Social support and well-being of adolescent learners: a latent profile analysis

Abstract

This study focused on identifying adolescent learners’ social support profiles based on their self-reports regarding three sources of social support (parental, teacher, and close friend support) and associations of such profiles with their well-being. A biographical questionnaire, the Social Support Scale for Children, and the Mental Health Continuum – Short Form were administered to 770 Grades 10 and 11 learners from previously disadvantaged schools in South Africa. Five social support profiles were identified using latent profile analyses of parental, teacher, and close friend support: weakly supported, adult-supported, peer-supported, moderately supported, and integrated support. Social support profiles were associated with well-being. Flourishing, the most desirable well-being outcome, was associated with the integrated support profile (high parent, teacher, and close friend support). In contrast, the lowest level of well-being was observed in the weakly supported profile (low parent, teacher, and close friend support). The results confirmed that support from parents, teachers, and close friends was vital for adolescent learners’ well-being.

Keywords

Adolescents latent profile analysis mental health social support South Africa well-being

Introduction

Adolescence is a critical period for emotional and social development (Sawyer et al., 2018). In sub-Saharan Africa, many adolescents face pressures from economic hardship, violence, and social instability (Mhongera & Lombard, 2020). Moreover, adolescent learners are at risk of mental health problems (including depression, anxiety disorders, emotional and behavioural problems, post-traumatic stress disorder, and suicidal ideation), substance abuse, HIV/AIDS, and teenage pregnancy (Hendricks et al., 2019; Jörns-Presentati et al., 2021). A supportive environment helps them navigate this phase, during which they form their identities, build their self-esteem, and learn to manage complex social interactions (Keyes, 2024). Moreover, social support can offer a safe space for emotional growth, resilience-building, enhancing educational outcomes, dealing with socio-economic challenges, and promoting health and well-being (Hendricks et al., 2019).

Different sources and types of social support can help alleviate some of the stressors experienced by adolescents and promote well-being (Tennant et al., 2015). Research has shown that interpersonal relationships with parents, teachers, and close friends are valuable social support sources contributing to adolescents’ well-being and development (Camara et al., 2017). However, we found only one study (Ciarrochi et al., 2017) that applied a person-centred approach (latent profile analysis, LPA) to study combinations of social support sources. According to Harter (1985), classmate support can be added as a dimension of social support. This study contributes to the literature on the social support and well-being of Grade 10 and 11 learners in a non-Western, educated, industrialised, rich, and democratic (WEIRD) society (Henrich et al., 2010). The social environments in which adolescent learners negotiate their lives in South Africa are substantially different from those in a WEIRD society (e.g., in the study by Ciarrochi et al., 2017).

Social support

Regulatory effectiveness of support (RES) theory postulates that social support motivates adolescents to feel effective in different dimensions of their lives (Zee et al., 2020). Support helps adolescents understand and manage challenging situations. This, in turn, empowers them to develop behavioural and psychological responses such as motivation, utilising past best practices, and regulating their moods to navigate their needs. Therefore, social support affects adolescent learners’ mental health.

The four types of social support in an adolescent’s schooling life (parental, teacher, classmate, and close friend) may combine in multiple ways (Ciarrochi et al., 2017). Morin et al. (2011) and Morin and Wang (2016) highlight the necessity of a person-centred approach (in contradistinction to the traditional variable-centred approach) in understanding social support among adolescents. This perspective recognises the existence of distinct social support profiles within the population, such as variations in support from parents and teachers versus close friends and classmates. It emphasises that individuals may belong to subpopulations with qualitatively and quantitatively different configurations of social support, suggesting a more nuanced analysis of social support dynamics. Given that support from one source of social support is associated with support from other sources, it is conceivable for one profile to experience high levels of social support across all sources. At the same time, another may encounter uniformly low levels of support from all sources (Demaray et al., 2005). However, social support profiles may also be divergent, reflecting high support from some sources and low support from others (Jager, 2011). Divergence occurs when some adolescent learners find their social support needs more adequately fulfilled by parents (and teachers) than by close friends and classmates, or vice versa.

‘Parental social support’ benefits adolescents’ psychological well-being (PWB; Chentsova Dutton et al., 2020). Parental support entails the extent to which parents understand their children, care about their feelings, want to listen to their problems, treat them like they matter, like them as they are, and make them feel that what they do is important (Harter, 1985). Parents can support adolescents emotionally (through expressions of love, empathy, warmth, and concern), informally (through information, guidance, or advice), and by offering material assistance (Chentsova Dutton et al., 2020). Parent support acts as a protective measure against mental health risks such as depression, antisocial behaviour, and delinquency (Attar-Schwartz, 2015) and promotes positive learning outcomes (Hendricks et al., 2019).

Biological parents in South Africa are not necessarily the primary caregivers for adolescent learners. Grandparents take on responsibilities to complement parental care and, in some cases, take on the duties of the primary caregiver (Ben-Shlomo & Taubman-Ben-Ari, 2016; Hoffman, 2019). Research has shown that children raised by grandparents (compared with those raised by single mothers) have lower socioemotional and PWB levels, often leading to behavioural problems in the children (Pilkauskas & Dunifon, 2016).

