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

Learn More →

Associations Between Attentional Bias and Interpretation Bias and Change in School Concerns and Anxiety Symptoms During the Transition from Primary to Secondary School

Associations Between Attentional Bias and Interpretation Bias and Change in School Concerns and... The transition from primary to secondary school is often associated with a period of heightened anxiety and worry. For most children, any feelings of anxiety subside relatively quickly but for a small minority, emotional difficulties can continue into the first year of secondary school and beyond. This study recruited 109 children and measured their anxiety symptoms and school concerns toward the end of primary school and again at the end of their first term of secondary school. We investigated for the first time whether pre-transition measures of attentional and interpretation bias, and the magnitude of change in attentional bias toward and away from threat stimuli were associated with pre- and post-transition measures of anxiety and school concerns, and the change in these measures over time. Over 50% of the current sample exceeded clinical levels of anxiety at pre-transition. However, anxiety symptoms and school concerns had significantly reduced by post-transition. Higher levels of pre-transition anxiety or school concerns, and a greater magnitude of change in attentional bias towards threat stimuli predicted a larger reduction in anxiety symptoms and school concerns across the transition period. A greater interpretation bias toward threat was associated with higher pre-transition anxiety symptoms and school concerns but not post-transition scores, or the change in these scores. While many children experience heightened anxiety prior to school transition, this appears to be largely temporary and self-resolves. Nonetheless, the current findings highlight the importance of monitoring children’s anxiety and concerns, and related cognitive processes during this important transition period. . . . Keywords School transition Anxiety Attentional Bias Interpretation Bias In the UK, most children move from primary to secondary School transition has been shown in some studies to school at 11 years of age. At the time of transition, many negatively impact children’s emotional wellbeing children experience worry and anxiety, which typically sub- (Anderson et al. 2000; Smith et al. 2008) although such sides relatively quickly over the first term of secondary school studies are relatively sparse and results somewhat incon- (Rice et al. 2011; Stradling and MacNeil 2001; Zeedyk et al. sistent (Evans et al. 2018; Grills-Taquechel et al. 2010). 2003). For a minority of children, emotional difficulties en- Nonetheless, school concerns and anxiety symptoms ap- dure, continuing into and beyond the first year of secondary pear strongly correlated at primary and secondary school education (Zeedyk et al. 2003). (Rice et al. 2011) and emotional difficulties around the transition period are associated with concurrent and pro- Electronic supplementary material The online version of this article spective school attainment (Riglin et al. 2014), suggesting (https://doi.org/10.1007/s10802-019-00528-3) contains supplementary school transition may have a longer term negative impact material, which is available to authorized users. on child outcomes for some. Children with enduring emo- tional symptoms following the school transition may rep- * Kathryn J. Lester resent a particularly vulnerable group (Riglin et al. 2013). k.lester@sussex.ac.uk The school transition period thus represents a critical time for identifying children at increased risk for poor adjust- School of Psychology, University of Sussex, Pevensey Building, ment, and for nurturing children’s mental health, especial- Brighton BN1 9QH, UK ly since this developmental stage is associated with the Institute of Psychiatry, Psychology and Neuroscience, King’s onset of many anxiety disorders (Kessler et al. 2005). College London, De Crespigny Park, London SE5 8AF, UK 1522 J Abnorm Child Psychol (2019) 47:1521–1532 Numerous risk and protective factors for academic, behav- modification task) early in the semester predicted a larger in- ioural and emotional adjustment to school transition have crease in anxiety in students by the end of their first semester at been identified (Evans et al. 2018). Being younger, female, university. A second study revealed that change in attentional having lower socioeconomic status, being less academically bias away from threat was not predictive of subsequent chang- able, and experiencing stressful life events predict greater risk es in anxiety by the end of the semester. This suggests that for poor adjustment (Anderson et al. 2000;Riceet al. 2011; increases in anxiety in response to mild extended stress were West et al. 2010). Likewise peer factors (e.g. lack of peer not explained by a high level of general attentional plasticity or support, experiences of victimisation, stability in friendships) a tendency to develop attentional avoidance of threat but were and family relationships (maternal depressive symptoms, par- specifically determined by the degree to which individuals enting styles characterised by warmth) are also significant were prepared to acquire an attentional bias toward threat. predictors of children’s adjustment and academic attainment While Clarke and colleagues’ conclusions are compelling, (Hirsch and DuBois 1992; Ng-Knight et al. 2016;Rice etal. a stronger test of their competing hypotheses requires a 2015;Westet al. 2010). With regard to emotional predictors, within-participant design where all participants are exposed one study showed that heightened symptoms of generalised to both bias contingencies (toward threat, away from threat). anxiety prior to transition was associated with greater con- If individuals who show the largest change in attentional bias cerns about school transition both before and after transition toward threat also show the largest change in attentional bias (Rice et al. 2011). A meta-analysis also found that emotional away from threat then this pattern might indicate that some difficulties significantly predicted poorer school attainment individuals’ attentional processing is generally more mallea- (Riglin et al. 2014). ble and responsive to environmental contingencies favouring There is reasonable evidence that heightened symptoms of selective attention both toward and away from threat. The anxiety are associated with cognitive biases in attention and inclusion of both contingency conditions also allows us to interpretation that favour the selective processing of threat unpick whether change in anxiety symptoms/school concerns (Bar-Haim et al. 2007; Dudeney et al. 2015; Lau and Waters is best explained by a.) the degree to which individuals are 2017; Suarez and Bell-Dolan 2001). These biases are thought prepared to acquire an attentional bias bias specifically to- to contribute causally to the onset and/or maintenance of wards threat, b.) the degree to which individuals acquire an symptoms (Van Bockstaele et al. 2014), and to impact emo- attentional bias characterised by avoidance of threat, or c.) is tional response to stressors (Hakamata et al. 2010;Osinsky indexed equally well by both, indicating a more general atten- et al. 2012), although this evidence is somewhat more mixed tional malleability effect. A further rationale for including the in child samples (Dodd et al. 2012; Dudeney et al. 2015). This avoid threat condition was to explore direction of effects. is the first study to investigate whether attentional and inter- Some studies have shown that attentional avoidance of threat pretation biases assessed toward the end of primary school are is associated with heightened anxiety (Brown et al. 2013; associated with anxiety symptoms and school concerns over Stirling et al. 2006); whereas, others have shown that atten- the school transition period. tional avoidance of threat is associated with a reduction in The degree to which people respond to stressors with in- symptoms of anxiety (e.g. Legerstee et al. 2009). creases in anxiety differs substantially (Clarke et al. 2008). This study investigated the impact of the school transition The basis for this variability is poorly understood but it has period on children’s concerns about school and their anxiety been argued that exposure to stressful life events may make it symptoms. To do so, we recruited children during the final year temporarily adaptive to develop an attentional bias for threat of primary school and assessed the degree to which school con- cues that may signal genuine danger (MacLeod 1999). As a cerns and anxiety symptoms changed between the first measure- consequence, individuals who demonstrate the greatest ment in primary school and the second measurement at the end change in this capacity to selectively direct attention toward of the first term at secondary school. As noted previously, a small threat might be expected to display the most intense and number of studies have shown that school transition negatively sustained anxiety reactions to a subsequent stressor. impacts on emotional wellbeing, although any emotional diffi- Attentional bias modification studies which use experimental culties typically subside relatively quickly. In light of this, and in contingencies to encourage selective attention to threat stimuli line with prior research (Rice et al. 2011) we began by testing the have shown that those individuals who acquire a larger atten- hypothesis that school concerns, and anxiety symptoms would tional bias for threat also report greater anxiety in response to a decrease significantly over the transition period and that anxiety subsequent experimental stressor (MacLeod and Clarke and school concern scores at pre-transition would correlate sig- 2015). However, findings have been more equivocal in child nificantly with scores at post-transition (Hypothesis 1). Given the samples (Cristea et al. 2015). relatively extended nature of the school transition period and the Consistent with research using lab-based stressors, Clarke time period between assessments in the present study, we focused et al. (2008) found that a greater change in attentional bias on anxiety symptoms using the Screen for Child Anxiety Related toward threat stimuli (evoked using an attentional bias Disorders (SCARED), rather than on trait or state anxiety per se. J Abnorm Child Psychol (2019) 47:1521–1532 1523 Prior research indicates that the SCARED taps into elements of Children attended two 45-min sessions, 1 week apart dur- both trait and state anxiety, but may be more strongly correlated ing the last term of primary school. One hundred and six with measures of trait anxiety (Monga et al. 2000). participants completed both sessions, and 79 participants Given the reasonable evidence that cognitive biases (75% response rate) returned a follow-up questionnaire to- favouring the selective processing of threat are associated with wards the end of their first term of secondary school. heightened levels of anxiety (e.g. Lau and Waters 2017), we Compared to responders, non-responders to the follow-up hypothesised that a stronger tendency to interpret ambiguous questionnaire did not differ significantly on age, sex, ethnicity, situations as threatening (Hypothesis 2) and a greater atten- anxiety symptoms, school concerns, interpretation bias or at- tional bias toward threat (Hypothesis 3) prior to transition tentional bias at pre-transition (all p values > 0.05). would be associated with higher school concerns, and anxiety A priori power calculations were performed for the associ- symptoms before and after school transition. Past research has ation between change in attentional bias toward threat and also suggested that individuals who show a greater change in change in anxiety scores across the school transition. Effect their capacity to selectively direct attention toward threat (i.e. size estimates were taken from Clarke et al. (2008) who ob- to acquire an attentional bias toward threat) but not away from served a correlation of r= 0.47 between attentional bias threat (i.e. to acquire an attentional bias away from threat) may change and anxiety change in their sample of students attend- display more intense anxiety reactions to stressors like school ing the first semester at University. With an effect size of r = transition (Clarke et al. 2008). Uniquely, we explored in the 0.47, and to achieve 80% power with α = 0.008 (to allow for same sample whether the magnitude of change in a.) atten- multiple testing corrections, see BStatistical Analysis^ section tional bias toward threat and b.) attentional bias away from below) required a minimum N =50. threat was associated with school concerns, and anxiety symp- toms before and after school transition (Hypothesis 4). Finally, we explored whether our pre-transition measures of anxiety Measures symptoms, school concerns, interpretation bias, attentional bi- as and change in attentional bias toward and away from threat Anxiety Symptoms Anxiety symptoms (e.g. BI worry about predicted change in a.) school concerns and b.) anxiety symp- other people liking me, When I get frightened I feel dizzy^) toms across the transition period (Hypothesis 5). in the preceding 3 months were measured using the 41-item Screen for Child Anxiety Related Emotional Disorders (SCARED, (Birmaher et al. 1999). Responses were made Method using a 3-point scale (‘not true or hardly ever true’–‘very true or often true’). Internal consistency was excellent (α = Participants 0.85 and α = 0.90 at pre- and post-transition, respectively). One hundred and nine children (mean age = 10.7 years, SD = School Concerns Concerns about secondary school were mea- 0.5, 53% female) were recruited from seven mainstream sured using the 17-item School Concerns Questionnaire schools. Schools were invited pragmatically on the basis of (Thomasson et al. 2006). Children rated their degree of con- being mainstream primary schools situated within a relatively cern for each item (e.g. ‘being bullied’, ‘following a timeta- broad area of Greater London but within a reasonable travel- ble’) using a 10-point scale (1 = ‘not worried’ to 10 = ‘ex- ling distance using public transport from King’s College tremely worried’). Internal consistency was good (α = 0.92 London. Ethical approval was granted by King’s College and 0.79 at pre- and post-transition). London Psychiatry, Nursing and Midwifery Research Ethics Committee (PNM/14/15–66). Parents gave written informed Ambiguous Situations Questionnaire – School Transition Ten consent and children verbal assent. Inclusion criteria included ambiguous scenarios were presented describing typical sec- being in the final year of primary school. Parents were advised ondary school situations (see Supplementary Materials for that the study was not suitable for children with significant measure development). For example, BYou are in a PE lesson learning disabilities or insufficient understanding of the at your new school. Your teacher chooses two team captains English language to comprehend the study materials. Forty- and asks them to pick teams for a basketball game. You wait six percent of children identified as White British/other White for yournametobecalledout^. Participants responded to an background, 14% as African, and 7% as other mixed/multiple open-ended question designed to elicit their interpretation of ethnic background. Fifty-five percent of parents were educat- the scenario (BWhen do you think your name will be ed to university degree level or above. Overall the sample was called?^). Scenarios were presented randomly on a laptop more highly educated compared to the UK national average screen using EPrime 2.0. Free choice responses were record- but was ethnically diverse and broadly representative with ed, transcribed and coded using an established coding ap- regard to ethnicity for London. proach (K. J. Lester et al. 2010). 