Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Concurrent and Prospective Associations Between Social Anxiety and Responses to Stress in Adolescence

Concurrent and Prospective Associations Between Social Anxiety and Responses to Stress in... Several studies have investigated the relationship between adolescents’ responses to stress and general anxiety and depres- sion, but only few studies addressed the relationship between responses to stress and social anxiety. The current three-wave longitudinal study, that covered a period of 5 years with a time interval of on average two years between waves, examined concurrent as well as prospective relations between adolescents’ self-reported stress responses, including coping responses, and self-perceived social anxiety. Both the predictive power of social anxiety for different stress responses and, reversely, of stress responses for social anxiety were evaluated. Participants were 331 youth (170 boys) aged 9 to 17 years old at Wave 1. Self-report questionnaires were used to measure social anxiety, responses to social stress, and depressive symptoms. Results showed signic fi ant concurrent relations between social anxiety and maladaptive stress responses. Moreover, the study yielded evidence for social anxiety predicting stress responses across time as well as stress responses predicting social anxiety, although evidence for the former link is stronger. The findings suggest that a relative lack of adaptive stress responses may heighten social anxiety and social anxiety in turn may trigger maladaptive as well as adaptive responses to social problems. The relevance of these findings for social anxiety prevention and intervention purposes are discussed. Keywords Adolescence · Social anxiety · Coping · Stress responses Adolescence is a period of significant changes in many If adolescents do not cope with social stress in an effec- domains of life, physiological, cognitive, emotional, and tive way, internalizing problems related to social anxiety social. In adolescents’ social life, relationships with peers may develop (Grant et al., 2004; Richey et al., 2019). In become more, and those with parents less, important turn, these problems may then further hinder the use of (Steinberg, 2010). Adolescents increasingly share their adequate stress responses possibly resulting in even more experiences and emotions with peers and rely on these social stress, anxiety symptoms, and ultimately social anxi- peers for social support (Furman & Buhrmester, 1992). ety disorder. Thus, adolescents’ responses to stress may They also become more sensitive to their own position be risk factors as well as consequences of social anxiety in the peer group. How they are evaluated by peers and problems (Compas et al., 2001; Wright et al., 2010). particularly if they are accepted by peers are important To our knowledge, no study to date has addressed the factors in the lives of adolescents (LaGreca & Prinstein, prospective relation between adolescent stress responses 1999; Sentse et al., 2010). This need for peer acceptance, and social anxiety, despite its importance for understanding for being “one of them”, may result in increasing social social anxiety development. The available studies used sam- stress levels (Ollendick & Hirshfeld-Becker, 2002; Van ples of children or early adolescents. Because adolescence den Bos et  al., 2014), especially in the case of adverse is a period in which social anxiety symptoms increase and peer experiences (Blöte et al., 2015; Reijntjes et al., 2010). social anxiety disorder has its onset (Magee et al., 1996; Wittchen & Fehm, 2003), this age group is of particular interest for studying the role of stress responses in relation * Anke W. Blöte to social anxiety development. The present study therefore bloete@fsw.leidenuniv.nl tried to fill the gap in the literature by paying special atten- Institute of Psychology, Unit of Developmental tion to the potential bi-directional relation between stress and Educational Psychology, Leiden University, responses and social anxiety problems in adolescents. Wassenaarseweg 52, 2333 AK Leiden, The Netherlands Vol.:(0123456789) 1 3 Research on Child and Adolescent Psychopathology Responses to stress are categorized as either voluntary or Clark & Wells, 1995; Spence & Rapee, 2016) pay only lim- involuntary (Compas et al., 1997; Lazarus & Folkman, 1984). ited attention to the relation of different coping styles with Voluntary responses – generally referred to as coping– are the development (Spence & Rapee, 2016) and maintenance regarded as under the person’s control and requiring conscious (Clark & Wells, 1995; Rapee & Heimberg, 1997) of social effort, whereas involuntary responses to stress are automatic, anxiety, although some coping strategies such as cognitive not under the control of the person concerned. With regards restructuring and safety behaviors are addressed. Cogni- to stress responses in youth, Compas et al. (2001) presented tive restructuring, that addresses the replacement of anxi- three dimensions along which the different stress responses ety laden biased cognitions with more realistic cognitions may differ; the just mentioned distinction between voluntary (McLellan et al., 2015), is a form of secondary engagement and involuntary responses characterizes the first dimension. coping. Safety behaviors such as avoiding eye contact and Voluntary responses are consciously aimed at changing the keeping quiet in interactions with others, are intended to stressful situation or one’s own response to it, be it cognitive, avoid aversive social experiences. As such they are a type behavioral, or emotional. Examples of voluntary responses of disengagement response. In the Spence and Rapee (2016) are among others, problem solving, seeking distraction, and model on social anxiety development in adolescents, both emotion regulation. Involuntary or automatic responses to negative social cognitions and safety behaviors are included stress, such as physiological arousal and rumination, are not as risk factors of social anxiety development. (directly) aimed at regulating the situation or one’s emotional Recently, a new model of adolescent social anxiety, Sen- responses to it and they may or may not be within conscious sitivity Shift Theory (SST), was presented that more explic- awareness. The second dimension makes a distinction based itly pays attention to the link between social anxiety and on the person’s engagement with or disengagement from the stress responses (Richey et al., 2019). SST describes the stressful situation. Engagement responses are responses that development from an inhibited temperament to social stress approach the stressful event or unpleasant emotion, whereas and social anxiety, and from there to social anhedonia. SST disengagement responses are oriented away from them and are emphasizes the importance of coping in this development characterized by avoidance. The third dimension is specific by stating that the success of socially anxious adolescents’ to voluntary engagement responses and addresses primary coping responses in reducing stress determines whether control versus secondary control strategies. Primary control they will continue to put effort into their coping with future coping is directed at changing the stressful situation or the stressors, or give up, stop expending energy and start to person’s emotional state caused by the situation, while sec- avoid social situations. When the latter occurs, the positive ondary control coping refers to the person adapting to the affect normally associated with social situations disappears, problem, for example by cognitive restructuring and accept- reducing the chance that successful coping strategies are ance of the situation or emotions involved. effectuated in future social situations and maintaining or Some stress responses are considered adaptive because even further increasing social anxiety. This final stage in they help to diminish the stress, and other stress responses the model describes a condition named social anhedonia that as maladaptive because they do not diminish it and conse- is primarily characterized by disengagement from stressful quently may lead to internalizing problems (Compas et al., social interactions. 2017; Connor-Smith et  al., 2000). Compas et  al. (2001; Empirical research investigating the concurrent rela- 2017) concluded, based on extensive literature reviews, that tion between adolescents’ stress responses and social anxi- engagement coping in youth is adaptive because it is asso- ety is limited and used relatively young participants from ciated with psychological adjustment and well-being. For late childhood and early adolescence. Furthermore, the example, studies found that engagement coping is negatively results of these studies yielded equivocal results. In the fol- related to internalizing problems in adolescents both when lowing review studies used self-report to assess the main using self-reported and parent-reported measures (Connor- variables; exceptions to this rule are explicitly noted. In a Smith et al., 2000) and engagement coping with peer vic- sample aged 8 t o11 years old, Wright et al. (2010) found timization in socially anxious adolescents is linked with a significant, although weak, positive link of social anxiety social competence (Kaeppler & Erath, 2017). In contrast, symptoms with problem solving and support seeking, both disengagement from stressful situations or one’s own feel- primary engagement coping responses, and a significant and ings of stress is linked with poor psychological adjustment strong positive link with involuntary engagement responses. and is therefore considered maladaptive. Richardson et al. (2020) reported a significant association The role of coping in theories of social anxiety seems of social anxiety symptoms with avoidance coping and a somewhat underrepresented. In light of the distinction significant but small negative association with problem between adaptive and maladaptive coping strategies and solving responses in 10 to 12-year olds. Parent-reported their links with emotional adjustment, it is a bit surprising measures yielded similar results. In contrast, Erath et al. to see that important models of social anxiety disorder (e.g., (2007) did not find significant correlations between social 1 3 Research on Child and Adolescent Psychopathology anxiety symptoms and either engagement or disengagement component in it. Social anxiety and depression share a num- responses in early adolescents (sixth and seventh graders). ber of characteristics with regards to risk factors, and associ- To our knowledge, only two prospective studies have been ated and consequent variables. Identifying what is specific conducted on the relation between social anxiety and stress to a disorder is important for several reasons, among them responses and these studies used children or young adoles- the development of treatment and prevention interventions cents, 8 to11 year olds in the Wright et al. (2010) study (Epkins & Heckler, 2011; Starr & Davila, 2008). It is also and 10 to12 year olds in the Richardson et al. (2020) study. possible that depression and social anxiety each have differ - The Wright et al. (2010) study found that over a period of ent relations with certain stress responses as the Wright et al. nine months, social anxiety symptoms predicted a rela- (2010) study made clear. This study revealed that depres- tive increase of social support seeking, a form of primary sion did not explain any relation between social anxiety and engagement coping, controlling for the effect of depression. stress responses. Remarkably, depression and social anxi- Social anxiety was also related to relative increases in invol- ety even had opposite effects on children’ s stress responses. untary responses (e.g., worrying, not sleeping). Specifically, For example, depression predicted a decrease of distraction for children who experienced lower peer rejection, social seeking whereas social anxiety predicted an increase in dis- anxiety predicted an increase in distraction seeking, a form traction seeking (for children who experienced lower peer of secondary control engagement coping. Stress responses, rejection). These opposing effects might have obscured the in turn, did not predict subsequent changes in social anxi- relationship between social anxiety and stress responses if ety symptoms. In contrast, Richardson et al. (2020) found the effect of depression had not been taken in account. This that social anxiety, controlled for depression, did not predict finding further stresses the importance of paying attention increases in the different coping responses (problem solv - to the role of depression in the link between social anxiety ing, social support seeking, and avoidant coping) measured and stress responses. one year later. Reversely, avoidant coping predicted relative increases in social anxiety symptoms. Parent-reported meas- ures