Emotion dysregulation in adult ADHD: Introducing the Comprehensive Emotion Regulation Inventory (CERI)

ABSTRACT

Introduction

Previous research has shown that Attention-Deficit Hyperactivity Disorder (ADHD) is linked to emotion dysregulation, but we still do not know enough about the specific nature of this deficit. The aim of the present study was therefore to study emotion dysregulation in adult ADHD using a new self-rating instrument, the Comprehensive Emotion Regulation Inventory (CERI).

Methods

The study included 390 participants, both adults diagnosed with ADHD and a comparison group. The CERI was created based on Gross’ process model of emotion. Unlike instruments already available, the CERI enables a detailed analysis of different strategies for regulating both positive and negative emotions and of the impact emotion dysregulation has on daily life.

Results

Results showed that the CERI has good psychometric properties, and a 5-factor solution was identified. Adults with ADHD differed from the comparison group with regard to how often they use different strategies, although effect sizes were small to medium. However, large effect sizes were found between ADHD and the comparison group with regard to problems implementing the use of various regulatory strategies and the negative impact on daily life functioning with regard to friendships, work/studies, or romantic relations. Within-group differences showed that adults with ADHD use the less advanced strategy suppression significantly more often compared to the more advanced strategies attentional deployment and reappraisal.

Conclusions

The CERI appears to be a valuable instrument for gaining more in-depth insights into emotion dysregulation in adults with mental disorders. Future research needs to examine what aspects of emotion dysregulation are shared between different mental disorders and what aspects might be specifically related to ADHD.

KEYWORDS:

• ADHD
• emotion regulation
• emotional reactivity
• psychometric properties
• Gross’ process model

An increasingly large body of research on Attention-Deficit Hyperactivity Disorder (ADHD) has emphasized that emotion dysregulation should be regarded as a prominent feature of this disorder in both childhood/adolescence and adulthood (reviews by Bunford, Evans, & Wymbs, 2015b; Faraone et al., 2019; Martel, 2009; Shaw et al., 2014). However, most previous studies assessing emotion dysregulation in ADHD have used general measures of emotional functioning (i.e., including items such as “overacts emotionally”) or symptom levels related to other psychiatric symptoms (“e.g., anxiety and depression), rather than investigating the use of specific strategies for regulating emotions. As emphasized in a recent review by Faraone et al. (2019), there is a need to develop new measurements that capture the multidimensional nature of emotion dysregulation. In the present study, we therefore introduce the Comprehensive Emotion Regulation Inventory (CERI), which enables more detailed investigation of the following aspects of emotion regulation: (1) the use of different strategies for regulating emotions, (2) problems implementing strategies for regulating emotions, and (3) the impact of poor emotion regulation on daily life. In addition, the CERI includes items measuring emotional reactivity as it has been argued that it is important to distinguish between reactivity and regulation (Cole et al., 2004). By including these different aspects, we hope to provide a more in-depth understanding of the association between emotion dysregulation and ADHD. This should be considered important, as ADHD is a heterogeneous disorder, and it is likely that there are subgroups of the disorder which differ with regard to emotional functioning. In addition, emotion dysregulation is a transdiagnostic construct. Hopefully, an instrument providing a more comprehensive assessment would also allow future research to examine what aspects of emotion dysregulation are shared between different mental disorders and what aspects might be specifically related to ADHD.

ADHD and emotional dysregulation

Previous research has made it clear that emotion dysregulation is strongly related to ADHD among children, adolescents, and adults, also when comorbid disorders are taken into account (reviews by Bunford, Evans, & Becker, 2015a; Martel, 2009; Shaw et al., 2014). It has also been shown that although emotion dysregulation is significantly related to other neuropsychological deficits associated with ADHD (e.g., executive function deficits and delay aversion), emotion dysregulation has an independent effect on ADHD when controlling for these other deficits (Sjöwall et al., 2015, 2013). Another reason for why it should be considered important to assess emotion dysregulation in ADHD is that we know from previous research that individuals with ADHD who have problems regulating their emotions have poorer daily life functioning compared to those with ADHD without emotion dysregulation with regard to, for example, comorbid symptom levels, social relations, and quality of life (e.g., R. Barkley & Fischer, 2011; Bodalski et al., 2019; Bunford et al., 2015b; Bunford, Evans, & Langberg, 2018; Surman et al., 2013). Emotion dysregulation has also been found to be associated with the persistence of ADHD symptom levels over time (e.g., Biederman et al., 2012). Based on these findings, it can be argued that strategies for improving emotion regulation should be a vital part of behavioral treatments for ADHD, regardless of whether they improve core ADHD symptom levels. In order for these treatments to be effective, it is important to have more detailed knowledge about what specific emotion regulation deficits are present – knowledge that would allow us to tailor the treatment to patients’ individual needs. In sum, we know from previous research that ADHD is strongly linked to emotion dysregulation, but we still know very little about how to characterize the specific nature of emotion dysregulation in ADHD (cf. Faraone et al., 2019)

