The DOI matches its Crossref record; 82 of the 82 references it cites were found in a registry or an index; 21 of the 47 sources its claims cite were checked for support.
Examining a model of anxiety in autistic adults
Open accessAbstract
Anxiety disorders commonly occur in autism. Existing studies implicate intolerance of uncertainty, alexithymia, sensory processing differences and emotion regulation difficulties as influencing factors of anxiety in autism. To date, a few studies have considered the combination of these factors within the same sample. This study used structural equation modelling to test the prediction that intolerance of uncertainty and emotion regulation constitute more direct causes of anxiety in autism that mediate the influences of sensory processing difference and alexithymia as more sequential contributing factors. Autistic (n = 86) and non-autistic adults (n = 100) completed a battery of self-report questionnaires. Only when applied to each group separately, the broad predictions of the model were confirmed for the autistic group following data-driven additions of paths between sensory processing difference and anxiety and alexithymia implying that sensory processing difference contribute indirectly as well as directly to individual differences in anxiety. For the non-autistic group, model fit could only be achieved after removing autism-related traits and sensory processing differences as predictors of anxiety. These results suggest that aetiology and expression of anxiety in autism partially overlap with what is observed in the general population except that sensory processing differences appear to play a relatively unique role in the context of autism. Lay abstract: Anxiety disorders are common in autism. Research studies have identified factors that influence anxiety in autism, such as difficulties with uncertain situations, difficulties understanding own emotions, differences in processing sensory input (related to our senses) and difficulties regulating emotions. To date, a few studies have considered the combination of these factors within the same sample. This study used structural equation modelling to test the contribution of these factors in autism. Autistic (n = 86) and non-autistic adults (n = 100) completed a battery of self-report questionnaires. Only when applied to each group separately, the broad predictions of the model were confirmed for the autistic group. The model confirmed that difficulties with uncertain situations and in regulating emotions play a central role in anxiety in autism. Difficulties understanding own emotions and differences in processing sensory input both contribute to anxiety indirectly through their respective interrelation with the other two factors (difficulties with uncertain situations and in regulating emotions). Importantly, the results imply that sensory processing differences contribute not only indirectly but also directly to individual differences in anxiety. For the non-autistic group, model fit could only be achieved after removing autism-related traits and sensory processing differences as predictors of anxiety. These results suggest that cause/development and expression of anxiety in autism partially overlap with what is observed in the general population except that sensory processing differences appear to play a relatively unique role in the context of autism. © The Author(s) 2023.
Overview
CoChat Summary
## TL;DR
This study tested a structural equation model of anxiety in 86 autistic and 100 non-autistic adults, focusing on autistic-like traits, sensory processing differences, alexithymia, intolerance of uncertainty, and emotion regulation. The model fit the autistic group only after data-driven changes, suggesting that intolerance of uncertainty and emotion regulation are central to anxiety in autism, with sensory processing differences also having a direct role.
## Key findings
- In group comparisons, autistic adults showed significantly higher levels of anxiety, emotion regulation difficulties, alexithymia, sensory processing differences, and intolerance of uncertainty than non-autistic adults.
- In the autistic group, the final modified SEM achieved adequate fit: χ2(11) = 16.196, p = .134 (Bollen–Stine p = .203), χ2/df = 1.472, RMSEA = 0.075, CFI = 0.974, SRMR = 0.065.
- In that autistic-group model, autistic-like traits predicted more alexithymia (β = 0.32), intolerance of uncertainty (β = 0.39), and sensory processing differences (β = 0.55); intolerance of uncertainty predicted more anxiety (β = 0.61), emotion regulation ratio predicted more anxiety (β = 0.34), and sensory processing differences also directly predicted more anxiety (β = 0.40).
- In the non-autistic group, adequate fit was only achieved after removing autistic-like traits and sensory processing differences; the resulting model showed alexithymia predicted anxiety directly (β = 0.49) and indirectly via intolerance of uncertainty (β = 0.27), while emotion regulation ratio was not significantly related to anxiety.
## Methods
- Cross-sectional questionnaire study comparing 86 autistic adults and 100 non-autistic adults (total sample = 186), aged 18–70 years; autistic participants were recruited from an autism clinic in Dresden and non-autistic participants from prior studies, advertisements, and institutional contacts.
- Participants completed 12 questionnaires online or on paper; the key measures used in analyses were AQ, GSQ, TAS-20, IU-18, ERQ, LSAS, and HADS anxiety.
- Analyses included group comparisons (chi-square, t-tests, Mann–Whitney U), Pearson correlations, and structural equation modelling in SPSS/AMOS with maximum-likelihood estimation and bootstrap corrections.
## Limitations
- The paper states that the study did not include autistic individuals with significant language impairments and/or learning disabilities, so findings may not generalise across the autism spectrum.
- It also states that the study relied exclusively on self-report measures, none of which were developed and validated specifically for autistic respondents, which may miss autism-specific anxiety characteristics.
- The authors state the sample size was relatively small and limited what additional variables or pathways could be tested; evident from the text, the design was cross-sectional and therefore cannot establish causality, and the non-autistic anxiety subgroup was small (13 with anxiety disorders).
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures



Verification
Claims the paper makes
18 Valid · 1 Not valid · 12 Not checked · 66 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“there is consistent evidence for a link between ALX and anxiety in both autistic as well as non-autistic individuals (Milosavljevic et al., 2016; Oakley et al., 2020).”
