Jump to ContentJump to Main Navigation
Child poverty, evidence and policyMainstreaming children in international development$

Nicola A. Jones

Print publication date: 2011

Print ISBN-13: 9781847424464

Published to Policy Press Scholarship Online: March 2012

DOI: 10.1332/policypress/9781847424464.001.0001

Show Summary Details
Page of

PRINTED FROM POLICY PRESS SCHOLARSHIP ONLINE (www.policypress.universitypressscholarship.com). (c) Copyright Policy Press, 2021. All Rights Reserved. An individual user may print out a PDF of a single chapter of a monograph in PPSO for personal use.date: 05 August 2021

Knowledge generation and child poverty and well-being

Knowledge generation and child poverty and well-being

(p.25) Two Knowledge generation and child poverty and well-being
Child poverty, evidence and policy

Nicola Jones

Andy Sumner

Policy Press

Abstract and Keywords

This chapter extends 3D thinking on child poverty and well-being to consider knowledge and evidence generation. It explores the knowledge base that underpins dominant understandings of childhood poverty and well-being, given this book's focus on the knowledge–policy interface around childhoods in the developing world. The chapter reviews trends in thinking about the generation of evidence in development studies, and then discusses how these broader debates have played out in the case of childhood poverty and well-being. It focuses particularly on emerging thinking about 3D approaches via combining research methods, and explores the particular challenges and opportunities of such approaches to child well-being.

Keywords:   3D thinking, child poverty, knowledge–policy interface, generation of evidence

2.1 Introduction

In this chapter, we extend our 3D thinking on child poverty and well-being to consider knowledge and evidence generation. Given this book’s focus on the knowledge–policy interface around childhoods in the developing world, it is critical that we explore the knowledge base that underpins dominant understandings of childhood poverty and well-being. Questions that we need to consider include: how are multiple forms of ‘expertise’ accommodated? Whose ‘evidence’ prevails – that is, is perceived to be most credible – in policy debates and why? What accounts for the prevailing underinvestment in evidence about children, especially in developing countries? This chapter begins by briefly reviewing trends in thinking about the generation of evidence in development studies, and then discusses how these broader debates have played out in the case of childhood poverty and well-being. We focus particularly on emerging thinking about 3D approaches via combining research methods, and explore the particular challenges and opportunities of such approaches to child well-being.1

This chapter is structured as follows. Section 2.2 discusses the concepts of evidence-based policy. Section 2.3 is about knowledge generation and children. Section 2.4 discusses 3D approaches in general via mixing methods. Section 2.5 applies this debate to child poverty and well-being, and Section 2.6 concludes.

2.2 The evidence-based policy movement and its critics

To understand the interaction between knowledge and policy processes, it is important to look at the role of different types of knowledge in development policy. Early work on the link between knowledge and policy focused predominantly on the rational role of science and research (Laswell and Lerner, 1951), conceptualising the policy arena as one that ideally facilitates the production and application of technical expertise to solve policy problems. Models then began to incorporate an (p.26) understanding of the pragmatic, political and often opportunistic ways in which policymakers draw on different sources. Here, the evidence-based policy movement sought to develop frameworks to understand the drivers of and barriers to research ‘uptake’ (Cracknell, 2001; Landry et al, 2003; Brehaut and Juzwishin, 2005; Young, 2005; Ammons and Rivenbark, 2008; Moynihan et al, 2008) with the normative goal of increasing the influence of research on policymaking. However, this school of thought has come under criticism for ignoring the political and epistemological dynamics of the production and use of particular sources of knowledge (Luke, 2003; Marston and Watts, 2003; Sanderson, 2004; Simons, 2004). By focusing on evidence (instead of the more complex meaning of knowledge), a value-free world versus the value-laden world of politics has been overemphasised:

[W]e need to work within a broader conception of rationality to recognize the validity of the range of forms of intelligence that underpin ‘practical wisdom’, to acknowledge the essential role of fallible processes of crafting judgement in assembling what is to be accepted as ‘evidence’, and to incorporate deliberation, debate and argumentation in relation to the ends of policy and the ethical and moral implications of alternative courses of action. (Sanderson, 2004: 376)

As such, there has been a tendency for evidence-based policy studies, especially in the health sciences and economics, to prioritise some research techniques over others, setting experimental methods as the ‘gold standard’2 and paying lesser attention to more qualitative and participatory sources, such as public service users’ views and local knowledge (Tilley and Laycock, 2000; Rycroft-Malone et al, 2004; see also Box 2.1). As we discuss later, there has been a move to address this bias, with participatory development explicitly seeking to counterbalance the top-down production of evidence by enabling bottom-up data generation. In this vein, rural development and poverty-reduction policies (such as PRSP processes) have seen systematic attempts to integrate local knowledge into the policymaking process (Jones and Villar, 2008; see also Chapter 4).



