Introduction
Methods
Search Strategy
Study Selection
Inclusion criteria | Exclusion criteria |
---|---|
Participant | |
• Have a clinical diagnosis of Autism Spectrum Disorder or equivalent (e.g. childhood autism (ICD-10)/Autistic Disorder (DSM-IV), Asperger’s Syndrome, Pervasive Developmental Disorder – Not Otherwise Specified) • Clinical diagnosis should be provided (ideally but not exclusively) by a healthcare or educational professional via clinical assessment measures | • Does not include individuals with clinical diagnosis of Autism Spectrum Disorder or equivalent |
Exposure | |
• Include at least one instrument to measure social anxiety symptom severity, but not limited to the tests and measures stated below, which have been identified by a systematic review by Leigh and Clark (2018) and Spain et al. (2018) ADIS-IV = Anxiety Disorders Interview Schedule for DSM-IV BFNE = Brief Fear of Negative Evaluation Scale CAPA = Child and Adolescent Psychiatric Assessment CASI = Child and Adolescent Symptom Inventory K-SADS-PL = Schedule for Affective Disorder sand Schizophrenia for School Aged Children, present and lifetime version LSAS = Liebowitz Social Anxiety Scale MASC = Multidimensional Anxiety Scale for Children MINI = Mini Interpersonal Neuropsychiatric Interview RCADS = Revised Children’s Anxiety and Depression Scale SAS (-A; -CR) = Social Anxiety Scale (for Adolescents, for Children-Revised) SASPA = Social Anxiety Scale for People With ASD SCARED = Screen for Child Anxiety Related Emotional Disorder SCAS = Spence Children’s Anxiety Scale SCID = Structured Clinical Interview for DSM-IV SDS = Social Desirability Scale SIAS = Social Interaction Anxiety Scale SPAI (-C) = Social Phobia Anxiety Inventory (for Children) SPIN = Social Phobia Inventory SPS = Social Phobia Scale SSRS = Social Skills Rating Scale SWQ = Social Worries Questionnaire | • Does not include any instrument to measure social anxiety |
Comparison (optional) | |
• May include age-matched sample of neurotypical or non-autistic individuals with or without social anxiety symptoms as a comparison group. If the study meets the requirement under Participant and Exposure of the PECO criteria, absence of a comparison group will not lead to the exclusion of the study in the systematic review, as a comparison group is optional and not required to address the stated research question | • Not applicable |
Outcome | |
For Research Question 1: • Must meet the inclusion criteria for Participant and Exposure, and include at least one measure of at least one psychological variable identified in the cognitive model of social anxiety by Clark and Wells (1995). A list of psychological variables is found in Clark and Wells (1995) and in the systematic review by Leigh and Clark (2018): 1) Negative social attitudes and cognitions 2) Negative interpretation bias 3) Self-focussed attention 4) Reduced processing of external social cues 5) Negative observer-perspective social images 6) Use of internal information 7) Use of safety behaviours 8) Pre- and post-event processing For Research Question 2: • Must meet the inclusion criteria for Participant and Exposure and may include at least one measure of the following processes that have shown associations with social anxiety symptom severity amongst autistic individuals: 1) Social skills, knowledge, motivation and competence beyond autism symptom severity (Lerner et al., 2012) 2) Social camouflaging / masking (Hull et al., 2021) 5) Interoceptive awareness and self-regulation (Pickard et al., 2020) 6) Alexithymia (Pickard et al., 2020) | • Does not include any measure of psychological variables identified in Clark and Wells (1995) model of social anxiety OR any of the mechanisms outlined under research question 2 that include more autism specific factors to consider |
Quality Appraisal
Data Extraction
Aim 1 – Clark and Well model (n = 901; 13 studies; 13 independent samples) | Aim 2 – Autism related factors (n = 2248; 37 studies; 32 independent samples) | |||
---|---|---|---|---|
M (SD) | Range | M (SD) | Range | |
Sample size | 69.31 (50.92) | 13—192 | 70.25 (73.21) | 8—354 |
% Male | 69.53 (21.89) | 28 – 100 | 74.20 (20.61) | 27.87—100 |
Age (Years) | 20.17 (9.84) | 11.91 – 49.05 | 19.96 (11.34) | 8.44 – 49.05 |
FSIQ | (7 studies) | (21 studies, 19 independent samples) | ||
104.99 (6.33) | 98.16 – 108.38 | 102.65 (15.65) | 45 – 118.65 | |
Ethnicity | (%—4 studies) | (%—9 studies) | ||
Caucasian | 81.06 (3.72) | 77.78 – 85.94 | 76.61 (17.39) | 36.67 – 90.08 |
Mixed/Other ethnicity | 12.07 (1.25) | 10.71 – 13.58 | 5.97 (9.98) | 0 – 30 |
Black | 2.16 (2.54) | 0 – 4.92 | 3.84 (5.74) | 0 – 17.39 |
Asian | 4.71 (4.30) | 1.56 – 10.71 | 14.69 (8.63) | 4.96 – 33.33 |
Study quality | 0.82 (0.09) | 0.64 – 0.91 | 0.92 (0.03) | 0.61 – 0.95 |
Recruitment | (n = studies) | (n = studies, samples in brackets) | ||
Clinical sites (hospitals / clinic) | 7 | 23 (19 samples) | ||
Community settings | 5 | 16 (15 samples) | ||
School/ University | 8 | 10 (8 samples) | ||
Online | 2 | 4 (4 samples) | ||
Longitudinal/cohort datasets | 0 | 1 (1 sample) | ||
Comorbidities | (n = participants; 6 studies) | (n = participants; 12 studies; 10 independent samples) | ||
ADHD | 51 | 74 | ||
Depression | 21 | 202 | ||
Anxiety (unspecified) | 5 | 9 | ||
GAD | 44 | 272 | ||
SAD | 122 | 148 | ||
Simple phobia | 2 | 2 | ||
Separation Anxiety | 14 | 49 | ||
Panic | 22 | 36 | ||
PTSD | 2 | 2 | ||
OCD | 32 | 62 | ||
Eating Disorder | 6 | 6 | ||
Learning/Language related diagnosis (including dyslexia, dyspraxia) | 37 | 41 | ||
Selective Mutism | - | 38 | ||
Social Anxiety Measure | (n = studies) | (n = studies) | ||
Self-report | 13 | 26 | ||
Caregiver report | 2 | 14 | ||
Clinician/Teacher report | 1 | 4 | ||
Mechanism Measure | (n = studies) | (n = studies) | ||
Self-report | 10 | 14 | ||
Caregiver report | 4 | 16 | ||
Clinician/Teacher measure | 2 | 4 | ||
Task-based performance measure | 2 | 13 |
Authors (Year, Country) | Sample Size (N, male) Recruitment Channel | Diagnosis (Criteria/ Measure) | Age (M; SD) IQ (M; SD) Ethnicity (Caucasian N; %) Co-occurring conditions in ASD group (N; %) | Social Anxiety Measure | Mechanism Measure | Outcome | Quality Score |
---|---|---|---|---|---|---|---|
A. Negative Social Cognitions (n = 3) | |||||||
Lei et al. (2023); UK | ASD: 61 (17) NT: 54 (11) Clinic; School/University; Online | Clinical diagnosis AutSev (ASD): AQ-28 77.93 (10.23) | Age: ASD: 16.34 (1.69); NT: 16.02 (1.56) IQ: none reported Ethnicity: ASD: 50 (81.97); NT: 37 (68.52) Co-occurring: 9 ADHD, 22 GAD, 19 SAD, 21 OCD, 1 Panic, 2 PTSD, 17 Depression, 6 Eating disorder | SPIN (self) | ASCQ (self) | Covariates controlled for: GAD and depression symptom severity, groups matched on SA Social anxiety: ASD and Non-ASD matched Social cognitions: ASD and Non-ASD matched Interaction: ASD: ↑ SA = ↑ Social anxiety related social cognitions (r = .69, p < .001) | 0.91 |
Wilson et al. (2023); UK | ASD: 192 (56) Non-ASD: 69 (26) Self-identify ASD: 51 (14) Broad Aut Phenotype (BAP): 24 (9) Community; Online | Clinical diagnosis; AQ AutSev (ASD): AQ-10 7.89 (1.79) | Age: ASD: 40.95 (13.62); N-ASD: 40.30 (17.28); S-ASD: 41.98 (13.95); BAP: 32.82 (11.16) IQ: none reported Ethnicity: ASD: 165 (86); N-ASD: 61 (88); S-ASD: 42 (82); BAP: 17 (71) Co-occurring: 58 SAD, 36 ADHD, 13 dyslexia, 15 dyspraxia, 9 language/learning-related diagnosis | LSAS (self) | SCogQ (self) | Covariates controlled for: none reported; no association between outcome variables and age/gender Social anxiety: ASD > non-ASD (Cohen’s d = 1.01) Social cognitions: ASD > non-ASD (Cohen’s d = 0.8) Interaction: ASD: ↑ SA = ↑ Social anxiety related social cognitions (r = .56, p < .05) | 0.95 |