§ Examining a model of anxiety in autistic adults > Abstract > Lay abstract Anxiety disorders are common in autism. Research studies have identified factors that influence anxiety in autism, such as difficulties with uncertain situations, difficulties understanding own emotions, differences in processing sensory input (related to our senses) and difficulties regulating emotions. To date, a few studies have considered the combination of these factors within the same sample. This study used structural equation modelling to test the contribution of these factors in autism. Autistic (n = 86) and non-autistic adults (n = 100) completed a battery of self-report questionnaires. Only when applied to each group separately, the broad predictions of the model were confirmed for the autistic group. The model confirmed that difficulties with uncertain situations and in regulating emotions play a central role in anxiety in autism. Difficulties understanding own emotions and differences in processing sensory input both contribute to anxiety indirectly through their respective interrelation with the other two factors (difficulties with uncertain situations and in regulating emotions). Importantly, the results imply that sensory processing differences contribute not only indirectly but also directly to individual differences in anxiety. For the non-autistic group, model fit could only be achieved after removing autism-related traits and sensory processing differences as predictors of anxiety. These results suggest that cause/development and expression of anxiety in autism partially overlap with what is observed in the general population except that sensory processing differences appear to play a relatively unique role in the context of autism. Autistic individuals commonly experience co-occurring anxiety disorders, with a life time prevalence of up to 42% (Hollocks et al., 2019). A recent study suggests that young autistic adults have a risk of an anxiety diagnosis two and a half times higher compared with non-autistic persons from the general population, with autistic individuals without intellectual disability (ID) appearing to have the highest risk (Nimmo-Smith et al., 2020). In combination with aspects of the defining clinical characteristics of autism (difficulties in social interaction, communication and repetitive interests and behaviours; American Psychiatric Association, 2013), anxiety disorders negatively impact upon quality of life and are, therefore, an important target for interventions (van Steensel et al., 2012). Despite an emerging consensus about the factors that contribute to high levels of anxiety in autism, questions remain about how multiple factors interact in the development and maintenance of anxiety in autism. Existing research has identified alexithymia (ALX), emotion regulation difficulties, intolerance of uncertainty (IU) and sensory processing differences as some of the most significant contributors to anxiety in autism (Cai et al., 2018; Glod, 2017; Lidstone et al., 2014; Maisel et al., 2016; Pickard et al., 2020; South & Rodgers, 2017; Wigham et al., 2015). ALX describes a trait characterised by difficulties identifying and describing one’s own emotional states (cf. Wittchen & Hoyer, 2011). Compared to around 10% in the general population, about 50% of autistic individuals report high levels of ALX (Kinnaird et al., 2019), and there is consistent evidence for a link between ALX and anxiety in both autistic as well as non-autistic individuals (Milosavljevic et al., 2016; Oakley et al., 2020). Emotion regulation (ER) comprises all the strategies people use to modulate their emotions in order to reach personal goals (Thompson, 1994). The literature on ER distinguishes between maladaptive (e.g. suppression) and adaptive (e.g. reappraisal) strategies on the basis of their longer-term influences on mental health and wellbeing (see, for example, Gross & John, 2003). Autistic individuals tend to use adaptive strategies less often and maladaptive strategies more often (Samson et al., 2012). This pattern presents a well-established risk factor for anxiety in autism (White et al., 2014) and has also been linked to ALX (Samson et al., 2012). The concept of IU is defined as the tendency to experience uncertain situations as aversive and, thus, avoid them. The construct was first identified as an important contributing factor to the development and maintenance of generalised anxiety in the general population (Dugas et al., 1998) and a considerable literature now demonstrates that IU is significantly elevated in autism where it consistently predicts higher levels of anxiety (Boulter et al., 2014; Hwang et al., 2019; Jenkinson et al., 2020; Maisel et al., 2016). Finally, sensory processing differences (SPDs) such as hypo- or hyper-sensitivity to sensory stimuli, which are now considered a defining characteristic of autism (American Psychiatric Association, 2013), have also been linked to anxiety in a number of studies (Gillott & Standen, 2007). Although the links between each of the constructs outlined above (ALX, ER, IU and SPD) and anxiety have been well established in autism, questions remain about their relative and unique contributions as well as their causal relations. For example, South and Rodgers (2017) suggest that SPD and ALX could lead to anxiety relatively indirectly by rendering external and internal sensory experiences less predictable and comprehensible, thereby contributing to greater levels of IU. This suggestion has received support from a number of studies which have demonstrated that IU mediates the relationship between SPD and anxiety (Hwang et al., 2019; Wigham et al., 2015), as well as between ALX and anxiety (Moore et al., 2021; Ozsivadjian et al., 2021) in autism spectrum disorder (ASD). However, Pickard et al. (2020) recently found that SPD and IU could independently predict individual differences in social anxiety in autistic youth, and Neil et al. (2016) have shown that the interrelations among SPD, IU and anxiety can also be understood under alternative causal assumptions. Rather than considering SPD as a contributing cause of anxiety, the authors show that anxiety can also be considered a mediator of the relationship between IU and SPD on the basis that anxiety resulting from IU might lead to hypervigilance and, thus, hypersensitivity to the sensory environment (see Green & Ben-Sasson (2010) and Normansell-Mossa et al. (2021) for further discussion). In addition to debates concerning the interplay among SPD, IU and anxiety, findings concerning the interactions among IU, anxiety, ALX and ER are also not entirely consistent. For example, Maisel et al. (2016) found that ALX, alongside emotional acceptance (an adaptive ER strategy often associated with mindfulness-based practices), fully accounted for higher levels of anxiety in autistic adults, with IU explaining no further individual differences in anxiety over and above these factors. Morie et al. (2019) also found that ALX and ER fully mediated the relationship between core autism traits and anxiety, although IU was not measured in this study. By contrast, Pickard et al. (2020) found that ALX played no independent role in predicting social anxiety once IU and SPD were controlled for, and Cai et al. (2018) found that IU fully mediated the relationship between ER difficulties and anxiety in autistic adolescents and young adults. Finally, a recent study of autistic adults with ID by Sáez-Suanes et al. (2020) showed that both IU and ER played independent mediating roles between autism-related behavioural traits and anxiety. As the overview above illustrates, despite a clear consensus concerning some of the key contributing factors to anxiety in autism, the precise nature of their interactions remains unclear, in part because very few studies have considered all of the relevant factors in the same sample. In fact, to the best of our knowledge, only the study by Pickard et al. (2020) has done so to date. This study was specifically concerned with the influences of ALX, SPD, IU and ER on social anxiety in autism, and the analyses were somewhat limited to examining binary associations and simple mediations. Nevertheless, such studies are critical for the development and further refinement of theoretical models such as the one proposed by South and Rodgers (2017), which will ultimately be important for guiding clinical practice as well as much needed longitudinal research in this area. This study, therefore, builds on the work outlined above with the aim of testing some key predictions that emerge from the literature to date. Specifically, we tested the validity of the theoretical model outlined in Figure 1, which represents an extension of the model proposed by South and Rodgers (2017) based on the combination of findings outlined above. Specifically, this model predicts that IU and ER constitute the most proximal causes of anxiety in autism, with SPD and ALX contributing more indirectly as sources of greater IU and ER.
