Table 3 Study characteristics, findings, and methodological quality

Field et al., [4]

United States

Cross-sectional

Setting: school

N = 79 (% female not reported)

Age (range): high-school seniors (17–18 years)

1Relationship with parents/friends were assessed with questions adapted from previous study [99]

Depression was assessed using the Center for Epidemiological Studies Depression Scale (CES-D [100]).

Depressed adolescents (CES-D ≥ 19) reported inferior levels of friendship quality, fewer friends, and being less popular.

6

De Matos et al., [57]

Portugal

Cross-sectional

Setting: school

N = 6903 (53% female)

Age (mean): 14.1 years (SD: 1.71)

Peer relationships were assessed with questions adapted from the Portuguese version of Health behavior in School-Aged Children (HSBC) 1998 survey.

1Depression was assessed with 2 questions.

Participants classified as depressed had significantly lower mean scores for peer relationships.

7

Kullik and Petermann [58]

Germany

Cross-sectional

Setting: school

N = 248 (51.2% female)

Age (mean): 14.4 years (SD: 1.39)

Peer attachment was assessed using the German short version of Inventory of Parent and Peer Attachment  [101].

Depression was measured using the German version of the CES-D [102].

-ve:

Depressive symptoms were significantly correlated with attachment to peers for both females (β = −.31, p < 0.001) and males (β = −.21, p < .05).

Internal- and external-functional emotion regulation were used as mediators.

For females, internal-functional emotion regulation partially mediated the relationship between attachment to peers and depressive symptoms.

For males, both internal- and external-functional emotion regulation partly mediated the relationship between peer attachment and depressive symptoms.

9

Fadda et al., [73]

Italy

Cross-sectional

Setting: school

N = 1193 (50.3% female)

Age (mean): 16.85 years (SD: 1.74)

Relationship with peers was assessed with three items from the Health Behaviour Questionnaire [103]

Positive psychological functioning was assessed using the revised Oxford Happiness Inventory (OHI [104])

+ve

A direct significant positive association was found between peer relationships and life satisfaction (β = .17, p < .01).

An indirect positive association was also observed for peer relationships and life satisfaction partially mediated by self-esteem (β = .06, p < .01).

9

Wong [59]

United states

Cross-sectional

Setting: school

N = 144 (50.7% female)

Age (mean): 15.7 years (SD: 1.35)

Peer attachment was assessed using the Inventory of Parent and Peer Attachment (IPPA [101]).

Psychological wellbeing was assessed with the CES-D.

-ve

Peer relationships were negatively correlated with depressive symptoms (β = −.29, p = .01). Participants who reported more positive peer relationships experienced lower levels of depression.

8

Afifi et al., [60]

Oman

Cross-sectional

Setting: school

N = 5409 (49.7% female)

Age (mean): 17.11 years

Relationship with friends

Depressive symptoms were measured using the Child Depression Inventory [105].

-ve

Positive friendship quality was significantly associated with lower odds of having at least mild depressive symptoms (OR = 0.69, 95% CI: 0.63–0.76).

8

Vanhalst et al., [68]

Belgium

Cross-sectional

Setting: school

N = 884 (68% female)

Age (mean): 15.79 years (SD: 1.33)

Friendship quality was assessed with the Friendship Qualities Scale (FQS [106]). the other one of us, we Self-esteem was measured using the Rosenberg Self-esteem Scale [107].

Shyness was measured using peer nomination with participants selecting other members of their class who they regarded as “shy and/or socially withdrawn”. Received nominations were counted for each individual.

Friendship quality was used as a direct predictor and as a mediator between loneliness and self-esteem and shyness.

Peer-related Loneliness was assessed using the Loneliness and Aloneness Scale for Children and Adolescents (LACA [108]).

-ve:

Friendship quality was significantly negatively associated with loneliness (β = −.24, p < .001).

Friendship quality acted as a partial mediator between shyness and loneliness (β = .03, 95% CI: 0.01–0.05), and also between self-esteem and loneliness (β = −.07, 95% CI: − 0.11--0.04).

10

Whitehouse et al., [56]

Australia

Case-control

Setting: school

Cases:

35 adolescents with Asperger’s Syndrome (AS), (17.1% female)

Age (mean): 14.2 (SD: 0.80)

Controls:

35 adolescents (20% female)

Age (mean): 14.4 (SD: 0.10)

Friendship quality was assessed using the Friendship Quality Questionnaire (FQQ [109]).

