Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2022 Sep 1.
Published in final edited form as: J Clin Child Adolesc Psychol. 2020 Jan 7;50(5):596–608. doi: 10.1080/15374416.2019.1703710

Psychometric Properties of the Emotion Dysregulation Inventory in a Nationally Representative Sample of Youth

Carla A Mazefsky a,*, Lan Yu b, Paul A Pilkonis a
PMCID: PMC7781089  NIHMSID: NIHMS1546413  PMID: 31910035

Abstract

Objective:

The Emotion Dysregulation Inventory (EDI) is an informant questionnaire developed based on the Patient-Reported Outcomes Measurement Information System (PROMIS®) Scientific Standards and refined through factor analyses and item response theory (IRT) analyses. Although it was developed to improve measurement of emotion dysregulation in youth with autism spectrum disorder, emotion dysregulation has transdiagnostic significance. Therefore, the aim of this study was to evaluate the EDI’s psychometric properties and to establish IRT-based scores for a general population of youth.

Methods:

Data were collected from a sample of 1000 caregivers of 6- to 17-year old youth matched to the US census on age, gender, race/ethnicity, years of education, and region. Confirmatory factor analyses and IRT analyses using the two-parameter graded response model were performed to evaluate the EDI’s structure and psychometric properties.

Results:

Analyses supported the original two-factor structure of the EDI, reflecting factors for Reactivity and Dysphoria. Simulations of computerized adaptive testing supported use of the same items for a Reactivity short form as those that emerged as most informative in the original autism psychometric analyses. IRT co-calibration with commonly used measures of emotion regulation and irritability in child clinical or community samples indicated the EDI scales provide more information across a wider range of emotion dysregulation. Validity was supported by moderate correlations with measures of related constructs and expected known-group differences.

Conclusions:

The EDI is an efficient and precise measure of emotion dysregulation for use in general community and clinical samples as well as samples of youth with ASD.

Keywords: Emotion Dysregulation, Irritability, Dysphoria, Reactivity, Item response theory

Introduction

Study of emotion regulation is prominent in psychology, psychiatry, and related disciplines (Gross, 2015). While several emotion regulation theories focus on specific strategies (Gross & Thompson, 2007), others focus on emotion dysregulation—general deficits in one’s ability to modulate the intensity or duration of emotional responses (Sloan et al., 2017). Emotion dysregulation has been implicated as a core feature or associated characteristic of nearly all psychiatric disorders, maladaptive behaviors, and persistent irritability (Aldao, Nolen-Hoeksema, & Schweizer, 2010; Compas et al., 2017; Leibenluft, 2011). As such, emotion dysregulation is a common mechanistic target or treatment outcome in clinical trials for various forms of child and adolescent psychopathology (Gratz, Weiss, & Tull, 2016).

Youth with autism spectrum disorder have high risk for severe and impairing emotion dysregulation (Joshi et al., 2018). However, a lack of validated measures of emotion dysregulation that are suitable for use in autism spectrum disorder has been a barrier to research and treatment development focused on emotion dysregulation in this population (Weiss, Thomson, & Chan, 2014). Primary concerns include whether established measures of emotion dysregulation are appropriate for youth with highly variable verbal and cognitive abilities and whether they are sufficiently sensitive to the range of observed emotion dysregulation in autism spectrum disorder (Mazefsky, Taylor et al., 2018). Widely used emotion dysregulation measures have been most extensively studied in typically-developing populations (Adrian, Zeman, & Veits, 2011). For some measures, the primary psychometric sample excluded youth with autism, with only limited representation (e.g., n = 5) of youth with autism in a secondary sample (e.g., Stringaris et al., 2012). Finally, many of the emotion dysregulation measures with good psychometric properties outside of autism were designed to measure stable characteristics or traits and are not ideal as outcome measures in the context of clinical trials.

Therefore, we developed and validated the Emotion Dysregulation Inventory (EDI), a change-sensitive, informant-report measure of emotion dysregulation in a sample of children and adolescents with autism spectrum disorder (Mazefsky, Day, et al., 2018; Mazefsky, Yu, et al., 2018). The EDI was created using the Scientific Standards of the Patient-Reported Outcomes Measurement Information System (PROMIS®). PROMIS is a National Institutes of Health Roadmap Initiative to develop outcome measures using state-of-the-art methods. The PROMIS Scientific Standards outline steps for item generation, item refinement, and psychometric analyses, including the use of models from item response theory (IRT; PROMIS Standards Committee, 2013). IRT models provide information about how well single items, as well as a complete “bank” of items, discriminate between people with differing degrees of severity of the construct being measured (Reeve et al., 2007).

