Introduction

Few would disagree that having happy and satisfied children is a worthy goal to pursue in its own right. Well-being, often conceptualised as being satisfied with one’s life, has been linked to a number of positive downstream social outcomes for children, such as high levels of self-worth, strong immune systems, original thinking, and higher academic attainment (Cárdenas et al., 2022; Kansky et al., 2016; Lyubomirsky et al., 2005). These studies demonstrate the importance of promoting positive well-being in children, with evidence suggesting that for some health outcomes, increasing positive aspects of well-being may be more impactful than minimising negative aspects (Holder, 2012). This observation holds particular relevance for children growing up in socioeconomically disadvantaged communities, who are at an elevated risk of experiencing adverse outcomes (Low et al., 2021).

A greater understanding of the determinants of well-being has the potential to influence education policy and teaching practice through informing school-based intervention efforts that promote the well-being of those children most likely to encounter adverse life events. When examining these determinants, it is important to adopt an integrated approach that considers how both social contextual and individual factors might work together to predict well-being. Conceptual frameworks (e.g., Connell & Wellborn, 1991; Ryan & Deci, 2009) and empirical studies (Holfve-Sabel, 2014; Lin et al., 2022; Oberle, 2018; Povedano-Diaz et al., 2019) underscore the significant connections between social interactions and relationships in the classroom and mental health and well-being outcomes in children. However, we still know little about the processes or mechanisms that might account for these links. While multiple pathways are likely, we propose that emotion regulation is a particularly promising mechanism worth considering (Jacobs & Gross, 2014). Given that emotions are often generated and regulated during social interactions (Gross et al., 2006), classroom contextual factors are likely to exert an influence on children’s emotion regulation. Consequently, being able to flexibly employ a variety of emotion regulation strategies to effectively navigate classroom challenges is expected to result in changes to positive and negative affect, as well as shape children’s perceptions and evaluations of their satisfaction with classroom experiences, the wider school environment, and beyond. The present study aimed to develop a better understanding of the key determinants of child well-being by examining how supportive teacher behaviours and children’s emotion regulation work together to predict well-being in children living in socioeconomically disadvantaged communities.

Children’s Well-Being

The well-being literature spans a range of disciplines and is riddled with different terminologies, definitions, and theoretical models (MacLeod, 2015). In the current study, we use Diener’s (1984) conceptualisation of subjective well-being, which includes an affective component (positive and negative affect) and a cognitive component (life satisfaction). We chose this global approach to assessing life satisfaction as we wanted the participant responses to be based on what they themselves consider to be most important to their lives, rather than using a well-being model that lists factors based on what participants from different cultural contexts consider to be important. Additionally, as our study includes a focus on school and classroom processes, we included a domain-specific measure of school satisfaction.

Recent international data (OECD, 2023) suggest that only 34% of 15-years-old from OECD countries report high satisfaction with their lives. Of equal concern is the recent downward trend that has been observed in both children and young people. In the UK, the Children’s Society has been measuring youth (ages 10–17 years) well-being since 2005 and have found that children have become increasingly unhappy. In their most recent survey, 11% of children reported low global life satisfaction (The Children’s Society, 2022). The present study was carried out in New Zealand, where data indicates that children experience some of the lowest levels of health and life satisfaction among OECD countries, ranking 35th out of 38 countries (Gromada et al., 2020).

There is also evidence to suggest significant within-country differences in well-being, with disadvantaged students being more likely to report low levels of life satisfaction (OECD, 2021; The Children’s Society, 2022). In particular, children and young people living in disadvantaged communities face a wide range of stressors and are at greater risk for negative outcomes such as substance abuse, suicide, and long-term problems of mental health (Buckner et al., 2009; Peverill et al., 2021). Data from international studies have shown that high levels of well-being in children and young people may act as a buffer against many of the negative effects of stress and challenging circumstances (Lyubomirsky et al., 2005; Suldo & Huebner, 2004). This further highlights the importance of examining the determinants of children’s well-being, particularly in at-risk or disadvantaged populations.

