J Autism Dev Disord DOI 10.1007/s10803-009-0725-4
ORIGINAL PAPER
Reactions to Ostracism in Adolescents with Autism Spectrum Conditions Catherine Sebastian Æ Sarah-Jayne Blakemore Æ Tony Charman
Ó Springer Science+Business Media, LLC 2009
Abstract Little is known about how adolescents with autism spectrum conditions (ASC) experience the initial impact of ostracism. This study investigated whether a mild, short-term episode of experimentally induced ostracism (Cyberball) would affect self-reported anxiety, mood, and the extent to which four social needs (self-esteem, belonging, control and meaningful existence) were threatened in adolescents with ASC and matched controls. Anxiety and the four needs were negatively affected by ostracism in both groups. However, ostracism did not modulate mood in the ASC group, and a number of possible interpretations of this group difference are discussed. In general, the results of this study suggest that normative models of ostracism are applicable to ASC. Keywords Autism spectrum conditions (ASC) Ostracism Adolescence Self-report
Introduction Individuals with autism spectrum conditions (ASC) often report a desire to take part in social interactions and C. Sebastian (&) S.-J. Blakemore Institute of Cognitive Neuroscience, University College London, 17 Queen Square, London WC1N 3AR, UK e-mail:
[email protected] T. Charman UCL Institute of Child Health, Guilford Street, London WC1N 3EH, UK Present Address: T. Charman Department of Psychology and Human Development, Institute of Education, 20 Bedford Way, London WC1H 0AL, UK
develop friendships (Frith 2004), but are often rejected due to their poor social skills (Attwood 1998). This is likely a contributing factor to high rates of affective disorder in this group, particularly in high functioning individuals who would be expected to be most aware of their difference from peers (Tantam 2000). Adolescence can be a particularly difficult time for individuals with ASC. Not only do individuals themselves wish to socialise more, but successful interaction also requires increasingly sophisticated social skills (Attwood 1998; Howlin 2003; Lawson 2001). Ostracism, a potent form of social aggression, is defined as being ignored or rejected by an individual or group (Williams 2007). However, it is not known how adolescents with ASC react to ostracism, and whether models of ostracism developed from data on typical adults are applicable to adolescents with ASC. Individuals with high-functioning ASC are often aware of their poor social skills, and of their difference from peers. These differences become particularly salient during adolescence (Howlin 2003). A recent study found that 10– 16 year olds with Asperger syndrome (AS) often compared themselves unfavourably with others on social dimensions. Individuals who perceived their peer group membership to be lower than other people’s were more likely to report depressive symptoms (Hedley and Young 2006). However, perceived social attractiveness and rank did not predict depressive scores, in contrast to the normative literature (Allan and Gilbert 1995). These results suggest that perceived inclusion or ostracism from the peer group is at least partially preserved in AS, as is the importance of group membership for psychological wellbeing. Several other studies on adolescents with ASC indicate that this group may be particularly vulnerable to the effects of ostracism. Kuusikko et al. (2008) found that children and adolescents with high functioning autism/AS report
123
J Autism Dev Disord
significantly more social anxiety symptoms than do their TD peers, and that these symptoms increase as individuals get older, in contrast to the pattern in TD peers. Additionally, Williamson et al. (2008) found that adolescents with AS perceive themselves to hold lower peer approval than TD controls, while placing the same emphasis on its importance. However, studies to date have not compared ASC and TD adolescent reactions to ostracism in an experimental setting, nor have they looked at these reactions in terms of well-validated models of ostracism. Several models of the affective and behavioural impact of ostracism have been developed using normative adult data (see Williams 2007 for a review). These are based on experimentally induced, short term, one-off instances of ostracism by others, often strangers. The focus in the current study is Williams’ (1997, 2001) Need Threat account, in which ostracism is proposed to threaten four fundamental psychological needs: self-esteem, belonging, control and a sense of meaningful existence. Several studies have found that these needs are threatened (a construct termed ‘need threat’) following ostracism in typical adults (Williams et al. 2000; Zadro et al. 2004, 2006), and that typical adults attempt to refortify these needs, for example, by conforming more to group norms (Williams et al. 2000). In these studies, ostracism was manipulated using Williams et al.’s (2000) ‘Cyberball’ paradigm. In this task, participants play a ball game over the internet with two other