critical appraisal

Sociology 2016, Vol. 50(4) 673 –694

© The Author(s) 2015 Reprints and permissions:

sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0038038515570144

soc.sagepub.com

The Effects of Financial Vulnerability and Mothers’ Emotional Distress on Child Social, Emotional and Behavioural Well-Being: A Structural Equation Model

Morag Treanor University of Edinburgh, UK

Abstract This article aims to understand the pathways through which financial vulnerability affects children’s social, emotional and behavioural (SEB) well-being and whether that impact is directly experienced or, as hypothesised, indirectly through their mothers’ emotional well-being. It uses data from Growing Up in Scotland – a longitudinal birth cohort study of 5217 children born in 2004–2005. The results show that maternal emotional distress is strongly associated with financial vulnerability, more so than with income, and that child SEB well-being is negatively associated with financial vulnerability and maternal emotional distress, with two-thirds of the effect of financial vulnerability being experienced indirectly through maternal emotional distress. While the qualitative evidence shows that financial vulnerability adversely affects older children directly, through the comparisons they make to their reference group, the quantitative finding is that young children are also negatively affected but predominantly via the effect of financial vulnerability on their mothers’ emotional distress.

Keywords children, financial vulnerability, income, maternal emotional distress, poverty, quantitative methods, Scotland, social, emotional and behavioural (SEB) well-being, structural equation modelling (SEM)

Background

Runciman (1966: 9), in the theory of relative deprivation, posited that ‘people’s attitudes, aspirations and grievances largely depend on the frame of reference within which they

Corresponding author: Morag Treanor, Social Policy/Q-Step, School of Social and Political Science, University of Edinburgh, 2.25 Chrystal Macmillan Building, 15a George Square, Edinburgh EH8 9LD, UK. Email: [email protected]

570144 SOC0010.1177/0038038515570144SociologyTreanor research-article2015

Article

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

674 Sociology 50(4)

are conceived’. The frame of reference, or reference group, to which an individual compares themselves ‘can be a class, a country, a single person or even an abstract idea’ (Runciman, 1966: 11). The theory of relative deprivation suggests that when an individual makes negative comparisons between themselves and their reference group it can lead to ‘subjective deprivation’ experienced as ‘subjective injustice’ and ‘emotional distress’ (Ragnarsdóttir et al., 2013: 756). Who an individual chooses as their reference group will invariably affect how the individual feels about their situation, position and status. Merton (1968: 14) states that ‘some similarity in status attributes between the individual and the reference group must be perceived or imagined for the comparison to occur at all’. For those living in poverty and financial vulnerability, the reference group can be those in the same socioeconomic position or others who are similar in other characteristics but who have a different socioeconomic position. For mothers living in poverty and financial vulnerability the reference group may be other mothers in their community with whom they perceive an attributional similarity, for example, the status of motherhood, but who may have dissimilar levels of financial vulnerability. Comparing themselves to mothers with different socioeconomic realities may result in feelings of subjective disadvantage, which ‘can appear as emotional distress manifested through anger and depression’ (Ragnarsdóttir et al., 2013: 758; Smith et al., 2012).

It is widely theorised that economic crises and financial vulnerability lead to distress directly, through the economic problems they bring (Lister, 2004), and also indirectly through the subjective injustice and emotional distress they trigger in the individual through comparisons of self to the reference group (Ragnarsdóttir et al., 2013). While the theory of relative deprivation has commonly been applied to sudden economic crises, such as recessions (Ragnarsdóttir et al., 2013), it has an apposite application to the study of chronic economic conditions such as poverty and financial vulnerability. Merton (1968: 201) noted that ‘“poverty” is not an isolated variable which operates in precisely the same fashion wherever found; it is only one in a complex of identifiably interdependent social and cultural variables’. Thus, poverty per se, and not just the experience of poverty, is relational to its social context (Townsend, 1979: 132).

Experiences of poverty can be transient, that is, mild and alleviable by existing or acquired resources; or acute, severe, chronic and persistent. This dynamic aspect of poverty is an increasing focus in the study of poverty with the recognition that poverty, particularly transient poverty, is more common than cross-sectional studies would suggest, indicating greater financial vulnerability than previously realised (Berthoud and Bryan, 2011; Fouarge and Layte, 2005; Jenkins et al., 2001). Yet, financial vulner- ability is a term that is often used erroneously and synonymously with poverty. Chambers (1989: 1) emphasises that vulnerability is not the same as poverty. He explains that where poverty indicates lack or want, vulnerability is defined by ‘insecurity, and expo- sure to risk, shocks and stress’ (1989: 6). That the financially vulnerable increasingly includes those who are in-work with insecure livelihoods implies consequences that are presently under-recognised (Shildrick et al., 2013).

Within a family milieu, how financial vulnerability directly, and its lived experience indirectly, creates subjective distress is not widely studied. How mothers, who have various frames of reference, and children, who have their peers as their reference group, are affected has not been widely studied in relation to the mothers and children together. The

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 675

theory of relative deprivation suggests, and the empirical evidence supports the theory, that social comparisons reduce emotional well-being for those with financial vulnerability. What is less understood are the pathways through which poverty and financial vulnera- bility affect the subjective distress of children and mothers. Older children show that they experience the distressing effects of poverty and financial vulnerability directly, through their own social comparisons; however, there is no research that studies the impact of their mothers’ distress on the young people’s experience. Moreover, studies on the pathways through which financial vulnerability has an impact on young children, who would not be expected to have either a reference group or the ability to make comparisons, are few and are not situated within a sociological framework. The gap that this research aims to fill is to understand these pathways through which financial vulner- ability has an impact on young children and whether that impact is directly experienced or, as hypothesised, indirectly through their mothers’ emotional well-being.

Reference Groups and Relative Deprivation

Relative deprivation theory states that subjective comparisons are made between the individual and their reference group. Such comparisons influence how people feel about their circumstances: negative social comparisons can lead to ‘invidious self-depreciation’ and ‘personal inadequacy’ (Merton, 1968: 294). Furthermore, the salience of the subjective comparisons can be greater than the objective reality of a given situation (Ragnarsdóttir et al., 2013). Within the study of poverty, this resonates in the relatively low level of overlap between subjective feelings of poverty and objective measures of poverty (Bradshaw and Finch, 2003). There are several reasons why the relationship between subjective and objective measures of poverty is imperfect: subjective deprivation; false consciousness; intra-familial transfer of resources; low aspirations or expectations; measurement error; and the lagged effect of income poverty on living standards (Bradshaw and Finch, 2003). Financial vulnerability, in contrast to measures of objective poverty, has an inherently subjective element.

Financial Vulnerability

Chambers (1989) argues that definitions of poverty conceived by professionals overlook vulnerability despite it being a primary concern to poor people themselves. He asserts that poverty, as measured by low income, can be reduced by borrowing, but that the resulting debt makes households more vulnerable (Chambers, 1989). People living in poverty have a fear of debt and are more aware than poverty professionals of the trade-offs between poverty and vulnerability. Chambers (1989: 6) posits that ‘poor people all over the world are reluctant to take debts which increase their vulnerability’.

Whelan and Maitre (2005, 2008) used the European Community Household Panel (ECHP) data to create a concept translated directly from Chambers’ work that they call ‘economic vulnerability’. They conceptualise vulnerability as insecurity and exposure to risk and shock rather than directly measured economic deprivation. Their measure of economic vulnerability includes objective risk of deprivation and subjective sense of insecurity (Whelan and Maitre, 2008: 640). They compared the groups identified as being

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

676 Sociology 50(4)

economically vulnerable with social class and found that financial vulnerabilities operate along traditional social class lines; those from a higher social class ‘had very high levels of protection from economic vulnerability’ whereas those from the traditionally lower social classes experienced persistent economic vulnerability (Whelan and Maitre, 2008: 655).