‘Teacher support’ refers to the extent to which adolescent learners’ teachers help them when they are upset, assist them in doing their best, care about them, are perceived to be fair to them, and treat them with respect (Harter, 1985). Teacher support has three characteristics: positive affect, balanced power, and complete reciprocity (Ibrahim & Zaatari, 2020). Positive teacher support is associated with lower levels of conflict, greater engagement, increased school liking, and greater academic achievement among learners (Tennant et al., 2015).

‘Classmate support’ refers to the extent to which adolescent learners’ classmates like them the way they are, are friendly, listen to what they say, refrain from embarrassing them, and invite them to participate in group activities (Harter, 1985). Classmate support assists adolescent learners in fulfilling their needs, particularly those for relatedness and acceptance (Brown, 2004). Demir and Leyendecker (2018) reported that classmate support positively affected learners’ quality of life, school-related engagement, and self-concept.

‘Close friend support’ is different from the other three types of support. It pertains to whether the adolescent learner has a close friend who responds in a supportive way (Harter, 1985). Close friend support concerns whether adolescent learners have a close friend with whom they can share their problems, who shows understanding, with whom they can share things that bother them, who listen to what they say, and with whom they can spend time.

Not all studies have revealed positive correlations between social support networks and adolescent learners’ mental health. Rueger et al. (2010) found that perceived peer social support was associated with lower levels of depression. However, Windle’s (1992) research indicated that peer support was not associated with depression. Furthermore, Meadows’s (2007) study suggested that peer support could have adverse effects, depending on the nature of the peer group (e.g., delinquents versus nondelinquents).

Research has shown that gender influences adolescents’ perceptions of support from their networks. Compared with male adolescents, female adolescents perceive receiving more support from classmates and close friends. However, both genders perceive receiving similar support from parents and teachers (Malecki & Demaray, 2003). There is no gender difference in the perception of teacher support. However, females perceive receiving classmate support more strongly than males do. This could be due to norms that females should invest emotionally in relationships (Kendrick et al., 2012).

Mental health: emotional, psychological, and social well-being

Keyes (2005, 2024) introduced a dual-continuum model consisting of mental health and mental illness as two related, but distinct, dimensions. Individuals can experience high levels of mental health while also dealing with mental illness (Keyes et al., 2020; Magyar & Keyes, 2019). Conversely, they can have low mental health without necessarily being mentally ill (Keyes, 2024). A person’s emotional well-being (EWB), PWB, and social well-being (SWB) are at the core of mental health (Keyes, 2024; Schotanus-Dijkstra et al., 2016).

EWB refers to the presence or absence of positive feelings about life (Magyar & Keyes, 2019).

PWB entails adolescents’ perceptions of their functioning regarding autonomy, environmental mastery, personal growth, purpose in life, positive relations with others, and self-acceptance (Magyar & Keyes, 2019). SWB refers to fulfilling tasks in social structures and consists of five dimensions: social integration, social contribution, social coherence, social actualisation, and social acceptance (Magyar & Keyes, 2019). Flourishing is a state of optimal mental health characterised by high levels of EWB, PWB, and SWB (Schotanus-Dijkstra et al., 2016). Languishing, the opposite of flourishing, represents stagnation and emptiness rather than mental illness (Keyes, 2024).

Languishing is a critical risk factor for developing mental illness (Keyes, 2024; Keyes et al., 2020). Moreover, languishing presents a high risk, especially for adolescents (Keyes, 2024). Languishing adolescent learners are at risk for academic and behavioural problems at school (Antaramian et al., 2010). Research has shown that as levels of flourishing increase, the potential for mental illness decreases (Keyes et al., 2020). Nurturing adolescents’ flourishing by encouraging positive emotions, fostering a loving parent–adolescent relationship, and creating a supportive school environment (Keyes et al., 2020) is essential.

This study

Studies (Camara et al., 2017; Ciarrochi et al., 2017) have shown that social support leads to desirable adolescent well-being outcomes. According to Ciarrochi et al. (2017), the effects of social support on adolescents’ flourishing can be studied using variable- and person-centred approaches. Person-centred approaches (such as LPAs) assume that subgroups of samples (profiles) come from a heterogeneous population (Morin & Wang, 2016). Different sources of social support have compensatory or mutually reinforcing effects (Ng & Sorensen, 2008).

A literature review showed that one study (Ciarrochi et al., 2017) focused on the combined effects of different sources of social support on adolescent learners’ well-being. Ciarrochi et al. (2017) found that one-third of adolescent learners received no social support from any source. Some learners found social support from some sources, while others received social support from all sources. Furthermore, difficulties were associated with the socially isolated profile, while increasing the benefits was associated with moving to more sources of social support. Leveraging diverse social support sources, including parental guidance, teacher mentorship, and supportive peer interactions, is crucial for adolescent well-being. Parental involvement enhances emotional stability, while mentorship boosts academic performance and resilience. Empathetic friendships reduce isolation, promoting overall flourishing (Biglan, 2015; Ciarrochi et al., 2017).

This study adopted a person-centred approach to (a) identify adolescent support based on parent, teacher, classmate, and close friend support; and (b) examine the associations between social support profiles, demographic variables, and adolescent learners’ well-being.