Table 2.1 Alternative quality criteria

Traditional scientific criteria (i.e. positivist)

Social constructivist criteria (i.e. relativist)

Artistic and evocative criteria

Critical change criteria (neo-Marxist, some feminist)

Evaluation standards and principles

  • Objectivity (attempts to minimise bias)

  • Validity of data

  • Systematic rigour of fieldwork practices

  • Triangulation (for consistency of findings)

  • Reliability of coding and pattern analysis (multiple coders)

  • Correspondence of findings to reality

  • Strength of evidence supporting causal hypotheses

  • Generalisability

  • Contributions to theory

  • Subjectivity acknowledged and embraced

  • Trustworthiness and authenticity – fairness and coverage of others’ perspectives

  • Triangulation (to capture multiple perspectives)

  • Reflexivity and praxis – understanding one’s own background and how to act in the world

  • Particularity – doing justice to unique cases

  • Contributions to dialogue – encouraging multiple perspectives

  • Opens the world to us in some way

  • Creativity

  • Aesthetic quality

  • Interpretive vitality

  • Flows from self – embedded in lived experience

  • Stimulating

  • Provocative

  • Connects with and moves the audience

  • Voice is distinct and expressive

  • Feels ‘true’, ‘authentic’, and ‘real’

  • Case studies become literary works, blurring boundaries

  • Critical perspectives – increases consciousness of injustice

  • Identifies nature and sources of inequalities and injustice

  • Represents the perspective of the less powerful

  • Makes visible the ways in which those with more power exercise and benefit from this power

  • Engages those with less power respectfully and collaboratively

  • Builds capacity of those involved to take action

  • Identifies potential changemaking strategies

  • Consequential or catalytic validity

  • Utility – no point doing it if it is not going to be useful to some audience

  • Feasibility – no point if not practically or politically doable

  • Propriety – fair and ethical

  • Accuracy

  • Systematic inquiry

  • Integrity/honesty and respect for people

  • Responsibility to general public welfare

Source: Sumner and Tribe (2008), adapted from Patton (2002: 544).

(p.29) A second important critique of the evidence-based policy approach concerns the understanding of power. By emphasising the instrumental role of research and its direct impact on policy decisions, there is a danger of obscuring the ‘invisible’ ways in which power is exercised. Gaventa (2006), based on the seminal work of Lukes (1974), focuses on the importance of the ‘third dimension’ of power – that is, the way in which power operates to shape people’s preferences – and the role of policy discourses in influencing the parameters of policy decisionmaking. As such, both scholars and practitioners in this field increasingly appreciate that scientific knowledge is but one of several rival forces shaping the policy process, and subscribe to an evidence-inspired policy-making approach (Duncan, 2005), making research a key building block of policy formulation (Nutley et al, 2002).

Finally, there is a growing debate over the politicisation of science – as highlighted by recent heated debates on the nature and intensity of climate change. Monaghan (2008: 145) argues that evidence-based policy may at times be policy-based evidence, which he defines as ‘the cherry-picking of favourable evidence to support an established policy position’.

2.3 Approaches to evidence generation on child poverty and well-being

Historically evidence about childhood and childhood poverty in developing countries has been dominated by quantitative assessments of children’s human capital development, such as nutritional status and school enrolment. Over the last decade and a half, in part inspired by the UNCRC, a much richer body of knowledge has emerged, including child-centred approaches that incorporate children’s participation. As a result, the debates about evidence quality and what knowledge counts as ‘legitimate’ or ‘most persuasive’ discussed in the previous section have also surfaced in the field of childhood studies.

Quantitative researchers have focused on measuring the extent and causes of childhood poverty, especially infant mortality rates, child malnutrition (using anthropometric data), educational attainment and achievement,3 and involvement in harmful forms of child labour (recent noteworthy examples include Cockburn, 2002; Gordon et al, 2004). These researchers have sought to address the disjuncture between childhood and adult/household-level poverty, especially as ‘traditional’ proxy monetary measures of poverty and sources of data such as income and consumption are problematic for children for the reasons noted in Chapter 1 (such as the possibility that non-market channels are (p.30) more important in shaping childhood poverty; children’s access to, and control of, income can be extremely marginal; and the likelihood that resources and power are distributed unequally within households and thus disaggregated data are essential – see Box 2.2).

Qualitative researchers have, by contrast, engaged less with discourses of poverty reduction, instead focusing more on aspects of well-being, including care, nurture, resilience, capabilities, rights, social capital, the creation of gendered identities and opportunities for participation and decision-making, among others (noteworthy examples include Graue and Walsh, 1998; Woodhead, 1999; Lloyd-Smith and Tarr, 2000; White, S., 2002). The ‘thick description’ and nuanced insights of qualitative analysis provide an understanding of the intra-household dynamics and/or social processes behind the numbers. In the case of participatory research, it also enables an understanding of children’s experiences and perceptions of various forms of deprivation and vulnerability. This is critical as it shifts policy debates from preparing for children’s future (p.31) ‘well becoming’ to working towards their current ‘well-being’ (Ben-Arieh, 2006).4

Until relatively recently, quantitative and qualitative approaches to researching child poverty tended to be published in different types of journals – economics, epidemiology and development studies versus childhood studies, sociology, anthropology and gender studies, respectively – with relatively little crossover between the two. Gradually, however, links are being forged between the two disciplinary/methodological clusters, and mixed methods research on childhood well-being is emerging.

A central argument of this chapter is that the distinctiveness of children’s well-being (noted in Chapter 1) means research on this topic in particular benefits from mixing methods and combining quantitative and qualitative analysis. In Chapter 1 we concluded that relational and subjective dimensions of children’s well-being have been underemphasised, and that they matter when considering children’s voice, vision and visibility and play a role in determining material dimensions of children’s well-being. To recap, we concluded that childhood poverty and well-being are distinct from adult poverty and well-being on a number of levels. Better understanding of the dynamics and processes that might reinforce or reverse patterns of disadvantage or benefit is a matter of urgency in light of the growing body of scholarship on the life-course and the intergenerational impacts of childhood poverty.

