A nomological network for misophonia in two German samples using the S-Five model for misophonia

Nico Remmert; Antonia Jebens; Rebecca Gruzman; Jane Gregory; Silia Vitoratou

doi:10.3389/fpsyg.2022.902807

. 2022 Dec 21;13:902807. doi: 10.3389/fpsyg.2022.902807

A nomological network for misophonia in two German samples using the S-Five model for misophonia

Nico Remmert ^1,^*, Antonia Jebens ², Rebecca Gruzman ³, Jane Gregory ⁴, Silia Vitoratou ²

PMCID: PMC9811822 PMID: 36619047

Abstract

The Selective Sound Sensitivity Syndrome Scale (S-Five) is a contemporary and multidimensional self-report instrument measuring different aspects of misophonia. The five-factor scale consists of 25 items measuring the severity of the misophonic experience. The items capture misophonia in relation to internalising and externalising appraisals, perceived threat, aggressive behavior (outbursts), and adverse impact on individuals’ lives. It is complemented by a trigger checklist (S-Five-T), measuring the emotional nature and intensity of reactions to sensory triggers. In this work, we administered the S-Five in two German samples with a majority of individuals with significant misophonia. The S-Five and the supplementary S-Five-T were both translated into German using a rigorous translation procedure (i.e., TRAPD) and were separately tested in large German community samples. Psychometric analyses included the evaluation of the factor structure, measurement invariance with respect to age and gender, reliability (internal consistency and stability over time), and an extensive examination of the construct validity in a proposed nomological network. The nomological network we explore in this work consists of several constructs including different misophonic manifestations, anger and aggression, disgust propensity, anxiety sensitivity, depression, obsessive–compulsive traits, and functional impairment in different life domains. Results indicate evidence in line with the nomological network as demonstrated by strong correlations between the S-Five dimensions and convergent measures. All S-Five dimensions strongly correlated with overall misophonic symptoms (r ≥ 0.53). Internalising appraisals were highly associated with insight into excessive or disproportionate reactions to sounds (r ≥ 0.59), externalising appraisals with anger and irritability (r ≥ 0.46), threat with trait anxiety and dysregulation facets (r ≥ 0.62), aggressive behavior (outbursts) with anger and behavioral dysregulation (r ≥ 0.70), and impact with distress and functional impairment (r ≥ 0.64). The results demonstrate that the S-Five has a robust five-factor structure and allows to draw reliable and valid conclusions about misophonic experiences in German samples. The proposed nomological network gives an initial insight into the nature of misophonia and provides a formalized fundament to develop and test further hypotheses about misophonia in a more sophisticated and symptom-oriented way.

Keywords: misophonia, nomological network, psychometrics, selective sound sensitivity syndrome scale S-Five, construct validation

Introduction

Misophonia is a disorder related to decreased tolerance to certain sounds (Swedo et al., 2022), most commonly sounds related to eating, nose and throat sounds, and repetitive environmental sounds (Vitoratou et al., 2021a). Individuals with misophonia can experience profound distress and functional impairment from their emotional, physical and behavioral responses to these sounds (e.g., Jastreboff and Jastreboff, 2014; Brout et al., 2018; Jager et al., 2020).

For the assessment of the multidimensional experience of misophonia, Vitoratou et al. (2021b) developed a five-factor model scale known as the S-Five-E (Selective Sound Sensitivity Syndrome Scale - Experiences). The S-Five was constructed based on the responses and feedback of a large sample of English-speaking self-identified misophonic individuals, over four sampling waves. The resulting scale consists of 25-items corresponding to five dimensions: (1) internalising appraisals attributing blame for reactions to oneself (e.g., believing to be an unlikable or angry person), (2) externalising appraisals blaming other people (e.g., believing others to be rude and inconsiderate), (3) perceived emotional threat (e.g., feeling distress, trapped and helpless), (4) having or fearing having verbal or physical outbursts, and (5) the impact of misophonia on the ability to do things they would like to do. Along with the main scale, the S-Five has a supplementary trigger checklist (S-Five-T), which captures the emotional nature and intensity of the responses to sounds (Vitoratou et al., 2021b, 2022a). The format of the S-Five-T allows the researcher or clinician to modify the trigger sounds list and the response types, in line with changes in the growing literature on the field and individual presentations of the disorder. The five-factor model of the S-Five has been replicated in a large sample representative of the UK population (Vitoratou et al., 2022a). Excellent psychometric properties have been shown for the scale in English (Vitoratou et al., 2021b) and Mandarin (Vitoratou et al., 2022b), with cross-cultural replication of the five-factor model. A German translation, however, is still pending.

To our knowledge, the only genuine German questionnaire measuring misophonic symptoms is the Berlin Misophonia Questionnaire Revised (BMQ-R; Remmert et al., 2022). The BMQ-R reflects the proposed diagnostic criteria of misophonia by Jager et al. (2020). However, the BMQ-R is a long and comprehensive diagnostical instrument comprising 77 items. In comparison, given the S-Five’s inductive scale construction approach and resulting five core dimensions of misophonia, this scale measures typical misophonic experiences in a more efficient manner. Further, the S-Five allows to investigate the emotional nature and intensity of triggers. The strengths of the S-Five would thus certainly complement the measurement of misophonia in German samples. We therefore see merit in providing a valid German translation of the S-Five and in utilizing the strengths of both the S-Five and BMQ-R to investigate associations between misophonic symptoms.

Albeit evidence for the construct validity of the BMQ-R and the S-Five has been gathered, neither those two scales, nor any other misophonia questionnaire can be considered fully validated. Thus, construct validation plays a principal role in developing misophonia scales and in translating existing questionnaires. A widely used method for corroborating construct validity is showing evidence in line with nomological networks (Cronbach and Meehl, 1955). In nomological networks theoretical associations of constructs are to be empirically demonstrated and new constructs (e.g., misophonic symptoms) are to be placed in the proposed associational structure. To this end, hypotheses about relationships between attributes which are measured by a new instrument (e.g., the (German) S-Five) and convergent or discriminant constructs are formulated and tested. However, for relatively new constructs, such as misophonia, there is few and limited information on theoretical associations between constructs (i.e., misophonic symptoms or experiences). This does not imply the lack of a nomological network, but rather that it needs to be explored gradually. This study is a first and partially exploratory attempt to develop such a nomological network. The remainder of this introduction presents the descriptive and theoretical background for the development of the proposed nomological network of misophonia, followed by specific hypotheses and aims of the study.