Financial Vulnerability and Subjective Distress

Financial vulnerability, as measured by debt, money worries and managing on current income is qualitatively shown to have a negative impact on emotional well-being, stress, anxiety and depression (Green, 2007; Magadi and Middleton, 2007). One study conducted longitudinal, qualitative research on families living in poverty over the course of a year in 2008. Its overarching finding does not relate to poverty but to the depth and extent of financial vulnerability: in particular of debt and the inability to cope with unexpected bills and expenses (Harris et al., 2009; Stewart, 2009). It high- lights the negative impact of financial vulnerability on adult emotional health with families reporting this as their main cause of anxiety, depression and relationship conflict. A further study emphasises that debt ‘compounds vulnerability and negatively affects emotional wellbeing’ (Whitham, 2012: 5). The quantitative evidence is less abundant. One study explores the financial vulnerability on emotional distress, using data from c.6000 adults aged 16 to 64 years in Sweden. It shows that women are twice as likely, and men three times as likely, to experience anxiety, depression and reduced psychological well-being if they are experiencing financial vulnerability (Starrin et al., 2009).

As regards children and young people, qualitative studies clearly show that older children (usually aged eight and above) feel ashamed, excluded and stigmatised by their family’s economic disadvantage (Holscher, 2008). This subjective distress children and young people experience is said to occur because they are unable to participate in the social, leisure and celebratory activities of their peer group, which can adversely affect their friendships and self-esteem (Ridge, 2002, 2009). Additionally, children and young people are reported as being aware of, and worried about, the financial pressure their family is under, which has further detrimental effects on their subjective distress (Whitham, 2012). This suggests that, at least in part, poverty and financial vulnerability have direct negative impacts on children and young people’s emotional well-being due to negative social comparisons.

There are a variety of causal pathways proposed for the impacts on financial vulner- ability on child social, emotional and behavioural (SEB) well-being. The stress induced by financial vulnerability, be it the result of directly experienced resource deprivation, or the subjective deprivation induced by subjective comparisons to the reference group, is postulated to have adverse impacts on mothers’ emotional well-being which in turn have adverse effects on child well-being, creating an indirect path from financial vulnerability to child well-being via maternal well-being (Conger et al., 2010; Yeung et al., 2002). This article uses structural equation modelling to decompose the hypothesised relationship between financial vulnerability, child SEB well-being and maternal emotional distress.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 677

Data and Methods

The analysis is conducted using data from the Growing Up in Scotland (GUS) study, a birth and child cohort study, with an annual frequency of data collection, initiated to record the characteristics, circumstances, health and behaviours of children in the early years in Scotland (Anderson et al., 2007). A stratified, clustered sampling strategy was used: a named sample was selected on the basis of the children’s dates of birth using UK Child Benefit records, chosen because 97 per cent of all eligible fami- lies were registered for this, then universal, benefit (2005). The response rate across the birth and child cohorts was 80 per cent of all in-scope children producing an achieved sample for the birth cohort of 5217 babies at sweep 1 (Anderson et al., 2007: 196) reducing to 3833 children at sweep 5; see Table 1 for more details on survey response rates. The main carer, usually the mother (approximately 98 per cent), is the respondent in GUS. The analysis is undertaken using Stata 13 and the ‘surveyset’ procedure is tested to take account of the complex sampling, that is, strata, clustering, sample selection weights and longitudinal attrition weights. For more information see the annual user guides on GUS (Bradshaw et al., 2009, 2010; Corbett et al., 2005, 2006, 2007).

The Dependent Variable – Child SEB Well-Being

The child SEB outcome is measured by the age-appropriate Strengths and Difficulties Questionnaire (SDQ) available for children aged four to 17 years old. For the children in this study (aged four/five) the questionnaire was completed by mothers. There are five dimensions to the SDQ: conduct problems; emotional symptoms; hyperactivity; peer relationships; and pro-social behaviour (Goodman, 1997: 581). The first four of these are summed to provide a total score. The fifth dimension, pro-social behaviour, cannot be incorporated into the total score ‘since the absence of pro-social behaviours is conceptually different from the presence of psychological difficulties’ (Goodman, 1997: 582). The resulting score from the summed first four dimensions is then reversed and standardised, giving a mean of zero and a standard deviation of one. Positive scores denote higher than average SEB well-being and negative scores denote lower SEB well-being.

Table 1. Sweep information for the birth cohort.

Sweep Year Achieved sample Response rate (all eligible cases) (%)

Response rate (as % of sweep 1)

1 2005–2006 5217 80 100 2 2006–2007 4512 88 87 3 2007–2008 4193 91 80 4 2008–2009 3994 91 77 5 2009–2010 3833 92 73

Source: GUS sweeps 1–5.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

678 Sociology 50(4)

Financial Vulnerability

GUS collects data in sweep 5 on financial vulnerability. The exact wording of the questions and their possible responses are set out below (Bradshaw et al., 2010):

•• Money worries: how often would you say you have been worried about money during the last few weeks?

○ 1 almost all the time, ○ 2 quite often, ○ 3 only sometimes, ○ 4 never

•• Household debt: thinking back over the past 12 months, how often would you say you have had trouble with debts that you found hard to repay?

○ 1 almost all the time, ○ 2 quite often, ○ 3 only sometimes, ○ 4 never

•• Manage financially: taking everything together, which of the phrases on this card best describes how you and your family are managing financially these days?

○ 1 Manage very well ○ 2 Manage quite well ○ 3 Get by all right ○ 4 Don’t manage very well ○ 5 Have some financial difficulties ○ 6 Are in deep financial trouble

These three variables form the latent construct ‘financial vulnerability’ estimated in the first measurement part of the structural equation model discussed in the methods section.

Maternal Emotional Distress (SF-12)

In sweep 5 data are collected using the SF-12 health survey form, which is the instrument of choice for large scale and longitudinal survey studies (Jenkinson et al., 1997; Ware et al., 1996). For this paper seven questions from the SF-12 pertaining to maternal emotional distress are used (EMD1–EMD7). These are:

•• EMD1 – In general, would you say your health is excellent, very good, good, fair or poor?

○ 1 Excellent ○ 2 Very good ○ 3 Good ○ 4 Fair ○ 5 Poor ○ 6 Can’t say

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 679

•• EMD2 – During the past four weeks, have you accomplished less than you would like as a result of any emotional problems, such as feeling depressed or anxious?

○ 1 Yes ○ 2 No

•• EMD3 – During the past four weeks, did you not do work or other regular activi- ties as carefully as usual as a result of any emotional problems, such as feeling depressed or anxious?

○ 1 Yes ○ 2 No

•• EMD4 – How much time during the past four weeks have you felt calm and peaceful?

○ 1 All of the time ○ 2 Most of the time ○ 3 A good bit of the time ○ 4 Some of the time ○ 5 A little of the time ○ 6 None of the time

•• EMD5 – How much of the time during the past four weeks did you have a lot of energy?

○ 1 All of the time ○ 2 Most of the time ○ 3 A good bit of the time ○ 4 Some of the time ○ 5 A little of the time ○ 6 None of the time

•• EMD6 – How much of the time during the past four weeks have you felt down?

○ 1 All of the time ○ 2 Most of the time ○ 3 A good bit of the time ○ 4 Some of the time ○ 5 A little of the time ○ 6 None of the time

•• EMD7 – During the past four weeks, how much of the time has your physical health or emotional problems interfered with your social activities like visiting with friends, relatives, etc?

○ 1 All of the time ○ 2 Most of the time ○ 3 A good bit of the time ○ 4 Some of the time ○ 5 A little of the time ○ 6 None of the time

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

680 Sociology 50(4)

These seven variables form the latent construct ‘maternal emotional distress’ estimated in the second measurement part of the structural equation model discussed in the methods section.