Children have differing needs, wants and capacities depending on their stage of childhood (e.g. infancy, early childhood, middle childhood and adolescence). Over the course of the first decades of life, children undergo certain physical and neurological transformations that give rise to evolving capacities over time (e.g. Lansdown, 2005). These are, however, experienced in diverse ways as children are a heterogeneous group living in divergent socio-economic conditions with distinct needs and concerns. As Wood (1985: 2) argues, ‘Children become “cases” which are “disorganised” from their own context and “reorganised” into the categories given by development intervention’. Whereas there is broad acceptance of ‘gender’ and ‘sexuality’ as social divisions that are not natural or ‘god-given’, but culturally constructed, recognition of childhood as a culturally constructed phenomenon, whereby children in different cultural contexts have divergent sets of rights and responsibilities, is more recent and little explored outside childhood studies circles (e.g. James and Prout, 1990; Platt, 2003). As such, processes of evidence generation need to take account of the (p.32) complexities of childhood biological, neurological, social and moral development (e.g. Ridge, 2002; Yaqub, 2002).

We can posit that a further key differentiating experience is that childhood poverty and well-being are more intensely relational in nature: there is greater reliance on ‘others’, typically adults, for care and nurture; greater physiological and psychological vulnerability; and reduced autonomy/power – that is, children’s conventional voicelessness has a particular quality and intensity. In recent years scholars have paid increasing attention to the relational nature of well-being and the importance of care, especially for young children and the elderly (e.g. Lewis, 2002; White, S., 2002; Folbre and Bittman, 2004). In order to understand child well-being, exploring intra-household dynamics and arrangements of care is critical given children’s greater vulnerability and reliance on (usually) adult care (Marshall, 2003). However, as research on child-headed households and the gendered dimensions of child work has underscored, intra-household dynamics often entail children, especially girls, shouldering part of the care management burden, especially in large impoverished households (e.g. Kabeer, 2003). Although analyses of care dynamics usually lend themselves more readily to qualitative approaches, feminist economists are increasingly seeking to explore the impacts of intra-household allocations of resources and power quantitatively in order to draw greater policy attention to the political economy of care (e.g. Folbre, 2006).

Furthermore, a considerable body of research evidence has emphasised the ways children are situated and influenced not only by their household environment, but also by their neighbourhood, school and society (e.g. Becker et al, 1998; Ruel et al, 1999). Although the current emphasis on children as ‘participant agents’ in social relations (Mayall, 2002) who shape their circumstances and social structure is a necessary correction to conceptualisations of children as passive targets of social intervention, child well-being is ultimately more dependent on community and social influences than that of adults due to their unequal power and prevailing decision-making relations. As Sarah White (2002: 2) argues, ‘“child-centred” development practices must not be “child-only”: social and economic justice for poor children must be tackled in the context of their families and communities’. In light of the above, this book argues that a 3D approach to knowledge generation about child poverty and well-being is essential to accurately capture and understand the complexities of child poverty and well-being.

(p.33) 2.4 Mixed-methods knowledge on childhood

To accept and promote cross-disciplinary approaches implies openness to the use of all available insights to gain a better understanding of phenomena.5 Labels such as ‘qual–quant’, ‘q-squared’ or ‘q-integrated’ might suggest that mixed methods simply entails taking a quantitative method and adding a qualitative method, giving equal weight to each. However, there are numerous possible combinations, each imbued with assumptions regarding the respective roles, relative importance and desired sequencing of qualitative or quantitative methods.

At the outset, it is worth reminding ourselves what the terms ‘qualitative’ and ‘quantitative’ are used to refer to:

  • types of methodology – the overall research strategy used to address the research questions or hypotheses;

  • types of methods of data collection – that is, the specific methods;

  • types of data collected – that is, the raw data;

  • types of data analysis – that is, the techniques of analysis; and

  • types of data output – that is, the data in the final report or study.

With regard to poverty research, Carvalho and White (1997: 1) characterise quantitative and qualitative approaches as follows:

The quantitative approach … typically uses random sample surveys and structured interviews to collect the data – mainly, quantifiable data – and analyzes it using statistical techniques. By contrast, the qualitative approach … typically uses purposive sampling and semi-structured or interactive interviews to collect the data – mainly, data relating to people’s judgment, preferences, priorities, and/or perceptions about a subject – and analyzes it usually through sociological or anthropological research techniques.

Qualitative methods can also produce quantitative data, although the opposite is not true. Moser (2003), for instance, has championed the need for ‘apt illustration’ (as compared to anecdotal evidence) through quantifiable qualitative research:

[There is a need to shift] goalposts as to the definition of robustness so that it becomes more ‘inclusive’ of quantifiable qualitative research. Only this can ensure that social issues do not remain confined to anecdotal boxes, but provide (p.34) information of equal comparability in poverty assessments. (2003: 82)

Moser’s work on violence in Colombia and Guatemala, which quantifies and categorises insights from participatory research with hundreds of urban poor people, was designed to break down the divide between researchers and policymakers and make information about the complexities of people’s experiences ‘accessible to more policymakers not only within the research countries but also in a broader context’ (2004: 3).6

However, there is no guarantee that different approaches, methods or data will be immediately synthesisable or even comparable. It is interesting to question how one adjudicates situations when the evidence is contradictory. Mixing might have different functions – to enrich, explain or even contradict, rather than to confirm or refute. It may even tell ‘different stories’ on the same subject, because quantitative methods are good for specifying relationships (i.e. describing) and qualitative methods for explaining and understanding relationships (Thomas and Johnson, 2002: 1).

Brannen (2005: 12–14) lists four functions of combining methods:7

  • elaboration or expansion (‘the use of one type of data analysis adds to the understanding being gained by another’);

  • initiation (‘the use of a first method sparks new hypotheses or research questions that can be pursued using a different method’);

  • complementary (‘the data analyses from the two methods are juxtaposed and generate complementary insights that together create a bigger picture’); and

  • contradictions (‘simply juxtapose the contradictions for others to explore in further research’).