A reasonable starting point for a nomological network of misophonic symptoms is the proposed diagnostic criteria put forward by Schröder et al. (2013) and revised by Jager et al. (2020). Based on a large sample of participants with misophonia, Jager et al. (2020) proposed five main symptom domains in their diagnostic criteria for clinically significant misophonia: (1) aversive emotional and physical reactions to sounds, with (2) insight into the excessive and disproportional nature of responses, (3) loss of self-control, (4) avoidance behavior, and (5) functional impairment. This description largely coincides with the recently published consensus definition of misophonia (Swedo et al., 2022). Nevertheless, several of the symptoms reported in the literature are not covered by those symptom domains, such as externalising and internalising appraisals as described by Vitoratou et al. (2021b) and Vitoratou et al. (2021a) and misophonic beliefs as described by Rosenthal et al. (2021). We therefore identified further symptom domains based on phenomenological similarities, and explored how these domains relate to different nomological aspects of misophonia. Recognition of similar psychological processes (e.g., reactions to sounds or influences on reactions; Swedo et al., 2022) and functions of symptoms (e.g., emotional regulation) is pivotal for the broadening of main symptom areas. This means grouping symptoms not necessarily by symptom type (e.g., a domain related to behavioral, cognitive, etc. symptoms), but rather by the function of the symptom (e.g., a domain for behavior used for the function of avoiding sounds or associated perceived threat, as separate from a domain for behavior used for the function of emotion regulation).

Based on the symptoms reported in contemporary misophonia literature (Jastreboff and Jastreboff, 2014; Brout et al., 2018; Potgieter et al., 2019; Jager et al., 2020; Vitoratou et al., 2021b; Swedo et al., 2022) we identified five main symptom domains: (1) misophonic appraisals, (2) misophonic emotional experiences, (3) misophonia-specific dysregulation, (4) misophonic avoidance, and (5) misophonic impairment. Critically, these symptom domains serve to give the nomological network a broader structure by clustering symptoms. In order to better understand this clustering attempt, the individual symptom domains and their associated symptoms are described in further detail below.

Misophonic appraisals encompass symptoms associated with the subjective meaning or evaluation placed on or knowledge about one’s own reactions to sounds and the circumstances in which they occur (i.e., attributional styles and clinical insight; Vitoratou et al., 2021b). These are meta-cognitive processes or beliefs about misophonic symptoms, rather than thoughts in response to misophonia triggers. The initial item pool for the Duke Misophonia Questionnaire (Rosenthal et al., 2021) included cognitive responses in the moment of triggers, but the items that were retained after factor analysis seemed to relate more to the state of urgency and intensity that occurs in that moment (e.g., “I would do anything to make it stop,” was retained), than to an appraisal of the situation (e.g., “They do not care how this sound affects me,” was not retained). In the symptom severity composite scale of the DMQ, these cognitive responses clustered together with other symptoms physical and emotion symptoms, not as a separate “cognitive” factor. That is, cognitions relating to the anguish of the moment were part of a dimension of physical, emotional and cognitive distress, and cognitions relating to assumptions about the moment did not seem to be part of the latent variable of misophonia symptom severity. We therefore focused this dimension on appraisals reflecting more general beliefs about the meaning of their symptoms, rather than appraisals in the moment.

The domain includes internalising and externalising appraisals (blaming for symptom experience; Rosenthal et al., 2021; Vitoratou et al., 2021b) as well as clinical insight (Jager et al., 2020; Swedo et al., 2022). Clinical insight included recognition of excess and recognition of disproportionality (e.g., see Jager et al., 2020). A broader definition of clinical insight includes the comprehension of one’s own symptoms (i.e., symptom coherence; e.g., Moss-Morris et al., 2002; Witteman et al., 2011). However, symptom coherence is a characteristic that has not been studied in the context of misophonia yet and is thus entirely exploratory in our study.

Misophonic emotional experiences entail all immediate emotional and physical reactions and experiences to misophonic triggers (i.e., anger, irritability, aggression, disgust, anxiety, and corresponding physical symptoms). Note that aggression entails different phenomenological aspects. We follow Buss and Perry (1992) distinguishing anger related to aggression, verbal and physical aggression as well as hostility. Physical reactions or symptoms are clustered within this domain since physical symptoms are part of the emotional misophonic response (i.e., autonomic stress response or emotional arousal; e.g., Edelstein et al., 2013). Although it has been shown that physical symptoms can be modelled as a separate misophonic factor (e.g., Dibb et al., 2021; Rinaldi et al., 2021), we do not see the benefit in separating physical reactions from the domain emotional experiences.

Misophonic emotional experiences are to be distinguished from misophonia-specific dysregulation, which is defined as an extension of loss of self-control (Remmert et al., 2022) as an incapability to cope with emotional experiences as well as uncontrolled behavioral manifestations for elevated levels of emotional arousal and negative affectivity. This also means disentangling various aspects of impaired self-control, including behavioral dysregulation (e.g., verbal or physical aggression), cognitive, and emotional dysregulation (i.e., loss of control over emotional experiences; e.g., Swedo et al., 2022). This domain is a category into which failed coping attempts fit (e.g., Guetta et al., 2022). It is not yet clear which domain the S-Five construct of perceived emotional threat fits into, as it includes items related to experiencing anxiety and distress, which may fit in the emotional experiences domain, as well as items related to feeling trapped and helpless (i.e., lack of regulative strategies to cope with misophonic experiences), which may align with the dysregulation domain.

Misophonic avoidance includes dysfunctional behavioral coping strategies to either prevent being exposed to misophonic sounds (anticipated avoidance) or escaping such situations (reactive avoidance; e.g., Remmert et al., 2022). Both avoidance behaviors form part of the definition of misophonia (Swedo et al., 2022). Rosenthal et al. (2021) showed that anticipated avoidance is the most prominent coping strategy before being faced with triggers, whereas reactive avoidance is the most prominent coping strategy when being triggered. Although avoidance behavior is a coping strategy, it can be distinguished from dysregulation because it serves the purpose of (re-)gaining control over the stimuli and is not the incapability to control emotional reactions. It may also include behaviors intended to prevent feared consequences of emotional dysregulation.

The fifth domain is misophonic impairment, which entails symptoms associated with the suffering and limitations caused by misophonic experiences (e.g., Wu et al., 2014; Jager et al., 2020; Swedo et al., 2022). Functional impact can be assigned to life domains or activities in which the impact occurs: e.g., cognitive impact, social impact, and impact on daily routine (WHO, 2001). Further, this domain entails distress as a consequence of misophonic symptoms including depressive mood and emotional burden (e.g., Jager et al., 2020; Remmert et al., 2022).

Note that this clustering of symptoms into domains is intended to facilitate the investigation of misophonic symptoms, rather than a strict classification. The domains may thus naturally overlap in some characteristics, while grouping misophonic symptoms reasonably. After having defined the broader structure of the nomological network, the following section outlines theoretical, empirical, and exploratory assumptions about how the specific symptoms are associated with each other, both within and across symptom domains (see Figure 1). Since there are 190 possible correlations between symptoms, we pragmatically concentrated on the core nomological structure, which predominantly involves symptoms being measured by the S-Five (indicated in grey boxes in the network in Figure 1). This is also due to the fact that the German S-Five is the focus of the presented studies. The proposed assumptions on associations are drawn from both misophonia research and the broader literature on mental disorders.