Hollocks et al. (2016); UK | ASD: 21 (21) ASDAnx: 34 (34) Non-ASD: 28 (28) Clinical site | Clinical diagnosis, ADI-R/ADOS, SCQ AutSev: SCQ 19.4 (5.7) for ASD, 24.7 (5.8) for ASDAnx | Age: ASD: 13 (1.9); ASDAnx: 12.7 (1.9); Non-ASD: 13.9 (1.8) IQ: ASD: 103 (16.7); ASDAnx: 99.7 (10.9); Non-ASD: 116 (9.5) Ethnicity: none reported Co-occurring: 62% ≥ 1 anxiety; 18% ≥ 3 anxiety; 21 panic/agoraphobia, 22 GAD, 14 separation, 2 simple phobia, 4 SAD, 11 OCD | SCAS-P (caregiver) SCAS-C (self) | Interpretation bias: Ambiguous Situation Interview (self) | Covariates controlled for: IQ Social anxiety: ASDAnx > ASD > non-ASD Interpretation bias: ASDAnx > non-ASD on negative interpretations for social and non-social situations, no group differences in physical threat situations Interaction: SA rated by parent and child not associated with total interpretation bias (r = .05 to .17, p > .05), social interpretation bias (r = .1 to .16, p > .05), or physical interpretation bias (r = -.11 to .28, p > .05) | 0.86 |
B. Perceived Social Danger (including fear of negative evaluation from other) (n = 4) | |||||||
Boulton et al. (2021); Australia | ASD: 102 (73) SAD: 316 (178) ASD + SAD: 60 (40) Clinical site | Clinical diagnosis; ADOS (DSM-5); ADIS DSM-IV (for SAD) AutSev: ADOS-2: 9.59 (2.79) for ASD group; 9.08 (2.48) for ASD + SAD group | Age: ASD: 23.23 (6.61); SAD: 24.64 (7.11); ASD + SAD: 23.57 (3.29) IQ: ASD: 108.38 (7.22); SAD: 111.21 (3.35); ASD + SAD: 107.58 (6.40) Ethnicity: none reported Co-occurring: SAD | LSAS-SR (self) SIAS (self) SPS (self) | BFNE (self) | Covariates controlled for: IQ Social anxiety: n above clinical cut-off for each measure: ASD: 67/102 on LSAS-SR; 73/102 on SIAS; 57/102 on SPS SAD: 227/316 on LSAS-SR; 263/316 on SIAS; 210/316 on SPS Interaction: ASD group: ↑ FNE = ↑ SA (r = 0.64, p < .001); fear (r = .70, p < .001), ↑ avoidance (r = .52, p < .001); ↑ social interaction anxiety (r = 0.67, p < .0017) ↑ being scrutinised by others (r = 0.62, p < .001) SA group: ↑ FNE = ↑ SA (r = 0.61, p < .001); fear (r = .66, p < .001) ↑ avoidance (r = .54, p < .001); ↑ social interaction anxiety (r = .69, p < .001) ↑ being scrutinised by others (r = 0.61, p < .001) | 0.82 |
Gaziel-Guttman et al. (2023); Israel | ASD: 33 (28) Non-ASD: 38 (32) Community | Clinical diagnosis by DSM-IV-TR or DSM-5; ADOS AutSev: AQ 30.7 (3.89) | Age: ASD: 23.55 (2.37); Non-ASD: 24.29 (2.18) IQ: ASD: 116.7 (6.94); Non-ASD: 118.58 (6.52) Ethnicity: none reported Co-occurring: 6 ADHD, 5 anxiety, 4 depression | LSAS (self) | GASP (self) | Covariates controlled for: none reported; no group diff in gender, age, IQ Social Anxiety: ASD > non-ASD on total, fear, avoidance Shame: ASD < non-ASD on total and negative-self-evaluation subscale; no group diff on withdraw subscale Interaction: • ASD: ↑ Avoidance = ↓ Negative self-evaluation (r = -.39, p < .05) ASD vs. Non-ASD: ↑ Shame = ↑ SA only in non-ASD group (fear subscale: r = .53, p < .001; withdraw subscale: r = .42, p < .001); ↑ Shame = ↑ Withdraw in both groups (ASD: r = .85, p < .001; non-ASD: r = .89, p < .001) | 0.91 |
Maddox et al. (2015); USA | ASD: 28 (15) SAD: 26 (13) Non-Clinical: 25 (12) University | Clinical diagnosis, ADOS-2 and ADI-R AutSev: SRS-2-A 68.43 (9.82) | Age: ASD: 23.93 (6.92); SAD: 25.96 (7.12); NC: 24.78 (7.31) IQ: ASD: 106.68 (16.58); SAD: 108.85 (10.51); NC: 114.24. (10.78) Ethnicity: ASD: 22 (78.6); SAD: 20 (76.9); NC: 17 (68) Co-occurring: SAD (14, 50%) | ADIS-IV: SAD module SASPA (clinician) SIAS (self) | BFNE (self) | Covariates controlled for: no between group diff in gender, age and IQ not associated with social anxiety Social anxiety: 14 (50%) of ASD group > SAD cut-off; Most highly endorsed items on ADIS-IV: ASD + SAD (n = 14): 1) attending parties; 2) formal speaking; 3) speaking with unfamiliar people SAD: 1) formal speaking; 2) being assertive to ask others to change behaviours; 3) speaking with unfamiliar people Interaction: • ASD + SAD > SAD on social interaction anxiety – especially on “making eye contact”; “talk about myself or my feelings”, “mix comfortably with peers”, “making friends of my own age”, “talking with others” • ASD + SAD vs. SAD: ↑ SA during middle childhood • 50% of ASD + SAD and 8% of SAD group attributed autism related social skills impairment to the development and maintenance of SA | 0.86 |
Pickard et al. (2020); UK | ASD: 61 (42) NT: 62 (26) School | Clinical and school diagnosis; SCQ-L and SRS-2 AutSev (ASD): SRS-2 77.47 (10.65); SCQ lifetime 20.94 (6.98) | Age: ASD: 13.46 (1.77); NT: 13.52 (1.57) IQ: ASD: 98.16 (13.99); NT: 100.76 (11.55) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | BFNE (self) | Covariates controlled for: sex Social anxiety: 52.4% of ASD vs. 54.8% NT > SA cut-of FNE: no difference between NT and ASD (d = .22, p > .05) Interaction: • ↑ FNE = ↑ SA in ASD (r = .72, p < .001) and NT (r = .66, p < .001) • ↑ FNE = ↑ autistic traits in NT (r = .31, p < .05) and not in ASD (r = .18, p > .05) | 0.91 |
C. Processing of Self as Social Object (n = 2) | |||||||
Wilson et al. (2023); UK | ASD: 192 (56) Non-ASD: 69 (26) Self-identify ASD: 51 (14) Broad Aut Phenotype (BAP): 24 (9) Community Online | Clinical diagnosis; AQ AutSev (ASD): AQ-10 7.89 (1.79) | Age: ASD: 40.95 (13.62); N-ASD: 40.30 (17.28); S-ASD: 41.98 (13.95); BAP: 32.82 (11.16) IQ: none reported Ethnicity: ASD: 165 (86); N-ASD: 61 (88); S-ASD: 42 (82); BAP: 17 (71) Co-occurring: 58 SAD, 36 ADHD, 13 dyslexia, 15 dyspraxia, 9 language/learning-related diagnosis | LSAS (self) | FAQ (self) | Covariates controlled for: none reported; no association between outcome variable sand age/gender Social anxiety: ASD > non-ASD (Cohen’s d = .1.01) Self-focussed attention: ASD > non-ASD (Cohen’s d = 0.6) Interaction: ASD: Self-focussed attention not associated with SA (r = 0.41, p > .05) | 0.95 |
Wood et al. (2022); UK | Study 1: High SA: 41 (29) Low SA: 30 (26) Study 2: 76 (44) Community; School/University | Clinical diagnosis AutSev: none reported | Study 1: Age: High SA: 18.17 (2.46); Low SA: 17.33 (1.42) IQ: none reported Ethnicity: none reported Study 2: Age: 17.91 (1.93) IQ: none reported Ethnicity: none reported Co-occurring: none reported for either | SAS-A (self) SPIN (self) | FAQ (self) SCS (self) Self-imagery: Performance scale | Covariates controlled for: none reported. No between group diff on demographics Social anxiety: High SA > Low SA on FNE, anxiety in general and new situations Social performance: High SA < Low SA on self- and observer-rated social performance Interaction: • ↑ SA = ↓ Self-rated performance (r = -.415, p < .001) (fully mediated by greater interoceptive sensibility; partially mediated by public social consciousness/trait SFA) • ↑ SA = ↑ Self-focussed attention (r = .332, p = .003) | 0.82 |
D. Physical arousal and interoceptive sensibility (use of internal information, n = 3) | |||||||
Bellini et al. (2006) | ASD: 41 (35) School | Clinical diagnosis AutSev: none reported | Age: 14.22 FSIQ: 99.94 (18.81) Ethnicity: none reported Co-occurring: none reported | SAS-A (self) | MASC-Physical Symptoms Scale (self) | Covariates controlled for: none reported Interaction: ↑ physical arousal = ↑ SA (β = .52, p < .01) | 0.64 |
Pickard et al. (2020); UK | ASD: 61 (42) NT: 62 (26) School | Clinical and school diagnosis AutSev (ASD): SRS-2 77.47 (10.65); SCQ lifetime 20.94 (6.98) | Age: ASD: 13.46 (1.77); NT: 13.52 (1.57) IQ: ASD: 98.16 (13.99); NT: 100.76 (11.55) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | BPQ-Awareness (self) | Covariates controlled for: sex Social anxiety: 52.4% of ASD vs. 54.8% NT > SA cut-off Interoceptive sensibility: NT > ASD (d = .58, p < .01) Interaction: ASD: ↑ SA = ↑ Interoceptive sensibility (r = .46, p < .001) NT: ↑ SA = ↑ Interoceptive sensibility (r = .33, p < .009) No difference in magnitude of correlations between groups (Zs = -.38–1.64, ps > .101) | 0.91 |
Wood et al. (2021); UK | Study 1: High SA: 41 (29) Low SA: 30 (26) Study 2: 76 (44) Community; School/University | Clinical diagnosis AutSev: none reported | Study 1: Age: High SA: 18.17 (2.46); Low SA: 17.33 (1.42) IQ: none reported Ethnicity: none reported Study 2: Age: 17.91 (1.93) IQ: none reported Ethnicity: none reported Co-occurring: none reported for either | SAS-A (self) SPIN (self) | APQ (self) | Covariates controlled for: none reported. No between group diff on demographics Social anxiety: High SA > Low SA on FNE, anxiety in general and new situations Interoceptive sensibility: ↑ interoceptive sensibility = ↑ state self-focussed attention (r = .78, p < .001) and trait self-focussed attention (r = .35, p = .002) Interaction: ↑ SA = ↑ interceptive sensibility (r = .61, p < .001) Association between ↑ SA and ↓ self-rated performance was fully mediated by ↑ interoceptive sensibility ↑ SA endorsed ↑ interoceptive sensibility in bodily temperature, muscle tension, and heart rate | 0.92 |