Equivalised Income

Economies of scale, where people pool their resources, that is, share their wealth or their poverty with other family or household members (Alcock, 2006), are taken account of in the measurement of income through equivalence scales, which assign a ‘weight’ to each member. The equivalence scale used here is the modified Organisation for Economic Co-operation and Development (OECD) equivalence scale, which gives a weight of 1.0 for the first adult in a household, 0.5 for an additional person aged 15 years or over and 0.3 for any children aged 0–14 years (Chanfreau and Burchardt, 2008), and is a continuous measure taken at sweep 5.

Control Variables

The existing research identifies factors that should be controlled for in analysis that includes socioeconomic disadvantage and child SEB well-being (Kiernan and Huerta, 2008; Kiernan and Mensah, 2009; Schoon et al., 2010, 2012). These factors are: child’s gender; family composition; maternal education; maternal employment; birth order of the child; and the age of the mother at the birth of her first child.

Child’s Gender

The gender of a child is found to be associated with his/her SEB well-being: being a boy is associated with lower scores on this developmental outcome (Blair et al., 2004). Research shows that boys mature more slowly than girls (Cohn, 1991), that girls are more content than boys to sit still and listen in school and that boys are more physical and active (King and Gurian, 2006); all of which may affect perceptions of boys’ SEB well-being. The gender variable is a straightforward binary girl/boy variable.

Family Composition

The existing evidence on the impact of family composition on child well-being is often contradictory. Furthermore, family composition often focuses on the differences between married and unmarried parents and not on family transitions, that is, moving from a couple to a lone parent family or vice versa. As the data here are longitudinal, using all five sweeps, the family composition variable can focus on transitions. The derived variable is categorical and the categories are:

•• ‘stable couple family’, where a couple has been together since the start of the study (reference category);

•• ‘stable lone parent family’, where the respondent is the sole adult in the household in each of the five years of the study;

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 681

•• ‘lone parents who have re-partnered’ – there is no distinction in the measure on the point at which the respondent re-partners;

•• ‘couple families who have separated’ – the same caveat applies as before; and •• ‘separation(s) and re-partnering(s)’ – this category does not differentiate between

those who may be separating and re-partnering with the same or with different partners.

Maternal Education

Maternal education is highly significantly associated with both the socioeconomic posi- tion of a family and with the developmental outcomes of a child (Hansen et al., 2010; Melhuish et al., 2008). Maternal education is relatively stable in GUS and so the measure taken at sweep 5 is used in the analysis. The categories are:

•• Degree or equivalent (reference category) •• Vocational qualification below degree •• Higher grade/A-level •• Standard grade/GCSE •• No qualifications

Maternal Employment

Maternal employment is significantly associated with the socioeconomic position of a family, with maternal mental well-being and with child outcomes (Thomas et al., 2005; Zick et al., 2001). The measure for maternal employment taken at sweep 5 is used in the analysis. The categories are:

•• Full-time employed (reference category) •• Part-time employed •• Not in paid work

Study Child’s Birth Order

Birth order is posited to be associated with higher levels of child development (Bradshaw, 2011). However, it is not clear what affect birth order has on SEB well-being as it is not well documented. The birth order variable is a simple binary first born/not first born measure.

Age of Mother at First Child’s Birth

Having a younger mother is associated in the literature with lower SEB well-being (Bradshaw and Tipping, 2010; Bromley, 2009). Younger mothers are also at increased risk of living in poverty and for living in poverty for longer periods of time (Barnes et al., 2010). The age of the mother at the birth of her first child is a continuous variable in the model. The summary statistics for all variables can be found in Table 2.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

682 Sociology 50(4)

Methods

Structural equation modelling (SEM) is used to understand the pathways through which financial vulnerability has an impact on child SEB well-being, and to ascertain whether these impacts are experienced directly or indirectly via maternal emotional distress, while controlling for confounding variables. SEM comprises two components in one model: the measurement part and the structural part (Acock, 2013). The measurement

Table 2. Summary of all variables.

Variables Count Mean SD Min Max

Child SEB 3537 0.00 1.00 −4.93 1.65 Financial vulnerability 3537 0.00 1.00 −1.02 2.38 Financial vulnerability variables: How often money worries (FV1) 3537 1.81 0.79 1.00 3.00 How manage financially (FV2) 3537 1.57 0.65 1.00 3.00 How often debts hard to pay (FV3) 3537 1.40 0.66 1.00 3.00 Maternal emotional distress variables: General health (EMD1) 3537 2.32 0.99 1.00 6.00 Time felt calm (EMD2) 3537 2.99 1.10 1.00 6.00 Time felt energetic (EMD3) 3537 3.03 1.14 1.00 6.00 Time felt down (EMD4) 3537 4.85 1.06 1.00 6.00 Time health interfered socially (EMD5) 3537 5.40 1.12 1.00 6.00 Limited accomplishments (EMD6) 3537 1.85 0.36 1.00 2.00 Limited work/activities (EMD7) 3537 1.88 0.33 1.00 2.00 Income (equivalised) 3537 24676.96 12442.59 1286.77 68965.52 Birth order (ref: first born child) 3537 0.51 0.50 0.00 1.00 Male 3537 0.51 0.50 0.00 1.00 Family transition dummy variables: Couple 3537 0.78 0.41 0.00 1.00 Lone parent 3537 0.08 0.27 0.00 1.00 Re-partnered lone parent 3537 0.05 0.21 0.00 1.00 Separated couple 3537 0.06 0.24 0.00 1.00 Separations/Re-partnerings 3537 0.03 0.18 0.00 1.00 Age at birth of first child 3537 27.02 5.92 11.00 46.00 Maternal employment dummy variables: Full-time 3537 0.77 0.42 0.00 1.00 Part-time 3537 0.12 0.33 0.00 1.00 Not in paid work 3537 0.10 0.31 0.00 1.00 Maternal education dummy variables: No qualifications 3537 0.06 0.24 0.00 1.00 Standard grade (GCSE) 3537 0.14 0.34 0.00 1.00 Higher grade (A level) 3537 0.07 0.26 0.00 1.00 Vocational 3537 0.40 0.49 0.00 1.00 Degree 3537 0.33 0.47 0.00 1.00

Source: GUS sweeps 1–5.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 683

part comprises confirmatory factor analysis which allows a latent variable to be esti- mated from its manifest indicator variables while accurately isolating any measurement error (Acock, 2013). The predictive power of the model is stronger when measurement error is removed as it is assumed to be a random error with no explanatory power (Acock, 2013). For this article the latent construct financial vulnerability is estimated using three ordinal variables previously described: ‘money worries’, ‘household debt’ and ‘manage financially’; and maternal emotional distress is estimated using a measurement model and the seven variables, EMD1–EMD7, described previously.

The structural part of the model is a path analysis that can be decomposed into direct and indirect effects pathways. The structural part of the model can show theoretically causal linkages which can provide evidence to support a causal theory (Acock, 2013). Figure 1 presents the full structural equation model which shows the measurement parts of the model with direct pathways leading from financial vulnerability to maternal emotional distress and child SEB well-being; from maternal emotional distress to child SEB well-being; and an indirect path from financial vulnerability to SEB well-being via maternal emotional distress. The control variables described earlier have pathways to maternal emotional distress, to financial vulnerability and to child SEB well-being.