Determinants of Children’s Well-Being

Supportive Teacher Behaviours

A large body of literature has focused on the importance of teacher–pupil interactions for children’s cognitive development (De Pol et al., 2010; Vandenbroucke et al., 2018). Far less attention has been paid to how these interactions influence other school-related outcomes, such as school satisfaction or more global measures of children’s well-being. One framework that examines teacher–pupil interactions and includes a focus on well-being is Self-Determination Theory (SDT; Ryan & Deci, 2009). SDT posits that in order to experience an ongoing sense of well-being, three basic psychological needs must be satisfied: the need for autonomy, competence, and relatedness. Autonomy refers to being self-regulating of one’s own actions, competence refers to understanding how to achieve certain outcomes, and relatedness involves developing connections with others.

Connell and Wellborn (1991) extended SDT to include dimensions of teacher behaviour that foster the fulfilment of these three basic psychological needs. In their model, children’s need for autonomy is promoted when interactions with the teacher are autonomy supportive. Autonomy support in this context refers to the amount of freedom a child is given to determine their own behaviour. Teacher behaviours might include being responsive (e.g., listening to pupils), providing pupils with choice, and making connections between school activities and children’s interests. The provision of structure in the classroom is identified as the contextual factor that nurtures the child’s need for competence. This includes clear expectations, consistent consequences, and adequate help ensuring children understand their schoolwork. Finally, the teacher behaviour dimension that nurtures the need for relatedness is involvement. Involvement refers to the quality of the interpersonal relationship between the teacher and the child, and also the teacher’s willingness to dedicate time and resources to the child.

Empirical research examining these different approaches to supporting a child in the classroom reveals a large body of evidence indicating that autonomously motivated students excel in educational settings, and that students benefit from teacher behaviours that support their autonomy (Guay, 2022; Reeve, 2016). Similarly, several lines of evidence suggest that clear expectations and consistency in the classroom lead to positive school outcomes (Franklin & Harrington, 2019; Hattie, 2009; Skinner & Belmont, 1993). Oberle (2018) found that consistent teacher support was a significant predictor of both life satisfaction, and (to a lesser extent) positive affect. The importance of strong interpersonal relationships is also substantiated by a number of studies, underscoring the significance of pupils’ relationships with their teachers (Pianta et al., 2003). These effects seem to be particularly powerful for at-risk pupils as they appear to buffer the negative effects associated with disadvantage (Luthar et al., 2000).

Emotion Regulation

One individual-level factor that has been linked with well-being, both theoretically and empirically is emotion regulation. Emotion regulation refers to individuals’ attempts to increase, maintain, or decrease emotions (Gross, 1998). Gross’ Process Model of Emotion Regulation (1998) posits that different types of emotion regulation strategies alter emotion trajectories in different ways. For example, cognitive reappraisal involves modifying one’s appraisal of an emotion-eliciting situation. This strategy has typically been viewed as an adaptive strategy as it influences an individual’s subjective emotional experience—what some describe as their conscious emotional feelings (LeDoux & Hofmann, 2018). A number of studies examining emotion regulation in childhood and adolescence have found that in terms of reducing or down-regulating negative emotion, cognitive reappraisal strategies lead to decreased levels of negative emotional experience (Schaefer et al., 2017; Somerville & Whitebread, 2019; Willner et al., 2022) and decreased activation in emotion generative areas of the brain, such as the amygdala (Pitskel et al., 2011). As emotion is considered a key element of well-being, frequent use of this strategy is thought to have cumulative benefits for well-being. Conversely, an emotion regulation strategy such as suppression only targets an individual’s behavioural expression of emotion rather than their subjective experience of emotion (as reappraisal does). This strategy is often viewed as a maladaptive strategy with empirical studies linking frequent use of suppression in children and adolescents with experiencing more negative emotion, increased risk for anxiety and depressive symptoms, and lower levels of social well-being (Chervonsky & Hunt, 2019; Compas et al., 2017; Gross & Cassidy, 2019; Schaefer et al., 2017).

Emotion regulation strategies develop from an early age. During the pre-school years, children are already able to use strategies such as self-distraction to regulate their negative emotions (Blankson et al., 2017). As they develop, advances in cognitive capacities allow them to employ more sophisticated cognitive emotion regulation strategies (Harrington et al., 2020). Much of the literature on emotion regulation in childhood has focused on the role of parents as being primary agents in socialising children’s emotion regulation through supportive interaction (See Spinrad & Eisenberg, 2024, for a recent review). This has often been conceptualised through a Vygotskian framework (1978), moving from interpersonal (or co-) regulation to intrapersonal (or self-) regulation of emotion.