E. Safety Behaviours (n = 3) | |||||||
Lei et al. (2023); UK | ASD: 61 (17) Non-ASD: 54 (11) Clinic; School/University; Online | Clinical diagnosis AutSev (ASD): AQ-28 77.93 (10.23) | Age: ASD: 16.34 (1.69); Non-ASD: 16.02 (1.56) IQ: none reported Ethnicity: ASD: 50 (81.97); Non-ASD: 37 (68.52) Co-occurring: 9 ADHD, 22 GAD, 19 SAD, 21 OCD, 1 Panic, 2 PTSD, 17 Depression, 6 Eating disorder | SPIN (self) | ASBQ (self) | Covariates controlled for: GAD and depression symptom severity, groups matched on SA Social anxiety: ASD and Non-ASD matched Social behaviours: ASD and Non-ASD matched on avoidance and impression management Interaction: ASD: ↑ SA = ↑ Safety behaviours total (r = .63, p < .001), including ↑ avoidance (r = .71, p < .001) and ↑impression management (r = .44, p < .001) | 0.91 |
Perry et al. (2015) | ASD: 13 (12) Non-ASD: 13 (13) Community | Clinical diagnosis, ADOS, ADI-R AutSev: none reported | Age: ASD: 25 (1.24); NT: 24 (0.46) IQ: Not reported Ethnicity: not reported Co-occurring: none reported | LSAS (self) | Social avoidance: interpersonal distance measured by stop-distance paradigm (self) | Covariates controlled for: none reported Social anxiety: ASD and non-ASD matched Social avoidance: ASD and non-ASD matched on mean preferred distance (in meters); ASD > non-ASD on variance on preferred distance Interaction: ↑ SA = ↑ Preferred interpersonal distance in ASD (r = .59, p < .05) but not in non-ASD group | 0.77 |
Wilson et al. (2023); UK | ASD: 192 (56) Non-ASD: 69 (26) Self-identify ASD: 51 (14) Broad Aut Phenotype (BAP): 24 (9) Community; Online | Clinical diagnosis; AQ AutSev (ASD): AQ-10 7.89 (1.79) | Age: ASD: 40.95 (13.62); N-ASD: 40.30 (17.28); S-ASD: 41.98 (13.95); BAP: 32.82 (11.16) IQ: none reported Ethnicity: ASD: 165 (86); N-ASD: 61 (88); S-ASD: 42 (82); BAP: 17 (71) Co-occurring: 58 SAD, 36 ADHD, 13 dyslexia, 15 dyspraxia, 9 language/learning-related diagnosis | LSAS (self) | SBQ (self) | Covariates controlled for: none reported; no association between outcome variable sand age/gender Social anxiety: ASD > non-ASD (Cohen’s d = .1.01) Safety behaviours: ASD > non-ASD (Cohen’s d = 0.9) Interaction: ASD < non-ASD in association between safety behaviours and SA ASD: ↑ SA = ↑ Safety behaviours (r = .54, p < .05) | 0.95 |
F. Social Experiences (n = 6) | |||||||
a) Peer Victimisation and Social Problems (n = 3) | |||||||
Ambler et al. (2015); Australia | ASD: 52 (42) non-ASD: 52 (42) School | Clinical diagnosis of Asperger’s, high functioning autism, or autistic disorder without intellectual disability AutSev: none reported | Age: ASD: 14.5 (1.77); non-ASD: 14.35 (1.68) IQ: none reported Ethnicity: none reported Co-occurring: none reported | RCMAS-2 – social anxiety scale (self) | School behaviours: SBS (teacher) | Covariates controlled for: none reported; groups matched on gender and age Social anxiety: ASD > non-ASD Social problems: ASD > non-ASD Interaction: ASD: ↑ SA = ↑ social problems (r = .29, p < .05) (i.e. avoids social interaction in class, ignored/rejected by peers) Non-ASD: no significant association between SA and social problems (r = .17, p > .05) | 0.77 |
Ung et al. (2016); USA | ASD: 81 (62) Clinical site, community | Clinical diagnosis AutSev: none reported | Age: 11.91 (2.32) IQ: 104.1 (14.24) Ethnicity: 63 (77.8) Co-occurring: none reported | SASC-R (self) | PEQ-R (self, caregiver) | Covariates controlled for: none reported Peer victimisation: In past 12 months, self-report of peer victimisation on average “a few times”, relational victimisation on average “a few times” and overt victimisation on average “a few times”. Cyberbullying reported by 9 youths as happening on average “a few times a week”. No sex differences Interaction: Child reports: ↑ peer victimisation = ↑ social avoidance (b = .13, p < .05), not with fear of negative evaluation (FNE; b = .03, p = .79) Parent reports: ↑ peer victimisation = ↑ social avoidance (b = .14, p < .05), not with FNE (b = .07, p = .14) | 0.77 |
Van Schalkwyk et al. (2018); USA | ASD: 35 (23) Clinical site, school | Clinical diagnosis by DSM-IV or DSM-5 AutSev: SRS-2 total 87.3 (20.9) | Age: 16.4 (1.58) IQ: none reported Ethnicity: none reported Co-occurring: none reported | MASC-2-P (caregiver) MASC-2-C (self) | Peer victimisation: report on bullying/victimisation (caregiver, self) | Covariates controlled for: none reported Bullying: parent- and child-rated bullying showed non-significant negative association (r = -.27, p = .156) Interaction: ↑Parent rated bullying = ↑child-rated SA (r = .56, p = .001) and parent-rated SA (r = .37, p = .038) Child rated bullying not associated with child (r = -.34, p = .06) or parent rated (r = -.21, p = .24) SA | 0.68 |
b) Adverse Childhood Experiences (n = 1) | |||||||
Wilson et al. (2023); UK | ASD: 192 (56) Non-ASD: 69 (26) Self-identify ASD: 51 (14) Broad Aut Phenotype (BAP): 24 (9) Community; Online | Clinical diagnosis; AQ AutSev (ASD): AQ-10 7.89 (1.79) | Age: ASD: 40.95 (13.62); N-ASD: 40.30 (17.28); S-ASD: 41.98 (13.95); BAP: 32.82 (11.16) IQ: none reported Ethnicity: ASD: 165 (86); N-ASD: 61 (88); S-ASD: 42 (82); BAP: 17 (71) Co-occurring: 58 SAD, 36 ADHD, 13 dyslexia, 15 dyspraxia, 9 language/learning-related diagnosis | LSAS (self) | VEQ-subset (self) | Covariates controlled for: none reported; no association between outcome variable sand age/gender Social anxiety: ASD > non-ASD (Cohen’s d = .1.01) Interaction: Negative early social experiences did not explain additional variances between autism traits and greater SA; SA not significantly associated with adverse childhood social experiences (r = .22) | 0.95 |
Authors (Year, Country) | Sample Size (N, male) Recruitment Channel | Diagnosis (Criteria/ Measure) | Age (M; SD) IQ (M; SD) Ethnicity (Caucasian N; %) Co-occurring conditions in ASD group (N; %) | Social anxiety measure | Mechanism measure | Outcome | Quality Score |
---|---|---|---|---|---|---|---|
A. Social Anhedonia (Social Situation, n = 1) | |||||||
†Gadow et al. (2020); USA | ASD: 268 (233) Non-ASD psychiatry referrals: 641 (447) Clinical site | Clinical diagnosis by DSM-IV AutSev: SCQ | Age: ASD: 10.5 (3.3); Non-ASD: 12.2 (3.4) IQ (n/% < 70): ASD: 71 (26.4%); Non-ASD: 18 (3.1%) Ethnicity: 229 (81.5) Co-occurring: none reported | CASI-4R Social Anxiety subscale (parent) | Social motivation: CASI-4R Social Anhedonia subscale (parent) | Covariates controlled for: age, gender, intellectual disability Social anxiety: ASD and non-ASD with ↑ Social Anhedonia = ↑ SA; Social anhedonia + preference to be alone > social anhedonia – preference to be alone (ASD: d = .53; non-ASD: d = .82) Social anhedonia: 32% of autistic youths with social anhedonia preferred to be alone, compared to 7% of non-autistic youths Interaction: SA contributes to odds of meeting criteria for social anhedonia even when accounting for reciprocal social interaction skills (OR = 1.46, p < .001) For both groups: ↑ SA = ↑ Social Anhedonia (r = .38, p < .001) | 0.95 |
B. Intolerance of Uncertainty (Social Situation to Activate Assumptions, n = 1) | |||||||
Pickard et al. (2020); UK | ASD: 61 (42) NT: 62 (26) School | Clinical and school diagnosis AutSev (ASD): SRS-2 77.47 (10.65); SCQ lifetime 20.94 (6.98) | Age: ASD: 13.46 (1.77); NT: 13.52 (1.57) IQ: ASD: 98.16 (13.99); NT: 100.76 (11.55) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | IUS (self, caregiver) | Covariates controlled for: sex Social anxiety: 52.4% of ASD vs. 54.8% NT > SA cut-off IU-Parent: ASD > NT (r = .49, p < .001) Interaction: ASD—IU-child: ↑ SA = ↑ IU (r = .71, p < .001), IU not associated with autism traits (r = .14, p > .05) NT – IU-child: ↑ SA = ↑ IU (r = .62, p < .001), ↑ IU = ↑ autism traits (r = .41, p < .001) ASD – IU-parent: ↑ SA = ↑ IU (r = .53, p < .001), ↑ IU = ↑ autism traits (r = .46, p < .001); IU mediated association between SA and Autistic traits NT – IU-parent: ↑ SA = ↑ IU (r = .40, p < .01), ↑ IU = ↑ autism traits (r = .75, p < .001) Stronger association between autistic traits and parent-report IU in NT vs ASD (Z = -2.53, p = .011) | 0.91 |