The use of structural equation modelling in this study brings manifold advantages over the traditional regression modelling techniques. The first is that the latent constructs financial vulnerability and maternal emotional distress can be estimated as integral parts of the model allowing the measurement error to be estimated. The second is that SEM

.73*** .75*** .73***

.43***

-.08***

-.32***

.58***

.52***

.55***

-.78***

-.74***

-.71***

-.68***

Financial vulnerability

1

Household debt

Money worries

Manage financially

ε 2 ε 3 ε 4

Child well-being ε 5

Control variables (see table for

details)

RSMEA= 0.020 CFI = 0.970 CD = 0.351

ε 1

Maternal emotional distress

1

ε 6

EmD2 ε 8

EmD3

EmD5

EmD6

EmD4

EmD7

EmD1 ε 7

ε 10

ε 9

ε 11

ε 12

ε 13

Figure 1. SEM model.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

684 Sociology 50(4)

allows for direct and indirect effects to see what effects, if any, are direct to the child and which operate through the pathway of maternal emotional distress. The third is that the model allows pathways from all of the control variables to the two measurement models (financial vulnerability and maternal emotional distress) and the outcome variable (child SEB) to be estimated separately, which decomposes the influence of these confounding factors, while providing additional evidence to further the substantive knowledge in this area. SEM relies on correlations between the variables in the model for its estimation procedures. A correlation of all the single variables (including dummy variables) used in this analysis can be found in Table 3.

The model was estimated using Stata 13. The svyset weighting facility was tested in the modelling framework but was not used in the final model of this article as it made little difference to the coefficients or standard errors but did reduce the goodness of fit statistics available post-estimation.

Results

The three financial vulnerability indicator variables load highly onto the latent construct ‘financial vulnerability’ with respective loadings of ‘money worries’ 0.75, ‘household debt’ 0.73 and ‘managing financially’ 0.73, all of which are significant at the 0.1 per cent level. The seven SF-12 indicator variables load highly and moderately highly onto the latent construct ‘maternal emotional distress’ with respective loadings of ‘EMD1’ 0.52, ‘EMD2’ 0.58, ‘EMD3’ 0.55, ‘EMD4’ –0.78, ‘EMD5’ –0.74, ‘EMD6’ –0.71, ‘EMD7’ –0.68, all of which are significant at the 0.1 per cent level.

The goodness-of-fit measures, shown in Table 4, are as follows: the root mean squared error of approximation (RMSEA) is 0.020, much lower than the recommended 0.5 and closer to the ideal of zero (Acock, 2013; Hu and Bentler, 1999; Statacorp, 2013); com- parative fit index (CIF) is 0.970, higher than the recommended 0.95; the Tucker-Lewis index (TLI) at 0.981 is higher than the recommended 0.95; and the coefficient of deter- mination (r-squared, ranging between 0 and 1) at 0.351 is moderate (Statacorp, 2013).

Table 5 shows the direct and indirect effects in the model. The first section of the model shows that maternal emotional distress has the largest negative effect on child SEB well-being, accounting for almost a third of a standard deviation decrease (β = −0.32). Financial vulnerability, as hypothesised, has a direct (β = −0.08) and an indirect (β = −0.14) negative effect on child SEB well-being through the pathway of maternal emotional distress (total effect size = −0.22). Other variables that have an impact on child SEB well-being are: (1) income, direct β = 0.05 and indirect β = 0.06: higher income is associated with more positive levels of child SEB well-being; (2) not being the first born, that is, having siblings, is associated with higher SEB well-being, (direct β = 0.11); (3) boys, as expected, have lower SEB well-being (direct β = −0.13); (4) family transitions – stable lone parent (indirect β = −0.02), re-partnered lone parent (indirect β = −0.02), separated couple (indirect β = −0.03), through their effect on maternal emotional distress, and separations and re-partnerings (direct β = −0.07), all compared to stable couple parents. This means that having a mother with repeated separations and re-partnerings is the only family formation that has a direct, negative association with SEB well-being; (5) age of the mother at first birth (direct β = 0.06); (6) not being in paid work, direct β = −0.05 and indirect β = −0.07; and (7) all education levels

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

T ab

le 3

. C

or re

la tio

n m

at ri

x of

a ll

va ri

ab le

s (in

cl ud

in g

du m

m ie

s) .

SE B

FV 1

FV 2

FV 3

M M

H 1

M M

H 2

M M

H 3

M M

H 4

M M

H 5

M M

H 6

M M

H 7

In co

m e

Fi rs

t bo

rn M

al e

C ou

pl e

Lo ne

pa

re nt

Re -p

ar t

ne r

Se p

co up

le Se

p+

Re -p

ar t

A ge

Fu ll-

tim e

Pa rt

- tim

e U

ne m

- pl

oy ed

N o

qu al

s St

an d

H ig

he r

Vo ca

- tio

na l

D eg

re e

SE B

1

FV 1

−0 .2

26 1

FV

2 −0

.2 09

0. 56

6 1

FV

3 −0

.2 43

0. 53

8 0.

52 3

1

M M

H 1

−0 .2

69 0.

22 7

0. 22

9 0.

24 1

1

M M

H 2

−0 .2

99 0.

23 7

0. 20

8 0.

17 9

0. 34

7 1

M

M H

3 −0

.2 62

0. 20

7 0.

21 3

0. 15

6 0.

46 0

0. 59

0 1

M

M H

4 0.

32 1

−0 .2

92 −0

.2 52

−0 .2

74 −0

.3 59

−0 .4

92 −0

.4 30

1

M M

H 5

0. 28

4 −0

.2 41

−0 .2

04 −0

.2 56

−0 .4

03 −0

.3 65

−0 .3

86 0.

58 9

1

M M

H 6

0. 25

1 −0

.2 65

−0 .2

16 −0

.2 62

−0 .3

59 −0

.4 16

−0 .3

90 0.

55 0

0. 55

3 1

M

M H

7 0.

22 6

−0 .2

39 −0

.2 03

−0 .2

56 −0

.3 29

−0 .3

79 −0

.3 53

0. 51

4 0.

54 8

0. 75

4 1

In

co m

e 0.

24 1

−0 .3

31 −0

.3 81

−0 .3

77 −0

.2 15

−0 .0

73 −0

.0 70

0. 18

1 0.

21 4

0. 17

2 0.

16 9

1

Fi rs

t bo

rn 0.

06 0

0. 03

3 0.

04 0

0. 04

4 0.

01 0

0. 05

2 0.

01 6

−0 .0

32 −0

.0 49

−0 .0

44 −0

.0 45

−0 .1

43 1

M

al e

−0 .1

45 0.

01 3

0. 02

4 −0

.0 07

0. 01

1 0.

03 3

0. 01

3 −0

.0 26

−0 .0

29 −0

.0 24

−0 .0

26 −0

.0 10

0. 01

4 1

C

ou pl

e 0.

19 7

−0 .2

39 −0

.1 87

−0 .3

06 −0

.1 64

−0 .0

43 −0

.0 48

0. 17

6 0.

18 6

0. 16

5 0.

15 6

0. 40

1 0.

05 2

−0 .0

25 1

Lo

ne p

ar en

t −0

.1 15

0. 12

9 0.

13 5

0. 19

5 0.

11 0

0. 00

6 0.

02 2

−0 .1

04 −0

.1 47

−0 .1

31 −0

.1 28

−0 .3

02 −0

.0 70

0. 03

2 −0

.5 50

1

R e-

pa rt

Lo ne

−0 .0

74 0.

08 9

0. 04

7 0.

12 7

0. 06

4 0.

01 8

0. 01

4 −0

.0 44

−0 .0

54 −0

.0 25

−0 .0

38 −0

.1 22

−0 .0

38 −0

.0 22

−0 .4

25 −0

.0 65

1

Se p

co up

le −0

.0 50

0. 13

9 0.

10 3

0. 13

0 0.

07 7

0. 04

4 0.

04 2

−0 .1

26 −0

.0 91

−0 .1

05 −0

.0 95

−0 .1

60 0.

05 4

0. 01

3 −0

.4 73

−0 .0

73 −0

.0 56

1

Se ps

/R e-

pa rt

s −0

.1 26

0. 06

8 0.

