Poverty adversely impacts on the realisation of children’s rights to health and development. These rights are enshrined in international law in the United Nations Convention on Rights of the Child.[[1]] In developed countries, poverty can mean reduced access to nutritious food, lack of quality clothing such as warm clothes and sturdy shoes and reduced ability to see a doctor when needed.[[2]] Compared with their wealthier peers, children living in poverty are more likely to have poorer cognitive outcomes and school performance and are at an increased risk of antisocial behaviour and mental disorders.[[3]] These disadvantages cause flow-on effects such as poorer health and reduced employment prospects, resulting in broader social and economic costs due to expenditure on welfare, healthcare and criminal justice.

Poverty is a significant health and equity issue in Aotearoa New Zealand. Like many developed countries, poverty rates in Aotearoa New Zealand have been defined as the percentage of households that have disposable income less than 50% of the national median after housing costs.[[4]] Using this measure, 235,400 Aotearoa New Zealand children (20.8%) lived in poverty in 2018/19.[[4]] While Aotearoa New Zealand child poverty rates have gradually decreased since 2014/15,[[4]] child rates remain consistently higher than most other age groups and are almost twice the rate experienced in the 1980s.[[5]] Poverty in Aotearoa New Zealand is also ethnically patterned with Māori (the Indigenous population of Aotearoa New Zealand) and Pacific (mostly second generation migrants from Pacific Islands) children experiencing rates almost twice as high as NZ European children.[[4]] This consistent—and inequitable—burden of child poverty has been attributed to several social and economic factors, including the dominance of neoliberal (free trade) economics since the 1980s and a tendency for poverty to be framed as a minority ethnic issue.[[5]] The result is an environment that fails to support many children’s right to healthy development and breaches principles of Te Tiriti o Waitangi, Aotearoa New Zealand’s founding document, most notably the promotion of best health outcomes for Māori[[6]] and the United Nations Declaration on the Rights of Indigenous Peoples.[[7]]

Given the many consequences of child poverty, multidimensional research is important for documenting its effects.[[2]] Accounts of child poverty are largely based on self-report, report by adult caregivers, or household income surveys that typically only track the prevalence of child poverty. Often these accounts do not highlight children’s poverty from the child’s perspective. Moreover, the plight of children living in poverty is often difficult to convey to politicians and the wider population, as abstract numbers and references to disease and disadvantage do not translate easily into the consciousness of people for whom poverty is not a reality.

A new methodology developed by this research team, Kids’Cam, uses wearable cameras and Global Positioning Systems (GPS) to provide an objective and easily communicable account of children’s lived experiences.[[8,9]] Using this data set, we aimed to assess aspects of child poverty, including children’s access to household resources, household harms, their behaviours, and the nature of their built environments.

Methods

The Kids’Cam study

Kids’Cam was a cross-sectional observational study conducted in 2014/15. Full details on Kids’Cam’s methodology, including the recruitment strategy and sample, are available elsewhere.[[8,9]] In brief, 168 randomly selected children aged 11–13 years were recruited from 16 randomly selected schools in the Wellington Region of Aotearoa New Zealand. Recruitment was stratified by ethnicity and school decile based on aggregate school enrolment data from the Ministry of Education to enable equal explanatory power for ethnicity (Māori, Pacific and NZ European) and socio-economic deprivation subgroups.[[9]] Each child wore a wearable camera and GPS recorder around their neck for four consecutive days, recording an image every seven seconds and GPS location every five seconds. Children were asked to wear the devices for all waking hours, and to remove the camera in situations where privacy could be expected, if they felt uncomfortable, when swimming or playing vigorous sport, or if requested.[[9]] Ethical approval was obtained to study all aspects of children’s lives relevant to public health.[[8]]

Study sample

To compare the lived reality of child poverty, we selected a sub-sample of 78 Kids’Cam participants (46.4% of the original sample) from the lowest and highest levels of household socio-economic deprivation (NZiDep quintiles 1 and 5, respectively). NZiDep is based on eight questions relating to material and social deprivation,[[10]] categorised as: 1—zero deprivation characteristics (least deprived); 2—one deprivation characteristic; — two deprivation characteristics; 4—three or four deprivation characteristics; and 5—five or more deprivation characteristics (most deprived). There were more participants in the low socio-economic deprivation group (n=52) than the high socio-economic deprivation group (n=26), which reflects national trends for socio-economic deprivation.[[2]] There were similar age and gender distributions between groups, but fewer NZ European participants in the high deprivation group (Table 1).