did not corroborate this finding. The Present Study Two other studies examined the prospective links between stress responses and more general adjustment problems. One In sum, it is clear that adolescent anxiety and depression are study with 9–15 years old participants, three assessment associated with maladaptive responses to stress, but it is not waves over a period of two years, and general anxiety and clear if social anxiety shows a similar pattern of relations to depression as respective internalizing variables revealed sev- stress responses. Moreover, there is a gap in our knowledge eral predictive links from stress responses to anxiety (Flynn about the prospective links between anxiety/depression in & Rudolph, 2011). The combination of low engagement general and social anxiety in particular, with stress responses coping and high involuntary stress responses (both in rela- (Compas et al., 2017). Therefore, the present study sought tion to social situations) predicted relatively high levels of to answer the following question. What is the relationship anxiety and depression as diagnosed in interviews with the between social anxiety and responses to social stress (a) con- participants and their primary caregiver. Reversely, anxiety currently and (b) over time? and depression did not predict stress responses. Based on the literature about general anxiety and depres- The second study evaluated coping with poverty-related sion we expected that concurrently social anxiety symptoms family stress in adolescents with a mean age of 14 years are positively related to maladaptive stress responses (dis- (Wadsworth & Berger, 2006). This study, that used two engagement coping and involuntary stress responses), and time-points eight months apart, addressed self-reported negatively related to adaptive responses (engagement cop- emotional adjustment in general, without further distinguish- ing). There are as yet too few longitudinal studies to guide ing between anxiety and depression symptoms, and found any specific expectations for the prospective effects of social that emotional adjustment did not predict changes in coping anxiety symptoms and stress responses over time. Moreover, strategies, but that coping strategies did predict changes in the studies we know of yielded equivocal results. Theoreti- emotional adjustment. For adolescents with relatively high cally, we would expect that a low level of engagement coping initial family stress levels, primary control engagement cop- and a high level of disengagement coping and involuntary ing predicted higher emotional adjustment. For adolescents stress responses predict relative increases of social anxiety with relatively low initial adjustment, disengagement coping over time, because these responses will not help to solve the predicted even lower adjustment. social stress related problems and therewith prevent further Because of the high comorbidity between social anxiety distress related to these problems (Richey et al., 2019). and depression (Epkins & Heckler, 2011), it may not be clear Because disengagement coping (Silk et al., 2003; Wright whether links between social anxiety and stress responses are et al., 2010) and involuntary stress responses (Silk et al., specific to social anxiety or are explained by the depression 2003) not only have been linked to (social) anxiety but also 1 3 Research on Child and Adolescent Psychopathology to depression, links between social anxiety and these stress the timepoints. The T1 social anxiety of participants with responses may be caused by the depressive component in missing data caused by attrition or incidental events (that social anxiety. For this reason, we evaluated whether the is, all missing data except the ones related to age) did not putative links between stress responses and social anxiety significantly differ from that of participants with complete could at least partly be attributed to depression. data. A repeated measures ANOVA on social anxiety meas- ured at the three timepoints yielded a non-significant group effect between participants with and without missing data, Method F(1, 231) = 3.28, n.s., as well as a non-significant Group x Time interaction effect, F(2, 231) = 0.72, n.s. Design and Procedure Instruments Data were drawn from the Social Anxiety and Normal Development (SAND) study, a community study that The Dutch translation (H. Koot and E. Utens, unpub- selected students from two primary schools and one sec- lished) of the Social Anxiety Scale for Adolescents (SAS-A; ondary school in an urban area in the Netherlands (Miers LaGreca & Lopez, 1998) was used to measure social anxiety. et al., 2013; Westenberg et al., 2009). Data relevant to the The SAS-A contains 18 statements linked to social anxiety present study were collected at the first, third, and fourth (e.g., “I worry that others don’t like me”). Participants indi- waves (referred to as T1, T2, and T3, respectively) of this cate for each statement how true it is for themselves using four-wave longitudinal study. (At the second wave, only a a 5-point Likert scale (1 = not at all, 5 = all the time). Total restricted number of variables were measured). The time scores range between 18 and 90 with total score > 50 as the interval between these waves was two years on average with recommended criterion for clinically significant levels of intervals varying between one and three years. Severe psy- social anxiety (LaGreca, 1998). The SAS-A has shown good chological problems or physical illness as registered by the internal consistency with Cronbach’s alpha > 0.90 and cor- school were a contra-indication for participation in the study. relates strongly with other social anxiety measures (LaGreca Participants individually completed a battery of assess- & Lopez, 1998; Storch et al., 2004). Cronbach’s alpha for the ment forms at the university laboratory including the three three timepoints of the present study was > 0.92. In the data questionnaires used in the present study. The SAND study analysis, sum scores over the 18 items were used. was approved by the university’s Medical Ethical Commit- We measured depression with the Dutch translation tee. Parents gave their written consent and youth their writ- (Timbremont & Braet, 2002) of the Children’s Depression ten assent for participation in the study. Inventory (CDI; Kovacs, 1985). For each of the 27 items of the questionnaire participants indicate which of three state- Participants ments best describes how they felt in the last two weeks. For example, “I do most things OK”, “I do many things wrong”, The SAND study started at T1 with 331 primary and sec- and “I do everything wrong.” Scores range from 0 to 2 (most ondary school students, 170 boys and 161 girls (Miers et al., depressed). The Dutch version of the CDI has good internal 2013). Originally, 434 students from Grades 4 to 9 had been consistency (Cronbach’s alpha > 0.80) and shows a strong contacted of whom 75 students did not assent or their parents correlation with DSM-oriented depression measures (Roelofs did not consent to participation in the study. Other students et al., 2010; Timbremont & Braet, 2002). Total scores may (n = 28) did not participate for various other reasons, mainly range between 0 and 54 with a cut-off point of 16 for clinically because it was not possible to invite them to the laboratory significant cases. For ethical reasons, the item asking about within the available time slots. Out of the 331 students, 248 suicide was not presented to the participants. Data analysis and 236 (75% and 71%) respectively, still participated at was therefore based on the sum scores over 26 items. Cron- T2 and T3. At T1, participants’ ages ranged between 9 and bach’s alpha was 0.80, 0.84, and 0.83 for the three respective 17 years, M = 13.34, SD = 2.25. At T3, the mean age of the timepoints. participants was M = 17.48, SD = 2.72. The T3 sample con- The Responses to Stress Questionnaire (RSQ; Connor- sisted of 121 boys and 115 girls. Apart from the missing Smith et al., 2000) translated into Dutch by H. Ouwehand data caused by attrition at T2 and T3, other missing data (unpublished) was used to measure participants’ stress occurred for the responses to stress measure because it was responses. The RSQ offers the possibility to measure ado- not completed by primary school students (the measure was lescents’ stress responses related to different domains. In specifically designed for adolescents). This resulted in 126 the present study we referred to stressful situations in the age-related missing values on this measure at T1 with 32 of social domain, namely, having problems with other kids. those still missing at T2. Furthermore, seven participants For example, “When problems with other kids come up, had missing values on one of the three measures at any of I can't stop thinking about how I am feeling.” The RSQ 1 3 Research on Child and Adolescent Psychopathology has five scales: Primary control engagement coping (9 Results items covering problem solving, emotional regulation, and emotional expression responses), secondary con- Descriptive Analyses trol engagement coping (12 items covering acceptance, distraction, cognitive restructuring, and positive think- Means and SDs of the study variables at T1 are presented ing responses), disengagement coping (9 items cover- in Table 1. The table shows that the social anxiety level of ing denial, avoidance, and wishful thinking responses), the present sample is largely comparable to that of other involuntary engagement (15 items covering rumination, community samples (Ranta et  al., 2012; Storch et  al., intrusive thoughts, emotional and physiological arousal, 2004). The SAS-A cut-off score of 50 for clinical levels and impulsive action responses), and involuntary dis- of social anxiety (LaGreca, 1998) lies within 1SD above engagement (12 items covering emotional numbing, the mean indicating that high levels of social anxiety were inaction, escape, and cognitive interference). Items are not uncommon in the sample. The CDI mean is relatively scaled from 1 to 4 (1 = not at all, 4 = a lot), according to low (Roelofs et al., 2010). how often the participant says to use the voluntary cop- ing or experience the involuntary stress response. The RSQ scales of the social stress version have adequate internal consistency (Cronbach’s alpha’s ranging from Concurrent Relations 0.73—0.89). It was found that the primary engagement and disengagement coping scales correlate with the COPE At T1, social anxiety was strongly (Hemphill, 2003) posi- and the involuntary stress responses scales with heart- tively related to the three maladaptive responses to stress, rate reactivity (Connor-Smith et al., 2000). In the present namely disengagement coping, involuntary engagement, study, Cronbach’s alpha of the scales ranged from 0.71 and involuntary disengagement (see Table 1). Social anxi- for disengagement coping to 0.88 for involuntary engage- ety was not associated with primary or secondary control ment at T1, from 0.72 for secondary engagement coping engagement coping. Social anxiety and depression were to 0.89 for involuntary engagement at T2, and from 0.73 relatively strongly related and their respective correlations for secondary engagement coping to 0.89 for involuntary with the responses to stress variables were largely similar. engagement at T3. After controlling for the effect of depression, the correla- tions between social anxiety and the respective maladap- tive stress responses were still significant at a p < 0.001 Data Analysis level albeit the partial correlations were somewhat lower than the zero-order correlations. That is, the partial cor- First, descriptives and correlations between all study relations of social anxiety with disengagement coping, variables were computed for the different timepoints. involuntary engagement, and involuntary disengagement Second, cross-lagged panel analysis was used to test a were r = 0.459, r = 0.546, and r = 0.429, respectively, com- model describing the effect of social anxiety at T1 on pared to the zero-order correlations r = 0.592, r = 0.656, participants’ stress responses at T2, of social anxiety at and r = 0.572. T2 on participants’ stress responses at T3, and reversely, Age at T1 was significantly correlated with primary of the stress responses of T1 and T2 on the respective engagement coping, r = 0.170, p < 0.05. (At T3, it was social anxiety levels at T2 and T3. In order to control for significantly correlated with disengagement coping, r = depression, we also tested a model in which depression -0.143, p < 0.05, and at T4 both with primary engagement levels measured at the three time-points were added in coping, r = 0.143, p < 0.05, and disengagement coping, all regressions either as a dependent or independent vari- r = -0.134, p < 0.05). Controlling for age did not affect able together with social anxiety. Finally, because the age the other correlations, as the partial correlations of social range