To a lesser extent, this framework has been applied to educational contexts and the role of the teacher in facilitating the development of young children’s emotion regulation (Denham et al., 2012). Many of the supportive, coaching, and modelling behaviours identified in the home context have been found to be equally important in classroom context; this is particularly the case for early years settings. For example, classroom observational studies have found that pre-school teachers frequently engage in a range of strategies to co-regulate children’s emotion. These include problem solving strategies that help to modify or change an emotionally challenging situation, supporting children in the interpretation of a situation, and helping them to modify their behavioural or emotional response (Kurki et al., 2016).

Far less is known about these co-regulatory emotion processes in primary school settings (Kostøl & Cameron, 2022). Some argue this may be due to the increased focus on academic outcomes during middle childhood (Jacobs & Gross, 2014). It could also be related to the different methodologies used to examine emotion regulation across the different age groups. Observational methods are more frequently used with pre-school aged children; an approach that has the benefit of capturing real-time social interactions and contextually rich data relating to the emotion-eliciting situations. However, a key limitation of this approach is that you cannot assess a child’s internal thoughts or subjective experience relating to their emotions. (Adrian et al., 2011). A common approach to capturing this type of data, is through self-report methodologies. While pre-school children may not be able to reliably report on their affective internal states, primary-aged children are likely to have developed a more sophisticated understanding of their internal thoughts and subjective affective experiences and are able to describe the strategies they use to regulate their emotions.

Of the studies that do use self-report methodologies with primary-aged children, the large majority focus on how emotion regulation relates to negative outcomes, such as internalising/externalising problem behaviours (e.g., Schaefer et al., 2017), rather than positive indicators of well-being. This is an important difference, as evidence suggests the correlates for mental illness and positive indicators of well-being in this age group are largely distinct (Patalay & Fitzsimons, 2016). In a recent systematic review of emotion regulation in primary children, Schlesier et al. (2019) identified just one study that included a specific focus on well-being; however, the primary well-being measures in this study were described by the authors as “internalizing and externalizing symptoms and behaviors”, so it did not focus on positive indicators of well-being (Crescentini et al., 2016, p. 3). The current study aims to address this gap in the literature with an explicit focus on how emotion regulation strategies in primary school-aged children relate to a number of positive indicators of well-being, such as life satisfaction, school satisfaction, and positive affect; in addition to examining links between emotion regulation and negative affect.

Socioeconomic Factors

There is research evidence to suggest that socioeconomic factors, such as income, education, and occupation, play a role in determining individual levels of well-being. Greater socioeconomic disadvantage has been linked with lower life satisfaction in children (Viñas et al., 2019), lower levels of self-worth (Evans & English, 2002; McLeod & Owens, 2004) and less satisfaction with school (Loft & Waldfogel, 2021; The Children’s Society, 2022). In New Zealand, where the present study took place, the Office of the Children’s Commissioner and Oranga Tamariki—the Ministry for Children (2019) used questionnaires, interviews and focus groups to capture the views of more than 6000 New Zealand children to inform the government’s well-being strategy. When asked what was needed for children to have a good life, one of the most frequent responses was that families need enough money for basics like food, clothing, and housing. Child poverty was also the area identified by the children as being in most need of urgent action from the government.

There is also evidence indicating connections between socioeconomic factors and emotion regulation. These studies have consistently found that children from lower socioeconomic backgrounds perform less well on emotion regulation measures (See Razman & Mohd Hoesni, 2023 for a review). These findings suggest that socioeconomic factors are likely to play a role in shaping children’s emotional experiences and how they evaluate different aspects of their lives. Consequently, it is important to include these factors in models exploring children’s well-being to gain a comprehensive understanding of potential predictors and how they might work together.

The Present Study

The present study examined the extent to which supportive teacher behaviours and emotion regulation predicted primary school pupils’ well-being. The schools sampled for this study were selected for being in socioeconomically disadvantaged areas where there is an overrepresentation of ethnic minority groups (Loring et al., 2022; Maré et al., 2001). We argue that it is especially important to understand how teacher support might be promotive of good wellbeing in these areas where well-being is lower on average (Sutcliffe et al., 2023). Indeed, teacher support might be even more important for students in the most socioeconomically disadvantaged areas.