Gross’ model of emotion regulation

Emotional dysregulation is a very complex construct, including not only the regulation of several different types of emotions, but also regulation through several different types of strategies. In contrast to other neuropsychological deficits associated with ADHD, such as inhibition, emotion regulation is not a matter of inhibiting or not inhibiting a response, but rather a process that could include several different steps, each of which could be more or less effective. In the present study, we used Gross (2001) process model of emotional regulation as our theoretical framework when developing the CERI. Gross model is one of the most well-established models of emotion regulation and it is to our knowledge the only available model that includes a detailed description of different strategies for regulating emotions. In this introduction, we only describe the parts that are relevant to the present study and refer to other publications for more details on the extended version of this model (e.g., Sheppes et al., 2015). The original process model (Gross, 2001) includes five steps. Step 1 (“situation selection”) involves encountering an emotionally relevant situation. Problems at the first step include not being able to foresee which situations will cause strong emotions, and therefore need regulation, or not being able to avoid a situation even though one wishes to do so. Step 2 (“situation modification”) involves changing the situation in some way when one is already in an emotionally challenging situation. Step 3 (“attentional deployment”) involves diverting one’s attention away from an emotional situation. Step 4 (“cognitive change,” often referred to as “reappraisal”) involves interpreting the situation in an alternative way so as to reduce strong emotions. Finally, Step 5 (“response modulation,” often referred to as “suppression”) includes hiding one’s emotions from others. Individuals vary in the extent to which they employ these different strategies, and this variation could result from actively choosing not to use a certain strategy or having problems implementing a certain strategy despite wanting to do so.

Gross’ model is a general model that was not created specifically for understanding emotional dysregulation in patients with different mental health problems, and much of the previous research on this model has been conducted on university students. However, a few studies have tested parts of this model in patients with psychiatric disorders, such as bipolar disorder (Johnson et al., 2016), social anxiety disorder (Blalock et al., 2016; Werner et al., 2011) and autism spectrum disorders (Samson et al., 2015). A general finding from these studies is that individuals with mental health problems have a tendency to use less effective strategies for regulating emotions (e.g., emotional suppression) compared to more effective strategies (e.g., cognitive reappraisal). These findings further emphasize the need to not only investigate whether or not emotions are regulated, but to also look at the extent to which specific regulatory strategies are used.

Few previous studies on ADHD have used Gross’ theoretical framework. However, Sheppes et al. (2015) hypothesized that individuals with ADHD have problems implementing the most appropriate strategy due to temporal discounting, which means that they prioritize short-term benefits over long-term goals. In one of the few empirical studies on adult ADHD that has addressed specific regulatory strategies, ADHD symptom levels were not significantly related to either cognitive reappraisal or emotional suppression (Bodalski et al., 2019). However, this previous study found that higher levels of ADHD symptoms were related to higher levels of cognitive and behavioral avoidance (i.e., avoidance in general, not measured as a specific strategy for regulating emotions), which in turn was related to a range of negative outcomes.

Measuring emotion dysregulation in adults

A large amount of research has demonstrated that individuals with ADHD often have higher levels of other psychiatric symptoms such as those associated with depression, anxiety, and different forms of antisocial behavior (e.g., review by Katzman et al., 2017). These are all disorders associated with strong negative emotional reactions. When creating the CERI, we therefore wanted to focus on strategies for emotion regulation (i.e., how exactly emotions are regulated) rather than emotional reactivity (i.e., how often and strongly emotions are displayed). Naturally, it is difficult to separate emotional dysregulation and emotional reactivity. However, an individual with infrequent and weak emotional reactions may display poor regulation and an emotional individual may be a relatively good regulator. Emotion dysregulation and emotional reactivity have therefore been conceptualized as two separate phenomena (e.g., review by Cole et al., 2004). In support for this distinction, previous research on children has demonstrated that emotional reactivity and dysregulation have different longitudinal correlates (Eisenberg et al., 1995). In previous ADHD research on children, it has been shown that emotion dysregulation is a better predictor of peer preference compared to emotional reactivity (Melnick & Hinshaw, 2000; Thorell et al., 2017).

Previous research on emotional functioning in adult ADHD has largely focused on emotional reactivity, sometimes also referred to as emotional lability. One reason for this could be the lack of instruments that focus specifically on emotion dysregulation. However, there are at least three exceptions: (1) the Emotion Regulation Questionnaire (ERQ; Gross & John, 2003), which focuses on reappraisal and suppression, (2) the Cognitive Emotion Regulation Questionnaire (CERQ; Garnefski et al., 2001), which focuses on different forms of cognitive regulation such as self-blame, rumination and catastrophizing, and (3) the Coping Inventory for Stressful Situations (CISS; Endler & Parker, 1990), which includes emotional reactivity, but also avoidance and items referred to as “task-oriented behaviors” (e.g., “Do what I think is best” and “Work to understand the question”). Thus, these questionnaires do not include subscales targeting different strategies for regulating emotions, without also including emotional reactivity, and none of them include all the five steps included in Gross process model.