1524 J Abnorm Child Psychol (2019) 47:1521–1532 Participants were also presented simultaneously with a questionnaire administration was randomised, followed threat (BYou think your name will be called near the end as by completion of the ambiguous situations questionnaire, the team captains won’t want you on their team^) and non- then attentional bias change task. In session 2, participants threat (BYou think your name will be called very soon as the completed the opposite attentional contingency phase team captains will want you on their team^) forced-choice condition to that given in session 1. Participants received interpretation. Participants indicated which interpretation they a follow-up questionnaire toward the end of their first thought was most likely to be the outcome of each situation term of secondary school. Participants received a small using counterbalanced response keys. Free and forced-choice craft gift and a £10 gift card at the end of sessions 1 response formats were highly correlated, r (109) = 0.67, and 2, respectively, and a £10 gift card if they returned p < 0.001. Therefore, a combined threat interpretation score the follow-up questionnaire. (range 0–100%) was computed: (total free + forced-choice threat interpretations/total number of valid interpretations) × Data Processing Changes in school concerns (SCQ ) and anxiety CHANGE Attentional Bias Change Task This task comprised of three symptoms (SCARED ) were computed by CHANGE phases, each using a dot-probe or modified dot-probe task. subtracting pre-transition scores from post-transition The first and third phases (the pre- and post-contingency bias scores. We used an established approach to data cleaning assessment phases) assessed relative attentional allocation to of dot-probe reaction time data with children, and to deal- threat and neutral stimuli. The second phase (the attentional ing with participants with outlying responses compared to contingency block) exposed participants to a contingency de- the sample mean or high error rates (Brown et al. 2014). signed to elicit an attentional bias favouring threat stimuli This approach was decided upon prior to data collection (attentional contingency – toward threat condition) or and was used to remove trials with extreme (short or long) favouring neutral stimuli (attentional contingency – avoid response times compared to the individuals mean reaction threat condition). time, thus representing trials on which anticipatory re- Stimuli comprised of forty models (20 males and 20 fe- sponses were made or delayed responses, which may re- males) portraying angry and neutral facial expressions select- flect distraction/lack of attentiononthattrial. Meanreac- ed from established face sets (Biehl et al. 1997; Langner et al. tion times (RTs) were calculated for pre- and post- 2010; Tottenham et al. 2009). The stimuli were divided into contingency assessments of attentional bias, after remov- four subsets balanced for male and female faces. The alloca- ing errors (2.8% of trials), and data values >2.5 SDs from tion of face sets and order of contingency conditions was individual means, or < 100 ms (4.1% of trials). Bias scores counterbalanced using random block allocation, and blinded were calculated for pre- and post-contingency assessment to the experimenter. Participants received the same face set in phases by subtracting mean RTs for probes presented in the pre- and post-contingency bias assessment phases (within the angry face locus from mean RTs for probes presented session) and a different subset for each attentional contingen- in the neutral face locus, resulting in a measure of partic- cy block. ipant’s attentional bias prior to contingency exposure During the attentional contingency phase, participants (AB ), and after contingency exposure PRE-CONTINGENCY viewed 280 trials (240 angry-neutral, 40 neutral-neutral trials). (AB ). A positive score indicates a bias POST-CONTINGENCY The task procedure is depicted in Fig. 1 (full details in the toward threat while a negative score indicates a bias away Supplementary Information). In the toward threat condition, from threat. Participants who made incorrect or outlying the probe (< or >) consistently appeared in the location previ- responses on more than 25% of trials were excluded from ously occupied by the angry face of angry-neutral face pairs. analyses using bias score data. Three participants were In the away from threat condition, the probe consistently ap- excluded from analyses in which pre-attentional bias peared in the location previously occupied by the neutral face. scores were correlated with pre- or post-transition anxiety The pre- and post-contingency assessment phases comprised or school concern scores or the change in those scores. of 120 trials (40 neutral-neutral and 80 angry-neutral trials) Outliers comprised of participants who had mean bias where the probe appeared with equal probability behind the scores exceeding 2.5 SD above or below the sample mean angry and neutral stimulus. Reaction times (RTs) and accuracy or participants who made incorrect or outlying responses of responses was recorded. on more than 25% of trials used to calculate bias scores. Change in attentional bias (AB ) was computed for CHANGE Procedure the Battend toward threat^ condition (AB CHANGE–TOWARD )bysubtractingAB scores from THREAT PRE-CONTINGENCY The procedure is outlined in Fig. 2. Sessions 1 and 2 were AB scores. For the Battend away from POST-CONTINGENCY conducted in a quiet study space. In session 1, threat^ condition (AB )AB CHANGE–AVOID THREAT POST- J Abnorm Child Psychol (2019) 47:1521–1532 1525 Fig. 1 Attentional bias change task procedure Fixation cross: displayed for 500ms Face pair: angry-neutral Inter-trial interval: or neutral-neutral 500ms displayed for 500ms Probe (< or >): remained on screen until mouse button response detected scores were subtracted from AB approximated a normal distribution. Sensitivity analyses were CONTINGENCY PRE- scores. In both conditions larger positive scores conducted in which participants who were outliers for atten- CONTINGENCY on the AB index reflected a greater degree of change tional bias scores or for AB or CHANGE CHANGE-TOWARD THREAT in attentional bias in the direction encouraged during the con- AB scores were not removed from anal- CHANGE-AVOID THREAT tingency phase. For analyses correlating AB scores yses. Overall, the substantive conclusions remained un- CHANGE with pre-transition measures of anxiety symptoms or school changed despite minor fluctuations in effect size and p value concerns, eight participants were excluded due to having (see Supplementary Materials). mean change scores exceeding 2.5 SD above or below the sample mean or because they made incorrect or outlying re- Statistical Analysis sponses on more than 25% of trials used to calculate bias scores. For analyses correlating AB scores with post- We examined bivariate correlations between all variables CHANGE transition measures of anxiety symptoms or school concerns both within and across time points and between anxiety or change in these outcomes, six participants were excluded symptom and school concern change scores and all pre- for the same reasons. After data processing all variables transition variables. Bonferroni corrections were applied Fig. 2 Experimental procedure First term of secondary Last term of primary school school Session 1 (N = 109) Session 2 (N = 106) Follow-Up Questionnaire (N = 79) Anxiety (SCARED) AB Change Task School summer holidays School Concerns (SCQ) Anxiety (SCARED) Interpretation Bias School Concerns (SCQ) (ASQ) Attentional Bias AB Change Task Mean interval between session 1 and session 2 = 7.5 days, SD = 1.7. Mean interval between school sessions and follow-up questionnaire =187 days, SD = 50.6. 1526 J Abnorm Child Psychol (2019) 47:1521–1532 to control for the number of tests performed. Changes in Results anxiety symptoms and school concerns were analysed using ordinary least squares (OLS) linear regression. Hypothesis 1: Correlations Between Anxiety Assumptions for OLS regression were carefully checked. Symptoms and School Concerns at Pre- First, scatterplots were checked and showed that relation- and Post-Transition ships between the independent and dependent variables were linear. Second, there was no evidence of We hypothesised that school concerns and anxiety symptoms multicollinearity (correlations between independent vari- would decrease significantly over the transition period and ables were all lower than 0.8 and variance inflation fac- that anxiety and school concern scores at pre-transition would tors ranged 1.0–1.9). Third, plots of standardised resid- correlate significantly with scores at post-transition. Pre- uals vs. standardised predicted values showed no obvious transition mean scores for anxiety and school concerns (see signs of funnelling suggesting the assumption of homo- Table 1) were approximately 0.5 SD above reported norms scedasticity was met. Fourth, all Cook’s distance values while scores at post-transition were comparable to previous were under 1 (range 0.0–0.5) suggesting that no individ- reports for school concerns (Rice et al. 2011) and were 0.25 ual cases were unduly influencing the model. Finally, SD below the norms for non-anxiety cases (Birmaher et al. normal probability plots of the standardised residuals 1999). Consistent with this hypothesis, we observed a signif- for each model indicated some mild deviations from nor- icant reduction in school concerns and anxiety symptoms mality. While mild deviations from normality are unlike- from pre- to post-transition (Table 1). Nonetheless, there was ly to impact on the validity of our findings, as a sensi- substantial inter-individual variability. Anxiety symptoms and tivity analysis we re-ran the regression models using school concerns were moderately correlated at pre- and post- bootstrapped 95% bias corrected and accelerated confi- transition (see Table 2) and, consistent with our hypothesis, dence intervals (with 1000 bootstrap iterations). These there was continuity such that individuals with higher scores analyses gave results that were consistent to the models at pre-transition also reported higher scores at post-transition. without bootstrapping and did not change our substantive conclusions (full results available on request). Hypothesis 2: Associations Between Interpretation We included as covariates pre-transition variables that Bias and Anxiety Symptoms and School Concerns were significantly associated with the outcome at Pre- and Post-Transition (SCARED and SCQ ) in the bivariate cor- CHANGE CHANGE relations (p < 0.05). When a significant covariate is iden- We hypothesised that a greater interpretation bias favouring tified, it is significant over and above the other covariates threat would predict higher levels of anxiety symptoms and in the model. In all analyses involving SCARED CHANGE school concerns at pre- and post-transition. In support of this and SCQ scores we controlled for the pre- CHANGE hypothesis we found that a stronger tendency toward transition level of the outcome. For the change score interpreting ambiguous situations as threatening was signifi- ΔY = Y − Y ,weincludeY as a covariate in the model; 2 1 1 cantly associated with greater school concerns and anxiety for bivariate correlations, we estimate partial correlations, symptoms prior to school transition (see Table 2). However controlling for Y Controlling for pre-transition scores is threat interpretation bias did not significantly correlate with important because of correlations between pre-transition anxiety symptoms or school concern scores at post-transition. scores on the outcome and exposure (i.e. Y and X ). 1 1 Pre-transition outcome scores (Y e.g. pre-transition SCARED scores) are therefore correlated with both pre- Hypothesis 3: Associations Between Attentional Bias transition exposures (X e.g. interpretation bias scores) and Anxiety Symptoms and School Concerns at Pre- and the post-transition outcome (Y e.g. post-transition and Post-Transition SCARED scores). This represents a form of confounding that must be controlled for (Pearl 2016). Failure to ac- We hypothesised that a greater attentional bias for threat count for correlations between Y and X will lead to would predict higher levels of anxiety symptoms and school 1 1 biased results. This issue only arises when Y and X concerns at pre- and post-transition. The mean bias score was 1 1 are correlated (i.e. when Y predicts both X and ΔY −3.6 (SD = 36.2) indicating no significant attentional bias in 1 1 (e.g. SCARED score), and thus, confounds the either direction (difference from 0, t (109) = −1.03, p =0.304). CHANGE association between X and ΔY). Where X and Y are There were substantial individual differences in bias scores 1 1 1 uncorrelated (or are weakly correlated), there is no con- (range: −127.33 to 120.18). Contrary to our hypothesis, how- founding due to Y , and the results are not affected by ever, pre-transition attentional bias was not significantly asso- controlling for Y . However, for consistency, all estimates ciated with pre-or post-transition anxiety symptom or school have been adjusted for baseline outcome scores (Y ). concern scores (see Table 2). 1 J Abnorm Child Psychol (2019) 47:1521–1532 1527 Table 1 Descriptive and test statistics for school concerns, and anxiety symptoms at pre- and post-school transition Measure Pre-transition Post-transition Change score t (df) p Mean (SD) Mean (SD) Mean (SD) [95% CI] School concerns (SCQ) 78.21 (32.36) 46.30 (24.58) −29.32 (27.97) −9.32 (78) < 0.001 (n =106) (n =78) [−35.58 – −23.05] Anxiety symptoms (SCARED) 26.70 (14.72) 16.10 (13.79) −9.17 (12.51) −6.43 (76) < 0.001 (n =106) (n =76) [−12.01 – −6.33] Interpretation Bias (ASQ ) 32.93 (19.13) –– – – PRE (n =109) Attentional Bias (S1 AB ) −3.63 (36.19) –– – – PRE-CONTINGENCY (n =106) AB −0.05 (51.45) –– – – CHANGE – TOWARD THREAT (n =101) AB −6.04 (56.43) –– – – CHANGE – AVOID THREAT (n =98) SCQ School Concerns Questionnaire, SCARED Screen for Child Anxiety Related Emotional Disorders, ASQ Ambiguous Situations Questionnaire, AB Attentional Bias Hypothesis 4: Associations Between Attentional Bias (see Table 2). Larger changes in attentional bias toward threat Change and Anxiety Symptoms and School Concerns were significantly associated with lower anxiety symptoms at at Pre- and Post-Transition post-transition (but not school concerns), although this effect did not survive multiple testing corrections. AB CHANGE–AVOID We tested the hypothesis that the magnitude of change in scores were unrelated to pre-transition school con- THREAT attentional bias toward threat (AB ) cerns. However, higher anxiety symptoms at pre-transition CHANGE–TOWARD THREAT and magnitude of change in attentional bias away from threat was significantly associated with a smaller magnitude of at- (AB ) would be associated with school tentional bias change away from threat, although this too was CHANGE–AVOID THREAT concerns and anxiety symptoms at pre- and post-transition. no longer statistically significant after corrections for multiple AB scores were not significantly re- testing. No significant associations were observed with post- CHANGE– TOWARD THREAT lated to pre-transition school concerns or anxiety symptoms transition school concerns, or anxiety symptoms. No Table 2 Bivariate correlations between school concerns, anxiety symptoms and cognitive measures at pre- and post-school transition SCARED SCQ SCARED SCQ SCARED ASQ S1 AB AB AB PRE POST POST CHANGE CHANGE PRE PRE- CHANGE- CHANGE- CONTINGENCY TOWARD AVOID THREAT THREAT a ab a ab SCQ 0.687 0.509 0.024 -0.609 −0.022 0.542 0.013 −0.053 −0.053 PRE <0.001 <0.001 <0.001 <0.001 0.851 <0.001 0.895 0.600 0.606 ab a ab a SCARED – 0.250 0.633 0.229 -0.447 0.533 −0.140 −0.012 −0.219 PRE 0.030 <0.001 0.047 <0.001 0.000 0.153 0.908 0.030 ab ab ab ab SCQ – 0.672 –– -0.187 0.119 -0.159 -0.233 POST <0.001 0.106 0.309 0.189 0.054 ab ab ab ab SCARED –– – -0.170 0.123 -0.314 -0.076 POST 0.146 0.300 0.009 0.540 ab ab ab ab SCQ –– -0.153 0.065 -0.277 0.009 CHANGE 0.186 0.579 0.020 0.944 ab ab ab ab SCARED – -0.079 0.147 -0.392 0.078 CHANGE 0.502 0.215 0.001 0.531 SCQ School Concerns Questionnaire, SCARED Screen for Child Anxiety Related Emotional Disorders, ASQ Ambiguous Situations Questionnaire, AB a b Attentional Bias. Partial correlation coefficients controlling for interval in days between in-school sessions and return of follow-up questionnaires. Partial correlation coefficients controlling for interval in days between in-school