C. Autistic Identity (Activate Assumptions, n = 1) | |||||||
ASD: 121 (82) Community | Clinical diagnosis; SRS-S AutSev: SRS-S: 18.5 (6.1) | Age: 17.6 (1.1) IQ: not reported Ethnicity: 109 (90) Co-occurring: none reported | SAS-A (self) | Adapted version of Social Identification scale (Leach et al., 2008) (self) | Covariates controlled for: age, gender, autism traits Social anxiety: ↑ Autism traits = ↑ SA (r = .57, p < .001) Autism Identity: ↑ Autism traits = ↑ central autism identity (r = .34, p < .001) and ↑ solidarity with other autistic people (r = .26, p < .01) Interaction: Accounting for gender, age, autism traits: ↑ Autism satisfaction = ↓ SA (ß = -0.2, p < .05) Accounting for age and gender: ↑ SA = ↑ Autism Centrality (r = .22, p < .05), not with autism solidarity (r = .15, p > .05), autism satisfaction (r = .19, p > .05), autism self-stereotyping (r = .15, p > .05), or autism homogeneity (r = .06, p > .05) | 0.95 | |
D. Attentional Bias / Threat Monitoring (Activating Assumption to Perceived Social Danger, n = 8) | |||||||
Corden et al. (2008); UK | ASD: 21 (16) Non-ASD: 21 (16) Community | Clinical diagnosis; ADOS, ADI-R, 3Di AutSev (ASD): ADOS: 9.8 (3.75) | Age: ASD: 33.8 (13.6); Non-ASD: 32.1 (11.58) IQ: ASD: 117.9 (11.67); Non-ASD: 117.2 (8) Ethnicity: none reported Co-occurring: none reported | SPAI (self) | Ekman-Friesen test of facial affect recognition (self) | Covariates controlled for: none reported; matched on age, IQ, visual-perceptual ability Social anxiety: ASD > non-ASD Facial Recognition: ASD < non-ASD: recognising fearful and sad faces Interaction: ASD: ↑ SA = ↓ fear recognition (r = -.51, p = .04), and ↓ time spent fixating on eyes (r = -.5, p = .04) Non-ASD: no significant interaction | 0.82 |
Hollocks et al. (2013); UK1 | ASD: 38 (38) Control A: 21 (21) Control B: 20 (20) School | Clinical diagnosis and enrolment in specialist ASD school AutSev: SCQ 22 (5.6) | Age: ASD: 12.9 (1.4); Control A: 13.6 (1.2); Control B: 14.2 (1.8) IQ: ASD: 95.6 (13.3); Control A: 107 (9.99); Control B: 117 (8.89) Ethnicity: none reported Co-occurring: none reported | SCAS-P (caregiver) SCAS-C (self) | Attention: Dot probe emotional faces/words Emotion recognition: affect recognition & inhibition from NEPSY-II | Covariates controlled for: IQ Social anxiety: ASD < Control A and Control B on parent-rated SA Attention: ASD slower than Control A and B on face and word tasks, but no difference in attention bias Interaction: Parent rated SA not associated with threat face bias (r = .04, p = .85), social word bias (r = .13, p = .53), or physical word bias (r = .09, p = .67) | 0.77 |
Hollocks et al. (2016); UK1 | ASD: 21 (21) ASDAnx: 34 (34) Non-ASD: 28 (28) Clinical site | Clinical diagnosis, ADI-R/ADOS, SCQ AutSev: SCQ 19.4 (5.7) for ASD, 24.7 (5.8) for ASDAnx | Age: ASD: 13 (1.9); ASDAnx: 12.7 (1.9); Non-ASD: 13.9 (1.8) IQ: ASD: 103 (16.7); ASDAnx: 99.7 (10;9); Non-ASD: 116 (9.5) Ethnicity: none reported Co-occurring: 62% ≥ 1 anxiety; 18% ≥ 3 anxiety; 21 panic/agoraphobia, 22 GAD, 14 separation, 2 simple phobia, 4 SAD, 11 OCD | SCAS-P (caregiver) SCAS-C (self) | Attention: Dot-Probe (self) Emotion recognition: emotional face and words (self) | Covariates controlled for: IQ Social anxiety: ASDAnx > ASD > non-ASD Attention: ASDAnx > ASD, non-ASD on threat-bias towards angry faces Interaction: ASD and ASDAnx groups: ↑ SA (child) = ↑ threat face bias (r = .40, p = .003) SA rated by parent and child not associated with social threat words (r = .05 to .08, p > .05), physical threat words (r = -.17 to -.04, p > .05) | 0.86 |
*Kanat et al. (2017); Germany | ASD: 29 (29) NT: 30 (30) Clinical site | Clinical diagnosis by DSM-IV, ADOS-4 AutSev (ASD): AQ 37.8 (8.1) | Age: ASD: 38.2 (10.6); NT: 32.1 (12.3) VIQ: no intellectual disability Ethnicity: none reported Co-occurring: 18 depression, 7 ADHD, 6 anxiety disorders | SIAS (self) | Attention: dot-probe task using face/house (self) | Covariates controlled for: social anxiety symptom severity Intervention: self-administered 3 puffs of oxytocin or placebo nasal spray 45 min prior to computer task Interaction: ASD-Placebo: ↑ SA = ↓ attention to faces (r = -.56, p = .002) ASD-oxytocin: SA not associated with attention to faces (r = -.2, p = .289), significant decrease from placebo condition (z = -1.5, p = .07) NT: weak negative association between SA and attention to faces under placebo (r = -.03, p = .12), no association following oxytocin (r = -.56, p = .79) | 0.82 |
*Kang et al. (2019); USA | ASD: 53 (38) Clinical site | Clinical diagnosis, ADOS-2 AutSev: none reported | Age: 11.6 (2.96) IQ: 103.49 (15.4) Ethnicity: none reported Co-occurring: none reported | MASC-2-P (caregiver) MSAC-2-C (self-report) | Threat monitoring: Flanker task (error-rated negativity; ERN) (self) | Covariates controlled for: baseline scores and intervention condition Intervention: 1.5 h/week ten-week RCT of a group Social Skills Intervention (treatment vs. active control). Treatment included theatre-based activities. Control included group games or projects Social anxiety: Intervention group: 39% parents and 52% young people-reported significant ↓ in SA Interaction: Baseline: ↑ SA = ↑ negative ERN (r = -.345, p = .17), ↑ negative ERN predicted ↑ improvement | 0.85 |
*Luxford et al.(2017); UK | ASD: 35 (31) INT: 18 WLC: 17 School | Clinical diagnosis AutSev: SCQ INT: 18.61 (4.33) WLC: 19.06 (4.94) | Age: 13.2 (1.1) IQ: INT: 105.44 (17.83); WLC: 102 (11.3) Ethnicity: none reported Co-occurring: none reported | SWQ-C (Self) SWQ-T (teacher) | Attention control: Erikson flanker task Attention to threat: Emotional stroop colour matching schematic face task (self) | Covariates controlled for: baseline anxiety for pre-post analyses Intervention: ‘Exploring Feelings’ manualised CBT – 6 weekly 90 min sessions in small groups of 4–6 participants in school, weekly homework, 6-week follow-up Social anxiety: INT = WLC: self-report ↓ SA at end of intervention and follow-up INT > WLC: teacher-report ↓ SA at end of intervention and follow-up Attention control: INT < WLC on conflict score; Attention to threat: both groups—angry bias > happy Interaction: T1: neither self nor teacher-reported SA were associated with attention control (r = -.08 to -0.06, p > .05), attention to threat for happy bias (r = -0.18 to .04, p > .05), angry bias (r = -0.01 to .18, p > .05), fear bias (r = -0.09 to .21, p > .05) | 0.79 |
May et al. (2015); Australia | ASD: 45 (22) NT: 45 (22) Clinical site, community | Clinical diagnosis by DSM-IV-TR, SRS AutSev (ASD): SRS 100.1 (29.7) | Age: ASD: 10.27 (1.55); NT: 9.86 (1.55) VIQ: ASD: 99.2 (14.1); NT: 106.9 (14.3) Ethnicity: none reported Co-occurring: none reported | SCAS Social Phobia scale (Parent) | Threat bias: visual dot probe emotional faces (self) | Covariates controlled for: VIQ Social anxiety: ASD group scored similarly to children with anxiety disorders (n = 333) in Nauta et al. (2004) study Attention/Threat bias: No differences between groups on threat or happy congruent/incongruent/neutral trials Interaction: No significant association between SA and facilitated or disengagement from either threat or happy faces on the dot probe task for ASD and NT group | 0.86 |
Spain et al. (2016); UK | ASD: 50 (50) Clinical site, community (from a larger study) | Clinical diagnosis by ICD-10, ADI-R, ADOS-G, AQ AutSev: ADOS-G total 10.2 (4.7) | Age: 26.3 (5.8) IQ: 108 (14.9) Ethnicity: none reported Co-occurring: none reported | LSAS-SR (self) BFNE (self) SPS (self) SIAS (self) | KDEF (self) | Covariates controlled for: none reported Social anxiety: 26 scored above cut-off on LSAS-SR Interaction: No significant correlations between any of the SA measures and any socio-emotional tests (r < .24, p > .05). Comparing those who scored > and < SA cut-off, no differences in performances on any socio-emotional tasks | 0.82 |