Loneliness was measured using the De Jong-Gierveld Loneliness Scale [110].

Depressive symptoms were measured using the children’s version of the CES-D (CES-DC) [111].

Friendship quality and loneliness:

-ve

In the Typical group, overall friendship score was negatively associated with loneliness (r = − 0.36, p < 0.05). This was the same for the AS group, but the strength of the association was smaller (r = − 0.28, p < 0.05).

In the regression analysis, overall friendship score predicted levels of loneliness in the Typical group (R² = 0.12, F = 4.84, p < 0.05) but not the AS group.

Friendship quality and depressive symptoms:

Overall friendship quality did not predict depressive symptoms in either group.

7

Jose, [83, 86]

New Zealand

Longitudinal

Setting: school

N = 1774 (52% female)

Age (mean): 12.21 (SD: 1.75)

Period: one year

Peer connectedness was assessed with seven items [112].

Positive affect or happiness was assessed with the CES-D.

Negative affect or depressive symptoms was assessed the CES-D.

Peer connectedness and positive affect (happiness):

+ve

Peer connectedness at T1 was significantly positively associated with positive affect at T2 (β = 0.12, p = 0.0002).

This relationship was bidirectional, with individuals who reported higher positive affect at T1 also reported an increase in peer connectedness at T2 (β = 0.05, p = 0.037).

Peer connectedness and negative affect (depressive symptoms):

-ve

Peer connectedness was significantly negatively associated with negative affect (β = − 0.07, p = 0.011).

13

Oberle et al., [74]

Canada

Cross-sectional

Setting: school

N = 1402 (47% female)

Age (mean): 14.6 (SD: 1.03)

Positive peer relationships was assessed using the Resilience Inventory [113, 114].

Life satisfaction was assessed using the Satisfaction with Life Scale for Children (SWLS-C [115]).

+ve

Positive peer relationships were positively associated with life satisfaction (β = 0.12, t = 4.00, p < 0.001).

10

Lodder et al., [69]

Netherlands

Cross-sectional

Setting: school

N = 1172 (50.9% female)

Age (mean): 12.81 (SD: 0.43)

Best friendship quality was assessed using the Investment Model Scale [116]. friendship quality (α = 0.88).

Network friendship quality was assessed using the Peer subscale of the IPPA.

Loneliness was assessed using the Louvain Loneliness and Aloneness Scale for Children and Adolescents [108].

-ve

Best Friendship Quality and Loneliness:

Best friendship quality was negatively associated with loneliness (β = −.24, p < 0.001).

Network friendship quality and loneliness:

Network friendship quality was negatively associated with loneliness (β = −.45, p < 0.001).

10

Lambert et al., [81]

New Zealand

Cross-sectional

Setting: school

N = 9107 (46.4% female)

Age (range): 13 or under – 17 and over (SD: n/a)

Friend/peer connection was assessed with nine items.

Happiness was measured using the WHO-5 Wellbeing Index [117].

+ve:

Friend/peer connection was positively associated with happiness (β = 1.49, p < 0.0001).

11

Laghi et al., [75]

Italy

Cross-sectional

Setting: school

N = 1211 (59.6% female)

Age (mean): 17.31 (SD: 1.09)

Peer attachment was assessed with the IPPA.

Life satisfaction was assessed using the Satisfaction with Life Scale (SWLS [118]).

+ve

Peer attachment was moderately positively associated with life satisfaction (β = 0.09, p < 0.01).

9

Sasikala and Cecil, [78]

India

Cross-sectional

Setting: school

N = 97 (59.8% female)

Age (mean): 16.9 (SD: n/a)

Peer attachment was measured using the IPPA.

Self-esteem was measured using the Rosenberg Self-esteem Scale.

Psychological wellbeing was measured using the General Health Questionnaire [119].

+ve

Peer attachment was positively correlated with self-esteem (β = 0.23, p < 0.05) and subjective wellbeing (β = 0.37, p < 0.01).

9

Balluerka et al., [79]

Spain

Cross-sectional

Setting: school

N = 2182 (51.6% female)

Age (mean): 14.51 (SD: 1.55)

Peer attachment was assessed using the IPPA.