Method

Participants

The sample comprised 770 adolescent learners from four secondary schools in previously disadvantaged areas. Stratified systematic sampling was used to select male and female learners from Grades 10 and 11 from each school. The learners’ ages spanned the adolescent development stages. The characteristics of the participants are reported in Table 1.

Table 1.

Characteristics of the participants (n = 770).

Item	Category	Frequency	Percentage
Gender	Male	302	39.2
	Female	367	47.7
	Did not disclose	101	13.1
Age in years	13–14	22	2.9
	15–16	390	50.6
	17–18	303	39.4
	19 and over	53	6.9
	Did not disclose	2	0.3
Grade	10	385	50.0
	11	385	50.0
Guardian	Grandparent(s)	63	8.2
	Both parents	281	36.5
	Father	38	4.9
	Mother	352	45.7
	Relative(s) and siblings	27	3.5
	Did not disclose	9	1.2

Measuring instruments

The Social Support Scale for Children (SSSC; Harter, 1985) was used to measure the support and positive regard that adolescent learners felt they received from people in their lives. The measure has 24 questions and four subscales: parents, teachers, classmates, and close friends. Each subscale investigates different contents. Each statement is coded from 1 (really true for me) or 2 (sort of true for me), when one of the options of the first part of the statement was chosen, and 3 (sort of true for me) or 4 (really true for me), in the event of the second statement being chosen. The scale has excellent content, criterion, and construct validity; internal consistency scores range from 0.72 to 0.88 (Harter, 1985).

The Mental Health Continuum – Short Form (MHC-SF; Keyes, 2009) was used to assess flourishing versus languishing among adolescent learners. The MHC-SF consists of three items measuring EWB, including questions such as ‘How often does the learner feel happy?’. Six-item scales measure the six dimensions of PWB. One of the questions posed in this subsection is ‘How often does the learner like his/her personality for the most part?’. Finally, there are five items that measure SWB. These include questions about how often the learner believes his or her life has meaning. The items are each rated from 1 to 6 on a Likert-type scale. The MHC-SF test–retest reliability is 0.68 over 3 months and 0.65 over 9 months (Lamers et al., 2010). Research by Keyes et al. (2008) in a South African context has shown that the original three-dimensional factor structure fits the data the best. The MHC-SF also shows excellent internal consistency and discriminant validity in adolescents (aged 12 to 18) in South Africa (Keyes, 2005; Lamers et al., 2010).

Research procedure and ethics clearance

As part of the research procedure and ethics considerations before the commencement of the study, permission was obtained from the Gauteng Department of Education (GDE) and the Department of Education of KwaZulu-Natal for research to be conducted in Gauteng West and Newlands, respectively. The Basic and Social Sciences Research Ethics Committee (BaSSREC) of North-West University granted ethical approval for the study (Reference: NWU-HS-2018-0125). Approval letters from the GDE to conduct the study were issued from 4 February 2019 (the first school quarter) to 30 September 2019 (the third school quarter).

The logistics of the study were discussed with key stakeholders (principals, gatekeepers, and participants) from the four secondary schools within the stipulated regions. All the data collectors were trained in data collection. Consent was obtained before the questionnaires were administered. Participants could withdraw from the research at any point.

Statistical analysis

The descriptive statistics were computed with SPSS 26.0 (IBM Corp, 2017). Full information maximum likelihood (FIML) was used to handle missing values in the dataset in Mplus 8.10 (Muthén & Muthén, 2009–2023). LPA was utilised to determine whether distinctive profiles relating to perceived social support could be found in the data (Wang & Wang, 2020). The competing models were compared based on their Bayesian information criterion (BIC), Akaike information criterion (AIC), and sample-size adjusted BIC (ABIC) values. Entropy was used to assess the level of profile verification in the LPAs (West et al., 2023). The Lo–Mendell–Rubin likelihood ratio test (LMR LR), the adjusted Lo-–Mendell–Rubin test (ALMR), and the bootstrapped likelihood ratio test (BLRT) were employed to determine the optimal number of profiles (Wang & Wang, 2020).

The measurement models of social support and well-being were tested using Mplus 8.10 (Muthén & Muthén, 1998–2023) by conducting exploratory factor analysis (EFA) and confirmatory factor analysis (CFA), respectively. The weighted least squares mean and variance estimator (WLSMV), appropriate for analysing categorical variables, was utilised to investigate the SSSC responses. The robust maximum likelihood estimator (MLR) was used to estimate the MHC-SF model. Model fit was assessed using chi-square statistics, the standardised root mean residual (SRMR), the root mean square error of approximation (RMSEA), the Tucker–Lewis index (TLI), and the comparative fit index (CFI). Alpha coefficients were computed to assess scale reliability. A cut-off score of 0.70 was used. To determine the mean of a distal continuous outcome across latent profiles, the automatic Bolck, Croon, and Hagenaars (BCH) method was employed (Asparouhov & Muthén, 2014).