Another important limitation is that available questionnaires, except for the ERQ, do not include items related to dysregulation of positive emotions. This should be seen as a limitation, as it has been demonstrated that, at least in studies on children, ADHD diagnosis and ADHD symptom levels are related to dysregulation of both negative emotions and happiness/exuberance (Forslund et al., 2016; Rydell et al., 2003; Sjöwall et al., 2013). In addition, one previous study of children found that dysregulation of happiness/exuberance is a better predictor of peer problems compared to dysregulation of negative emotions (Thorell et al., 2017). To our knowledge, previous research has not investigated the association between dysregulation of positive emotions in daily life functioning in adults with ADHD. Based on the limitations of the previous research described above, there is a need for an adult self-rating instrument that (1) distinguishes between reactivity and regulation, (2) includes both positive and negative emotions, and (3) includes a more detailed analysis of the extent to which different emotion regulation strategies are used and whether a given individual experiences problems with implementing these strategies. The CERI was developed to meet these needs.

Aims of the present study

The aim of the present study was to study emotion dysregulation in adult ADHD using a new self-rating instrument, the Comprehensive Emotion Regulation Inventory (CERI), which enables a detailed analysis of how often different strategies are used for regulating both positive and negative emotions, of the extent to which an individual experiences problem in implementing different strategies, and of the negative impact emotion dysregulation has on daily life. More specifically, we examined the following research questions:

1. Is it possible to distinguish between the five different strategies for regulating emotions included in Gross’ process model using the CERI?

2. What are the psychometric properties of the CERI in terms of reliability (i.e., internal consistency and test–retest reliability) and validity (i.e., associations between the CERI and the ERQ, an already available instrument targeting emotion regulation)?

3. What is the association between the use of different emotion regulation strategies and the impact that emotion dysregulation has on daily life?

4. Are adults with ADHD more emotionally dysregulated compared to adults without any psychiatric disorders as measured by the CERI?

Methods

Participants and procedure

The present study consisted of a total of 390 adults between 18 and 60 years of age. In order to increase the power in our analyses, we included twice as many individuals in the comparison group (n = 260) compared to the ADHD group (n = 130). Power analysis showed that when using twice as many individuals in the comparison group, a sample size of 131 in the clinical group is needed in order to be able to detect also small effects with a power of.80. Descriptive data are presented in Table 1. The ADHD group consisted of clinically referred patients with ADHD (age 18 to 59 years) who were recruited in collaboration with three adult psychiatry clinics. All patients with ADHD were shown to meet the full diagnostic criteria for ADHD as specified in the fifth version of the Diagnostic and Statistical Manual of Mental Disorders (DSM‐5; American Psychiatric Association [APA], 2013). During the diagnostic assessment, the patients met both a licensed clinical psychologist and a psychiatrist and the assessment included a detailed anamnesis, neuropsychological testing, a diagnostic interview, and standardized rating instruments. Information was collected from both the patient him-/herself and a significant other (i.e., a partner, parent, or sibling). The second version of the Diagnostic Interview for ADHD in Adults (DIVA 2.0; Kooij & Francken, 2010) was used to assess of ADHD symptoms and impairment in five areas of functioning (i.e., education, work, family, social/relationships, and self-confidence) in both childhood and at the present time. With regard to ratings, childhood ADHD symptom levels were assessed using the Childhood Symptom Scale (R. A. Barkley & Murphy, 1998), the Autism – Tics, AD/HD, and other Comorbidities inventory (A-TAC; Larson et al., 2010), or the Wender Utah Rating Scale (WURS; Ward et al., 1993). ADHD symptom levels in adulthood were assessed using the Adult ADHD Self-Report Scale (ASRS; Kessler et al., 2005) or the Adult Self-report Scale (ASR; Achenbach & Rescorla, 2003). For significant others, the Adult Behavior Checklist (ABCL; Achenbach & Rescorla, 2003) was used most often. Comorbid symptoms were assessed using the Mini International Neuropsychiatric Interview (M.I.N.I.; Sheehan et al., 1998). If comorbid symptoms were identified during this structured interview, it was complemented with one or several standardized rating instruments, depending on the identified symptoms. Patients with a full-scale intelligence score of <70, most often assessed using the fourth edition of the Wechsler Adult Intelligence Scale (WAIS‐IV; Wechsler, 2008), were excluded from the study.

Table 1. Descriptive data for sex, age, and comorbid disorders for the individuals with ADHD and the comparison group.