players, whose actions are pre-programmed. However, participants are led to believe they are playing with real individuals. After a few throws, the other players stop throwing the ball to the participants. This paradigm reliably induces feelings of rejection, while representing a relatively mild and controlled form of ostracism (Eisenberger et al. 2003; Sebastian et al. 2008; Williams et al. 2000; Zadro et al. 2004, 2006), and was therefore chosen for use in the current study. As these theories are based on the typical adult population, it is unknown whether adolescents with ASC react to ostracism in the same way. A recent study investigating the Need Threat model in TD adolescent girls showed that this group are more sensitive to some aspects of the ostracism experience than are adults (Sebastian et al. 2008). The current study also used the Need Threat model as a framework for investigating reactions to ostracism in ASC compared with matched TD controls, and examined whether aspects of this model might need to be adapted for the ASC population. The consequences of ostracism were measured via selfreport scales, and included measures of anxiety and mood as well as the Need Threat scale developed by Williams et al. (2000). Although individuals with ASC can lack insight into their own mental states and emotions (e.g. Hill et al. 2004), several recent studies have shown that self-
123
report can be a useful methodology for this group when reporting on their everyday experience (Baron-Cohen et al. 2001; Berthoz and Hill 2005; Chalfant et al. 2007; Hedley and Young 2006; Kim et al. 2000; Kuusikko et al. 2008; Williamson et al. 2008). We predicted that need threat would be similarly high after ostracism in both TD and ASC adolescents. This is because need threat is argued to be an automatic, reflexive response to ostracism (Williams 2007), and there is currently no evidence to suggest that individuals with ASC lack this response. There are several possible predictions that could be made for the effect of ostracism on mood and anxiety in ASC. It is possible that mood and anxiety in the ASC group would be more affected by ostracism than in the control group, since this group finds social interaction stressful, report high levels of social anxiety, and compare themselves negatively with others in the social domain (Hedley and Young 2006; Kuusikko et al. 2008; Tantam 2000; Williamson et al. 2008). On the other hand, there is evidence that individuals with ASC underestimate the extent of their physiological emotional responses (Ben Shalom et al. 2006; Silani et al. 2008). Therefore, it is possible that anxiety and mood in the ASC group would be less affected by ostracism than in the control group.
Method Participants The sample comprised 29 participants (all male): 13 adolescents with ASC (mean age 16.9, SD 0.3) and 16 typically developing (TD) controls (mean age 16.9, SD 0.7). Data from three controls were excluded from the analysis due to distraction during testing. The participants were a convenience subsample of the SNAP cohort (Baird et al. 2006) with a minimum IQ of 90. All participants in the ASC group had received a consensus clinical ICD-10 diagnosis of childhood autism (N = 6) or ‘other ASC’ (N = 7; ‘PDD other’ in six cases due to sub-threshold symptomatology and in one case ‘PDD unspecified’ due to lack of information on early development) following an assessment including the autism diagnostic interviewrevised (ADI-R; Lord et al. 1994), the ADOS-G (Lord et al. 2000) as well as IQ, language and adaptive behaviour measures. TD participants were recruited from a database held by the Institute of Child Health and none had dyslexia, epilepsy, or any other neurological conditions by parent and self-report. All participants gave informed consent to take part in the study, which was approved by the local ethics committee. The groups (after excluding the three TD participants) were matched on age and IQ. IQ was measured using the
J Autism Dev Disord
Wechsler Abbreviated Scale of Intelligence-UK (WASI: Wechsler 1999). Mean (SD) full scale IQ was 100.46 (7.49) in the ASC group, and 102.46 (8.62) in the control group. All participants had an IQ of over 90.
proportion of the time they had received the ball, given that 33% was an equal share.
Design
The extent to which inclusion and ostracism affected the four needs was measured using the Need Threat questionnaire (Williams et al. 2000). This scale consists of 12 statements; three questions pertaining to each need (selfesteem, belonging, control, meaningful existence). The self-esteem measure correlates highly with the Rosenberg (1968) self-esteem scale (Kipling D. Williams, personal communication), and so this section of the questionnaire was used to measure baseline self-esteem, in addition to the impact of Cyberball. The other three needs were specific to the experience of inclusion and ostracism, and so were only measured after the two Cyberball games. Participants rated how much each statement described their reactions during each game on a scale from 1 to 5.