Several steps were used to generate and refine the EDI item pool (Mazefsky, Day, et al., 2018). First, a comprehensive literature review was conducted to identify established measures related to emotion dysregulation. Next, a conceptual model was developed based on the first author’s clinical experience with emotionally dysregulated youth with autism and research on emotion regulation in autism. Because emotion regulation research in autism was limited at the time the EDI was developed, the conceptual model was also heavily informed by the literature on emotion regulation outside of ASD.

Once items were drafted, they were assigned to specific facets in the conceptual model, including increased negative affect, decreased positive affect, disturbed behavior, decreased vitality, nervousness and fear, hyperarousal, rapid emotion escalation, emotional intensity, emotional lability, sustained emotional reactions, and poor modulation of emotional duration or intensity. Only observable indicators of emotion dysregulation were included, and items related to the communication of affective experiences to others or socially-mediated aspects of emotion regulation (e.g., seeking help when distressed) were omitted in order to avoid potential confounds with verbal ability or social impairments.

Once the initial 67-item pool was developed, interviews were completed with 19 parents of youth and adults with autism spectrum disorder to assess their understanding of the items and their decision-making processes when selecting their responses (Mazefsky, Day, et al., 2018). Information generated from these interviews was utilized to revise the items, directions, and response options. No items were added at this stage, though some wording was modified, and one was deleted. The complete draft of the EDI was reviewed by a panel clinicians and researchers with expertise in measure development, emotion dysregulation, and/or autism spectrum disorder from more than 10 academic institutions.

Caregivers of 1,755 youth with autism spectrum disorder (ages 4–20, though 98.2% were 6–17; M = 12.3, SD =3.2) completed the 66 candidate EDI items used for field testing (Mazefsky, Yu, et al., 2018). To ensure that the EDI would be sensitive to the most extreme forms of emotion dysregulation, participants from specialized autism psychiatric inpatient units were recruited from the Autism Inpatient Collection (n = 432; Siegel et al., 2015). The remainder of the sample (n = 1323) was recruited from a national autism registry (Interactive Autism Network, 2017) and was intended to be representative of the full population of youth with parent-reported professional ASD diagnoses in the United States. Approximately 28% of the combined sample had co-occurring intellectual disability and 21% were female, consistent with national autism surveillance statistics (Baio et al., 2018).

The final 30 items of the EDI were selected based on classical test theory (factor analysis) and IRT analyses (Mazefsky, Yu, et al., 2018). Both exploratory and confirmatory factor analyses identified two factors. Factor One (Reactivity) included items capturing rapidly escalating, intense, and labile negative affect characterized by anger and irritability, as well as difficulty down-regulating that affect. Factor Two (Dysphoria) included items that reflect sadness, unease, anhedonia, and low motivation. We next conducted two rounds of IRT analysis, examined the two assumptions of IRT (model misfit and local dependency), and eliminated items with low information yield. The final result was a primary scale of 24 Reactivity items and a secondary scale of six Dysphoria items. Simulations of computerized adaptive testing were used to select items for a Reactivity Short Form, which included seven items and was correlated .99 with the full Reactivity item bank. We also used IRT to assess differential item functioning (DIF) based on pertinent participant characteristics. The final EDI items did not have any DIF based on gender, age, intellectual ability, or verbal ability.

The validity of the EDI was supported by expert review, expected group differences (higher scores in a psychiatric inpatient sample versus an autism spectrum disorder community sample), and expected correlations with measures of related constructs. Further, the EDI was shown to have strong test-retest reliability in youth with stable treatments over a 4-week period, as well as change-sensitivity from admission to discharge in psychiatric inpatients. Later work demonstrated the EDI’s change-sensitivity in response to an emotion regulation-focused psychosocial treatment in adolescents with autism spectrum disorder (Conner et al., 2018).

Although the EDI was initially developed to address a gap in available measures for autism research and clinical care, there is also a need for sensitive and psychometrically sound measures of emotion dysregulation for use in other clinical and community populations. For example, a lack of high quality measures has been identified as a major barrier in studies of irritability (Vidal-Ribas, Brotman, Valdivieso, Leibenluft, & Stringaris, 2016). By developing the EDI to be sensitive to the full range of emotion dysregulation in autism spectrum disorder, it is possible that the EDI has advantages in terms of breadth of content coverage and precision across a wide range of emotion dysregulation severity.