Results

Testing measurement models of social support and mental health

Testing measurement models of social support

Before conducting exploratory factor analyses, we used one-way analysis of variance to test whether statistically significant differences existed in terms of parental support experienced by adolescent learners with different types of parents (including guardians if they did not have biological parents). No differences were found between adolescents with biological parents, a mother only, a father only, a grandmother/grandfather, or other family members as guardians (parents; F = 2.22, df = 4, p = .07).

The factor structure of the SSSC was explored using EFA in Mplus 8.10 (Muthén & Muthén, 2009–2023). The eigenvalues (> 1) showed that five factors could be extracted. However, a four-factor structure for the SSSC was interpretable. The following fit statistics were obtained for the four-factor model: χ² = 427.23 (df = 186, p < .001); RMSEA = 0.04 [0.04, 0.05], p = .998; CFI = 0.94; TLI = 0.91; SRMR = 0.04. The standardised loadings for the four types of social support varied as follows: from 0.38 (Item 21) to 0.63 (Item 17) for parental support, from –0.01 (Item 6) to 0.44 (Item 22) for classmate support, from 0.33 (Item 7) to 0.60 (Item 11) for teacher support, and from 0.30 (Item 20) to 0.78 (Item 8) for close friend support. Given the poor standardised regression coefficients for classmate support, it was decided to remove this subscale from the remaining analyses.

Testing the measurement invariance of the MHC-SF for Grades 10 and 11

Configural, metric, and scalar invariance were computed to determine the MHC-SF measurement invariance for learners in Grades 10 and 11. The measurement models for Grades 10 and 11 were specified: EWB consisted of three observed variables, PWB consisted of six observed variables, and SWB consisted of five observed variables. The three variables were allowed to correlate. Table 2 presents the results of the measurement invariance testing.

Table 2.

Measurement invariance for the mental health dimensions for learners in Grades 10 and 11.

Models compared	χ²	df	p
Metric against configural	14.40	11	.211
Scalar against configural	25.71	22	.265
Scalar against metric	11.07	11	.437

Table 2 shows evidence for configural, metric, and scalar invariance. Based on this finding, the data for Grades 10 and 11 were pooled for the remaining analyses.

Testing the mental health measurement model in the pooled sample

CFA was used to test the measurement model for mental health. The following fit statistics were obtained: AIC = 38,062.81; BIC = 38,271.90; χ² = 205.14 (df = 74, p < .001); RMSEA = 0.05 [0.04, 0.06], p = .653; CFI = 0.89; TLI = 0.87; SRMR = 0.05. Given the lack of fit of the measurement model to the data, the standardised regression coefficients and modification indices of the items of the MHC-SF were inspected. The standardised beta for item 4 (‘ . . . feel that you had something important to contribute to society’) was low (β = 0.19). The modification indices (MIs) showed a high error correlation between Items 4 and 5 (MI = 38.92). Therefore, the measurement model was respecified without Item 4, resulting in an acceptable fit of the measurement model of mental health: AIC = 35,125.89; BIC = 35,321.04; χ² = 142.44 (df = 73, p < .001); RMSEA = 0.04 [0.03, 0.05], p > .05; CFI = 0.93; TLI = 0.91; SRMR = 0.04. The standardised loadings for the three scales of the MHC-SF varied as follows: from 0.43 to 0.63 for EWB, 0.25 to 0.64 for PWB, and 0.15 to 0.65 for SWB.

Descriptive statistics, reliabilities, and correlations

Table 3 shows the descriptive statistics, alpha coefficients, and Pearson correlation coefficients between the variables.

Table 3.

Descriptive statistics, reliabilities, and Pearson correlations.

Item	Mean	SD	α	1	2	3	4	5	6
EWB	3.64	0.99	0.87	–	–	–	–	–	–
PWB	3.84	0.85	0.77	0.41**	–	–	–	–	–
SWB	2.45	1.18	0.58	0.35**	0.32**	–	–	–	–
Flourish	3.36	0.75	0.74	0.69**	0.80**	0.76**	–	–	–
PS	2.95	0.69	0.66	0.26**	0.18**	0.23**	0.28**	–	–
TS	2.84	0.62	0.62	0.18**	0.17**	0.23**	0.25**	0.36**	–
CS	3.07	0.65	0.65	0.10**	0.18**	0.09*	0.17**	0.21**	25**

EWB: emotional well-being; PWB: psychological well-being; SWB: social well-being; PS: parent support; TS: teacher support; CS: close friend support.

p < .05; **p < .01.

Identifying latent profiles

Latent profile analysis

The responses of the 770 learners on the SSSC were subjected to LPAs using Mplus 8.10. Table 4 shows the fit statistics for the different profiles.

Table 4.

Comparison of different latent profile analysis models.

Model (profiles)	AIC	BIC	ABIC	LMR LR testp-value	ALMR LR testp-value	BLRTp-value	Smallest class membership (%)
1	4578.28	4606.16	4587.10	N/A	N/A	N/A	–
2	4423.83	4470.29	4438.54	.0001**	.0001**	.0001**	45.32
3	4375.59	4440.64	4396.19	.0001**	.0001**	.0001**	10.39
4	4366.31	4449.78	4392.78	.487	.497	.0001**	3.89
5	4344.26	4446.49	4376.63	.062	.067	.0001**	4.12
6^a	4317.26	4438.07	4355.50	.200	.211	.0001**	4.22

AIC: Akaike information criterion; BIC: Bayesian information criterion; ABIC: adjusted Bayesian information criterion; LMR LR: Lo–Mendell–Rubin likelihood ratio test; ALMR LR = adjusted Lo–Mendell–Rubin likelihood ratio test; BLRT = bootstrapped likelihood ratio test.