One concrete example of mixing can be taken from poverty researchers who have sought to combine quantitative approaches (to determine the amount of poverty and where it is) and qualitative approaches (to identify the causes and dynamics of poverty). They have done so by combining household surveys and case studies from participatory poverty assessments (PPAs). Table 2.2 sets out selected generic strengths and weaknesses of surveys and PPAs.

Combination may take place during data collection by simultaneously conducting a survey and a PPA in the same sample, or during the data-analysis stage by merging results and/or synthesising findings into one set of recommendations (see Table 2.3).


Table 2.2 Selected possible generic strengths and weaknesses of PPAs and surveys




  • Richer definition of poverty

  • More insights into causal processes

  • Holistic – a set of relationships as a whole, not pre-selected attributes

  • Scope for attention to processes as well as snapshots of the situation

  • Feedback loop – new/more interviews for interrogating data

  • Focus on context and people’s experiences

  • Lack of generalisability (but the sample can be made more or less representative of the population)

  • Difficulty of verifying information

  • Limited systematic disaggregation

  • Possibility of unrepresentative participation

  • Potential for agenda framing by facilitators

  • Pitfalls in attitudinal data – arrival of a PPA team changes people’s behaviour

Household surveys

  • Aggregation and comparisons possible across time and with other data sets

  • Reliability of results measurable

  • Credibility of numbers with policymakers

  • Credibility of national statistics with policymakers

  • Allows simulation of different policy options

  • Correlations identify associations, raising questions of causality

  • Misses what is not easily quantifiable

  • Sampling frame may miss significant members of the population

  • May fail to capture intra-household resource allocation

  • Assumes that numbers are objective and conclusive

  • Assumes that one question means the same thing in different cultural contexts

Sources: Appleton and Booth (2001), Carvalho and White (1997) and Chambers (2003).

Table 2.3 Selected examples of combining qualitative and quantitative data collection and analysis




Stage of research process

Data collection

  1. Conduct a simultaneous survey and PPA in the same sample (ideally nationally representative)

  1. Use surveys to identify subgroups for PPAs or use PPAs to identify survey questions

Data analysis

  1. Synthesise findings into one set of results or merge outcomes from mixed teams of qualitative and quantitative researchers

  • Use PPAs to confirm or refute the validity of surveys (or vice versa)

  • Use PPAs to enrich or explain information on processes in survey variables (or vice versa)

Sources: Constructed and expanded from text in Carvalho and White (1997), Shaffer (2003) and Thorbecke (2003).

At a more sophisticated level, integration might take place at the data-collection stage by using surveys to identify subgroups within PPAs or using PPAs to identify survey questions. At the data-analysis stage, integration could take place by PPAs and surveys confirming or (p.36) refuting each other (e.g. using PPAs to confirm the validity of surveys, or vice versa), or by PPAs and surveys enriching/explaining each other’s findings (e.g. using PPAs to obtain information on processes underpinning survey variables, or vice versa). In sum, the researcher needs to consider two questions, both of which are informed by the type of research problem and the question and/or hypothesis under investigation. First, is the ‘dominant’ method – that which will yield most of the data – qualitative or quantitative? Second, are the methods to be mixed sequentially or simultaneously?

So how do these methodological concerns play out in research on child well-being? We begin by discussing the relatively new field of participatory research with children, and then discuss examples of research where the mixing of methods was used to ‘initiate’ (generate new hypotheses), ‘expand’, ‘combine’ or ‘contradict’ the findings generated through a different methodological approach.

In order to capture the particular quality of children’s ‘voices’, qualitative researchers interested in childhood have used participatory research methods such as play, song, drawing and photography to explore conventionally silenced perspectives (Alfini, 2006; see also Table 2.5 at the end of this chapter). As Selener (1997: 2) argues:

The inclusion of direct testimony in the development debate can help to make it less of a monologue and more of a dialogue, as people’s testimony begins to require answers and as their voices force the development establishment to be more accountable for their actions.

As discussed in Chapter 1, while adult researchers may emphasise children’s health, nutritional and scholastic outcomes, participatory research with children suggests that relational and subjective deprivations are often equally important concerns, including insufficient time to play, lack of affection from family members, feelings of social exclusion by peers and shabby and/or dirty clothing.

The degree and duration of children’s involvement in research varies considerably. Cahill (2007) argues that research with children is distinct from research on children, an argument resonating with the claim that children as project/research beneficiaries are quite different from children genuinely involved in decision-making and/or knowledge-generation processes. The extent of children’s involvement in the research process depends largely on the perceptions of, and the methodology adopted by, the researcher(s). The majority of research is never truly participatory in the sense that children are not conferred (p.37) equal ‘co-researcher’ status – partnerships are usually characterised by power imbalances (see Cahill, 2007). This is perhaps largely due to embedded discourses within the academy regarding children’s capacities and reliability in a research context (Ben-Arieh, 2005), logistical obstacles to such long-term equitable engagement and, interestingly, the way in which participation is defined and interpreted by development agencies and professionals (White and Choudhury, 2007). In the latter case, it is suggested that as participation does not (as is sometimes claimed) unproblematically bypass issues of power, but rather introduces a new set of power relations, it can seldom be said that what is heard is an ‘authentic’ children’s voice as discussed in Chapter 1.