sessions and return of follow-up questionnaires, and pre-transition level of the outcome. Correlation coefficients in bold survive multiple testing corrections: Bonferroni corrected p value for correlations with SCARED and PRE SCQ = 0.01; for correlations with SCARED and SCQ = 0.008; for correlations with SCARED and SCQ =0.008 PRE POST POST CHANGE CHANGE 1528 J Abnorm Child Psychol (2019) 47:1521–1532 significant correlation was observed between AB predicted a reduction of 7 points on the SCQ between pre- and CHANGE– and AB scores (r post-transition (β = −0.24; p =0.009). TOWARD THREAT CHANGE–AVOID THREAT (92) = −0.17, p = 0.110). Change in anxiety symptoms was significantly predicted by pre-transition anxiety symptoms and AB CHANGE-TOWARD scores. Higher pre-transition anxiety predicted a great- THREAT Hypothesis 5: Correlates and Prediction of Change er reduction in anxiety symptoms across the transition period. in Anxiety Symptoms and School Concerns A one SD increase in pre-transition anxiety score (14.23) pre- from Pre-Transition Variables dicted a reduction of 6 points on the SCARED between pre- and post-transition (β = −0.45; p < 0.001). Likewise, a one SD Our final hypothesis investigated whether pre-transition as- increase in AB scores (51.2ms) pre- CHANGE-TOWARD THREAT sessments (of anxiety symptoms, school concerns, interpreta- dicted a reduction of 4 points on the SCARED between pre- tion bias, attentional bias and change in attentional bias toward and post-transition (β = −0.35; p =0.001). and away from threat) predicted change in school concerns and anxiety symptoms over the transition period. Overall, we found good support for this hypothesis. Greater school con- cerns at pre-transition was significantly associated with a larg- Discussion er reduction in school concerns over the transition period (see Table 2) with the same pattern observed between anxiety This study explored whether attentional biases, the malleability scores. No significant correlations were observed between of attentional biases, and interpretation bias measured before pre-transition measures of interpretation bias, attentional bias, school transition explained individual differences in changes in or AB scores and SCQ and anxiety symptoms and school concerns over the transition peri- CHANGE-AVOID THREAT CHANGE SCARED scores. However, a greater change in atten- od. In support of hypothesis 1, we observed a significant decrease CHANGE tional bias toward threat stimuli (AB ) in anxiety symptoms and school concerns from pre- to post- CHANGE-TOWARD THREAT was significantly associated with a larger reduction in school transition. Anxiety and school concern scores at pre-transition concerns and anxiety symptoms over the transition period. also correlated significantly with post-transition scores. The correlation with SCQ did not survive multiple Hypothesis 2 was also partly supported with a greater threat CHANGE testing corrections. interpretation bias associated with higher pre-transition anxiety To test hypothesis 5, changes in school concerns and anx- symptoms and school concerns but not post-transition scores. iety symptoms were analysed using multiple regression. Only Hypothesis 3 was not supported as no significant associations pre-transition variables that were significantly associated with were observed between pre-transition attentional bias and pre- or the outcome (SCARED and SCQ ) in the bi- post- anxiety or school concern scores. We also found no con- CHANGE CHANGE variate correlations reported above (p < 0.05) were included in vincing support for hypothesis 4; attentional bias change toward each model. Change in school concerns was significantly as- andawayfromthreatwerenot significantly correlated with pre- sociated with pre-transition school concerns, pre-transition or post- transition anxiety symptom or school concern scores anxiety symptoms and AB scores after multiple testing corrections were applied. In support of hy- CHANGE-TOWARD THREAT (see Table 3). Higher pre-transition school concerns signifi- pothesis 5, we identified a small number of significant predictors cantly predicted a reduction in school concerns scores across of change in anxiety symptoms and school concerns across the the transition period (β = −0.75). Higher pre-transition anxiety transition period. Higher levels of pre-transition anxiety or school scores were positively associated with SCQ (β = concerns were associated with greater reductions in severity CHANGE 0.28). Increases in attentional bias toward threat predicted a across transition. A larger increase in attentional bias toward decrease in school concerns across the transition period. A one threat significantly predicted a larger reduction in anxiety symp- SD increase in AB scores (51.2ms) toms and school concerns across the school transition period. CHANGE–TOWARD THREAT Table 3 Regression analyses for change in school concerns and anxiety symptoms SCQ SCARED CHANGE CHANGE b [95% CI] β tp b [95% CI] β tp SCQ −0.75 [−0.97 – −0.52] −0.81 −6.69 <0.001 – ––– PRE SCARED 0.59 [0.08–1.10] 0.28 2.32 0.023 −0.39 [−0.57 – −0.22] −0.45 −4.47 <0.001 PRE AB −0.14 [−0.24 – −0.04] −0.24 −2.69 0.009 −0.09 [−0.13 – −0.04] −0.35 −3.49 0.001 CHANGE-TOWARD THREAT SCQ School Concerns Questionnaire, SCARED Screen for Child Anxiety Related Emotional Disorders, AB Attentional Bias J Abnorm Child Psychol (2019) 47:1521–1532 1529 The period towards the end of primary school appears to be predicts who will naturally develop an attentional bias for threat an especially stressful time with around half of our sample when exposed to a real-life extended stressor, in turn predicting exceeding the suggested clinical cut-off on the SCARED. who will experience increases in anxiety. However, unlike the We also observed continuity of symptoms, such that higher young adults who transitioned to university, children who anxiety symptoms at pre-transition were correlated with transitioned to secondary school reported a significant reduc- retaining heightened anxiety symptoms at post-transition. tion in anxiety over time. Notably, the current sample of chil- This is consistent with prior research showing that anxiety dren were considerably more anxious at the study outset com- symptoms in primary school predict later symptoms (Lester pared to the sample of young adults in Clarke et al. (2008). A et al. 2013). Higher anxiety scores pre-transition predicted a later study assessed change in attentional bias toward threat decrease in anxiety scores over time. The same effects were among individuals with social anxiety disorder prior to cogni- observed for school concerns. This pattern of results may re- tive behaviour therapy (Clarke et al. 2012). They also found, as flect a combination of regression toward the mean and partic- we did with children, that in a sample of adults with initially ipants with higher pre-transition scores having greater room to elevated anxiety levels, participants with the largest change in shift downwards on each measure over time. However, while attentional bias toward threat showed the greatest reductions in children with higher anxiety at pre-transition appeared to have anxiety across treatment. They argue that this is because pre- a steeper slope of change in anxiety over the time period, they paredness to acquire a threat bias reflects a general attentional still retained higher levels of anxiety symptoms at post- plasticity effect whereby individuals who most readily acquire a transition. bias favouring threat in response to a contingency making a For most children heightened symptoms of anxiety and threat bias adaptive will also be most likely to adopt the reverse school concerns were relatively short-lived, with mean scores processing bias when exposed to environmental conditions significantly reduced by post-transition relative to pre-transi- such as therapy, which make this adaptive and so reduce anx- tion. One possible explanation is that by the end of their first iety. However, our findings that readiness to acquire a bias term of secondary school, most children had adjusted to toward and away from threat were not correlated and change changes in the physical and personal school environment, in in anxiety not predicted by changeinbiasawayfromthreat turn leading to reductions in anxiety symptoms and school argues against such a general plasticity account. Further re- concerns. However, changes in these measures may also be search is needed to better understand the differences in patterns accounted for by other processes that occur in the time interval of effects seen, and to fully explain the mechanisms underpin- between the end of primary school, and the start of secondary ning why preparedness to acquire a threat attentional bias was school that are independent of the transition to a new school adaptive in this instance (and in Clarke et al. 2012) but predict- (Lohaus et al. 2004). Lohaus and colleagues suggest that the ed maladaptive increases in anxiety in Clarke et al. (2008). CBT for social anxiety and school transition have in com- reduction in stress and symptoms observed may reflect a re- covery effect over the summer break from school. This recov- mon the fact that on average participants begin with elevated ery effect may outweigh any possible stress-inducing effects anxiety levels which decrease over the course of treatment/ of the transition for the majority of children resulting in the time. Furthermore, the process of CBT for social anxiety and observed reductions in anxiety symptoms and school concerns arguably the school transition experience typically involve by post-transition. However, there was a subset of children some form of repeated exposure involving confrontation with with elevated anxiety symptoms at pre-transition who contin- a feared object, situation, or anxiety-provoking thought. The ued to experience heightened anxiety post-transition even af- content of fears around changes in social environment and ter any recovery effect of the summer holidays. Around half of relationships/experiences with others may also be similar for children who exceeded the clinical cut-off for anxiety at pre- individuals receiving CBT for social anxiety, and children transition retained clinically severe anxiety levels at post-tran- transitioning to secondary school. Within CBT, and perhaps sition. Furthermore, 22.1% and 29.9% of the current sample during school transition, successive exposures to a feared reported either no change or an increase in anxiety symptoms, stimuli or situation in the absence of any aversive conse- or school concerns by post-transition. quences, should result in the individual learning that their A larger change in attentional bias toward threat significantly feared object is not predictive of an aversive outcome and predicted a greater reduction in anxiety symptoms and school ultimately anxiety is reduced (Craske et al. 2014). An en- concerns across time. This was unexpected and at odds with hanced capacity to attend to, identify and engage with threat prior research that observed a larger change in attentional bias as a consequence of more readily acquiring an attentional bias toward threat at the start of the first semester of university pre- for threat might increase the opportunity for an individual to dicted a greater increase in anxiety by the end of the semester learn that their feared object is not necessarily predictive of an (Clarke et al. 2008). It is argued that this is because a heightened aversive outcome, thus facilitating extinction of fears and re- preparedness to acquire a threat bias in response to an experi- ducing anxiety (Barry et al. 2015). A small number of studies mental contingency that favours selective threat processing have indeed shown that a stronger tendency to attend to threat 1530 J Abnorm Child Psychol (2019) 47:1521–1532 relative to attending away from threat or no bias is predictive With three or more measurements we could have tracked of improved response to exposure treatment (Legerstee et al. changes in symptoms more precisely, and could also have ex- 2010; Price et al. 2011; Waters et al. 2012). Similar processes plicitly modeled the correlation between intercept and slope of may account for the present pattern of findings: A greater change rather than simply controlling for pre-transition out- readiness to acquire an attentional bias toward threat may have come scores. With a relatively modest N of 79, we were 80% facilitated children learning that a potentially feared and powered to detect a correlation of 0.31. While most of the key threatening school situation was not necessarily predictive of findings in our study exceeded this effect size, we were under- something aversive occurring leading to a reduction in anxiety poweredtodetectsmallereffects, and Type II errors may be symptoms and school concerns. present. We did not have diagnostic measures or data on treat- Change in attentional bias away from threat was not signif- ment use, and cannot discount the possibility that diagnostic icantly associated with change in anxiety or school concerns. and treatment status may have moderated the association be- The correlation between change in anxiety symptoms and tween attentional and interpretation bias and symptom change change in attentional bias away from threat was in the opposite across time. We used a measure of anxiety symptoms assessed direction and was significantly different from the correlation over a 3-month period rather than separate measures of state coefficient between change in anxiety and attentional bias and trait anxiety. The SCARED likely taps elements of both change toward threat (Z (57) = −2.18, p = 0.032). This suggests trait and state anxiety but appears to more highly correlated that reductions in anxiety across transition were specifically with trait anxiety (Monga et al. 2000). However, we were un- predicted by the degree to which individuals acquired an atten- able to unpick the impact of school transition on state and trait tional bias toward threat and not by a high level of general anxiety independently or to fully investigate whether partici- attentional plasticity. Measures of change in attentional bias pants with high trait anxiety differ in their trajectory and corre- toward and away from threat were not significantly correlated. lates of change from those participants who reported only being This is inconsistent with a small number of studies which have high state but not high trait anxious. Notwithstanding this, the observed that malleability in attentional bias toward and away findings in the subset of the sample exceeding clinical cut-off from threat are equally predicted by other factors, including scores for anxiety were very similar in direction and magnitude variation in the 5HTTLPR gene (Fox et al. 2011) and a mea- of effects to those observed in the entire sample (see sure of attentional control (Basanovic et al. 2017). Supplementary Materials). There are many other factors that When faced with ambiguous school situations, children with we could have measured using not only child but also parent heightened anxiety and school concerns at pre-transition were or teacher-report measures, (e.g. stressful life events, peer rela- more likely to resolve that ambiguity in a threatening way. tionships, bullying and victimisation, parenting styles and psy- Interpretation bias was not significantly associated with anxiety chopathology) and which may predict emotional adjustment or school concerns at post-transition or the change in these across school transition directly, or indirectly by influencing measures. Our findings are more consistent with interpretation change in children’s attentional responses to emotional stimuli. bias being a consequence or epiphenomenon of anxiety (Dodd For most children there were no persistent negative effects et al. 2012), and with prospective studies, which have found no of school transition. However, a concerning proportion of or minimal significant evidence for a longitudinal relationship children reported clinical levels of anxiety, and importantly, between interpretation bias and anxiety symptoms (Creswell higher anxiety symptoms at pre-transition were associated et al. 2011;Doddet al. 2012; Muris et al. 2004). with the retention of higher anxiety