How School-Based Wellbeing Interventions, Socioeconomic Status, and Gender Impact Anxiety Development in Middle Childhood

Abstract

The current study examined whether delivering wellbeing interventions in schools impacted anxiety development in middle childhood. Schools have an important role to play in addressing children’s anxiety through daily practice in classrooms and by implementing specialist wellbeing interventions that support children’s coping. A nationally representative sample of Irish primary school children (N = 2313; 48.9% male) were surveyed when they were in second class (M age 8.09 years, SD = 0.39) and fourth class (M age 9.96 years, SD = 0.40) of primary school. Teachers reported on wellbeing interventions in schools. Key findings from moderation analyses suggest that family affluence and being female predicted a greater increase in anxiety development and that wellbeing interventions interacted with family affluence to reduce anxiety development for more affluent children. These findings can be used to strengthen the design and implementation of wellbeing interventions for a more efficacious approach to minimising children’s anxiety development.

1. Introduction

Across childhood, anxiety is typically low and stable; however, for some children, anxiety can gradually increase, rising to clinically high levels [1]. Risk factors for developing anxiety include being female [2,3] and experiencing the environmental stressors that come with socioeconomic disadvantage [4,5,6]. If left untreated, problems often deteriorate; therefore, early treatment of anxiety is of the utmost importance. Across the lifespan, experiencing higher levels of anxiety is associated with numerous adverse outcomes, including reduced perceived quality of life, lower educational attainment, and higher instances of substance abuse in young people and adults [7]. However, only a small minority of children with anxiety receive mental health support [8]. As schools offer the opportunity to reach large numbers of children at key developmental stages, they are regarded as appropriate settings for the early treatment of anxiety [9]. Early intervention may aid in improving the present as well as future wellbeing of the individual child, and it may also be beneficial to society at large given that it may reduce societal costs relating to anxiety problems across the lifespan, such as school drop-out, medication use, unemployment, and so on [8].

Wellbeing interventions, also known as social and emotional learning interventions, are used in schools to further children’s capabilities to enhance their social, emotional, and behavioural development and improve the quality of their relationships [10]. Social, emotional, and behavioural skills are integral to managing stressors and coping [11]. Wellbeing intervention programmes such as the “FRIENDS for life” intervention programme have been found to be efficacious in the treatment of anxiety in children [5,12]. Key mechanisms of change in successful interventions include emotional coaching and help with managing bullies [12]. Efficacy studies have been conducted globally and have shown positive results leading to a decrease in anxiety symptoms immediately following completion of such programmes as well at 1–3 year follow-up assessments [12,13,14]. Due to the aforementioned evidence, the World Health Organisation recommends wellbeing intervention programmes in the treatment of anxiety in children [15].

The results of Neil and Christensen’s (2009) systematic literature review of school-based interventions for anxiety revealed that most interventions were successful in reducing symptoms of anxiety in children and teenagers, while in their meta-analysis and systematic literature review, Werner-Seidler et al. (2017) found a small effect for school-based interventions on children’s anxiety levels at post intervention, which was maintained at both short- and medium-term follow-ups. A significant limitation of the research on school-based wellbeing interventions is a lack of statistical power [16], and most reviews have been conducted within the United States of America, which makes cross-cultural generalisation difficult to achieve [17]. This highlights the need for further international research on the efficacy of wellbeing interventions for reducing anxiety in childhood with larger, nationally representative samples.

The current study focused on whether school-based wellbeing interventions impacted the development of anxiety in middle childhood in a nationally representative sample of children in Ireland. Taking a sociocultural perspective, the study also explored whether child gender and the socioeconomic status of individual children and of the child’s school could moderate anxiety development both directly and in interaction with wellbeing interventions. Accordingly, the study gives insight into how wellbeing interventions delivered in schools impacted children’s anxiety in middle childhood in relation to socioeconomic disadvantage and gender.