03 7

0. 09

2 0.

03 8

0. 01

2 0.

00 7

−0 .0

34 −0

.0 27

−0 .0

19 0.

00 2

−0 .1

17 −0

.0 41

0. 01

9 −0

.3 53

−0 .0

54 −0

.0 42

−0 .0

47 1

A

ge 0.

19 7

−0 .2

19 −0

.1 98

−0 .3

19 −0

.1 57

−0 .0

20 −0

.0 27

0. 13

6 0.

16 1

0. 11

5 0.

12 1

0. 51

0 −0

.2 17

−0 .0

27 0.

36 5

−0 .2

11 −0

.2 08

−0 .1

06 −0

.1 38

1

Fu ll-

tim e

0. 20

1 −0

.2 55

−0 .2

68 −0

.3 21

−0 .1

82 −0

.0 69

−0 .0

60 0.

19 5

0. 20

6 0.

19 0

0. 19

6 0.

52 3

−0 .0

12 −0

.0 15

0. 56

4 −0

.4 66

−0 .0

25 −0

.3 26

−0 .1

44 0.

31 3

1

Pa rt

-t im

e −0

.0 46

0. 12

6 0.

13 4

0. 13

6 0.

04 5

0. 01

7 0.

00 0

−0 .0

78 −0

.0 45

−0 .0

51 −0

.0 64

−0 .2

55 −0

.0 15

−0 .0

15 −0

.3 01

0. 20

0 0.

01 1

0. 24

4 0.

06 2

−0 .0

94 −0

.6 92

1

U ne

m pl

oy ed

−0 .2

27 0.

21 5

0. 22

4 0.

29 5

0. 20

2 0.

07 6

0. 08

2 −0

.1 83

−0 .2

35 −0

.2 07

−0 .2

01 −0

.4 43

0. 03

3 0.

03 7

−0 .4

49 0.

42 4

0. 02

2 0.

18 5

0. 13

1 −0

.3 29

−0 .6

27 −0

.1 29

1

N o

qu al

s −0

.1 47

0. 09

1 0.

09 8

0. 12

6 0.

10 6

0. 03

3 0.

03 8

−0 .1

13 −0

.1 21

−0 .0

94 −0

.0 91

−0 .2

47 0.

08 4

0. 00

1 −0

.1 65

0. 11

7 0.

10 4

0. 01

2 0.

06 5

−0 .2

00 −0

.2 14

0. 04

4 0.

24 6

1

St an

da rd

−0 .0

80 0.

05 2

0. 06

1 0.

10 5

0. 06

4 −0

.0 18

−0 .0

04 −0

.0 38

−0 .0

59 −0

.0 30

−0 .0

19 −0

.1 89

0. 04

0 0.

01 9

−0 .1

29 0.

11 2

0. 06

5 0.

02 0

0. 02

6 −0

.1 91

−0 .1

25 0.

02 9

0. 14

1 −0

.1 01

1

H ig

he r

0. 02

3 −0

.0 19

−0 .0

02 −0

.0 05

−0 .0

07 −0

.0 12

0. 01

2 0.

01 4

0. 00

8 0.

02 1

0. 01

7 −0

.0 34

0. 01

6 0.

00 5

0. 01

2 0.

00 9

−0 .0

01 −0

.0 09

−0 .0

28 −0

.0 38

0. 00

7 0.

01 0

−0 .0

21 −0

.0 71

−0 .1

11 1

V

oc at

io na

l −0

.0 20

0. 06

0 0.

07 4

0. 08

9 0.

03 6

−0 .0

33 −0

.0 24

0. 01

3 −0

.0 03

−0 .0

02 −0

.0 15

−0 .1

50 0.

00 2

−0 .0

30 −0

.0 57

0. 02

1 0.

03 2

0. 03

4 0.

01 7

−0 .1

22 −0

.0 15

0. 03

3 −0

.0 16

−0 .2

07 −0

.3 25

−0 .2

28 1

D

eg re

e 0.

14 2

−0 .1

36 −0

.1 70

−0 .2

31 −0

.1 34

0. 03

8 0.

00 2

0. 06

4 0.

10 2

0. 06

0 0.

06 6

0. 43

8 −0

.0 82

0. 01

5 0.

23 0

−0 .1

68 −0

.1 33

−0 .0

51 −0

.0 55

0. 38

9 0.

21 1

−0 .0

84 −0

.1 99

−0 .1

78 −0

.2 80

−0 .1

96 −0

.57 4

1

So ur

ce : G

U S

sw ee

ps 1

–5 .

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

686 Sociology 50(4)

(except Higher/A-level grade) compared to degree educated mothers with direct betas of −0.07 (no qualifications), –0.05 (standard grade/GCSE) and vocational level (–0.04). As the coefficients for this model are additive, this means that the combined direct and indirect coefficients for income, financial vulnerability and employment sum to 0.45 of a standard deviation decrease in child SEB well-being, which, when combined with the direct effect of maternal emotional distress (0.32) amounts to a −0.77 of a standard deviation decrease in SEB well-being.

In the second section, the model shows the variables that are directly associated with financial vulnerability as there are no indirect effects on financial vulnerability. The largest effect comes from income (β = −0.34) which accounts for 34 per cent of a standard deviation increase in financial vulnerability. Other variables statistically associated with increased levels of financial vulnerability are: (1) being a re-partnered lone parent (β = 0.06) or being a separated couple (β = 0.07), compared to being a stable couple; (2) the age of the mother at the birth of first child (β = 0.08) with financial vulnerability decreasing as maternal age increases; and (3) being in either part-time work (β = 0.08) or not in paid work (β = 0.14) compared to being employed full-time.

In the third section, the model shows that financial vulnerability has the largest effect, greater than income, on maternal emotional distress (β = 0.43), accounting for 43 per cent of a standard deviation increase. Income has an insignificant direct effect on maternal emotional distress; however, it has a strong indirect effect mediated through its association with financial vulnerability (β = −0.15). This suggests that financial vulnerability is a more important factor in increased emotional distress than income alone and that it is important irrespective of the level of income a family has. Other variables statistically significantly associated with increased maternal emotional distress are: having an employment status of ‘not in paid work’ (β = 0.18), and having been part of a couple that has since separated (β = 0.08): each of these has a direct pathway and an indirect pathway through its association with financial vulnerability. The remaining significant variables have an indirect pathway through financial vulner- ability: (1) being a re-partnered lone parent (β = 0.03); (2) having part-time compared to full-time employment (β = 0.03); and (3) the age of the mother at the birth of first child (β = −0.03).

Table 4. Goodness-of-fit statistics.

Goodness-of-fit statistics Value

Root mean squared error of approximation (RMSEA) 0.020 Comparative fit index (CIF) 0.970 Tucker-Lewis index (TLI) 0.981 Coefficient of determination (CD) 0.351 N 3583 d.f. 355

Source: GUS sweeps 1–5.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 687

T ab

le 5

. D

ir ec

t an

d in

di re

ct e

ffe ct

s.

O ut

co m

e D

ir ec

t ef

fe ct

s In

di re

ct e

ffe ct

s

C oe

f. St

d. E

rr .

P- va

lu e ⩽

St d.

C oe

f. C

oe f.

St d.

E rr

. P-

va lu

e ⩽

St d.

C oe

f.

C hi

ld S

EB ←

M at

er na

l e m

ot io

na l d

is tr

es s

− 0.

62 0.

04 2

0. 00

0 −

0. 32

– –

– –

Fi

na nc

ia l v

ul ne

ra bi

lit y

− 0.

13 0.

04 0

0. 00

1 −

0. 08

− 0.

24 0.

01 5

0. 00

0 −

0. 14

In

co m

e 0.

00 0.

00 0

0. 01

4 0.

05 0.