There are many potential applications of measures of emotion dysregulation that are psychometrically robust across populations. Because emotion dysregulation is implicated transdiagnostically and associated with high comorbidity (Aldao, 2016; Sloan et al., 2017), it may be beneficial to focus on it dimensionally across diagnoses (Kelly, Clarke, Cryan, & Dinan, 2018). Further, because emotion dysregulation is a prominent feature of many psychiatric disorders, measures of emotion dysregulation have potential applicability to disorder-specific studies. Finally, identifying trajectories of emotion dysregulation may provide insight into etiological pathways to psychopathology (Ip, Jester, Sameroff, & Olson, 2019).

Therefore, this study aimed to evaluate the EDI’s psychometric properties in a large sample of US census-matched youth. Analyses were conducted to evaluate its factor structure, item properties, precision, and validity in comparison to commonly used measures of emotion dysregulation in general child psychology and psychiatry. We hypothesized that the EDI factor structure found with youth with ASD would be replicated, that the items would have high information yield in a sample representative of US children and adolescents, and that its validity would be supported by expected correlations with measures of related constructs and differences between known groups.

Methods

Participants

Participants were recruited through YouGov, a global public opinion and polling company. YouGov has over two million registrants in the US, thereby providing a feasible means to collect a large, nationally representative sample. The 1,055 respondents that YouGov initially contacted were reduced to a sample of 1,000 to produce the final dataset representative of the US census from a non-randomly selected pool of respondents. The sample represents the US population of adults with related children in the household, ages six to 17. The matched cases were weighted to the sampling frame using propensity scores that included age, gender, race/ethnicity, years of education, and region in ratios derived from full 2016 American Community Survey (United States Census Bureau, 2016). The propensity scores were grouped into deciles of the estimated propensity score in the frame and post-stratified according to these deciles. The weights were then post-stratified on a four-way stratification of gender, age (four categories), race (four categories), and education (four categories), to produce the final weight. The matched cases and the frame were combined, and a logistic regression was estimated for inclusion in the frame.

The final sample had a mean age of 12.1 (SD = 3.6) and was 49.3% female (n = 492). In terms of ethnicity, 17.5% were Hispanic (n = 175); 78.9% (n = 651) were White and 17.6% (n = 145) were African-American. The final sample was invited to complete a questionnaire battery including the EDI online. They received YouGov points for their participation, which are redeemable for an incentive of their choosing, such as gift cards, cash via PayPal, or a charity donation.

Measures

Emotion Dysregulation Inventory (EDI; Mazefsky, Day, et al., 2018, Mazefsky, Yu, et al., 2018).

The EDI is an informant report measure of emotion dysregulation over the past seven days, rated on a five-point scale from “not at all” to “very severe.” The EDI produces scores for Reactivity (rapidly escalating, intense, and poorly regulated negative affect characterized by anger/irritability) and Dysphoria (sadness, unease, low motivation, anhedonia). Reactivity scores are available from a 24-item bank or seven-item short form and the Dysphoria scale is six items. Cronbach’s alpha internal consistency was .97 for Reactivity, .92 for the Reactivity short form, and .90 for Dysphoria in this sample. The EDI scales, scoring instructions including tables to convert raw scores to theta or t-scores based on both the original autism and YouGov samples, and interpretive information (e.g., clinical cut-offs based on the YouGov sample) are available without charge from the first author.

Child Behavior Checklist (CBCL; Achenbach & Ruffle, 2000).

The CBCL is a caregiver report of psychiatric symptoms for six- to 18-year-old youth, rated on a three-point scale from “not at all” to “often true or very true” over the past month. It produces several empirically-derived symptom scales as well as scales meant to map onto clinical diagnoses, with t-score conversions based on age and gender. The CBCL is part of the Achenbach System of Empirically Based Assessment (ASEBA) suite of measures which were developed and validated in large, nationally representative samples of youth with careful attention to their factor structure, reliability, and validity. CBCL scales were selected for inclusion in analyses based on their relevance to the EDI scales, including: (a) CBCL Anxious/Depressed diagnostic scale; (b) CBCL Withdrawn/Depressed syndrome scale; and (c) a sum of 3 CBCL items (“temper tantrums or hot temper,” “stubborn, sullen or irritable,” and “sudden changes in mood or feelings”) which is used as an index of irritability and has been shown to be both reliable and valid (Evans et al., 2019). Cronbach’s alpha internal consistency was .87 for CBCL Anxious/Depressed, .83 for CBCL Withdrawn/Depressed, and .81 for CBCL Irritability in this sample.