	ADHD (n = 130)	Comparison group (n = 260)
Sex, n (%) males	45 (33.1)	80 (30.8)
Age, mean (SD)	33.11 (9.93)	37.30 (7.77)
Education
9 years of schooling or less	21 (16.3)	3 (1.2)
High school	64 (49.6)	43 (16.6)
University/college degree	44 (34.1)	213 (82.2)
ADHD presentation
Combined presentation, n (%)	109 (85.4)	–
Inattentive presentation, n (%)	19 (15.4)	–
Comorbid disorders (%)
Autism Spectrum Disorder	3.1	0
Depression/Anxiety Disorders	17.0	0
Obsessive Compulsive Disorder	1.5	0
Phobias	1.5	0
Eating Disorders	3.1	0
Borderline Personality Disorder	1.5	0
Bipolar disorders	2.3	0

The comparison group consisted of 260 adults (age 18 to 48 years) without a psychiatric disorder. This was a convenience sample recruited through advertisements at universities, schools, stores, and health-care centers. Among the individuals in the ADHD group, about half of the participants completed the questionnaire during a visit to the clinic and half in the home setting. The comparison group received the questionnaire by mail and returned it in a pre-stamped envelope. To examine test–retest reliability, a small subsample of the controls (n = 44) of the individuals in the comparison group were asked to complete the CERI twice, with a 2- to 3-week interval between the first and second occasion and 91% of them (n = 40) agreed to do this. Moreover, ratings using the ERQ (i.e., the measure used for assessing convergent validity) were only available for the same subsample (n = 44). The male/female ratio did not differ significantly between adults with ADHD and the comparison group (χ² = .21, p = .64). However, individuals in the comparison group were significantly older than those in the ADHD group (t = 4.56, p < .001), and had a significantly higher educational level (χ² = 95.97, p < .001). The study was approved by the local ethics committee.

Measures

Comprehensive Emotion Regulation Inventory

The Comprehensive Emotion Regulation Inventory includes a total of 42 items divided into seven sections (see Table 2 and Supplementary material). Ratings are made on a 5-point Likert scale from 1 (”very rarely/never”) to 5 (“very frequently/always”) and the scores were calculated as mean values for the items in each scale. The first section (reactivity) assesses how often participants experience different emotions (i.e., anger, fear, sadness, and happiness/excitement). In the next five sections, participants are asked with what frequency they use different emotion regulatory strategies. Each of these sections represents one of the following five strategies included in Gross’ model: situation selection, modification, attentional deployment, reappraisal, and suppression. In addition to asking the participants how often they used different strategies, the CERI includes items assessing the extent to which the participants have problems using a specific strategy. This section is referred to as “implementation.” The last section consists of questions asking participants whether their emotions pose a problem that it has a negative impact on their daily life in general or with regard to friendships, work/studies, or romantic relationships; this section is referred to as “negative impact.” The CERI is freely available for both researchers and clinicians, and the complete questionnaire is presented as supplementary material.

Table 2. Description of the different sections included in the CERI.

Name	# items	Example item
Emotional reactivity	4	“I often feel angry/sad/scared/excited”
Situation selection	5	“I avoid situations that could make me angry”
Modification	5	When I get scared, I change the situation to calm myself down”
Attentional deployment	5	“When I get sad, I think of something else to calm myself down”.
Reappraisal	5	“When I get too excited, I re-evaluate the situation to calm myself down”.
Suppression	5	“When I get angry, I keep it to myself”
Implementation	7	“I find it difficult to regulate my emotions by thinking of something else”.
Negative Impact	6	“My difficulties with regulating emotions pose a problem in my daily life”.

Emotion Regulation Questionnaire

The Emotion Regulation Questionnaire (ERQ; Gross & John, 2003) was included to enable investigation of convergent validity between an already available questionnaire and the newly developed CERI. The ERQ consists of 10 items rated on a 7-point scale from 1 (strongly disagree) to 7 (strongly agree). It measures two specific strategies for regulating emotions: (1) Cognitive reappraisal (six items; e.g., “When I want to feel less negative emotion such as sadness or anger, I change what I’m thinking about”) and (2) Suppression (four items; e.g., “I control my emotions by not expressing them”). The scores used in the present study were mean values for the two subscales.

Statistical analysis

First, a factor analysis was conducted including the items meant to capture the five strategies for regulating emotions (i.e., selection, modification, attention, reappraisal, and suppression). Based on the fact that the factors all measured different types of emotion regulation, they were expected to be significantly correlated and we, therefore, chose to use an oblique factor rotation (i.e., Direct Oblimin). The number of factors was determined by investigating the scree test and the number of factors with an eigenvalue ≥1. In addition, parallel analysis was run utilizing the rawpar.sps script developed by O’Connor (2000). Factor loadings greater than .40 were considered to load on each respective factor. As recommended by, for example, Hu and Bentler (1999), a combination of fit indices should be used. In the present study, we report the following indices: (1) normed chi-square (χ²/df), (2) the standardized root-mean-square residual (SRMR), (3) the root-mean-square error of approximation (RMSEA), (4) the comparative fit index (CFI). Model fit is considered as acceptable if the χ²/df value is below 5, the values of the SRMR and the RMSEA are close to or below 0.08 and the value of the CFI is greater than 0.90 (>0.90 acceptable, >0.95 excellent).