of the participants was relatively large, we tested anxiety with disengagement coping, involuntary engage- a model that controlled for age. The cross-lagged panel ment, and involuntary disengagement show, respectively, analyses were performed with Lavaan (Rosseel, 2012) in r = 0.587, r = 0.650, and r = 0.586. R (R-Core-Team, 2019). With regards to missing data, full The two engagement responses were positively interre- information maximum likelihood estimation (Schafer & lated. Primary control engagement coping was also signifi- Graham, 2002) was employed thus all available data were cantly positively associated with involuntary engagement. used. As goodness of fit criteria we used a comparative fit Furthermore, the three maladaptive stress responses, dis- index (CFI) of around 0.95 and root-mean-square error of engagement coping and involuntary engagement and dis- approximation (RMSEA) of around 0.05 (Kline, 2005). engagement, were strongly interrelated. At T2 and T3, the 1 3 Research on Child and Adolescent Psychopathology Table 1 Ms (SDs), and M (SD) n SAS CDI Prim Sec Dis InvE InvD Correlations of Study Variables ** ** ** ** at T1 SAS 40.671 - 0.502 0.003 0.019 0.592 0.656 0.575 (12.750) n = 328 ** ** ** CDI 8.881 - -0.054 -0.062 0.489 0.483 0.508 (5.361) n = 327 ** ** Prim 22.427 - 0.316 -0.069 0.276 0.068 (4.922) n = 205 ** Sec 28.259 - 0.198 0.030 0.111 (5.300) n = 203 ** ** Dis 15.499 - 0.545 0.618 (3.691) n = 205 ** InvE 25.515 - 0.707 (6.886) n = 204 InvD 18.512 - (4.676) n = 204 SAS social anxiety scale, CDI children’s depression inventory, Prim primary engagement, Sec secondary engagement, Dis disengagement, InvE involuntary engagement, InvD involuntary disengagement ** Correlation is significant at the 0.01 level (2-tailed) pattern of correlations between the variables was mainly (compared to adolescents with higher primary control engage- similar to that at T1 (see supplementary Tables 1 and 2). ment coping) of T2 social anxiety. Reversely, T1 social anxiety predicted a relative increase of secondary engagement coping and disengagement coping at T2. From T2 to T3, none of the Relations Over Time stress responses predicted social anxiety, but social anxiety predicted a relative increase of all four stress responses. The Because the correlation between involuntary engagement and significant effects in the model are depicted in Fig.  1. disengagement was very high, with rs > 0.70 at the different The second cross-lagged analysis that added depression to time points, and we wished to improve the number of partici- the model, showed that the path from T1 social anxiety to T2 pants/number of variables ratio in the cross-lagged panel anal- disengagement coping was no longer significant (see Table  3 yses, we chose to exclude involuntary disengagement from of the supplementary tables). Instead, depression predicted these analyses. Based on attention received in the literature to disengagement coping. In other respects, the pattern of results date, involuntary engagement, which includes rumination and was unaffected by the addition of depression to the model. emotional and physiological arousal, is the most important. From T2 to T3, as compared with from T1 to T2, social anxi- It should be noted that the involuntary disengagement scale ety predicted disengagement coping even after depression was was added to the RSQ purely for model-theoretical reasons, added to the model. namely, to have an engagement-disengagement dimension that The third model that included age as covariate (see Table 4 applies to voluntary as well as involuntary responses (Connor- of the supplementary tables) did not meaningfully change the Smith et al., 2000). results as found in the first model. The first cross-lagged panel model (that did not include depression or age) had a good fit, CFI = 0.978 and RMSEA = 0.055. The results showed rather large autoregres- Discussion sive effects of social anxiety as well as the different stress responses indicating that individual differences on these vari- This study found that adolescents’ self-perceived social ables are relatively stable over time (see Table 2). Low primary anxiety and stress responses are linked both concurrently control engagement coping at T1 predicted a relative increase as well as over time. Social anxiety showed concurrent 1 3 Research on Child and Adolescent Psychopathology Table 2 Standardized Estimates responses (Compas et al., 2001; 2017) was replicated spe- Standard P of Cross-lagged Analysis cifically for social anxiety. estimates Prospectively, we found bi-directional effects between T2SAS social anxiety and stress responses. That is to say, initially   T1SAS 0.465 0.00 from T1 to T2, low primary control engagement coping   T1Prim -0.173 0.01 (e.g., low problem solving, low emotion regulation) pre-   T1Sec -0.027 0.66 dicted social anxiety, and social anxiety predicted second-   T1Dis -0.057 0.47 ary engagement coping (e.g., distraction seeking, cognitive   T1InvE 0.147 0.12 restructuring) and disengagement coping (e.g., avoidance, T2Prim denial). Later on, in the development of social anxiety from   T1SAS 0.011 0.85 T2 to T3, stress responses had no predictive role any more.   T1Prim 0.548 0.00 Reversely, social anxiety did predict a relative increase T2Sec of stress responses. Surprisingly, this involved all stress   T1SAS 0.176 0.00 responses tested in the model.   T1Sec 0.382 0.00 The role of depression in the link between social anxiety T2Dis and stress responses seems limited, both concurrently and   T1SAS 0.138 0.04 prospectively. The only finding was that in the prospective   T1Dis 0.423 0.00 relation from T1 to T2 (but not from T2 to T3) depression T2InvE explained the link between social anxiety and disengage-   T1SAS 0.017 .83 ment coping.   T1InvE 0.444 0.00 Our finding that adolescents who show low primary T3SAS engagement coping initially (i.e., from T1 to T2) are more   T2SAS 0.506 0.00 likely to increase in social anxiety relative to peers who   T2Prim -0.024 0.68 show higher engagement coping, corroborates the results   T2Sec -0.101 0.06 of the two studies on general anxiety and adjustment prob-   T2Dis 0.012 0.86 lems in relation to stress responses (Flynn & Rudolph, 2011;   T2InvE 0.087 0.21 Wadsworth & Berger, 2006). However, it is not in line with T3Prim the Wright et al. (2010) and Richardson et al. (2020) stud-   T2SAS 0.112 0.03 ies on social anxiety. There are various differences between   T2Prim 0.604 0.00 the studies that can possibly explain the inconsistent results, T3Sec among which are the difference in stress responses addressed   T2SAS 0.215 0.00 and age differences between the participants. In particular   T2Sec 0.460 0.00 the age difference seems important. That is, the Wright T3Dis et al. (2010) study used a sample of children with the oldest   T2SAS 0.223 0.00 participants aged 11 years old. As social anxiety symptoms   T2Dis 0.406 0.00 tend to increase in early adolescence and social anxiety dis- T3InvE order often has its onset in this period of life (Ollendick &   T2SAS 0.218 0.00 Hirshfeld-Becker, 2002; Stein et al., 2017), it is likelier that   T2InvE 0.424 0.00 predictors of social anxiety development will be found in Significant estimates in bold this particular age group than in a sample of children. SAS social anxiety scale, Prim Our finding that adolescents with higher levels of social primary engagement, Sec second- anxiety relatively more often develop adaptive as well as ary engagement, Dis disengage- maladaptive stress responses later on, is consistent with the ment, InvE involuntary engage- Wright et al. (2010) study, but not with others that failed to ment find such a relation (Flynn & Rudolph, 2011; Richardson et al., 2020; Wadsworth & Berger, 2006). When compar- associations with disengagement coping (e.g., denial and ing results between the studies one should keep in mind avoidance responses), involuntary engagement (e.g., rumi- that the cited studies had various additional variables (e.g., nation, emotional arousal), and involuntary disengagement experienced stress, stress reactivity, general anxiety) in (e.g., inaction, emotional numbing), all of them maladap- their regression analyses, which may have influenced the tive responses. This pattern of associations was very sta- strength of (social)anxiety/adjustment as a predictor of stress ble over the three timepoints. Thus, the frequently reported responses, and for that matter stress responses as predictors link between internalizing problems and maladaptive stress of (social) anxiety/adjustment. 1 3 Research on Child and Adolescent Psychopathology Fig. 1 Significant prospective T1 T2 T3 relations between social anxiety and stress responses. Autore- SOC ANX SOC ANX SOC ANX gressive effects not depicted -.17 .11 for clarity reasons. SOC ANX: social anxiety; PRIM ENG: primary engagement cop- PRIM ENG PRIM ENG PRIM ENG ing; SEC ENG: secondary .18 PRIM ENG .22 engagement coping; DISENG: disengagement coping; INVOL ENG: involuntary engagement .14 .22 SEC ENG SEC ENG SEC ENG response DISENG DISENG DISENG INVOL INVOL INVOL ENG ENG ENG The present study’s findings suggest that instead of seek - stressful social situations may be brought under their control ing solutions for stressful social situations or regulating their (Davey, 1994). With regard to involuntary engagement, which distressing emotions, socially anxious adolescents tend to is characterized by emotional and physiological arousal and go into denial, practice wishful thinking, and/or avoid the rumination responses, it suggests that social anxiety over time stressful situation or emotion. In addition, they experience may also lead to increased emotional distress in response to responses to the stressor that are not under their control, for social stressors. example, repetitive thoughts about a negative social event The present study’s results may contribute to distinctive they encountered, physiological and emotional arousal, and approaches for prevention and intervention efforts respec - feelings of numbness and helplessness. The reason for ado- tively. At an early stage, when adolescents have not yet lescents becoming socially anxious may be that they did not developed high levels of social anxiety, but a deficiency (sufficiently) use adequate, primary engagement strategies, in adaptive coping strategies may make them prone to its such as problem solving and emotion regulation, to cope development, prevention efforts may address the practicing with the social problems they encountered. However, this of adaptive coping responses. By experiencing a reduction of seems only the case early on in the development as sug- stress as a result of helpful coping strategies the link between gested by the fact that T2 but not T3 social anxiety was the two will be strengthened. These helpful coping responses predicted by low primary engagement coping. Because may protect adolescents from developing chronic social individual differences in social anxiety may have been more anxiety (Richey et al., 2019). However, if they do develop or less crystalized after T2 with many participants having chronic social anxiety, interventions should also address the reached middle or late adolescence, there was less room for unhelpful stress responses they are inclined to. In particular, stress responses to influence changes in social anxiety. It it may be effective if they unlearn their tendency to respond should be noted, however, that the relatively large age range with disengagement strategies, such as denial and avoidance. of participants in the study makes interpretations like these At the same time, they need to be supported in practicing inconclusive. engagement strategies. Remarkably, the range of stress responses predicted by This prospective study is the first to describe the pro- social anxiety broadened from T2 to T3, although in both spective links between social anxiety and stress responses instances adaptive as well as maladaptive responses were in adolescence. A strong point of the study is its longitudi- involved. Increased primary engagement coping and invol- nal design with three measurement waves. There are also untary engagement appeared as new consequences of social a number of limitations of this study. First, all measures anxiety. Taking primary engagement coping, this occurrence used adolescents’ self-report, which poses the problem of may be associated with the growing flexibility and diversity of single source bias. Although social anxiety and responses coping responses in older adolescents (Zimmer-Gembeck & to stress mostly represent internal processes that only partly Skinner, 2011). Possibly, they learn these adaptive responses are accessible to others, information obtained from other from their friends (Reindl et  al., 2016) seeing that some sources such as parents, clinical interviews and behavioral 1 3 Research on Child and Adolescent Psychopathology (Eds.), Social Anxiety and Phobia in Adolescents (pp. 151–181). observations, would have added to the validity of the study. Springer International Publishing.  