Affective Reactivity Index (ARI; Stringaris et al., 2012).

The ARI is a widely used six-item scale of chronic irritability (a mood characterized by being easily annoyed or angered and temper outbursts) over the past six months, rated on a three-point scale from “not true” to “certainly true.” The ARI produces a single irritability score with strong internal consistency and evidence of reliability. Evidence for the ARI’s validity comes from moderate and significant correlations with emotional and conduct problems. Further, the ARI has evidence of known-groups validity, with scores for those with severe mood dysregulation (characterized by irritability) the highest, followed in order by bipolar disorder, those with a family history of bipolar disorder, and then healthy volunteers. Cronbach’s alpha internal consistency was .90 for the ARI in this sample.

Emotion Regulation Checklist (ERC; Shields & Cicchetti, 1997).

The ERC is a 24-item scale rated on a scale of one (never) to four (always). The ERC yields two subscales: (a) Lability/Negativity, which assesses inflexibility, lability, and dysregulated negative affect, and (b) Emotion Regulation, which measures appropriate emotional expression, empathy, and emotional self-awareness. It was developed using Q-sort methodology, and its initial validity was documented in a study of maltreated and non-maltreated children, including assessment of convergent, divergent, and construct validity. Cronbach’s alpha internal consistency was .86 for Lability/Negativity and .76 for Emotion Regulation in this sample.

Psychometric Analysis

Confirmatory factor analysis.

The goal of the factor analyses was to evaluate whether the original EDI factor structure was replicated in a general population, allowing use of the same version in non-ASD samples. Therefore, we focused on confirmatory factor analyses as opposed to evaluation of alternate factor structures. We expected that the two-facture structure of the 24-item EDI Reactivity (EDI-R) scale and the 6-item EDI Dysphoria (EDI-D) scale would be confirmed in the YouGov sample. In order to test this hypothesis, we performed confirmatory factor analysis on EDI-R and EDI-D using Mplus 7.0 (Muthén & Muthén, 2007). Fit indices including RMSEA, CFI, and TLI and factor loadings were used to evaluate the unidimensionality assumption.

IRT calibration.

After the CFAs of EDI-R and EDI-D, we evaluated the item-level properties of the EDI-R and EDI-D. First, we calculated frequency distributions for each item across response categories. Because the highest response options (severe and very severe) were endorsed at a relatively low rate (see Results), we performed two rounds of separate IRT calibrations: a) first, based on the original five response categories, and b) then based on four response categories by collapsing the top two response categories. These analyses were intended to determine whether there was any benefit to combining the two highest response options. EDI items were calibrated with the two-parameter graded response model (GRM; Reise & Yu, 1990) using IRTPRO 3.1(Cai, Thissen, & Du Toit, 2011).

Static short form development.

To test if the same seven short form items selected for EDI-R based on prior work in the clinical sample (Mazefsky, Yu, et al., 2018) were still appropriate in the YouGov sample, we performed separate computerized adaptive testing (CAT) simulations. We used the same four criteria to rank order EDI items: discrimination parameters, the percentage of times the item would have been selected in a simulated CAT using the YouGov sample, expected information under the standard normal distribution with a mean of zero and SD of one, and expected information under a normal distribution with a larger SD (i.e., a mean of zero and SD of 1.5). The CAT simulations were performed using the Firestar program (Choi, 2009).

Concurrent calibrations with established measures.

In order to examine how established measures of related constructs performed on the same latent trait scale of EDI-R and EDI-D, we fixed the final item parameters for EDI-R and EDI-D based on the YouGov sample, and calibrated the relevant established measures with these fixed parameters using the GRM. Specifically, we co-calibrated the Emotion Regulation Checklist - Emotion Regulation (ERC-ER), Affective Reactivity Index (ARI), Emotion Regulation Checklist -Emotional Lability/Negativity (ERC-L/N), and CBCL Irritability (CBCL-Irr) with EDI-R and its short form. We co-calibrated the ERC-L/N, ERC-ER, CBCL Anxious/Depressed (CBCL A/D), and CBCL Checklist Withdrawn/Depressed (CBCL W/D) with EDI-D.

Validity Analyses.