1.1. Anxiety in Middle Childhood

The term anxiety refers to “a person’s conscious state of worry over a future unwanted event, or fear of an actual situation” [3] p. 189. The Diagnostic and Statistical Manual of Mental Disorders—5th Edition [DSM-5] [America Psychological Association, 2013] and International Classification of Diseases—11th Edition [ICD-11] [18] provide systems for classifying anxiety problems that consider the developmental timing of their manifestation, the type of stimuli that provoke the anxious response, the pervasiveness and nature of responses to stimuli, and the role of identifiable factors in the development of anxiety. Both the DSM-V and the ICD-11 identify the condition of generalised anxiety disorder [GAD], whereby a variety of stimuli provoke an anxious response. GAD is especially common in childhood [19]. Girls are more likely to experience anxiety disorders than boys, although this gender difference is more salient in clinical samples as opposed to community samples, as girls are also more likely to receive treatment for their anxiety [3]. It is widely recognised that anxiety disorders have the potential to cause impaired psychological and social functioning. Studies have revealed a high comorbidity between the various types of anxiety disorders and between anxiety and other types of psychiatric disorders, particularly depression and substance abuse [20]. Chronic anxiety is also associated with adverse outcomes in childhood such as impaired social and emotional development and poorer academic outcomes [11]. Despite the impairments caused by anxiety across multiple domains, the number of children who seek treatment for anxiety disorders remains relatively low [21]. Middle childhood refers to the period of individual development between approximately five and twelve years of age [19,22]. Children become more vulnerable to developing anxiety as their cognitive capacities increase, meaning that middle childhood is a window wherein anxiety can manifest [23]. A meta-analysis of the age of onset of anxiety disorders identified that certain types of anxiety disorder have their average onset in middle childhood and early adolescence, including separation anxiety disorder [age 10.6 years], social phobia [age 11.0 years], and social phobia [age 14.3 years] [24]. School settings may contribute to social-oriented anxiety in middle childhood, with increasing emphasis on academic performance in the later years of primary/middle school, at age 11/12 years, interacting with children’s heightened awareness of themselves in comparison to others and leading to worries about their academic performance and social standing [19]. Therefore, research on the development of anxiety in middle childhood in school-based settings may be useful to conduct.

1.2. Anxiety and Socioeconomic Status

Socioeconomic status [SES] is challenging to define and measure, with great variation in how SES is conceptualised across the literature [25]. The American Psychological Association [APA] defines SES as the social standing or class of an individual or group of people and highlights education, income, and occupation as key indicators of SES. Examinations of SES often reveal inequities in access to resources plus issues related to privilege, power, and control [26]. At an individual level, the social stressors occurring because of limited social and financial resources can enrich the conditions for promoting anxiety in children [4]. However, children from families with higher SES can exhibit higher levels of anxiety than their lower-SES counterparts in some studies [27,28,29]. Qualitative research indicates that girls in higher-SES families can feel anxious about perceived pressure to produce an ideal image of themselves that is a match with their middle- and upper-class family status [29], and children in higher-SES schools have also reported higher anxiety possibly related to perceived pressures to achieve high academic results during standardised testing [28].

1.3. Socioeconomic Status in Schools and Families in Ireland

In Ireland, the setting of the current study, SES is also conceptualised as a multidimensional state of resource deprivation at a geographic level. For example, the Pobal Haase–Pratschke [HP] deprivation index [30] was developed as a means of classifying geographic localities by their SES. Indicators of geographic SES used by the index include the prevalence of lower educational outcomes, single-parent families, overcrowding in households, lower occupational status, and unemployment rates. In Ireland, schools located in areas with greater social and economic deprivation are given increased resources by the government to try to resolve issues of educational capital inequality. Those schools are classified under the Delivering Equality of Opportunity in Schools scheme and are referred to as DEIS schools. Recent research has identified that in Ireland, attending DEIS schools has a negative impact on college enrolment over and above individual-level SES [31]. Possibly, differential effects of individual SES and school-level SES can also be observed for anxiety development, perhaps due to qualitatively different stressors being present in the low-income school versus family environment.