Did not converge.

p < .01.

The fit indices showed a significantly better fit for every profile than the previous one (see Figure 1).

Figure 1.

Graph of the AIC, BIC, and ABIC values of the six profiles.

The entropy values for the different profiles were as follows: two profiles = 0.50; three profiles = 0.73; four profiles = 0.73; five profiles = 0.73. The average latent class assignment probabilities for the five profiles were 0.78, 0.73, 0.85, 0.82, and 0.85. Learners were classified with moderate to high certainty into their most likely latent profile. The BLRTs were significantly different (p < .0001) for the five-profile model compared to the four-profile model. Despite the lower AIC, BIC, and ABIC values of Profile 6, a substantial number of starting values in this profile did not converge. Therefore, the five-profile profile model was preferred (see Figure 2).

Figure 2.

Types of social support: The three latent profiles.

The profiles were labelled according to their mean scores on each social support construct: parental support, teacher support, and close friend support. Profile 1 – Weakly supported (4.16% of the learners) showed low parental, teacher, and close friend support. Profile 2 – Adult-supported (4.54% of the learners) showed moderate to high adult (parent and teacher) support and low peer support. Profile 3 – Peer-supported (8.31% of the learners) showed low parent support, moderate teacher support, and high peer support. Profile 4 – Moderately supported (48.31% of the learners) showed moderate scores on parent, teacher, and close friend support. Profile 5 – Integrated support (34.68% of the learners) showed high scores on parent, teacher, and close friend support.

Associations between latent profiles and mental health

Table 5 shows the differences in well-being among the different social support profiles. There were statistically significant differences in mental health among the different profiles (χ² = 70.00, p < .001). First, the weakly supported profile obtained statistically significantly lower flourishing scores than the moderately supported (χ² = 12.06, p < .001) and integrated support (χ² = 34.87, p < .001) profiles. Second, the adult-supported profile obtained statistically significantly lower flourishing scores than the peer-supported (χ² = 4.72, p = .030), moderately supported (χ² = 15.42, p < .001), and integrated support (χ² = 29.35, p < .001) profiles. In the third place, the peer-supported profile obtained a statistically significantly lower flourishing score than the integrated support profile (χ² = 9.74, p = .002), while the moderately supported profile obtained a lower score than the integrated support profile (χ² = 15.42, p < .001).

Table 5.

Latent profiles and well-being.

Profile comparison	Mean	SE
Profile 1 – Weakly supported	2.91	0.12
Profile 2 – Adult-supported	2.49	0.21
Profile 3 – Peer-supported	3.08	0.17
Profile 4 – Moderately supported	3.36	0.05
Profile 5 – Integrated support	3.66	0.05
	χ²	p
Overall test	70.00	< .001**
Weakly supported versus peer-supported	2.75	.097
Weakly supported versus adult-supported	0.73	.393
Weakly supported versus moderately supported	12.06	< .001**
Weakly supported versus integrated support	34.87	< .001**
Adult-supported versus peer-supported	4.72	.030*
Adult-supported versus moderately supported	15.42	< .001**
Adult-supported versus integrated support	29.35	< .001**
Peer-supported versus moderately supported	2.43	.119
Peer-supported versus integrated support	9.74	.002**
Moderately supported versus integrated support	15.20	< .001**

p < .05; **p < .01.

Differences between demographic groups in different profiles

An auxiliary model that included the latent profile, covariates, and distal outcomes was constructed using the BCH approach (Wang & Wang, 2020). The model consisted of two separate runs. First, the LPA was estimated using only the latent profile indicators (i.e., the sources of social support). The BCH weights were computed and saved together with the distal outcome (i.e., well-being) and three covariates (namely, sex, age category, and grade). Second, the saved data from the first run were retrieved for further analysis. Two regression models were tested using the MLR estimator. The first regression was a multinomial logit model in which the three covariates were used to predict latent profile membership. The second model was a linear regression in which the three covariates were utilised to predict mental health. The regression slope coefficients and associated parameters were restricted to be the same across all latent profiles in the latter model.

Grade (being in Grade 10) was negatively associated with flourishing in the moderately supported profile (estimate = –0.26, p = .021). Gender (male; estimate = 0.25, p = .025) was positively associated, and grade (being in Grade 10) was negatively associated (estimate = 0.26, p = .027) with flourishing in the integrated support profile. More males than females were in the weakly supported profile compared with the peer-supported profile (estimate = 2.32, p = .005). More females than males were in the moderately supported profile compared with the weakly supported profile (estimate = –1.62, p = .028). Finally, more Grade 11 than Grade 10 learners were in the integrated support compared with the moderately supported profile (estimate = 0.97, p < .001). No other statistically significant effects were found.