To evaluate convergent validity, we correlated IRT-calibrated theta scores on the EDI-R, EDI-R-SF, and EDI-D with the established measures. We compared means between groups expected to differ via analysis of variance. The following groups were expected to have higher EDI scores than those not in these groups: participants with a history of one or more psychiatric hospitalizations (n = 49), participants who had an in-home evaluation, emergency room visit, or police contact for a behavioral crisis in the past two months (n = 46), participants who were currently taking an antipsychotic, mood stabilizer, or antidepressant medication (n = 59), participants who were currently in individual or group therapy (n = 142), participants with current depressive or anxiety disorder diagnoses (n = 112), and participants with current diagnoses of disruptive mood dysregulation disorder, oppositional defiant disorder, or bipolar disorder (n = 47). Cohen’s d was calculated to estimate effect sizes for group comparisons, using the following interpretive guidelines: d = 0.2 is considered a ‘small’ effect size, 0.5 represents a ‘medium’ effect size and 0.8 a ‘large’ effect size (Cohen, 1988; Lakens, 2013). Finally, we performed receiver operating characteristics (ROC) analyses using the six known-groups described above to further support validity (Youngstrom, 2013).

Results

Confirmatory Factor Analysis

As hypothesized, the single-factor CFAs on EDI-R and EDI-D confirmed their strong unidimensionality (see Table 1). For EDI-R, the factor loadings ranged from 0.76 to 0.90. Fit indices also demonstrated unidimensionality: RMSEA = 0.065, CFI = 0.975, and TLI = 0.973. For EDI-D, factor loadings ranged from 0.72 to 0.91. Fit indices for Dysphoria also supported strong unidimensionality: RMSEA = 0.070, CFI = 0.995, and TLI = 0.991.

Table 1.

Factor Loadings from Confirmatory Factor Analyses

EDI Reactivity
Item Stem Estimate SE
EDI42S Seems to be in rage 0.903 0.012
EDI36S Has trouble calming himself/herself down 0.880 0.010
EDI19S Has extreme or intense emotional reactions 0.873 0.011
EDI34S Emotions go from 0 to 100 instantly 0.873 0.011
EDI46S Reactions are usually more severe than the situation calls for 0.867 0.012
EDI59S Easily triggered/upset (you have to walk on eggshells around him/her) 0.862 0.013
EDI21S Hard to calm him/her down when mad or upset 0.861 0.013
EDI24S Reactions are so intense that he/she has had to be removed from an activity or place 0.852 0.016
EDI28S When upset or angry, he/she stays that way for a long time 0.847 0.014
EDI50S Becomes upset without a clear reason 0.837 0.016
EDI3S Has explosive outbursts 0.834 0.013
EDI58S Cannot change his/her mood even with your best efforts 0.834 0.016
EDI53S Difficult to distract if he/she is frustrated or upset 0.831 0.014
EDI4S Cries or stays angry for 5 minutes or longer 0.818 0.014
EDI6S Cannot calm down without help from someone else 0.818 0.016
EDI38 Tense or agitated and unable to relax 0.808 0.016
EDI27S Seems on edge 0.805 0.016
EDI52S Has mood swings 0.800 0.015
EDI10S Destroys property on purpose 0.796 0.022
EDI13S Breaks down (crying, screaming) if told he/she canť do something 0.794 0.017
EDI1S Appears angry or irritable 0.787 0.016
EDI26S Physically attacks people 0.772 0.023
EDI8S Frustrates easily 0.769 0.016
EDI7S Suddenly switches to an opposite emotion (e.g. from sad to happy) 0.762 0.018
EDI Dysphoria
EDI31S Does not seem to enjoy anything 0.912 0.011
EDI43S Very little makes him/her happy 0.884 0.016
EDI64S Appears uneasy through the day 0.873 0.015
EDI63S Seems sad or unhappy 0.871 0.013
EDI57S Not responsive to praise or good things happening 0.840 0.017
EDI51S Refuses to leave the house or go to school or activities unless forced 0.721 0.026
Open in a new tab

Note. Reactivity Short Form items are in boldface.