Next, intraclass correlations (2-way mixed effects model, absolute agreement) were used to investigate test–retest reliability for the factors obtained from the factor analyses, as well as for the measures of emotional reactivity, implementation, and negative impact. According to guidelines (Cicchetti, 1994), an ICC < .50 is considered to be poor, .50 to .74 moderate, .75 to .89 good and ≥ .90 excellent. As a further measure of reliability, internal consistency was investigated. Due to the low number of items in each scale, mean inter-item correlations were used instead of Cronbach’s alpha. According to guidelines, adequate internal consistency is found when the mean inter-item correlation range between .15 and .50 (e.g., Clark & Watson, 1995). Internal consistency was not measured for emotional reactivity as reactivity of different emotions (including both negative and positive emotions) cannot be expected to measure the same construct and the four items of emotional reactivity were therefore included as separate measures in all analyses.

With regard to validity, we first examined correlations between the CERI and the ERQ, a well-established instrument targeting emotion regulation. The two ERQ subscales measure only reappraisal and suppression. Thus, by investigating correlations between the two ERQ subscales and all of the subscales in the CERI, we were able to investigate both convergent and divergent validity. Spearman correlations were used to investigate associations between the use of different strategies for regulating emotions and the impact emotion dysregulation has on daily life.

Finally, we wanted to investigate associations with ADHD. Several of the CERI subscales were significantly correlated with age, rs ranging between .02 and .34. Analyses of Covariances (ANCOVAs), using age as a covariate, were therefore used to compare patients diagnosed with ADHD and adults without any psychiatric disorder. However, it should be noted that results did not differ with or without control for age. Effect sizes for the group differences were calculated using Cohen’s d, where .20 is considered a small effect, .50 a medium-sized effect, and .80 a large effect (Cohen, 1988). Paired-samples t-tests, with separate analyses conducted for adults with ADHD and the comparison group, were also used to examine whether the generally less effective strategy of suppression was used significantly more often than any of the other four strategies. Finally, logistic regression analysis was conducted to investigate which CERI subscales contributed independently (i.e., when controlling for the overlap between subscales) to discriminating between individuals with ADHD and the comparison group. Before conducting the regression analysis, Variance Inflation Factors (VIF) were computed to check for multicollinearity as the CERI subscales were expected to be inter-correlated.

Results

Factor analysis

The Kaiser-Meyer-Olkin measure of sampling adequacy was .84, and Bartlett’s test of sphericity was significant, χ² = 9545.97, p < .001. Thus, the data appeared suitable for conducting a factor analysis. The results showed that seven factors emerged with an eigenvalue ≥1. However, the scree plot indicated that a five-factor solution could be the best option and the parallel analysis also showed five significant factors. The 7-factor solution did not provide a good fit for the data as several items loaded onto more than one factor. A 5-factor solution was then examined, and this solution appeared to fit the data well, except that one item (i.e., suppression of happiness/exuberance) did not show a high loading on any of the factors. However, when removing this item, five clear factors were found, each representing one of the five strategies included in Gross model (see Table 1). This solution explained a total of 53.91% of the variance. All fit indices indicated an acceptable fit with χ²/df = 2.77, CFI = .92, SRMR = .05, and RMSEA = .07.

Reliability

When examining test–retest reliability, intraclass correlations indicated that the reliability of the CERI was good for all five emotion regulation strategies: situation selection (.78), modification (.82), attentional deployment (.80), reappraisal (.78), and suppression (.77). Test–retest reliability was also good for implementation (.82) and negative impact (.89). With regard to emotional reactivity, the results showed that the test–retest reliability was good for sadness (.84) and happiness/exuberance (.80), but questionable for anger (.74) and fear (.69).

The internal consistency, measured with mean inter-item correlations, was shown to be good for all five emotion regulation strategies: situation selection (r = .41), modification (r = .47), attentional deployment (r = .48) and reappraisal (r = .47), and suppression (r = .41). The items measuring implementation (r = .49) and negative impact (r = .41) also showed high mean inter-item correlations.

Convergent and divergent validity

To examine convergent and divergent validity, correlations were calculated between the two subscales from Gross’ instrument ERQ and all of the CERI subscales. Results showed strong convergent validity, as the two subscales measuring reappraisal were strongly related to one another (r = .57, p < .001) as were the two subscales measuring suppression (r = .65, p < .001). There was also support for divergent validity as the two subscales from the ERQ were not significantly related to any of the other CERI subscales, except for a modest, although significant, correlation between the ERQ reappraisal and the CERI modification subscales, r = .30, p < .05.

Interrelations between use of strategies and impact on daily life

When investigating associations between the CERI subscales measuring the use of different strategies and the subscale negative impact, results showed that, among adults with ADHD, higher negative impact on daily life was associated with lower use of the strategy modification (r = – .23, p < .01), attentional deployment (r = – .28, p < .001) and reappraisal (r = – .29, p < .001). Among the individuals in the comparison group, high negative impact was found to be associated with low use of modification (r = – .19, p < .01) and reappraisal (r = – .17, p < .01), but high use of situation selection (r = .22, p < .001).