Discussion

This study utilised a person-centred multidimensional conceptualisation of social support (parents, teachers, and close friends) as perceived by Grades 10 and 11 adolescent learners in previously disadvantaged schools in the South African context. Studies (Harter, 1985; Lipski et al., 2014) have shown that while the dimensions of social support are distinctive, they are positively related. However, few person-centred studies have been conducted (see Ciarrochi et al., 2017 for an exception) to focus on the joint effects of different sources of perceived social support on adolescent learners. The results showed that adolescents perceived social support that was generalised across parental, teacher, and close friend support. However, the results showed that adolescents also experienced different levels of social support across sources. Different social support profiles were associated with different levels of mental health.

Our results revealed an inequality in perceived social support. Most learners (48.31%) were moderately supported by parents, teachers, and close friends. Moreover, only 34.68% of the learners were placed in the integrated support profile, showing high scores on parent, teacher, and close friend support. Furthermore, the learners (4.16%) in the weakly supported profile received low support from parents, teachers, and close friends. The adult-supported profile (4.54% of the learners) showed high parent and teacher support, but low close friend support. Conversely, learners in the peer-supported profile (8.31% of the sample) showed low parent and teacher support, but high close friend support.

Profiles that were moderate to high regarding parent support were also moderate to high in teacher and close friend support for 83% of the learners. However, a high level of peer support was not always associated with high adult support. We found no evidence for a teacher-only support profile. Therefore, high levels of perceived teacher support appeared to be accompanied by satisfactory support from learners’ parents and peers.

Social support profiles were associated with flourishing. The most desirable flourishing outcome was associated with the integrated support profile (high parent, teacher, and close friend support). In contrast, the lowest levels of flourishing were observed in the weakly supported profile (low parent, teacher, and close friend support). Studies have confirmed that emotional support from teachers is negatively associated with school-related problems, internalising problems, disengagement, and emotional issues (Tennant et al., 2015). Research has shown that parental support has a greater influence on externalised learner behaviour than peer support (Attar-Schwartz, 2015; Demir & Leyendecker, 2018).

The weakly supported profile was associated with the worst well-being outcomes in this study, while the integrated support profile was associated with the best well-being outcomes. Therefore, the well-being of adolescent learners is related to the presence or absence of social support. The sources of social support, specifically having support from parents, teachers, and close friends, are critical for the flourishing of adolescent learners.

The peer-supported profile had a well-being advantage over the adult-supported profile. The most apparent difference between the peer-supported and adult-supported profiles was parental support. Learners in the peer-supported profile had high close friend support, moderate teacher support, and low parental support. The adult-supported profile showed low close friend support, moderate teacher support, and high parent support.

Previous research has shown that social support has a low positive association with well-being. More specifically, support from teachers and school personnel is more strongly associated with adolescent learners’ well-being (Chu et al., 2010). Biglan (2015) also suggests that supportive adults (e.g., teachers) can increase positive learner outcomes. In our study, teacher support seemed linked to well-being because the scores were moderate to high in each profile where well-being was high. However, teacher support was not the most substantial in any of the profiles.

Our results, furthermore, suggested that peers were able to compensate, to some extent, for a lack of adult support. For example, the peer-supported profile perceived lower levels of parental support than the weakly supported profile and scored significantly higher in well-being. This observation is consistent with past research, suggesting that peers can be a valuable source of support (Chu et al., 2010).

Limitations and recommendations

This study was performed at the beginning of the first term, which may have produced a skewed analysis of learner well-being. It was not easy to conduct stratified random sampling for Grades 10 and 11 due to school timetables. This study showed that although parental support was important, it was essential to capitalise on the support important to adolescents and the support that was effective for them. Current programmatic interventions focusing only on one of the stakeholders – usually either parental or peer programmes – are insufficient and should be researched. Research is needed to better understand the costs and benefits associated with peer support as a compensatory mechanism for a lack of parent support.

Conclusion

This study focused on adolescents’ social support profiles and well-being relationships. Using a measure of parent, teacher, and close friend support, five social support profiles were identified: weakly supported, adult-supported, peer-supported, moderately supported, and integrated support. High parental, teacher, and close friend support showed the strongest associations with the flourishing of adolescent learners.

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Sebastiaan Rothmann

References

Antaramian

S. P.

Scott Huebner

Hills

K. J.

Valois

R. F.

(2010). A dual-factor model of mental health: Toward a more comprehensive understanding of youth functioning. American Journal of Orthopsychiatry, 80(4), 462–472. https://doi.org/10.1111/j.1939-0025.2010.01049.x

Asparouhov

Muthén

(2014). Auxiliary variables in mixture modeling: Using the BCH method in Mplus to estimate a distal outcome model and an arbitrary secondary model. Mplus Web Notes, 21(2), 1–22. https://doi.org/10.1080/10705511.2014.915181

Attar-Schwartz

(2015). Emotional closeness to parents and grandparents: A moderated mediation model predicting adolescent adjustment. American Journal of Orthopsychiatry, 85(5), 495–503. https://doi-org.nwulib.nwu.ac.za/10.1037/ort0000082

Ben-Shlomo

Taubman-Ben-Ari

(2016). Grandparenthood – Grand generativity. In Findler

Taubman-Ben-Ari

(Eds.), Grandparents of children with disabilities: Theoretical perspectives of intergenerational relationships (pp. 1–18). Springer.