00 0.

00 0

0. 00

0 0.

06

Fi rs

t bo

rn 0.

22 0.

03 1

0. 00

0 0.

11 −

0. 02

0. 01

3 0.

21 4

− 0.

01

M al

e −

0. 25

0. 03

0 0.

00 0

− 0.

13 −

0. 02

0. 01

2 0.

20 1

− 0.

01

Fa m

ily t

ra ns

iti on

s (r

ef : s

ta bl

e co

up le

):

St ab

le lo

ne p

ar en

t 0.

09 0.

06 9

0. 19

3 0.

02 −

0. 06

0. 02

9 0.

05 1

− 0.

02

R

e- pa

rt ne

re d

lo ne

p ar

en t

− 0.

06 0.

07 2

0. 42

7 −

0. 01

− 0.

09 0.

03 1

0. 00

3 −

0. 02

Se pa

ra te

d co

up le

0. 09

0. 07

0 0.

19 2

0. 02

− 0.

13 0.

03 0

0. 00

0 −

0. 03

Se pa

ra tio

ns a

nd r

e- pa

rt ne

ri ng

s −

0. 39

0. 08

6 0.

00 0

− 0.

07 0.

00 0.

03 6

0. 97

1 0.

00

A ge

o f m

ot he

r at

fi rs

t bi

rt h

0. 01

0. 00

3 0.

00 1

0. 06

0. 00

0. 00

1 0.

09 7

0. 01

Em

pl oy

m en

t st

at us

( re

f: fu

ll- tim

e w

or k)

Pa

rt -t

im e

w or

k 0.

02 0.

05 4

0. 70

1 0.

01 −

0. 06

0. 02

3 0.

01 2

− 0.

02

N

ot in

p ai

d w

or k

− 0.

17 0.

06 7

0. 01

3 −

0. 05

− 0.

22 0.

03 0

0. 00

0 −

0. 07

Ed

uc at

io n

st at

us (

re f:

de gr

ee o

r ab

ov e)

N

o qu

al ifi

ca tio

ns −

0. 30

0. 07

3 0.

00 0

− 0.

07 −

0. 05

0. 03

0 0.

10 1

− 0.

01

St

an da

rd /G

C SE

− 0.

15 0.

05 3

0. 00

5 −

0. 05

0. 02

0. 02

2 0.

39 9

0. 01

H ig

he r/

A -le

ve l

− 0.

02 0.

06 3

0. 80

0 0.

00 0.

04 0.

02 6

0. 15

8 0.

01

V

oc at

io na

l b el

ow d

eg re

e −

0. 07

0. 03

8 0.

05 4

− 0.

04 0.

01 0.

01 6

0. 35

3 0.

01 Fi

na nc

ia l v

ul ne

ra bi

lit y ←

In co

m e

0. 00

0. 00

0 0.

00 0

− 0.

34 –

– –

–

Fi rs

t bo

rn −

0. 01

0. 02

1 0.

55 3

− 0.

01 –

– –

–

M al

e 0.

01 0.

02 0

0. 78

8 0.

00 –

– –

– (C on

tin ue

d)

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

688 Sociology 50(4)

O ut

co m

e D

ir ec

t ef

fe ct

s In

di re

ct e

ffe ct

s

C oe

f. St

d. E

rr .

P- va

lu e ⩽

St d.

C oe

f. C

oe f.

St d.

E rr

. P-

va lu

e ⩽

St d.

C oe

f.

Fa

m ily

t ra

ns iti

on s

(r ef

: s ta

bl e

co up

le ):

St

ab le

lo ne

p ar

en t

0. 05

0. 04

6 0.

31 1

0. 02

– –

– –

R e-

pa rt

ne re

d lo

ne p

ar en

t 0.

17 0.

04 8

0. 00

1 0.

06 –

– –

–

Se

pa ra

te d

co up

le 0.

17 0.

04 7

0. 00

0 0.

07 –

– –

–

Se

pa ra

tio ns

a nd

r e-

pa rt

ne ri

ng s

0. 08

0. 05

7 0.

14 8

0. 03

– –

– –

A

ge o

f m ot

he r

at fi

rs t

bi rt

h −

0. 01

0. 00

2 0.

00 0

− 0.

08 –

– –

–

Em pl

oy m

en t

st at

us (

re f:

fu ll-

tim e

w or

k)

Pa rt

-t im

e w

or k

0. 14

0. 03

5 0.

00 0

0. 08

– –

– –

N ot

in p

ai d

w or

k 0.

26 0.

04 4

0. 00

0 0.

14 –

– –

–

Ed uc

at io

n st

at us

( re

f: de

gr ee

o r

ab ov

e)

N o

qu al

ifi ca

tio ns

0. 00

0. 04

9 0.

96 0

0. 00

– –

– –

St an

da rd

/G C

SE −

0. 01

0. 03

5 0.

79 1

− 0.

01 –

– –

–

H

ig he

r/ A

-le ve

l −

0. 04

0. 04

2 0.

39 9

− 0.

02 –

– –

–

V

oc at

io na

l b el

ow d

eg re

e 0.

04 0.

02 5

0. 16

3 0.

03 –

– –

– M

at er

na l E

m ot

io na

l d is

tr es

s (E

M D

) ←

Fi na

nc ia

l v ul

ne ra

bi lit

y 0.

38 0.

02 5

0. 00

0 0.

43

In

co m

e 0.

00 0.

00 0

0. 14

1 0.

04 0.

00 0.

00 0

0. 00

0 −

0. 15

Fi

rs t

bo rn

0. 03

0. 01

8 0.

07 0

0. 03

0. 00

0. 00

8 0.

55 3

0. 00

M

al e

0. 02

0. 01

7 0.

20 3

0. 02

0. 00

0. 00

8 0.

78 8

0. 00

(C on

tin ue

d)

T ab

le 5

. (C

on tin

ue d)

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 689

O ut

co m

e D

ir ec

t ef

fe ct

s In

di re

ct e

ffe ct

s

C oe

f. St

d. E

rr .

P- va

lu e ⩽

St d.

C oe

f. C

oe f.

St d.

E rr

. P-

va lu

e ⩽

St d.

C oe

f.

Fa

m ily

t ra

ns iti

on s

(r ef

: s ta

bl e

co up

le ):

St

ab le

lo ne

p ar

en t

0. 06

0. 04

0 0.

12 4

0. 03

0. 02

0. 01

8 0.

31 2

0. 01

R e-

pa rt

ne re

d lo

ne p

ar en

t 0.

05 0.

04 3

0. 27

0 0.

02 0.

06 0.

01 9

0. 00

1 0.

03

Se

pa ra

te d

co up

le 0.

11 0.

04 1

0. 01

0 0.

05 0.

07 0.

01 8

0. 00

0 0.

03

Se

pa ra

tio ns

a nd

r e-

pa rt

ne ri

ng s

− 0.

05 0.

05 1

0. 31

2 −

0. 02

0. 03

0. 02

2 0.

15 0

0. 01

A

ge o

f m ot

he r

at fi

rs t

bi rt

h 0.

00 0.

00 2

0. 63

1 0.

01 0.

00 0.

00 1

0. 00

0 −

0. 03

Em

pl oy

m en

t st

at us

( re

f: fu

ll- tim

e w

or k)

Pa

rt -t

im e

w or

k 0.

01 0.

03 1

0. 80

1 0.

01 0.

05 0.

01 4

0. 00

0 0.

03

N

ot in

p ai

d w

or k

0. 19

0. 04

0 0.

00 0

0. 12

0. 10

0. 01

8 0.

00 0

0. 06

Ed

uc at

io n

st at

us (

re f:

de gr

ee o

r ab

ov e)

N

o qu

al ifi

ca tio

ns 0.

08 0.

04 3

0. 05

7 0.