This study sought to address the aforementioned knowledge gaps by using self-report measures to examine the relations among perceived teacher support, emotion regulation, and indicators of well-being in primary school pupils. Drawing on both theory and empirical research indicating links between teacher behaviours and positive pupil outcomes (Connell & Wellborn, 1991; Oberle, 2018; Pianta et al., 2003; Ryan & Deci, 2009), we hypothesised that perceived teacher support would predict pupil well-being (Hypothesis 1a). As supportive teacher behaviours, including co-regulation approaches, have been shown to facilitate the development of children’s emotion regulation (Kurki et al., 2016), we also expected teacher support to predict emotion regulation (Hypothesis 2a).

Moreover, we predicted that teacher support would be a stronger predictor of both well-being (Hypothesis 1b) and emotion regulation (Hypothesis 2b) for those children attending schools situated in the most socioeconomically disadvantaged communities. This prediction is consistent with the compensatory hypothesis (McClelland et al., 2017), which suggests that children who lack access to optimal resources, such as those living in socioeconomically disadvantaged communities, have more to gain from additional teacher support than those who are already well supported.

Despite close conceptual links between emotion regulation and subjective well-being, very few empirical studies have examined this link in children. However, evidence indicates that, consistent with adult samples, adaptive strategies such as cognitive reappraisal are effective in reducing negative affect in middle childhood (Willner et al., 2022), and strategies often described as maladaptive such as suppression have been positively associated with anxiety and depressive symptoms, and negatively associated with social well-being (family and friends; Chervonsky & Hunt, 2019). We thus expected emotion regulation to predict subjective well-being in our sample (Hypothesis 3). Moreover, extending past research linking classroom contextual factors and well-being (Oberle, 2018), we included children’s emotion regulation in our models as a potential candidate mechanism that might account for these connections, and hypothesised that teacher support would indirectly predict well-being, through emotion regulation (Hypothesis 4).

Methods

Participants

The study involved 508 children aged between 8 and 12 years (M = 9.91, SD = 0.63), of whom 243 were male (age: M = 9.95, SD = 0.64) and 265 were female (age: M = 9.87, SD = 0.62). Ethnicity was self-reported, with 54% of participants identifying as Pasifika, 11% as Māori, 5% as Asian, and 29% as mixed ethnicity. The participants were enrolled in Year 5/6 composite classrooms situated in low-socioeconomic urban communities in Auckland, New Zealand. The Ministry of Education (2023) calculates a socioeconomic decile for every school, based on national census data. Decile 10 schools draw their students from high-socioeconomic communities and Decile 1 schools from low-socioeconomic areas. The schools in the present study were all classified as Decile 1 (see Table 1). Within Decile 1, schools have additional subcategories denoted as 1A, 1B, or 1C. These subcategories further differentiate the level of disadvantage within the Decile 1 range, with 1A indicating the highest level of disadvantage and 1C indicating a relatively lower level of disadvantage within the Decile 1 grouping.

Table 1 Descriptive statistics of continuous and categorical study variables in the sample
Full size table

Procedure

Invitation emails were sent to all 51 Decile 1 schools in the Auckland education region, and we recruited 8 schools (with 31 classrooms) to the study. Information sheets and consent forms were sent to all parents/carers of the children in the selected classrooms prior to data collection. Child-friendly information sheets and consent forms were also given to those children who had parental permission to participate in the study. In the majority of classrooms (n = 31), all children participated. Of the 31 classrooms, 6 had fewer than 10 participants due to mixed year classrooms or non-response.

The information was read to the participants by the first author. Participants were given the opportunity to ask questions, decline to participate, or withdraw from the study at any time. No incentives were given to participate. The paper and pencil questionnaires were completed during morning class time in one sitting with short breaks between each scale (completion time was approximately 30 min). Participants were asked to raise their hand at any time if they did not understand the instructions or questionnaire items. Following data collection, data were transferred manually from the paper questionnaires to an electronic database for analysis.