Measures

We coded for a range of household resources, household harms, behaviours and built environment characteristics (see Appendix 1), using images captured on Thursday and Saturday during children’s leisure time. We defined leisure time as “all hours outside school time”, which includes Thursday morning before school, Thursday afternoon after school and Saturday. All data were recorded in a pre-formatted Excel spreadsheet in 2019.

Household resources included fruit, vegetables, educational materials, cognitive stimulation materials, “personal items” such as cell phones and games, physical activity equipment and play spaces, and the presence of computers, heating and fixed heating. We also coded for children’s sleeping arrangement (own room vs shared room). Household harms included structural deficiencies and mould in participants’ homes and the presence of alcohol. Behaviours included children’s consumption of “core” and “non-core” foods (defined using a nutrient profiling model)[[11]] and children’s participation in educational activities, structured physical activity and unstructured physical activity. Built environment characteristics included physical disorder in children’s neighbourhoods (defined as the area 500m—as the crow flies—from their residential address, based on children’s GPS data that is detailed elsewhere).[[12]] We coded for three types of disorder: fixed (vacant or dilapidated buildings); semi-fixed (graffiti and dilapidated lots with more easily fixed elements); and moveable (litter and abandoned items), based on definitions in a previous study.[[13]]

Statistical analysis

We used Stata/IC 15 for all statistical analyses. To compare differences by household socio-economic deprivation, we used negative binomial regression models to estimate mean rates and rate ratios for each poverty variable, using low socio-economic deprivation children as the reference group. For count-based variables, rates represent the mean number of unique items of each variable per household. Count-based variables include fruit, vegetables, educational materials, cognitive stimulation, structural deficiencies, physical activity equipment, play spaces and “personal items”. Our analysis of these variables represents “variety” in a household, as each item type was only counted once. For binary variables, rates represent the proportion of children’s households that had the variable present. Binary variables include alcohol presence, computer access, heating, fixed heating and “own room for sleeping”. For behaviours (food consumption and children’s participation in educational activities, structured physical activity and unstructured physical activity), rates represent the mean frequency of each behaviour divided by recording time (rescaled as a mean rate per hour). Likewise, mean rates for neighbourhood physical disorder represent participant’s frequency of exposure to disorder divided by the recording time in outdoor settings, rescaled as a mean rate per hour spent outdoors. We also explored the association between children’s exposure to physical disorder and neighbourhood-level deprivation (NZiDep), using low deprivation neighbourhoods as the reference group. Differences in poverty variables by ethnicity are described, but further analysis was not undertaken due to low numbers. All analyses accounted for the differential probability of selection into the study, using Stata’s svy commands and associated sampling weights.

Results

Children in this study captured a mean of 2,482 images during leisure time in the two-day recording period, equivalent to a mean of 4.8 hours, including 3.0 hours in homes and 1.8 hours in other settings (Table 1). Low socio-economic deprivation children captured more images inside their homes compared to high socio-economic deprivation children (Table 1). However, this difference was driven by several outliers in the low socio-economic deprivation group who captured substantial data in these settings. There was no difference in median photos captured between deprivation groups (nonparametric equality of medians test: p=0.230).

Household resources and harms

Children living in conditions of high socio-economic deprivation captured fewer types of fruit (RR=0.46, 95%CI 0.25–0.85), vegetables (RR=0.25, 95%CI 0.14–0.58), educational materials (RR=0.49, 95%CI 0.37–0.65), physical activity equipment (RR=0.66, 95%CI 0.45–0.96) and ‘personal interest’ items (RR=0.66, 95%CI 0.48, 0.90) on camera than children living in conditions of low socio-economic deprivation (Table 2). Fruit and vegetables were observed stored in fridges, pantries, freezers and counter tops, and during food consumption. In low socio-economic deprivation households, fruit was more often positioned in locations visible to children, e.g., in “fruit bowls” on spacious countertops (Figure 1). More structural deficiencies (RR=4.50, 95%CI 2.48–8.15) and mould were observed in high socio-economic deprivation households compared to low socio-economic deprivation households (no mould was observed in low socio-economic deprivation households) (Figure 1). Children living in conditions of high socio-economic deprivation were less likely to have computer access (RR=0.45, 95%CI 0.25–0.80), less likely to sleep in their own room (RR=0.50, 95%CI 0.28–0.91) and less likely to have a fixed source of heating (RR=0.27, 95%CI 0.10–0.71). Alcohol presence was not associated with socio-economic deprivation. Children living in conditions of low socio-economic deprivation usually had their own “personal space”, including their own bedroom, a desk for studying and collections of “personal items” such as books, posters and games (Figure 1). In contrast, children living in conditions of high deprivation children had less defined “personal spaces” and less material possessions.