The design was a 2 9 3 factorial with between-subjects factor group (ASC, Control), and within-subjects factor condition (baseline, inclusion, ostracism). Ostracism Manipulation Cyberball (Williams et al. 2000; Williams and Jarvis 2006) was used as the ostracism manipulation. In order to deflect attention from the true motivation of the study, participants were told that the aim of the task was to look at ‘mental visualisation ability’. Ethical permission for this minor deception was obtained from the local research ethics committee. Players were represented on the computer screen by cartoon drawings, with the participant’s character always located at the bottom centre (Fig. 1a). They could choose to throw the ball to the players on either their left or their right by pressing corresponding keyboard buttons. The game comprised 70 throws, lasting around 3 min. The probability that the other players would throw the ball to the participant systematically varied according to condition. Inclusion always preceded ostracism in order to avoid negative spill-over effects, which would have been theoretically more problematic than spill-over effects from inclusion in terms of the hypotheses. This same fixed order strategy has been used whenever Cyberball has been employed in within-subjects design (Eisenberger et al. 2003, 2007). In the inclusion condition, participants were in possession of the ball 33% of the time (equal inclusion). In the ostracism condition the confederates were initially programmed to throw the ball to the participant with equal probability; however, after the first eight throws, they stopped throwing it to the participant altogether for the remainder of the game (around 50 throws). Dependent Measures Manipulation Check To make sure participants realised they were being included or ostracised, each participant was asked to rate how much he agreed with two statements on a scale from 1 to 5 after each run of Cyberball (‘‘I was ignored’’ and ‘‘I was excluded’’). Participants were also asked to estimate what
Need Threat
Anxiety Anxiety levels were measured using the State/Trait Anxiety Inventory (Spielberger 1983). This consists of 20 statements for each subscale [State (STAI-S) and Trait (STAI-T)]. Participants rated how much each statement described them on a scale from 1 to 4. Trait scores were measured at baseline, and state anxiety was measured at baseline as well as after inclusion and ostracism conditions. Mood Participants rated how good/bad, happy/sad, friendly/ unfriendly and tense/relaxed they were currently feeling, on a scale of 1–7. These anchors comprised the mood section of Williams et al.’s (2000) Need Threat questionnaire. Procedure See Fig. 1b for a schematic of the experimental procedure. The ADOS-G was completed in a separate session, prior to the study. Participants initially completed the WASI IQ measure; then baseline self-report measures of state and trait anxiety and the self-esteem and mood components of the need threat questionnaire. Participants then played Cyberball (inclusion condition). Participants were told that they would be playing with two other boys over the internet, and that it was important for them to visualise the experience as much as possible. After this condition, participants filled out the state anxiety and need threat questionnaires. Participants then played Cyberball in the ostracism condition, and again completed the
123
J Autism Dev Disord Fig. 1 The Cyberball game and a schematic of the experimental procedure: a The Cyberball game: participants were represented by a cartoon at the bottom of the screen, with computerised characters on either side. In the ostracism condition, the other characters did not throw the ball to the participant after the first few throws. b A schematic of the time course of the experiment. On arrival, participants completed an IQ test and the baseline measures. They then played Cyberball (first inclusion, then ostracism). After each run of Cyberball they completed a battery of selfreport measures
(a)
Participant’s name
(b)
Cyberball (Inclusion)
Cyberball (Ostracism)
Debriefing
Timeline IQ test
Baseline measures: - Mood - State/trait anxiety - Self-esteem
questionnaires. At the end of the experiment, they completed the manipulation check measures, and were then fully debriefed as to the aims of the study. It was made clear that the computerised characters in Cyberball were not real. No participant expressed regret at having taken part.
