MODA acknowledges that children’s needs differ across their life-cycle. In addition, the Bristol approach is specifically concentrated on the ‘counting’ method, while MODA goes beyond this and also incorporates analyses by deprivation ratios and deprivation overlaps. For each of these types of analysis a profile of the most deprived children is created, which assists in determining their geographical and social position. The deprivation overlaps and profiles reveal specific characteristics of deprived children and can help to point towards mechanisms for effective policy design. Moreover, MODA includes an additional important dimension of child well-being which is lacking in the Bristol approach, namely child protection (protection from violence).

MODA has parallels with the Multidimensional Poverty Index MPI and integrates several innovations of MPI into the analysis. Both methodologies focus on people experiencing multiple deprivations; both aggregate various deprivation dimensions into indices; and both approaches allow indices to be decomposed by subgroups and dimensions to show the contribution of each group and dimension to the overall deprivation level. However, the most evident difference is that MODA focuses only on child-relevant dimensions (health, nutrition, education, water, sanitation, shelter, information and violence), whereas the MPI looks at three dimensions that are relevant for the entire population (living standards, health, and education). In addition, unlike the MPI, MODA adopts a life-cycle approach, analyzing children from different age groups separately by looking specifically at children below the age of five and children aged 5-17 years, or smaller age-groups if the data allows. The unit of analysis in MODA is the child, whereas for the MPI the unit of analysis is the household. Moreover, MODA puts more emphasis on “counting” the number of deprivations experienced and analyzing the types of deprivations that are experienced simultaneously by children (“overlap analysis”). In addition, MODA focuses on the identification of the most deprived children by creating a profile of (multiply) deprived children, which is based on individual and household characteristics (profiling analysis).

Both “deprivation” and “monetary poverty” are important concepts in equity analyses. Their complementarity is important in general and even more important for children. Distinguishing the two concepts (poverty and deprivation) is pragmatically useful and conceptually illuminating, with monetary ‘poverty’ measuring the lack of financial means by which deprivations could be addressed; ‘deprivations’ are then the outcome of lack of financial means (among other things). For children the distinction between financial poverty and deprivation is even more important since (monetary) poverty is measured in terms of the financial wealth of the households they live in; deprivation, however, is measured as deprivations they themselves experience (for an elaboration of the arguments see the Office of Research Working Paper Lost in Dimensions). Existing research has shown that the group of people identified as income (monetary/financially) poor does not fully coincide with the group that is found to be multidimensionally poor, based on the basic needs approach (for further information, see ‘Step-by-step Guidelines to MODA’). Additionally, with regards to children it has been argued that the use of household or adult poverty measures to represent child well-being is inaccurate or incomplete, since children’s needs differ from the needs of their parents and households. This argument refers in particular to access to social services or public goods (e.g. health care, education), which are not necessarily related to the level of income in the household.

Both types of poverty analysis make valid contributions to the objective of poverty reduction. MODA emphasizes the deprivation element of the analysis capturing those aspects of child well-being that cannot be expressed in monetary terms. Whenever data are available (and that is not the case in CC-MODA – see next question), MODA includes an analysis of financial poverty and studies the overlap between children who are deprived and children living in monetary poor families. The possibility of profiling each of the four subgroups (deprived and poor; deprived but not poor; poor but not deprived; and neither deprived nor poor) is one of the strengths of MODA.

Ideally, MODA should include both child deprivation analysis and monetary poverty analysis. These analyses complement each other and can make valid contributions to the objective of poverty reduction. Additionally, including both would provide the opportunity to analyze the overlap between the two fields of child well-being. However, income and consumption indicators are not included in CC-MODA due to data constraints: DHS and MICS do not collect information on the financial means of the households. Survey data allowing the simultaneous analysis of poverty and deprivation does exist for some countries. In many countries, however, income and consumption poverty data comes from surveys that often do not have detailed information on dimensions such as health, nutrition and child protection, which are relevant for child well-being. From most survey data in low income countries it is not possible to identify whether the same children living in income poor households are also deprived in the child well-being dimensions analyzed in MODA. With improved survey measures in low and middle income countries it is possible that in the near future we will be able to identify both income/consumption poor children and children deprived in child well-being dimensions. Moreover, recent experiments with methodologies to link household living standard surveys and DHS/MICS are encouraging and may lead to the possibility of analyzing poverty and deprivation simultaneously for a larger group of countries.

Cross-country MODA (CC-MODA) does not include the wealth index as one of its dimensions of child well-being because wealth and deprivation are seen as two conceptually different aspects of poverty. Also, the wealth index cannot be compared across countries due to the way it has been constructed. Moreover, certain elements of the wealth index such as the type of housing and the source of drinking water are also included in CC-MODA as separate dimensions, because they refer directly to children’s well-being. Thus, the wealth index would correlate highly to the water, sanitation, and housing dimensions that are used in the analysis.