1.4. Wellbeing Interventions and Anxiety in Middle Childhood

Wellbeing can be defined as experiencing a positive emotional state [e.g., happiness], and personal accomplishment and fulfilment. Maintaining personal wellbeing can be challenging when environments offer more risk than opportunity and when children are ill prepared to cope with stressors that undermine their ability to maintain happiness [32]. Wellbeing interventions [a term commonly used in Europe] are also known as social and emotional learning [SEL] interventions [a term commonly used in the United States]. Wellbeing interventions aim to support children in developing their competencies for emotional regulation, goal setting and striving, perspective taking, communication, decision making, and managing interpersonal conflicts [33]. In schools, wellbeing interventions are often applied at the whole-school or whole-year group level, with the notion that these should have a positive impact for all students. The evidence base for wellbeing interventions has grown considerably since rigorous trials first established their positive impact on social, emotional, and behavioural outcomes [33] This compelling evidence base, along with the requirement for schools to teach social and emotional skills to children as stipulated in most contemporary policies and curricula worldwide, means that more schools are implementing wellbeing interventions now than ever before [34]. Because anxiety is widespread in childhood and adolescence, wellbeing interventions present an exciting new approach to protecting children against anxiety using schools as the medium and a holistic intervention targeting social and emotional development as the mode. Furthermore, because the targets of wellbeing interventions are diverse, more research is needed on how wellbeing interventions impact anxiety specifically. Further research is also needed on whether wellbeing interventions have similar impacts across diverse groups of students. This is because wellbeing interventions are manualised with few variations available to suit students with different types of learning needs or social backgrounds. Of interest to the current study, the potential for wellbeing interventions to protect children from developing anxiety might vary depending on children’s gender and on the social context of children’s schools and families. Put simply, do wellbeing interventions have the same impact for both genders and for children regardless of their family SES or low-income schooling?

1.5. Developmental Perspective on Anxiety in Middle Childhood

In helping to understand children’s anxiety development in relation to social context, it is important to consider the many factors that interact between person and environment across time. Bronfenbrenner’s process-person-context-time [PPCT] model [Bronfenbrenner and Evans, 2000] emphasises how individual characteristics, combined with aspects of the environment, influence human development. Proximal processes, described as “the engines of development” [35] p. 118, refer to the daily interactions between people and their immediate environments. The impact of proximal processes on individual development varies according to individual characteristics, environmental contexts, and the period over which the processes occur [36]. Bronfenbrenner recognised the potential of proximal processes to promote positive or dysfunctional outcomes for individuals, concluding that stable and advantageous contexts result in a greater possibility of producing positive outcomes [37]. Accordingly, the current study aims to examine proximal processes by exploring how anxiety develops in relation to individual characteristics [namely gender] and the contextual factors of individual and school SES and wellbeing interventions. In addition to Bronfenbrenner’s PPCT theory, the developmental psychopathology framework contributes to an understanding of how anxiety develops across childhood. The developmental psychopathology framework is important to consider, as it has become the dominant approach for helping to understand the origin of mental health difficulties in young people over the past decade [38]. The developmental psychopathology approach aligns with Bronfenbrenner’s bioecological model in its position on psychopathology development being a complex interplay of psychobiological vulnerabilities with contextual risk factors such as low-income schooling and lower family SES.

1.6. The Current Study

As outlined above, anxiety is prevalent in childhood and has been systematically related to gender and to the challenges of socioeconomic deprivation. Wellbeing interventions offer an opportunity for all children to develop coping skills to help prevent increases in anxiety across time. However, the impact of wellbeing interventions on anxiety in early childhood is not often explored in larger samples, with nationally representative capability. Furthermore, the role of social context in the effectiveness of wellbeing interventions should be explored, in line with Bronfenbrenner’s PPCT and bioecological models [35,36]. To generate knowledge in these areas, the current study has three aims. The first aim is to test whether children’s anxiety development is moderated by socioeconomic status, building on studies that find a relationship between socioeconomic status and anxiety development in childhood [4]. The current study extends the previous literature by testing whether anxiety development is moderated by [a] low-income schooling and [b] individual SES within the same model. This aim sets the scene for testing the role of wellbeing interventions in children’s anxiety development in the context of varying socioeconomic circumstances. The second aim tests whether school-based wellbeing interventions moderate children’s anxiety development and whether this effect differs according to children’s gender. This aim is informed by commonly found gender differences in children’s anxiety [3] and inconclusive evidence on whether wellbeing intervention effects differ by children’s gender [39] while testing the main pathway of interest: the impact of wellbeing interventions on children’s anxiety development. The third aim builds on the first and second aims by testing whether the impact of socioeconomic status on children’s anxiety development is moderated by school-based wellbeing interventions. This third aim question builds on aims 1 and 2 to identify whether the potential associations between SES and anxiety development could be influenced by wellbeing interventions. This aim is informed by previous research where wellbeing interventions are observed to have potential to prevent or delay the manifestation of mental health difficulties [16]. We endeavoured to investigate whether wellbeing interventions were equally as effective for children of both low and high SES, while we were also interested to see whether a negative SES–anxiety relationship would be observed.