Results Data from 13 ASC and 13 control participants were analysed. Data were analysed using mixed-model ANOVAs, with group (ASC, control) as the between-subjects factor, and condition (baseline, inclusion, ostracism) as the withinsubjects factor. Greenhouse Geisser corrections were used to correct for non-sphericity where necessary. Descriptive statistics are reported in Table 1. ASC symptoms as measured by the ADOS-G did not correlate with any measure, and will not be further discussed. Manipulation Check One participant in the control group did not complete the manipulation check measure. Responses to the manipulation check items ‘I felt ignored’ and ‘I felt excluded’ were amalgamated. Ratings showed that both groups realised they had been included and ostracised in the appropriate conditions. On a scale of 1–5, with five meaning participants felt fully included, means (SD) in the inclusion condition were 4.4 (.6) in the ASC group, and 4.5 (.6) in the control group. In the ostracism condition, mean ratings were 1.7 (.9) in the ASC group and 1.6 (.9) in the control group. Thus, both groups realised they had been ostracised
123
Paul
Tom
Inclusion measures: - Mood - State anxiety - Need threat
Ostracism measures: - Mood - Need threat - State anxiety - Manipulation check
in the ostracism condition. In a 2 9 2 ANOVA between group (ASC, control) and condition (inclusion, ostracism), there was a main effect of condition (F(1,23) = 123.08, p \ .001, g2p = .84) with higher mean ratings after inclusion than after ostracism. However, there was no main effect of group (F(1,23) \ .001, p = .99, g2p \ .001) and no interaction (F(1,23) = .197, p = .661, g2p = .008). On average participants were in possession of the ball 33% of the time during inclusion, and 11% during ostracism. Mean (SD) estimates of possession were 38.9% (13.7%) for inclusion and 10.1% (6.1%) for ostracism in the ASC group, and 36.0% (7.3%) for inclusion and 9.4% (4.7%) for ostracism in the control group. These estimates did not differ from actual possession in either condition or group in one-sample t-tests (all ps [ .15). In a 2 9 2 (condition 9 group) ANOVA with possession estimates as the dependent variable, there was a main effect of condition (F(1,23) = 201.48, p \ .001, g2p = .90), with higher estimates for inclusion than ostracism, but no main effect of group (F(1,23) = .38, p = .55, g2p = .02) or interaction (F(1,23) = .34, p = .57, g2p = .01). This suggests that the groups did not differ in their ratings of the objective aspects of the two conditions. Need Threat Self-esteem Self-esteem was measured at all three time points (baseline, inclusion, ostracism). A 2 9 3 (group 9 condition) ANOVA showed a main effect of condition (F(2,48) = 32.57, p \ .001, g2p = .58). Post hoc tests showed that selfesteem was significantly lower after ostracism than both
J Autism Dev Disord Table 1 Mean (SD) self-report data for ASC and control groups on the State/Trait Anxiety Inventory, and Williams et al’s (2000) Need Threat and Mood questionnaires ASC (n = 13) Baseline
Control (n = 13) Inclusion
Ostracism
Baseline
Inclusion
Ostracism
Anxiety Trait
40.5 (10.0)
–
–
39.0 (8.4)
–
–
State Need threat
33.2 (7.9)
31.2 (7.4)
32.9 (7.6)
33.4 (7.7)
31.9 (6.3)
33.1 (8.1)
3.8 (.86)
3.6 (.8)
2.5 (.9)
4.1 (.7)
3.5 (.7)
2.3 (.8)
Self-esteem Belonging
–
4.2 (1.0)
2.4 (1.3)
–
4.6 (.8)
2.3 (1.6)
Control
–
3.0 (1.1)
2.0 (1.2)
–
2.8 (1.3)
1.4 (.9)
Meaningful existence
–
4.5 (1.0)
2.4 (1.2)
–
4.5 (.7)
3.5 (1.4)
Good/bad
5.8 (.6)
5.4 (1.0)
5.2 (1.2)
6.0 (.7)
6.1 (1.0)
4.8 (1.4)
Happy/sad
5.4 (.9)
5.5 (.8)
5.3 (1.0)
5.6 (1.1)
5.8 (1.1)
4.8 (1.3)
Friendly/unfriendly
6.1 (1.0)
6.2 (.8)
5.3 (1.4)
5.9 (1.3)
5.9 (1.1)
5.1 (.1.6)
Tense/relaxed
4.8 (2.0)
5.0 (2.1)
5.7 (1.3)
5.5 (1.8)
5.8 (1.1)
4.6 (1.8)
Overall mood
5.5 (.7)
5.5 (.6)
5.4 (.8)
5.8 (.8)
5.9 (1.0)
4.8 (1.3)
Mood
Need threat is measured on a scale from 1 to 5, with 1 indicating most threat; mood is measured on a scale from 1 to 7, with 1 indicating lowest mood
baseline and inclusion in both groups (ps \ .001). There was no main effect of group (F(1,24) = .002, p = .97, g2p \ .001), or interaction between group and condition (F(1,24) = .65, p = .53, g2p = .03). Belonging After each game of Cyberball, participants reported how much they had felt part of the group while playing (inclusion and ostracism conditions). A 2 9 2 (group 9 condition) ANOVA revealed a main effect of condition (F(1,24) = 44.19, p \ .001, g2p = .65), due to lower ratings of belonging after ostracism compared to inclusion across groups (p \ .001). There was no main effect of group (F(1,24) = .10, p = .75, g2p = .004), or interaction between group and condition (F(1,24) = .77, p = .39, gp2 = .03).
Meaningful Existence On this measure, participants were asked whether they felt visible and that their life had meaning during each of the Cyberball games. On a 2 9 2 (group 9 condition) ANOVA, there was a main effect of condition (F(1,24) = 35.55, p \ .001, g2p = .60), due to lower ratings of meaningful existence after ostracism across groups (p \ .001). There was no main effect of group (F(1,24) = 2.67, p = .11, g2p = .10); however, there was an interaction between group and condition (F(1,24) = 3.59, p = .051, g2p = .15). Simple effects analysis showed that both groups showed significantly lower ratings after ostracism than inclusion (for ASC, p \ .001; for controls, p = .011). However, while there were no group differences in ratings after inclusion (p = .88), ratings after ostracism were significantly lower in the ASC group than in the TD group (p = .041, Fig. 2).