As an alternative, an asset index has been constructed and applied in CC-MODA. This index is used as a profiling variable (i.e., as one characteristic of a child’s household), rather than as one of the dimensions of children’s well-being. The asset index is constructed using the same method as the wealth index, but includes only those elements that represent households’ material well-being. The variables used to construct the dimensions included in CC-MODA are not part of the asset index to avoid self-explanatory results and co-linearity.

The main advocacy objectives of MODA are: (1) to complement monetary poverty analyses, because MODA’s basic rights approach improves understanding of child well-being, and (2) to complement sector analyses, because MODA’s whole-child approach places each child rather than each sector at the center of the analysis, which reveals how the multiple facets of child poverty are experienced simultaneously and gives insight into children’s deprivation severity and the characteristics of the multiply deprived, unmasking inequities.

Key messages emerging from both CC-MODA and N-MODA are targeted at decision makers who are able to adjust or create new policies with an aim to benefit children. The target audience includes government ministers, heads of state, and members of parliament, donors, leaders of civil society organizations and the private sector. MODA provides a tool for UNICEF country offices and other UN agencies and child rights institutions to engage with institutions and individuals who can influence policy makers, including teachers and students, youth, and the general public.

MODA is a child centered methodology, which analyzes the fulfillment of children's needs according to their stage in life. MODA prefers to use individual level data which allows for analyzing age-specific needs and intra-household differences. Nevertheless, due to data limitations many of the dimensions and indicators currently adopted in MODA are measured on a household level. Household level indicators mask possible inequalities within the household. In addition, indicators that refer to household facilities might double-count the required policy response (e.g. when a household with two children uses an unimproved toilet type both children are identified as deprived, while only one improved toilet is required to improve the situation of these children). As for all empirical studies, MODA is limited by the availability of trustworthy data. No data are collected on many aspects of children’s lives across their life-cycle in most countries. Filling the data gaps is an important task for the future. MODA points to the existing gaps in the available data and therefore can serve as a source of inspiration to engage in new survey designs.

One of the lessons learned while applying the MODA methodology is that data quality is often a constraint to perform a child-level deprivation analysis. MODA follows an approach which places the child at the center of the analysis (i.e., a child-centered approach). This approach requires that data that is relevant to child well-being is available at an individual level. Moreover, due to the adoption of the life-cycle approach, there is a need for information which is age-specific and applies to all children within a given age-group. Many datasets (also the ones currently used for CC-MODA) face limitations, such as high proportions of missing values on variables, the lack of data on an individual level, and the use of sub-samples to measure certain variables.

The CC-MODA study faces additional data requirements due to the need for standardization and international comparability. While many relevant aspects of child well-being are currently included in CC-MODA, concessions have been made due to data limitations. For example, health, although relevant for all children, has not been included for children over the age of five in this comparative study, due to missing information on access to health care for this age group.

UNICEF’s equity agenda stipulates the need to focus on the most vulnerable and disadvantaged children and to place emphasis on demonstrating equitable results and value for money. The equity approach encourages analysis that highlights the situation of disadvantaged groups. MODA is developed to highlight exactly that by systematically examining patterns of inequities. MODA enhances the equity focus by unmasking inequities hidden in sector-based analyses as highly multiple deprived children have distinguishable profiles. Firstly, the analysis presents results by different levels of deprivation intensity and by the simultaneous experience of deprivations, and can therefore zoom in on the most deprived children in a given society. Secondly, the “profiling function” of MODA helps to identify notable characteristics of the deprived children in a given country or society. These two features contribute to the equity agenda, because they provide information on how many (multiply) deprived children there are, who they are, and where they live.

MODA contributes to MoRES on two different levels. Firstly, it is the backbone of an equity focused situation analysis which is the backbone of Level-1 monitoring under MoRES. MODA makes the identification of children in need - and those most in need - prominent, and allows profiling of their and their households’ characteristics. MODA thus serves as a guide in equity focused policy design and country programming.

Secondly, MODA will play an expanding role in Level-3 monitoring under MoRES. The results of policy interventions have to be understood ultimately in terms of reducing the (overlapping) deprivations experienced by children. MODA and MoRES teams work closely together to expand this part of the analysis and explore for example how ‘real-time monitoring’ information can be aligned with MODA results.