Nonetheless, researchers undertaking studies with children have made significant efforts to even out the power differentials inherent in all research contexts. These include: involving children in the initial design of the research project, thereby transferring a sense of ownership over the process; employing not just one participatory technique, but a range of methods to ensure that children with differing skills, capacities and experiences can participate (particularly important for preventing the reproduction of patterns of social exclusion); tailoring the degree of adult facilitation in accordance with the nature of the research/techniques used; and taking creative and innovative approaches to overcome cultural and language barriers, such as the use of visual methods (Biggeri et al, 2006). A number of studies also speak of the importance of continued child involvement after the ‘formal’ research process has ended. Through a reciprocal approach (Pham and Jones, 2005), it is possible to involve child participants in processes of active research dissemination whereby findings are communicated directly to both local communities and key decision-makers (van Blerk and Ansell, 2007). If participation is genuinely about empowerment, and more importantly sustained, long-term empowerment, then researchers must ensure not only that children attain valuable skills and experiences through the process, but also that they – and their families and communities – gain a better understanding of the issues researched, and are given opportunities to use the findings to push for action and change.

Participatory approaches can be mixed with other approaches, resulting in – we would argue – a 3D approach to researching child well-being whereby research seeks to capture where, how and/or why children’s well-being may be distinct from adult well-being in the material, relational and subjective dimensions (see Table 1.5). In this regard, a 3D understanding of child well-being clearly needs to pay particular attention to the temporal dimensions of child outcomes (p.38) and experiences. This is necessary if researchers are to advance understanding about children’s evolving capacities, as well as life-course and intergenerational poverty transfers. These research areas are methodologically challenging, especially as there are frequently significant longitudinal data limitations in the developing world. However, examples drawing on Northern longitudinal data sets suggest that a combination of quantitative analysis of panel data with qualitative analysis of oral life histories from a purposefully selected subsample can be a fruitful approach to capturing both objective and subjective changes in well-being over different life stages (Holland et al, 2006).8 One of the better-known examples of such an approach is Thompson’s (2004) research on stepfamilies, which he argues ‘brings together the strengths of both qualitative and quantitative methods in a middle way, using two eyes instead of one, embedded in a dichotomized approach’ (quoted in Holland et al, 2006: 13). More specifically, he combines a life-course study using in-depth interviews and mental health histories from adults who grew up in stepfamilies, as well as a census-based national quota sample and data from a quantitative study on coping strategies used by adults and children living in stepfamilies drawn from the UK’s longitudinal National Child Development Study (initiated in the 1950s). This mixed methods approach enabled Thompson to identify key life moments linked to experiences in stepfamily environments that were largely missed in the quantitative surveys, but which shed valuable new insights on quantitative data patterns.

In researching the relational dimension of well-being, capturing the complexities of intra-household and intra-community relations necessitates a multi-pronged methodological approach and multiple data sources. Two examples from the National Poverty Centre at the University of Michigan provide creative solutions that could be adapted to developing-country contexts (see Table 2.4). The first tackles the influence of intra-household distribution of resources and power on child material well-being. Magnuson and Smeeding’s (2005) work on the relative impact of different sources of income, state benefits and intra- and inter-family transfers in lifting young families out of poverty drew on a nationally representative birth cohort study (Fragile Families and Child Well-being Study). They complemented their quantitative analysis with a follow-up qualitative study involving in-depth, semi-structured interviews with parents, first as couples and then individually. Whereas the quantitative data provided a robust picture of household economic trends over time, the qualitative research explored the complexities and subjective effects of co-residing with parents, where (p.39)

Table 2.4: Examples of combining qualitative and quantitative data collection and analysis on childhood well-being




Stage of research process

Data collection

  1. Children from migrant households’ educational experiences in Mongolia (Batbaatar et al, 2005)

  • Impacts of poverty reduction strategies on child work and education in Ethiopia (Woldehanna et al, 2005a, 2005b)

  • Relative impact of different sources of income, state benefits, and intra- and inter-family transfers in lifting young families out of poverty (Magnuson and Smeding, 2005)

Data analysis

  1. Impacts of poverty reduction strategies on child work and education in Ethiopia (Woldehanna et al, 2005a, 2005b)

  • The role of neighbourhood poverty status in shaping youth risk behaviour (Clampet-Lundquist et al, 2005)

  • Experiences of growing up in stepfamilies (Thompson, 2004)

much-needed financial support was balanced against loss of space, privacy and, in some cases, decision-making power.

A second example focuses on the role that children’s and young people’s communities play in shaping their subjective well-being. In order to better understand the relative importance of neighbourhood poverty on youth risk behaviour, Clampet-Lindquist et al (2005) combined a longitudinal panel study with a random stratified subsample of retrospective qualitative interviews. These focused on different dimensions of male and female youth experiences in moving from highly deprived to less poor neighbourhoods. The data were creatively complemented by interviews with a control group (youth who had not moved), as well as friends of the ‘movers’, to explore similarities and differences in behavioural patterns. Whereas the quantitative data showed that moving had no or even a negative impact for males (but not females), the qualitative methods identified key additional themes such as the protective role that gender norms play in keeping girls closer to the house and under closer supervision, and the negative stereotypes to which young African-American men are subject to and react against. The researchers then used these themes to generate hypotheses for more detailed follow-up work.

In researching the subjective domain, careful ethnographic and participatory research has an important role to play in highlighting the diversity of and, especially, the cultural constructedness of childhood. James et al (1997) emphasise that such work needs to be approached in (p.40) a sensitive manner in order to balance cultural relativism and universal principles. Here a mixed methods approach might be able to provide the authority, moral weight and nuanced approach that James et al advocate. For instance, quantitative survey data on the incidence of child labour can be used to draw attention to the extent of involvement in harmful forms of child work, while qualitative research can capture the complex ways in which children, their families and communities ascribe meaning to work and the intra-household and socio-economic dynamics that need to be taken into account to eradicate exploitative forms of work in an effective and sustainable way. Woldehanna et al’s (2005a) work on children’s paid and unpaid work in Ethiopia is one example of such a mixed methods approach, which was used for policy engagement purposes during the country’s second PRSP (Jones et al, 2009).