What the sources say
- Supportedcites Milosavljevic et al., 2016 source
The section ‘The association of co-occurring alexithymia with emotional difficulties’ reports higher trait anxiety in adolescents with ASD and alexithymia than in those with ASD without alexithymia, and says this is in line with findings in non-ASD samples. The paper thus supports a link in both groups; page numbers were not supplied in our copy. (LLM, conf=0.93)
- Not matchedcites Oakley et al., 2020
Couldn’t be matched to a paper in Crossref, so it can’t be checked.
boostedstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
Every source the paper cites for this claim was checked and supports it.
In the paper
“In fact, to the best of our knowledge, only the study by Pickard et al. (2020) has done so to date.”
§ Examining a model of anxiety in autistic adults > Abstract > Lay abstract Anxiety disorders are common in autism. Research studies have identified factors that influence anxiety in autism, such as difficulties with uncertain situations, difficulties understanding own emotions, differences in processing sensory input (related to our senses) and difficulties regulating emotions. To date, a few studies have considered the combination of these factors within the same sample. This study used structural equation modelling to test the contribution of these factors in autism. Autistic (n = 86) and non-autistic adults (n = 100) completed a battery of self-report questionnaires. Only when applied to each group separately, the broad predictions of the model were confirmed for the autistic group. The model confirmed that difficulties with uncertain situations and in regulating emotions play a central role in anxiety in autism. Difficulties understanding own emotions and differences in processing sensory input both contribute to anxiety indirectly through their respective interrelation with the other two factors (difficulties with uncertain situations and in regulating emotions). Importantly, the results imply that sensory processing differences contribute not only indirectly but also directly to individual differences in anxiety. For the non-autistic group, model fit could only be achieved after removing autism-related traits and sensory processing differences as predictors of anxiety. These results suggest that cause/development and expression of anxiety in autism partially overlap with what is observed in the general population except that sensory processing differences appear to play a relatively unique role in the context of autism. Autistic individuals commonly experience co-occurring anxiety disorders, with a life time prevalence of up to 42% (Hollocks et al., 2019). A recent study suggests that young autistic adults have a risk of an anxiety diagnosis two and a half times higher compared with non-autistic persons from the general population, with autistic individuals without intellectual disability (ID) appearing to have the highest risk (Nimmo-Smith et al., 2020). In combination with aspects of the defining clinical characteristics of autism (difficulties in social interaction, communication and repetitive interests and behaviours; American Psychiatric Association, 2013), anxiety disorders negatively impact upon quality of life and are, therefore, an important target for interventions (van Steensel et al., 2012). Despite an emerging consensus about the factors that contribute to high levels of anxiety in autism, questions remain about how multiple factors interact in the development and maintenance of anxiety in autism. Existing research has identified alexithymia (ALX), emotion regulation difficulties, intolerance of uncertainty (IU) and sensory processing differences as some of the most significant contributors to anxiety in autism (Cai et al., 2018; Glod, 2017; Lidstone et al., 2014; Maisel et al., 2016; Pickard et al., 2020; South & Rodgers, 2017; Wigham et al., 2015). ALX describes a trait characterised by difficulties identifying and describing one’s own emotional states (cf. Wittchen & Hoyer, 2011). Compared to around 10% in the general population, about 50% of autistic individuals report high levels of ALX (Kinnaird et al., 2019), and there is consistent evidence for a link between ALX and anxiety in both autistic as well as non-autistic individuals (Milosavljevic et al., 2016; Oakley et al., 2020). Emotion regulation (ER) comprises all the strategies people use to modulate their emotions in order to reach personal goals (Thompson, 1994). The literature on ER distinguishes between maladaptive (e.g. suppression) and adaptive (e.g. reappraisal) strategies on the basis of their longer-term influences on mental health and wellbeing (see, for example, Gross & John, 2003). Autistic individuals tend to use adaptive strategies less often and maladaptive strategies more often (Samson et al., 2012). This pattern presents a well-established risk factor for anxiety in autism (White et al., 2014) and has also been linked to ALX (Samson et al., 2012). The concept of IU is defined as the tendency to experience uncertain situations as aversive and, thus, avoid them. The construct was first identified as an important contributing factor to the development and maintenance of generalised anxiety in the general population (Dugas et al., 1998) and a considerable literature now demonstrates that IU is significantly elevated in autism where it consistently predicts higher levels of anxiety (Boulter et al., 2014; Hwang et al., 2019; Jenkinson et al., 2020; Maisel et al., 2016). Finally, sensory processing differences (SPDs) such as hypo- or hyper-sensitivity to sensory stimuli, which are now considered a defining characteristic of autism (American Psychiatric Association, 2013), have also been linked to anxiety in a number of studies (Gillott & Standen, 2007). Although the links between each of the constructs outlined above (ALX, ER, IU and SPD) and anxiety have been well established in autism, questions remain about their relative and unique contributions as well as their causal relations. For example, South and Rodgers (2017) suggest that SPD and ALX could lead to anxiety relatively indirectly by rendering external and internal sensory experiences less predictable and comprehensible, thereby contributing to greater levels of IU. This suggestion has received support from a number of studies which have demonstrated that IU mediates the relationship between SPD and anxiety (Hwang et al., 2019; Wigham et al., 2015), as well as between ALX and anxiety (Moore et al., 2021; Ozsivadjian et al., 2021) in autism spectrum disorder (ASD). However, Pickard et al. (2020) recently found that SPD and IU could independently predict individual differences in social anxiety in autistic youth, and Neil et al. (2016) have shown that the interrelations among SPD, IU and anxiety can also be understood under alternative causal assumptions. Rather than considering SPD as a contributing cause of anxiety, the authors show that anxiety can also be considered a mediator of the