Biglan

(2015). The nurture effect: How the science of human behavior can improve our lives and our world. New Harbinger Publications.

Brown

(2004). Adolescents’ relationships with peers. In Lerner

Steinberg

(Eds.), Handbook of adolescent psychology (pp. 363–394). Wiley.

Camara

Bacigalupe

Padilla

(2017). The role of social support in adolescents: Are you helping me or stressing me out? International Journal of Adolescence and Youth, 22(2), 123–136. https://doi.org/10.1080/02673843.2013.875480

Chentsova Dutton

Y. E.

Choi

. (2020). Perceived parental support and adolescents’ positive self-beliefs and levels of distress across four countries. Frontiers in Psychology, 11(353), 1–11. https://doi.org/10.3389/fpsyg.2020.00353

Chu

P. S.

Saucier

D. A.

Hafner

(2010). Meta-analysis of the relationships between social support and well-being in children and adolescents. Journal of Social and Clinical Psychology, 29(6), 624–645. https://doi.org/10.1521/jscp.2010.29.6.624

10.

Ciarrochi

Morin

A. J. S.

Sahdra

B. K.

Parker

P. D.

(2017). A longitudinal person-centered perspective on youth social support: Relations with psychological wellbeing. Developmental Psychology, 53(6), 1154–1169. https://dx-doi-org.web.bisu.edu.cn/10.1037/dev0000315

11.

Demaray

Malecki

Davidson

Hodgson

Rebus

(2005). The relationship between social support and student adjustment: A longitudinal analysis. Psychology in the Schools, 42, 691–706. https://doi.org/10.1002/pits.20120

12.

Demir

Leyendecker

(2018). School-related social support is associated with school engagement, perceived self-competence and health-related quality of life (HRQoL) in Turkish immigrant students. Frontiers in Education, 3, Article 83. https://doi.org/10.3389/feduc.2018.00083

13.

Harter

(1985). The social support scale for children and adolescents: Manual. University of Denver.

14.

Hendricks

Lake

Sanders

Shung-King

(Eds.). (2019). South African Child Gauge 2019. Children’s Institute, University of Cape Town.

15.

Henrich

Heine

Norenzayan

(2010). Most people are not WEIRD. Nature, 466, 29. https://doi.org/10.1038/466029a

16.

Hoffman

(2019). Second-parenthood realities, third-age ideals: (Grand) parenthood in the context of poverty and HIV/AIDS. In Timonen

(Ed.), Grandparenting practices around the world (pp. 1–16). Policy Press; Bristol University Press; University of Chicago Press.

17.

IBM Corp. (2017). IBM SPSS Statistics for Windows (Version 25).

18.

Ibrahim

Zaatari

W. E.

(2020). The teacher–student relationship and adolescents’ sense of school belonging. International Journal of Adolescence and Youth, 25(1), 382–395. https://doi.org/10.1080/02673843.2019.1660998

19.

Jager

(2011). Convergence and nonconvergence in the quality of adolescent relationships and its association with adolescent adjustment and young-adult relationship quality. International Journal of Behavioral Development, 35, 497–506. https://doi.org/10.1177/0165025411422992

20.

Jörns-Presentati

Napp

Dessauvagie

A. S.

Stein

D. J.

Jonker

Breet

Charles

Swart

R. L.

Lahti

Suliman

Jansen

Seedat

Groen

(2021). The prevalence of mental health problems in sub-Saharan adolescents: A systematic review. PLOS ONE, 16(5), Article e0251689. https://doi.org/10.1371/journal.pone.0251689

21.

Kendrick

Jutengren

Stattin

(2012). The protective role of supportive friends against bullying perpetration and victimization. Journal of Adolescence, 35(4), 1069–1080. https://doi.org/10.1016/j.adolescence.2012.02.014

22.

Keyes

C. L. M.

(2005). Mental illness and/or mental health? Investigating axioms of the complete state model of health. Journal of Consulting and Clinical Psychology, 73(3), 539–548. https://doi.org/10.1037/0022-006X.73.3.539

23.

Keyes

C. L. M.

(2009). Atlanta: Brief description of the Mental Health Continuum – Short Form (MHC-SF). https://peplab.web.unc.edu/wp-content/uploads/sites/18901/2018/11/MHC-SFoverview.pdf

24.

Keyes

C. L. M.

(2024). Languishing: How to feel alive again in a world that wears us down. Penguin Books.

25.

Keyes

C. L. M.

Wissing

Potgieter

J. P.

Temane

Kruger

van Rooy

(2008). Evaluation of the Mental Health Continuum–Short Form (MHC–SF) in Setswana-Speaking South Africans. Clinical Psychology & Psychotherapy, 15(3), 181–192. https://doi.org/10.1002/cpp.572

26.

Keyes

C. L. M.

Yao

Hybels

C. F.

Milstein

Proeschold-Bell

R. J.

(2020). Are changes in positive mental health associated with increased likelihood of depression over a two-year period? A test of the mental health promotion and protection hypotheses. Journal of Affective Disorders, 270, 136–142. https://doi.org/10.1016/j.jad.2020.03.056

27.