• To what extent does socioeconomic status impact the development of children’s anxiety? To investigate this issue, we were interested in examining the impact of both individual family SES and low-income schooling, in line with Bronfenbrenner’s bioecological model of multiple nested developmental contexts [35,36]
• To what extent do wellbeing interventions and gender impact the development of children’s anxiety? We were interested in examining whether wellbeing interventions had a stronger impact for girls versus boys, based on inconclusive findings in the literature on the effectiveness of wellbeing interventions in relation to gender [39]
• To what extent do wellbeing interventions moderate the impact of socioeconomic status on the development of children’s anxiety? This third research question builds on questions 1 and 2 to identify whether the potential associations between SES and anxiety development could be influenced by wellbeing interventions. This question builds on previous research where wellbeing interventions were observed to have potential to prevent or delay the manifestation of mental health difficulties [16] We endeavoured to investigate whether wellbeing interventions were equally as effective for children of both low and high SES, while we were also interested to see whether a negative SES–anxiety relationship would be observed.

2. Materials and Methods

2.1. Participants and Procedures

The study is a secondary data analysis of the Irish cohort study of primary schooling: Children’s School Lives (CSL) [40]. In CSL, a random sample of primary schools was drawn from all primary schools in Ireland. Schools were recruited into the study until the target sample of 100 schools was reached for each of two age cohorts. Data are collected annually with children, their class teachers, school principals, and parents. The current study analyses longitudinal cohort B data from wave 1 (May–June 2019) and wave 3 (May–June 2021). Wave 2 data (May–June 2020) were not included due to the small sample size resulting from COVID-19 school closures. A dataset was requested from the CSL study team that included the variables measuring children’s self-reported anxiety at waves 1 and 3, child gender, child self-reported family affluence (the most complete measure of childhood SES) at wave 3, Department of Education statistics on low-income schooling status, and teacher-reported use of wellbeing interventions at wave 3. The total sample of children across both waves (N = 2313) included 50% male (n = 1132) and 50% female (n = 1147), with missing data on gender for 34 children. Most children (87%) were born in Ireland and spoke English or Irish at home (86%). At wave 1, children were aged 7–9 years (M = 8.07, SD = 0.39), and at wave 3, children were aged 9–11 years (M = 9.96, SD = 0.40). The number of children with self-reported data on anxiety was 1863 at wave 1 and 1978 at wave 3. The difference in number is due in part to non-attendance on the day of data collection, children withdrawing from the study because of a change of schools, children not assenting for their self-report data to be used, and additional children being recruited into the study in wave 3 due to school mobility. Trained fieldworkers administered child questionnaires in children’s classrooms in the form of paper workbooks at each wave of data collection. In wave 3, 94 teachers completed an online survey where they reported on whether they had used a wellbeing intervention with their class during the current year. Of those teachers, 25% (n = 23) were men, and 75% (n = 72) were women, whereas 3 teachers did not report their gender. Most teachers were aged between 26 and 45 years old (79.5%). All teachers spoke Irish or English as their first language. Most teachers either had a bachelor’s degree (33%), a postgraduate diploma (34%), or a masters’ degree (26%) as their highest qualification. The CSL study received ethical approval from the UCD Human Research Ethics Committee (HREC) and is compliant with the General Data Protection Regulation (GDPR) Dublin, Ireland. Parents received consent forms and other information pertinent to the CSL study and signed consent forms were returned to the CSL study team. Children who had parental consent to participate were involved in a 2.5 h workshop that aimed to give them a greater understanding of the research process and what they would have to do if they wished to participate in the CSL study. After the workshop, children signified whether they would like their answers in the child workbook to be used in the CSL study, which ensured their informed assent was gained. Participating teachers also provided informed, written consent to the CSL researchers (16).

2.2. Measures

2.2.1. Anxiety

Children’s self-reported anxiety was measured at wave 1 and wave 3 using three questions adapted from the Generalised Child Self-report subscale of the Revised Children’s Anxiety and Depression Scale (RCADS). The three questions were as follows: “I worry about things”; “I worry something bad will happen to me”; and “I worry about what is going to happen”. Each question was measured on a five-point Likert scale (1 = never to 5 = always). The anxiety measure was reliable at wave 1 (α = 0.76) and wave 3 (α= 0.85).