Group differences between ADHD and the comparison group

The results of the ANCOVAs showed that significant group differences were found for all five strategies, except for suppression. However, as can be seen in Table 4, the direction of the effects varied. For situation selection, patients with ADHD reported using this strategy more often compared to individuals in the comparison group, whereas they reported using modification, attentional deployment, and reappraisal significantly less often. In addition, patients with ADHD reported significantly more problems with implementing strategies for regulating emotions as well as higher negative impact of emotion dysregulation on daily life. Effect sizes were small to medium for all significant effects related to the use of strategies, but well above the level considered a large effect for implementation and negative impact.

Table 3. Results of the factor analysis for the CERI. Factor loadings smaller than .30 are not displayed in the table.

		Factor
A priori scale	Item description	1	2	3	4	5
E	When I get sad, I reassess the situation to calm myself down	.90
E	When I get upset, I reassess the situation to calm myself down	.75
E	When I get scared, I reassess the situation to calm myself down	.74
E	When I get angry, I reassess the situation to calm myself down	.72
E	When I get too excited, I reassess the situation to calm myself down	.57
B	I choose to avoid situations that may make me angry		.90
B	I choose to avoid situations that may upset me		.86
B	I choose to avoid situations that may make me sad		.60
B	I choose to avoid situations that may make me scared		.55
B	I choose to avoid situations that may make me too excited		.41
C	When I get angry, I change the situation to calm myself down			.93
C	When I get upset, I change the situation to calm myself down			.90
C	When I get scared, I change the situation to calm myself down			.61
C	When I get scared, I think about something else to calm myself down			.58
C	When I become too excited, I change the situation to calm myself down			.42
F	I keep my negative emotions to myself	.			.73
F	When I get angry, I keep it to myself				.65
F	When I get sad, I keep it to myself				.60
F	When I get scared, I keep it to myself				.53
D	When I get upset, I think about something else to calm myself down					.83
D	When I get angry, I think about something else to calm myself down					.72
D	When I get scared, I think about something else to calm myself down					.54
D	When I become too excited, I think about something else to calm myself down					.52
D	When I get sad, I think about something else to calm myself down					.46
	Explained variance	27.28	11.20	5.79	5.25	4.39

B = Situation selection; C = Modification; D = Attentional deployment; E = Reappraisal; F = Suppression.

Table 4. Means (M), standard deviations (SD), results of ANCOVAs, and effect sizes (d) comparing individuals with ADHD and the comparison group on all CERI subscales.

	ADHD M (SD) (n = 130)	Comparison group M (SD) (n = 260)	Group F	d	Effect sizes5% CI
Selection	3.01 (.82)	2.54 (.83)	21.16***	.57	.35 –.78
Modification	2.72 (.89)	3.06 (.92)	8.90*	.37	.16 –.58
Attention	2.37 (.90)	2.92 (.83)	35.19***	.64	.43 –.86
Reappraisal	2.48 (.98)	3.04 (.82)	32.86***	.64	.42 –.85
Suppression	2.87 (1.01)	2.86 (.77)	0.01	.01	−. 20 −.20
Happiness/excitement	1.88 (.98)	2.30 (.96)	21.16***	.43	.22 –.65
Problems with implementation	3.31 (.89)	2.25 (.76)	127.57***	1.31	1.08–1.54
Negative impact on daily life	2.75 (.92)	1.49 (.67)	203.66***	1.66	1.42–1.90
Reactivity
Anger	3.30 (1.16)	2.71 (.90)	26.57***	.60	.39 –.82
Fear	2.43 (1.24)	2.28 (1.16)	0.44	.13	−.08 –.34
Sadness	3.47 (1.10)	2.63 (.90)	48.06***	.87	.65–1.09
Happiness	3.58 (1.07)	3.32 (.96)	12.06***	.26	.05 –.47

*p < .05, *** p < .00.

When investigating within-group differences, paired t-tests showed that adults with ADHD used suppression significantly more often compared to attentional deployment and reappraisal (both ts > 3.40, p < .001). The use of suppression did not differ significantly from the use of situation selection or modification. In contrast, adults in the comparison group used suppression less often than modification and reappraisal (ts > 2.74, p < .01), but more often than situation selection (t = 5.08, p < .001).

As suppression of happiness/exuberance was not included in the suppression subscale based on the results from the factor analysis, a separate analysis examined the group difference for this variable was conducted. The results showed that adults with ADHD suppressed their happiness/exuberance significantly less than the individuals in the control group, although the effect size was small (see Table 4). As can also be shown in Table 4, adults with ADHD had significantly higher emotional reactivity with regard to anger, sadness, and happiness, whereas the group difference was not significant for reactivity of fear.