Lamers

S. M.

Westerhof

G. J.

Bohlmeijer

E. T.

ten Klooster

P. M.

Keyes

C. L.

(2010). Evaluating the psychometric properties of the mental health continuum-short form (MHC-SF). Journal of Clinical Psychology, 67(1), 99–110. https://doi.org/10.1002/jclp.20741

28.

Lipski

D. M.

Sifers

S. K.

Jackson

(2014). A study of the psychometric properties of the Social Support Scale for Children. Journal of Psychopathology and Behavioral Assessment, 36(2), 255–264. https://doi.org/10.1007/s10862-013-9383-0\

29.

Magyar

J. L.

Keyes

C. L. M.

(2019). Defining, measuring, and applying subjective well-being. In Gallagher

M. W.

Lopez

S. J.

(Eds.), Positive psychological assessment: A handbook of models and measures (2nd ed., pp. 389–415). American Psychological Association. https://doi.org/10.1037/0000138-025

30.

Malecki

C. K.

Demaray

M. K.

(2003). What type of support do they need? Investigating student adjustment as related to emotional, informational, appraisal, and instrumental support. School Psychology, 18(3), 231–252. https://doi.org.nwulib.nwu.ac.za/10.1521/scpq.18.3.231.22576

31.

Meadows

S. O.

(2007). Evidence of parallel pathways: Gender similarity in the impact of social support on adolescent depression and delinquency. Social Forces, 85, 1143–1167. https://doi.org/10.1353/sof.2007.0048

32.

Mhongera

P. B.

Lombard

(2020). Pathways of resilience for children facing socio-economic adversities: Experiences from Future Families’ OVC programmes in South Africa. Children and Youth Services Review, 108, 104657. https://doi.org/10.1016/j.childyouth.2019.104657

33.

Morin

A. J. S.

Morizot

Boudrias

Madore

(2011). A multifoci person-centered perspective on workplace affective commitment: A latent profile/factor mixture analysis. Organizational Research Methods, 14, 58–90. https://doi.org/10.1177/1094428109356476

34.

Morin

A. J. S.

Wang

(2016). A gentle introduction to mixture modeling using physical fitness performance data. In Ntoumanis

Myers

(Eds.), An introduction to intermediate and advanced statistical analyses for sport and exercise scientists (pp. 183–210). Wiley.

35.

Muthén

L. K.

Muthén

B. O.

(1998–2023). Mplus users’ guide (10th ed.).

36.

T. W.

Sorensen

K. L.

(2008). Towards a further understanding of the relationships between perceptions of support and work attitudes: A meta-analysis. Group and Organization Management, 33, 243–268. https://doi.org/10.1177/1059601107313307

37.

Pilkauskas

N. V.

Dunifon

R. E.

(2016). Understanding grand-families: Characteristics of grandparents, nonresident parents, and children. Journal of Marriage and Family, 78(3), 623–633. https://doi.org/10.1111/jomf.12291

38.

Rueger

S. Y.

Malecki

C. K.

Demaray

M. K.

(2010). Relationship between multiple sources of perceived social support and psychological and academic adjustment in early adolescence: Comparisons across gender. Journal of Youth and Adolescence, 39, 47–61. https://doi.org/10.1007/s10964-008-9368-6

39.

Sawyer

S. M.

Azzopardi

P. S.

Wickremarathne

Patton

G. C.

(2018). The age of adolescence. The Lancet: Child & Adolescent Health, 2(3), 223–228. https://doi.org/10.1016/S2352-4642(18)30022-1

40.

Schotanus-Dijkstra

Pieterse

M. E.

Drossaert

C. H. C.

Westerhof

G. J.

de Graaf

ten Have

Walburg

J. A.

Bohlmeijer

E. T.

(2016). What factors are associated with flourishing? Results from a large representative national sample. Journal of Happiness Studies, 17(4), 1351–1370. https://doi.org/10.1007/s10902-015-9647-3

41.

Tennant

J. E.

Demaray

M. K.

Malecki

C. K.

Terry

M. N.

Clary

Elzinga

(2015). Students’ ratings of teacher support and academic and social–emotional well-being. School Psychology Quarterly, 30(4), 494–511. https://doi.org/10.1037/spq0000106

42.

Wang

(2020). Structural equation modeling: Applications using Mplus. Wiley. https://doi.org/10.1002/9781118356258

43.

West

S. G.

McNeish

Savord

(2023). Model fit in structural equation modelling. In Hoyle

(Ed.), Handbook of structural equation modeling (2nd ed., pp. 184–205). Guilford Press.

44.

Windle

(1992). Temperament and social support in adolescence: Relations with depressive symptoms and delinquent behaviors. Journal of Youth and Adolescence, 21, 1–21. https://doi.org/10.1007/BF01536980

45.

Zee

K. S.

Bolger

Higgins

E. T.

(2020). Regulatory effectiveness of social support. Journal of Personality and Social Psychology, 119(6), 1316–1358. https://doi.org/10.1037/pspi0000235