Control
Anxiety
Participants also retrospectively reported how much they felt they were in control of the interaction during each game. Again, in a 2 9 2 (group 9 condition) ANOVA there was a main effect of condition (F(1,24) = 14.34, p = .001, gp2 = .37), due to lower ratings of control after ostracism compared to inclusion across groups (p \ .001). There was no main effect of group (F(1,24) = 1.63, p = .21, g2p = .06), or interaction between group and condition (F(1,24) = .23, p = .63, g2p = .01).
An independent samples t-test showed no group differences in trait anxiety, measured at baseline: t(24) = .43, p = .68. Means (SD) were 40.54 (9.99) in the ASC group and 39.0 (8.38) in the control group. A 2 9 3 (group 9 condition) ANOVA conducted on state anxiety showed a trend towards a main effect of condition (F(2,48) = 2.95, p = .062, g2p = .11; Fig. 3). Post hoc tests showed that state anxiety was lower after inclusion than both baseline (p = .030) and ostracism (p = .027).
123
7
5
ASD
ASD
*
Control
Overall Mood (1-7 scale)
Meaningful Existence Rating (1-5 scale)
J Autism Dev Disord
4
3
2
1
*
Control
6
5
4
3
Inclusion
Baseline
Ostracism
*p < .05
*p < .05
Condition
Fig. 2 Interaction between group and condition for self-reported need threat for meaningful existence, on a scale of 1–5. Significantly lower scores after ostracism than inclusion in both groups indicate that sense of meaningful existence was threatened. However, this effect was greater in the autism spectrum condition (ASC) group, with significantly lower ratings on this measure than the control group following ostracism 40 ASD 38
*
*
Control
36 State Anxiety
*
34
Inclusion
Ostracism
Condition
Fig. 4 Overall mood ratings for each condition. Mood did not differ between conditions for the autism spectrum condition ASC group. However, mood was significantly lower in the control group after ostracism than at baseline or after inclusion
(F(1.5,36.3) = 3.03, p = .074, g2p = .11). There were no differences between any of the conditions in the ASC group (p [ .70). However, in the control group, there was a significant reduction in mood between both baseline and ostracism (p = .01) and inclusion and ostracism (p = .004), while there was no reduction in mood between baseline and inclusion conditions (p = .49; Fig. 4).
32 30
Discussion
28 26 24 *p < .05
Baseline
Inclusion
Ostracism
Condition
Fig. 3 Self-reported state anxiety (possible scores range from 20 to 80). Anxiety was reduced in both groups after inclusion, and increased back to baseline levels following ostracism
There was no main effect of group (F(1,24) = .02, p = .89, g2p = .001), or interaction (F(2,48) = .07, p = .94, g2p = .003). Mood The four mood questions were collapsed into an overall mood score (as has been reported previously; Williams et al. 2000; Zadro et al. 2004), and a 2 9 3 (group 9 condition) ANOVA was conducted. There was a main effect of condition (F(1.5,36.3) = 4.80, p = .022, g2p = .17), due to overall lower mood after ostracism than after baseline (p = .035) and inclusion (p = .020). There was no main effect of group (F(1,24) = .03 p = .87, g2p = .001). There was a marginal interaction between group and condition
123
The current study sought to investigate whether adolescents with ASC would respond in the same way as TD control adolescents to an experimentally induced episode of ostracism. The study used the Cyberball paradigm (Williams et al. 2000), which to our knowledge has not been used with individuals with ASC before. The results suggest that individuals with ASC are as able as controls to recognise when they are being excluded from a social situation; they also report similar effects of ostracism on anxiety and on need threat (this effect was exaggerated in the case of meaningful existence ratings). In contrast, participants with ASC did not report any differences in overall mood following ostracism, while mood was lowered by ostracism in the control group in line with previous studies (Williams et al. 2000; Zadro et al. 2004). The groups did not differ on any manipulation check measure. Both groups fairly accurately estimated their possession of the ball in each game, and also recognised that they had been included and excluded in the appropriate conditions. This suggests that group differences on other dependent measures cannot be explained by basic differences in perception of the Cyberball game.