https:// doi. org/ 10. 1007/ Second, the study used a relatively homogeneous sample 978-3- 319- 16703-9_7 of Dutch middle-class youth. It is therefore not clear if the Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. results can be generalized to a wider population. Third, the In R. G. Heimberg, M. R. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: Diagnosis, assessment, and treatment (pp. study did not select adolescents with SAD and therefore can- 69–93). Guilford Press. not draw conclusions about the relationship between stress Compas, B. E., Connor-Smith, J. K., Saltzman, H., Thomsen, A. H., & responses and clinical levels of social anxiety. Fourth, the Wadsworth, M. E. (2001). Coping with stress during childhood study did not take into consideration specific stressors that and adolescence: Problems, progress, and potential in theory and research. Psychological Bulletin, 127(1), 87. cause stress for adolescents, such as being neglected or being Compas, B. E., Connor, J., Osowiecki, D., & Welch, A. (1997). Effort- bullied. Including responses to specific stressors may add to ful and involuntary responses to stress. Coping with chronic stress a deeper understanding of the link between stress responses (pp. 105–130). Springer. and social anxiety in adolescence. Finally, as noted by an Compas, B. E., Jaser, S. S., Bettis, A. H., Watson, K. H., Gruhn, M. A., Dunbar, J. P., et al. (2017). Coping, emotion regulation, and anonymous reviewer, the present study cannot answer the psychopathology in childhood and adolescence: A meta-analysis question to what degree the study’s results are specific to and narrative review. Psychological Bulletin, 143(9), 939. social anxiety or apply to anxiety in general. Connor-Smith, J. K., Compas, B. E., Wadsworth, M. E., Thomsen, A. In conclusion, the present study suggests that adolescents’ H., & Saltzman, H. (2000). Responses to stress in adolescence: Measurement of coping and involuntary stress responses. Jour- responses to stress and social anxiety development are inter- nal of Consulting and Clinical Psychology, 68(6), 976. woven. Adding stress responses to developmental models of Davey, G. C. (1994). Trait factors and ratings of controllability as social anxiety may result in a more complete understanding predictors of worrying about significant life stressors. Personal- of social anxiety in adolescence. ity and Individual Differences, 16(3), 379–384. Epkins, C. C., & Heckler, D. R. (2011). Integrating etiological mod- els of social anxiety and depression in youth: Evidence for a Supplementary Information The online version contains supplemen- cumulative interpersonal risk model. Clinical Child and Family tary material available at https://doi. or g/10. 1007/ s10802- 021- 00880-3 . Psychology Review, 14(4), 329–376. Erath, S. A., Flanagan, K. S., & Bierman, K. L. (2007). Social anxi- Ackowledgment We thank Marjolein Fokkema for her assistance in ety and peer relations in early adolescence: Behavioral and cog- conducting the cross-lagged panel analysis. nitive factors. Journal of Abnormal Child Psychology, 35(3), 405–416. Funding No funds, grants, or other support was received. Flynn, M., & Rudolph, K. D. (2011). Stress generation and adolescent depression: Contribution of interpersonal stress responses. Jour- nal of Abnormal Child Psychology, 39(8), 1187–1198. Compliance with Ethical Standards Furman, W., & Buhrmester, D. (1992). Age and sex differences in perceptions of networks of personal relationships. Child Develop- Ethics Approval The Medical Ethical Committee of Leiden University ment, 63(1), 103–115. approved the study. Grant, K. E., Compas, B. E., Thurm, A. E., McMahon, S. D., & Gipson, P. Y. (2004). Stressors and Child and Adolescent Psychopathology: Consent to Participate Parents gave their written consent and youth Measurement Issues and Prospective Effects. Journal of Clinical their written assent for participation in the study. Child & Adolescent Psychology, 33(2), 412–425. https:// doi. org/ 10. 1207/ s1537 4424j ccp33 02_ 23 Conflicts of Interests The authors have no relevant financial or non- Hemphill, J. F. (2003). Interpreting the magnitudes of correlation coef- financial interests to disclose. ficients. American Psychologist, 58(1), 78–79. https:// doi.or g/10. 1037/ 0003- 066X.58.1. 78 Open Access This article is licensed under a Creative Commons Attri- Kaeppler, A. K., & Erath, S. A. (2017). Linking social anxiety with social bution 4.0 International License, which permits use, sharing, adapta- competence in early adolescence: Physiological and coping mod- tion, distribution and reproduction in any medium or format, as long erators. Journal of Abnormal Child Psychology, 45(2), 371–384. as you give appropriate credit to the original author(s) and the source, Kline, R. B. (2005). Principles and practice of structural equation provide a link to the Creative Commons licence, and indicate if changes modeling: Methodology in the social sciences. Guilford Press. were made. The images or other third party material in this article are Kovacs, M. (1985). The Children Depression Inventory (CDI). Psy- included in the article's Creative Commons licence, unless indicated chopharmacology Bulletin, 21, 995–998. otherwise in a credit line to the material. If material is not included in LaGreca, A. M. (1998). Social anxiety scales for children and adoles- the article's Creative Commons licence and your intended use is not cents: Manual and instructions for the SASC, SASC–R, SAS–A permitted by statutory regulation or exceeds the permitted use, you will (adolescents), and parent versions of the scales. University of need to obtain permission directly from the copyright holder. To view a Miami, Department of Psychology. copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . LaGreca, A. M., & Lopez, N. (1998). Social anxiety among adoles- cents: Linkages with peer relations and friendships. Journal of Abnormal Child Psychology, 26(2), 83–94. References LaGreca, A. M., & Prinstein, M. J. (1999). Peer group In W. K. Silverman, & T. H. Ollendick (Eds.), Developmental issues in the clinical treat- Blöte, A. W., Miers, A. C., Heyne, D. A., & Westenberg, P. M. (2015). ment of children (pp. 171–198). Allyn & Bacon. Social Anxiety and the School Environment of Adolescents. In Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. K. Ranta, A. M. La Greca, L.-J. Garcia-Lopez, & M. Marttunen Springer publishing company. 1 3 Research on Child and Adolescent Psychopathology Magee, W. J., Eaton, W. W., Wittchen, H.-U., McGonagle, K. A., & Silk, J. S., Steinberg, L., & Morris, A. S. (2003). Adolescents’ emotion Kessler, R. C. (1996). Agoraphobia, simple phobia, and social regulation in daily life: Links to depressive symptoms and prob- phobia in the National Comorbidity Survey. Archives of General lem behavior. Child Development, 74(6), 1869–1880. Psychiatry, 53(2), 159–168. Spence, S. H., & Rapee, R. M. (2016). The etiology of social anxi- McLellan, L. F., Alfano, C. A., & Hudson, J. L. (2015). Cognition- ety disorder: An evidence-based model. Behaviour Research and focused interventions for social anxiety disorder among adoles- Therapy, 86, 50–67. cents. Social Anxiety and Phobia in Adolescents (pp. 225–250). Starr, L. R., & Davila, J. (2008). Differentiating interpersonal corre- Springer. lates of depressive symptoms and social anxiety in adolescence: Miers, A. C., Blöte, A. W., de Rooij, M., Bokhorst, C. L., & Westenberg, Implications for models of comorbidity. Journal of Clinical Child P. M. (2013). Trajectories of Social Anxiety during Adolescence & Adolescent Psychology, 37(2), 337–349. and Relations with Cognition, Social Competence, and Tempera- Stein, D. J., Lim, C. C., Roest, A. M., De Jonge, P., Aguilar-Gaxiola, S., ment. Journal of Abnormal Child Psychology, 41(1), 97–110. Al-Hamzawi, A., et al. (2017). The cross-national epidemiology https:// doi. org/ 10. 1007/ s10802- 012- 9651-6 of social anxiety disorder: Data from the World Mental Health Ollendick, T. H., & Hirshfeld-Becker, D. R. (2002). The developmental Survey Initiative. BMC Medicine, 15(1), 143. psychopathology of social anxiety disorder. Biological Psychiatry, Steinberg, L. (2010). A dual systems model of adolescent risk-taking. 51(1), 44–58. https:// doi. org/ 10. 1016/ s0006- 3223(01) 01305-1 Developmental Psychobiology: THe Journal of the International R-Core-Team. (2019). R: A language and environment for statisti- Society for Developmental Psychobiology, 52(3), 216–224. cal computing. R Foundation for Statistical Computing. https:// Storch, E. A., Masia-Warner, C., Dent, H. C., Roberti, J. W., & Fisher, www.R- proje ct. org/. Accessed. P. H. (2004). Psychometric evaluation of the Social Anxiety Scale Ranta, K., Junttila, N., Laakkonen, E., Uhmavaara, A., La Greca, A. for Adolescents and the Social Phobia and Anxiety Inventory for M., & Niemi, P. M. (2012). Social Anxiety Scale for Adolescents Children: Construct validity and normative data. Journal of Anxi- (SAS-A): Measuring social anxiety among Finnish adolescents. ety Disorders, 18(5), 665–679. Child Psychiatry & Human Development, 43(4), 574–591. Timbremont, B., & Braet, C. (2002). Children’s Depression Inventory: Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral Nederlandstalige versie [Children’s Depression Inventory: Dutch model of anxiety in social phobia. Behaviour Research and Ther- version. Swets & Zeitlinger. apy, 35(8), 741–756. Van den Bos, E., De Rooij, M., Miers, A. C., Bokhorst, C. L., & Westenberg, Reijntjes, A., Kamphuis, J. H., Prinzie, P., & Telch, M. J. (2010). Peer P. M. (2014). Adolescents’ Increasing Stress Response to Social Evalu- victimization and internalizing problems in children: A meta- ation: Pubertal Effects on Cortisol and Alpha-Amylase During Public analysis of longitudinal studies. Child Abuse & Neglect, 34(4), Speaking. Child Development, 85(1), 220–236. https://doi. or g/10. 1111/ 244–252. https:// doi. org/ 10. 1016/j. chiabu. 2009. 07. 009 cdev. 12118 Reindl, M., Gniewosz, B., & Reinders, H. (2016). Socialization of emo- Wadsworth, M. E., & Berger, L. E. (2006). Adolescents coping with tion regulation strategies through friends. Journal of Adolescence, poverty-related family stress: Prospective predictors of coping 49, 146–157. and psychological symptoms. Journal of Youth and Adolescence, Richardson, C. E., Magson, N. R., Fardouly, J., Oar, E. L., Forbes, M. 35(1), 54–67. K., Johnco, C. J., et al. (2020). Longitudinal associations between Westenberg, P. M., Bokhorst, C. L., Miers, A. C., Sumter, S. R., Kallen, coping strategies and psychopathology in pre-adolescence. Jour- V. L., van Pelt, J., et al. (2009). A prepared speech in front of a nal of Youth and Adolescence, 1–16. pre-recorded audience: Subjective, physiological, and neuroendo- Richey, J. A., Brewer, J. A., Sullivan-Toole, H., Strege, M. V., Kim-Spoon, crine responses to the Leiden Public Speaking Task. Biological J., White, S. W., et al. (2019). Sensitivity shift theory: A develop- Psychology, 82(2), 116–124. https://doi. or g/10. 1016/j. biops y cho. mental model of positive affect and motivational deficits in social 2009. 06. 005 anxiety disorder. Clinical psychology review, 101756. Wittchen, H. U., & Fehm, L. (2003). Epidemiology and natural course Roelofs, J., Braet, C., Rood, L., Timbremont, B., Van Vlierberghe, L., of social fears and social phobia. Acta Psychiatrica Scandinavica, Goossens, L., et al. (2010). Norms and screening utility of the 108(s417), 4–18. Dutch version of the Children’s Depression Inventory in clinical Wright, M., Banerjee, R., Hoek, W., Rieffe, C., & Novin, S. (2010). and nonclinical youths. Psychological Assessment, 22(4), 866. Depression and social anxiety in children: Differential links with Rosseel, Y. (2012). Lavaan: An R package for structural equation mod- coping strategies. Journal of Abnormal Child Psychology, 38(3), eling and more. Version 0.5–12 (BETA). Journal of statistical 405–419. software, 48(2), 1–36. Zimmer-Gembeck, M. J., & Skinner, E. A. (2011). The development of Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the coping across childhood and adolescence: An integrative review state of the art. Psychological Methods, 7(2), 147. and critique of research. International Journal of Behavioral Sentse, M., Lindenberg, S., Omvlee, A., Ormel, J., & Veenstra, R. Development, 35(1), 1–17. (2010). Rejection and acceptance across contexts: Parents and peers as risks and buffers for early adolescent psychopathology. Publisher's Note Springer Nature remains neutral with regard to The TRAILS study. Journal of abnormal child psychology, 38(1), jurisdictional claims in published maps and institutional affiliations. 119–130. 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Abnormal Child Psychology Springer Journals

Concurrent and Prospective Associations Between Social Anxiety and Responses to Stress in Adolescence

Loading next page...