White et al. (2015); USA | ASD: 15 (8) NT: 18 (10) Clinical site, Community | Clinical diagnosis; ADOS, ADI-R AutSev (ASD): SRS total 87.53 (16.69) | Age: ASD: 14.88 (1.55); NT: 14.33 (1.52) IQ: no ID reported by parents Ethnicity: ASD: 12 (80); NT: 17 (94.4) Co-occurring: 6 SAD | SWQ-P (caregiver) SWQ-C (self) BFNE (self) | Attention/threat-bias: gaze patterns for positive/negative face (self) | Covariates controlled for: autism traits Social anxiety: ASD and non-ASD group matched Interaction: ASD: ↑ FNE (child-report) = ↑ gaze to anger (r = .75, p < .01) and ↑ gaze to disgust (r = .68, p < .01) faces, but not happy faces (r = .4, p > .05) NT: no association between FNE and gaze to anger faces (r = .33, p > .05), disgust faces (r = .25, p > .05) or happy faces (r = .16 p > .05) | 0.91 |
E. Assertiveness (Fear of Negative Evaluation; n = 3) | |||||||
Bellini (2004); USA2 | ASD: 41 (35) School | Clinical diagnosis AutSev: none reported | Age: 14.22 FSIQ: 99.94 (18.81) Ethnicity: none reported Co-occurring: none reported | SAS-A (self) MASC-Social Anxiety (humiliation/performance fears) (self) | SSRS (self, caregiver) | Covariates controlled for: none reported Social anxiety: 20 out of 41 > cut-off for SA; 5 out of 41 scored below low SA Interaction: • ↓ assertion = ↑ SA and distress in new (r = -.31, p < .05) and general situations (r = -.39, p < .01), but not fear of negative evaluation (r = -.13, p > .05) • ↓ assertion = ↑ performance fears (r = -.33, p < .05), but not with humiliation fears (r = -.25) | 0.77 |
Bellini (2006); USA2 | ASD: 41 (35) School | Clinical diagnosis AutSev: none reported | Age: 14.22 FSIQ: 99.94 (18.81) Ethnicity: none reported Co-occurring: none reported | SAS-A (self) | SSRS (self, caregiver) | Covariates controlled for: none reported Interaction: ↓ assertion = ↑ SA (β = -1.68, p < .01) | 0.64 |
Chang et al. (2012) | ASD: 53 Clinical site, community, school | Clinical diagnosis; ADI-R; ADOS AutSev: none reported | Age: 9.55 (1.73) IQ: none reported Ethnicity: none reported Co-occurring: 49 separation anxiety, 55 GAD, 54 SAD, 30 OCD | ADIS-C/P (clinician) | SSRS (parent) | Covariates controlled for: none reported Social anxiety: 92.5% > cut-off Interaction: ↑ SA = ↑ social skills difficulties including assertion (R2 = .12) and responsibility (R2 = .12) | 0.77 |
F. Emotion Regulation (Somatic/Cognitive Symptoms to Processing of Self as Social Object, n = 1) | |||||||
Swain et al. (2015); USA | ASD: 69 (49) Clinical site | Clinical diagnosis AutSev: SRS-2 69.67 (10.13 | Age: 20.5 (2) IQ: none reported Ethnicity: 41 (59.42) Co-occurring: none reported | SAS (self) | DERS (self) | Covariates controlled for: none reported Interaction: ↑ caregiver rated SA = ↑ difficulty in selecting effective emotion regulation strategies (r = .25, p < .05); ↑ emotion dysregulation (r = .25, p < .05) ↑ caregiver rated FNE = ↑ non-acceptance of negative emotions (r = .26, p < .05), ↑ difficulty with goal-directed behaviours for negative emotions (r = .26, p < .05), ↑ total emotion dysregulation (r = .26, p < .05) ↑ self-rated SA = ↑ non-acceptance of negative emotions (r = .61, p < .01), ↑ inability to act in goal-directed way (r = .47, p < .01), ↑ total emotion dysregulation (r = .61, p < .01) ↑ self-rated FNE = ↑ non-acceptance of negative emotions (r = .54, p < .01), ↑difficulty with goal-directed behaviour for negative emotions (r = .47, p < .01), ↑limited access to strategies for emotion regulation (r = .59, p < .01), ↑ total emotion dysregulation (r = .58, p < .05) | 0.86 |
G. Alexithymia (Emotion Regulation, n = 5) | |||||||
Albantakis et al. (2020); Germany | ASD: 122 (83); non-ASD: 62 (37); NT: 261 (37) Clinical site | DSM-5 criteria; ADOS-2 AutSev: ADOS: 7.01 (3.17) AQ: 36.25 (8.16) | Age: ASD: 33.46 (10.40); non-ASD: 35.15 (12.62); NT: 26.41 (7.8) IQ: none reported Ethnicity: none reported Co-occurring: none reported | LSAS (self) | TAS-20 (self) | Covariates controlled for: age, sex Social anxiety: ASD > non-ASD > NT Alexithymia: ASD > non-ASD > NT Interaction: ASD: alexithymia (step 1; R2 = 11.8%) no longer significantly associated with SA when accounting for autism traits (step 2; R2 = 10.3%). Autism traits (step 1; R2 = 21.1%) significantly associated with SA when accounting for alexithymia (step 2; R2 = 1%) Non-ASD: alexithymia (step 1; R2 = 30.1%) significantly associated with SA when accounting for autism traits (step 2; R2 = 3.6%). Autism traits (step 1; R2 = 17.4%) significantly associated with SA when accounting for alexithymia (step 2; R2 = 16.2%) NT: alexithymia (step 1; R2 = 17.2%) significantly associated with SA when accounting for autism traits (step 2; R2 = 6.2%). Autism traits (step 1; R2 = 15.5%) significantly associated with SA when accounting for alexithymia (step 2; R2 = 7.9%) | 0.91 |
Antezana et al. (2023); USA | ASD: 19 (19) Community | ADOS-2 AutSev: ADOS-2 comparison: 5 (1.79) | Age: 21.82 (3.83) FSIQ-2: 108.28 (16.47) Ethnicity: 16 (84.21) Co-occurring: none reported | LSAS (self) | TAS-20 (self) | Covariates controlled for: age, IQ Social anxiety: 13 out of 19 > cut-off Alexithymia: 7 out of 19 > cut-off Interaction: • ↑ SA = ↓ overall facial emotion recognition accuracy, more in the low IQ (79–106) group (r = -.78, p < .01) and not in high IQ (107–14) group (r = -.43, p = .25) • SA not associated with alexithymia (r = .31, p = .197) | 0.91 |
Pickard et al. (2020); UK | ASD: 61 (42) NT: 62 (26) School | Clinical and school diagnosis; SCQ-L and SRS-2 AutSev (ASD): SRS-2 77.47 (10.65); SCQ lifetime 20.94 (6.98) | Age: ASD: 13.46 (1.77); NT: 13.52 (1.57) IQ: ASD: 98.16 (13.99); NT: 100.76 (11.55) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | TAS-20 (self) | Covariates controlled for: sex Social anxiety: 52.4% of ASD vs. 54.8% NT > SA cut-off Interaction: ASD: ↑ SA = ↑ Alexithymia (r = .63, p < .001), and ↑ Alexithymia = ↑ autism traits (r = .36, p < .01) | 0.91 |
Ringold et al. (2022); USA | ASD: 57 (44) Developmental Coordination Disorder (DCD): 26 (15) NT: 53 (31) Clinical site, community, school | Clinical diagnosis, ADI-R, ADOS-2 AutSev: none reported | Age: ASD: 11.89 (2.29); DCD: 11.75 (2.31); NT: 11.75 (2.13) IQ: ASD: 107.51 (16.82); DCD: 109.69 (17.13); NT: 118.28 (13.74) Ethnicity: none reported Co-occurring: none reported | SCARED-P (Caregiver) | AQC (Caregiver) | Covariates controlled for: age, sex, IQ Social anxiety: ASD > DCD, NT Alexithymia: ASD > DCD, NT Interaction: ASD: ↑ sensory over-responsivity = ↑ alexithymia (r = .30, p < .01); ↑ sensory over-responsivity = ↑ SA (r = .47, p < .01) DCD: ↑Sensory over-responsivity = ↑ SA (r = .54, p < .01) No effect in TD group (r = .24) | 0.91 |
Wilson et al. (2023); UK | ASD: 192 (56) Non-ASD: 69 (26) Self-identify ASD: 51 (14) Broad Aut Phenotype (BAP): 24 (9) Community Online | Clinical diagnosis; AQ AutSev (ASD): AQ-10 7.89 (1.79) | Age: ASD: 40.95 (13.62); N-ASD: 40.30 (17.28); S-ASD: 41.98 (13.95); BAP: 32.82 (11.16) IQ: none reported Ethnicity: ASD: 165 (86); N-ASD: 61 (88); S-ASD: 42 (82); BAP: 17 (71) Co-occurring: 58 SAD, 36 ADHD, 13 dyslexia, 15 dyspraxia, 9 language/learning-related diagnosis | LSAS (self) | GAFS-8 (self) | Covariates controlled for: none reported; no association between outcome variable sand age/gender Social anxiety: ASD > non-ASD (Cohen’s d = .1.01) Alexithymia: ASD > non-ASD (Cohen’s d = 1.28) Interaction: Alexithymia did not account for additional variance between ↑ autism traits and ↑ SA. SA and Alexithymia’s association: r = .18 | 0.95 |
H. Hyper/Hypo Sensitivity (Emotion Regulation, n = 5) | |||||||