Measures

Demographic Characteristics

Participants were asked to self-report their age, school year, gender, and ethnicity. School SES was defined according to the level of disadvantage of the school according to the government’s classification of schools. Within the most disadvantaged decile, there are three subcategories denoted as 1A (most disadvantaged), 1B and 1C (least disadvantaged). SES scores for each school are based on a set of factors including household income, occupation, household crowding, educational qualifications, and income support. Our disadvantaged schools were categorised as being High (least disadvantaged), Medium and Low (most disadvantaged).

Well-Being

Four wellbeing facets were examined in this study: Positive Affect, Negative Affect, Life Satisfaction and School Satisfaction. Positive and Negative Affect Schedule for Children (Laurent et al., 1999) was used to capture the frequency with which specific positive (e.g., joyful) and negative (e.g., scared) emotions were experienced. Participants were asked to rate how much they had experienced each emotion in the past few weeks in school using a 5-point scale from 1 (very slightly) to 5 (extremely). To facilitate understanding, a glossary was provided which included brief definitions of each emotion word. This is a widely used measure, has good convergent and discriminant validity with existing mental health self-report instruments, and high internal consistency reliabilities for the current sample for both the positive (12 items, α = 0.86) and negative (15 items, α = 0.85) scales.

Huebner’s (1991) 7-item Students’ Life Satisfaction Scale was administered to capture a global evaluation of participants’ satisfaction with their lives (e.g., ‘my life is going well’). Responses were rated on a 4-point scale from 1 (never) to 4 (almost always). The scale correlates predictably with existing measures and internal consistency was 0.71 for the current sample.

The School subscale of the Multidimensional Students’ Life Satisfaction Scale (Huebner, 2001) was administered to capture school satisfaction. This is an 8-item scale that specifically targets participants’ satisfaction with their school experience (e.g., ‘I like being in school’). As with the life satisfaction scale, responses were rated on a 4-point scale from 1 (never) to 4 (almost always). Internal consistency was 0.81 for the current sample.

Emotion Regulation

The Emotional Control subscale of the Emotion Regulation Index for Children and Adolescents (ERICA; MacDermott et al., 2010) was administered to assess emotion regulation. This is a 7-item scale that measures the ability to regulate negative emotions and restrain from exhibiting inappropriate behaviours (e.g., ‘I get angry when adults tell me what I can and cannot do’). Responses were rated on a 5-point scale from 1 (strongly disagree) to 5 (strongly agree). Higher scores on this scale indicate a greater ability to regulate negative emotions. Internal consistency was 0.71 for the current sample.

Teacher Support

To assess teacher support, the student-report version of the Teacher as Social Context (Skinner & Belmont, 1993) was used. The TASC includes three subscales: autonomy support which reflects teacher behaviours that promote student autonomy (e.g., ‘my teacher gives me a lot of choices about how I do my schoolwork’); structure which reflects teacher consistency and predictability of response (e.g., ‘every time I do something wrong, my teacher acts differently’), and involvement which reflects teacher behaviours that help children feel related to their teachers (e.g., ‘my teacher spends time with me’). The measure consists of 24 items (8 items for each subscale), each with a 4-point response scale of not at all true, not very true, sort of true, and very true. Internal consistencies for the subscales using the current sample were low (α = 0.40 to 0.59). We used the full scale TASC score in our analyses (α = 0.74).

Data Analytic Strategy

First, descriptive data were generated for the main variables and covariates in our study. Second, we conducted correlational analyses between our main variables. Finally, we examined the relationships between teacher support, emotional regulation and well-being by fitting multilevel linear models whereby pupil outcomes were specified to vary randomly by classroom (Snijders & Bosker, 1999). Our 508 pupils were nested within 31 classrooms. The number of children in each classroom ranged from 3 to 24, with the average number being 16.4. Pupils in the same classroom are likely to share characteristics due to the shared classroom environment as well as classroom (and school) selection bias.Footnote 1 Children’s outcomes (including well-being and emotion regulation) may therefore be correlated within classrooms. If this is ignored, the standard errors of the coefficient estimates will be underestimated, increasing the probability of finding a statistically significant effect where one does not exist. Initially, we conducted a variance components (or ‘null’) model where the variation in well-being and emotion regulation was modelled with a random intercept term for classroom (Level 2) and a random error term for pupil (Level 1). For well-being variables, the intraclass correlations (ICC) were 0.040 (Life satisfaction), 0.092 (School satisfaction), 0.046 (Positive affect) and 0.058 (Negative affect) suggesting that 4.0–9.2% of the variance in well-being scores were attributable to classrooms. For emotion regulation, the ICC was 0.012 suggesting that only 1.2% of the variance in emotion regulation was attributable to classrooms. We used the xtmixed command in Stata 17.0 for all multilevel models.