 
/lp/springer-journals/concurrent-and-prospective-associations-between-social-anxiety-and-P6EbPCSCu9
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2021
ISSN
0091-0627
eISSN
2730-7174
DOI
10.1007/s10802-021-00880-3
Publisher site
See Article on Publisher Site

Abstract

Several studies have investigated the relationship between adolescents’ responses to stress and general anxiety and depres- sion, but only few studies addressed the relationship between responses to stress and social anxiety. The current three-wave longitudinal study, that covered a period of 5 years with a time interval of on average two years between waves, examined concurrent as well as prospective relations between adolescents’ self-reported stress responses, including coping responses, and self-perceived social anxiety. Both the predictive power of social anxiety for different stress responses and, reversely, of stress responses for social anxiety were evaluated. Participants were 331 youth (170 boys) aged 9 to 17 years old at Wave 1. Self-report questionnaires were used to measure social anxiety, responses to social stress, and depressive symptoms. Results showed signic fi ant concurrent relations between social anxiety and maladaptive stress responses. Moreover, the study yielded evidence for social anxiety predicting stress responses across time as well as stress responses predicting social anxiety, although evidence for the former link is stronger. The findings suggest that a relative lack of adaptive stress responses may heighten social anxiety and social anxiety in turn may trigger maladaptive as well as adaptive responses to social problems. The relevance of these findings for social anxiety prevention and intervention purposes are discussed. Keywords Adolescence · Social anxiety · Coping · Stress responses Adolescence is a period of significant changes in many If adolescents do not cope with social stress in an effec- domains of life, physiological, cognitive, emotional, and tive way, internalizing problems related to social anxiety social. In adolescents’ social life, relationships with peers may develop (Grant et al., 2004; Richey et al., 2019). In become more, and those with parents less, important turn, these problems may then further hinder the use of (Steinberg, 2010). Adolescents increasingly share their adequate stress responses possibly resulting in even more experiences and emotions with peers and rely on these social stress, anxiety symptoms, and ultimately social anxi- peers for social support (Furman & Buhrmester, 1992). ety disorder. Thus, adolescents’ responses to stress may They also become more sensitive to their own position be risk factors as well as consequences of social anxiety in the peer group. How they are evaluated by peers and problems (Compas et al., 2001; Wright et al., 2010). particularly if they are accepted by peers are important To our knowledge, no study to date has addressed the factors in the lives of adolescents (LaGreca & Prinstein, prospective relation between adolescent stress responses 1999; Sentse et al., 2010). This need for peer acceptance, and social anxiety, despite its importance for understanding for being “one of them”, may result in increasing social social anxiety development. The available studies used sam- stress levels (Ollendick & Hirshfeld-Becker, 2002; Van ples of children or early adolescents. Because adolescence den Bos et  al., 2014), especially in the case of adverse is a period in which social anxiety symptoms increase and peer experiences (Blöte et al., 2015; Reijntjes et al., 2010). social anxiety disorder has its onset (Magee et al., 1996; Wittchen & Fehm, 2003), this age group is of particular interest for studying the role of stress responses in relation * Anke W. Blöte to social anxiety development. The present study therefore bloete@fsw.leidenuniv.nl tried to fill the gap in the literature by paying special atten- Institute of Psychology, Unit of Developmental tion to the potential bi-directional relation between stress and Educational Psychology, Leiden University, responses and social anxiety problems in adolescents. Wassenaarseweg 52, 2333 AK Leiden, The Netherlands Vol.:(0123456789) 1 3 Research on Child and Adolescent Psychopathology Responses to stress are categorized as either voluntary or Clark & Wells, 1995; Spence & Rapee, 2016) pay only lim- involuntary (Compas et al., 1997; Lazarus & Folkman, 1984). ited attention to the relation of different coping styles with Voluntary responses – generally referred to as coping– are the development (Spence & Rapee, 2016) and maintenance regarded as under the person’s control and requiring conscious (Clark & Wells, 1995; Rapee & Heimberg, 1997) of social effort, whereas involuntary responses to stress are automatic, anxiety, although some coping strategies such as cognitive not under the control of the person concerned. With regards restructuring and safety behaviors are addressed. Cogni- to stress responses in youth, Compas et al. (2001) presented tive restructuring, that addresses the replacement of anxi- three dimensions along which the different stress responses ety laden biased cognitions with more realistic cognitions may differ; the just mentioned distinction between voluntary (McLellan et al., 2015), is a form of secondary engagement and involuntary responses characterizes the first dimension. coping. Safety behaviors such as avoiding eye contact and Voluntary responses are consciously aimed at changing the keeping quiet in interactions with others, are intended to stressful situation or one’s own response to it, be it cognitive, avoid aversive social experiences. As such they are a type behavioral, or emotional. Examples of voluntary responses of disengagement response. In the Spence and Rapee (2016) are among others, problem solving, seeking distraction, and model on social anxiety development in adolescents, both emotion regulation. Involuntary or automatic responses to negative social cognitions and safety behaviors are included stress, such as physiological arousal and rumination, are not as risk factors of social anxiety development. (directly) aimed at regulating the situation or one’s emotional Recently, a new model of adolescent social anxiety, Sen- responses to it and they may or may not be within conscious sitivity Shift Theory (SST), was presented that more explic- awareness. The second dimension makes a distinction based itly pays attention to the link between social anxiety and on the person’s engagement with or disengagement from the stress responses (Richey et al., 2019). SST describes the stressful situation. Engagement responses are responses that development from an inhibited temperament to social stress approach the stressful event or unpleasant emotion, whereas and social anxiety, and from there to social anhedonia. SST disengagement responses are oriented away from them and are emphasizes the importance of coping in this development characterized by avoidance. The third dimension is specific by stating that the success of socially anxious adolescents’ to voluntary engagement responses and addresses primary coping responses in reducing stress determines whether control versus secondary control strategies. Primary control they will continue to put effort into their coping with future coping is directed at changing the stressful situation or the stressors, or give up, stop expending energy and start to person’s emotional state caused by the situation, while sec- avoid social situations. When the latter occurs, the positive ondary control coping refers to the person adapting to the affect normally associated with social situations disappears, problem, for example by cognitive restructuring and accept- reducing the chance that successful coping strategies are ance of the situation or emotions involved. effectuated in future social situations and maintaining or Some stress responses are considered adaptive because even further increasing social anxiety. This final stage in they help to diminish the stress, and other stress responses the model describes a condition named social anhedonia that as maladaptive because they do not diminish it and conse- is primarily characterized by disengagement from stressful quently may lead to internalizing problems (Compas et al., social interactions. 2017; Connor-Smith et  al., 2000). Compas et  al. (2001; Empirical research investigating the concurrent rela- 2017) concluded, based on extensive literature reviews, that tion between adolescents’ stress responses and social anxi- engagement coping in youth is adaptive because it is asso- ety is limited and used relatively young participants from ciated with psychological adjustment and well-being. For late childhood and early adolescence. Furthermore, the example, studies found that engagement coping is negatively results of these studies yielded equivocal results. In the fol- related to internalizing problems in adolescents both when lowing review studies used self-report to assess the main using self-reported and parent-reported measures (Connor- variables; exceptions to this rule are explicitly noted. In a Smith et al., 2000) and engagement coping with peer vic- sample aged 8 t o11 years old, Wright et al. (2010) found timization in socially anxious adolescents is linked with a significant, although weak, positive link of social anxiety social competence (Kaeppler & Erath, 2017). In contrast, symptoms with problem solving and support seeking, both disengagement from stressful situations or one’s own feel- primary engagement coping responses, and a significant and ings of stress is linked with poor psychological adjustment strong positive link with involuntary engagement responses. and is therefore considered maladaptive. Richardson et al. (2020) reported a significant association The role of coping in theories of social anxiety seems of social anxiety symptoms with avoidance coping and a somewhat underrepresented. In light of the distinction significant but small negative association with problem between adaptive and maladaptive coping strategies and solving responses in 10 to 12-year olds. Parent-reported their links with emotional adjustment, it is a bit surprising measures yielded similar results. In contrast, Erath et al. to see that important models of social anxiety disorder (e.g., (2007) did not find significant correlations between social 1 3 Research on Child and Adolescent Psychopathology anxiety symptoms and either engagement or disengagement component in it. Social anxiety and depression share a num- responses in early adolescents (sixth and seventh graders). ber of characteristics with regards to risk factors, and associ- To our knowledge, only two prospective studies have been ated and consequent variables. Identifying what is specific conducted on the relation between social anxiety and stress to a disorder is important for several reasons, among them responses and these studies used children or young adoles- the development of treatment and prevention interventions cents, 8 to11 year olds in the Wright et al. (2010) study (Epkins & Heckler, 2011; Starr & Davila, 2008). It is also and 10 to12 year olds in the Richardson et al. (2020) study. possible that depression and social anxiety each have differ - The Wright et al. (2010) study found that over a period of ent relations with certain stress responses as the Wright et al. nine months, social anxiety symptoms predicted a rela- (2010) study made clear. This study revealed that depres- tive increase of social support seeking, a form of primary sion did not explain any relation between social anxiety and engagement coping, controlling for the effect of depression. stress responses. Remarkably, depression and social anxi- Social anxiety was also related to relative increases in invol- ety even had opposite effects on children’ s stress responses. untary responses (e.g., worrying, not sleeping). Specifically, For example, depression predicted a decrease of distraction for children who experienced lower peer rejection, social seeking whereas social anxiety predicted an increase in dis- anxiety predicted an increase in distraction seeking, a form traction seeking (for children who experienced lower peer of secondary control engagement coping. Stress responses, rejection). These opposing effects might have obscured the in turn, did not predict subsequent changes in social anxi- relationship between social anxiety and stress responses if ety symptoms. In contrast, Richardson et al. (2020) found the effect of depression had not been taken in account. This that social anxiety, controlled for depression, did not predict finding further stresses the importance of paying attention increases in the different coping responses (problem solv - to the role of depression in the link between social anxiety ing, social support seeking, and avoidant coping) measured and stress responses. one year later. Reversely, avoidant coping predicted relative increases in social anxiety