2.5 Conclusions

This chapter highlighted the importance of understanding debates about the process of generating evidence or knowledge that underpins key policy and practice decisions, and how these play out with regard to childhood poverty and well-being in developing country contexts. It emphasised that evidence is not a neutral concept, but is embedded within a set of power relations between knowledge producers and knowledge users. This is arguably particularly true in the case of evidence about childhood well-being, as children’s perspectives are too often hidden or silenced in mainstream development debates. Indeed, the very process of being involved in a participatory research endeavour may open up new and potentially profound possibilities for children and change how they interact in their social worlds.

In order to capture the complexities of children’s material, relational and subjective well-being, the second half of the chapter argued that a mixed methods approach combining quantitative and qualitative (both participatory and ethnographic) methodologies can best capture the multidimensionality and heterogeneity of childhoods and, indeed, lend new weight to the urgency of investing in genuinely cross-disciplinary approaches. In this vein, there is a particular need to develop more sophisticated methodologies for capturing intra-household dynamics, community–child relations and macro–micro policy linkages.

(p.41) (p.42) (p.43) (p.44) (p.45) (p.46) (p.47) (p.48) (p.49) (p.50)

Table 2.5: Selected studies on children’s participation in research and the ways in which power relations are understood

Factors mediating children’s participation

Reference (author(s)/date)

Definition/meaning of participation taken by author(s)

Type of study: conceptual or empirical (if empirical then nature of study methodology) and scope (i.e. country coverage)

Power as narratives/discourses

Power as actors/networks

Power as context/institutions

Altken and Herman (2009)

  • Children should be involved in research design: the number of children living independently of adults is evidence of the potential competence of children and their agency in plans for improving livelihood

  • Participation can take many forms and operate at different levels in the research process

  1. Conceptual

  • Perceptions of children’s ‘reduced capacity’ to understand the consequence of actions, including participation in research activities

  • Conceptualisations of children as legal subjects

  • Potential disregard of children because their involvement does not fit the ideals of what is ‘normal’ and ‘good’ for a child

  • Operational ambiguity of ‘participation’

  • Perceptions that provision and protection should be privileged over participation

  • Alms and motivations of researchers

  • Consent of children/parents

  • Who is participation for?

  • Types of methods implemented by researchers: do children get a choice?

  • Relevancy of questions asked to children

  • Payment of participants

  • Skills of children as researchers

  • Ethical considerations

  • Issues of confidentiality and privacy

  • Stage at which children are involved in the research process

  • Children’s relations to wider population

Ben-Arieh (2005)

  • Children’s active involvement in the study of their well-being – a natural consequence of the concept of child rights

  • Conceptual

  • Perceptions that: i) children should not take part in research due to age and low cognitive ability; and ii) children are unreliable and research with them produces low response rates

  • Whether follow-up activities are conducted by researchers

  • Parental consent

  • Power disparities between children and researchers

  • Stage at which children are involved in the research

  • Are children involved in the dissemination process?

  • Role of child participants (e.g. data collection, data analysis, etc.)

  • Anonymity of research

Biggeri et al (2006)

  • Capability-based approach: children have and can define their capabilities

  • Taking children’s views into account is important for overcoming assumptions inherent to adult-based understandings of children’s capabilities

  • Empirical: mixed methods approach looking at children’s conceptualisations of their own capabilities

  • At 2004 ‘Children’s World Congress on Child Labour’, 104 child delegates first surveyed, then participatory focus group with South Asian child delegates, and case studies through in-depth interviews

  • Prevailing attitudes towards children, based on the view that adults know best and act in children’s best interests

  • Age of child (different capabilities)

  • Motivations of adults: Using children to promote their own goals

  • Representativeness of child delegates/participants

  • Capacities of parents which influence those of their children

  • Decisions and actions of parents, guardians, and teachers

  • Expectations of researchers

  • Training/skills of interviewers

  • Children’s education

  • Ground rules of the research

  • Language barriers

Cahill (2007)

  • Involvement of participants (co-researchers) throughout every stage of the research

  • Participants are active in all decision-making processes from deciding on the focus of the project to agenda setting and data analysis

  • Built on the idea that all people develop social theory through life experience

  • Conceptual/empirical: smallscale participatory action research project in New York involving six collaborators/coresearchers at every stage of the research

  • Dominant discourses embedded within academia about research with children

  • Methods for expression employed by researcher

  • Power relations between researcher and participants

  • Role of the facilitator

  • Capacities of participants

  • Tensions between producing academically robust and ‘legitimate’ results and effecting action

Camfield et al (2008)

  • A child-focused perspective that emphasises participation provides a better understanding of children’s experiences, and creates opportunities for them to contribute to discussions and interventions that affect their lives

  • Participation represents a valuable way of exploring children’s understandings of well-being

  • Conceptual

  • Contrasting perspectives of children and adults

  • Stereotyping and idealising the vulnerable and invisible

  • Power imbalances between child participants

  • Power of researcher

  • Types of methods utilised

  • Number of interviews conducted

  • Children’s role in the research

  • Is research quantitative or qualitative?

  • Power relations associated with setting the agenda and defining research questions

Crivello et al (2009)

  • Capturing children’s ‘standpoints’ in contexts of poverty and considering their views and experiences informs more effective and integrated interventions.