following transition. Pre-transition attentional bias for threat was not significantly These findings reiterate the importance of monitoring chil- associated with anxiety symptoms and school concerns, or the dren’s emotional wellbeing during a time of heightened stress, change in these measures. There is very limited research investi- which corresponds with a sensitive period for the develop- gating prospective associations between attentional biases and ment of anxiety disorders. This is the first study to demon- anxiety in child samples (Morales et al. 2015), and there is evi- strate an association between magnitude of change in atten- dence to suggest that attentional biases in children may only be tional bias toward threat and change in anxiety symptoms and observed at clinical levels of anxiety (Bar-Haim et al. 2007; school concerns in response to school transition. More re- Dudeney et al. 2015). We found only very minimal evidence search is needed to unravel the mechanisms underpinning this for any association with anxiety symptoms even when we con- association, and to determine whether this relationship can be fined our analyses to those children exceeding the clinical cut-off. exploited in intervention approaches. A recent meta-analysis found the effect size for the association between child anxiety and attentional bias was smaller compared Compliance with Ethical Standards to adults, less robust, and sensitive to important methodological Funding This research was funded by a UK Medical Research Council details such as task type and format (Dudeney et al. 2015). Experimental Medicine Grant awarded to Dr. Kathryn Lester and This study has several limitations. We had no control group Professor Thalia Eley (MR/J011762/1). T. C Eley is part funded by a of non-transitioning children and only two assessment points. J Abnorm Child Psychol (2019) 47:1521–1532 1531 program grant from the UK Medical Research Council (MR/M021475/ Brown, H. M., Eley, T. C., Broeren, S., Macleod, C., Rinck, M., Hadwin, 1). E. Carr and T. C. Eley are funded by the by the National Institute for J. A., & Lester, K. J. (2014). Psychometric properties of reaction Health Research (NIHR) Biomedical Research Centre at South London time based experimental paradigms measuring anxietyrelated and Maudsley NHS Foundation Trust and King’s College London. The information-processing biases in children. Journal of Anxiety views expressed are those of the authors and not necessarily those of the Disorders, 28(1), 97–107. https://doi.org/10.1016/j.janxdis.2013. NHS, the NIHR or the Department of Health. 11.004. Clarke, P. J. F., MacLeod, C., & Shirazee, N. (2008). Prepared for the worst: readiness to acquire threat bias and susceptibility to elevate Conflict of Interest The authors declare that they have no conflict of trait anxiety. Emotion, 8(1), 47–57. https://doi.org/10.1037/1528- interest. 3542.8.1.47. Clarke, P. J. F., Chen, N. T. M., & Guastella, A. J. (2012). Prepared for the Ethical Approval The study was ethically approved by King’s College best: readiness to modify attentional processing and reduction in London Psychiatry, Nursing and Midwifery Research Ethics Committee anxiety vulnerability in response to therapy. Emotion, 12(3), 487– (PNM/14/15-66). 494. https://doi.org/10.1037/a0025592. Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. Informed Consent All participants provided informed consent (parent) (2014). Maximizing exposure therapy: an inhibitory learning ap- and verbal assent (child). proach. Behaviour Research and Therapy, 58(Supplement C, 10– 23. https://doi.org/10.1016/j.brat.2014.04.006. Creswell, C., Shildrick, S., & Field, A. P. (2011). Interpretation of ambi- guity in children: a prospective study of associations with anxiety Open Access This article is distributed under the terms of the Creative and parental interpretations. Journal of Child and Family Studies, Commons Attribution 4.0 International License (http:// 20(2), 240–250. https://doi.org/10.1007/s10826-010-9390-7. creativecommons.org/licenses/by/4.0/), which permits unrestricted use, Cristea, I. A., Mogoase, C., David, D., & Cuijpers, P. (2015). Practitioner distribution, and reproduction in any medium, provided you give appro- review: cognitive bias modification for mental health problems in priate credit to the original author(s) and the source, provide a link to the children and adolescents: a meta-analysis. JournalofChild Creative Commons license, and indicate if changes were made. Psychology and Psychiatry and Allied Disciplines, 56(7), 723– 734. https://doi.org/10.1111/jcpp.12383. Dodd,H.F.,Hudson,J.L.,Morris, T. M.,&Wise,C.K.(2012). Interpretation bias in preschool children at risk for anxiety: a pro- spective study. Journal of Abnormal Psychology, 121(1), 28–38. https://doi.org/10.1037/a0024589. Dudeney, J., Sharpe, L., & Hunt, C. (2015). Attentional bias towards References threatening stimuli in children with anxiety: a meta-analysis. Clinical Psychology Review, 40,66–75. https://doi.org/10.1016/j. Anderson, L. W., Jacobs, J., Schramm, S., & Splittgerber, F. (2000). cpr.2015.05.007. School transitions: Beginning of the end or a new beginning? Evans, D., Borriello, G. A., & Field, A. P. (2018). A review of the aca- International Journal of Educational Research, 33(4), 325–339. demic and psychological impact of the transition to secondary edu- https://doi.org/10.1016/S0883-0355(00)00020-3. cation. Frontiers in Psychology, 9, 1482. https://doi.org/10.3389/ Bar-Haim, Y., Lamy, D., Lee, P., Bakermans-Kranenburg, M. J., & van fpsyg.2018.01482. Ijzendoorn, M. H. (2007). Threat-related attentional bias in anxious Fox, E., Zougkou, K., Ridgewell, A., & Garner, K. (2011). The serotonin and nonanxious individuals: a meta-analytic study. Psychological transporter gene alters sensitivity to attention bias modification: ev- Bulletin, 133(1), 1–24. https://doi.org/10.1037/0033-2909.133.1.1. idence for a plasticity gene. Biological Psychiatry, 70(11), 1049– Barry, T. J., Vervliet, B., & Hermans, D. (2015). An integrative review of 1054. https://doi.org/10.1016/j.biopsych.2011.07.004. attention biases and their contribution to treatment for anxiety dis- Grills-Taquechel, A. E., Norton, P., & Ollendick, T. H. (2010). A longi- orders. Frontiers in Psychology, 6(968). https://doi.org/10.3389/ tudinal examination of factors predicting anxiety during the transi- fpsyg.2015.00968. tion to middle school. Anxiety, Stress, and Coping, 23(5), 493–513. Basanovic, J., Notebaert, L., Grafton, B., Hirsch, C. R., & Clarke, P. https://doi.org/10.1080/10615800903494127. (2017). Attentional control predicts change in bias in response to Hakamata, Y., Lissek, S., Bar-Haim, Y., Britton, J. C., Fox, N. A., attentional bias modification. Behavior Research and Therapy, 99, Leibenluft, E., ... Pine, D. S. (2010). Attention Bias modification 47–56. https://doi.org/10.1016/j.brat.2017.09.002. treatment: a meta-analysis toward the establishment of novel treat- Biehl, M., Matsumoto, D., Ekman, P., Hearn, V., Heider, K., Kudoh, T., & ment for anxiety. Biological Psychiatry, 68(11), 982–990. https:// Ton, V. (1997). Matsumoto and Ekman’s Japanese and Caucasian doi.org/10.1016/j.biopsych.2010.07.021. facial expressions of emotion (JACFEE): reliability data and cross- Hirsch, B. J., & DuBois, D. L. (1992). The relation of peer social support national differences. Journal of Nonverbal Behavior, 21,2–21. and psychological symptomatology during the transition to junior https://doi.org/10.1023/A:1024902500935. high school: a two-year longitudinal analysis. American Journal of Birmaher, B., Brent, D. A., Chiappetta, L., Bridge, J., Monga, S., & Community Psychology, 20(3), 333–347. https://doi.org/10.1007/ Baugher, M. (1999). Psychometric properties of the screen for child BF00937913. anxiety related emotional disorders (SCARED): a replication study. Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Journal of the American Academy of Child and Adolescent Walters, E. E. (2005). Lifetime prevalence and age-of-onset distri- Psychiatry, 38(10), 1230–1236. https://doi.org/10.1097/00004583- butions of DSM-IV disorders in the national comorbidity survey 199910000-00011. replication. Archives of General Psychiatry, 62(6), 593–602. https://doi.org/10.1001/archpsyc.62.6.593. Brown, H. M., McAdams, T. A., Lester, K. J., Goodman, R., Clark, D. M., & Eley, T. C. (2013). Attentional threat avoidance and familial Langner, O., Dotsch, R., Bijlstra, G., Wigboldus, D. H. J., Hawk, S. T., & van Knippenberg, A. (2010). Presentation and validation of the risk are independently associated with childhood anxiety disorders. Journal of Child Psychology and Psychiatry, 54(6), 678–685. Radboud faces database. Cognition and Emotion, 24(8), 1377– https://doi.org/10.1111/jcpp.12024. 13 88. https://doi.org/10.1080/02699930903485076. 1532 J Abnorm Child Psychol (2019) 47:1521–1532 Lau, J. Y. F., & Waters, A. M. (2017). Annual research review: an ex- therapy for social phobia. Depression and Anxiety, 28(4), 349– 353. https://doi.org/10.1002/da.20791. panded account of information-processing mechanisms in risk for child and adolescent anxiety and depression. Journal of Child Rice, F., Frederickson, N., & Seymour, J. (2011). Assessing pupil con- Psychology and Psychiatry, 58(4), 387–407. https://doi.org/10. cerns about transition to secondary school. British Journal of 1111/jcpp.12653. Educational Psychology, 81(Pt 2, 244–263. https://doi.org/10. Legerstee, J. S., Tulen, J. H. M., Kallen, V. L., Dieleman, G. C., Treffers, 1348/000709910X519333. P. D. A., Verhulst, F. C., & Utens, E. M. W. J. (2009). Threat-related Rice, F., Frederickson, N., Shelton, K., McManus, C., Riglin, L., & Ng- selective attention predicts treatment success in childhood anxiety Knight, T. (2015). Identifying factors that predict successful and disorders. Journal of the American Academy of Child and difficult transitions to secondary school. Nuffield Foundation. Adolescent Psychiatry, 48(2), 196–205. https://doi.org/10.1097/ Riglin, L., Frederickson, N., Shelton, K. H., & Rice, F. (2013). A longi- CHI.0b013e31819176e4. tudinal study of psychological functioning and academic attainment Legerstee, J. S., Tulen, J. H. M., Dierckx, B., Treffers, P. D. A., Verhulst, at the transition to secondary school. Journal of Adolescence, 36(3), F. C., & Utens, E. M. W. J. (2010). CBT for childhood anxiety 507–517. https://doi.org/10.1016/j.adolescence.2013.03.002. disorders: differential changes in selective attention between treat- Riglin, L., Petrides, K. V., Frederickson, N., & Rice, F. (2014). The ment responders and non-responders. Journal of Child Psychology relationship between emotional problems and subsequent school and Psychiatry, 51(2), 162–172. https://doi.org/10.1111/j.1469- attainment: a meta-analysis. Journal of Adolescence, 37(4), 335– 7610.2009.02143.x. 346. https://doi.org/10.1016/j.adolescence.2014.02.010. Lester, K. J., Seal, K., Nightingale, Z. C., & Field, A. P. (2010). Are Smith, J. S., Akos, P., Lim, S., & Wiley, S. (2008). Student and stake- children’s own interpretations of ambiguous situations based on holder perceptions of the transition to high school. The High School how they perceive their mothers have interpreted ambiguous situa- Journal, 91(3), 32–42. tions for them in the past? Journal of Anxiety Disorders, 24,102– Stirling, L. J., Eley, T. C., & Clark, D. M. (2006). Preliminary evidence 108. https://doi.org/10.1016/j.janxdis.2009.09.004. for an association between social anxiety symptoms and avoidance Lester, L., Waters, S., & Cross, D. (2013). The relationship between of negative faces in school-age children. Journal of Clinical Child & school connectedness and mental health during the transition to Adolescent Psychology, 35(3), 431–439. https://doi.org/10.1207/ secondary school: a path analysis. Australian Journal of Guidance s15374424jccp3503_9. and Counselling, 23(2), 157–171. https://doi.org/10.1017/jgc.2013. Stradling, R., & MacNeil, M. (2001). Moving on : The emotional well- being of young people in transition from primary to secondary Lohaus, A., Elben, C., Ball, J. E., & Klein-hessling, J. (2004). School school. Inverness: Highland Health Board. transition from elementary to secondary school: changes in psycho- Suarez, L., & Bell-Dolan, D. (2001). The relationship of child worry to logical adjustment. Educational Psychology, 24(2), 161–173. cognitive biases: threat interpretation and likelihood of event occur- https://doi.org/10.1080/0144341032000160128. rence. Behavior Therapy, 32(3), 425–442. https://doi.org/10.1016/ MacLeod, C. (1999). Anxiety and anxiety disorders. In T. Dalgleish & M. S0005-7894(01)80029-0. Power (Eds.), Handbook of cognition and emotion (pp. 447–478). Thomasson, R., Field, L., O'Donnell, C., & Woods, S. (2006). School Chichester: England: Wiley. concerns questionnaire. Buckinghamshire: Buckinghamshire MacLeod, C., & Clarke, P. J. F. (2015). The attentional bias modification County Council. approach to anxiety intervention. Clinical Psychological Science, 3(1), 58–78. Tottenham, N., Tanaka, J. W., Leon, A. C., McCarry, T., Nurse, M., Hare, Monga, S., Birmaher, B., Chiappetta, L., Brent, D., Kaufman, J., Bridge, T. A., ... Nelson, C. (2009). The NimStim set of facial expressions: J., & Cully, M. (2000). Screen for child anxiety-related emotional Judgments from untrained research participants. Psychiatry disorders (SCARED): Convergent and divergent validity. Research, 168(3), 242–249. https://doi.org/10.1016/j.psychres. Depression and Anxiety, 12(2), 85–91. https://doi.org/10.1002/ 2008.05.006. 1520-6394(2000)12:2<85::AID-DA4>3.0.CO;2-2. Van Bockstaele, B., Verschuere, B., Tibboel, H., De Houwer, J., Morales, S., Pérez-Edgar, K. E., & Buss, K. A. (2015). Attention biases Crombez, G., & Koster, E. H. W. (2014). A review of current evi- towards and away from threat mark the relation between early dys- dence for the causal impact of attentional bias on fear and anxiety. regulated fear and the later emergence of social withdrawal. Journal Psychological Bulletin, 140(3), 682–721. https://doi.org/10.1037/ of Abnormal Child Psychology, 43(6), 1067–1078. https://doi.org/ a0034834. 10.1007/s10802-014-9963-9. Waters, A. M., Mogg, K., & Bradley, B. P. (2012). Direction of threat Muris, P., Jacques, P., & Mayer, B. (2004). The stability of threat percep- attention bias predicts treatment outcome in anxious children receiv- tion abnormalities and anxiety disorder symptoms in non-clinical ing cognitive-behavioural therapy. Behaviour Research and children. Child Psychiatry and Human Development, 34(3), 251– Therapy, 50(6), 428–434. https://doi.org/10.1016/j.brat.2012.03. 265. https://doi.org/10.1023/B:CHUD.0000015000.13845.9d. Ng-Knight, T., Shelton, K. H., Riglin, L., McManus, I. C., Frederickson, West, P., Sweeting, H., & Young, R. (2010). Transition matters: pupils’ N., & Rice, F. (2016). A longitudinal study of self-control at the experiences of the primary-secondary school transition in the West transition to secondary school: considering the role of pubertal status of Scotland and consequences for well-being and attainment. and parenting. JournalofAdolescence, 50,44–55. https://doi.org/ Research Papers in Education, 25(1), 21–50. https://doi.org/10. 10.1016/j.adolescence.2016.04.006. 1080/02671520802308677. Osinsky, R., Lösch, A., Hennig, J., Alexander, N., & MacLeod, C. Zeedyk, M. S., Gallacher, J., Henderson, M., Hope, G., Husband, B., & (2012). Attentional bias to negative information and 5-HTTLPR Lindsay, K. (2003). Negotiating the transition from primary to sec- genotype interactively predict students’ emotional reactivity to first ondary school. School Psychology International, 24(1), 67–79. university semester. Emotion, 12(3), 460–469. https://doi.org/10. https://doi.org/10.1177/0143034303024001010. 1037/a0026674. Pearl, J. (2016). Lord’s paradox revisited – (oh Lord! Kumbaya!). Journal of Causal Inference, 4(2). https://doi.org/10.1515/jci-2016-0021. Publisher’sNote Springer Nature remains neutral with regard to Price, M., Tone, E. B., & Anderson, P. L. (2011). Vigilant and avoidant jurisdictional claims in published maps and institutional affiliations. attention biases as predictors of response to cognitive behavioral http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Abnormal Child Psychology Springer Journals