symptoms. Parent-reported meas- ures did not corroborate this finding. The Present Study Two other studies examined the prospective links between stress responses and more general adjustment problems. One In sum, it is clear that adolescent anxiety and depression are study with 9–15 years old participants, three assessment associated with maladaptive responses to stress, but it is not waves over a period of two years, and general anxiety and clear if social anxiety shows a similar pattern of relations to depression as respective internalizing variables revealed sev- stress responses. Moreover, there is a gap in our knowledge eral predictive links from stress responses to anxiety (Flynn about the prospective links between anxiety/depression in & Rudolph, 2011). The combination of low engagement general and social anxiety in particular, with stress responses coping and high involuntary stress responses (both in rela- (Compas et al., 2017). Therefore, the present study sought tion to social situations) predicted relatively high levels of to answer the following question. What is the relationship anxiety and depression as diagnosed in interviews with the between social anxiety and responses to social stress (a) con- participants and their primary caregiver. Reversely, anxiety currently and (b) over time? and depression did not predict stress responses. Based on the literature about general anxiety and depres- The second study evaluated coping with poverty-related sion we expected that concurrently social anxiety symptoms family stress in adolescents with a mean age of 14 years are positively related to maladaptive stress responses (dis- (Wadsworth & Berger, 2006). This study, that used two engagement coping and involuntary stress responses), and time-points eight months apart, addressed self-reported negatively related to adaptive responses (engagement cop- emotional adjustment in general, without further distinguish- ing). There are as yet too few longitudinal studies to guide ing between anxiety and depression symptoms, and found any specific expectations for the prospective effects of social that emotional adjustment did not predict changes in coping anxiety symptoms and stress responses over time. Moreover, strategies, but that coping strategies did predict changes in the studies we know of yielded equivocal results. Theoreti- emotional adjustment. For adolescents with relatively high cally, we would expect that a low level of engagement coping initial family stress levels, primary control engagement cop- and a high level of disengagement coping and involuntary ing predicted higher emotional adjustment. For adolescents stress responses predict relative increases of social anxiety with relatively low initial adjustment, disengagement coping over time, because these responses will not help to solve the predicted even lower adjustment. social stress related problems and therewith prevent further Because of the high comorbidity between social anxiety distress related to these problems (Richey et al., 2019). and depression (Epkins & Heckler, 2011), it may not be clear Because disengagement coping (Silk et al., 2003; Wright whether links between social anxiety and stress responses are et al., 2010) and involuntary stress responses (Silk et al., specific to social anxiety or are explained by the depression 2003) not only have been linked to (social) anxiety but also 1 3 Research on Child and Adolescent Psychopathology to depression, links between social anxiety and these stress the timepoints. The T1 social anxiety of participants with responses may be caused by the depressive component in missing data caused by attrition or incidental events (that social anxiety. For this reason, we evaluated whether the is, all missing data except the ones related to age) did not putative links between stress responses and social anxiety significantly differ from that of participants with complete could at least partly be attributed to depression. data. A repeated measures ANOVA on social anxiety meas- ured at the three timepoints yielded a non-significant group effect between participants with and without missing data, Method F(1, 231) = 3.28, n.s., as well as a non-significant Group x Time interaction effect, F(2, 231) = 0.72, n.s. Design and Procedure Instruments Data were drawn from the Social Anxiety and Normal Development (SAND) study, a community study that The Dutch translation (H. Koot and E. Utens, unpub- selected students from two primary schools and one sec- lished) of the Social Anxiety Scale for Adolescents (SAS-A; ondary school in an urban area in the Netherlands (Miers LaGreca & Lopez, 1998) was used to measure social anxiety. et al., 2013; Westenberg et al., 2009). Data relevant to the The SAS-A contains 18 statements linked to social anxiety present study were collected at the first, third, and fourth (e.g., “I worry that others don’t like me”). Participants indi- waves (referred to as T1, T2, and T3, respectively) of this cate for each statement how true it is for themselves using four-wave longitudinal study. (At the second wave, only a a 5-point Likert scale (1 = not at all, 5 = all the time). Total restricted number of variables were measured). The time scores range between 18 and 90 with total score > 50 as the interval between these waves was two years on average with recommended criterion for clinically significant levels of intervals varying between one and three years. Severe psy- social anxiety (LaGreca, 1998). The SAS-A has shown good chological problems or physical illness as registered by the internal consistency with Cronbach’s alpha > 0.90 and cor- school were a contra-indication for participation in the study. relates strongly with other social anxiety measures (LaGreca Participants individually completed a battery of assess- & Lopez, 1998; Storch et al., 2004). Cronbach’s alpha for the ment forms at the university laboratory including the three three timepoints of the present study was > 0.92. In the data questionnaires used in the present study. The SAND study analysis, sum scores over the 18 items were used. was approved by the university’s Medical Ethical Commit- We measured depression with the Dutch translation tee. Parents gave their written consent and youth their writ- (Timbremont & Braet, 2002) of the Children’s Depression ten assent for participation in the study. Inventory (CDI; Kovacs, 1985). For each of the 27 items of the questionnaire participants indicate which of three state- Participants ments best describes how they felt in the last two weeks. For example, “I do most things OK”, “I do many things wrong”, The SAND study started at T1 with 331 primary and sec- and “I do everything wrong.” Scores range from 0 to 2 (most ondary school students, 170 boys and 161 girls (Miers et al., depressed). The Dutch version of the CDI has good internal 2013). Originally, 434 students from Grades 4 to 9 had been consistency (Cronbach’s alpha > 0.80) and shows a strong contacted of whom 75 students did not assent or their parents correlation with DSM-oriented depression measures (Roelofs did not consent to participation in the study. Other students et al., 2010; Timbremont & Braet, 2002). Total scores may (n = 28) did not participate for various other reasons, mainly range between 0 and 54 with a cut-off point of 16 for clinically because it was not possible to invite them to the laboratory significant cases. For ethical reasons, the item asking about within the available time slots. Out of the 331 students, 248 suicide was not presented to the participants. Data analysis and 236 (75% and 71%) respectively, still participated at was therefore based on the sum scores over 26 items. Cron- T2 and T3. At T1, participants’ ages ranged between 9 and bach’s alpha was 0.80, 0.84, and 0.83 for the three respective 17 years, M = 13.34, SD = 2.25. At T3, the mean age of the timepoints. participants was M = 17.48, SD = 2.72. The T3 sample con- The Responses to Stress Questionnaire (RSQ; Connor- sisted of 121 boys and 115 girls. Apart from the missing Smith et al., 2000) translated into Dutch by H. Ouwehand data caused by attrition at T2 and T3, other missing data (unpublished) was used to measure participants’ stress occurred for the responses to stress measure because it was responses. The RSQ offers the possibility to measure ado- not completed by primary school students (the measure was lescents’ stress responses related to different domains. In specifically designed for adolescents). This resulted in 126 the present study we referred to stressful situations in the age-related missing values on this measure at T1 with 32 of social domain, namely, having problems with other kids. those still missing at T2. Furthermore, seven participants For example, “When problems with other kids come up, had missing values on one of the three measures at any of I can't stop thinking about how I am feeling.” The RSQ 1 3 Research on Child and Adolescent Psychopathology has five scales: Primary control engagement coping (9 Results items covering problem solving, emotional regulation, and emotional expression responses), secondary con- Descriptive Analyses trol engagement coping (12 items covering acceptance, distraction, cognitive restructuring, and positive think- Means and SDs of the study variables at T1 are presented ing responses), disengagement coping (9 items cover- in Table 1. The table shows that the social anxiety level of ing denial, avoidance, and wishful thinking responses), the present sample is largely comparable to that of other involuntary engagement (15 items covering rumination, community samples (Ranta et  al., 2012; Storch et  al., intrusive thoughts, emotional and physiological arousal, 2004). The SAS-A cut-off score of 50 for clinical levels and impulsive action responses), and involuntary dis- of social anxiety (LaGreca, 1998) lies within 1SD above engagement (12 items covering emotional numbing, the mean indicating that high levels of social anxiety were inaction, escape, and cognitive interference). Items are not uncommon in the sample. The CDI mean is relatively scaled from 1 to 4 (1 = not at all, 4 = a lot), according to low (Roelofs et al., 2010). how often the participant says to use the voluntary cop- ing or experience the involuntary stress response. The RSQ scales of the social stress version have adequate internal consistency (Cronbach’s alpha’s ranging from Concurrent Relations 0.73—0.89). It was found that the primary engagement and disengagement coping scales correlate with the COPE At T1, social anxiety was strongly (Hemphill, 2003) posi- and the involuntary stress responses scales with heart- tively related to the three maladaptive responses to stress, rate reactivity (Connor-Smith et al., 2000). In the present namely disengagement coping, involuntary engagement, study, Cronbach’s alpha of the scales ranged from 0.71 and involuntary disengagement (see Table 1). Social anxi- for disengagement coping to 0.88 for involuntary engage- ety was not associated with primary or secondary control ment at T1, from 0.72 for secondary engagement coping engagement coping. Social anxiety and depression were to 0.89 for involuntary engagement at T2, and from 0.73 relatively strongly related and their respective correlations for secondary engagement coping to 0.89 for involuntary with the responses to stress variables were largely similar. engagement at T3. After controlling for the effect of depression, the correla- tions between social anxiety and the respective maladap- tive stress responses were still significant at a p < 0.001 Data Analysis level albeit the partial correlations were somewhat lower than the zero-order correlations. That is, the partial cor- First, descriptives and correlations between all study relations of social anxiety with disengagement coping, variables were computed for the different timepoints. involuntary engagement, and involuntary disengagement Second, cross-lagged panel analysis was used to test a were r = 0.459, r = 0.546, and r = 0.429, respectively, com- model describing the effect of social anxiety at T1 on pared to the zero-order correlations r = 0.592, r = 0.656, participants’ stress responses at T2, of social anxiety at and r = 0.572. T2 on participants’ stress responses at T3, and reversely, Age at T1 was significantly correlated with primary of the stress responses of T1 and T2 on the respective engagement coping, r = 0.170, p < 0.05. (At T3, it was social anxiety levels at T2 and T3. In order to control for significantly correlated with disengagement coping, r = depression, we also tested a model in which depression -0.143, p < 0.05, and at T4 both with primary engagement levels measured at the three time-points were added in coping, r = 0.143, p < 0.05, and disengagement coping, all regressions either as a dependent or independent vari- r = -0.134, p < 0.05). Controlling for age did not affect able together with social anxiety. Finally, because the age the other correlations, as the partial