Core Nomological Network of Misophonic Symptoms. Circles represent proposed misophonic symptom domains. Squares represent specific misophonic symptoms. Grey squares represent symptoms being measured by the S-Five scales. Dashed lines around symptoms within symptoms domains shall indicate that symptom domains are not strictly separated. Strong associations are indicated by black arrows, weaker associations are grey, negative associations are blue with less negative associations in light blue, dashed arrows indicate uncertainty.

Associations in the nomological network

Principally, we expect that misophonic symptoms, regardless of their proposed symptom domain, are significantly positively associated with each other. This is based on the understanding that these symptoms together constitute the higher-order construct of misophonia and are hence naturally associated with each other, which is widely empirically supported (e.g., Rinaldi et al., 2021; Rosenthal et al., 2021; Vitoratou et al., 2021a,b, 2022a,b; Remmert et al., 2022). Moreover, misophonic symptoms within the proposed symptom domains are expected to be strongly associated since they are identified on account of sharing common characteristics and functions. Any exceptions to these two principles, as well as specific hypotheses about associations between symptoms from different symptom domains, are outlined in the following sections. An additional principle of our validation approach is that explicit convergent measures (i.e., measures that exactly measure the same symptom) are assumed to correlate strongly.

Misophonic appraisals

Misophonic attributional styles (internalising and externalising appraisals) have been shown to correlate moderately with each other (Vitoratou et al., 2021b), which we assume to replicate in this study. Beyond this, the relationships of interest for misophonic appraisals are with clinical insight (i.e., recognition of excess and disproportionality, and symptom coherence), functional impact and distress, as well as misophonia-specific aggressive behavior (S-Five outbursts).

Individuals with misophonia often recognize that their behavior is excessive or disproportionate to the situation (Hadjipavlou et al., 2008; Jager et al., 2020; Swedo et al., 2022). Although the relationship between attributional styles and dimensions of clinical insight have not been investigated yet for misophonia, it is reasonable to assume that those recognizing their reactions as excessive or disproportionate would be more likely to attribute blame to themselves (internalising) than to other people (externalising). Critically, these relationships have been substantiated for psychiatric disorders (e.g., schizophrenia; Cotton et al., 2012) and neurodevelopmental conditions (e.g., autistic spectrum conditions; Didehbani et al., 2012). We therefore expect a higher correlation between recognition of excess or disproportionality and internalising appraisals compared with externalising appraisals. Further, it has been found that a good understanding of the nature and cause of obsessive–compulsive symptoms (i.e., symptom coherence) is associated with internal attributions, but not with external, environmental attributions (Pedley et al., 2019). Based on this, we likewise expect individuals with higher levels of symptom coherence for misophonic symptoms to be less likely to blame themselves, and instead understand that the source of the problem is not the individual, but the condition of misophonia itself. Thus, a negative correlation is expected between misophonic symptom coherence and internalising appraisals and a less negative or non-significant correlation with external appraisals.

Another characteristic of internal attribution is that it is strongly associated with depression, distress and daily impact, whereas external attribution has been shown to be less strongly associated (e.g., Peterson et al., 1981; Hu et al., 2015), which has also been shown for misophonia (Vitoratou et al., 2021b). We therefore expect strong associations to emerge between misophonic distress symptoms and functional impact with internal appraisals, but substantially less with external appraisals.

Finally, regarding the relationship with misophonia-specific aggression (outbursts), the original validation of the S-Five (Vitoratou et al., 2021b) found that the outbursts factor was moderately correlated with internalising and externalising appraisals. Surprisingly, outbursts were more strongly correlated with internalising than with externalising appraisals and both appraisal factors had low positive correlations with anger reactions to trigger sounds. While other research indicates that anger, aggression, and aggressive behavior are more frequent in those who blame others than themselves for their reactions (e.g., Averill, 1983; Quigley and Tedeschi, 1996), this appears to be have been the case with misophonia (Vitoratou et al., 2021b). We anticipate that both types of appraisals will be associated with higher levels of anger, aggression, behavioral dysregulation, and outbursts. Since irritability shares common emotional characteristics with anger and aggression (e.g., Stringaris, 2011) we assume it will also be associated with internalising and externalising appraisals.

Misophonic emotional experiences

Misophonia can cause a strong physical reaction (Edelstein et al., 2013; Kumar et al., 2017), which is most strongly associated with emotional reactions (Rosenthal et al., 2021). Accordingly, strong correlations between emotional misophonic responses (i.e., anger, aggression, irritability, disgust, and anxiety) and physical symptoms are assumed. Whilst anxiety may co-occur in misophonia, it is different in the psychological process compared to other emotional reactions (i.e., anger, aggression, irritability, and disgust). Anxiety is a rather anticipatory emotion caused by perceived threat whereas anger and related emotions (aggression and irritability) as well as disgust are rather reactive emotions caused by violations of personal needs, integrity or boundaries. For misophonia, anger is the most prominent reactive emotion whereas anxiety, if present, is rather anticipatory (e.g., Jager et al., 2020). Since anxiety is different from other emotional reactions in some features and does not necessarily need to co-occur, it is assumed to correlate lower (but still moderately) with other emotional reactions.

Misophonia-specific dysregulation

As experiencing anger, aggression and irritability when confronted with sounds might manifest in behavioral dysregulation such as aggressive outbursts (e.g., Swedo et al., 2022), these symptoms are particularly expected to correlate. Likewise, behavioral dysregulation and outbursts are likely to be related to functional impact, with this behavior naturally contributing to social conflicts and negative consequences in daily life (Wu et al., 2014). It is further hypothesized that emotional dysregulation is linked to the concept of threat as measured by the S-Five, which includes experiences of feeling trapped and helpless (i.e., expressions of dysregulated threatening emotions). Experiencing threat is conceptually and empirically related to anxiety and heightened autonomic arousal (i.e., physical symptoms; Vitoratou et al., 2021b) and therefore expected to be associated with anxiety and physical symptoms. Moreover, threat entails aspects of failed avoidance strategies in the sense that threat emerges when triggers cannot be avoided. We expect that experiencing threat motivates increased avoidance behavior in order to circumvent the feared consequences of being triggered. Thus, positive associations between threat, anxiety and avoidance strategies are expected. Threat has further been shown to be strongly correlated with functional impact (Vitoratou et al., 2021b) and is likewise expected to cause significant distress in individuals’ lives.