Associations Between Attentional Bias and Interpretation Bias and Change in School Concerns and Anxiety Symptoms During the Transition from Primary to Secondary School

Download PDF
12 pages

Loading next page...
 
/lp/springer-journals/associations-between-attentional-bias-and-interpretation-bias-and-SCRKXZ6xxb

References (57)

Publisher
Springer Journals
Copyright
Copyright © 2019 by The Author(s)
Subject
Psychology; Child and School Psychology; Neurosciences; Public Health
ISSN
0091-0627
eISSN
1573-2835
DOI
10.1007/s10802-019-00528-3
Publisher site
See Article on Publisher Site

Abstract

The transition from primary to secondary school is often associated with a period of heightened anxiety and worry. For most children, any feelings of anxiety subside relatively quickly but for a small minority, emotional difficulties can continue into the first year of secondary school and beyond. This study recruited 109 children and measured their anxiety symptoms and school concerns toward the end of primary school and again at the end of their first term of secondary school. We investigated for the first time whether pre-transition measures of attentional and interpretation bias, and the magnitude of change in attentional bias toward and away from threat stimuli were associated with pre- and post-transition measures of anxiety and school concerns, and the change in these measures over time. Over 50% of the current sample exceeded clinical levels of anxiety at pre-transition. However, anxiety symptoms and school concerns had significantly reduced by post-transition. Higher levels of pre-transition anxiety or school concerns, and a greater magnitude of change in attentional bias towards threat stimuli predicted a larger reduction in anxiety symptoms and school concerns across the transition period. A greater interpretation bias toward threat was associated with higher pre-transition anxiety symptoms and school concerns but not post-transition scores, or the change in these scores. While many children experience heightened anxiety prior to school transition, this appears to be largely temporary and self-resolves. Nonetheless, the current findings highlight the importance of monitoring children’s anxiety and concerns, and related cognitive processes during this important transition period. . . . Keywords School transition Anxiety Attentional Bias Interpretation Bias In the UK, most children move from primary to secondary School transition has been shown in some studies to school at 11 years of age. At the time of transition, many negatively impact children’s emotional wellbeing children experience worry and anxiety, which typically sub- (Anderson et al. 2000; Smith et al. 2008) although such sides relatively quickly over the first term of secondary school studies are relatively sparse and results somewhat incon- (Rice et al. 2011; Stradling and MacNeil 2001; Zeedyk et al. sistent (Evans et al. 2018; Grills-Taquechel et al. 2010). 2003). For a minority of children, emotional difficulties en- Nonetheless, school concerns and anxiety symptoms ap- dure, continuing into and beyond the first year of secondary pear strongly correlated at primary and secondary school education (Zeedyk et al. 2003). (Rice et al. 2011) and emotional difficulties around the transition period are associated with concurrent and pro- Electronic supplementary material The online version of this article spective school attainment (Riglin et al. 2014), suggesting (https://doi.org/10.1007/s10802-019-00528-3) contains supplementary school transition may have a longer term negative impact material, which is available to authorized users. on child outcomes for some. Children with enduring emo- tional symptoms following the school transition may rep- * Kathryn J. Lester resent a particularly vulnerable group (Riglin et al. 2013). k.lester@sussex.ac.uk The school transition period thus represents a critical time for identifying children at increased risk for poor adjust- School of Psychology, University of Sussex, Pevensey Building, ment, and for nurturing children’s mental health, especial- Brighton BN1 9QH, UK ly since this developmental stage is associated with the Institute of Psychiatry, Psychology and Neuroscience, King’s onset of many anxiety disorders (Kessler et al. 2005). College London, De Crespigny Park, London SE5 8AF, UK 1522 J Abnorm Child Psychol (2019) 47:1521–1532 Numerous risk and protective factors for academic, behav- modification task) early in the semester predicted a larger in- ioural and emotional adjustment to school transition have crease in anxiety in students by the end of their first semester at been identified (Evans et al. 2018). Being younger, female, university. A second study revealed that change in attentional having lower socioeconomic status, being less academically bias away from threat was not predictive of subsequent chang- able, and experiencing stressful life events predict greater risk es in anxiety by the end of the semester. This suggests that for poor adjustment (Anderson et al. 2000;Riceet al. 2011; increases in anxiety in response to mild extended stress were West et al. 2010). Likewise peer factors (e.g. lack of peer not explained by a high level of general attentional plasticity or support, experiences of victimisation, stability in friendships) a tendency to develop attentional avoidance of threat but were and family relationships (maternal depressive symptoms, par- specifically determined by the degree to which individuals enting styles characterised by warmth) are also significant were prepared to acquire an attentional bias toward threat. predictors of children’s adjustment and academic attainment While Clarke and colleagues’ conclusions are compelling, (Hirsch and DuBois 1992; Ng-Knight et al. 2016;Rice etal. a stronger test of their competing hypotheses requires a 2015;Westet al. 2010). With regard to emotional predictors, within-participant design where all participants are exposed one study showed that heightened symptoms of generalised to both bias contingencies (toward threat, away from threat). anxiety prior to transition was associated with greater con- If individuals who show the largest change in attentional bias cerns about school transition both before and after transition toward threat also show the largest change in attentional bias (Rice et al. 2011). A meta-analysis also found that emotional away from threat then this pattern might indicate that some difficulties significantly predicted poorer school attainment individuals’ attentional processing is generally more mallea- (Riglin et al. 2014). ble and responsive to environmental contingencies favouring There is reasonable evidence that heightened symptoms of selective attention both toward and away from threat. The anxiety are associated with cognitive biases in attention and inclusion of both contingency conditions also allows us to interpretation that favour the selective processing of threat unpick whether change in anxiety symptoms/school concerns (Bar-Haim et al. 2007; Dudeney et al. 2015; Lau and Waters is best explained by a.) the degree to which individuals are 2017; Suarez and Bell-Dolan 2001). These biases are thought prepared to acquire an attentional bias bias specifically to- to contribute causally to the onset and/or maintenance of wards threat, b.) the degree to which individuals acquire an symptoms (Van Bockstaele et al. 2014), and to impact emo- attentional bias characterised by avoidance of threat, or c.) is tional response to stressors (Hakamata et al. 2010;Osinsky indexed equally well by both, indicating a more general atten- et al. 2012), although this evidence is somewhat more mixed tional malleability effect. A further rationale for including the in child samples (Dodd et al. 2012; Dudeney et al. 2015). This avoid threat condition was to explore direction of effects. is the first study to investigate whether attentional and inter- Some studies have shown that attentional avoidance of threat pretation biases assessed toward the end of primary school are is associated with heightened anxiety (Brown et al. 2013; associated with anxiety symptoms and school concerns over Stirling et al. 2006); whereas, others have shown that atten- the school transition period. tional avoidance of threat is associated with a reduction in The degree to which people respond to stressors with in- symptoms of anxiety (e.g. Legerstee et al. 2009). creases in anxiety differs substantially (Clarke et al. 2008). This study investigated the impact of the school transition The basis for this variability is poorly understood but it has period on children’s concerns about school and their anxiety been argued that exposure to stressful life events may make it symptoms. To do so, we recruited children during the final year temporarily adaptive to develop an attentional bias for threat of primary school and assessed the degree to which school con- cues that may signal genuine danger (MacLeod 1999). As a cerns and anxiety symptoms changed between the first measure- consequence, individuals who demonstrate the greatest ment in primary school and the second measurement at the end change in this capacity to selectively direct attention toward of the first term at secondary school. As noted previously, a small threat might be expected to display the most intense and number of studies have shown that school transition negatively sustained anxiety reactions to a subsequent stressor. impacts on emotional wellbeing, although any emotional diffi- Attentional bias modification studies which use experimental culties typically subside relatively quickly. In light of this, and in contingencies to encourage selective attention to threat stimuli line with prior research (Rice et al. 2011) we began by testing the have shown that those individuals who acquire a larger atten- hypothesis that school concerns, and anxiety symptoms would tional bias for threat also report greater anxiety in response to a decrease significantly over the transition period and that anxiety subsequent experimental stressor (MacLeod and Clarke and school concern scores at pre-transition would correlate sig- 2015). However, findings have been more equivocal in child nificantly with scores at post-transition (Hypothesis 1). Given the samples (Cristea et al. 2015). relatively extended nature of the school transition period and the Consistent with research using lab-based stressors, Clarke time period between assessments in the present study, we focused et al. (2008) found that a greater change in attentional bias on anxiety symptoms using the Screen for Child Anxiety Related toward threat stimuli (evoked using an attentional bias Disorders (SCARED), rather than on trait or state anxiety per se. J Abnorm Child Psychol (2019) 47:1521–1532 1523 Prior research indicates that the SCARED taps into elements of Children attended two 45-min sessions, 1 week apart dur- both trait and state anxiety, but may be more strongly correlated ing the last term of primary school. One hundred and six with measures of trait anxiety (Monga et al. 2000). participants completed both sessions, and 79 participants Given the reasonable evidence that cognitive biases (75% response rate) returned a follow-up questionnaire to- favouring the selective processing of threat are associated with wards the end of their first term of secondary school. heightened levels of anxiety (e.g. Lau and Waters 2017), we Compared to responders, non-responders to the follow-up hypothesised that a stronger tendency to interpret ambiguous questionnaire did not differ significantly on age, sex, ethnicity, situations as threatening (Hypothesis 2) and a greater atten- anxiety symptoms, school concerns, interpretation bias or at- tional bias toward threat (Hypothesis 3) prior to transition tentional bias at pre-transition (all p values > 0.05). would be associated with higher school concerns, and anxiety A priori power calculations were performed for the associ- symptoms before and after school transition. Past research has ation between change in attentional bias toward threat and also suggested that individuals who show a greater change in change in anxiety scores across the school transition. Effect their capacity to selectively direct attention toward threat (i.e. size estimates were taken from Clarke et al. (2008) who ob- to acquire an attentional bias toward threat) but not away from served a correlation of r= 0.47 between attentional bias threat (i.e. to acquire an attentional bias away from threat) may change and anxiety change in their sample of students attend- display more intense anxiety reactions to stressors like school ing the first semester at University. With an effect size of r = transition (Clarke et al. 2008). Uniquely, we explored in the 0.47, and to achieve 80% power with α = 0.008 (to allow for same sample whether the magnitude of change in a.) atten- multiple testing corrections, see BStatistical Analysis^ section tional bias toward threat and b.) attentional bias away from below) required a minimum N =50. threat was associated with school concerns, and anxiety symp- toms before and after school transition (Hypothesis 4). Finally, we explored whether our pre-transition measures of anxiety Measures symptoms, school concerns, interpretation bias, attentional bi- as and change in attentional bias toward and away from threat Anxiety Symptoms Anxiety symptoms (e.g. BI worry about predicted change in a.) school concerns and b.) anxiety symp- other people liking me, When I get frightened I feel dizzy^) toms across the transition period (Hypothesis 5). in the preceding 3 months were measured using the 41-item Screen for Child Anxiety Related Emotional Disorders (SCARED, (Birmaher et al. 1999). Responses were made Method using a 3-point scale (‘not true or hardly ever true’–‘very true or often true’). Internal consistency was excellent (α = Participants 0.85 and α = 0.90 at pre- and post-transition, respectively). One hundred and nine children (mean age = 10.7 years, SD = School Concerns Concerns about secondary school were mea- 0.5, 53% female) were recruited from seven mainstream sured using the 17-item School Concerns Questionnaire schools. Schools were invited pragmatically on the basis of (Thomasson et al. 2006). Children rated their degree of con- being mainstream primary schools situated within a relatively cern for each item (e.g. ‘being bullied’, ‘following a timeta- broad area of Greater London but within a reasonable travel- ble’) using a 10-point scale (1 = ‘not worried’ to 10 = ‘ex- ling distance using public transport from King’s College tremely worried’). Internal consistency was good (α = 0.92 London. Ethical approval was granted by King’s College and 0.79 at pre- and post-transition). London Psychiatry, Nursing and Midwifery Research Ethics Committee (PNM/14/15–66). Parents gave written informed Ambiguous Situations Questionnaire – School Transition Ten consent and children verbal assent. Inclusion criteria included ambiguous scenarios were presented describing typical sec- being in the final year of primary school. Parents were advised ondary school situations (see Supplementary Materials for that the study was not suitable for children with significant measure development). For example, BYou are in a PE lesson learning disabilities or insufficient understanding of the at your new school. Your teacher chooses two team captains English language to comprehend the study materials. Forty- and asks them to pick teams for a basketball game. You wait six percent of children identified as White British/other White for yournametobecalledout^. Participants responded to an background, 14% as African, and 7% as other mixed/multiple open-ended question designed to elicit their interpretation of ethnic background. Fifty-five percent of parents were educat- the scenario (BWhen do you think your name will be ed to university degree level or above. Overall the sample was called?^). Scenarios were presented randomly on a laptop more highly educated compared to the UK national average screen using EPrime 2.0. Free choice responses were record- but was ethnically diverse and broadly representative with ed, transcribed and coded using an established coding ap- regard to ethnicity for London. proach (K. J. Lester et al. 2010). 