2.2.2. Individual Socioeconomic Status

Children’s self-reported family SES was measured at wave 3 using five questions from the Family Affluence Scale (FAS). Children responded to “Do you have your own bedroom for yourself?” (1 = no, 2 = yes), “Do you have a dishwasher at home?” (1 = no, 2 = yes), “Does your family own a car, van or truck?” (1 = no, 2 = one, 3 = two, 4 = more than two), “How many computers does your family own?” (1 = none, 2 = one, 3 = two, 4 = more than two), and “How many bathrooms (rooms with a bath/shower or both) are in your home?” (1 = none, 2 = one, 3 = two, 4 = more than two). A total score for the scale was computed. Possible scores ranged from 5 to 16, with a higher score denoting a higher family SES.

2.2.3. Low-Income Schooling

Low-income schooling was represented using schools’ DEIS classification, which was collected from Department of Education statistics during the sampling procedure and updated at each wave with principal report. The item was scored 1 = non-DEIS, 2 = DEIS.

2.2.4. Wellbeing Interventions

Use of wellbeing interventions was measured at wave 3 in the teacher online questionnaire. Wellbeing interventions were taken to be specifically designed programmes with manualized course materials that schools can opt into rather than the standard curricular support for wellbeing provided by the Irish social, personal, and health education curriculum. The reason that the presence of optional wellbeing interventions was chosen for analysis is because the PHSE curriculum is present in all schools, meaning that there would be no variance in a measurement of PHSE. All items were constructed by the CSL team. Teachers were asked to indicate which wellbeing interventions commonly used in Irish primary schools that they had used with their class that year. Teachers were given the options of “none”, “Weaving Wellbeing”, “Nurture Groups”, “Incredible Years”, “Roots of Empathy”, “Zippy’s Friends”, “FRIENDS for Life”, “Making Meaning”, and “Other”. “Making Meaning” was a fabricated programme included in the answer options to ensure that teachers were reliably reporting on their use of wellbeing interventions. No teacher reported the use of “Making Meaning”, indicating a high accuracy of responses. Wellbeing interventions were thereafter grouped into a binary variable, with 0 assigned as a code when no wellbeing intervention was used and 1 assigned as a code when a wellbeing intervention was used. There are limitations associated with this approach; for example, the unique structure and content of each programme were not assessed for their effectiveness at reducing anxiety levels, so the differential impact of programmes was not captured in this study. In addition, some programmes may be more effective at reducing anxiety levels than others, for example, those with a CBT theoretical foundation such as FRIENDS for Life. The decision to group all wellbeing interventions together means that the nuances of individual programmes are not accounted for, which reduces the granularity of the results. It also limits our interpretive potential, as only broad statements about wellbeing interventions in general can be made, meaning some interventions may be poorly represented by the conclusions made within this study.

2.3. Analysis Plan

Descriptive statistics and bivariate correlations were computed using SPSS Version 26. As the three research questions regarded the impact of moderating variables on the relationship between an independent and dependent variable, moderation analysis was chosen as the most appropriate method. Statistically, moderation is where the nature of the relationship between a predictor variable and an outcome variable is altered according to the value of a moderating variable or variables [41]. The PROCESS Macro for SPSS (28) was used to compute the moderation analyses. To address the first aim regarding socioeconomic deprivation and the development of children’s anxiety, the moderating impact of individual student SES (M) on the relationship between low-income schooling (X) and the development of child anxiety between wave one (covariate) and wave three (Y) was analysed. Next, to address the second aim about the impact of wellbeing interventions on anxiety development, the moderating impact of gender (M) on the relationship between the use of a wellbeing intervention (X) and the development of anxiety between waves one (covariate) and three (Y) was examined. Finally, to address the third aim about whether wellbeing interventions moderate the impact of SES on children’s anxiety development, a moderation analysis was conducted to explore whether the use of a wellbeing intervention (M) moderated the relationship between SES (X) and the development of anxiety between wave one (covariate) and wave three (Y). In each of the three models, we controlled for the dependent variable of anxiety at wave one before examining the impact of the main and secondary moderators (socioeconomic status, wellbeing, and gender) on the dependent variable of anxiety at wave three. This allowed us to identify the impact of the moderators on change in anxiety that occurred between the two waves.

3. Results

Descriptive statistics and bivariate correlations are presented in

Table 1. Descriptive Statistics.

	Variable Names	Score	N	%
1	Gender		2279	100
	Boys	1	1140	51
	Girls	2	1139	49
2	Low-income schooling		2294	100
	Children in low-income schools	1	1537	67
	Children not in low-income schools	2	757	33
		Min–Max	N	M	SD
3	Child SES	6–16	1947	12.85	1.91
4	Anxiety W1	1–5	1863	2.64	1.05
5	Anxiety W3	1–5	1978	2.88	1.01
6	Wellbeing intervention	0–1	1548	0.51	0.50

min, minimum; max, maximum; N, total number; M, mean; SD, standard deviation; %, percentage; W1, W3, data collection waves one and three; SES, socioeconomic status.

and

Table 2. Bivariate Correlations for Study Variables.