Misophonic avoidance

Experiencing threat and anxiety is generally associated with pronounced avoidance behavior causing significant distress and social isolation (Abramowitz et al., 2019). Considering the frequent reports of both anxiety and avoidance behavior in misophonia (Wu et al., 2014; Jager et al., 2020; Swedo et al., 2022), we assume that perceived threat, anxiety and avoidance behavior will be strongly correlated. Avoidance behavior can also contribute to the maintenance of symptoms (e.g., Spinhoven et al., 2017) and poor treatment outcomes, thus elevating symptom burden (e.g., Wheaton et al., 2018). Hence, we assume strong associations with functional impact.

Misophonic impairment

Most of the associations for symptoms from this domain have already been described in the previous sections. In summary, all misophonic symptoms being measured by the S-Five except for impact (i.e., externalising appraisals, internalising appraisal, outbursts and threat) are expected to highly correlate with symptoms from the domain misophonic impairment.

Associations with symptoms of other mental disorders and traits

To further explore the extension of the nomological network, we also investigated associations between the S-Five and S-Five-T scores with related psychological constructs. In particular, anxiety sensitivity, which is a relatively stable trait fear of arousal-related sensations (Hovenkamp-Hermelink et al., 2019), has been shown to be related to misophonic symptoms (Cusack et al., 2018; McKay et al., 2018; Schadegg et al., 2021). Higher anxiety sensitivity was found to strengthen the relationship between misophonia and aggression (Schadegg et al., 2021). Cusack et al. (2018) found that the relationship between anxiety sensitivity and misophonia was partially mediated by obsessive–compulsive symptoms. An association between obsessive–compulsive symptoms and misophonia has also been reported elsewhere (Wu et al., 2014; Erfanian et al., 2019; Jager et al., 2020). Misophonia has been associated with symptoms of depression (Erfanian et al., 2019), particularly in relation to internalising appraisals and impact (Vitoratou et al., 2021b). This fits with the notion that internal attributional appraisals are strongly associated with depression and distress (Hu et al., 2015). Therefore, we assume high correlations between internalising appraisals and impact, and depressive symptoms. The associations between anxiety sensitivity, obsessive–compulsive symptoms and misophonic symptoms are exploratory because the associations with misophonia have only been shown for overall misophonic symptoms. However, experiencing threat when confronted with sounds entails aspects of anxiety and heightened arousal (see Misophonia-specific dysregulation), so it is likely that threat is associated with higher levels of anxiety sensitivity.

Hypotheses

We expect to find equivalent psychometric properties for the German S-Five compared with the original version. Specifically, we hypothesize configural invariance between German-speaking and English-speaking populations, high internal reliability and high stability in time (>0.75 in agreement coefficients). We further expect to find similar intercorrelations between symptoms measured by the S-Five compared to the original validation study (Vitoratou et al., 2021b), which are outlined in Table 1. In relation to the nomological network, we outlined our hypotheses in the preceding section and summarize them in Table 2.

Table 1.

Internal Consistencies and Intercorrelations of the S-Five from the Original Validation Study.

Measure	EXT	INT	IMP	OUT	THR
1. S-Five: External Appr.	(0.85)	-	-	-	-
2. S-Five: Internal Appr.	0.21	(0.88)	-	-	-
3. S-Five: Impact	0.29	0.50	(0.83)	-	-
4. S-Five: Outbursts	0.30	0.40	0.39	(0.84)	-
5. S-Five: Threat	0.27	0.32	0.51	0.33	(0.83)

Open in a new tab

N = 828. S-Five = Selective Sound Sensitivity Syndrome Scale; EXT = Externalising Appraisals, INT = Internalising Appraisals, IMP = Impact, OUT = Outbursts, THR = Threat. Cronbach’s α estimates are in parentheses on the main diagonal. The depicted correlations are rounded. All correlations are significant at p < 0.01.

Table 2.

Predicted Associations between Misophonic Symptoms in the Nomological Network.

Symptom domain	Misophonic symptom (measures)	Predicted associations with other misophonic symptoms
Misophonic Appraisals	Internalising appraisals (S-Five)	Positively correlated with: externalising appraisals, recognition of excess and recognition of disproportionality (BMQ-R), functional impact, distress, irritability, anger, aggression, behavioral dysregulation/outburstsNegatively correlated with: symptom coherence (IPQ-MH)
Misophonic Appraisals	Externalising appraisals (S-Five)	Positively correlated with: internalising appraisals, irritability, anger, aggression, behavioral dysregulation/outburstsIn comparison with internalising appraisals less correlated with: recognition of excess and recognition of disproportionality, symptom coherence, functional impact and distress
Misophonic Emotional Experiences	Anger (BMQ-R, AQ)	Positively correlated with: physical symptoms, externalising appraisals, internalising appraisals, and behavioral dysregulation/outbursts
	Irritability (BMQ-R, BITe)	Positively correlated with: physical symptoms, externalising appraisals, internalising appraisals, and behavioral dysregulation/outbursts
	Aggression (AQ)	Positively correlated with: physical symptoms, externalising appraisals, internalising appraisals, and behavioral dysregulation/outbursts
	Anxiety (BMQ-R, STICSA)	Positively correlated with: physical symptoms and threat Less correlated with anger, irritability, aggression, and disgust than their correlations with each other
	Physical symptoms (BMQ-R, STICSA)	Positively correlated with: anger, aggression, irritability, disgust (BMQ-R, DPSS-R), anxiety and threat
Misophonia-specific Dysregulation	Behavioral dysregulation/outburst (BMQ-R, S-Five, DERS)	Positively correlated with: anger, aggression, irritability, externalising appraisals, internalising appraisals, functional impact, and distress
	Emotional dysregulation (BMQ-R, DERS)	Positively correlated with: threat
	Threat (S-Five)	Positively correlated with: emotional dysregulation, anxiety, physical symptoms, reactive avoidance, anticipated avoidance, functional impact, distress
Misophonic Avoidance	Reactive avoidance (BMQ-R, NAQ, BEAQ)	Positively correlated with: threat, functional impact
Misophonic Avoidance	Anticipated avoidance (BMQ-R, NAQ, BEAQ)	Positively correlated with: threat, functional impact
Misophonic Impairment	Functional impact (BMQ-R, S-Five, WHODAS 2.0)	Internalising appraisals, externalising appraisals, behavioral dysregulation/outbursts, threat, reactive avoidance, anticipated avoidance
Misophonic Impairment	Distress (BMQ-R)	Internalising appraisals, externalising appraisals, behavioral dysregulation/outbursts, threat, functional impact
Other	Depressive symptoms (PHQ-9)	Positively correlated with: internalising appraisals, functional impact In comparison with internalising appraisals less correlated with: externalising appraisals.