Finally, logistic regression analysis was conducted to investigate which CERI subscales contributed independently (i.e., when controlling for the overlap between subscales) to discriminating between adults with ADHD and adults without a psychiatric disorder. As the group difference for the subscale suppression had not been significant, only suppression of happiness/exuberance was included in the regression analysis. The variance inflation factors (VIF) were all <2.4, which indicated that multicollinearity was not a problem in this analysis. Results showed that the model was significant (χ² = 187.92, p < .001, R² = .54) and the following subscales contributed independently: situation selection (Wald = 5.55, p < .05), suppression of happiness/exuberance (Wald = 11.58, p < .001) and negative impact (Wald = 36.05, p < .001). There were also marginally significant effects for attentional deployment (Wald = 3.26, p = .07) and implementation (Wald = 3.03, p = .08). However, the effects for modification (Wald = 1.06), and reappraisal (Wald = .13) were not significant.

Discussion

The aim of the present study was to investigate emotion dysregulation in adults with ADHD using a newly developed self-rating instrument called the Comprehensive Emotion Regulation Inventory (CERI). The CERI included questions related to the five different strategies for regulating emotions included in Gross (2001) process model of emotion regulation (i.e., situation selection, modification, attentional deployment, reappraisal, and suppression). However, factor analysis identified three factors, with situation selection and suppression forming two separate factors and the other three strategies forming one common factor. With regard to the psychometric properties of the CERI, results showed adequate to good internal consistency and high test–retest reliability for all subscales. Convergent validity was established by showing high correlations between the ERQ and corresponding subscales within the CERI. In support for divergent validity, the ERQ subscales were not significantly related to the other CERI subscales, except for a significant correlation between ERQ reappraisal and CERI modification. When investigating group difference, individuals with ADHD used selection more often than individuals in the comparison group, but modification, attentional deployment, and reappraisal less often, although the effect sizes were small to medium. However, large effect sizes were found between adults with ADHD and adults without any psychiatric disorder with regard to problems implementing the different regulatory strategies and negative impact on daily life functioning. Lower use of the more cognitively challenging strategy modification, attentional deployment, and reappraisal was significantly related to a higher negative impact of emotion dysregulation on daily life.

Reliability of the CERI

With regard to reliability, all CERI subscales showed comparable or higher internal consistency and higher test–retest reliability compared to other emotion regulation questionnaires such as the ERQ (Gross & John, 2003). The four items measuring emotional reactivity were not considered a scale as it cannot be expected that a few items of reactivity, including both positive and negative emotions, should be highly correlated. In addition, previous research points to the importance of using separate assessment of negative and positive emotional reactivity (e.g., Becerra et al., 2019). The relatively low test–retest reliability for reactivity of anger and fear suggests that it is not advisable to investigate emotional reactivity as briefly as done in the present study. Thus, the CERI should be regarded as an instrument for emotion regulation that needs to be complemented with other instruments for assessing emotional reactivity.

CERI does not yet exist in a version for significant others, and it was therefore not possible to examine inter-rater reliability. It could be argued that emotion regulation is a construct that is best rated by the individual in question. Still, it would be of value for future research to also include ratings provided by a significant other, especially because previous studies of adults examining other neuropsychological constructs have found that the correlation between raters is often modest (e.g., Holst & Thorell, 2018; Roth et al., 2005)

Factorial validity of the CERI

Emotional functioning is an umbrella term that includes various more or less related constructs such as emotional lability, emotional intensity, emotional reactivity, emotion recognition, and emotion regulation. As argued by, for example, Shaw et al. (2014), future research needs to be much more precise in operationalizing the constructs under study. For this reason, the CERI was created based on several different subscales, and factor analyses showed that it was possible to distinguish between the different emotion regulation strategies included in Gross (2001) process model. It is interesting to note that it was possible to distinguish between all five strategies using the CERI, which is a relatively brief instrument. That said, it should be noted that emotion dysregulation is a complex construct. After identifying patients having poor emotion regulation using the CERI, it will therefore most likely be necessary to obtain even more detailed information about different strategies for regulating emotions using another format than self-report. The Emotion Regulation Interview (ERI; Werner et al., 2011) was developed to examine the frequency and self-efficacy with which emotions are regulated based on both a laboratory speech task and recent real-life situations. To our knowledge, this interview has not been used with patients with ADHD. However, it should be possible to use in other patient groups, providing relevant tasks can be identified that trigger strong emotions in a laboratory setting.