		1	2	3	4	5
1	Gender (Girl)	1
2	Low-income schooling	−0.15 **	1
3	Child SES	−0.01	−0.16 **	1
4	Anxiety W1	0.06 *	0.02	−0.06 *	1
5	Anxiety W3	0.12 **	0.04	0.02	0.29 **	1
6	Wellbeing intervention	−0.02	0.10 **	0.01	−0.02	−0.03

* p < 0.05; ** p < 0.01. W1, W3, data collection waves one and three; SES, socioeconomic status.

for the whole sample.

3.1. Changes in Anxiety across Time

As can be seen in Table 1, children’s self-reported anxiety increased between wave 1 (M = 2.64, SD = 1.05) and wave 3 (M = 2.88, SD = 1.01). A paired-samples t-test identified this change as statistically significant (t(1600) = −8.00, p < 0.001). The independent-samples t-tests revealed no significant differences in the anxiety levels of children depending on whether they attended a low-income school at wave 1 (t(1130) = −0.98, p = 0.33) or wave 3 (t(1174) = −1.51, p = 0.13). Child gender was associated with anxiety at wave 1 (t(1861) = −2.42, p = 0.02) and wave 3 (t(1974) = −5.12, p < 0.001). Compared to boys, girls had higher levels of anxiety at wave 1 (girls M = 2.70, SD = 1.05; boys M = 2.58, SD = 1.05) and wave 3 (girls M = 2.99, SD = 1.02; boys M = 2.76, SD = 1.00).

3.2. Moderation Analysis One

The first moderation analysis examined whether low-income schooling impacted children’s anxiety development and whether this impact was moderated by individual family SES. All regression coefficients are unstandardized (b). The overall model was significant (F(4, 1565) = 40.79, p < 0.001, R² = 0.09), with the effect size suggesting that 9% of the variance in the model was explained by the predictor variable (low-income schooling), correlate (anxiety wave 1), and moderator (individual level SES). There was no main effect for low-income schooling on the development of anxiety (b = 0.57, t(1565) = 1.63, p = 0.10) and no interaction between low-income schooling and individual family SES on the development of anxiety (b =−0.04, t(1565) =−1.43, p = 0.15). Results revealed a main effect for individual family SES on the development of anxiety (b = 0.08, t(1565) = 2.03, p = 0.04) so that for every one unit increase in individual family SES, anxiety increased by 0.08 of a point on the scale. Figure 1 clearly illustrates this. Although not statistically significant, Figure 1 also shows that children attending low-income schools experience similarly high levels of anxiety irrespective of their individual SES, while as SES increased for those in higher-income schools, so did anxiety.

3.3. Moderation Analysis Two

A second moderation analysis was conducted to see if wellbeing interventions impacted anxiety development and whether this pathway was moderated by child gender. The overall model was significant (F(4, 1122) = 38.19, p < 0.001, R² = 0.12) and suggested that 12% of the variance in anxiety at wave 3 could be explained by the predictor variable (wellbeing interventions), covariate (anxiety wave 1), and moderator (gender). The results revealed that the use of a wellbeing intervention did not impact the development of anxiety (b = 0.13, t(1122) = 0.73, p = 0.47), and neither did the interaction of wellbeing interventions and gender (b = −0.12, t(1122) = −1.08, p = 0.28). A significant main effect for gender on the development of child anxiety was found (b = 0.33, t(1122) = 4.13, p < 0.001). As can be seen in Figure 2, females have significantly higher anxiety levels than males. Although not statistically significant, Figure 2 also suggests that females who engaged with a wellbeing intervention experienced lower levels of anxiety than those who did not, while there was no significant difference in the anxiety levels of males depending on whether they engaged with a wellbeing intervention.

3.4. Moderation Analysis Three

A third moderation analysis was conducted to explore whether the use of a wellbeing intervention moderated the relationship observed in the first moderation analysis between individual family SES and the development of anxiety. This model was also significant (F(4, 1098) = 33.76, p < 0.001, R² = 0.11), with 11% of the variance being explained by the predictor variable (individual family SES), the covariate (anxiety wave 1), and the moderator (wellbeing intervention). The results revealed a significant main effect for individual family SES (b =0 0.06, t(1098) = 2.52, p = 0.01) and the use of a wellbeing intervention (b = 0.98, t(1098) = 2.41, p = 0.02). SES and the use of a wellbeing intervention also significantly interacted to predict the development of anxiety (b = −0.08, t(1098) = −2.60, p = 0.01).