†Black et al. (2017); Canada | ASD: 39 (30) Non-ASD: 40 (16) Clinical site, school | Clinical diagnosis; ADOS, AQ-C AutSev (ASD group): AQ: 97.7 (13.6) | Age: ASD: 12.08 (2.63); Non-ASD: 11.03 (3.03) VIQ: ASD: 45 (13.3); Non-ASD: 51.9 (7.1) Ethnicity: none reported Co-occurring: none reported | SCAS-P (caregiver) | SSP (caregiver) | Covariates controlled for: gender, VIQ, PIQ Social anxiety: No difference between ASD and non-ASD groups Interaction: Hypersensitivity not associated with SA in ASD group (r = .26, p > .05) nor TD group (r = .07, p > .05) | 0.86 |
Ludlow et al. (2023); UK | ASD + Selective Mutism (SM): 38 (21) SM: 37 (11) | Clinical diagnosis (n = 32) or referred for ASD diagnosis (n = 6) AutSev (ASD): AASQ 23.76 (9.26) | Age: SM + ASD: 9 (4.39); SM: 10.51 (4.25) IQ: none reported Ethnicity: none reported Co-occurring: 38 selective mutism | RCADS-P Social Anxiety subscale (caregiver) | SP2 (caregiver) | Covariates controlled for: GAD symptom severity Social anxiety: SM + ASD > SM Sensory sensitivity: SM + ASD > SM on sensory seeking, sensitivity, avoidance, and low registration Interaction: Group differences on SA partially mediated by levels of sensory avoidance (β = .47, p < .001), even when controlling for GAD | 0.77 |
MacLennan et al. (2020); UK | ASD: 41 (28) Clinical site, community, online | Clinical diagnosis of ASD by DSM-IV or DSM-5; AQ AutSev: AQ 35.5 (5.8) | Age: 8.44 (2.86) IQ: 109.03 (16.92) Ethnicity: none reported Co-occurring: none reported | SCAS-P (caregiver) PAS (caregiver) | SPSI (caregiver) | Covariates controlled for: autism symptom severity; age and IQ did not affect significance of associations Social anxiety: 35 (62.9%) showed elevated symptoms Interaction: When controlling for autism severity, ↑ SA = ↓ hyporeactivity (Spearman’s rho = -.397, p < .05); SA not associated with hyperreactivity (Spearman’s rho = .08, p > .05) | 0.86 |
Pickard et al. (2020); UK | ASD: 61 (42) NT: 62 (26) School | Clinical and school diagnosis; SCQ-L and SRS-2 AutSev (ASD): SRS-2 77.47 (10.65); SCQ lifetime 20.94 (6.98) | Age: ASD: 13.46 (1.77); NT: 13.52 (1.57) IQ: ASD: 98.16 (13.99); NT: 100.76 (11.55) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | AASP (self) | Covariates controlled for: sex Social anxiety: 52.4% of ASD vs. 54.8% NT > SA cut-off Hyposensitivity: NT > SAD (d = .41, p < .05) Interaction: ASD-Hypersensitivity: ↑ SA = ↑ hypersensitivity (r = .62, p < .001), ↑ SA = ↑ autistic traits (r = .33, p < .05), hypersensitivity mediated association between SA and autistic traits NT-Hypersensitivity: ↑ SA = ↑ hypersensitivity (r = .67, p < .001), ↑ SA = ↑ autistic traits (r = .36, p < .01) ASD-Hyposensitivity: ↑ SA = ↑ hyposensitivity (r = .29, p < .05) NT-Hyposensitivity: no association between hyposensitivity and SA (r = .09) nor autistic traits (r = -.09) | 0.91 |
Ringold et al. (2022); USA | ASD: 57 (44) Developmental Coordination Disorder (DCD): 26 (15) NT: 53 (31) Clinical site, community, school | Clinical diagnosis, ADI-R, ADOS-2 AutSev: none reported | Age: ASD: 11.89 (2.29); DCD: 11.75 (2.31); NT: 11.75 (2.13) IQ: ASD: 107.51 (16.82); DCD: 109.69 (17.13); NT: 118.28 (13.74) Ethnicity: none reported Co-occurring: none reported | SCARED-P (Caregiver) | SSP-2 (Caregiver) SensOR (Caregiver) | Covariates controlled for: age, sex, IQ Social anxiety: ASD > DCD, NT Sensory differences: ASD > DCD, NT on sensory seeking (partial η2 = .32), avoidance (partial η2 = .46), sensitivity (partial η2 = .51), oversensitivity on tactile (partial η2 = .39) and auditory (partial η2 = .36) Interaction: ASD: ↑Sensory over-responsivity = ↑ SA (r = .47, p < .01), ↑ Alexithymia (r = .30, p < .01) DCD: ↑Sensory over-responsivity = ↑ SA (r = .54, p < .01), not Alexithymia (r = .18) NT: No effect between sensory over-responsivity and SA (r = .24), nor Alexithymia (r = -.17) | 0.91 |
I. Camouflaging (Safety Behaviours and Processing of Self as Social Object, n = 4) | |||||||
Hull et al. (2019); UK3 | ASD: 354 (108) Non-ASD: 478 (192) Clinical site, Community | Clinical diagnosis AutSev: BAPQ | Age: ASD: 41.93 (13.55); Non-ASD: 30.24 (13.72) IQ: none reported Ethnicity: none reported Co-occurring: none reported | LSAS (self) | CAT-Q (self) | Covariates controlled for: none reported Interaction: In ASD group: ↑ SA = ↑ camouflaging (r = .44, p < .001), ↑ compensation (r = .30, p < .001), ↑ masking (r = .19, p < . 01), ↑ assimilation (r = .60, p < .001) In non-ASD group: ↑ SA = ↑ camouflaging (r = .60, p < .001), ↑ compensation (r = .46, p < .001), ↑ masking (r = .35, p < . 001), ↑ assimilation (r = .69, p < .001) | 0.86 |
Hull et al. (2021); UK3 | ASD: 305 (104) Clinical site | Clinical diagnosis AutSev: BAPQ 4.32 | Age: 41.9 IQ: none reported Ethnicity: none reported Co-occurring: 25% have 1 dx, 14% with 2 dxs, 9% ≥ 3 dxs; 173 GAD, 166 depression, 7 SAD | LSAS (self) | CAT-Q (self) | Covariates controlled for: age, autistic traits Interaction: Social camouflaging accounted for additional variance in SA symptom severity in a linear fashion (β = .21, p < .001), relationship did not differ by gender identity | 0.91 |
Lei et al. (2023); UK | ASD: 61 (17) NT: 54 (11) Clinical site, School/University; Online | Clinical diagnosis AutSev (ASD): AQ-28 77.93 (10.23) | Age: ASD: 16.34 (1.69); NT: 16.02 (1.56) IQ: none reported Ethnicity: ASD: 50 (81.97); NT: 37 (68.52) Co-occurring: 9 ADHD, 22 GAD, 19 SAD, 21 OCD, 1 Panic, 2 PTSD, 17 Depression, 6 Eating disorder | SPIN (self) | CAT-Q (self) | Covariates controlled for: GAD and depression symptom severity, groups matched on SA Social anxiety: ASD and Non-ASD matched Social camouflaging: ASD and Non-ASD groups did not differ on compensation and assimilation; ASD < non-ASD on masking Interaction: ASD: ↑ SA = ↑ camouflaging total (r = .51, p < .001), ↑ compensation (r = .47, p < .001), ↑ assimilation (r = .61, p < .001), but not with masking (r = .25, p = .051) | 0.91 |
Wilson et al. (2023); UK | ASD: 192 (56) Non-ASD: 69 (26) Self-identify ASD: 51 (14) Broad Aut Phenotype (BAP): 24 (9) Community Online | Clinical diagnosis; AQ AutSev (ASD): AQ-10 7.89 (1.79) | Age: ASD: 40.95 (13.62); N-ASD: 40.30 (17.28); S-ASD: 41.98 (13.95); BAP: 32.82 (11.16) IQ: none reported Ethnicity: ASD: 165 (86); N-ASD: 61 (88); S-ASD: 42 (82); BAP: 17 (71) Co-occurring: 58 SAD, 36 ADHD, 13 dyslexia, 15 dyspraxia, 9 language/learning-related diagnosis | LSAS (self) | CAT-Q (self) | Covariates controlled for: none reported; no association between outcome variable sand age/gender Social anxiety: ASD > non-ASD (Cohen’s d = 1.01) Camouflaging: ASD > non-ASD (Cohen’s d = 1.10) Interaction: Adding camouflaging did not explain additional variance between ↑ autism traits and ↑ SA. Camouflaging and SA’s association: r = .35 | 0.95 |
Data Synthesis
Results
Overview of Included Studies
Quality Appraisal
Participant Characteristics
Review Question 1: Research Evidence Underpinning Clark and Wells’ (1995) Cognitive Model of Social Anxiety
Negative Social Cognitions
Perceived Social Danger (including Fear of Negative Evaluation from others)
Processing of Self as Social Object (Self Focussed Attention)
Somatic and cognitive symptoms
Safety Behaviours
Social Experiences (Peer victimisation and adverse childhood experiences)
Review Question 2: Additional mechanisms associated with SA in autistic individuals beyond the Clark and Wells’ (1995) model
Authors (Year, Country) | Sample Size (N, male) Recruitment Channel | Diagnosis (Criteria/ Measure) | Age (M; SD) IQ (M; SD) Ethnicity (Caucasian N; %) Co-occurring conditions in ASD group (N; %) | Social anxiety measure | Mechanism measure | Outcome | Quality score |
---|---|---|---|---|---|---|---|
A. General Social Skills (n = 7) | |||||||
*Deckers et al. (2016); Netherlands4 | ASD: 52 (47) Waitlist: 26 (23) Soc Skills: 26 (24) Clinical site | Clinical diagnosis by DSM-IV-TR AutSev: none reported | Age: ASD: 10.1 (1.27); Waitlist: 10 (1.1); Soc Skills: 10.2 (1.43) IQ: not reported Ethnicity: none reported Co-occurring: 30 has co-occurring diagnosis, including 21 with ADHD | SCARED (caregiver) | SSO (caregiver, teacher) | Intervention: 12 weekly 1 h child sessions and three 1 h parent sessions. Each Social Skills Group has 4 autistic children, delivered by psychologist + co-therapist Covariates controlled for: none reported Social skills: Soc Skills > WLC on parent/teacher ratings of social skills improvement Interaction: ↑SA = ↓social skills (β = -2.56, p = .009). SA did not moderate change in social skills rated by teachers or parents | 0.83 |