relationship between IU and SPD on the basis that anxiety resulting from IU might lead to hypervigilance and, thus, hypersensitivity to the sensory environment (see Green & Ben-Sasson (2010) and Normansell-Mossa et al. (2021) for further discussion). In addition to debates concerning the interplay among SPD, IU and anxiety, findings concerning the interactions among IU, anxiety, ALX and ER are also not entirely consistent. For example, Maisel et al. (2016) found that ALX, alongside emotional acceptance (an adaptive ER strategy often associated with mindfulness-based practices), fully accounted for higher levels of anxiety in autistic adults, with IU explaining no further individual differences in anxiety over and above these factors. Morie et al. (2019) also found that ALX and ER fully mediated the relationship between core autism traits and anxiety, although IU was not measured in this study. By contrast, Pickard et al. (2020) found that ALX played no independent role in predicting social anxiety once IU and SPD were controlled for, and Cai et al. (2018) found that IU fully mediated the relationship between ER difficulties and anxiety in autistic adolescents and young adults. Finally, a recent study of autistic adults with ID by Sáez-Suanes et al. (2020) showed that both IU and ER played independent mediating roles between autism-related behavioural traits and anxiety. As the overview above illustrates, despite a clear consensus concerning some of the key contributing factors to anxiety in autism, the precise nature of their interactions remains unclear, in part because very few studies have considered all of the relevant factors in the same sample. In fact, to the best of our knowledge, only the study by Pickard et al. (2020) has done so to date. This study was specifically concerned with the influences of ALX, SPD, IU and ER on social anxiety in autism, and the analyses were somewhat limited to examining binary associations and simple mediations. Nevertheless, such studies are critical for the development and further refinement of theoretical models such as the one proposed by South and Rodgers (2017), which will ultimately be important for guiding clinical practice as well as much needed longitudinal research in this area. This study, therefore, builds on the work outlined above with the aim of testing some key predictions that emerge from the literature to date. Specifically, we tested the validity of the theoretical model outlined in Figure 1, which represents an extension of the model proposed by South and Rodgers (2017) based on the combination of findings outlined above. Specifically, this model predicts that IU and ER constitute the most proximal causes of anxiety in autism, with SPD and ALX contributing more indirectly as sources of greater IU and ER.
What the sources say
- Supportedcites Pickard et al. (2020) source
The Study aims section says Pickard et al. investigated cognitive, emotional and sensory correlates of social anxiety in autistic and neurotypical adolescents. The Method (Participants) section identifies the adolescents studied, supporting that these factors were considered in the same sample. The citing authors qualify the claim that it was the only such study as their own assessment, not a finding attributed to Pickard et al. (LLM, conf=0.87)
hedged- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“the findings are in line with growing evidence which suggests that interventions for anxiety in autism that target ER strategy use, ALX and IU generally proof to be effective.”
§ Examining a model of anxiety in autistic adults > Discussion This study investigated the role of SPD, ALX, IU and ER in the relationship between autistic-like traits and anxiety in autistic and non-autistic adults with average intelligence. In line with existing evidence (Hwang et al., 2019; Kinnaird et al., 2019; Lehnhardt et al., 2011; Samson et al., 2012; Takayama et al., 2014), participants in the autistic group scored significantly higher on measures for autistic-like traits, ALX, IU, SPD and anxiety. The autistic group also demonstrated a more maladaptive pattern of ER skills that was characterised primarily by significantly lower reported use of reappraisal relative to suppression strategies compared to the non-autistic group. No significant difference was observed between groups in the use of suppression strategies per se, which some previous studies suggest are sometimes used more frequently in autism (Samson et al., 2012). As Cai et al. (2018) suggest, however, the relative use of adaptive versus maladaptive strategies often has greater implications for mental health than the absolute degree to which either strategy might be used, which was supported by the current findings in terms of the correlation between an ER ratio score and measures of anxiety. More broadly, the pattern of correlations among measures also confirmed previous observations of significant interrelations among measures of autistic-like traits, SPD, ALX, IU and ER difficulties that this study sought to shed further light on. Specifically, our primary aim was to test a specific model of these interrelations, which predicts that IU and ER difficulties constitute relatively proximal causes of anxiety in autism that mediate the impact of ALX and SPD. When testing the predicted model against the combined data across both groups, no adequate fit could be achieved, even when adjusting model paths based on data-driven principles. When the model was tested against the autistic group data alone, however, adequate fit was achieved after some data-driven adjustments. The resulting model confirmed that IU, ER and SPD play a relatively central role in anxiety in autism. In addition, SPD and ALX contribute to anxiety through respective associations with IU and ER difficulties. This pattern fits well with an emerging consensus view about the role of IU and ER in anxiety in autism (Boulter et al., 2014; South & Rodgers, 2017) and further suggests that each of these factors can contribute independently to experiences of anxiety. The findings also support ongoing developments in clinical practice, where interventions focussed on building tolerance of uncertainties (Rodgers et al., 2017, 2018) and ER skills (e.g. Conner et al., 2019; Weiss et al., 2018) are often effective in the treatment of anxiety in autism. The required adjustments to the model to achieve adequate fit have important implications for understanding the role of SPD and ALX. Based on a model by South and Rodgers (2017), and subsequent findings (Moore et al., 2021; Ozsivadjian et al., 2021), we predicted that ALX would contribute to anxiety not only by increasing ER difficulties but also by rendering internal emotional states confusing and, thereby, contributing to IU. Our model did not support this prediction, suggesting instead that ALX contributes relatively independently to anxiety only via associations with ER difficulties. By contrast, SPD did not only impact upon anxiety through indirect