  • Children are both competent social actors and ‘experts in their own lives’, and therefore constitute valuable sources of data

  • Empirical: examination of three core qualitative methods for researching child wellbeing as adopted by Young Lives in Ethiopia, Peru, Vietnam, and India

  • Discourses informing construction of methodologies (i.e. emphasis on survival and deprivation, or on resilience and agency)

  • Age and varying capacities of child participants

  • Children’s memories

  • Types of methods/techniques employed by researchers (including icebreakers, games, physical movement)

  • Degree of facilitation by adults

  • Extent of interaction among children in exercises

  • Translation of concepts such as ‘well-being’ into different languages and cultures

  • Ethical issues associated with information-sharing in a group setting

  • If longitudinal study, importance of building rapports

Fattore et al (2007)

  • Children should be involved in defining their understandings, perspectives, and standpoints of and on well-being

  • Giving these views serious consideration, and incorporating them into the identification of key domains that can be leveraged to monitor and measure important aspects of wellbeing at a population level

  • Empirical: collaborative research project between New South Wales Commission for Children and Young People and University of Sydney (Australia)

  • Individual/group semistructured interviews with 126 children; in-depth interviews with 95 (of the original 126) children; task-oriented project with 56 children

  • Approaches to conceptualising/researching well-being

  • Policymakers’ views towards significance of research carried out with children

  • General reluctance to involve children at each stage of research and policy processes

  • Factors associated with local contexts

Gallagher (2008)

  • Foucauldian understanding of participation: power is relational and always involves a relationship between two entities

  • Participation does not entail the straightforward transfer of power from adult to child, but rather constitutes a messy, fraught and ambiguous process

  • Instances of participation are always different

  • Conceptual/empirical: author’s own experience in a Glaswegian school

  • Concept of the ‘powerless’ child

  • Perceptions of a ‘dominating adult–subordinate child’ relationship

  • Children may perceive participatory initiatives as unwelcome interventions

  • Romanticisations of child agency

  • Multiple complex relations in research settings among adult researchers and child participants

  • Power relations between child participants

  • Risk of domineering children’s views being privileged

  • Multiple points of resistance exist in research contexts

Huebner (2004)

  • Involving children and youth in studying and researching life satisfaction

  • Conceptual

  • ‘Scientific’ and ‘traditional’ psychological discourses informing methodological approaches to life satisfaction research

  • Novelty of the research field (i.e. if new, then many studies will be exploratory and appropriate methods may not yet have been refined)

Hill et al (2004)

  • Multidimensional participation that enhances children’s involvement in decisionmaking and combats social exclusion/promotes social inclusion

  • Key distinction drawn between participation and consultation

  • Conceptual (UK-oriented)

  • Effective child participation requires sustained links between research and policy domains

  • Quantitative research more likely to inform policy: less capacity for child involvement?

Mannion (2007)

  • Participation entails discovering and understanding children’s views (various rationales for this are proposed, including enlightenment, empowerment, and citizenship), and, importantly, is about child–adult relations

  • Conceptual/empirical: three UK case studies (school grounds changes, arts centre, ‘child-free’ zones)

  • Tendency to analyse children’s voices in isolation from their socio-spatial contexts

  • Nature of child–adult relations

  • Extent of intergenerational dialogues

Naker (2007)

  • Drawn from Hart’s ladder: The concept of child participation manifests in a variety of adult-child engagements within development practice

  • Power (re)distribution is key

  • Partnerships based on joint initiation and direction of processes

  • Empirical: assessment of child participation through an investigation of a project in Uganda involving an Advisory Committee of 40 children, aged 10–17, half girls and half boys

  • Understandings of the concept of participation

  • Adults’ uncertainties about their feelings regarding sharing decision-making power with children

  • Scepticism among children

  • Difficulties of integrating child participation into practice after years of NGOs operating in a particular way

  • Skills of facilitators and group leaders

  • Disagreements among researchers about the role of the committee

  • The ‘African way’ of relating to children: set social hierarchies

Powell and Smith (2009)

  • If children are to be involved in research, their views must be heard and taken seriously, even (or especially) if they challenge dominant protectionist and paternalistic discourses/systems. Participation is defined as taking part and the sense of knowing that one’s actions are taken note of and may be acted upon (see Boyden and Ennew, 1997, in Morrow, 1999: 149).

  • Empirical : email interviews with 12 researchers working on/with children, predominantly in New Zealand

  • Conceptions of children among researchers and gatekeepers

  • Infiltration of a protectionist discourse with regard to children

  • Failure to take child agency into account

  • Parental consent

  • Gatekeepers and subsequent selection problems

  • Types of methods used

  • Communication skills of researchers; ability to build relationships

  • Power imbalances in research

  • Choice of research topic–too sensitive to involve children?

  • Structural vulnerabilities

Redmond (2008)

  • Children are capable social actors whose perspectives serve an important policy purpose in terms of problem and coping strategy identification

  • Conceptual

  • Perception among policymakers and researchers that children are passive, apolitical, and uninvolved in strategy-making

  • Discourses of professionalism

  • Children exclude other children

  • Children’s family environment – does it promote resilience?

  • Outlook and confidence of children

  • Do researchers create ‘childfriendly’ spaces?