Open in a new tab

Aims

The study has five specific aims:

Provide a rigorous German translation of the S-Five and S-Five-T instruments.
Replicate the results from the original S-Five in German.
Scrutinize the psychometric properties of the scales (i.e., internal consistency, model-based reliability, and test–retest reliability as well as evidence on construct validity including the factorial structure).
Utilize the S-Five to investigate an associational network of misophonic symptoms to demonstrate evidence for the construct-valid measurement of misophonic symptoms using the S-Five.
Provide a structural and theoretical basis for further explorations of misophonic symptoms and their associations through a nomological network.

Materials and methods

Study overview

Two studies were conducted. The first study was part of a larger validation study that investigated a nomological network for misophonia using the responses to the 25-items of the S-Five and to the items of a new diagnostical instrument for misophonia, the Berlin Misophonia Questionnaire (BMQ-R; Remmert et al., 2022). The purpose of the second study was to provide a (partial) replication of study 1, to include the S-Five-T measure, to evaluate the stability of the German versions of the S-Five and S-Five-T, and to extend the nomological network.

Participants

For both studies, individuals at least aged 16 or older were included in the analyses. Further eligibility criteria were not having been diagnosed with a severe learning disability or intellectual disability and having sufficient self-reported German language skills for answering the survey. Data protection guidelines were met and participants gave informed consent before completing the surveys. The studies were approved by the Ethics Committee at the Department of Education and Psychology of the Freie Universität Berlin, Germany (document number: 029/2020) and by the PNM Research Ethics Panel, King’s College London (RESCM-19/20-11,826).

In study 1, we further administered items assessing participants’ attention in line with DeSimone et al. (2015) and chose 80% correct answers as an inclusion cut-off. To check for aberrant response behavior we calculated a response pattern index as proposed by Meade and Craig (2012) excluding participants with more than 30% consecutive equal answers. The first study aimed at the evaluation of the dimensionality of the S-Five. Forero et al. (2009) suggested a sample size of N > 200–500 for using latent trait models for ordinal data using the WLSMV estimator. We collected data from 952 individuals, of which N = 639 (67.12%) completed the S-Five and met the inclusion criteria. For study 2, we recruited 322 participants, of which N = 235 (73.0%) met the inclusion criteria and completed at least the S-Five. The second study focused on the translation of the S-Five-T complementary trigger checklist along with providing a confirmation dataset for the factor structure of the S-Five scale and evaluating stability.

Both studies were conducted using social media platforms in Germany (e.g., Facebook and Instagram) as well as university mailing lists. Groups with individuals identifying as having misophonia as well as unspecific recruitment groups and groups with individuals suffering from any form of impaired hearing or disorders related to hearing (e.g., tinnitus, hyperacusis, etc.) were included in the sampling frame. The recruitment language was German. As an incentive participants could participate in a lucky draw for 10 × 5 Euro Amazon voucher and psychology students received course credit. In the second study motivation was provided in terms of a lucky draw for 25 × 20 Euro amazon vouchers. A test-retest study was conducted two to 4 weeks later.

Translation procedure

The translations of the scales from their respective language (i.e., Polish or English) to German was conducted by applying the TRAPD procedure (Harkness, 2003). TRAPD is an acronym for the following steps ensuring the quality of questionnaire translation: translation, review, adjudication, pretesting, and documentation. Two translators, who are fluent or native speakers of the respective languages, independently translated the items of each scale. The translated items were then reviewed with the translators and authors of the study (three of whom are German native speaker and fluent in English). Objects of the review were content, wording, and authenticity (i.e., evaluation of how natural or native the translation is) of the items. In this part, alterations of items were implemented if indicated.

Measures

All measures are described in detail below. For both studies, three scales measuring aspects of misophonia were administered: S-Five, BMQ-R, and MisoQuest.

For study 1, non-misophonia specific scales were administered, each with its instructions contextualized for the respondent to answer in relation to misophonia. At the beginning of the study, participants were asked to think about the sounds that bothered them most or, if not applicable, about typical misophonic sounds (i.e., eating, swallowing, and sniffing) and were instructed to consider either the presence or impact of those sounds in relation to each scale. For instance, we added the accessory sentence: “[…] when you are confronted with bothersome sounds.” This procedure aimed at minimizing between-person variability and within-person inconsistency due to thinking about different contexts when giving a response and thus aimed at increasing validity (cf. Lievens et al., 2008). The scales contextualized for misophonic sounds in study 1 were the Aggression Questionnaire (AQ), Brief Irritability Test (BITe), State–Trait Inventory for Cognitive and Somatic Anxiety (STICSA), Difficulties in Emotion Regulation Scale (DERS), Noise Avoidance Questionnaire (NAQ), Brief Experiential Avoidance Questionnaire (BEAQ), World Health Organization Disability Assessment Schedule 2.0 (WHODAS 2.0), Illness Perception Questionnaire Mental Health (IPQ-MH). Note that items from each instrument were administered randomly in blocks. For the BMQ-R, items from each symptom area were presented in randomized blocks (see Remmert et al., 2022).

For study 2, non-misophonia-specific scales were not contextualized for misophonia, because we aimed at investigating associations with adjacent clinical constructs not only limited to misophonic contexts. Further, the three constructs are not described as misophonic symptoms, so it is not reasonable to contextualize them accordingly. These were the Anxiety Sensitivity Index 3 (ASI-3), Patient Health Questionnaire 9 (PHQ-9), and the Dimensional Obsessive–Compulsive Scale (DOCS). Additionally, the S-Five-T was included in study 2, but not in study 1. For the S-Five, the order of questions (five S-Five items per question), and the order of item presentation in each question were randomized. Answering the S-Five-T trigger checklist was optional and participants were given the opportunity to skip to the following section after each trigger sound presented. This was done to minimize the potential distress and discomfort experienced when reading about misophonic triggers. Moreover, participants were randomly evenly allocated to either the BMQ-R or MisoQuest. The order of presentation for the BMQ-R items was randomized, as well as the order of the PHQ-9, DOCS, and ASI-3 thereafter. The three non-misophonia scales were optional. The links for the test-retest were sent out two to 4 weeks after initial participation in the survey. The follow-up survey contained the S-Five, S-Five-T, and basic demographic data such as a unique participant identification number and age.

Measures of misophonia

The Selective Sound Sensitivity Syndrome Scale (S-Five; Vitoratou et al., 2021b) is a self-report instrument measuring misophonic symptoms which consists of 25 items corresponding to five subscales: internalising appraisals, externalising appraisals, perceived threat, outbursts, and impact. Items are rated on an 11-point rating scale (0 = not at all true to 10 = completely true). A supplementary trigger checklist, the S-Five-T, consists of 37 misophonic triggers and both the emotional response (e.g., anger and disgust) and the intensity (from 0 to 10) of the response to triggers. Three indices can be derived: Trigger Count (i.e., number of triggers; TC), Frequency/Intensity of Reactions Score (i.e., total value of the intensity of triggers; FIRS), Relative Intensity of Reactions Score (i.e., intensity of reactions relative to the number of triggers; RIRS). The German and English S-Five can be found in the Supplementary material.