Deckers et al. (2017); Netherlands4 | ASD: 73 (62) ADHD: 76 (54) Non-clinical: 106 (62) Clinical site, community, school | Clinical diagnosis by DSM-IV-TR AutSev (ASD): none reported | Age: ASD: 11.22 (2.42); ADHD: 11.79 (2.48); Non-clinical: 11.61 (2.63) IQ: > 70 Ethnicity: predominantly Caucasian Co-occurring: 1 adjustment disorder; 17 ADHD; 9 anxiety disorders; 2 eating disorders; 4 learning disorders; 1 mood disorders; 23 relational problems; 2 other disorder or diagnosis deferred | SCARED (caregiver) | SSO (caregiver, teacher) WSIS (caregiver, teacher) | Covariates controlled for: gender Social anxiety: ASD > ADHD, Non-clinical Social skills: ASD < ADHD, non-clinical (parent/teacher ratings) Interaction: ASD: ↑SA = ↓social skills rated by parents (r = -.47, p < .001) Clinical control: ↑SA = ↓social skills rated by parents (r = -.46, p < .001) Non-clinical control: no significant effect (r = -.47) | 0.91 |
Espelöer et al. (2021); Germany | ASD: 23 (17) SAD: 68 (28) NT: 25 (10) Clinical site, community | Clinical diagnosis by ICD-10 AutSev: none reported | Age: ASD: 44 (10.55); SAD: 37 (10); NT: 38.8 (10.41) IQ: ASD: 118.65 (15.58); SAD: > 80; NT: 110.28 (14.11) Ethnicity: none reported Co-occurring: none reported | SASKO—SA subscale (speaking and being in focus of attention; rejected by others) (self) | Social Skills: SASKO—Social Competence (interaction deficits, deficits in processing of social information) (self) | Covariates controlled for: gender Social anxiety: No group differences between ASD and SAD, both groups > NT on SA Social skills: ASD > SAD in difficulties in interaction and processing of social information (p < .001) | 0.82 |
*Kaboski et al. (2015); USA | ASD: 8 (8) Non-ASD: 8 (8) Community | Clinical diagnosis by DSM-5, ADOS-2, SCQ AutSev (ASD): SCQ-lifetime 17.9 (4.7) | Age: ASD: 14.05 (1.73); Non-ASD: 13.83 (1.45) IQ: ASD: 106 (18.56); Non-ASD: 112 (10.77) Ethnicity: none reported Co-occurring: none reported | SASC-R (self) SAS-A (self) | SSIS (parent) | Intervention: Two consecutive weeklong camps (3 h/day for 5 consecutive days) including career skills and robotics instructions, programming, and pair work Covariates controlled for: none reported, groups matched on age, gender, grade in school, IQ and language skills Social anxiety: ASD > non-ASD (d = 1.1), especially in SAD-General subscale (d = 1.57) ASD: significant ↓ in SA post-intervention (d = .74) Social skills: ASD < non-ASD on social skills (d = 2.7) at baseline ASD: no significant changes in social skills post-intervention (d = .17) Interaction: At baseline: ↓ social skills = ↑SA (r = -.55, p < .05), ↑SA in new situations (r = -.54, p < .05), ↑SA in general situations (r = -.59, p < .05), though not with fear of negative evaluation (r = -.40, p > .05) | 0.79 |
Scharfstein et al. (2011); USA | ASD: 30 (26) SAD: 30 (23) NT: 30 (22) Clinical site | Clinical diagnosis;, ADI-R AutSev: none reported | Age: ASD: 10.57 (1.6); SAD: 10 (1.8); NT: 10.6 (2) IQ: ASD: 114 (14.08) Ethnicity: ASD: 27 (90); SAD: 18 (60); NT: 11 (36.7) Co-occurring: none reported | SPAI-C (self) SAM (self) | BAT (observer) | Covariates controlled for: race (though results did not differ with/without covariate) Social anxiety: SAD > ASD, NT (partial η2 = .18) on self-reports and blind observer ratings (partial η2 = .18) Social behaviours: • SAD < ASD, NT on overall social skills (partial η2 = .12) • SAD < NT on pragmatic social behaviours (partial η2 = .12) • SAD < ASD, NT on appropriate speech and prosodic social behaviours (partial η2 = .17) • SAD and ASD < NT on vocal intensity (partial η2 = .43) • SAD < ASD and NT on vocal intensity variability (partial η2 = .18) • SAD > ASD on vocal pitch (partial η2 = .08) and vocal pitch variability (partial η2 = .09) Interaction: Vocal characteristics significantly contributed towards predicting group membership when accounting for observer-rated social skill differences (effect size = .98) | 0.92 |
Zukerman et al. (2020); Israel | ASD: 53 (49) University | Clinical diagnosis of ASD; AQ AutSev: AQ 43.79 (4.5) | Age: 23.53 (2.81) IQ: WAIS comprehension 87.97 (9.18) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | Social skills: Comprehension/Social adaptive behaviour gap (C-SOC) (self) | Covariates controlled for: age Interaction: ↑ SA = ↓ social adaptive behaviours (r = -.42, p < .01), and ↑ gap between comprehension and social adaptive behaviours (r = .40, p < .01) ↑SA avoidance = ↑ C-SOC gap (β = .49, p < .001) | 0.77 |
B. Neurocognitive Differences (n = 7) | |||||||
a) Empathy (n = 3) | |||||||
Bellini (2004); USA2 | ASD: 41 (35) School | Clinical diagnosis AutSev: none reported | Age: 14.22 FSIQ: 99.94 (18.81) Ethnicity: none reported Co-occurring: none reported | SAS-A (self) MASC-Social Anxiety (humiliation/performance fears) (self) | SSRS (self, caregiver) | Covariates controlled for: none reported Social anxiety: 20 out of 41 > cut-off for SA; 5 out of 41 scored below low SA Interaction: Empathy shows curvilinear (inverted U) relationship with Fear of Negative Evaluation (η2 = .40), and distress in new (η2 = .27), general (η2 = .33) situations, and performance fears (η2 = 0.19) | 0.77 |
Bellini (2006); USA2 | ASD: 41 (35) School | Clinical diagnosis AutSev: none reported | Age: 14.22 FSIQ: 99.94 (18.81) Ethnicity: none reported Co-occurring: none reported | SAS-A (self) | SSRS (self, caregiver) | Covariates controlled for: none reported Interaction: ↓ empathy = ↑ SA (β = -13.35, p < .01) | 0.64 |
Demetriou et al. (2020)5 | ASD: 62 (41) SAD: 83 (55) Early Psychosis: 48 (22) NT: 43 (22) Clinical site, community | Clinical diagnosis; ADOS-2 and ADI-R AutSev: ASD Group: SRS-2 Total = 72.64 (13.23) ASD + SAD Group: SRS-2 Total = 71.96 (13.12) | Age: ASD: 22.63 (5.55); SAD: 22.34 (6.15); EP: 23.08 (5.76); NT: 23.21 (5.84) IQ: > 70 Ethnicity: none reported Co-occurring: none reported | ADIS-IV/V (clinician) | EQ (self) | Covariates controlled for: none reported, no group differences in age/gender Empathy: ASD and SAD groups distinguished by overall levels of empathy | 0.91 |
b) Executive Function (including cognitive flexibility) (n = 4) | |||||||
Blakeley-Smith et al. (2012); USA | ASD: 63 (56) Clinical site | Clinical diagnosis; ADOS, ADI-R, SCQ AutSev: none reported | Age: 10.83 (1.72) VIQ: 107 (16.3) Ethnicity: 54 (85.7) Co-occurring: none reported | SCARED-P (caregiver); SCARED-C (self) | BRIEF (caregiver) | Covariates controlled for: none reported Social anxiety: Child and parent ratings showed non-significant, though moderate levels of agreement (r = .59) Interaction: Metacognitive ability significantly correlated with parent–child agreement on the SA domain (r = .33) | 0.86 |
Demetriou et al. (2018); Australia5 | ASD: 60 (38) SAD: 76 (41) Early Psychosis: 58 (37) NT: 59 (31) Clinical site, community | Clinical Diagnosis; ADOS AutSev: none reported | Age: ASD: 24.11 (7.27); SAD: 22.11 (5.64); EP: 21.79 (4.05); NT: 24.88 (5.3) IQ: ASD: 107.56 (8.66); SAD: 111.28 (6.56); EP: 100.71 (9.37); NT: 106.99 (8.3) Ethnicity: none reported Co-occurring: none reported | ADIS (clinician) | Cognitive flexibility: CANTAB-IED, TMT-B, COWAT, BRIEF (self) | Covariates controlled for: IQ, education Cognitive Flexibility: ASD > EP, SAD and NT on EF impairments, including psychomotor speed, mental flexibility, sustained attention, and fluency tasks | 0.91 |
Kimura et al. (2020); Japan | ASD: 33 (24) Non-ASD: 35 (21) Clinical site | Clinical diagnosis by DISCO AutSev (ASD): SRS 112.18 (31.56) | Age: ASD: 27.63 (6.18); Non-ASD: 28.03 (5.88) VIQ: ASD: 105.94 (11.04); Non-ASD: 106.46 (11.43) Ethnicity: none reported Co-occurring: none reported | LSAS (self) | Cognitive flexibility: letter fluency (self) | Covariates controlled for: none reported, matched on FSIQ and VIA Social anxiety: ASD > non-ASD on total, fear and avoidance SA scores Cognitive flexibility: ASD < non-ASD on total number of correct responses Interaction: ASD: ↑ correct responses on letter fluency = ↓SA (Spearman’s rho = -.38, p = .029) Non-ASD: ↑ correct responses on letter fluency = ↓SA (avoidance) (Spearman’s rho = -.491, p = .003), ↓SA (fear/anxiety) (Spearman’s rho = -.542, p = .001) | 0.82 |