J Autism Dev Disord
As predicted, both groups reported significant need threat during ostracism for all four needs (self-esteem, belonging, control, and meaningful existence). This provides support for Williams’ (1997, 2001, 2007) Need Threat model, which argues that need threat occurs independent of personality or social/situational factors. The current findings suggest that this automatic response is preserved in individuals with ASC, and belies the characterisation of this group as lacking the same social needs as typically developing individuals. However, our results also show that there is some flexibility in the degree to which different needs are threatened in different populations. While the degree of self-reported need threat did not differ between groups for self-esteem, belonging and control, we found greater need threat in the ASC group, compared with controls, on the meaningful existence scale. Inspection of data suggests that this need was not threatened as much as the other three needs in the control group, rather than this need being disproportionately threatened in the ASC group. In the original study using the Cyberball paradigm, Williams et al. (2000) found only weak support for the effect of the ostracism game on meaningful existence, and suggested that this might be because of the short term nature of the ostracism experience. The TD participants are more likely to have a well developed support network that would prevent them feeling ‘invisible’, ‘meaningless’, or ‘non-existent’ during a short, virtual episode of ostracism, while the ASC group may have lacked this resource. In the current study, groups did not differ on trait anxiety or on state anxiety measured at baseline. This enabled us to explore whether anxiety after both a positive (inclusion) and a negative (ostracism) social encounter would differ between ASC and TD control groups with similar levels of baseline anxiety. State anxiety in both groups was reduced by the inclusion condition, and it increased back to baseline levels following ostracism. It seems, then, that inclusion had a mild anxiolytic effect on both groups. This supports various accounts of high functioning ASC in which individuals, like TD individuals, seek and value social inclusion (Attwood 1998; Frith 2004; Howlin 2003). While it is slightly surprising that anxiety after ostracism did not differ from baseline anxiety, it did increase significantly in both groups between inclusion and ostracism, showing modulation of self-reported anxiety in the expected direction in participants with ASC as well as controls. Participants were also asked to rate their current mood at baseline and after inclusion/ostracism. Mood in the control group was significantly lower after ostracism than at baseline and inclusion; in contrast mood ratings in the ASC group were very similar at all time points. Since previous studies tend to show that ostracism using the Cyberball paradigm reduces overall mood (Williams et al. 2000;
Zadro et al. 2004) this lack of an effect in the ASC group represents a potentially important difference in reactions to ostracism between ASC and TD adolescent groups. There are a number of possible explanations for this effect. We think it is unlikely that the ASC group responded perseveratively on the mood questionnaire, as they did not respond in this way on the anxiety or needs measures, and were able to respond appropriately to individual mood items that were reverse-coded, in line with previous selfreport studies in ASC (Berthoz and Hill 2005). One possibility is that this result represents a qualitative difference in the experience of ostracism between the groups. While need threat and anxiety were similarly affected in both groups, ostracism may not have impacted upon mood in quite the same way as in ASC participants, with this difference being accurately reported by participants. Alternatively, the ASC group may have been equally affected by the ostracism condition as TD controls, but might have lacked insight into how the experience affected their current mood. While it could be argued that the state anxiety inventory (on which no group differences were found) also requires participants to introspect on their current internal state, many of the STAI-S questions ask about physical states (‘I feel rested’; ‘I feel jittery’), which may be easier to gauge than the more abstract mood questions which refer to emotional states (‘I feel friendly’; ‘I feel sad’). Indeed, several previous studies have found that individuals with ASC may under-report their affective responses to emotional stimuli, in the context of intact physiological response (Ben Shalom et al. 2006; Silani et al. 2008). Additionally, individuals with ASC have been found to be more prone to alexithymia than the general population (meaning individuals report having difficulties interpreting their own emotional states: Hill et al. 2004). In light of this evidence, the latter explanation of the mood data is perhaps most plausible. However, we recognise that one of the limitations of self-report methodology is that it is difficult to arbitrate between these two competing explanations. However, in terms of assessing reactions to social phenomena such as ostracism, it has the strength of allowing direct access to the way in which a participant has gauged their own response to an experimental manipulation, which physiological or parent report methods do not allow. While these findings suggest that caution is necessary when interpreting self-report data in this group, they do not preclude the use of these measures. Indeed, several recent studies have used self-report methods in this group successfully (BaronCohen et al. 2001; Berthoz and Hill 2005; Chalfant et al. 2007; Hedley and Young 2006; Kim et al. 2000; Kuusikko et al. 2008; Williamson et al. 2008). The current study shows that the Need Threat model of ostracism can be applied and adapted to both TD