We then introduced the main variables of interest to the fixed part of the models. We specified a series of these models to address Hypotheses 1–4 (Models 1–4 were aligned with Hypotheses 1–4 respectively). Model 1a estimated the association between teacher support and pupil well-being controlling for covariates. Model 1b included the interaction effect of teacher support and school SES. Model 2a captured the association between teacher support and pupil emotion regulation adjusting for covariates and Model 2b included the interaction effect of teacher support and school SES. Models 1b and 2b therefore tested whether school SES moderated the link between teacher support and child well-being and teacher support and emotional regulation, respectively. Model 3 estimated the association between pupil emotion regulation and pupil well-being (negative affect, positive affect, life satisfaction, and school satisfaction), adjusting for covariates (i.e., school SES as well as pupil gender, age, and ethnicity). Finally, where we identified a significant coefficient for teacher support in Model 1a, we tested the mediating role of emotional regulation using ml_mediation in Stata (Model 4), which computes the direct and indirect effects in a multilevel model based on the xtmixed, reml command, an approach adapted from Krull and MacKinnon (2001). A total of 18 models were run (specified in Table 2).

Table 2 Multilevel model specification
Full size table

Results

Descriptive Statistics and Correlational Analyses

Descriptive statistics for the pupils in the sample are presented in Table 1. When compared with samples from the validation studies of the measures, the pupils in the current sample had on average higher levels of school satisfaction, life satisfaction, positive affect, negative affect, and emotion regulation, and lower levels of teacher support (Huebner & Gilman, 2007; Iglesias-García et al., 2019; Laurent et al., 1999; MacDermott et al., 2010).

All well-being domains, emotion regulation and teacher support were significantly weakly-to-moderately correlated in expected directions (Table 3). Correlation absolute effect sizes between emotion regulation and well-being variables ranged from 0.21 to 0.40; the highest correlation was with negative affect (−0 .40). Correlation absolute effect sizes between teacher support and well-being ranged 0.27–0.45 with the strongest association being with school satisfaction (0.45). Teacher support and emotion regulation were moderately positively correlated (0.31).

Table 3 Correlations between the main variables
Full size table

Multilevel Regression Modelling

Models 1a and 1b: Estimating the Association Between Teacher Support and Well-Being

First, we examined the role of teacher support in relation to pupil well-being, adjusting for school SES and pupil gender, age, and ethnic background (Model 1a, Table 4). Teacher support was significantly related to higher positive affect (b = 0.710, p < 0.001), lower negative affect (b =  − 0.562, p < 0.001), higher life satisfaction (b = 0.602, p < 0.001) and higher school satisfaction (b = 0.682, p < 0.001). In these models, attending a higher SES compared with a lower SES school was significantly associated with all domains of well-being. We then tested the interactions between teacher support and school SES (Model 1b) in relation to all four well-being outcomes and none of the interaction terms were significant (results available on request). With regard to the random effects, the between-classroom variance ranged from 0.003 (negative affect) to 0.062 (school satisfaction) in the fully adjusted model.

Table 4 Fixed effects, variance, and covariance estimates of Model 1a (n = 508)
Full size table

Models 2a and 2b: Estimating the Association Between Perceived Teacher Support and Emotion Regulation

Next, we examined the relationship between teacher support and pupil emotion regulation adjusting for school SES, and pupil age, gender, and ethnic background (Model 2a, Table 5). Teacher support was significantly predictive of emotion regulation (b = 0.743, p < 0.001) but school SES was not. The variance in emotion regulation between classrooms was nearly fully explained by the child and school covariates. Additionally, we found a significant interaction (Model 3b) between teacher support and school SES (Model 2b; middle SES: b = 0.743, p < 0.001; high SES: b = 0.743, p < 0.001) in relation to pupil emotion regulation. To illustrate the interaction between teacher support and school SES, we plotted the predicted values of emotion regulation for pupils with high (75th percentile) and low (25th percentile) teacher support by school SES (low, middle, and high; See Fig. 1). Predicted values were plotted for the averages of continuous covariates and the reference groups for categorical covariates. Figure 1 suggests that teacher support appears to be especially beneficial for pupils in middle and high SES schools compared with those in low SES schools. The gap is wider between high and low teacher support in those schools compared to that represented in the figure for low SES schools.