1524 J Abnorm Child Psychol (2019) 47:1521–1532 Participants were also presented simultaneously with a questionnaire administration was randomised, followed threat (BYou think your name will be called near the end as by completion of the ambiguous situations questionnaire, the team captains won’t want you on their team^) and non- then attentional bias change task. In session 2, participants threat (BYou think your name will be called very soon as the completed the opposite attentional contingency phase team captains will want you on their team^) forced-choice condition to that given in session 1. Participants received interpretation. Participants indicated which interpretation they a follow-up questionnaire toward the end of their first thought was most likely to be the outcome of each situation term of secondary school. Participants received a small using counterbalanced response keys. Free and forced-choice craft gift and a £10 gift card at the end of sessions 1 response formats were highly correlated, r (109) = 0.67, and 2, respectively, and a £10 gift card if they returned p < 0.001. Therefore, a combined threat interpretation score the follow-up questionnaire. (range 0–100%) was computed: (total free + forced-choice threat interpretations/total number of valid interpretations) × Data Processing Changes in school concerns (SCQ ) and anxiety CHANGE Attentional Bias Change Task This task comprised of three symptoms (SCARED ) were computed by CHANGE phases, each using a dot-probe or modified dot-probe task. subtracting pre-transition scores from post-transition The first and third phases (the pre- and post-contingency bias scores. We used an established approach to data cleaning assessment phases) assessed relative attentional allocation to of dot-probe reaction time data with children, and to deal- threat and neutral stimuli. The second phase (the attentional ing with participants with outlying responses compared to contingency block) exposed participants to a contingency de- the sample mean or high error rates (Brown et al. 2014). signed to elicit an attentional bias favouring threat stimuli This approach was decided upon prior to data collection (attentional contingency – toward threat condition) or and was used to remove trials with extreme (short or long) favouring neutral stimuli (attentional contingency – avoid response times compared to the individuals mean reaction threat condition). time, thus representing trials on which anticipatory re- Stimuli comprised of forty models (20 males and 20 fe- sponses were made or delayed responses, which may re- males) portraying angry and neutral facial expressions select- flect distraction/lack of attentiononthattrial. Meanreac- ed from established face sets (Biehl et al. 1997; Langner et al. tion times (RTs) were calculated for pre- and post- 2010; Tottenham et al. 2009). The stimuli were divided into contingency assessments of attentional bias, after remov- four subsets balanced for male and female faces. The alloca- ing errors (2.8% of trials), and data values >2.5 SDs from tion of face sets and order of contingency conditions was individual means, or < 100 ms (4.1% of trials). Bias scores counterbalanced using random block allocation, and blinded were calculated for pre- and post-contingency assessment to the experimenter. Participants received the same face set in phases by subtracting mean RTs for probes presented in the pre- and post-contingency bias assessment phases (within the angry face locus from mean RTs for probes presented session) and a different subset for each attentional contingen- in the neutral face locus, resulting in a measure of partic- cy block. ipant’s attentional bias prior to contingency exposure During the attentional contingency phase, participants (AB ), and after contingency exposure PRE-CONTINGENCY viewed 280 trials (240 angry-neutral, 40 neutral-neutral trials). (AB ). A positive score indicates a bias POST-CONTINGENCY The task procedure is depicted in Fig. 1 (full details in the toward threat while a negative score indicates a bias away Supplementary Information). In the toward threat condition, from threat. Participants who made incorrect or outlying the probe (< or >) consistently appeared in the location previ- responses on more than 25% of trials were excluded from ously occupied by the angry face of angry-neutral face pairs. analyses using bias score data. Three participants were In the away from threat condition, the probe consistently ap- excluded from analyses in which pre-attentional bias peared in the location previously occupied by the neutral face. scores were correlated with pre- or post-transition anxiety The pre- and post-contingency assessment phases comprised or school concern scores or the change in those scores. of 120 trials (40 neutral-neutral and 80 angry-neutral trials) Outliers comprised of participants who had mean bias where the probe appeared with equal probability behind the scores exceeding 2.5 SD above or below the sample mean angry and neutral stimulus. Reaction times (RTs) and accuracy or participants who made incorrect or outlying responses of responses was recorded. on more than 25% of trials used to calculate bias scores. Change in attentional bias (AB ) was computed for CHANGE Procedure the Battend toward threat^ condition (AB CHANGE–TOWARD )bysubtractingAB scores from THREAT PRE-CONTINGENCY The procedure is outlined in Fig. 2. Sessions 1 and 2 were AB scores. For the Battend away from POST-CONTINGENCY conducted in a quiet study space. In session 1, threat^ condition (AB )AB CHANGE–AVOID THREAT POST- J Abnorm Child Psychol (2019) 47:1521–1532 1525 Fig. 1 Attentional bias change task procedure Fixation cross: displayed for 500ms Face pair: angry-neutral Inter-trial interval: or neutral-neutral 500ms displayed for 500ms Probe (< or >): remained on screen until mouse button response detected scores were subtracted from AB approximated a normal distribution. Sensitivity analyses were CONTINGENCY PRE- scores. In both conditions larger positive scores conducted in which participants who were outliers for atten- CONTINGENCY on the AB index reflected a greater degree of change tional bias scores or for AB or CHANGE CHANGE-TOWARD THREAT in attentional bias in the direction encouraged during the con- AB scores were not removed from anal- CHANGE-AVOID THREAT tingency phase. For analyses correlating AB scores yses. Overall, the substantive conclusions remained un- CHANGE with pre-transition measures of anxiety symptoms or school changed despite minor fluctuations in effect size and p value concerns, eight participants were excluded due to having (see Supplementary Materials). mean change scores exceeding 2.5 SD above or below the sample mean or because they made incorrect or outlying re- Statistical Analysis sponses on more than 25% of trials used to calculate bias scores. For analyses correlating AB scores with post- We examined bivariate correlations between all variables CHANGE transition measures of anxiety symptoms or school concerns both within and across time points and between anxiety or change in these outcomes, six participants were excluded symptom and school concern change scores and all pre- for the same reasons. After data processing all variables transition variables. Bonferroni corrections were applied Fig. 2 Experimental procedure First term of secondary Last term of primary school school Session 1 (N = 109) Session 2 (N = 106) Follow-Up Questionnaire (N = 79) Anxiety (SCARED) AB Change Task School summer holidays School Concerns (SCQ) Anxiety (SCARED) Interpretation Bias School Concerns (SCQ) (ASQ) Attentional Bias AB Change Task Mean interval between session 1 and session 2 = 7.5 days, SD = 1.7. Mean interval between school sessions and follow-up questionnaire =187 days, SD = 50.6. 1526 J Abnorm Child Psychol (2019) 47:1521–1532 to control for the number of tests performed. Changes in Results anxiety symptoms and school concerns were analysed using ordinary least squares (OLS) linear regression. Hypothesis 1: Correlations Between Anxiety Assumptions for OLS regression were carefully checked. Symptoms and School Concerns at Pre- First, scatterplots were checked and showed that relation- and Post-Transition ships between the independent and dependent variables were linear. Second, there was no evidence of We hypothesised that school concerns and anxiety symptoms multicollinearity (correlations between independent vari- would decrease significantly over the transition period and ables were all lower than 0.8 and variance inflation fac- that anxiety and school concern scores at pre-transition would tors ranged 1.0–1.9). Third, plots of standardised resid- correlate significantly with scores at post-transition. Pre- uals vs. standardised predicted values showed no obvious transition mean scores for anxiety and school concerns (see signs of funnelling suggesting the assumption of homo- Table 1) were approximately 0.5 SD above reported norms scedasticity was met. Fourth, all Cook’s distance values while scores at post-transition were comparable to previous were under 1 (range 0.0–0.5) suggesting that no individ- reports for school concerns (Rice et al. 2011) and were 0.25 ual cases were unduly influencing the model. Finally, SD below the norms for non-anxiety cases (Birmaher et al. normal probability plots of the standardised residuals 1999). Consistent with this hypothesis, we observed a signif- for each model indicated some mild deviations from nor- icant reduction in school concerns and anxiety symptoms mality. While mild deviations from normality are unlike- from pre- to post-transition (Table 1). Nonetheless, there was ly to impact on the validity of our findings, as a sensi- substantial inter-individual variability. Anxiety symptoms and tivity analysis we re-ran the regression models using school concerns were moderately correlated at pre- and post- bootstrapped 95% bias corrected and accelerated confi- transition (see Table 2) and, consistent with our hypothesis, dence intervals (with 1000 bootstrap iterations). These there was continuity such that individuals with higher scores analyses gave results that were consistent to the models at pre-transition also reported higher scores at post-transition. without bootstrapping and did not change our substantive conclusions (full results available on request). Hypothesis 2: Associations Between Interpretation We included as covariates pre-transition variables that Bias and Anxiety Symptoms and School Concerns were significantly associated with the outcome at Pre- and Post-Transition (SCARED and SCQ ) in the bivariate cor- CHANGE CHANGE relations (p < 0.05). When a significant covariate is iden- We hypothesised that a greater interpretation bias favouring tified, it is significant over and above the other covariates threat would predict higher levels of anxiety symptoms and in the model. In all analyses involving SCARED CHANGE school concerns at pre- and post-transition. In support of this and SCQ scores we controlled for the pre- CHANGE hypothesis we found that a stronger tendency toward transition level of the outcome. For the change score interpreting ambiguous situations as threatening was signifi- ΔY = Y − Y ,weincludeY as a covariate in the model; 2 1 1 cantly associated with greater school concerns and anxiety for bivariate correlations, we estimate partial correlations, symptoms prior to school transition (see Table 2). However controlling for Y Controlling for pre-transition scores is threat interpretation bias did not significantly correlate with important because of correlations between pre-transition anxiety symptoms or school concern scores at post-transition. scores on the outcome and exposure (i.e. Y and X ). 1 1 Pre-transition outcome scores (Y e.g. pre-transition SCARED scores) are therefore correlated with both pre- Hypothesis 3: Associations Between Attentional Bias transition exposures (X e.g. interpretation bias scores) and Anxiety Symptoms and School Concerns at Pre- and the post-transition outcome (Y e.g. post-transition and Post-Transition SCARED scores). This represents a form of confounding that must be controlled for (Pearl 2016). Failure to ac- We hypothesised that a greater attentional bias for threat count for correlations between Y and X will lead to would predict higher levels of anxiety symptoms and school 1 1 biased results. This issue only arises when Y and X concerns at pre- and post-transition. The mean bias score was 1 1 are correlated (i.e. when Y predicts both X and ΔY −3.6 (SD = 36.2) indicating no significant attentional bias in 1 1 (e.g. SCARED score), and thus, confounds the either direction (difference from 0, t (109) = −1.03, p =0.304). CHANGE association between X and ΔY). Where X and Y are There were substantial individual differences in bias scores 1 1 1 uncorrelated (or are weakly correlated), there is no con- (range: −127.33 to 120.18). Contrary to our hypothesis, how- founding due to Y , and the results are not affected by ever, pre-transition attentional bias was not significantly asso- controlling for Y . However, for consistency, all estimates ciated with pre-or post-transition anxiety symptom or school have been adjusted for baseline outcome scores (Y ). concern scores (see Table 2). 1 J Abnorm Child Psychol (2019) 47:1521–1532 1527 Table 1 Descriptive and test statistics for school concerns, and anxiety symptoms at pre- and post-school transition Measure Pre-transition Post-transition Change score t (df) p Mean (SD) Mean (SD) Mean (SD) [95% CI] School concerns (SCQ) 78.21 (32.36) 46.30 (24.58) −29.32 (27.97) −9.32 (78) < 0.001 (n =106) (n =78) [−35.58 – −23.05] Anxiety symptoms (SCARED) 26.70 (14.72) 16.10 (13.79) −9.17 (12.51) −6.43 (76) < 0.001 (n =106) (n =76) [−12.01 – −6.33] Interpretation Bias (ASQ ) 32.93 (19.13) –– – – PRE (n =109) Attentional Bias (S1 AB ) −3.63 (36.19) –– – – PRE-CONTINGENCY (n =106) AB −0.05 (51.45) –– – – CHANGE – TOWARD THREAT (n =101) AB −6.04 (56.43) –– – – CHANGE – AVOID THREAT (n =98) SCQ School Concerns Questionnaire, SCARED Screen for Child Anxiety Related Emotional Disorders, ASQ Ambiguous Situations Questionnaire, AB Attentional Bias Hypothesis 4: Associations Between Attentional Bias (see Table 2). Larger changes in attentional bias toward threat Change and Anxiety Symptoms and School Concerns were significantly associated with lower anxiety symptoms at at Pre- and Post-Transition post-transition (but not school concerns), although this effect did not survive multiple testing corrections. AB CHANGE–AVOID We tested the hypothesis that the magnitude of change in scores were unrelated to pre-transition school con- THREAT attentional bias toward threat (AB ) cerns. However, higher anxiety symptoms at pre-transition CHANGE–TOWARD THREAT and magnitude of change in attentional bias away from threat was significantly associated with a smaller magnitude of at- (AB ) would be associated with school tentional bias change away from threat, although this too was CHANGE–AVOID THREAT concerns and anxiety symptoms at pre- and post-transition. no longer statistically significant after corrections for multiple AB scores were not significantly re- testing. No significant associations were observed with post- CHANGE– TOWARD THREAT lated to pre-transition school concerns or anxiety symptoms transition school concerns, or anxiety symptoms. No Table 2 Bivariate correlations between school concerns, anxiety symptoms and cognitive measures at pre- and post-school transition SCARED SCQ SCARED SCQ SCARED ASQ S1 AB AB AB PRE POST POST CHANGE CHANGE PRE PRE- CHANGE- CHANGE- CONTINGENCY TOWARD AVOID THREAT THREAT a ab a ab SCQ 0.687 0.509 0.024 -0.609 −0.022 0.542 0.013 −0.053 −0.053 PRE <0.001 <0.001 <0.001 <0.001 0.851 <0.001 0.895 0.600 0.606 ab a ab a SCARED – 0.250 0.633 0.229 -0.447 0.533 −0.140 −0.012 −0.219 PRE 0.030 <0.001 0.047 <0.001 0.000 0.153 0.908 0.030 ab ab ab ab SCQ – 0.672 –– -0.187 0.119 -0.159 -0.233 POST <0.001 0.106 0.309 0.189 0.054 ab ab ab ab SCARED –– – -0.170 0.123 -0.314 -0.076 POST 0.146 0.300 0.009 0.540 ab ab ab ab SCQ –– -0.153 0.065 -0.277 0.009 CHANGE 0.186 0.579 0.020 0.944 ab ab ab ab SCARED – -0.079 0.147 -0.392 0.078 CHANGE 0.502 0.215 0.001 0.531 SCQ School Concerns Questionnaire, SCARED Screen for Child Anxiety Related Emotional Disorders, ASQ Ambiguous Situations Questionnaire, AB a b Attentional Bias. Partial correlation coefficients controlling for interval in days between in-school sessions and return of follow-up