Woolard et al. (2021); Australia | ASD: 62 (40) Employed: 57 Not employed: 5 Clinical site | Clinical diagnosis by DSM-5, ADOS-2 AutSev: ADOS-2 9.71 (2.75) | Age: 23.27 (6.79) IQ: 107.17 (10.07) Ethnicity: 59 born in Australia Co-occurring: none reported | LSAS (self) | Cognitive flexibility: TMT-B; BRIEF-A (self) | Covariates controlled for: IQ, ADOS-2, employment status, age Interaction: SA not associated with cognitive flexibility (r = .22, p > .05) | 0.86 |
Social Anhedonia
Intolerance of Uncertainty (IU)
Autistic identity
Attention Bias/Threat Monitoring
Assertiveness
Emotion and Sensory Processing
Emotion Regulation
Alexithymia
Sensory Differences
Camouflaging
Other factors considered in autism in relation to social anxiety
General Social Skills (beyond core autism specific social communication differences)
Neurocognitive differences
Discussion
Generalisability of Findings
Limitations and Considerations
Research Implications
Current Evidence (Section) | Suggested Future Research Directions | |
---|---|---|
H1. Interpretation of external social events in an excessively negative fashion | RQ1 S1 RQ1 S5 RQ2 S2 RQ2 S4 | • Explore how SA related social cognitions and interpretation bias may be associated with autistic individuals’ lived experiences and social comprehension skills • Explore whether intolerance of uncertainty may uniquely contribute towards activating negative interpretation bias and biased assumptions about one’s own performance in social situations when accounting for its associations with symptoms of generalised anxiety and non-social specific worries • Explore real-time interpretation of external social cues using experimental paradigms in addition to questionnaire-based outcome measures |
H2. Greater self-focussed attention when anxious in social situations | RQ1 S3 | • Examine the relationship between autistic identity and fear of negative evaluation amongst autistic individuals and explore whether this association may be partially mediated by one’s negative social attitudes, cognitions, and interpretation bias in social situations • Use experimental paradigms to look at manipulation of focus of attention in social interaction task, paying attention to how switching between internal and external focus of attention may be influenced by executive function differences such as cognitive flexibility, interoceptive sensibility, and sensory differences of both internal and external sensory cues |
H3. Reduced processing of external social cues when anxious | RQ1 S1 RQ2 S4 | • Use experimental paradigms such as look at differences in processing time for face versus non-social stimuli under conditions that elicit high versus low SA in autistic and non-autistic individuals matched on SA • Compare recall of internal versus external information following social and non-social interactions in autistic individuals with high vs. low levels of co-occurring SA |
H4. Generate distorted observer-perspective images of how they think they appear to others when in feared social situations | RQ1 S2 RQ2 S3 | • Use cognitive interview approach to explore how autistic individuals are interpreting questionnaires that look at fear of negative evaluation and self-imagery in relation to their autistic identity and SA, to ensure that the questionnaire has good construct validity in assessing SA related cognitions • Explore quality of self-imagery generated by autistic individuals based on internal bodily cues and internal focus of attention, and how such imagery may influence maintenance of SA in autistic individuals |
H5. Use internal information made accessible by self-focussed attention to make (erroneous) inferences about how they appear to others | RQ1 S3 RQ2 S6 | • Explore how alexithymia and sensory processing differences moderate the relationship between emotion regulation difficulties endorsed by autistic young adults and self-/caregiver-reported SA • Explore the relationship between somatic and cognitive symptoms of anxiety and processing of self as a social object in autistic individuals • Explore how perceived social danger may be associated with changes in somatic and cognitive symptoms of SA in autistic individuals, how such symptoms may be used to inform processing of self in social situations, and whether associations are moderated by emotion regulation strategies |
H6. In situation safety behaviour and self-focussed attention prevent disconfirmation of negative beliefs and maintain social phobia | RQ1 S4 RQ2 S7 | • Explore differences in motivations for adopting behaviours that look like safety behaviours from SA literature, for example, camouflaging and hiding autistic traits to try to fit in with non-autistic individuals may provide an alternative rationale for engaging in masking behaviours analogous to impression management behaviours |
H7. In situation safety behaviours and self-focussed attention can contaminate social interactions by making them less appealing to others | RQ1 S4 RQ2 S7 | • Explore autistic individuals’ self-perceptions and associated beliefs about the role of different safety behaviours related to SA, and camouflaging behaviours related to masking autistic traits • Explore how dropping safety behaviours (associated with social anxiety) and camouflaging behaviours (associated with masking autistic traits) may differentially influence observer ratings of social performance on tasks that evoke high versus low levels of SA in autistic individuals • Examine the differential effects on social anxiety related cognitions, anxious feelings, and ratings by others of safety behaviours (both impression management and avoidance) and camouflaging |
H8. Reduced processing of external social cues is biased in favour of detection and recall of cues that could be interpreted as signs of disapproval from others | RQ2 S4 | • Use experimental paradigms to look at autistic individuals’ ability to detect and recall positive and negative audience behaviours during and after a performance-based task that evokes high levels of SA. Explore whether ability is associated with co-occurring SA symptoms. A matched non-autistic group with similar levels of SA can address whether differences in detection/recall of positive versus negative audience behaviours is related to SA or potential social communication differences associated with autistic traits |
H9. Engage in negatively biased anticipatory processing before entering feared social situations | None to date | • Use experimental paradigms to explore recall of positive and negative words associated with others’ perception of self, one’s own perception of self, versus perception of others, following performance task that evokes high SA (e.g. giving a talk) versus distraction task • Qualitative studies to explore anticipatory processes autistic individuals engage in prior to social situations, paying close attention to potential overlap with camouflaging/masking behaviours • Explore how anticipatory behaviours affect one’s performance in tasks that evoke high versus low levels of SA |
H10. Engage in prolonged, negatively biased, post-event processing | None to date | • Explore recall of positive and negative self-related information after performance tasks that involve stressful social interactions or evoke high levels of SA • Qualitative interviews to explore content and nature of imagery associated with post-event processing after stressful social situations. Pay close attention to the function of behaviour and specificity to SA, as post-event processing may be related gaining understanding of social situation (i.e. as a compensating behaviour for social communication differences), or related to co-occurring conditions such as depression/generalised anxiety |