123
J Autism Dev Disord
adolescent males and those with an ASC. It also adds to the growing body of evidence (Baron-Cohen et al. 2001; Berthoz and Hill 2005; Hedley and Young 2006; Kim et al. 2000; Kuusikko et al. 2008; Williamson et al. 2008) showing that self-report is a useful tool for gauging various types of response in the ASC population. This study provides an insight into how adolescents with ASC react to a single episode of ostracism. While this group may not report lowered mood, the need threat and anxiety data show that they are just as affected by the experience as TD adolescents. If this group lack insight into the affective consequences of an episode of rejection, this may compound the difficulties faced by this group in seeking reassurance and friendship elsewhere. However, we recognise a number of limitations of the study that might preclude generalisation of this result without further replication. The sample was relatively small, restricted to males, and the ASC group did not distinguish between participants with a diagnosis of childhood autism and that of ‘other ASC’. Future studies could extend these findings to a larger sample, and to females with an ASC. It may also be fruitful to focus on sub-categories of ASC. The effects of ostracism could also be studied in younger or older individuals with ASC to see whether adolescence is a particularly sensitive time for this group, as it is in the TD population. In addition to the selfreport measures used here, a combination of direct interviews with participants and indirect physiological measures such as skin-conductance response could be used to investigate the relationship between qualitative experience, physiological arousal, and responses on validated self-report measures. Behavioural consequences for subsequent conformity or self-regulation could also be investigated in ASC, as has been studied in typical adults (Baumeister et al. 2005; Williams et al. 2000). Finally, it would also be useful to study the relationship between short and longer term reactions to ostracism, since many adolescents with ASC experience chronic social rejection, and this is implicated in clinical depression and anxiety (Attwood 1998; Williams 2007). This study provides preliminary evidence that subtle modulation of anxiety and need threat occurs and can be reported by adolescents with ASC within an experimental context, even when the manipulation employed is relatively mild. This novel approach supports previous nonexperimental studies and anecdotal reports showing that adolescents with ASC are very sensitive to peer rejection despite their social impairments. Though it may not be feasible to eliminate exposure to ostracism for this group, one fruitful avenue for improving the quality of life might be to focus on ways of understanding and dealing with ostracism and its effects when it does occur.
123
Acknowledgments We would like to thank Odette Megnin for recruiting participants for this study, and Kipling D. Williams for his advice on using the Cyberball paradigm. This research was supported by the Royal Society and the BBSRC. SJB is a Royal Society Research Fellow and CS has a BBSRC PhD studentship.
References Allan, S., & Gilbert, P. (1995). A social comparison scale: Psychometric properties and relationship to psychopathology. Personality and Individual Differences, 19, 293–299. doi:10.1016/ 0191-8869(95)00086-L. Attwood, T. (1998). Asperger’s syndrome: A guide for parents and professionals. London: Jessica Kingsley. Baird, G., Simonoff, E., Pickles, A., Chandler, S., Loucas, T., Meldrum, D., et al. (2006). Prevalence of disorders of the autism spectrum in a population cohort of children in South Thames: The special needs and autism project (SNAP). Lancet, 368, 210– 215. doi:10.1016/S0140-6736(06)69041-7. Baron-Cohen, S., Wheelwright, S., Skinner, R., Martin, J., & Clubley, E. (2001). The autism-spectrum quotient (AQ): Evidence from Asperger syndrome/high-functioning autism, males and females, mathematicians and scientists. Journal of Autism and Developmental Disorders, 31, 5–17. doi:10.1023/A:1005653411471. Baumeister, R. F., DeWall, C. N., Ciarocco, N. J., & Twenge, J. M. (2005). Social exclusion impairs self-regulation. Journal of Personality and Social Psychology, 88(4), 589–604. doi:10.1037/ 0022-3514.88.4.589. Ben Shalom, D., Mostofsky, S. H., Hazlett, R. L., Goldberg, M. C., Landa, R. J., Faran, Y., et al. (2006). Normal physiological emotions but differences in expression of conscious feelings in children with high-functioning autism. Journal of Autism and Developmental Disorders, 36(3), 395–400. doi:10.1007/s10803006-0077-2. Berthoz, S., & Hill, E. L. (2005). The validity of using self-reports to assess emotion regulation abilities in adults with autism