Positive affect of psychological wellbeing was assessed using the Children’s Depression Scale (CDS [120]).

+ve

Peer attachment was significantly positively associated with positive affect (β = 0.16, p < 0.0001).

9

Biggs et al., [61]

United States

Longitudinal (but the relationship under investigation was assessed cross-sectionally)

Setting: school

N = 91 (56.5% female)

Age (mean): 15.5 (SD: 0.61)

Friendship quality was assessed using a short version of the FQQ.

Depression symptoms were assessed using the short form of the Children’s Depression Inventory (CDI-SF [121]).

Null

Positive friendship quality and friendship conflict were not significantly associated with depression at T2.

11

Al-Yagon, [70]

Israel

Cross-sectional

Setting: school

N = 280 (55% female)

Age (mean): 15.94 (SD: 0.70)

Participants were categorized into three groups: (i) adolescents with learning disability (LD); (ii) adolescents with LD and attention-deficit hyperactivity disorder (LD-ADHD); and (iii) typically developed adolescents (TD).

Friendship quality was assessed using the Hebrew adaptation of the FQQ [122].

Loneliness was measured using the Peer-Network Loneliness and Peer-Dyadic Loneliness Scale [123].

Positive affect/happiness was assessed using the Hebrew adaptation of the Affect Scale ( [61, 124, 125].

-ve

Adolescents (TD) who reported higher friendship quality reported significantly lower peer-network/peer-dyadic loneliness (β = − 0.40 and β = − 0.53 respectively).

In the other groups (LD, LD-ADHD), friendship quality was significantly negatively associated with peer-network/peer-dyadic loneliness (LD: β = − 0.37 and β = − 0.48 respectively; LD-ADHD: β = − 0.42 and β = − 0.59 respectively).

A significant association was found between positive affect and friendship quality in the LD-ADHD group only (β = 0.65).

10

Corsano et al., [71]

Italy

Cross-sectional

Setting: school

N = 330 (50.9% female)

Age (mean): 15.04 (SD: 2.47)

Interpersonal relationships were assessed using the Italian adaptation of the Assessment of Interpersonal Relations [126].

Loneliness was assessed using the Italian adaptation of the Louvain Loneliness Scale for Children and Adolescents [127].

-ve:

Interpersonal relations with male peers was significantly negatively associated with peer loneliness (r = − 0.16, p < 0.01).

Interpersonal relations with female peers was significantly negatively associated with peer loneliness (r = − 0.19, p < 0.001).

8

Kamper and Ostrov, [84]

United States

Longitudinal

Setting: school

N = 776 (49.6% female)

Age (mean): 11 (SD: 0.23)

Duration: 4 years (from age 11 to 15 years)

Friendship quality was assessed with the FQQ.

Positive and negative friendship quality

were investigated as potential mediators in the relationship between relational aggression and depressive symptoms.

Depressive symptoms were assessed using the CDI-SF.

+ve

Negative friendship quality positively predicted an increase in depressive symptoms (β = 0.31, p < 0.05).

Negative friendship quality also acted as a partial mediator between relational aggression and depression (F = 16.40, p < 0.001, R² = 0.10).

13

Lieb and Bohnert, [62]

United States

Cross-sectional

Setting: online groups and healthcare settings

N = 127 adolescents with autism spectrum disorder (ASD) (18.9% female)

Age (mean): 13.95 (SD: 1.60)

Friendship quality was assessed using the abbreviated version of the FQQ (FQQ-A).

Friendship quality was also examined as a potential mediator between executive functions and loneliness and depressive symptoms.

Loneliness was assessed with the Children’s Loneliness and Social Dissatisfaction Scale [128].

Depressive symptoms were assessed with Achenbach Youth Self Report Depression Scale (YSR-D; [129]).

Friendship quality and loneliness:

-ve

The results indicated that there was a significant negative bivariate correlation between friendship quality and loneliness (r = − 0.43, p < 0.01).

Friendship quality was also negatively associated with loneliness in a separate mediation model (β = − 0.26, p ≤ 0.001).

Friendship quality mediated the relationship between emotional control and loneliness (β = 0.65, p ≤ 0.001; Sobel test = 0.12, p ≤ 0.05).