influences on IU (Hwang et al., 2019; Wigham et al., 2015) but also through an independent direct pathway, which supports findings by Pickard et al. (2020). Interestingly, the study by Pickard et al. (2020) was concerned with predictors of social anxiety, whereas the studies by Hwang et al. (2019) and Wigham et al. (2015) examined predictors of more generalised anxiety. In this study, we derived a latent anxiety factor from measures of both generalised and social anxiety, and it is, therefore, possible that SPD plays a relatively greater role in social rather than more generalised forms of anxiety in autism. This suggestion aligns with evidence, which suggests that autistic individuals often experience social interactions as difficult because of associated sensory processing demands (Hilton et al., 2010). To examine this further is a task for future research. A further model adjustment that was required to achieve adequate fit and that has interesting implications for the literature is the addition of a pathway from SPD to ALX. Specifically, growing literature suggests that ALX is closely associated with aspects of interoception, which describes the processes involved in sensing internal physiological states such as changes in heart rate or arousal (Brewer et al., 2016). In addition, evidence suggests that autism may be associated with certain interoceptive difficulties (DuBois et al., 2016) that are closely linked to ALX (Shah et al., 2016), anxiety (Garfinkel et al., 2016; Palser et al., 2018) and SPD (Proff et al., 2021). Within the context of this wider literature, the findings of this study may, therefore, suggest that SPD contribute to ALX in autism by altering relevant underlying interoceptive processes and that ALX in turn contributes to anxiety by compromising adaptive patterns of ER. Clinically, the findings suggest that interventions targeting SPD could potentially reap benefits for anxiety in autism. Sensory-based interventions are already widely used, but the evidence base for their effectiveness is currently relatively weak, owing partly to methodological weaknesses in relevant trials and a lack of consensus regarding conceptualisations of SPD in autism as well as the desired outcomes of interventions (see, for example, Case-Smith et al., 2015). The final model adjustment that was necessary for the data of the autistic group was the introduction of a direct path from autism-related traits to IU, which suggests that IU is not exclusively linked to SPD in autism. Conceptually, it is linked to insistence on sameness and a recent study reported increased anxiety rates in autistic children that were preceded by elevated levels of insistence on sameness behaviour 1–2 years earlier (Baribeau et al., 2022). In addition, previous studies have shown that factors such as cognitive flexibility may also play an important role in the relationship between the core clinical characteristics of autism and IU (Ozsivadjian et al., 2021), and other studies point to a role of Theory of Mind (ToM) difficulties in the expression of anxiety in autism (Lei & Ventola, 2018). Further cognitive differences that have recently been suggested to play a role in the development of anxiety in autism are reduced predictive processing and ‘black and white thinking’ (Stark et al., 2021). Thus, the model tested in this study is unlikely to represent all of the factors that contribute to increased levels of anxiety in autism. Turning now to the finding that the predicted model could not be fit to the data across both groups combined, it is interesting that no adequate model fit could be achieved when the data were considered from the non-autistic group alone. In the context of the model that was ultimately derived for the autistic group, this may suggest that the mechanisms involved in anxiety in autism are, at least to some extent, distinct from those involved in anxiety in the general population. The finding that anxiety can often manifest differently in autism than in the general population supports this suggestion. For example, following detailed clinical interviews with parents of autistic children, Kerns et al. (2014) found that 15% of parents described patterns of anxiety in their children that would not fall under traditional definitions of anxiety, and a further 30% described combinations of such ‘atypical’ presentations of anxiety together with more typical presentations. Often, the autism-specific expression of anxiety reported in this study was related to worries about changes in routines (i.e. IU) or fears of particular sensory stimuli (i.e. SPD), which are not typically reported as sources of anxiety in the general population. In addition, factors such as cognitive inflexibility and ToM difficulties may also play relatively unique roles in the expression of anxiety in autism as compared to the general population (Lei & Ventola, 2018; Ozsivadjian et al., 2021). In the context of this wider literature, it is, therefore, perhaps not surprising to observe distinct patterns of interrelationships among relevant constructs. Based on the considerations above, we examined the data in the non-autistic group after removing autism-related traits and SPD as factors in the model, focussing instead on only those risk factors (IU, ALX and ER) that are well established to play a role in anxiety in the general population (Aldao et al., 2016; Carleton et al., 2012; Kranzler et al., 2016). With only these factors in the model, adequate fit was achieved, demonstrating that ALX and IU exert independent direct influences on anxiety. Although ALX was also associated with ER in this model, ER did not significantly contribute to anxiety either directly or indirectly. This observation is somewhat unexpected given the consistent evidence that links ER to anxiety and other mental health difficulties (e.g. Aldao et al., 2016). However, to the best of our knowledge, no previous studies have examined the combined contributions of ALX, IU and ER to individual differences in anxiety in the general population. In such general population samples, ALX and IU may play a relatively more important role in anxiety than ER. In this context, it is also relevant to note that the distribution of the ER ratio score, as well as both the HADS and LSAS anxiety scores, was skewed in our non-autistic group. The implications of the observations for models of clinical levels of anxiety are, therefore, somewhat unclear, and future studies should attempt to compare predictors of anxiety in autistic samples to the predictors of similar levels of anxiety in non-autistic anxious individuals. When considering our results, one should consider the transdiagnostic nature of our sample. Including individuals with co-occurring disorders in the autistic group and anxiety disorders in the non-autistic group will potentially increase the generalisability of our findings. In addition, for the same reason, we included a sample with a large age range. Therefore, age may have impacted on the