Table 5 Fixed effects, variance, and covariance estimates for emotion regulation (Model 2a and 2b; n = 508)
Full size table
Fig. 1
figure 1

Predicted emotion regulation for pupils by teacher support and school SES. Note: Predictions are plotted for the reference group for each categorical variable and at the mean of each continuous variable. Low and high teacher support defined at the 25th and 75th percentile

Full size image

Model 3: Estimating the Association Between Emotion Regulation and Well-Being

Emotion regulation was significantly associated with positive affect (b = 0.204, p < 0.001), negative affect (b =  − 0.335, p < 0.001), life satisfaction (b = 0.142, p < 0.001) and school satisfaction (b = 0.143, p < 0.001; Model 3, Table 6), adjusting for school SES and pupil gender, age, and ethnic background.

Table 6 Fixed effects, variance, and covariance estimates of Model 3 (n = 508)
Full size table

Model 4: Testing the Indirect Effect of Teacher Support on Well-Being Via Emotion Regulation

Finally, we tested whether teacher support predicted the well-being outcomes indirectly, through emotion regulation. Table 7 contains the coefficients of indirect, direct, and total effects for well-being outcomes. Both the total effects and the indirect effects were significant for all outcomes.

Table 7 Direct, indirect, and total effects
Full size table

Discussion

The importance of children’s well-being is well-established, with studies consistently linking higher levels of well-being with a range of positive downstream consequences (Cárdenas et al., 2022; Kansky et al., 2016). Evidence suggests well-being is particularly important for children and young people living in disadvantaged communities who face a wide range of stressors and are at greater risk for negative outcomes (Low et al., 2021). While past research indicates that supportive relationships and interactions in the classroom predict well-being in children, we still know little about the processes or mechanisms involved in these associations. Given that emotion is often generated and regulated during social interactions, and that well-being includes a core affective component, we argued that emotion regulation was a likely candidate mechanism to explain these connections and included it in our models. To the best of our knowledge, this study was the first to examine how classroom contextual factors and emotion regulation work together to predict children’s well-being.

Consistent with prominent conceptual models (Connell & Wellborn, 1991; Ryan & Deci, 2009), our findings support our Hypothesis 1 prediction that supportive teacher behaviours would be associated concurrently with both cognitive (i.e., school and life satisfaction) and affective (positive and negative affect) well-being outcomes. Teacher support was found to be a stronger concurrent predictor of school satisfaction than the more global measure of life satisfaction. This was not unexpected, given that life satisfaction is likely to be influenced by additional predictors beyond the school environment. Teacher support also significantly predicted positive and negative affect. In contrast to other studies (e.g., Oberle, 2018), teacher support was a stronger predictor of positive affect than the life and school satisfaction variables. It is important to note that these effect size differences are not large. It is possible that this difference in findings is due to the affective component of well-being being less stable than the cognitive components; for example, studies have shown that mood effects of the respondent have a greater impact on affective scores (Eid & Diener, 2004). It may also be related to differences in the cultural contexts of the studies. Large-scale multi-country studies have shown there to be considerable cross-cultural variation in predictors of these components of well-being (Oishi et al., 2009).

Establishing teacher support as a key predictor of well-being is important, however, to extend this literature we argue a better understanding of the mechanisms underpinning these pathways is needed. Our hypothesised mechanism of emotion regulation was initially supported by our finding that participants who perceived their teacher to be supportive indicated that they were better at regulating their emotions (Hypothesis 2). This is consistent with Denham’s conceptualisation of teachers as socialisers of children’s emotions (e.g., Kurki et al., 2016), however this conceptual framework, and the empirical studies looking at co-regulation in classrooms has had a strong focus on early-years settings, with few studies looking at the primary years. Developmentally, it would be expected that this older group would require less direct intervention from the teacher during emotionally challenging situations. Nonetheless, our findings indicate that teacher support remains important for children’s emotion regulation in this age group; though, a different form of scaffolding is likely to be beneficial for these children, such as using metacognitive prompting to encourage reflection on emotion generation and situation-appropriate regulation strategies (Silkenbeumer et al., 2018).