Friendship quality and depressive symptoms:

null

No significant relationship was found between friendship quality and depressive symptoms.

15

Raboteg-Saric and Sakic, [76]

Croatia

Cross-sectional

Setting: school

N = 401 (39.9% female)

Age (mean): 16.92 (SD: 1.16)

Friendship quality was assessed with the short form version of the FQS [130].

Happiness was measured using the Subjective Happiness Scale (SHS [131]).

Life satisfaction was assessed using the Student’s Life Satisfaction Scale (SLSS [132]).

Self-esteem was assessed using the Rosenberg Self-Esteem Scale.

Friendship quality and happiness:

Adolescents who reported higher levels of friendship quality also had greater levels of happiness (F = 18.43, p < 0.001) partial η² = 0.06).

Friendship quality and life satisfaction:

Adolescents with higher levels of friendship quality were more satisfied with life (F = 4.22, p = 0.041, partial η² = 0.01).

Friendship quality and self-esteem:

Adolescents who rated their friendship quality more favourably had marginally higher levels of self-esteem (F = 3.42, p < 0.065, η² = 0.01).

8

Oppenheimer and Hankin, [85]

United States

Longitudinal

Data collection at three time points across five-week intervals

Setting: school

N = 350 (57% female)

Age (mean): 14.5 (SD: 1.40)

Relationship quality was assessed using the short version of the Network of Relationships Inventory (NRI; [133]).

Depressive symptoms were assessed using the CDI.

Null:

Adolescents’ perception of friendship quality has no effect on their depressive symptoms while the inverse is true

Standardized path coefficients indicated that depressive symptoms (T1) were significantly negatively associated with positive relationship quality (T2) (β = − 0.13, p ≤ 0.001) and positively associated with negative relationship qualities (T2) (β = 0.14, p ≤ 0.01).

Likewise, depressive symptoms (T2) were significantly negatively associated with positive relationship quality (T3) (β = − 0.15, p ≤ 0.001) and positively associated with negative relationship qualities (T3) (β = 0.07, p ≤ 0.01).

13

Huang and Chen, [63]

Taiwan

Cross-sectional

Setting: school

N = 1325 (53.2% female)

Age (mean): 16.5 (SD: 0.90)

Negative peer relationships were assessed using the Taiwanese Relationship Inventory for Child and Adolescent [134].

Depressed mood was assessed using Chinese version of the Center of Epidemiological Studies Scale [135].

+ve:

The results indicated a significant positive bivariate correlation between negative peer relationships and depressed mood (r = 0.49, p < 0.001).

Result of the multiple linear regression model showed a significant positive association between negative peer relationships and depressed mood (β = 0.28, p < 0.001).

11

MacPhee and Andrews, [64]

Canada

Cross-sectional

Setting: home

N = 2014

Female percentage (n/a)

Age (range): 12–13

Peer relations were assessed using the Peer Relations scale of the Self-Description Questionnaire (SDS [136]).

Depressive symptoms were assessed using a shortened version of the CES-D.

+ve

Peer relations were significantly associated with depressive symptoms in the multivariate stepwise regression analysis (R² = 0.418, ∆ R² = 0.002, p < 0.001).

For males, peer relations were significantly associated with depressive symptoms (R² = 0.352, ∆ R² = 0.013, p < 0.001), however, this was not the case for females.

10

Newland et al., [77]

United States

Cross-sectional

Setting: school

N = 149 (47.7% female)

Age (mean): 13 (SD: n/a)

Peer relationships were assessed with an adapted version of the Children’s Worlds survey [137]

Life satisfaction was assessed with the Children’s Worlds survey (α = 0.90).

+ve:

The results showed a significant positive bivariate correlation between peer relationships and life satisfaction (r = 0.59, p < 0.001).

Results from the regression analysis indicated that adolescents who reported higher quality peer relationships had higher levels of life satisfaction (β = 0.32, p < 0.001).

10

Preddy and Fite, [65]

United States

Cross-sectional

Setting: home

N = 89 (44.0% female)

Age (mean): 10.4 (SD: 1.1)

Friendship quality was assessed with an abbreviated version of the FQQ to assess friendship quality with a best friend.

Friendship quality was also examined as a moderator between different subtypes of aggression and depressive symptoms.