The Berlin Misophonia Questionnaire Revised (BMQ-R; Remmert et al., 2022) is a multidimensional diagnostical instrument for measuring misophonic symptoms. It consists of 15 symptom-oriented scales (excluding scales on anticipated reactions to sounds) which can be assigned to their corresponding diagnostic criteria by Jager et al. (2020). The scales have been shown to be reliable with McDonald’s ω ranging from.72 to.94. Results from latent variable models as well as correlations with convergent and discriminant measures give substantive evidence regarding construct validity. In total, 67 items were used, which are rated on a 6-point rating scale (0 = does not apply at all to 5 = completely applies).

MisoQuest¹ (Siepsiak et al., 2020) is a unidimensional self-report instrument of misophonia with 14 items, rated on a 5-point Likert scale (1 = strongly disagree to 5 = strongly agree). The instrument was translated from Polish into German.

Emotion states and dispositions

The Aggression Questionnaire (AQ; Buss and Perry, 1992) comprises four dimensions: (1) physical aggression, (2) verbal aggression, (3) anger and (4) hostility with 29 items being rated on a 4-point rating scale (1 = does not apply to 4 = fully applies). We used the German version of the Aggression Questionnaire (Werner and von Collani, 2004)² in an optimized version for the measurement of misophonia (see Remmert et al., 2022).

The Disgust Propensity and Sensitivity Scale Revised³ (DPSS-R; Cavanagh and Davey, 2000) reduced-item version (Fergus and Valentiner, 2009), is a measure of disgust encompassing the dimensions disgust propensity and disgust sensitivity. The items measure the frequency of physical and emotional symptoms of disgust which are rated on a 5-point rating scale (1 = never to 5 = always). For this study, only the disgust propensity (DP) items were used (six items), which measure how easily an individual is repulsed. The instrument was translated from English into German.

The State–Trait Inventory for Cognitive and Somatic Anxiety³ (STICSA; Ree et al., 2008) measures dimensions of state and trait anxiety. Only the two trait scales were used for the present study, which capture a predisposition to experience anxiety in response to certain types of stressors, namely cognitive (10 items) and somatic (11 items) stressors. The items are rated on a 4-point rating scale (1 = not at all to 4 = very much so). The instrument was translated from English into German and optimized for the measurement of misophonia (see Remmert et al., 2022).

The Anxiety Sensitivity Index (ASI-3; Taylor et al., 2007) consists of 18 items that assess anxiety sensitivity, that is fear of anxiety-related sensations. It consists of three subscales: physical, cognitive, and social concerns. Responses are given on a 5-point rating scale from 0 = do not agree at all to 4 = fully agree. The German version was developed by Kemper et al. (2011).

The Brief Irritability Test⁴ (BITe; Holtzman et al., 2015) is a 5-item measure of irritability in the last 2 weeks. Items are rated on a 6-point rating scale (1 = never to 6 = always). We used the German version by Krey (2017).

The Dimensional Obsessive–Compulsive Scale (DOCS; Abramowitz et al., 2010) is a 20-item measure of obsessive–compulsive disorder. There are four categories of concerns: contamination, responsibility for harm, unacceptable thoughts, and “just right” concerns (denoted as symmetry). For each category there are five questions (rated from 0 to 4), asking about time occupied, avoidance behaviors, associated distress, functional impairment, and resistance to obsessions and compulsions. The German version by Fink-Lamotte et al. (2020) ⁵ was used.

The Patient Health Questionnaire (PHQ-9; Kroenke et al., 2001) is a 9-item measure of symptoms of depression. Respondents answer how often they were bothered by each symptom in the past 2 weeks, on a 4-point rating scale from 0 = not at all to 3 = almost every day. We used the German version by Gräfe et al. (2004).⁶

Emotion regulation

The Difficulties in Emotion Regulation Scale (DERS; Gratz and Roemer, 2004) is a measure of emotion regulation which consists of six subscales. For this study, we chose the following three subscales: (1) impulse control difficulties, (2) difficulties engaging in goal-oriented behavior, and (3) limited access to emotion regulation. These subscales consist of 19 items in total, of which 15 were chosen regarding their content validity to match the intended validation purpose. The items are rated on a 5-point rating scale regarding the experienced frequency (1 = almost never (0–10%) to 5 = almost always (91–100%)). The German version by Gutzweiler and In-Albon (2018)⁴ was used.

Avoidance behavior

The Noise Avoidance Questionnaire⁵ (NAQ; Bläsing and Kröner-Herwig, 2012) is a German self-report instrument measuring sound avoidance in daily life. It comprises 25 items of which 10 items describe specific situations that might be avoided. The remaining 15 items refer to specific behaviors related to sound avoidance. The more behavior-oriented items were chosen which are rated on a 5-point rating scale (1 = never to 5 = very often/always). We could not obtain the German items, so that the English items were translated. Since the statements are short and concise, we do not expect compromising effects due to translation, however, we optimized the item selection for the measurement of misophonia (see Remmert et al., 2022).

The Brief Experiential Avoidance Questionnaire (BEAQ; Gámez et al., 2014) is a 15-item measure of avoidance behavior. For this study, items from the original Behavioral Avoidance subscale of the German version (Böge et al., 2020)⁷ were relevant as they reflect situational avoidance of physical distress. Items are rated on a 6-point rating scale (1 = strongly disagree to 6 = strongly agree).

Impairment

The World Health Organization Disability Assessment Schedule 2.0⁸ (WHODAS 2.0; Üstün et al., 2010) is a clinical instrument based on the International Classification of Functioning, Disability and Health (ICF; WHO, 2001) which measures the impact of a given health condition in six domains of life: Cognition, mobility, self-care, getting along, life activities, and participation. Since mobility and self-care appear to be irrelevant for misophonia, these domains were not administered. The German self-report 36-item version (27 items after discarding the two domains) was optimized for the measurement of misophonia (see Remmert et al., 2022). Items are rated regarding the extent of difficulty individuals have doing the presented activities using a 5-point rating scale (1 = none to 5 = extreme or cannot do).

Clinical insight

Illness Perception Questionnaire Mental Health (IPQ-MH; Witteman et al., 2011) is an adapted version of the Illness Perception Questionnaire Revised (IPQ-R; Moss-Morris et al., 2002) measuring an individual’s perception of their mental health problem. Only the coherence subscale (five items), measuring the extent of an individual’s understanding of their mental health problem, was used for this study. The items are rated on a 5-point rating scale (1 = strongly disagree to 5 = strongly agree), with a higher score indicating more symptom coherence. The German version of the IPQ-R⁹ (Gaab et al., 2008) was used and adapted in line with Witteman et al. (2011) by replacing the term ‘illness’ with ‘problem’ in each item. There was one item from the coherence subscale of the IPQ-R that was removed from the scale for the final version of the IPQ-MH (“the symptoms of my condition are puzzling to me”). We included it in our survey as it had appeared in the German version of the IPQ-R and after initial psychometric examination showed a good fit, we retained the item in the measure.