ADHD and emotion dysregulation

The results of the present study once again confirm that ADHD is related to emotion dysregulation, even in adults. However, the effect sizes were relatively small with regard to how often the participants used different strategies, and the direction of the effect also varied with regard to the five different strategies. Concerning situation selection (i.e., avoiding an emotionally challenging situation), individuals with ADHD rated more frequent use of this strategy than the individuals in the comparison group did. This could be interpreted as higher capacity for emotion regulation among the ADHD group compared to the comparison group, which is not in line with expectations. However, as emphasized by, for example, Gross and Jazaieri (2014), we regulate our emotions to elicit feelings and behaviors that correspond with our goals for a given situation. Thus, using avoidance could be an adequate tactic if, in relation to our goals, the potential costs of attending the situation are outweighed by the potential benefits. When investigating how often different strategies are used, the CERI does not distinguish between adaptive and maladaptive avoidance. However, this instrument has the advantage of also including questions related to problems with implementing specific strategies. With regard to situation selection, the CERI includes both a question related to problems with avoiding situations, even if one wishes to do so, and a question related to foreseeing what situations could cause strong reactions. Individuals with ADHD clearly had much more difficulties with both these aspects compared to the individuals in the comparison group. Not knowing what situations to avoid means that individuals with ADHD are more often taken by surprise if they choose not to avoid a situation, making emotion regulation at a later stage (i.e., through the use of modification, attentional deployment, reappraisal, or suppression) more difficult. In sum, the type of avoidance that characterizes adult ADHD is largely problematic. This has been acknowledged within some cognitive behavior therapy (CBT) programs for adults with ADHD by including strategies for identifying barriers to acting, thereby attempting to decrease avoidance of challenging situations (review by Sprich et al., 2010).

Previous research has emphasized that emotion regulation is not just a matter of how much emotions are regulated, but also of what strategies are used, and individuals with mental health problems are less likely to use the strategies attentional deployment and reappraisal (e.g., Gross & John, 2003). These strategies include diverting one’s attention to something else to change the emotional impact of a situation or reinterpreting the meaning or importance of an emotionally challenging situation. Interestingly, lower use of these two strategies was found to be related to higher negative impact on daily life functioning. In addition, the present study showed that individuals with ADHD reported using suppression significantly more often than reappraisal and attentional deployment, whereas the opposite pattern was found among the individuals in the comparison group. These findings could be said to be in line with previous research showing that higher use of cognitive reappraisal is related to a range of positive outcomes such as closer interpersonal relationships, higher well-being, higher life satisfaction, better self-esteem, more optimism, and lower levels of depression (Gross & John, 2003). On the other hand, high levels of suppression have been found to be significantly related to lower levels of these positive outcomes. Gross and John (2003) argued that suppression should be regarded as a less effective strategy compared to modification, attentional deployment, and reappraisal, but it should be acknowledged that the more effective strategies also require higher cognitive skills. Previous research has shown that emotion dysregulation overlaps with other neuropsychological deficits (e.g., poor inhibitory control, working memory deficits, and delay aversion) in relation to ADHD in both children (Sjöwall et al., 2013; Sjöwall & Thorell, 2019) and adults (Sjöwall & Thorell, in press). Despite some overlap, these studies also show that emotion regulation is independently related to ADHD when controlling for the effect of both “cool” (e.g., working memory and inhibitory control) and “hot” (e.g., delay aversion) executive functions. However, there is a need for more research addressing how different strategies for regulating emotions are related to models of executive control (e.g., Petrovic & Castellanos, 2016; Zelazo, 2020), both in relations to different mental health conditions and daily life functioning.

Limitations, conclusions and future directions

First, it should be acknowledged that CERI is a self-report instrument and as such it has the limitations always associated with this type of measurement, the most important of which is probably rater bias. In addition, it would be naïve to think that we can truly capture all the complexities of emotion regulation using self-rating. Thus, it is our belief that the CERI should primarily be used as a screening tool for identifying difficulties that need to be further examined or as a research tool for trying to get a more nuanced picture of emotion dysregulation in either clinical or non-clinical samples. One important aspect to examine further is how different emotion regulation strategies are used to achieve a specific goal. In addition, it has been emphasized in previous research (e.g., Fernandez et al., 2016) that efforts to regulate emotion do not end with implementing a certain strategy. For successful emotion regulation, monitoring of the situation is essential, and flexibly in using several different emotion regulation strategies to achieve a more beneficial long-term goal is often needed.

An important avenue for future research would be to investigate what aspects of emotional dysregulation are shared between people with different mental disorders and what aspects might be disorder specific. Identifying subgroups of ADHD patients with different types of problems should also be considered important. In addition, previous research has emphasized the need to focus on transdiagnostic factors that may contribute to the development and maintenance of different kinds of mental health problems (e.g., Aldao et al., 2010; Buckholtz & Meyer-Lindenberg, 2012).

In conclusion, the present results show that the CERI appears to be a reliable and valid self-rating instrument that can be used for obtaining a more detailed understanding of emotion regulation problems in adults. The study adds valuable information by showing that adults with ADHD have great problems with implementing various emotion regulation strategies as well as revealing the negative impact emotion dysregulation has on daily life functioning. We, therefore, believe that the CERI could be valuable both in research and in clinical practice as a screening tool for assessing what areas of emotion dysregulation would be good treatment targets.

Acknowledgments

This study was supported by a grant from the Swedish Research Council to the first author.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental data

Supplemental data for this article can be accessed here.

Additional information

Funding

This work was supported by the Vetenskapsrådet [2017-01508].