Depressive symptoms were assessed using the CDI.

-ve:

The results indicated that there was a significant negative bivariate correlation between friendship quality and depressive symptoms (r = − 0.25, p < 0.05).

Regression analysis results indicated that friendship quality was negatively associated with depressive symptoms (β = − 0.11, p = 0.02).

Friendship quality did not moderate the relationship between aggression subtypes and depressive symptoms.

11

Ang, [72]

Malaysia

Cross-sectional

Setting: school

N = 618 (57.0% female)

Age (mean): 14.80 (SD: 0.99)

Friend connectedness was measured using the Hemingway Measure of Adolescent Connectedness (MAC [138]).

Loneliness was assessed using the UCLA Loneliness Scale [139]

Null

The results indicated no statistically significant association between friend connectedness and loneliness (β = − 0.02, p > 0.05).

11

Chang et al., [55]

Taiwan

Case-control

Setting: school, home, clinic, and community

Cases: 101 adolescents with ASD (20.2% female)

Age (mean): 15.6 (SD: n/a)

Controls: 101 neuro-typical adolescents (74.5% female)

Age (mean): 16.1 (SD: n/a)

Friendship quality was assessed using the FQS

Loneliness was assessed using the Chinese version of the short-form UCLA Loneliness Scale [134].

-ve:

The results showed a significant negative bivariate correlation between friendship quality and loneliness for the ASD group (r = − 0.229, p < 0.05) and neuro-typical group (r = − 0.220, p < 0.05).

8

McMahon et al., [82]

Ireland

Cross sectional

Setting: Home

N = 7527

(51.1% female)

Age (specific):

13 (98.4% of the sample)

(SD: n/a)

Peer attachment was assessed with the IPPA.

Psychological wellbeing was assessed with The Piers-Harris Self-Concept Scale [140]

-ve

Friendship quality was significantly associated with subjective wellbeing for (β = − 1.00, p < .001, [− 1.57, −.44], R2 = .006) but not for boys (β = −.27, p = .37, [−.85, .31]).

For girls, friendship quality partially meditated the association between stressful life events and subjective wellbeing (β = −.27, SE = −.27, [−.51, −.03])

For boys, friendship quality has no mediation effect on the association between stressful life events and psychological wellbeing

9

Smokowski et al., [66]

United States

Cross sectional

Setting: home

N = 4321 (53.0% female)

Age (mean): 12.8 (SD: n/a)

Friendship quality was assessed using the Negative peer relationships scale [141].

Depressive symptoms were assessed using the School Success Profile (SSP) [141].

Self-esteem was assessed using the Rosenberg Self-Esteem Scale.

Negative peer relationships and Depressive symptoms:

Results indicated that adolescents who reported higher negative peer relationships had higher levels of depressive symptoms (p < 0.001).

Negative peer relationships and self-esteem:

Adolescents who reported higher negative peer relationships had lower levels of self-esteem (p < 0.001).

8

Nyarko, [80, 87]

Ghana

Cross sectional

Setting: school

N = 100 (34.0% female)

Age (range):

15–18 (SD: n/a)

Peer relationship was measured using the revised class play method of peer assessment [142].

Self-esteem was assessed using the Rosenberg Self-Esteem Scale..

-ve:

The results showed a significant negative correlation between friendship quality and self-esteem (r = − 0.231, p < 0.05)

7

Spithoven et al., [67]

Belgium and Netherlands

Cross sectional

Two samples were used in this study but only one sample falls under our eligibility criteria

Sample of Dutch adolescents – 2nd sample):

Setting: school

N = 1361 (51.32% female)

Age (mean):

12.81 (SD: 0.42)

Friendship quality was assessed with theIPPA.

Depressive symptoms were measured using short form of the (CES-D [143]).

Peer-related Loneliness was assessed using the LACA.

Self-esteem was assessed using the Single Item Self-Esteem scale [144].

Happiness was assessed using a single item scale [145].