Statistical analyses

The subscales of the S-Five were jointly modelled in a confirmatory factor analysis (CFA) according to the specified measurement model by Vitoratou et al. (2021b). Therefore, we specified a correlated first-order factor model. Measurement models of the validation instruments were specified according to the original factor structure, but sometimes with an optimized set of items which adequately fit the measurement of misophonia (see Remmert et al., 2022). For the DERS scales we specified an S•I-1 model (Eid et al., 2017) with one item as the reference item (general dysregulation) and the other items as specific factors (dysregulation facets), which is different from the original. This procedure allows us to investigate associations of misophonic symptoms with general dysregulation and its facets rather than with the facets alone.

Non-normality and categorical indicators were taken into account using the weighted least square mean and variance adjusted (WLSMV; Muthén et al., 1997) estimator with ordered categories. For the S-Five (continuous indicators) we used maximum likelihood estimation with robust (Huber-White) standard errors (Huber, 1967; White, 1980). Item omissions were addressed using full information maximum likelihood estimates.

Model fit was evaluated by using absolute and relative fit indices. Namely, the exact relative χ² (that is the ratio of the χ² over the degrees of freedom) with values ranging from 2 (Hoelter, 1983; Ullman, 2001) to < 5 (Wheaton et al., 1977; West et al., 2012) indicating adequate fit (Schermelleh-Engel et al., 2003), the Root Mean Squared Error of Approximation (RMSEA) with values close to 0.06 indicating adequate fit (Hu and Bentler, 1999), the Standardized Root Mean Square Residual (SRMR) with values close to 0.08 indicating good fit (Hooper et al., 2008), McDonald’s Centrality Index (Mc) close to 0.90 (Hu and Bentler, 1999); as well as Comparative Fit Index (CFI) and Tucker Lewis Index (TLI) close to 0.97 (Schermelleh-Engel et al., 2003). The Expected Cross Validation Index (ECVI) was used to compare non-nested models (Browne and Cudeck, 1989).

The multiple indicator multiple causes model (MIMIC; Joreskog and Goldberger, 1975; Muthén, 1979) was used to assess measurement invariance in relation to gender and age. An item was regarded as non-invariant when the effect of the exogenous variable (age or gender) on the item directly (hereafter direct effect or de) was statistically significant.

Internal consistency was estimated with model-based McDonald’s ω (McDonald, 1999). The test-retest reliability was evaluated using the intraclass correlations coefficient (ICC; Shrout and Fleiss, 1979). ICC values > 0.75 are interpreted as good reliability, according to Koo and Li (2016).

Correlations were interpreted in line with Cohen (1988); i.e., r = |0.10|, r = |0.30|, r = |0.50| are considered weak, moderate and strong, respectively. Differences between correlations were statistically compared using Fisher’s z-test of dependent correlation or between an empirical and a hypothesized correlation (Fisher, 1956). We applied Bonferroni correction to significance-level α in order to address α-error inflation due to multiple testing (Bonferroni, 1936) and further decided to be as conservative as possible, thus correcting for all calculated correlations per study. In study 1 we calculated 820 correlations and therefore α = 0.00006 and for study 2 we calculated 465 correlations and therefore α = 0.0001. Additionally, we corrected for hypothesized correlation comparisons. For the 17 comparisons in study 1 α = 0.0029 and for the 17 comparisons in study 2 α = 0.0029. Ten comparisons between independent samples were tested using Fisher’s z-test of correlations in two independent samples (Fisher, 1956) with a corrected α = 0.005. Note that due to dropouts (respectively pairwise complete analyses) the sample sizes of dependent comparisons between correlation may vary within both studies, so we always selected the smallest overlapping sample size and still counted all comparisons within each study to adjust alpha-inflation, which is the most conservative method. The statistical software of Stata 16 (StataCorp, 2019), Mplus 8 (Muthen and Muthén, 1998-2017), and the “lavaan” package (version 0.6-9; Rosseel, 2012) in R (R Core Team, 2017) were used to carry out the analysis.

Results

Descriptive indices

In sample 1 with N = 639, most participants (86.2%) were female, and two individuals indicated non-binary gender. The mean age was M = 34.28 years (SD = 11.52, range 16 to 69). Approximately one-third (32.7%) were students. Further, one-third of the sample had a university degree and 46.7% had at least a college entrance qualification. A majority had a partner or was married (65.1%), whereas 33.9% did not have a partner or was living separated. Almost a third of the sample (30.3%) was either part-time or marginally employed, 37.3% was full-time employed, and 13.3% unemployed. About half of the sample (47.5%) fulfilled the diagnostic criteria by Jager et al. (2020) (with 24.6% having severe symptoms) as classified by the BMQ-R.¹⁰ According to the S-Five total score cut-off (i.e., total score of 87 or higher; cf. Vitoratou et al., 2022a), more than half of the sample (57.4%) had significant misophonia.

In sample 2, with N = 235, the majority of participants was female (85.1%), with two participants identifying as non-binary. The mean age was 35.8 years (SD = 11.8, range 19 to 80). The majority of the sample (95.3%) reported living in Germany or another German-speaking country (Germany 86.0%, Austria 5.5%, Switzerland 3.0, 5.9% rest of world). In terms of educational attainments, 6.4% had up to high school, 47.7% reported having done apprenticeships, 26.8% undergraduate degree, 14.5% postgraduate degree, and 4.7% doctoral or similar. Significant misophonia as indicated by the S-Five total cut-off was observed for 58.3% of the sample.

Structural validity and measurement invariance

The five-factor correlated model showed adequate fit to the data in both the first [χ²(265) = 850.93, p < 0.001, rel. χ² = 3.21, CFI = 0.94, TLI = 0.93, RMSEA = 0.07 [0.06–0.07], SRMR = 0.05, Mc = 0.55, ECVI = 1.87] and the second sample [χ²(265) = 452.15, p < 0.001, rel. χ² = 1.71, CFI = 0.94, TLI = 0.94, RMSEA = 0.06 [0.05–0.06], SRMR = 0.05]. An outline of the estimated model in sample 1 is shown Figure 2.

First-Order Factor Model of S-Five Scales. N = 639. All factor loadings, intercepts, and correlations are significant at p < 0.00006. Factor loadings and intercepts are completely standardized (both latent and observed variables). Unstandardized factor standard deviations are shown next to the latent variables.