04 0.

00 0.

01 9

0. 96

0 0.

00

St

an da

rd /G

C SE

− 0.

02 0.

03 1

0. 43

5 −

0. 02

0. 00

0. 01

3 0.

79 1

0. 00

H ig

he r/

A -le

ve l

− 0.

04 0.

03 7

0. 29

8 −

0. 02

− 0.

01 0.

01 6

0. 40

0 −

0. 01

V oc

at io

na l b

el ow

d eg

re e

− 0.

05 0.

02 3

0. 04

6 −

0. 04

0. 01

0. 01

0 0.

16 5

0. 01

So ur

ce : G

U S

sw ee

ps 1

–5 .

– no

in di

re ct

p at

h.

T ab

le 5

. (C

on tin

ue d)

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

690 Sociology 50(4)

Discussion

The results strongly demonstrate that experiencing financial vulnerability is signifi- cantly associated with higher levels of maternal emotional distress; an effect even greater than that of income. This corresponds to the literature where the presence of financial vulnerability is associated with high levels of stress, anxiety and depression in Swedish adults (Starrin et al., 2009). They show too that the broader, more subjective, concept of financial vulnerability is more salient than the objective measure of income, a finding premised in the theory of relative deprivation (Ragnarsdóttir et al., 2013). Financial vulnerability encapsulates the objective deprivation resulting from low income and the subjective deprivation associated with comparative aspects such as feel- ings towards coping on income. The results also suggest, as per Chambers’ (1989) assertion, that measures to alleviate low income, such as increased borrowing, increase vulnerability and that this vulnerability is keenly felt. This has relevance to academia, policy and practice and suggests that consideration should be given to financial vulner- ability when working with families experiencing poverty. For practice situations, financial vulnerability is an easily measured concept that could be employed to establish mothers’ heightened vulnerability and raised risk of emotional distress.

The results also strongly demonstrate that experiencing financial vulnerability is sig- nificantly associated with lower child SEB well-being which supports the findings in the qualitative literature (Harris et al., 2009; Whitham, 2012). The literature on older children indicates that the impact of financial vulnerability is directly experienced through the comparisons they make with their peer (reference) group (Ridge, 2002). It was hypothesised in this article that young children would be unable to make their own social comparisons and that the effect of financial vulnerability experienced by young children would be indirect, through the pathway of maternal emotional distress. This is indeed the case for almost two-thirds of the effect; however, the other third of the effect is experienced directly by the child despite their young ages. It is possible that the remaining third of the effect on the child may have another, unmeasured, pathway through a maternal characteristic or it may be that young children are directly affected by financial vulnerability. What is of note here is the young age of the children for whom this association is statistically visible.

These findings show that child SEB well-being is responsive to financial vulnerability and maternal emotional distress which suggests that were these conditions to change then child well-being might change too. That children of such a young age display lower SEB well-being when maternal vulnerabilities are high is a central finding of this article raising two thoughts: (1) children’s SEB well-being is highly sensitive to their mothers’ socioeconomic status and emotional distress; and (2) this implies that SEB is a malleable rather than a fixed trait that may respond to intervention.

There are two other results of incidental note; the effects of paid work on child and maternal well-being and the effects of family formation. A mother not being in paid work greatly increases her emotional distress and financial vulnerability. This in turn results in decreased child well-being. In a time of austerity measures that have disproportionately affected female employment the effect of not being in paid work on maternal emotional distress is substantively and statistically important. What is not modelled in this analysis is the nature of the effects of unpaid work on maternal emotional distress, that is, whether

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 691

unpaid care work leads to greater emotional distress or whether paid employment leads to reduced emotional distress, or a combination thereof. The literature in this area is large, varied and provides evidence to support both of these theses. This has not been modelled in this article but would make an interesting area of future study using the GUS data. The results in relation to family formation show that family transitions per se are not key to child SEB well-being but that they are key to financial vulnerability and, to a lesser extent, maternal emotional distress. The ‘couple who separated’ family transition was directly linked to poorer maternal emotional distress, which suggests that women experiencing separation would benefit from targeted emotional and financial support. The categories ‘stable lone parent’ and ‘re-partnered lone parent’ have no significant association with child SEB well-being directly, which suggests that the effect of a couple separating on maternal emotional distress is likely to be time-limited. In the end, only ‘separations and re-partnerings’ was directly linked to poorer child SEB well-being, which may indicate flux and uncertainty in a family’s life. As the literature using cross- sectional data often finds negative effects associated with lone-parenthood the longitudinal approach taken in this article benefits the analysis of the effects of family formation on child well-being.

Conclusions

In conclusion, this article shows that young children experience the effects of financial vulnerability indirectly via their mothers’ emotional distress and not through their own social comparisons as is the case with older children. It shows too that financial vulner- ability has a greater negative effect on maternal emotional distress than income per se suggesting that the subjective element of financial vulnerability is salient. It also shows that child SEB well-being is malleable and that children are emotional barometers responding to and corresponding with their mothers’ well-being.

A second conclusion is that actions taken to ameliorate the effects of income depriva- tion can paradoxically increase financial vulnerability, and that this is an added stressor for those living on low incomes, a factor which ought to be considered when working with, or legislating for, families living in socioeconomic disadvantage. Furthermore, not being in paid work affects maternal emotional distress and children’s SEB well-being independently, even after income and financial vulnerability are controlled for, which is important and may have implications for policy, for example, on access to affordable, high-quality childcare. Important too is the evidence elicited due to the longitudinal nature of the data that repeated separations and re-partnerings is the only category of family formation directly associated with reduced child SEB well-being. This provides evidence to counter the culture of blame towards ‘family breakdown’ common in current political discourse (DfE, 2012). It also has implications for practice: this evidence sug- gests that families who experience flux in their relationships are particularly vulnerable.

The implications of these findings are relevant to academia, policy and practice. It adds to the knowledge base in this field and makes the recommendation that financial vulnerability ought to be considered in tandem with income poverty and other measures of socioeconomic disadvantage.

The Growing Up in Scotland data are publicly available from the UK Data Service: http://ukdataservice.ac.uk/.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

692 Sociology 50(4)

Acknowledgements

I would like to thank my supervisors, Professor Kay Tisdall and Dr Paul Norris for their support.

Funding

I would like to thank the Economic and Social Research Council (ESRC) for the funding for my PhD (award number ES/HO13008/1) at the University of Edinburgh.

References

Acock AC (2013) Discovering Structural Equation Modeling Using Stata, Revised Edition. College Station, TX: Stata Press.

Alcock P (2006) Understanding Poverty. Basingstoke: Macmillan. Anderson S, Bradshaw P, Cunningham-Burley C, et al. (2007) Growing Up in Scotland: Sweep

One Overview Report. Edinburgh: Scottish Government. Barnes M, Chanfreau J and Tomaszewski W (2010) Growing Up in Scotland: The Circumstances

of Persistently Poor Children. Edinburgh: Scottish Government. Berthoud R and Bryan M (2011) Income, deprivation and poverty: A longitudinal analysis. Journal

of Social Policy 40: 135–156. Blair KA, Denham SA, Kochanoff A, et al. (2004) Playing it cool: Temperament, emotion regula-

tion, and social behavior in preschoolers. Journal of School Psychology 42: 419–443. Bradshaw J and Finch N (2003) Overlaps in dimensions of poverty. Journal of Social Policy 32:

513–525. Bradshaw P (2011) Growing Up in Scotland: Changes in Child Cognitive Ability in the Pre-School

Years. Edinburgh: Scottish Government. Bradshaw P and Tipping S (2010) Growing Up in Scotland: Children’s Social, Emotional and

Behavioural Characteristics at Entry to Primary School. Edinburgh: Scottish Government. Bradshaw P, Marryat L, Corbett J, et al. (2009) Growing Up in Scotland Sweep 4: 2008–2009 –

User Guide. Edinburgh: Scottish Government. Bradshaw P, Marryat L, Mabelis J, et al. (2010) Growing Up in Scotland Sweep 5: 2009–2010

– User Guide. Edinburgh: Scottish Government. Bromley C (2009) Growing Up in Scotland: The Impact of Children’s Early Activities on Cognitive

Development. Edinburgh: Scottish Government. Chambers R (1989) Editorial introduction: Vulnerability, coping and policy. IDS Bulletin 20: 1–7. Chanfreau J and Burchardt T (2008) Equivalence Scales: Rationales, Uses and Assumptions.