Friendship quality and depressive symptoms:

-ve:

Friendship quality inversely correlated with depressive symptoms (r = − 0.39, p < 0.001)

Friendship quality and loneliness:

-ve:

Friendship quality significantly negatively associated with loneliness (r = − 0.52, p < 0.001)

Friendship quality and self-esteem:

+ve

Friendship quality positively associated with self-esteem (r = 0.21, p < 0.001)

Friendship quality and happiness:

+ve

Friendship quality positively correlated with happiness (r = 0.36, p < 0.001)

9

O’Connor et al., [88]

Netherlands

Cross-sectional

Setting: Autism center, special education school, and public school

N = 306

Age (mean): 11.69 (SD: 1.33)

Autistic group:

N = 104 (17.3% female)

Typically developed group:

N = 202 (55.0% female)

Friendship Quality was assesses using the Best Friend Index (BFI [146];

Depressive symptoms were assessed using the Dutch version of CDI [147]

For females, negative friendship quality was positively associated with depressive symptoms in autistic (r = 0.48, 95% CI: 0.02–0.77) and typically developed adolescents (r = 0.28, 95% CI: 0.10–0.44), whereas positive friendship quality was negatively correlated among typically developed group only (r = − 0.24, 95% CI: − 0.41 - -0.06)

For males, positive friendship quality was negatively correlated with depressive symptoms only among autistic group (r = − 0.37, 95% CI: − 0.54 - -0.17)

7

Putri and Muttaqin, [89]

Indonesia

Cross-sectional

Setting: online survey

N = 450 (70.4% female)

Age (mean): 16.0 (SD: 2.58)

Friendship Quality was assessed using McGill Friendship Questionnaire-Friend’s Functions (MFQ-FF [148])

Life Satisfaction was assessed using Satisfaction with Life Scale (SWLS: [149])

Positive associations between friendship quality and life satisfaction were observed via four indirect paths:

1) Basic psychological need satisfaction (β = .330, p < 0.001)

2) Autonomy (β = 0.183, p < 0.001)

3) Competence (β = 0.224)

4) Relatedness (β = 0.303, p < 0.001)

9

Lim, [90]

South Korea

Longitudinal

Setting: School, home

N = 2250 (46.6% female)

Age (mean): 14.0 (SD: 0.03)

Negative peer relationships were assessed using Peer Relationship Quality Scale [150]

Life satisfaction was assessed using SWLS.

Standardized path coefficients indicated that negative peer relationships (T1) were significantly negatively associated with life satisfaction (T2) (β = − 0.043, p < 0.01).

Standardized path coefficient also indicated a significant negative relationship between negative peer relationships (T2) and life satisfaction (T3) (β = − 0.084, p < 0.01).

Negative peer relationships quality (T2) acted as a full mediator between smartphone dependence (T1) and life satisfaction (T3) (β = −  0.043, 95% CI: − 0.082- -0.020)

13

Kühner et al., [91]

China

Cross-Sectional

Setting: School

N = 1279 (47.8% female)

Age (mean): 11.47 (SD: 1.12)

Peer relationship was assessed using four questions developed and tested by the author

Subjective wellbeing was assessed using Children’s Worlds Subjective Well-Being Scale (CW-SWBS [151])

Overall life satisfaction was assessed using CW-SWBS

Peer relationship and subjective wellbeing:

Overall peer relationship score was positively associated with subjective wellbeing (β = .150, p < .001)

Peer relationship quality acted as a partial mediator between socioeconomic status and subjective wellbeing (β = 0.011, 95% CI: 0.003–0.021)

Peer relationship and overall life satisfaction:

Overall peer relationship score was positively associated with subjective wellbeing (β = .090, p < .001)

Peer relationship quality acted as a partial mediator between socioeconomic status and overall life satisfaction (β = 0.007, 95% CI: 0.001–0.014)

9

Choe and Yu, [92]

South Korea

Longitudinal

Setting: School

N = 1737 (48.4% female)

Age (mean): 14.0 (SD: 2.0)

Friendships was assessed using the School Life Adaptation Scale—Peer Relationships [152]

Depressive symptoms were assessed using Depression Scales of the Korean Mental Diagnostic Test [153]

Longitudinal bidirectional relationship between friendships and depressive symptoms was observed

Significant inverse relationships were found between friendships (T1) and depressive symptoms (T2) (β = − 0.148, p = 0.000), and inverse was true (β = − 0.091, p = 0.000)

Significant inverse relationships were found between friendships (T1) depressive symptoms (T2) (β = − 0.109, p = 0.000), and the inverse was true (β = − 0.066, p = 0.000).