Measurement invariance was explored with respect to gender and age, each adjusted for the other and the levels of the five factors. In sample 1, five items were directly affected from age but with negligible effect sizes (INT01: de = −0.018, p = 0.017; INT02: de = −0.020, p = 0.005, INT04: de = 0.020, p = 0.042, IMP03: de = 0.032, p = 0.001, IMP02: de = 0.670, p = 0.017). Two items were also affected from gender adjusted for age and the five misophonic dimensions of the misophonic experience (INT05: de = 0.698, p = 0.001; OUT03 de = 0.670, p = 0.017), with however less than one unit of effect on a 0 to 10 scale. Similar results emerged in the second sample, with three significant effects emerging for either age (INT02: de = −0.037, p = 0.002, INT04: de = −0.034, p = 0.007, OUT04: de = −0.032, p = 0.017) and gender (EXT01: de = 1.001, p = 0.013; IMP01: de = 1.068, p = 0.002; IMP03: de = 1.588, p < 0.001), with low magnitudes in either case.

An alternative bifactor S-1 model (Eid et al., 2017) was also fitted as from a theoretical perspective, the outburst factor comprises both verbally as well as physically aggressive behaviors. The bifactor S-1 model maintains a general outburst factor but takes the implied two-dimensionality of outbursts into account. The model was specified with physically aggressive behavior as the reference facet (G-factor) and verbally aggressive behavior as the specific factor yielding a model with good fit [χ²(4) = 11.84, p < 0.05, rel. χ² = 2.96, CFI = 0.99, TLI = 0.98, RMSEA = 0.07 [0.03–0.12], SRMR = 0.02, Mc = 0.99, ECVI = 0.08]. Likewise, an adapted five-factor correlated model integrating the presented bifactor S-1 approach for the factor outbursts demonstrated good fit [χ²(259) = 714.73, p < 0.001, rel. χ² = 2.76, CFI = 0.96, TLI = 0.95, RMSEA = 0.06 [0.05–0.06], SRMR = 0.05, Mc = 0.63, ECVI = 1.63]. Model comparison using a likelihood ratio test yielded a significantly better model fit of the bifactor S-1 model [Δχ²(6) = 115.29, p < 0.001, ΔCFI = 0.01, ΔRMSEA = 0.01]. For an outline of the alternative bifactor S-1 model see Figure 3.

Bifactor S-1 Model of S-Five Scales. N = 639. Factor loadings and intercepts are completely standardized (both latent and observed variables). Unstandardized factor standard deviations are shown next to the latent variables. ^zThe correlation is per definition set to zero. ^†n.s., ^*p < 0.05, ^**p < 0.01.

Internal consistency and intercorrelations of the S-Five

In both samples, the estimated internal consistencies of the factors were good to excellent, according to McDonald’s ω (ω ranged from 0.86 to 0.93; see Table 3). Descriptively, we found similar internal consistencies compared with the original validation study, except for impact and perceived threat, which were found to be slightly higher in our studies. In Study 1, the factor intercorrelations ranged from r = 0.40 to r = 0.72 with threat and impact being highest correlated (Table 1). Similarly, in study 2 the factor intercorrelations ranged from r = 0.51 to r = 0.79, with threat and impact again being most strongly correlated. All intercorrelations are significantly higher than in the original validation study (p < 0.005 for all comparisons), but we found almost the same correlational pattern. An exception was perceived threat which was comparably higher correlated with internalising appraisals and outbursts than other factors were correlated with internalising appraisals and outbursts. This aligns with the fact that perceived threat and outbursts were in general unproportionally highly correlated with other factors when compared to the original validation study (differences between 0.16 and 0.36).

Table 3.

Means and standard deviations, latent (Study 1) and Spearman’s (Study 2) intercorrelations, and reliability estimates for the S-Five.

Measure	Min–Max	Study 1		Study 2		EXT	INT	IMP	OUT	THR
		M	SD	M	SD
1. S-Five: External Appr.	0–50	28.97	15.26	27.31	15.70	(0.92/0.92)	0.51	0.59	0.57	0.58
2. S-Five: Internal Appr.	0–50	17.78	15.96	16.85	14.74	0.40	(0.93/0.88)	0.64	0.72	0.73
3. S-Five: Impact	0–50	12.19	13.83	14.76	14.08	0.48	0.61	(0.90/0.91)	0.65	0.79
4. S-Five: Outbursts	0–50	16.20	13.91	16.50	13.97	0.54	0.61	0.55	(0.87/0.86)	0.69
5. S-Five: Threat	0–50	26.85	16.55	28.98	17.28	0.56	0.62	0.72	0.63	(0.92/0.89)

Open in a new tab

N = 639 (Study 1); N = 235 (Study 2). Cells below the diagonal represent latent intercorrelations for Study 1; Cells above the diagonal represent Spearman’s correlation coefficients (ρ) for Study 2. S-Five = Selective Sound Sensitivity Syndrome Scale; EXT = Externalising Appraisals, INT = Internalising Appraisals, IMP = Impact, OUT = Outbursts, THR = Threat. McDonald’s ω (McDonald, 1999) based on the respective confirmatory factor analyses are in parentheses on the main diagonal (on left = Study 1, right = Study 2). All correlations for study 1 were significant at p < 0.00006 and for study 2 at p < 0.0001 (Bonferroni-corrected significance level, respectively). M = mean, SD = standard deviation; Min = scale minimum, Max = scale maximum; N = 633-636 (Study 1); N = 235 (Study 2). Means were calculated for manifest sum scores of the respective scale.

Test–retest reliability (study 2)

The S-Five items and scores all showed excellent agreement across the test and retest (N = 52), with ICC $\geq$ 0.86 in all cases and ICC = 0.90 for the total S-Five score. Similarly for the S-Five-T trigger scores, agreement was excellent with ICC $\geq$ 0.84 in all cases and ICC = 0.90 for the TC, FIRS, and RIRS.

Nomological network of misophonic symptoms: Construct validity

In this section, we report results regarding the proposed nomological network of misophonic symptoms for each symptom domain. In each section, we first describe associations within the respective symptom domain followed by associations between different domains and associations with symptoms of other mental disorders and traits. We additionally report associations between misophonic trigger scores and between S-Five scales and overall misophonic symptoms at the end of this section. An updated version of Figure 1 depicting the empirical nomological network can be found in Figure 4.

Empirical Nomological Network of Misophonic Symptoms. Circles represent proposed misophonic symptom domains. Squares represent specific misophonic symptoms. Grey squares represent symptoms being measured by the S-Five scales. Dashed lines around symptoms within symptoms domains shall indicate that symptom domains are not strictly separated. Strong associations are indicated by black arrows, weaker associations are grey, negative associations are blue with less negative associations in light blue, exploratory associations are red with weaker associations in light red, dashed arrows indicate uncertainty due to mixed results.