Edinburgh: Scottish Government. Cohn LD (1991) Sex differences in the course of personality development: A meta-analysis.

Psychological Bulletin 109: 252–266. Conger RD, Conger KJ and Martin MJ (2010) Socioeconomic status, family processes, and indi-

vidual development. Journal of Marriage and the Family 72(3): 685–704. Corbett J, Marryat L and Bradshaw P (2005) Growing Up in Scotland Sweep 1 User Guide. UK

data archive study number 5760. Edinburgh: Scottish Centre for Social Research. Corbett J, Marryat L and Bradshaw P (2006) Growing Up in Scotland Sweep 2 User Guide. UK

data archive study number 5760. Edinburgh: Scottish Centre for Social Research. Corbett J, Marryat L, Bradshaw P, et al. (2007) Growing Up in Scotland Sweep 3 User Guide. UK

data archive study number 5760. Edinburgh: Scottish Centre for Social Research. DfE (2012) Measuring Child Poverty: A Consultation on Better Measures of Child Poverty.

London: HM Stationery Office. Fouarge D and Layte R (2005) Welfare regimes and poverty dynamics: The duration and recur-

rence of poverty spells in Europe. Journal of Social Policy 34: 407–426.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

Treanor 693

Goodman R (1997) The Strengths and Difficulties Questionnaire: A research note. Journal of Child Psychology and Psychiatry 38: 581–586.

Green M (2007) Voices of People Experiencing Poverty in Scotland. York: Joseph Rowntree Foundation and the Poverty Alliance.

Hansen K, Joshi H and Dex S (2010) Children of the 21st Century. Bristol: Policy Press. Harris J, Treanor MC and Sharma N (2009) Below the Breadline. London: Barnardo’s. Holscher P (2008) Children and young people’s experiences of poverty and social exclusion. In:

Ridge T and Wright S (eds) Understanding Inequality, Poverty and Wealth. Bristol: Policy Press, 181–203.

Hu L and Bentler PM (1999) Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multi- disciplinary Journal 6: 1–55.

Jenkins SP, Rigg JA, Devicienti F, et al. (2001) The Dynamics of Poverty in Britain. Leeds: Corporate Document Services for the Department for Work and Pensions.

Jenkinson C, Layte R, Jenkinson D, et al. (1997) A shorter form health survey: Can the SF-12 rep- licate results from the SF-36 in longitudinal studies? Journal of Public Health 19: 179–186.

Kiernan KE and Huerta MC (2008) Economic deprivation, maternal depression, parenting and children’s cognitive and emotional development in early childhood. British Journal of Sociology 59: 783–806.

Kiernan KE and Mensah FK (2009) Poverty, maternal depression, family status and children’s cognitive and behavioural development in early childhood: A longitudinal study. Journal of Social Policy 38: 569–588.

King K and Gurian M (2006) Teaching to the minds of boys. Educational Leadership 64: 56–58. Lister R (2004) Poverty. Cambridge: Polity. Magadi M and Middleton S (2007) Severe Child Poverty in the UK. London: Save the Children. Melhuish EC, Phan MB, Sylva K, et al. (2008) Effects of the home learning environment and pre-

school center experience upon literacy and numeracy development in early primary school. Journal of Social Issues 64: 95–114.

Merton RK (1968) Social Theory and Social Structure. New York: Free Press/London: Collier- Macmillan.

Ragnarsdóttir BH, Bernburg JG and Ólafsdóttir S (2013) The global financial crisis and individual distress: The role of subjective comparisons after the collapse of the Icelandic economy. Sociology 47: 755–775.

Ridge T (2002) Childhood Poverty and Social Exclusion: From a Child’s Perspective. Bristol: Policy Press.

Ridge T (2009) Living with Poverty: A Review of the Literature on Children’s and Families’ Experiences of Poverty. London: Department for Work and Pensions.

Runciman WG (1966) Relative Deprivation and Social Justice: A Study of Attitudes to Social Inequality in Twentieth-Century England. London: Routledge & Kegan Paul.

Schoon I, Hope S, Ross A, et al. (2010) Family hardship and children’s development: The early years. Longitudinal and Life Course Studies 1: 209–222.

Schoon I, Jones E, Cheng H, et al. (2012) Family hardship, family instability, and cognitive devel- opment. Journal of Epidemiology & Community Health 66: 716–722.

Shildrick T, MacDonald R, Webster C, et al. (2013) Poverty and Insecurity [Electronic Resource]: Life in Low-Pay, No-Pay Britain. Bristol: Policy Press.

Smith HJ, Pettigrew TF, Pippin GM, et al. (2012) Relative deprivation: A theoretical and meta- analytic review. Personality & Social Psychology Review 16: 203–232.

Starrin B, Aslund C and Nilsson KW (2009) Financial stress, shaming experiences and psycho- social ill-health: Studies into the Finances–Shame Model. Social Indicators Research 91: 283–298.

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from

694 Sociology 50(4)

Statacorp (2013) Stata 13 Structural Equation Modelling Reference Manual. College Station, TX: Stata Press.

Stewart T (2009) Counting on Credit. London: Barnardos. Thomas C, Benzeval M and Stansfeld SA (2005) Employment transitions and mental health: An

analysis from the British Household Panel Survey. Journal of Epidemiology and Community Health 59: 243–249.

Townsend P (1979) Poverty in the United Kingdom: A Survey of Household Resources and Standards of Living. Harmondsworth: Penguin.

Ware JE Jr, Kosinski M and Keller SD (1996) A 12-item short-form health survey: Construction of scales and preliminary tests of reliability and validity. Medical Care 34: 220–233.

Whelan CT and Maitre B (2005) Vulnerability and multiple deprivation perspectives on economic exclusion in Europe: A latent class analysis. European Societies 7: 423–450.

Whelan CT and Maitre B (2008) Social class variation in risk: A comparative analysis of the dynamics of economic vulnerability. British Journal of Sociology 59: 637–659.

Whitham G (2012) Child Poverty in 2012: It Shouldn’t Happen Here. Manchester: Save the Children.

Yeung WJ, Linver MR and Brooks-Gunn J (2002) How money matters for young children’s devel- opment: Parental investment and family processes. Child Development 73(6): 1861–1879.

Zick CD, Bryant WK and Osterbacka E (2001) Mothers’ employment, parental involvement, and the implications for intermediate child outcomes. Social Science Research 30: 25–49.

Morag Treanor is a Nuffield-funded Q-Step lecturer in quantitative social policy at the University of Edinburgh. Her current research uses birth cohort data, mainly the Growing Up in Scotland study, to explore the impacts of longitudinal poverty and persistently low/high incomes on chil- dren’s developmental outcomes. Her research is concerned with the impact of family, peer and social relations on child well-being. She is an experienced quantitative and qualitative researcher in relation to socioeconomic inequalities and vulnerable populations. She is presently conducting longitudinal qualitative research with families affected by welfare reform with the Child Poverty Action Group in Scotland and is a member of their Expert Advisory Group.

Date submitted January 2014 Date accepted January 2015

at Glasgow University Library on September 12, 2016soc.sagepub.comDownloaded from