questionnaires. Partial correlation coefficients controlling for interval in days between in-school sessions and return of follow-up questionnaires, and pre-transition level of the outcome. Correlation coefficients in bold survive multiple testing corrections: Bonferroni corrected p value for correlations with SCARED and PRE SCQ = 0.01; for correlations with SCARED and SCQ = 0.008; for correlations with SCARED and SCQ =0.008 PRE POST POST CHANGE CHANGE 1528 J Abnorm Child Psychol (2019) 47:1521–1532 significant correlation was observed between AB predicted a reduction of 7 points on the SCQ between pre- and CHANGE– and AB scores (r post-transition (β = −0.24; p =0.009). TOWARD THREAT CHANGE–AVOID THREAT (92) = −0.17, p = 0.110). Change in anxiety symptoms was significantly predicted by pre-transition anxiety symptoms and AB CHANGE-TOWARD scores. Higher pre-transition anxiety predicted a great- THREAT Hypothesis 5: Correlates and Prediction of Change er reduction in anxiety symptoms across the transition period. in Anxiety Symptoms and School Concerns A one SD increase in pre-transition anxiety score (14.23) pre- from Pre-Transition Variables dicted a reduction of 6 points on the SCARED between pre- and post-transition (β = −0.45; p < 0.001). Likewise, a one SD Our final hypothesis investigated whether pre-transition as- increase in AB scores (51.2ms) pre- CHANGE-TOWARD THREAT sessments (of anxiety symptoms, school concerns, interpreta- dicted a reduction of 4 points on the SCARED between pre- tion bias, attentional bias and change in attentional bias toward and post-transition (β = −0.35; p =0.001). and away from threat) predicted change in school concerns and anxiety symptoms over the transition period. Overall, we found good support for this hypothesis. Greater school con- cerns at pre-transition was significantly associated with a larg- Discussion er reduction in school concerns over the transition period (see Table 2) with the same pattern observed between anxiety This study explored whether attentional biases, the malleability scores. No significant correlations were observed between of attentional biases, and interpretation bias measured before pre-transition measures of interpretation bias, attentional bias, school transition explained individual differences in changes in or AB scores and SCQ and anxiety symptoms and school concerns over the transition peri- CHANGE-AVOID THREAT CHANGE SCARED scores. However, a greater change in atten- od. In support of hypothesis 1, we observed a significant decrease CHANGE tional bias toward threat stimuli (AB ) in anxiety symptoms and school concerns from pre- to post- CHANGE-TOWARD THREAT was significantly associated with a larger reduction in school transition. Anxiety and school concern scores at pre-transition concerns and anxiety symptoms over the transition period. also correlated significantly with post-transition scores. The correlation with SCQ did not survive multiple Hypothesis 2 was also partly supported with a greater threat CHANGE testing corrections. interpretation bias associated with higher pre-transition anxiety To test hypothesis 5, changes in school concerns and anx- symptoms and school concerns but not post-transition scores. iety symptoms were analysed using multiple regression. Only Hypothesis 3 was not supported as no significant associations pre-transition variables that were significantly associated with were observed between pre-transition attentional bias and pre- or the outcome (SCARED and SCQ ) in the bi- post- anxiety or school concern scores. We also found no con- CHANGE CHANGE variate correlations reported above (p < 0.05) were included in vincing support for hypothesis 4; attentional bias change toward each model. Change in school concerns was significantly as- andawayfromthreatwerenot significantly correlated with pre- sociated with pre-transition school concerns, pre-transition or post- transition anxiety symptom or school concern scores anxiety symptoms and AB scores after multiple testing corrections were applied. In support of hy- CHANGE-TOWARD THREAT (see Table 3). Higher pre-transition school concerns signifi- pothesis 5, we identified a small number of significant predictors cantly predicted a reduction in school concerns scores across of change in anxiety symptoms and school concerns across the the transition period (β = −0.75). Higher pre-transition anxiety transition period. Higher levels of pre-transition anxiety or school scores were positively associated with SCQ (β = concerns were associated with greater reductions in severity CHANGE 0.28). Increases in attentional bias toward threat predicted a across transition. A larger increase in attentional bias toward decrease in school concerns across the transition period. A one threat significantly predicted a larger reduction in anxiety symp- SD increase in AB scores (51.2ms) toms and school concerns across the school transition period. CHANGE–TOWARD THREAT Table 3 Regression analyses for change in school concerns and anxiety symptoms SCQ SCARED CHANGE CHANGE b [95% CI] β tp b [95% CI] β tp SCQ −0.75 [−0.97 – −0.52] −0.81 −6.69 <0.001 – ––– PRE SCARED 0.59 [0.08–1.10] 0.28 2.32 0.023 −0.39 [−0.57 – −0.22] −0.45 −4.47 <0.001 PRE AB −0.14 [−0.24 – −0.04] −0.24 −2.69 0.009 −0.09 [−0.13 – −0.04] −0.35 −3.49 0.001 CHANGE-TOWARD THREAT SCQ School Concerns Questionnaire, SCARED Screen for Child Anxiety Related Emotional Disorders, AB Attentional Bias J Abnorm Child Psychol (2019) 47:1521–1532 1529 The period towards the end of primary school appears to be predicts who will naturally develop an attentional bias for threat an especially stressful time with around half of our sample when exposed to a real-life extended stressor, in turn predicting exceeding the suggested clinical cut-off on the SCARED. who will experience increases in anxiety. However, unlike the We also observed continuity of symptoms, such that higher young adults who transitioned to university, children who anxiety symptoms at pre-transition were correlated with transitioned to secondary school reported a significant reduc- retaining heightened anxiety symptoms at post-transition. tion in anxiety over time. Notably, the current sample of chil- This is consistent with prior research showing that anxiety dren were considerably more anxious at the study outset com- symptoms in primary school predict later symptoms (Lester pared to the sample of young adults in Clarke et al. (2008). A et al. 2013). Higher anxiety scores pre-transition predicted a later study assessed change in attentional bias toward threat decrease in anxiety scores over time. The same effects were among individuals with social anxiety disorder prior to cogni- observed for school concerns. This pattern of results may re- tive behaviour therapy (Clarke et al. 2012). They also found, as flect a combination of regression toward the mean and partic- we did with children, that in a sample of adults with initially ipants with higher pre-transition scores having greater room to elevated anxiety levels, participants with the largest change in shift downwards on each measure over time. However, while attentional bias toward threat showed the greatest reductions in children with higher anxiety at pre-transition appeared to have anxiety across treatment. They argue that this is because pre- a steeper slope of change in anxiety over the time period, they paredness to acquire a threat bias reflects a general attentional still retained higher levels of anxiety symptoms at post- plasticity effect whereby individuals who most readily acquire a transition. bias favouring threat in response to a contingency making a For most children heightened symptoms of anxiety and threat bias adaptive will also be most likely to adopt the reverse school concerns were relatively short-lived, with mean scores processing bias when exposed to environmental conditions significantly reduced by post-transition relative to pre-transi- such as therapy, which make this adaptive and so reduce anx- tion. One possible explanation is that by the end of their first iety. However, our findings that readiness to acquire a bias term of secondary school, most children had adjusted to toward and away from threat were not correlated and change changes in the physical and personal school environment, in in anxiety not predicted by changeinbiasawayfromthreat turn leading to reductions in anxiety symptoms and school argues against such a general plasticity account. Further re- concerns. However, changes in these measures may also be search is needed to better understand the differences in patterns accounted for by other processes that occur in the time interval of effects seen, and to fully explain the mechanisms underpin- between the end of primary school, and the start of secondary ning why preparedness to acquire a threat attentional bias was school that are independent of the transition to a new school adaptive in this instance (and in Clarke et al. 2012) but predict- (Lohaus et al. 2004). Lohaus and colleagues suggest that the ed maladaptive increases in anxiety in Clarke et al. (2008). CBT for social anxiety and school transition have in com- reduction in stress and symptoms observed may reflect a re- covery effect over the summer break from school. This recov- mon the fact that on average participants begin with elevated ery effect may outweigh any possible stress-inducing effects anxiety levels which decrease over the course of treatment/ of the transition for the majority of children resulting in the time. Furthermore, the process of CBT for social anxiety and observed reductions in anxiety symptoms and school concerns arguably the school transition experience typically involve by post-transition. However, there was a subset of children some form of repeated exposure involving confrontation with with elevated anxiety symptoms at pre-transition who contin- a feared object, situation, or anxiety-provoking thought. The ued to experience heightened anxiety post-transition even af- content of fears around changes in social environment and ter any recovery effect of the summer holidays. Around half of relationships/experiences with others may also be similar for children who exceeded the clinical cut-off for anxiety at pre- individuals receiving CBT for social anxiety, and children transition retained clinically severe anxiety levels at post-tran- transitioning to secondary school. Within CBT, and perhaps sition. Furthermore, 22.1% and 29.9% of the current sample during school transition, successive exposures to a feared reported either no change or an increase in anxiety symptoms, stimuli or situation in the absence of any aversive conse- or school concerns by post-transition. quences, should result in the individual learning that their A larger change in attentional bias toward threat significantly feared object is not predictive of an aversive outcome and predicted a greater reduction in anxiety symptoms and school ultimately anxiety is reduced (Craske et al. 2014). An en- concerns across time. This was unexpected and at odds with hanced capacity to attend to, identify and engage with threat prior research that observed a larger change in attentional bias as a consequence of more readily acquiring an attentional bias toward threat at the start of the first semester of university pre- for threat might increase the opportunity for an individual to dicted a greater increase in anxiety by the end of the semester learn that their feared object is not necessarily predictive of an (Clarke et al. 2008). It is argued that this is because a heightened aversive outcome, thus facilitating extinction of fears and re- preparedness to acquire a threat bias in response to an experi- ducing anxiety (Barry et al. 2015). A small number of studies mental contingency that favours selective threat processing have indeed shown that a stronger tendency to attend to threat 1530 J Abnorm Child Psychol (2019) 47:1521–1532 relative to attending away from threat or no bias is predictive With three or more measurements we could have tracked of improved response to exposure treatment (Legerstee et al. changes in symptoms more precisely, and could also have ex- 2010; Price et al. 2011; Waters et al. 2012). Similar processes plicitly modeled the correlation between intercept and slope of may account for the present pattern of findings: A greater change rather than simply controlling for pre-transition out- readiness to acquire an attentional bias toward threat may have come scores. With a relatively modest N of 79, we were 80% facilitated children learning that a potentially feared and powered to detect a correlation of 0.31. While most of the key threatening school situation was not necessarily predictive of findings in our study exceeded this effect size, we were under- something aversive occurring leading to a reduction in anxiety poweredtodetectsmallereffects, and Type II errors may be symptoms and school concerns. present. We did not have diagnostic measures or data on treat- Change in attentional bias away from threat was not signif- ment use, and cannot discount the possibility that diagnostic icantly associated with change in anxiety or school concerns. and treatment status may have moderated the association be- The correlation between change in anxiety symptoms and tween attentional and interpretation bias and symptom change change in attentional bias away from threat was in the opposite across time. We used a measure of anxiety symptoms assessed direction and was significantly different from the correlation over a 3-month period rather than separate measures of state coefficient between change in anxiety and attentional bias and trait anxiety. The SCARED likely taps elements of both change toward threat (Z (57) = −2.18, p = 0.032). This suggests trait and state anxiety but appears to more highly correlated that reductions in anxiety across transition were specifically with trait anxiety (Monga et al. 2000). However, we were un- predicted by the degree to which individuals acquired an atten- able to unpick the impact of school transition on state and trait tional bias toward threat and not by a high level of general anxiety independently or to fully investigate whether partici- attentional plasticity. Measures of change in attentional bias pants with high trait anxiety differ in their trajectory and corre- toward and away from threat were not significantly correlated. lates of change from those participants who reported only being This is inconsistent with a small number of studies which have high state but not high trait anxious. Notwithstanding this, the observed that malleability in attentional bias toward and away findings in the subset of the sample exceeding clinical cut-off from threat are equally predicted by other factors, including scores for anxiety were very similar in direction and magnitude variation in the 5HTTLPR gene (Fox et al. 2011) and a mea- of effects to those observed in the entire sample (see sure of attentional control (Basanovic et al. 2017). Supplementary Materials). There are many other factors that When faced with ambiguous school situations, children with we could have measured using not only child but also parent heightened anxiety and school concerns at pre-transition were or teacher-report measures, (e.g. stressful life events, peer rela- more likely to resolve that ambiguity in a threatening way. tionships, bullying and victimisation, parenting styles and psy- Interpretation bias was not significantly associated with anxiety chopathology) and which may predict emotional adjustment or school concerns at post-transition or the change in these across school transition directly, or indirectly by influencing measures. Our findings are more consistent with interpretation change in children’s attentional responses to emotional stimuli. bias being a consequence or epiphenomenon of anxiety (Dodd For most children there were no persistent negative effects et al. 2012), and with prospective studies, which have found no of school transition. However, a concerning proportion of or minimal significant evidence for a longitudinal relationship children reported clinical levels of anxiety, and importantly, between interpretation bias and anxiety symptoms (Creswell higher anxiety symptoms at pre-transition were associated et al. 2011;Doddet al. 2012; Muris et al. 2004). with the