MODA can be instrumental in informing UNICEF country programme cycles. UNICEF programming is traditionally aligned according to UNICEF sector work (health, education, child protection etc.). MODA challenges this traditional approach, encouraging country offices to take a whole child approach to programming. The needs of the most deprived children are multidisciplinary and need a network of services to address them simultaneously. While the result of sector-focused programming can be very effective in serving the specific cohort of children targeted, the approach has limitations. Many children experience multiple deprivations and while fragmented sector responses may deal with one of those problems, they rarely provide comprehensive solutions. Hence, a sector focused approach loses out on opportunities to provide more joined up, holistic and cost-effective support in reaching the most vulnerable and disadvantaged children. MODA demonstrates the child deprivation dimensions that most commonly overlap, allowing for UNICEF country offices to prioritize their programme interventions accordingly. MODA demonstrates that deprived children rarely experience only one deprivation, and this is particularly true for the dimensions that are measured directly at the child level. Specifically, children deprived of nutrition and health are commonly also deprived in other dimensions. For example, the Tanzania CC-MODA demonstrates that 16% of children under the age of five are simultaneously deprived in health, nutrition and water. The feedback loops between health, water and nutrition are by now well established, and for effective UNICEF programming the emphasis needs to be put on addressing all three simultaneously.

CC-MODA includes low- and middle-income countries that have publicly available DHS or MICS data from survey years 2007-2012. CC-MODA is an ongoing project, and is updated regularly with new results when analyses are performed and/or additional datasets become available.

N-MODA includes low- and middle-income countries that have performed a country-specific MODA analysis.

Some countries have started to carry out N-MODA studies for their countries, or have initiated an analysis similar to CC-MODA but with specific additional aspects such as studying the changes over time or allowing a simultaneous analysis for financial poverty and deprivations. The N-MODA part of the web-portal will gradually be filled with new analyses and data that become available for specific countries.

While for the countries with Multiple Indicator Cluster Survey (MICS) data information on child discipline is available, for those with Demographic and Health surveys (DHS) domestic violence towards female household members has been measured by using stand-alone, optional modules. This means that not all countries with DHS data that are included in CC-MODA have information with regards to domestic violence. Due to the importance of protection from violence to child well-being, it has been chosen to report the estimates of this dimension for those countries that have available data. Since CC-MODA requires a harmonization of dimensions to assure comparability, the cross-country analysis comprises two groups of countries: those with six dimensions, and those with five dimensions (the protection from violence dimension being unavailable for the latter). When comparing CC-MODA results across countries, these two groups of countries cannot be compared with each other as the number of dimensions used for the analysis is different.

Even though the main datasets (DHS and MICS) used for CC-MODA are harmonized to a large extent, differences exist in the availability of individual and household characteristics. DHS reports on household access to social insurance (mother’s health insurance coverage) and intra-household decision-making on spending (mother’s power in spending decisions). MICS does not have information on these household characteristics, but has four other profiling variables available at individual level. For children below the age of five, MICS provides information on the attendance of early childhood education (ECE) and child-adult interaction with regards to learning activities, such as reading, singing and story-telling. For children who are above the age of five, MICS reports on child labor and early marriage for girls.

Due to data limitations, the observations (i.e., the children included in the analysis) may have missing information on one or more indicator or dimension analyzed. The treatment of these missing values will unavoidably introduce a bias, depending on how these values are treated. In CC-MODA, for the single dimension analysis missing values are kept as missing by excluding the observations with missing data in order to provide the most precise and transparent results on each indicator and dimension as they are calculated separately. When carrying out multiple deprivation analysis, however, the missing values are treated as ‘non-deprived’. This is necessary in order to make the denominators of the dimensions comparable across dimensions. In this way, the analysis can be carried out for as many children as possible by using the available information of each child (unless information on all applicable dimensions is missing, in which case the observation is dropped). Moreover, observations are not deleted from the analysis to avoid introducing exclusion error. Dropping observations with missing values on one or more dimensions would introduce an unknown bias if the missing values were not distributed equally across the sample. In order not to distort the construction and the representativeness of the sample the deletion of observations is minimized.

It is acknowledged that the treatment of missing values creates an unavoidable bias in the analysis. The direction of the bias can however be assessed. The web portal provides a table showing the percentage of missing values per indicator and per dimension for each country in order to indicate the possible extent to which there may be an underestimation of the deprivation levels (see the Data availability section on the web portal). Additionally, countries or age groups that have an incidence of missing values of more than one third on any of the dimensions are excluded from the analysis. Such countries and age groups cannot be included because the comparability across countries can be seriously impaired if a large proportion of children is classified as non-deprived due to data treatment. Furthermore, a sensitivity analysis has been carried out showing how the results would change if the observations with missing values on two or more dimensions were excluded from the analysis, or if the missing values were treated as ‘deprived’ instead of ‘non-deprived’ (see the Sensitivity analysis section on the web portal).