  • Types of techniques used

  • General social barrier separating children and adults

Ridge (2002)

  • Children are active social agents whose views must be reflected in policy decisions that affect them

  • Participation is defined in contrast to social exclusion and is considered a core element of a child-centred approach to development

  • Empirical: quantitative and qualitative methodology

  • In-depth survey conducted in 1999 of 40 children living in South-east England, and unstructured interviews that children had the opportunity to shape and define

  • Embedded adult assumptions about children’s lives

  • Perceptions that uncovered child knowledge will ‘threaten’ these assumptions

  • Children as human ‘beings’ or ‘becomings’

  • Children’s backgrounds

  • Age of participants

  • Gender of participants

  • Ethical and moral dilemmas surrounding research with children

Pham and Jones (2005)

  • Only if children’s knowledges are ‘uncovered’ will policies and programmes be designed in a way that is ‘responsive and relevant to their concerns and needs’ (Boyden and Ennew, 1997: 10)

  • Participation must embrace the diversity of children’s experiences

  • Conceptual/empirical: case study material from Young Lives project in Vietnam (Children’s Fora and Young Journalist Clubs)

  • Dominant Western models of childhood that infiltrate policymakers’ perceptions

  • Presence of networks/ channels linking children’s knowledge generation to policymakers

  • Evaluations of whether outcomes of participatory initiatives influence policymaking

  • Is research reciprocal?

  • Relationships between practitioners/researchers and children and their families

  • Analytical and communication skills among children

  • Whether stakeholder ‘buy-in’ is secured (from the outset)

  • Relationships with local collaborators and project partners

  • Clientalism in southern contexts. Embracing context is far more important than seeking standardised praxis or a grand theory

Van Blerk and Ansell (2007)

  • Adults cannot presume to have insight into children’s social and cultural worlds

  • Children are the experts in this sense and therefore have much to offer researchers in terms of knowledge generation

  • Participation of children in the research process results in their empowerment

  • Empirical: participatory feedback and dissemination experiences in Lesotho and Malawi: participatory learning and action (PLA) workshops, dramas, community discussions, workshops with policymakers

  • Perceptions of competencies of children

  • Views on the importance of dissemination

  • Are policymakers interested in attending dissemination workshops?

  • Ways in which researchers involve children in dissemination processes

  • Understandings and utilisations of policy networks

  • Relationships with policymakers

  • Time delays between original research and dissemination processes.

  • Difficulties associated with accessing street children and out-of-school children

White and Choudhury (2007)

  • Ideally, participation is about raising children’s voice in development matters

  • In reality, participation is ‘produced’ through the development industry via ‘projectisation’ of participation and promotion of agencysponsored events and programmes

  • Conceptual/empirical: case study drawn from authors’ extensive experience with Amra, a children’s organisation in Bangladesh

  • Primary data collected through intensive group and individual interviews, consultation, and field visits with various members (100 children), in addition to interviews with parents and staff in an international agency

  • Understandings of what counts as ‘participation’ as determined by development agency staff

  • Framing of the ‘separate’ children’s world regulates what can be said by the researcher about what is taking place

  • Children’s agency constrained and determined by adults in development agencies (i.e. what can be said, when it should be said).

  • Development workers have the power to reject researchers’ interpretations of child participation if they do not coincide with their own

  • Participation as a new relations of power

Woodhead and Faulkner (2000)

  • UNCRC-based definition: entails a degree of power devolvement that goes beyond simply ‘listening to children’

  • Children become competent through participatory research

  • Conceptual

  • Perceptions of children’s roles and capacities

  • Objective and dispassionate scientific discourses that inform methodology (e.g. ‘traditional’ approaches to psychological research frame children as subjects rather than co-researchers)

  • Power differentials and differences in status

  • Researchers’ methodological approaches


(1) This chapter draws on and develops ideas in Sumner and Tribe (2008). An earlier version of part of this chapter was published as Jones and Sumner (2009).

(2) The influential work of organisations such as the Cochrane Collaboration (healthcare) and the Campbell Collaboration (broader social policy, most notably criminal justice) has been key to the endorsement of quantitative and experimental techniques as the model to be followed in evidence-based policy initiatives.

(3) Commonly researched educational indicators include rates of school enrolment for girls and boys, over-age enrolment and results on standardised scholastic achievement tests.

(4) We are grateful to Laura Camfield for this observation.

(5) This section draws on Sumner and Tribe (2009: ch 5).

(6) Holland and Abeyasekera’s (forthcoming) work on ‘participatory numbers’ presents another innovative approach to producing quantitative data from qualitative methods. See also Mayoux and Chambers (2005).

(7) Further, Brannen (2005: 14) identifies 12 specific conceivable combinations (see below). In each, there is a ‘dominant’ method (i.e. the method that gathers the majority of the data) and a ‘non-dominant’ method (i.e. the method that gathers the minority of the data). Capital letters denote the ‘dominant’ method (which will yield the majority of data); + denotes simultaneously occurring methods; and 〈 denotes temporal sequencing of methods.

Simultaneous research designs:

  1. (1.) QUAL + quan or

  2. (2.) QUAL + QUAN

  3. (3.) QUAN + quan or

  4. (4.) QUAN + QUAN

  5. (5.) QUAL + qual or

  6. (6.) QUAL + QUAL

Sequential research designs:

  1. (1.) QUAL 〈 qual or

  2. (2.) qual 〈 QUAL or

  3. (3.) QUAL 〈 QUAL

  4. (4.) QUAN 〈 quan or

  5. (p.52) (5.) quan 〈 QUAN or

  6. (6.) QUAN 〈 QUAN

  7. (7.) QUAL 〈 quan or

  8. (8.) qual 〈 QUAN or

  9. (9.) QUAL 〈 QUAN

  10. (10.) QUAN 〈 qual or

  11. (11.) quan 〈 QUAL or

  12. (12.) QUAN 〈 QUAL

(8) Holland et al (2006) provide a number of examples on school transitions, youth to work transitions, post-divorce life and so on.