13

Forgeron et al., [93]

Canada

Longitudinal (but the relationship under investigation was assessed cross-sectionally)

Setting: healthcare

N = 83 (% female not reported)

Age (mean): 15.29 (SD: 1.26)

Friendship quality was assessed using the FQS.

Depressed mood was assessed using the CES-D.

Loneliness was assessed using the Loneliness Scale [128].

Self-esteem was assessed the Rosenberg Self-Esteem Scale.

Friendship quality and depressed mood:

-ve

The results indicated that there was a significant bivariate correlation between depressed mood (T2) and negative (r = − 0.311, p < 0.01) and positive (r = − 0.293, p < 0.01) friendship quality (T2), but no significant relationships were identified between the same variables at T1

Friendship quality and loneliness:

-ve

There was a significant negative bivariate cross-sectional correlation solely between negative friendship quality and loneliness for each wave (T1: r = − 0.223, p 0.05; T2: r = − 0.270, p 0.05; T3: r = − 0.223, p 0.05).

Friendship quality and self-esteem:

+ve

The results indicated that there was a significant positive bivariate correlation between self-esteem (T2) and negative (r = 0.296, p < 0.01) and positive (r = 0.270, p < 0.05) friendship quality (T2), but no significant relationships were identified between the same variables at T1

10

Powell et al., [94]

United Kingdome

Longitudinal

Setting: School

N = 1712 (46.5% female)

Age (range): 11–12

Friendship quality of the best friend and top three friends was assessed using the FQS.

Depressive symptoms were assessed using the Short Moods and Feelings Questionnaire [154].

Friendship quality of best friend was negatively associated with later depressive symptoms (β = − 0.72, 95% CI: − 0.97 - -0.47, P < 0.001).

Friendship quality of top three friends friend was negatively associated with later depressive symptoms (β = − 0.69, 95% CI: − 0.94 - -0.44, P < 0.001).

14

Zhao et al., [95]

China

Cross-sectional

Setting: School

N = 1863

Participants were categorized into three groups bases on their parents’ migration status:

(i) Nonparent migrant group: N = 643 (49.6% female), age (mean) = 14.34 (SD) = 0.87

(ii) Both-parent migrant group: N = 409 (45.7% female), age (mean) = 14.33 (SD) = 1.19

(iii) Father-only migrant group: N = 750 (46.8% female), age (mean) = 14.32 (SD) = 0.84

Friendship quality was assessed by using the Chinese version of FQQ.

Depressive symptoms were assessed by using the Chinese version of the CDI

Friendship quality was negatively correlated with depressive symptoms among all adolescents’ groups irrespective of their fathers’ migration status (Nonparent migrant: r = − 35, both-parent migrant: r = −,32, Father-only migrant: r = −.29, p < .001)

10

Luijten et al., [96]

Netherlands

Longitudinal

Setting: School

N = 1298 (53.2% female)

Age (mean): 13.7 (SD: 1.1)

Friendship quality was assessed using the Network of Relationships Inventory.

Wellbeing was assessed using MHSC-SF.

No significant longitudinal association was found between friendship quality (T1) and adolescents’ wellbeing (T2)

12

Gautam et al., [97]

Nepal

Cross-sectional

Setting: School

N = 371 (49.1% female)

Age (mean): 17.4 (SD = 0.92)

Relationship with friend was assessed using The Patient Health Questionnaire (PHQ-9 [155];

Depressive symptoms were assessed using the PHQ-9.

Poor relationship with friend was significantly associated with higher depressive symptoms (β = 2.371, 95% CI: 1.078–5.215, p = 0.032)

11

Schwartz-Mette et al., [98]

United states

Longitudinal

Setting: School and community

N = 186 (69.9% female)

Age (mean): 15.68 (SD = 1.49)

Same-gender positive friendship quality was assessed using the revised revision of FQQ [156].

Depressive symptoms were assessed using CES-D.

Longitudinal bidirectional relationship between same-gender positive friendship quality and depressive symptoms was observed

Significant inverse relationships were found between positive friendship quality (T1) and depressive symptoms (T2) (β = − 0.15, p < 0.000), and the inverse, to a lesser extent, was also true (β = − 0.08, p = p < 0.05)

12