A simple slopes analysis provided further information about the interaction between individual family SES and the use of a wellbeing intervention. When a wellbeing intervention was not implemented, every one point increase in SES resulted in a 0.06 increase in the anxiety scale (b = 0.06, t(1098) = 2.52, p = 0.01). Where a wellbeing intervention was used, there was no relationship between SES and anxiety (b = −0.02, t(1098) = −1.13, p = 0.26). This is illustrated in Figure 3, whereby it is shown that children with higher SES experience higher anxiety levels when they are not exposed to a wellbeing intervention, while a lack of engagement with a wellbeing intervention does not result in higher levels of anxiety for children with lower SES. This suggests that wellbeing interventions may be more influential on the anxiety levels of higher-SES children while having less impact on the anxiety levels of children with lower SES.

4. Discussion

Anxiety is the most common type of mental health difficulty in children [6], and middle childhood is a window where specific types of anxiety disorder begin to emerge [24]. With wellbeing interventions becoming widespread in schools in Ireland, this presents an opportunity to test their efficacy for impacting anxiety development in middle childhood in consideration of children’s socioeconomic status and gender. Using an Irish national sample of children in primary schools, the current study found that higher family SES and being female significantly predicted the development of anxiety, with anxiety increasing more for children from higher SES backgrounds and for girls. In addition, individual family SES and the use of a wellbeing intervention significantly interacted to predict the development of anxiety, wherein wellbeing interventions were effective at reducing the anxiety levels of children with higher SES. Memon et al. (2019) emphasised the importance of returning to theory when interpreting results so that they are explained from a theoretical viewpoint, so the results are explained in relation to Bronfenbrenner’s PPCT theory [35,36] and the developmental psychopathology framework e.g., [42].

4.1. Anxiety Development and SES

In the current study, children with higher SES had greater anxiety development across time compared to children with lower SES, which is contrary to our assumptions that living with more socioeconomic stressors in childhood would equate with greater anxiety. Some studies have found similar results, for example, Farrell et al.’s (2009) study wherein children in higher-SES schools reported significantly higher levels of anxiety than their counterparts in lower-SES schools. Bronfenbrenner’s PPCT model provides a framework that prompts us to consider the many interacting factors that may help to explain why children from higher SES backgrounds emerged with greater anxiety development in this study. The PPCT model recognises that proximal processes, which are the reciprocal interactions between a developing human and the people and environments that surround them, can promote desirable or adverse outcomes for children [37]. Cultural differences exist in parents’ beliefs, goals, and parenting styles and practices that link to SES. For example, parents with higher SES provide earlier age estimates for when their children will achieve developmental milestones and higher estimates of young children’s capabilities than lower-SES parents [42] Qualitatively, Spencer et al. (2018) investigated the relationship between affluence and an increased vulnerability for psychosocial distress among teenage girls. It emerged that universal experiences of pressures to perform, narrow views of what constitutes success, and disparities in expectations between teenagers and their parents were some of the main sources of distress. Taken together, this evidence suggests that higher-SES parents may have higher expectations for their children, and this in turn may result in those children having greater anxiety. However, as both Bronfenbrenner’s PPCT model and the developmental psychopathology framework emphasise, development occurs via multi-faceted pathways in a nonlinear manner [43]. High SES alone is unlikely to explain elevated anxiety levels, meaning that an interplay of genetic and environmental risk factors likely contributed to the development and maintenance of anxiety in each individual [42].

4.2. Anxiety Development, Gender, and Wellbeing Interventions

There were significant differences in anxiety between boys and girls at both waves, with girls having higher levels of anxiety than boys each time. According to Bronfenbrenner’s PPCT model, gender is a person-level characteristic that has the potential to promote or stifle reactions from the social environment, which in turn influences proximal processes. The finding that girls are more vulnerable than boys to developing anxiety is strongly supported in the literature [44]. Zahn-Waxler et al. (2008) stipulated that gender is likely to play a role in differential mental health outcomes due to the frequent emergence of risk-by-gender interactions. They use a developmental psychopathology framework to explore early gender differences of different disorders from biological, physical, social, cognitive, and emotional perspectives. Despite such research on gender differences in the development of externalising and internalising problems, more research is needed to determine the exact role that gender plays in the manifestation of anxiety [44].