results. However, co-occurring neurological conditions whose risk increases with age are very unlikely in the current sample since participants were screened for those prior participation and only 13 participants in the total sample were 60 years of age and above. Before drawing this discussion to a close, it is important to acknowledge a number of potential limitations in this study. First, the study did not include autistic individuals with significant language impairments and/or learning disabilities and, therefore, the findings may not generalise across the autism spectrum, particularly as some evidence suggests that higher intellectual functioning may be associated with greater levels of anxiety in autism (e.g. Nimmo-Smith et al., 2020). Having said that, a recent study of autistic adults with ID found that IU and ER difficulties constituted independent mediators of the relationship between autism-related behaviours and anxiety (Sáez-Suanes et al., 2020), suggesting that the model presented in this article may generalise across the autism spectrum. Further studies involving representative samples are necessary, however, to further examine this issue. A second limitation of the study is that it relied exclusively on self-report measures to capture all constructs of interest, none of which were developed and validated specifically for autistic respondents. Although the internal consistencies of all measures used in this study were very good in both groups within the current sample, and all measures have been used in previous studies of anxiety in autism, there is a potential danger of missing autism-specific characteristics of anxiety (see, for example, Kerns et al., 2014). Recently, the first autism-specific self-report questionnaire for anxiety has been developed and validated for adult participants (Rodgers et al., 2020), which will help to address this concern in future research. In addition, however, it will also be important to develop more behavioural measures of key constructs of interest in order to bridge our understanding of the relationship between internalising and externalising symptoms of anxiety. Such a comprehensive measurement of anxiety will also be important for translating theoretical models such as the one described here into relevant interventions effectively. A third limitation concerns the sample size of this study, which may be considered as relatively small. Our sample size calculations were guided by previous literature. Maisel et al. (2016) included 76 autistic adults and 76 non-autistic individuals. In addition, we followed the general recommendation of 10 participants for every parameter (e.g. Hoe, 2008). Despite our greatest efforts in recruitment, we did not achieve this fully for the autistic group. Therefore, with our data, it was not possible to examine further variables or add additional pathways. The aim of this study was to establish and test a basic model of the development of anxiety in autism. One of the next steps could be to expand the current model to include returning pathways from anxiety to assess the impact of anxiety on autism-related symptoms. As mentioned in the introduction, Neil et al. (2016) have shown that anxiety can also be considered a mediator of the relationship between IU and SPD. To examine this and other additional pathways is a task for future research. Furthermore, the question arises whether different models of anxiety can be derived when considering different types of anxiety separately such as social and general anxiety. We chose a rather broad construct of anxiety including social and general anxiety in our latent variable for our basic model, which was done according to the previous literature such as Maisel et al. (2016). Future research testing larger samples could analyse and compare separate models for different anxiety types. A final limitation is that we implemented a variable-centred analysis in this study. The above reported associations may not apply to subgroups or at the individual level, that is, in some individuals IU or SPD may be the most relevant factor in driving their anxiety. However, the investigation of anxiety models in autistic subgroups could be a topic of future studies. In terms of the possible clinical implications of the current work, the findings are in line with growing evidence which suggests that interventions for anxiety in autism that target ER strategy use, ALX and IU generally proof to be effective. For example, recent reviews of cognitive behavioural therapies (CBTs; Spain et al., 2015) and mindfulness-based therapies (Cachia et al., 2016) suggest that they are effective in alleviating anxiety in autism. In addition, a newly developed intervention that utilises CBT and mindfulness-based practices to target IU specifically is yielding promising results as a means of managing anxiety in autism (Rodgers et al., 2018). Such work demonstrates the importance of tailoring interventions specifically for the autistic community, and the findings of this study may help by highlighting the importance that SPD may also play in this context. SPDs have for long been a target of interventions but rarely with a specific focus on anxiety as outcomes. For example, sensory integration therapy is often used to target broad functional skills and adaptive behaviours rather than anxiety and the evidence for its effectiveness is rather mixed (see Case-Smith et al., 2015; Lang et al., 2012). However, there are other types of interventions that may be fruitfully adapted and/or integrated with approaches such as CBT or mindfulness-based interventions to target anxiety in autism, such as systematic desensitisation (Koegel et al., 2004). In general, interventions targeting several factors influencing anxiety in combination will potentially be more beneficial rather than targeting each factor in isolation.
What the sources say
- Not checked — no full text of the sourcecites Spain et al., 2015 source
- Not checked — no full text of the sourcecites Cachia et al., 2016 source
- Partly supportedcites Rodgers et al., 2018 sourceExperimental — accuracy not yet measured
The Discussion reports promising changes in a four-person intervention targeting intolerance of uncertainty (IU), but says efficacy requires formal evaluation. The Introduction describes evidence for autism anxiety interventions as variable, and the paper does not establish that interventions targeting emotion-regulation strategy use or alexithymia (ALX) are generally effective. Page numbers were not available in our copy. (LLM, conf=0.87)
hedgedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“However, following stepwise data-driven additions of pathways from SPD directly to anxiety and ALX, and directly from AQ to IU, and the removal of our predicted pathway from ALX to IU, adequate model fit was achieved (compare Figures 1 and 2 for changes to the model) (χ2(11) = 16.196, p = .134 (Bollen–Stine p = .203), χ2/df = 1.472; RMSEA = 0.075, PCLOSE = 0.270; CFI = 0.974; SRMR = 0.065).”