correlations of social range of the participants was relatively large, we tested anxiety with disengagement coping, involuntary engage- a model that controlled for age. The cross-lagged panel ment, and involuntary disengagement show, respectively, analyses were performed with Lavaan (Rosseel, 2012) in r = 0.587, r = 0.650, and r = 0.586. R (R-Core-Team, 2019). With regards to missing data, full The two engagement responses were positively interre- information maximum likelihood estimation (Schafer & lated. Primary control engagement coping was also signifi- Graham, 2002) was employed thus all available data were cantly positively associated with involuntary engagement. used. As goodness of fit criteria we used a comparative fit Furthermore, the three maladaptive stress responses, dis- index (CFI) of around 0.95 and root-mean-square error of engagement coping and involuntary engagement and dis- approximation (RMSEA) of around 0.05 (Kline, 2005). engagement, were strongly interrelated. At T2 and T3, the 1 3 Research on Child and Adolescent Psychopathology Table 1 Ms (SDs), and M (SD) n SAS CDI Prim Sec Dis InvE InvD Correlations of Study Variables ** ** ** ** at T1 SAS 40.671 - 0.502 0.003 0.019 0.592 0.656 0.575 (12.750) n = 328 ** ** ** CDI 8.881 - -0.054 -0.062 0.489 0.483 0.508 (5.361) n = 327 ** ** Prim 22.427 - 0.316 -0.069 0.276 0.068 (4.922) n = 205 ** Sec 28.259 - 0.198 0.030 0.111 (5.300) n = 203 ** ** Dis 15.499 - 0.545 0.618 (3.691) n = 205 ** InvE 25.515 - 0.707 (6.886) n = 204 InvD 18.512 - (4.676) n = 204 SAS social anxiety scale, CDI children’s depression inventory, Prim primary engagement, Sec secondary engagement, Dis disengagement, InvE involuntary engagement, InvD involuntary disengagement ** Correlation is significant at the 0.01 level (2-tailed) pattern of correlations between the variables was mainly (compared to adolescents with higher primary control engage- similar to that at T1 (see supplementary Tables 1 and 2). ment coping) of T2 social anxiety. Reversely, T1 social anxiety predicted a relative increase of secondary engagement coping and disengagement coping at T2. From T2 to T3, none of the Relations Over Time stress responses predicted social anxiety, but social anxiety predicted a relative increase of all four stress responses. The Because the correlation between involuntary engagement and significant effects in the model are depicted in Fig.  1. disengagement was very high, with rs > 0.70 at the different The second cross-lagged analysis that added depression to time points, and we wished to improve the number of partici- the model, showed that the path from T1 social anxiety to T2 pants/number of variables ratio in the cross-lagged panel anal- disengagement coping was no longer significant (see Table  3 yses, we chose to exclude involuntary disengagement from of the supplementary tables). Instead, depression predicted these analyses. Based on attention received in the literature to disengagement coping. In other respects, the pattern of results date, involuntary engagement, which includes rumination and was unaffected by the addition of depression to the model. emotional and physiological arousal, is the most important. From T2 to T3, as compared with from T1 to T2, social anxi- It should be noted that the involuntary disengagement scale ety predicted disengagement coping even after depression was was added to the RSQ purely for model-theoretical reasons, added to the model. namely, to have an engagement-disengagement dimension that The third model that included age as covariate (see Table 4 applies to voluntary as well as involuntary responses (Connor- of the supplementary tables) did not meaningfully change the Smith et al., 2000). results as found in the first model. The first cross-lagged panel model (that did not include depression or age) had a good fit, CFI = 0.978 and RMSEA = 0.055. The results showed rather large autoregres- Discussion sive effects of social anxiety as well as the different stress responses indicating that individual differences on these vari- This study found that adolescents’ self-perceived social ables are relatively stable over time (see Table 2). Low primary anxiety and stress responses are linked both concurrently control engagement coping at T1 predicted a relative increase as well as over time. Social anxiety showed concurrent 1 3 Research on Child and Adolescent Psychopathology Table 2 Standardized Estimates responses (Compas et al., 2001; 2017) was replicated spe- Standard P of Cross-lagged Analysis cifically for social anxiety. estimates Prospectively, we found bi-directional effects between T2SAS social anxiety and stress responses. That is to say, initially   T1SAS 0.465 0.00 from T1 to T2, low primary control engagement coping   T1Prim -0.173 0.01 (e.g., low problem solving, low emotion regulation) pre-   T1Sec -0.027 0.66 dicted social anxiety, and social anxiety predicted second-   T1Dis -0.057 0.47 ary engagement coping (e.g., distraction seeking, cognitive   T1InvE 0.147 0.12 restructuring) and disengagement coping (e.g., avoidance, T2Prim denial). Later on, in the development of social anxiety from   T1SAS 0.011 0.85 T2 to T3, stress responses had no predictive role any more.   T1Prim 0.548 0.00 Reversely, social anxiety did predict a relative increase T2Sec of stress responses. Surprisingly, this involved all stress   T1SAS 0.176 0.00 responses tested in the model.   T1Sec 0.382 0.00 The role of depression in the link between social anxiety T2Dis and stress responses seems limited, both concurrently and   T1SAS 0.138 0.04 prospectively. The only finding was that in the prospective   T1Dis 0.423 0.00 relation from T1 to T2 (but not from T2 to T3) depression T2InvE explained the link between social anxiety and disengage-   T1SAS 0.017 .83 ment coping.   T1InvE 0.444 0.00 Our finding that adolescents who show low primary T3SAS engagement coping initially (i.e., from T1 to T2) are more   T2SAS 0.506 0.00 likely to increase in social anxiety relative to peers who   T2Prim -0.024 0.68 show higher engagement coping, corroborates the results   T2Sec -0.101 0.06 of the two studies on general anxiety and adjustment prob-   T2Dis 0.012 0.86 lems in relation to stress responses (Flynn & Rudolph, 2011;   T2InvE 0.087 0.21 Wadsworth & Berger, 2006). However, it is not in line with T3Prim the Wright et al. (2010) and Richardson et al. (2020) stud-   T2SAS 0.112 0.03 ies on social anxiety. There are various differences between   T2Prim 0.604 0.00 the studies that can possibly explain the inconsistent results, T3Sec among which are the difference in stress responses addressed   T2SAS 0.215 0.00 and age differences between the participants. In particular   T2Sec 0.460 0.00 the age difference seems important. That is, the Wright T3Dis et al. (2010) study used a sample of children with the oldest   T2SAS 0.223 0.00 participants aged 11 years old. As social anxiety symptoms   T2Dis 0.406 0.00 tend to increase in early adolescence and social anxiety dis- T3InvE order often has its onset in this period of life (Ollendick &   T2SAS 0.218 0.00 Hirshfeld-Becker, 2002; Stein et al., 2017), it is likelier that   T2InvE 0.424 0.00 predictors of social anxiety development will be found in Significant estimates in bold this particular age group than in a sample of children. SAS social anxiety scale, Prim Our finding that adolescents with higher levels of social primary engagement, Sec second- anxiety relatively more often develop adaptive as well as ary engagement, Dis disengage- maladaptive stress responses later on, is consistent with the ment, InvE involuntary engage- Wright et al. (2010) study, but not with others that failed to ment find such a relation (Flynn & Rudolph, 2011; Richardson et al., 2020; Wadsworth & Berger, 2006). When compar- associations with disengagement coping (e.g., denial and ing results between the studies one should keep in mind avoidance responses), involuntary engagement (e.g., rumi- that the cited studies had various additional variables (e.g., nation, emotional arousal), and involuntary disengagement experienced stress, stress reactivity, general anxiety) in (e.g., inaction, emotional numbing), all of them maladap- their regression analyses, which may have influenced the tive responses. This pattern of associations was very sta- strength of (social)anxiety/adjustment as a predictor of stress ble over the three timepoints. Thus, the frequently reported responses, and for that matter stress responses as predictors link between internalizing problems and maladaptive stress of (social) anxiety/adjustment. 1 3 Research on Child and Adolescent Psychopathology Fig. 1 Significant prospective T1 T2 T3 relations between social anxiety and stress responses. Autore- SOC ANX SOC ANX SOC ANX gressive effects not depicted -.17 .11 for clarity reasons. SOC ANX: social anxiety; PRIM ENG: primary engagement cop- PRIM ENG PRIM ENG PRIM ENG ing; SEC ENG: secondary .18 PRIM ENG .22 engagement coping; DISENG: disengagement coping; INVOL ENG: involuntary engagement .14 .22 SEC ENG SEC ENG SEC ENG response DISENG DISENG DISENG INVOL INVOL INVOL ENG ENG ENG The present study’s findings suggest that instead of seek - stressful social situations may be brought under their control ing solutions for stressful social situations or regulating their (Davey, 1994). With regard to involuntary engagement, which distressing emotions, socially anxious adolescents tend to is characterized by emotional and physiological arousal and go into denial, practice wishful thinking, and/or avoid the rumination responses, it suggests that social anxiety over time stressful situation or emotion. In addition, they experience may also lead to increased emotional distress in response to responses to the stressor that are not under their control, for social stressors. example, repetitive thoughts about a negative social event The present study’s results may contribute to distinctive they encountered, physiological and emotional arousal, and approaches for prevention and intervention efforts respec - feelings of numbness and helplessness. The reason for ado- tively. At an early stage, when adolescents have not yet lescents becoming socially anxious may be that they did not developed high levels of social anxiety, but a deficiency (sufficiently) use adequate, primary engagement strategies, in adaptive coping strategies may make them prone to its such as problem solving and emotion regulation, to cope development, prevention efforts may address the practicing with the social problems they encountered. However, this of adaptive coping responses. By experiencing a reduction of seems only the case early on in the development as sug- stress as a result of helpful coping strategies the link between gested by the fact that T2 but not T3 social anxiety was the two will be strengthened. These helpful coping responses predicted by low primary engagement coping. Because may protect adolescents from developing chronic social individual differences in social anxiety may have been more anxiety (Richey et al., 2019). However, if they do develop or less crystalized after T2 with many participants having chronic social anxiety, interventions should also address the reached middle or late adolescence, there was less room for unhelpful stress responses they are inclined to. In particular, stress responses to influence changes in social anxiety. It it may be effective if they unlearn their tendency to respond should be noted, however, that the relatively large age range with disengagement strategies, such as denial and avoidance. of participants in the study makes interpretations like these At the same time, they need to be supported in practicing inconclusive. engagement strategies. Remarkably, the range of stress responses predicted by This prospective study is the first to describe the pro- social anxiety broadened from T2 to T3, although in both spective links between social anxiety and stress responses instances adaptive as well as maladaptive responses were in adolescence. A strong point of the study is its longitudi- involved. Increased primary engagement coping and invol- nal design with three measurement waves. There are also untary engagement appeared as new consequences of social a number of limitations of this study. First, all measures anxiety. Taking primary engagement coping, this occurrence used adolescents’ self-report, which poses the problem of may be associated with the growing flexibility and diversity of single source bias. Although social anxiety and responses coping responses in older adolescents (Zimmer-Gembeck & to stress mostly represent internal processes that only partly Skinner, 2011). Possibly, they learn these adaptive responses are accessible to others, information obtained from other from their friends (Reindl et  al., 2016) seeing that some sources such as parents, clinical interviews and behavioral 1 3 Research on Child and Adolescent Psychopathology (Eds.), Social Anxiety and Phobia in Adolescents (pp. 151–181). observations, would have added to the validity of the study. Springer International Publishing.  