View Tables 1 & 2.

Child behaviours

Socio-economic deprivation was not associated with the total amount of food items consumed. However, children living in conditions of high socio-economic deprivation consumed more non-core food items (RR=1.39, 95%CI 1.03–1.98), including nearly twice as many non-core food items outside home (RR=1.94, 95%CI 1.18–3.20). This was due to increased consumption of sweets, ice creams and snack foods purchased from convenience stores and fast food outlets. They also consumed fewer core items outside home than children living in conditions of low socio-economic deprivation (RR=0.13, 95%CI 0.03–0.58).

Children living in conditions of high socio-economic deprivation appeared to participate in fewer educational activities and structured physical activities than children living in conditions of low socio-economic deprivation, but the results were not significant at the 95% confidence level. Socio-economic deprivation had little influence on children’s participation rate in unstructured physical activity. Backyards were popular spaces for unstructured physical activities, for both children living in high and low socio-economic deprivation (Figure 1).

Built environment characteristics

For each hour children spent outdoors in their neighbourhood, they were exposed to a mean of 0.9 fixed physical disorder items (mostly private dilapidated properties), 12.1 semi-fixed physical disorder items (mostly graffiti) and 6.1 moveable physical disorder items (mostly waste). Household socio-economic deprivation was not associated with children’s exposure to fixed, semi-fixed and moveable physical disorder. Children from high deprivation neighbourhoods were exposed to more disorder (4.5 times more fixed items, 1.5 times more semi-fixed items and 3.3 times more moveable items than children from low deprivation neighbourhoods), but these results were insignificant at the 95% confidence level.

Differences by ethnicity

Māori children captured fewer types of fruit (RR=0.59, 95%CI 0.37, 0.95), educational materials (RR=0.56, 95%CI 0.46, 0.68), cognitive stimulation materials (RR=0.41, 95%CI 0.24, 0.69), “personal items” (RR=0.76, 95%CI 0.24, 0.69) and physical activity equipment (RR=0.49, 95%CI 0.29, 0.84) on camera than NZ European children. Likewise, Pacific children captured fewer types of fruit (RR=0.44, 95%CI 0.24, 0.80), educational materials (RR=0.55, 95%CI 0.39, 0.78), cognitive stimulation materials (RR=0.27, 95% CI 0.16, 0.44), “personal interest” items (RR=0.64, 95% CI 0.48, 0.86) and physical activity equipment (RR=0.52, 95%CI 0.32, 0.84) than NZ European children, and were less likely than NZ European children to have a fixed source of heating (RR=0.67, 95%CI 0.49, 0.91).

View Figure 1.

Discussion

The images in this study illustrate that children living in poverty face disadvantages across many aspects of their lives. Children living in conditions of high socio-economic deprivation lived in households with fewer types of fruit, vegetables, educational materials, physical activity equipment and “personal items”, and more structural deficiencies and mould. They were also less likely to sleep in their own room and have access to a computer and fixed heating. These differences highlight a number of health concerns. Low availability of fruit and vegetables is associated with lower consumption of fruit and vegetables among children.[[14]] Fewer educational materials may present barriers for completing schoolwork and negatively affect school performance.[[15]] Moreover, poor housing conditions such as mould, structural deficiencies and lack of heating have several adverse implications for children’s health, including increased hospitalisation rates and stress.[[16,17]]

While most variables assessed have clear links to health, the health implications of having fewer “personal items” such as smartphones and toys can be both positive and negative. On one hand, a lack of material possessions can negatively affect children’s wellbeing[[18]] and reduce their capacity to pursue their own interests. On the other, excessive use of certain items, such as smartphones and videogames, can negatively affect the psychological health of some children.[[19,20]] Moreover, there is rising concern over increasing commercialisation of children’s environments,[[21,22]] which promotes materialism and may decrease life satisfaction among children.[[23]]

High deprivation children were more likely to consume non-core foods outside their home than low deprivation children. This could partly be explained by the obesogenic environment in which high deprivation children live.[[24]] Previous research shows that high socio-economic deprivation neighbourhoods in Aotearoa New Zealand have higher densities of unhealthy food outlets, such as convenience stores and fast food outlets.[[25]] Children in this study frequently purchased confectionery and sugary drinks from convenience stores, which may have particularly appealed to high deprivation children because of the typically low cost of these items.[[26]] Fast food may also appeal to low-income families because of its convenience and low cost compared with other eating out options.[[27]]