§ Examining a model of anxiety in autistic adults > Results > Structural equation modelling
plainstates an association- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“ER ratio significantly predicted more anxiety (β = 0.34, p < .001).”
§ Examining a model of anxiety in autistic adults > Results > Structural equation modelling
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
33 more claims
Create a free account to see all 97 claims and every source verdict.
- Probably not a substantial claim. 31
- It may state two separate claims as one. 35
- The quote only partly states this claim. 1
Statistics recomputed
Every reported test result, recomputed from its test statistic and degrees of freedom. A result shows as not adding up only once an independent check has confirmed it.
- Consistentt(184) = 0.19reported p = .849 · recomputed p = .850
“t(184.00) = 0.19, p = .849”
- Consistentχ²(1) = 1.65reported p = .2 · recomputed p = .199
“χ2(1) = 1.65, p = .20”
- Consistentχ²(12) = 103.376reported p = 0 · recomputed p < .001
“χ2(12) = 103.376, p = .000”
- Consistentχ²(12) = 49.031reported p = 0 · recomputed p < .001
“χ2(12) = 49.031, p = .000”
- Consistentχ²(11) = 16.196reported p = .134 · recomputed p = .134
“χ2(11) = 16.196, p = .134”
2 more recomputed tests
Create a free account to see every test the Stats check recomputed.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Self-check results appear here once an independent check has confirmed them.
Citations
Cites
Diagnostic and Statistical Manual of Mental Disorders
American Psychiatric Association (2013) · doi:10.1176/appi.books.9780890425596 · Cited by 116,369
Hu, L., & Bentler, P. M. (1999) · Structural Equation Modeling A Multidisciplinary Journal · doi:10.1080/10705519909540118 · Cited by 108,788 · cited 1×
“…model fit is regarded acceptable when χ 2 / df ⩽ 2.5, comparative fit index (CFI) ⩾ 0.90 (Homburg & Baumgartner, 1995), root mean square error of approximation (RMSEA) ⩽ 0.08 (Brown & Cudeck, 1993), PCLOSE > 0.05 and standardised root mean square residual (SRMR) ⩽ 0.08 (Hu & Bentler, 1999).”
BackgroundInfluentialnull (2016) · Philosophy Study · doi:10.17265/2159-5313/2016.09.003 · Cited by 28,317
Testing Structural Equation Models.
Mueller, C. W., Bollen, K. A., & Long, J. S. (1994) · Contemporary Sociology · doi:10.2307/2074955 · Cited by 14,120
Gross, J. J., & John, O. P. (2003) · Journal of Personality and Social Psychology · doi:10.1037/0022-3514.85.2.348 · Cited by 11,929 · cited 1×
“The literature on ER distinguishes between maladaptive (e.g. suppression) and adaptive (e.g. reappraisal) strategies on the basis of their longer-term influences on mental health and wellbeing (see, for example, Gross & John, 2003).”
BackgroundMeasures of multivariate skewness and kurtosis with applications
MARDIA, K. V. (1970) · Biometrika · doi:10.1093/biomet/57.3.519 · Cited by 4,792 · cited 1×
“Multivariate normality, assessed with the critical ratio to Mardia’s coefficient (Mardia, 1970), was violated, c.r. > 2.57.”
MethodInfluentialEMOTION REGULATION: A THEME IN SEARCH OF DEFINITION
Thompson, R. A. (2008) · Monographs of the Society for Research in Child Development · doi:10.1111/j.1540-5834.1994.tb01276.x · Cited by 2,963 · cited 1×
“Emotion regulation (ER) comprises all the strategies people use to modulate their emotions in order to reach personal goals (Thompson, 1994).”
BackgroundFearing the unknown: A short version of the Intolerance of Uncertainty Scale
Carleton, R. N., Norton, M. A. P. J., & Asmundson, G. J. G. (2007) · Journal of Anxiety Disorders · doi:10.1016/j.janxdis.2006.03.014 · Cited by 2,288 · cited 1×
“Because of varying factor solutions that were found in versions in different languages (Buhr & Dugas, 2002; Carleton et al., 2007; Freeston et al., 1994), only the total score was used.”
BackgroundFreeston, M. H., Rhéaume, J., Letarte, H., Dugas, M. J., & Ladouceur, R. (1994) · Personality and Individual Differences · doi:10.1016/0191-8869(94)90048-5 · Cited by 1,731 · cited 1×
“Because of varying factor solutions that were found in versions in different languages (Buhr & Dugas, 2002; Carleton et al., 2007; Freeston et al., 1994), only the total score was used.”
BackgroundThe intolerance of uncertainty scale: psychometric properties of the English version
Buhr, K., & Dugas, M. J. (2002) · Behaviour Research and Therapy · doi:10.1016/s0005-7967(01)00092-4 · Cited by 1,391 · cited 1×
“Because of varying factor solutions that were found in versions in different languages (Buhr & Dugas, 2002; Carleton et al., 2007; Freeston et al., 1994), only the total score was used.”