retention of higher anxiety following transition. Pre-transition attentional bias for threat was not significantly These findings reiterate the importance of monitoring chil- associated with anxiety symptoms and school concerns, or the dren’s emotional wellbeing during a time of heightened stress, change in these measures. There is very limited research investi- which corresponds with a sensitive period for the develop- gating prospective associations between attentional biases and ment of anxiety disorders. This is the first study to demon- anxiety in child samples (Morales et al. 2015), and there is evi- strate an association between magnitude of change in atten- dence to suggest that attentional biases in children may only be tional bias toward threat and change in anxiety symptoms and observed at clinical levels of anxiety (Bar-Haim et al. 2007; school concerns in response to school transition. More re- Dudeney et al. 2015). We found only very minimal evidence search is needed to unravel the mechanisms underpinning this for any association with anxiety symptoms even when we con- association, and to determine whether this relationship can be fined our analyses to those children exceeding the clinical cut-off. exploited in intervention approaches. A recent meta-analysis found the effect size for the association between child anxiety and attentional bias was smaller compared Compliance with Ethical Standards to adults, less robust, and sensitive to important methodological Funding This research was funded by a UK Medical Research Council details such as task type and format (Dudeney et al. 2015). Experimental Medicine Grant awarded to Dr. Kathryn Lester and This study has several limitations. We had no control group Professor Thalia Eley (MR/J011762/1). T. C Eley is part funded by a of non-transitioning children and only two assessment points. J Abnorm Child Psychol (2019) 47:1521–1532 1531 program grant from the UK Medical Research Council (MR/M021475/ Brown, H. M., Eley, T. C., Broeren, S., Macleod, C., Rinck, M., Hadwin, 1). E. Carr and T. C. Eley are funded by the by the National Institute for J. A., & Lester, K. J. (2014). Psychometric properties of reaction Health Research (NIHR) Biomedical Research Centre at South London time based experimental paradigms measuring anxietyrelated and Maudsley NHS Foundation Trust and King’s College London. The information-processing biases in children. Journal of Anxiety views expressed are those of the authors and not necessarily those of the Disorders, 28(1), 97–107. https://doi.org/10.1016/j.janxdis.2013. NHS, the NIHR or the Department of Health. 11.004. Clarke, P. J. F., MacLeod, C., & Shirazee, N. (2008). Prepared for the worst: readiness to acquire threat bias and susceptibility to elevate Conflict of Interest The authors declare that they have no conflict of trait anxiety. Emotion, 8(1), 47–57. https://doi.org/10.1037/1528- interest. 3542.8.1.47. Clarke, P. J. F., Chen, N. T. M., & Guastella, A. J. (2012). Prepared for the Ethical Approval The study was ethically approved by King’s College best: readiness to modify attentional processing and reduction in London Psychiatry, Nursing and Midwifery Research Ethics Committee anxiety vulnerability in response to therapy. Emotion, 12(3), 487– (PNM/14/15-66). 494. https://doi.org/10.1037/a0025592. Craske, M. G., Treanor, M., Conway, C. C., Zbozinek, T., & Vervliet, B. Informed Consent All participants provided informed consent (parent) (2014). Maximizing exposure therapy: an inhibitory learning ap- and verbal assent (child). proach. Behaviour Research and Therapy, 58(Supplement C, 10– 23. https://doi.org/10.1016/j.brat.2014.04.006. Creswell, C., Shildrick, S., & Field, A. P. (2011). Interpretation of ambi- guity in children: a prospective study of associations with anxiety Open Access This article is distributed under the terms of the Creative and parental interpretations. Journal of Child and Family Studies, Commons Attribution 4.0 International License (http:// 20(2), 240–250. https://doi.org/10.1007/s10826-010-9390-7. creativecommons.org/licenses/by/4.0/), which permits unrestricted use, Cristea, I. A., Mogoase, C., David, D., & Cuijpers, P. (2015). Practitioner distribution, and reproduction in any medium, provided you give appro- review: cognitive bias modification for mental health problems in priate credit to the original author(s) and the source, provide a link to the children and adolescents: a meta-analysis. JournalofChild Creative Commons license, and indicate if changes were made. Psychology and Psychiatry and Allied Disciplines, 56(7), 723– 734. https://doi.org/10.1111/jcpp.12383. Dodd,H.F.,Hudson,J.L.,Morris, T. M.,&Wise,C.K.(2012). Interpretation bias in preschool children at risk for anxiety: a pro- spective study. Journal of Abnormal Psychology, 121(1), 28–38. https://doi.org/10.1037/a0024589. Dudeney, J., Sharpe, L., & Hunt, C. (2015). Attentional bias towards References threatening stimuli in children with anxiety: a meta-analysis. Clinical Psychology Review, 40,66–75. https://doi.org/10.1016/j. Anderson, L. W., Jacobs, J., Schramm, S., & Splittgerber, F. (2000). cpr.2015.05.007. School transitions: Beginning of the end or a new beginning? Evans, D., Borriello, G. A., & Field, A. P. (2018). A review of the aca- International Journal of Educational Research, 33(4), 325–339. demic and psychological impact of the transition to secondary edu- https://doi.org/10.1016/S0883-0355(00)00020-3. cation. Frontiers in Psychology, 9, 1482. https://doi.org/10.3389/ Bar-Haim, Y., Lamy, D., Lee, P., Bakermans-Kranenburg, M. J., & van fpsyg.2018.01482. Ijzendoorn, M. H. (2007). Threat-related attentional bias in anxious Fox, E., Zougkou, K., Ridgewell, A., & Garner, K. (2011). The serotonin and nonanxious individuals: a meta-analytic study. Psychological transporter gene alters sensitivity to attention bias modification: ev- Bulletin, 133(1), 1–24. https://doi.org/10.1037/0033-2909.133.1.1. idence for a plasticity gene. Biological Psychiatry, 70(11), 1049– Barry, T. J., Vervliet, B., & Hermans, D. (2015). An integrative review of 1054. https://doi.org/10.1016/j.biopsych.2011.07.004. attention biases and their contribution to treatment for anxiety dis- Grills-Taquechel, A. E., Norton, P., & Ollendick, T. H. (2010). A longi- orders. Frontiers in Psychology, 6(968). https://doi.org/10.3389/ tudinal examination of factors predicting anxiety during the transi- fpsyg.2015.00968. tion to middle school. Anxiety, Stress, and Coping, 23(5), 493–513. Basanovic, J., Notebaert, L., Grafton, B., Hirsch, C. R., & Clarke, P. https://doi.org/10.1080/10615800903494127. (2017). Attentional control predicts change in bias in response to Hakamata, Y., Lissek, S., Bar-Haim, Y., Britton, J. C., Fox, N. A., attentional bias modification. Behavior Research and Therapy, 99, Leibenluft, E., ... Pine, D. S. (2010). Attention Bias modification 47–56. https://doi.org/10.1016/j.brat.2017.09.002. treatment: a meta-analysis toward the establishment of novel treat- Biehl, M., Matsumoto, D., Ekman, P., Hearn, V., Heider, K., Kudoh, T., & ment for anxiety. Biological Psychiatry, 68(11), 982–990. https:// Ton, V. (1997). Matsumoto and Ekman’s Japanese and Caucasian doi.org/10.1016/j.biopsych.2010.07.021. facial expressions of emotion (JACFEE): reliability data and cross- Hirsch, B. J., & DuBois, D. L. (1992). The relation of peer social support national differences. Journal of Nonverbal Behavior, 21,2–21. and psychological symptomatology during the transition to junior https://doi.org/10.1023/A:1024902500935. high school: a two-year longitudinal analysis. American Journal of Birmaher, B., Brent, D. A., Chiappetta, L., Bridge, J., Monga, S., & Community Psychology, 20(3), 333–347. https://doi.org/10.1007/ Baugher, M. (1999). Psychometric properties of the screen for child BF00937913. anxiety related emotional disorders (SCARED): a replication study. Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Journal of the American Academy of Child and Adolescent Walters, E. E. (2005). Lifetime prevalence and age-of-onset distri- Psychiatry, 38(10), 1230–1236. https://doi.org/10.1097/00004583- butions of DSM-IV disorders in the national comorbidity survey 199910000-00011. replication. Archives of General Psychiatry, 62(6), 593–602. https://doi.org/10.1001/archpsyc.62.6.593. Brown, H. M., McAdams, T. A., Lester, K. J., Goodman, R., Clark, D. M., & Eley, T. C. (2013). Attentional threat avoidance and familial Langner, O., Dotsch, R., Bijlstra, G., Wigboldus, D. H. J., Hawk, S. T., & van Knippenberg, A. (2010). Presentation and validation of the risk are independently associated with childhood anxiety disorders. Journal of Child Psychology and Psychiatry, 54(6), 678–685. Radboud faces database. Cognition and Emotion, 24(8), 1377– https://doi.org/10.1111/jcpp.12024. 13 88. https://doi.org/10.1080/02699930903485076. 1532 J Abnorm Child Psychol (2019) 47:1521–1532 Lau, J. Y. F., & Waters, A. M. (2017). Annual research review: an ex- therapy for social phobia. Depression and Anxiety, 28(4), 349– 353. https://doi.org/10.1002/da.20791. panded account of information-processing mechanisms in risk for child and adolescent anxiety and depression. Journal of Child Rice, F., Frederickson, N., & Seymour, J. (2011). Assessing pupil con- Psychology and Psychiatry, 58(4), 387–407. https://doi.org/10. cerns about transition to secondary school. British Journal of 1111/jcpp.12653. Educational Psychology, 81(Pt 2, 244–263. https://doi.org/10. Legerstee, J. S., Tulen, J. H. M., Kallen, V. L., Dieleman, G. C., Treffers, 1348/000709910X519333. P. D. A., Verhulst, F. C., & Utens, E. M. W. J. (2009). Threat-related Rice, F., Frederickson, N., Shelton, K., McManus, C., Riglin, L., & Ng- selective attention predicts treatment success in childhood anxiety Knight, T. (2015). Identifying factors that predict successful and disorders. Journal of the American Academy of Child and difficult transitions to secondary school. Nuffield Foundation. Adolescent Psychiatry, 48(2), 196–205. https://doi.org/10.1097/ Riglin, L., Frederickson, N., Shelton, K. H., & Rice, F. (2013). A longi- CHI.0b013e31819176e4. tudinal study of psychological functioning and academic attainment Legerstee, J. S., Tulen, J. H. M., Dierckx, B., Treffers, P. D. A., Verhulst, at the transition to secondary school. Journal of Adolescence, 36(3), F. C., & Utens, E. M. W. J. (2010). CBT for childhood anxiety 507–517. https://doi.org/10.1016/j.adolescence.2013.03.002. disorders: differential changes in selective attention between treat- Riglin, L., Petrides, K. V., Frederickson, N., & Rice, F. (2014). The ment responders and non-responders. Journal of Child Psychology relationship between emotional problems and subsequent school and Psychiatry, 51(2), 162–172. https://doi.org/10.1111/j.1469- attainment: a meta-analysis. Journal of Adolescence, 37(4), 335– 7610.2009.02143.x. 346. https://doi.org/10.1016/j.adolescence.2014.02.010. Lester, K. J., Seal, K., Nightingale, Z. C., & Field, A. P. (2010). Are Smith, J. S., Akos, P., Lim, S., & Wiley, S. (2008). Student and stake- children’s own interpretations of ambiguous situations based on holder perceptions of the transition to high school. The High School how they perceive their mothers have interpreted ambiguous situa- Journal, 91(3), 32–42. tions for them in the past? Journal of Anxiety Disorders, 24,102– Stirling, L. J., Eley, T. C., & Clark, D. M. (2006). Preliminary evidence 108. https://doi.org/10.1016/j.janxdis.2009.09.004. for an association between social anxiety symptoms and avoidance Lester, L., Waters, S., & Cross, D. (2013). The relationship between of negative faces in school-age children. Journal of Clinical Child & school connectedness and mental health during the transition to Adolescent Psychology, 35(3), 431–439. https://doi.org/10.1207/ secondary school: a path analysis. Australian Journal of Guidance s15374424jccp3503_9. and Counselling, 23(2), 157–171. https://doi.org/10.1017/jgc.2013. Stradling, R., & MacNeil, M. (2001). Moving on : The emotional well- being of young people in transition from primary to secondary Lohaus, A., Elben, C., Ball, J. E., & Klein-hessling, J. (2004). School school. Inverness: Highland Health Board. transition from elementary to secondary school: changes in psycho- Suarez, L., & Bell-Dolan, D. (2001). The relationship of child worry to logical adjustment. Educational Psychology, 24(2), 161–173. cognitive biases: threat interpretation and likelihood of event occur- https://doi.org/10.1080/0144341032000160128. rence. Behavior Therapy, 32(3), 425–442. https://doi.org/10.1016/ MacLeod, C. (1999). Anxiety and anxiety disorders. In T. Dalgleish & M. S0005-7894(01)80029-0. Power (Eds.), Handbook of cognition and emotion (pp. 447–478). Thomasson, R., Field, L., O'Donnell, C., & Woods, S. (2006). School Chichester: England: Wiley. concerns questionnaire. Buckinghamshire: Buckinghamshire MacLeod, C., & Clarke, P. J. F. (2015). The attentional bias modification County Council. approach to anxiety intervention. Clinical Psychological Science, 3(1), 58–78. Tottenham, N., Tanaka, J. W., Leon, A. C., McCarry, T., Nurse, M., Hare, Monga, S., Birmaher, B., Chiappetta, L., Brent, D., Kaufman, J., Bridge, T. A., ... Nelson, C. (2009). The NimStim set of facial expressions: J., & Cully, M. (2000). Screen for child anxiety-related emotional Judgments from untrained research participants. Psychiatry disorders (SCARED): Convergent and divergent validity. Research, 168(3), 242–249. https://doi.org/10.1016/j.psychres. Depression and Anxiety, 12(2), 85–91. https://doi.org/10.1002/ 2008.05.006. 1520-6394(2000)12:2<85::AID-DA4>3.0.CO;2-2. Van Bockstaele, B., Verschuere, B., Tibboel, H., De Houwer, J., Morales, S., Pérez-Edgar, K. E., & Buss, K. A. (2015). Attention biases Crombez, G., & Koster, E. H. W. (2014). A review of current evi- towards and away from threat mark the relation between early dys- dence for the causal impact of attentional bias on fear and anxiety. regulated fear and the later emergence of social withdrawal. Journal Psychological Bulletin, 140(3), 682–721. https://doi.org/10.1037/ of Abnormal Child Psychology, 43(6), 1067–1078. https://doi.org/ a0034834. 10.1007/s10802-014-9963-9. Waters, A. M., Mogg, K., & Bradley, B. P. (2012). Direction of threat Muris, P., Jacques, P., & Mayer, B. (2004). The stability of threat percep- attention bias predicts treatment outcome in anxious children receiv- tion abnormalities and anxiety disorder symptoms in non-clinical ing cognitive-behavioural therapy. Behaviour Research and children. Child Psychiatry and Human Development, 34(3), 251– Therapy, 50(6), 428–434. https://doi.org/10.1016/j.brat.2012.03. 265. https://doi.org/10.1023/B:CHUD.0000015000.13845.9d. Ng-Knight, T., Shelton, K. H., Riglin, L., McManus, I. C., Frederickson, West, P., Sweeting, H., & Young, R. (2010). Transition matters: pupils’ N., & Rice, F. (2016). A longitudinal study of self-control at the experiences of the primary-secondary school transition in the West transition to secondary school: considering the role of pubertal status of Scotland and consequences for well-being and attainment. and parenting. JournalofAdolescence, 50,44–55. https://doi.org/ Research Papers in Education, 25(1), 21–50. https://doi.org/10. 10.1016/j.adolescence.2016.04.006. 1080/02671520802308677. Osinsky, R., Lösch, A., Hennig, J., Alexander, N., & MacLeod, C. Zeedyk, M. S., Gallacher, J., Henderson, M., Hope, G., Husband, B., & (2012). Attentional bias to negative information and 5-HTTLPR Lindsay, K. (2003). Negotiating the transition from primary to sec- genotype interactively predict students’ emotional reactivity to first ondary school. School Psychology International, 24(1), 67–79. university semester. Emotion, 12(3), 460–469. https://doi.org/10. https://doi.org/10.1177/0143034303024001010. 1037/a0026674. Pearl, J. (2016). Lord’s paradox revisited – (oh Lord! Kumbaya!). Journal of Causal Inference, 4(2). https://doi.org/10.1515/jci-2016-0021. Publisher’sNote Springer Nature remains neutral with regard to Price, M., Tone, E. B., & Anderson, P. L. (2011). Vigilant and avoidant jurisdictional claims in published maps and institutional affiliations. attention biases as predictors of response to cognitive behavioral

Journal

Journal of Abnormal Child Psychology – Springer Journals

Published: Mar 20, 2019

There are no references for this article.