While insignificant at the 95% confidence level, it appeared that high socio-economic deprivation children were less likely to participate in structured physical activity. This is consistent with previous evidence[[28–30]] and is not surprising given the costs associated with participating in sport, e.g., training fees, and the purchasing of uniforms and equipment. In contrast, there was no association between socio-economic deprivation and unstructured physical activity. Backyards were popular spaces for unstructured physical activity among both socio-economic deprivation groups. This may reflect Aotearoa New Zealand’s traditional residential design, which appears to benefit children equally in terms of opportunities for outdoor physical activity and play.

Due to a history of colonisation, and institutional racism, child poverty is ethnically patterned in Aotearoa New Zealand, with Māori and Pacific children bearing a disproportionate burden.[[5]] While this study was underpowered to undertake detailed subgroup analyses, we found that Māori and Pacific captured fewer types of fruit (Māori only) educational materials, cognitive stimulation, “personal items”, physical activity equipment and fixed heating (Pacific only) in their homes than NZ European children.

This study has some limitations. First, we could not ascertain children’s perceptions of the variables under study. Although each variable was relevant for health (Appendix), their importance for children could depend on several factors, including external support (e.g., from schools) and cultural values (e.g., materialism and collectivism).[[31]] Second, wearable cameras were not suited for studying some important aspects of poverty, including clothing (because participant’s clothing was not usually visible) and overcrowding. Third, the method likely underestimated household resources that were stored out of sight, such as educational materials in drawers, food in cupboards and central heating. Finally, high deprivation children captured fewer images in home settings than low deprivation children, which may have resulted in an underestimation of their household resources. However, we do not believe this resulted in substantial bias, given that median recording time was similar between groups, children were highly mobile (thus enabling photo capture from different areas and vantages) and counts of household resources were not affected by the frequency with which they were captured (i.e., each resource type was only counted once).

Wearable cameras for poverty measurement: research implications

Wearable cameras have several advantages compared with population-based surveys of child poverty. The cameras enable poverty to be observed from the child’s perspective as they wear the cameras throughout their day. Underestimation is common in surveys, owing to report bias and the fact that people are often unaware of the factors around them. In contrast, this study demonstrates that wearable cameras can be used to capture a wide range of variables relevant to poverty. Moreover, the methodology allows the capture of contextual information such as sources of junk food consumed and the relative advantage of “personal space” for children to pursue their own interests.

Wearable cameras offer advantages and disadvantages compared to alternative visual methodologies such as Photovoice. Photovoice provides participants with cameras to help identify issues of concern and holds discussions with them to reflect upon these issues.[[32]] Compared to Photovoice, wearable cameras offer a more comprehensive and “objective” approach in that a wide range of aspects of participant’s lives can be studied. In this regard, it is analogous to passive momentary exposure assessment, such as wearable air pollution monitors used in environmental epidemiology.[[33]] In contrast, data from Photovoice is filtered through participants’ choices. While less detailed, it enables participants to identify the most salient features of their lived experience. Wearable camera research could be strengthened by use of qualitative interviews or group discussions, like Photovoice. Researchers should consider the use of wearable cameras to study child poverty from children’s perspectives.

Conclusions

This study illustrates that children in poverty face disadvantages across many aspects of their lives in breach of their rights under international law.[[1]] This “accumulation” of disadvantage can adversely affect their health and development and interfere with their right to an adequate standard of living. In the Aotearoa New Zealand context, the disproportionate burden experienced by Māori children breaches Te Tiriti o Waitangi and the United Nations Declaration on the Rights of Indigenous Peoples.[[7]] Pacific children also carry a heavier burden, as is true internationally of Indigenous and minority populations. From a policy perspective, the children in this study highlight the urgent need for comprehensive policies to improve outcomes for children in poverty. Although targeted policies (e.g., school food programmes)[[34]] can improve some consequences of child poverty, multi-pronged approaches are needed to help ensure equal living standards and opportunities for all children.[[2]] This should include community development initiatives that empower communities to address the “wicked” problem of child poverty.[[35]] While based in Aotearoa New Zealand, this study has relevance for child poverty in similar jurisdictions, particularly those with Indigenous and minority communities.

View Appendices.