When examining the moderating effect of SES (part of Hypotheses 1 and 2), this concurrent connection between teacher support and emotion regulation was stronger for those children attending schools situated in the middle and higher SES communities of our sample, when compared to the children attending schools in lower SES communities. This finding was in contrast to our prediction that participants attending lower decile schools would benefit more from the teacher support as they potentially face more risk factors and emotionally challenging situations in their daily lives (Reiss, 2013). However, it may in fact be that factors outside the classroom become more impactful in lower SES communities, due to the increased risk factors, reducing the effect of the relationships and interactions that take place inside the classroom or school. Attending a higher SES school also appeared to be linked to all domains of well-being. Though, SES did not remain a significant predictor when emotion regulation was included in the model. This suggests that emotion regulation and SES variables share common variance and that the effects of SES on well-being were subsumed by the inclusion of the emotion regulation variable.

Moreover, we found emotion regulation to be a significant predictor concurrently of all four well-being outcome variables, supporting Hypothesis 3. As with the teacher support literature, these connections have been understudied in this age group. Research examining developmental changes in emotion regulation highlights the importance of considering these associations in primary-aged children, and not inferring associations from adult or early years studies. Cognitive reappraisal, for example, has been observed in children as young as 3 years, however, this is typically in the context of adult scaffolding (Willner et al., 2022). Evidence of this strategy in older adolescents and adults indicates they use more sophisticated cognitive emotion regulation strategies than primary-aged children, and at a greater frequency (Riediger & Klipker 2014).

In terms of how emotion regulation related to the different well-being outcomes, we found it to be the strongest predictor of negative affect which was unsurprising given that negative emotions in the classroom are typically more in need of regulation than positive emotions, as negative emotional responses often have longer lasting and more intense consequences (Baumeister et al., 2001). Additionally, the emotion regulation measure we used (ERICA; MacDermott et al., 2010) includes more items related to the regulation of negative affect.

Finally, our mediation analysis gave further evidence for the indirect effect of teacher support on well-being, via emotion regulation. Supporting our prediction (Hypothesis 4), this indirect effect was significant for all four well-being outcomes. However, it is important to note that due to the cross-sectional design of our study, temporal order cannot be assumed here.

Limitations and Future Directions

The limitations of the current study should be noted when interpreting the findings. Firstly, the cross-sectional nature of the study restricts our ability to establish causal relationships or determine the direction of the observed effects. Secondly, both independent and dependent variables are based on child-report and therefore associations may be inflated due to shared method variance. Thirdly, we only collected data from Decile 1 schools, who draw their students from low-socioeconomic communities. Extending the sample across Deciles 1 to 10 would allow us to better understand how SES might work together with the other variables to predict well-being. Fourthly, given data collection constraints, we were unable to adjust for family SES or other adversities as potential confounders of the observed relationships.

Despite these limitations, the study can be seen as a first step in examining how classroom contextual factors and emotion regulation work together to predict well-being. To further expand our understanding of the underlying mechanisms through which teacher support influences children's well-being, longitudinal designs should be carried out to test the potential mediating role of emotion regulation. Additionally, exploring other potential mediating or moderating factors within the classroom, such as peer interactions and relations, could provide valuable insights and contribute to a deeper understanding of these pathways.

Finally, it is important to understand more about the specific processes that connects teacher support and emotion regulation. Currently, the emotion regulation literature predominantly focuses on competencies and strategies, paying less attention to the factors that drive individuals to regulate their emotions (Tamir & Mauss, 2011). It is plausible that teacher support influences these latter factors, such as the educational or social goals pursued by children and young people (Tamir et al., 2020), or their beliefs regarding their capacity to regulate emotions (Somerville et al., 2023).

Conclusion

Taken as a whole, the data from this study indicate that teacher support does matter for both emotion regulation and well-being. These findings have important implications for teacher educators, school leaders, and teachers themselves. Our results underscore the importance of teacher support behaviours such as cultivating warm and positive relationships, providing structure and consistency, and fostering pupil autonomy. Both teacher behaviours and emotion regulation are amenable to intervention; thus, a better understanding of how they work together to predict well-being is likely to inform future intervention efforts to promote children’s well-being.