spectrum disorder. European Psychiatry, 20(3), 291–298. doi: 10.1016/j.eurpsy.2004.06.013. Chalfant, A. M., Rapee, R., & Carroll, L. (2007). Treating anxiety disorders in children with high functioning autism spectrum disorders: A controlled trial. Journal of Autism and Developmental Disorders, 37, 1842–1857. doi:10.1007/s10803006-0318-4. Eisenberger, N. I., Lieberman, M. D., & Williams, K. D. (2003). Does rejection hurt? An fMRI study of social exclusion. Science, 302(5643), 290–292. doi:10.1126/science.1089134. Eisenberger, N. I., Way, B. M., Taylor, S. E., Welch, W. T., & Lieberman, M. D. (2007). Understanding genetic risk for aggression: Clues from the brain’s response to social exclusion. Biological Psychiatry, 61(9), 1100–1108. doi:10.1016/j.biopsych.2006.08.007. Frith, U. (2004). Emanuel Miller lecture: Confusions and controversies about Asperger syndrome. Journal of Child Psychology and Psychiatry and Allied Disciplines, 45, 672–686. doi:10.1111/ j.1469-7610.2004.00262.x. Hedley, D., & Young, R. (2006). Social comparison processes and depressive symptoms in children and adolescents with Asperger syndrome. Autism, 10, 139–153. doi:10.1177/1362361306062020. Hill, E., Berthoz, S., & Frith, U. (2004). Brief report: Cognitive processing of own emotions in individuals with autistic spectrum disorder and in their relatives. Journal of Autism and Developmental Disorders, 34(2), 229–235. doi:10.1023/B:JADD.0000022613.41399.14. Howlin, R. (2003). Asperger syndrome in the adolescent years. In L. Holliday Willey (Ed.), Asperger syndrome in adolescence:
J Autism Dev Disord Living with the ups and downs and things in between (pp. 13– 37). London: Jessica Kingsley. Kim, J. A., Szatmari, P., Bryson, S. E., Streiner, D. L., & Wilson, F. J. (2000). The prevalence of anxiety and mood problems among children with autism and Asperger syndrome. Autism, 4(2), 117– 132. doi:10.1177/1362361300004002002. Kuusikko, S., Pollock-Wurman, R., Jussila, K., Carter, A. S., Mattila, M. L., Ebeling, H., et al. (2008). Social anxiety in highfunctioning children and adolescents with autism and Asperger syndrome. Journal of Autism and Developmental Disorders, 38(9), 1697–1709. doi:10.1007/s10803-008-0555-9. Lawson, W. (2001). Understanding and working with the spectrum of autism: An insider’s view. London: Jessica Kingsley. Lord, C., Risi, S., Lambrecht, L., Cook, E., Leventhal, B., DiLavore, P. C., et al. (2000). The autism diagnostic observation schedule generic: A standard measure of social and communication deficits associated with the spectrum of autism. Journal of Autism and Developmental Disorders, 30(3), 205–223. doi:10.1023/A:10055 92401947. Lord, C., Rutter, M., & Le, C. A. (1994). Autism diagnostic interview-revised: A revised version of a diagnostic interview for caregivers of individuals with possible pervasive developmental disorders. Journal of Autism and Developmental Disorders, 24, 659–685. doi:10.1007/BF02172145. Rosenberg, M. (1968). Self-esteem questionnaire. In Society and the adolescent self-image. Princeton, NJ: Princeton University Press. Sebastian, C., Viding, E., Williams, K.D., Charman, T., & Blakemore, S.-J.(2008, April). Affect regulation after social exclusion is still developing in adolescence. Poster presented at the annual meeting of the Cognitive Neuroscience Society, San Francisco. Silani, G., Bird, G., Brindley, R., Singer, T., Frith, C., & Frith, U. (2008). Levels of emotional awareness and autism: An fMRI study. Social Neuroscience, 3(2), 97–112. doi:10.1080/17470 910701577020. Spielberger, C. D. (1983). Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press.
Tantam, D. (2000). Psychological disorder in adolescents and adults with Asperger syndrome. Autism, 4(1), 47–62. doi:10.1177/ 1362361300004001004. Wechsler, D. (1999). The Wechsler Abbreviated Scale of IntelligenceUK. San Antonio, TX: The Psychological Corporation. Williams, K. D. (1997). Social ostracism. In R. M. Kowalski (Ed.), Aversive interpersonal behaviors (pp. 133–170). New York: Plenum. Williams, K. D. (2001). Ostracism: The power of silence. New York: Guilford Press. Williams, K. D. (2007). Ostracism. Annual Review of Psychology, 58, 425–452. doi:10.1146/annurev.psych.58.110405.085641. Williams, K. D., Cheung, C. K., & Choi, W. (2000). Cyberostracism: Effects of being ignored over the Internet. Journal of Personality and Social Psychology, 79(5), 748–762. doi:10.1037/00223514.79.5.748. Williams, K. D., & Jarvis, B. (2006). Cyberball: A program for use in research on ostracism and interpersonal acceptance. Behavior Research Methods, Instruments, & Computers, 38, 174–180. Williamson, S., Craig, J., & Slinger, R. (2008). Exploring the relationship between measures of self-esteem and psychological adjustment among adolescents with Asperger syndrome. Autism, 14(4), 391–402. doi:10.1177/1362361308091652. Zadro, L., Boland, C., & Richardson, R. (2006). How long does it last? The persistence of the effects of ostracism in the socially anxious. Journal of Experimental Social Psychology, 42, 692– 697. doi:10.1016/j.jesp.2005.10.007. Zadro, L., Williams, K. D., & Richardson, R. (2004). How low can you go? Ostracism by a computer is sufficient to lower selfreported levels of belonging, control, self-esteem, and meaningful existence. Journal of Experimental Social Psychology, 40, 560–567. doi:10.1016/j.jesp.2003.11.006.
123