https:// doi. org/ 10. 1007/ Second, the study used a relatively homogeneous sample 978-3- 319- 16703-9_7 of Dutch middle-class youth. It is therefore not clear if the Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. results can be generalized to a wider population. Third, the In R. G. Heimberg, M. R. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: Diagnosis, assessment, and treatment (pp. study did not select adolescents with SAD and therefore can- 69–93). Guilford Press. not draw conclusions about the relationship between stress Compas, B. E., Connor-Smith, J. K., Saltzman, H., Thomsen, A. H., & responses and clinical levels of social anxiety. Fourth, the Wadsworth, M. E. (2001). Coping with stress during childhood study did not take into consideration specific stressors that and adolescence: Problems, progress, and potential in theory and research. Psychological Bulletin, 127(1), 87. cause stress for adolescents, such as being neglected or being Compas, B. E., Connor, J., Osowiecki, D., & Welch, A. (1997). Effort- bullied. Including responses to specific stressors may add to ful and involuntary responses to stress. Coping with chronic stress a deeper understanding of the link between stress responses (pp. 105–130). Springer. and social anxiety in adolescence. Finally, as noted by an Compas, B. E., Jaser, S. S., Bettis, A. H., Watson, K. H., Gruhn, M. A., Dunbar, J. P., et al. (2017). Coping, emotion regulation, and anonymous reviewer, the present study cannot answer the psychopathology in childhood and adolescence: A meta-analysis question to what degree the study’s results are specific to and narrative review. Psychological Bulletin, 143(9), 939. social anxiety or apply to anxiety in general. Connor-Smith, J. K., Compas, B. E., Wadsworth, M. E., Thomsen, A. In conclusion, the present study suggests that adolescents’ H., & Saltzman, H. (2000). Responses to stress in adolescence: Measurement of coping and involuntary stress responses. Jour- responses to stress and social anxiety development are inter- nal of Consulting and Clinical Psychology, 68(6), 976. woven. Adding stress responses to developmental models of Davey, G. C. (1994). Trait factors and ratings of controllability as social anxiety may result in a more complete understanding predictors of worrying about significant life stressors. Personal- of social anxiety in adolescence. ity and Individual Differences, 16(3), 379–384. Epkins, C. C., & Heckler, D. R. (2011). Integrating etiological mod- els of social anxiety and depression in youth: Evidence for a Supplementary Information The online version contains supplemen- cumulative interpersonal risk model. Clinical Child and Family tary material available at https://doi. or g/10. 1007/ s10802- 021- 00880-3 . Psychology Review, 14(4), 329–376. Erath, S. A., Flanagan, K. S., & Bierman, K. L. (2007). Social anxi- Ackowledgment We thank Marjolein Fokkema for her assistance in ety and peer relations in early adolescence: Behavioral and cog- conducting the cross-lagged panel analysis. nitive factors. Journal of Abnormal Child Psychology, 35(3), 405–416. Funding No funds, grants, or other support was received. Flynn, M., & Rudolph, K. D. (2011). Stress generation and adolescent depression: Contribution of interpersonal stress responses. Jour- nal of Abnormal Child Psychology, 39(8), 1187–1198. Compliance with Ethical Standards Furman, W., & Buhrmester, D. (1992). Age and sex differences in perceptions of networks of personal relationships. Child Develop- Ethics Approval The Medical Ethical Committee of Leiden University ment, 63(1), 103–115. approved the study. Grant, K. E., Compas, B. E., Thurm, A. E., McMahon, S. D., & Gipson, P. Y. (2004). Stressors and Child and Adolescent Psychopathology: Consent to Participate Parents gave their written consent and youth Measurement Issues and Prospective Effects. Journal of Clinical their written assent for participation in the study. Child & Adolescent Psychology, 33(2), 412–425. https:// doi. org/ 10. 1207/ s1537 4424j ccp33 02_ 23 Conflicts of Interests The authors have no relevant financial or non- Hemphill, J. F. (2003). Interpreting the magnitudes of correlation coef- financial interests to disclose. ficients. American Psychologist, 58(1), 78–79. https:// doi.or g/10. 1037/ 0003- 066X.58.1. 78 Open Access This article is licensed under a Creative Commons Attri- Kaeppler, A. K., & Erath, S. A. (2017). Linking social anxiety with social bution 4.0 International License, which permits use, sharing, adapta- competence in early adolescence: Physiological and coping mod- tion, distribution and reproduction in any medium or format, as long erators. Journal of Abnormal Child Psychology, 45(2), 371–384. as you give appropriate credit to the original author(s) and the source, Kline, R. B. (2005). Principles and practice of structural equation provide a link to the Creative Commons licence, and indicate if changes modeling: Methodology in the social sciences. Guilford Press. were made. The images or other third party material in this article are Kovacs, M. (1985). The Children Depression Inventory (CDI). Psy- included in the article's Creative Commons licence, unless indicated chopharmacology Bulletin, 21, 995–998. otherwise in a credit line to the material. If material is not included in LaGreca, A. M. (1998). Social anxiety scales for children and adoles- the article's Creative Commons licence and your intended use is not cents: Manual and instructions for the SASC, SASC–R, SAS–A permitted by statutory regulation or exceeds the permitted use, you will (adolescents), and parent versions of the scales. University of need to obtain permission directly from the copyright holder. To view a Miami, Department of Psychology. copy of this licence, visit http://cr eativ ecommons. or g/licen ses/ b y/4.0/ . LaGreca, A. M., & Lopez, N. (1998). Social anxiety among adoles- cents: Linkages with peer relations and friendships. Journal of Abnormal Child Psychology, 26(2), 83–94. References LaGreca, A. M., & Prinstein, M. J. (1999). Peer group In W. K. Silverman, & T. H. Ollendick (Eds.), Developmental issues in the clinical treat- Blöte, A. W., Miers, A. C., Heyne, D. A., & Westenberg, P. M. (2015). ment of children (pp. 171–198). Allyn & Bacon. Social Anxiety and the School Environment of Adolescents. In Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. K. Ranta, A. M. La Greca, L.-J. Garcia-Lopez, & M. Marttunen Springer publishing company. 1 3 Research on Child and Adolescent Psychopathology Magee, W. J., Eaton, W. W., Wittchen, H.-U., McGonagle, K. A., & Silk, J. S., Steinberg, L., & Morris, A. S. (2003). Adolescents’ emotion Kessler, R. C. (1996). Agoraphobia, simple phobia, and social regulation in daily life: Links to depressive symptoms and prob- phobia in the National Comorbidity Survey. Archives of General lem behavior. Child Development, 74(6), 1869–1880. Psychiatry, 53(2), 159–168. Spence, S. H., & Rapee, R. M. (2016). The etiology of social anxi- McLellan, L. F., Alfano, C. A., & Hudson, J. L. (2015). Cognition- ety disorder: An evidence-based model. Behaviour Research and focused interventions for social anxiety disorder among adoles- Therapy, 86, 50–67. cents. Social Anxiety and Phobia in Adolescents (pp. 225–250). Starr, L. R., & Davila, J. (2008). Differentiating interpersonal corre- Springer. lates of depressive symptoms and social anxiety in adolescence: Miers, A. C., Blöte, A. W., de Rooij, M., Bokhorst, C. L., & Westenberg, Implications for models of comorbidity. Journal of Clinical Child P. M. (2013). Trajectories of Social Anxiety during Adolescence & Adolescent Psychology, 37(2), 337–349. and Relations with Cognition, Social Competence, and Tempera- Stein, D. J., Lim, C. C., Roest, A. M., De Jonge, P., Aguilar-Gaxiola, S., ment. Journal of Abnormal Child Psychology, 41(1), 97–110. Al-Hamzawi, A., et al. (2017). The cross-national epidemiology https:// doi. org/ 10. 1007/ s10802- 012- 9651-6 of social anxiety disorder: Data from the World Mental Health Ollendick, T. H., & Hirshfeld-Becker, D. R. (2002). The developmental Survey Initiative. BMC Medicine, 15(1), 143. psychopathology of social anxiety disorder. Biological Psychiatry, Steinberg, L. (2010). A dual systems model of adolescent risk-taking. 51(1), 44–58. https:// doi. org/ 10. 1016/ s0006- 3223(01) 01305-1 Developmental Psychobiology: THe Journal of the International R-Core-Team. (2019). R: A language and environment for statisti- Society for Developmental Psychobiology, 52(3), 216–224. cal computing. R Foundation for Statistical Computing. https:// Storch, E. A., Masia-Warner, C., Dent, H. C., Roberti, J. W., & Fisher, www.R- proje ct. org/. Accessed. P. H. (2004). Psychometric evaluation of the Social Anxiety Scale Ranta, K., Junttila, N., Laakkonen, E., Uhmavaara, A., La Greca, A. for Adolescents and the Social Phobia and Anxiety Inventory for M., & Niemi, P. M. (2012). Social Anxiety Scale for Adolescents Children: Construct validity and normative data. Journal of Anxi- (SAS-A): Measuring social anxiety among Finnish adolescents. ety Disorders, 18(5), 665–679. Child Psychiatry & Human Development, 43(4), 574–591. Timbremont, B., & Braet, C. (2002). Children’s Depression Inventory: Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral Nederlandstalige versie [Children’s Depression Inventory: Dutch model of anxiety in social phobia. Behaviour Research and Ther- version. Swets & Zeitlinger. apy, 35(8), 741–756. Van den Bos, E., De Rooij, M., Miers, A. C., Bokhorst, C. L., & Westenberg, Reijntjes, A., Kamphuis, J. H., Prinzie, P., & Telch, M. J. (2010). Peer P. M. (2014). Adolescents’ Increasing Stress Response to Social Evalu- victimization and internalizing problems in children: A meta- ation: Pubertal Effects on Cortisol and Alpha-Amylase During Public analysis of longitudinal studies. Child Abuse & Neglect, 34(4), Speaking. Child Development, 85(1), 220–236. https://doi. or g/10. 1111/ 244–252. https:// doi. org/ 10. 1016/j. chiabu. 2009. 07. 009 cdev. 12118 Reindl, M., Gniewosz, B., & Reinders, H. (2016). Socialization of emo- Wadsworth, M. E., & Berger, L. E. (2006). Adolescents coping with tion regulation strategies through friends. Journal of Adolescence, poverty-related family stress: Prospective predictors of coping 49, 146–157. and psychological symptoms. Journal of Youth and Adolescence, Richardson, C. E., Magson, N. R., Fardouly, J., Oar, E. L., Forbes, M. 35(1), 54–67. K., Johnco, C. J., et al. (2020). Longitudinal associations between Westenberg, P. M., Bokhorst, C. L., Miers, A. C., Sumter, S. R., Kallen, coping strategies and psychopathology in pre-adolescence. Jour- V. L., van Pelt, J., et al. (2009). A prepared speech in front of a nal of Youth and Adolescence, 1–16. pre-recorded audience: Subjective, physiological, and neuroendo- Richey, J. A., Brewer, J. A., Sullivan-Toole, H., Strege, M. V., Kim-Spoon, crine responses to the Leiden Public Speaking Task. Biological J., White, S. W., et al. (2019). Sensitivity shift theory: A develop- Psychology, 82(2), 116–124. https://doi. or g/10. 1016/j. biops y cho. mental model of positive affect and motivational deficits in social 2009. 06. 005 anxiety disorder. Clinical psychology review, 101756. Wittchen, H. U., & Fehm, L. (2003). Epidemiology and natural course Roelofs, J., Braet, C., Rood, L., Timbremont, B., Van Vlierberghe, L., of social fears and social phobia. Acta Psychiatrica Scandinavica, Goossens, L., et al. (2010). Norms and screening utility of the 108(s417), 4–18. Dutch version of the Children’s Depression Inventory in clinical Wright, M., Banerjee, R., Hoek, W., Rieffe, C., & Novin, S. (2010). and nonclinical youths. Psychological Assessment, 22(4), 866. Depression and social anxiety in children: Differential links with Rosseel, Y. (2012). Lavaan: An R package for structural equation mod- coping strategies. Journal of Abnormal Child Psychology, 38(3), eling and more. Version 0.5–12 (BETA). Journal of statistical 405–419. software, 48(2), 1–36. Zimmer-Gembeck, M. J., & Skinner, E. A. (2011). The development of Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the coping across childhood and adolescence: An integrative review state of the art. Psychological Methods, 7(2), 147. and critique of research. International Journal of Behavioral Sentse, M., Lindenberg, S., Omvlee, A., Ormel, J., & Veenstra, R. Development, 35(1), 1–17. (2010). Rejection and acceptance across contexts: Parents and peers as risks and buffers for early adolescent psychopathology. Publisher's Note Springer Nature remains neutral with regard to The TRAILS study. Journal of abnormal child psychology, 38(1), jurisdictional claims in published maps and institutional affiliations. 119–130. 1 3

Journal

Journal of Abnormal Child PsychologySpringer Journals

Published: Oct 18, 2021

Keywords: Adolescence; Social anxiety; Coping; Stress responses

References