Background
Research works printed but not linked: 15
- Bach M., 1996, Psychotherapie, Psychosomatik, Medizinische Psychologie, 46, 23
- Bölte S., 2006, Diagnostisches Interview für Autismus-revidiert. [Autism Diagnostic Interview Revised] ADI-R
- Bühner M., 2006, Einführung in die Test-und Fragebogenkonstruktion (2., aktual. Und erw. Aufl.) [Introduction to test and questionnaire design]
- Glod M. (2017). Sensory profiles in autism spectrum disorder and Williams syndrome [PhD Thesis, Newcastle University]. http://theses.ncl.ac.uk/jspui/handle/10443/3897
- Herrmann-Lingen C., 2010, Hospital Anxiety and Depression Scale – Deutsche Version
- Petermann F., 2007, Hamburg-Wechsler-Intelligenztest für Kinder IV (HAWIK-IV) [The Wechsler Intelligence Scale for Children - Fourth Version]
- Poustka L., 2015, Diagnostische Beobachtungsskala für Autistische Störungen–2 (ADOS-2) [Autism Diagnostic Observation Schedule, Second Edition (ADOS-2)]
- Rudolf M., 2020, Multivariate Verfahren eine praxisorientierte Einführung mit Anwendungsbeispielen
- Rühl D., 2004, Diagnostische Beobachtungsskala für Autistische Störungen: ADOS [Autism Diagnostic Observation Schedule (ADOS)]
- Stangier U., 2003, Die Liebowitz Soziale Angst-Skala (LSAS) [Liebowitz Social Anxiety Scale (LSAS)]
- Tewes U., 1991, Hamburg-Wechsler-Intelligenztest für Erwachsene, revision 1991 [Wechsler Adult Intelligence Scale, revision]
- Tewes U., 1999, HAWIK-III; Manual
- Von Aster M., 2006, Wechsler-Intelligenztest für Erwachsene WIE. Deutschsprachige Bearbeitung und Adaptation des WAIS-III von David Wechsler [Wechsler Adult Intelligence Scale (WIE). German edition and adaptation of the WAIS-III from David Wechsler]
- Weiß R. H., 1997, Grundintelligenztest Skala 1-CFT 1 [Basic intelligence scale 1 - CFT]
- Zeisel A., BMC Psychiatry
Cited by
Kim, S.-I., Jang, S.-Y., Kim, T., Kim, B., Jeong, D., Noh, T., Jeong, M., Hall, K., Kim, M., Yoo, H. J., et al. (2023) · JMIR Formative Research · doi:10.2196/52157 · Cited by 34
“Specifically, workplace environments can amplify anxiety in individuals with autism, such as sensory overload, communication barriers, and unplanned interactions [4-8] as these environments are often designed with neurotypical expectations in mind.”
BackgroundThe differential diagnosis of autism spectrum disorder in adults
Carroll, H. M., Thom, R. P., & McDougle, C. J. (2025) · Expert Review of Neurotherapeutics · doi:10.1080/14737175.2025.2490533 · Cited by 22
“Social observations, effortful social responses in order to assimilate, and holding back from stimming, are all examples of behavioral strategies that adults with ASD might use to ‘camou-flage’ their autistic traits [101,102,107].”
Recio, P., Pozo, P., García‐López, C., & Sarriá, E. (2024) · Brain Sciences · doi:10.3390/brainsci14111088 · Cited by 14
Faustmann, L. L., & Altgassen, M. (2024) · Autism Research · doi:10.1002/aur.3165 · Cited by 9
Validation of the German Glasgow Sensory Questionnaire in autistic adults
Jakob, I. M., Roessner, V., & Ring, M. (2025) · BMC Psychiatry · doi:10.1186/s12888-025-06504-0 · Cited by 5
Self-citation
Metadata
- Authors
- Riedelbauch, S., Gaigg, S. B., Thiel, T., Roessner, V., & Ring, M.
- Affiliations
- Technische Universität Dresden (DE)
- City, University of London (GB)
- Corresponding author
- Ring, Melanie
- Venue
- Autism
- Publisher
- SAGE Publications
- Published
- 2023-06-16
- Volume
- 28
- Issue
- 3
- Pages
- 565-579
- ISSN
- 1362-3613 (ISSN-L)
- 1461-7005
- PMC
- PMC10913331
- OpenAlex
- W4381046726
- Cited by
- 19
- References
- 82
- Citations by year
- 2026: 5 · 2025: 9 · 2024: 3 · 2023: 2
- FWCI
- 2.84
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- Autism (publishedVersion, cc-by)
- PubMed Central (publishedVersion, cc-by)
- PubMed (publishedVersion)
- City Research Online (City University London) (acceptedVersion)
- Qucosa (Saxon State and University Library Dresden) (submittedVersion, cc-by)
- Funders
- Stiftung Irene
- Topics
- Life Sciences › Neuroscience › Cognitive Neuroscience › Autism Spectrum Disorder Research
- Social Sciences › Psychology › Experimental and Cognitive Psychology › Anxiety, Depression, Psychometrics, Treatment, Cognitive Processes
- Health Sciences › Medicine › Psychiatry and Mental health › Child Nutrition and Feeding Issues
- Keywords
- anxiety disorders, intolerance of uncertainty, alexithymia, sensory processing differences, emotion regulation, mediation model, autistic adults, individual differences in anxiety
- MeSH terms
- Autism Spectrum Disorder, Adult, Anxiety, Anxiety Disorders, Autistic Disorder, Emotions, Humans
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-11, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-11 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
How to cite
Cite CoChat for the checks on this page, and the paper for its own findings.
- This page
- CoChat. (2026). Citation check: Examining a model of anxiety in autistic adults. https://app.cochat.ai/papers/10.1177/13623613231177777
- The paper
- Riedelbauch, S., Gaigg, S. B., Thiel, T., Roessner, V., & Ring, M. (2024). Examining a model of anxiety in autistic adults. Autism. https://doi.org/10.1177/13623613231177777
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.