4.3. Anxiety Development, SES, and Wellbeing Interventions

The results of the final moderation analysis suggest that wellbeing interventions were only effective at reducing the anxiety levels of children with higher SES, meaning that children with lower levels of SES in general did not obtain the same benefits from wellbeing interventions for reducing their anxiety across time. When looking through the lens of Bronfenbrenner’s PPCT theory, it is indisputable that to create and implement culturally sensitive wellbeing interventions, the cultural and ecological influences on children’s development must be understood and considered [45]. Even though wellbeing interventions span several ecosystems and can result in improvements across a variety of developmental areas [46], the value, meaning, and efficacy of wellbeing interventions for diverse populations have recently been called into question. This is because most wellbeing interventions are not designed with diverse populations in mind and are therefore potentially less accessible for children who are not White and middle class [47]. This may help to explain why the wellbeing interventions in the current study were not effective at reducing the anxiety levels of children with lower SES. Children’s engagement in school-based wellbeing interventions is essential if we are to see a positive impact of these interventions because children will not experience positive effects if they are disengaged from the lesson material. Culturally responsive practices in social and emotional learning pedagogy, which involves using children’s lived experiences and frames of reference as well as incorporating their perspectives, interests, and needs, should help to increase engagement among diverse groups of students [48]. In particular, the accessibility of the language used in the interventions for low-SES children needs to be carefully considered. MacRuairc (2004) stipulates that middle-class linguistic capital is esteemed in Irish schools, which presents numerous challenges and barriers for children from other SES backgrounds who do not share this linguistic code. Irish teachers reported that making written and spoken language accessible to all students in their class presented a significant challenge, while they also reported difficulties with connecting new vocabulary and concepts in a variety of curriculum areas to the prior learning and cultural experiences of their students [49]. In this study, the language used in wellbeing interventions may not have been accessible to children with lower SES, and this may have diluted any positive outcomes arising from their engagement with the wellbeing interventions. This emphasises the need for profound, nurturing, and strategic pedagogical change that facilitates educators to work with children in a respectful manner that recognises the impact of ethnic influences, ensuring every child’s right to quality and equality in their educational experience is realised [40].

4.4. Limitations

There are some limitations to this study that need to be taken into consideration when interpreting results. First, data on children in the CSL study were collected using a self-report measure, which results in the risk of participant response bias [50]. Second, there are potential limitations due to using secondary data from the CSL study [51] rather than designing measures and data collection to exactly match the current study aims. Third, threats to validity also need to be considered when interpreting results, including the reliability of the anxiety and SES measures and the shifting nature of the CSL sample due to sample attrition and ongoing recruitment. Fourth, although the CSL sample is nationally representative to Ireland, the results of this study should not be generalised to children at different stages of development and to different cultural samples. Despite these limitations, several strengths also exist. Firstly, the current study used a large sample size, which helps to give adequate power to detect meaningful differences [52]. Furthermore, the CSL study used a random sampling frame to help define the target population. In addition, as the CSL study follows the same group of children over several years, it was possible to use data on anxiety from two different periods of time. This allowed for an exploration of the nature of developmental change [27] within this cohort of children.

5. Conclusions

This study investigated the role of individual- and school-level SES, gender, and wellbeing interventions on the development of anxiety in middle childhood in an Irish nationally representative sample of children in primary schools. The analysis revealed that anxiety increased more for girls and for children with higher individual family SES. However, attending a low-income school had no impact on the development of children’s anxiety when accounting for the impact of individual family SES. This indicates that the stressors provoking anxiety development in middle childhood are more salient in the family context than in that of schools. Furthermore, children with higher SES experienced less growth in anxiety when they were exposed to a wellbeing intervention in school. This finding is important, as it helps to contribute towards an understanding of the factors that impact the development of anxiety, which may contribute towards the evidence needed to expand existing theories such as the developmental psychopathology framework. The finding also helps to identify risk factors for the development of anxiety, which is important to consider when creating policies and initiatives to promote the healthy development of children. Finally, the results help wellbeing intervention designers and the people responsible for delivering lesson content to consider which factors might moderate the impact of wellbeing interventions on anxiety. This knowledge can be used to strengthen both design and implementation, resulting in better outcomes for children.