Item Frequencies

As expected, the item frequency distributions for each EDI-R and EDI-D item for the YouGov sample were more positively skewed than the distribution from the prior work in the autism sample (Mazefsky, Yu, et al., 2018). The combined frequencies of the top two response categories “severe” and “very severe” were less than 5% with the exception of the item, “frustrates easily.” However, responses of “moderate” or higher were commonly endorsed for all Reactivity items, with a range of 35.3% to 47.9% for short form items, and 13% to 71% for the full Reactivity item bank (see Table 2). Approximately 6% to 10% of the sample endorsed the Dysphoria items as “moderate” or higher (see Table 3). The Dysphoria items were rated as “mild” in 13% to 29% of cases, suggesting that they are still relevant but may be less common in more severe manifestations.

Table 2.

Item Parameter Estimates for Reactivity in Descending Order of the Slope Parameter

IRT parameters % of sample ≥ moderate
Item Stem a b1 b2 b3 b4 % (n)
EDI42s Seems to be in rage 3.62 0.98 1.64 2.38 3.34 17.4% (175)
EDI36s Has trouble calming himself/herself down 3.45 0.40 1.39 2.23 3.52 35.3% (354)
EDI19s Has extreme or intense emotional reactions 3.30 0.15 1.21 2.13 2.83 44.9% (449)
EDI34s Emotions go from 0 to 100 instantly 3.30 0.34 1.33 2.14 3.00 37.8% (377)
EDI46s Reactions are usually more severe than the situation calls for 3.21 0.24 1.20 1.90 2.85 41.5% (415)
EDI21s Hard to calm him/her down when mad or upset 3.14 0.23 1.24 2.20 3.14 41.7% (416)
EDI59s Easily triggered/upset (you have to walk on eggshells around him/her) 3.02 0.68 1.48 2.21 2.77 26.7% (266)
EDI24s Reactions are so intense that he/she has had to be removed from an activity or place 2.91 0.88 1.73 2.50 3.13 21.4% (214)
EDI28s When upset or angry, he/she stays that way for a long time 2.87 0.48 1.51 2.38 3.00 33.2% (332)
EDI53s Difficult to distract if he/she is frustrated or upset 2.77 0.39 1.51 2.27 3.12 36.1% (362)
EDI50s Becomes upset without a clear reason 2.74 0.63 1.56 2.57 3.38 28.8% (289)
EDI58s Cannot change his/her mood even with your best efforts 2.73 0.69 1.65 2.46 3.49 27.0% (270)
EDI3s Has explosive outbursts 2.69 0.30 1.33 2.50 3.44 39.6% (396)
EDI4s Cries or stays angry for 5 minutes or longer 2.59 0.06 1.2 2.36 3.60 47.9% (479)
EDI6s Cannot calm down without help from someone else 2.54 0.61 1.52 2.60 3.21 29.9% (297)
EDI52s Has mood swings 2.41 0.04 1.43 2.27 3.15 48.6% (487)
EDI27s Seems on edge 2.39 0.56 1.6 2.61 3.58 31.5% (315)
EDI1s Appears angry or irritable 2.38 −0.45 1.01 2.24 3.17 64.8% (648)
EDI13s Breaks down (crying, screaming) if told he/she can’t do something 2.37 0.40 1.45 2.34 3.44 36.5% (365)
EDI38s Tense or agitated and unable to relax 2.36 0.59 1.74 2.69 3.55 30.7% (307)
EDI10s Destroys property on purpose 2.33 1.20 1.87 2.69 3.77 15.9% (159)
EDI7s Suddenly switches to an opposite emotion (e.g. from sad to happy) 2.21 0.40 1.39 2.64 3.54 36.8% (367)
EDI8s Frustrates easily 2.19 −0.68 0.69 2.03 3.14 70.7% (707)
EDI26s Physically attacks people 2.13 1.40 2.18 3.10 4.04 12.8% (128)
Open in a new tab

Note. Short form items are in boldface. Column a displays the slope parameter (how well the item discriminates between respondents with low or high reactivity). Columns b1- b4 display threshold values for individual responses (low threshold values indicate that the item is sensitive to low severity levels and high threshold values indicate that the item is sensitive to high severity levels).

Table 3.

Item Parameter Estimates for Dysphoria in Descending Order of the Slope Parameter

IRT Parameters % of sample ≥ moderate
Item Stem a b1 b2 b3 b4 % (n)
EDI31s Does not seem to enjoy anything 3.64 0.68 1.46 2.26 3.06 9.6% (96)
EDI43s Very little makes him/her happy 3.22 0.67 1.56 2.36 3.05 9.3% (93)
EDI63s Seems sad or unhappy 3.17 0.35 1.57 2.48 3.08 8.6% (86)
EDI64s Appears uneasy through the day 2.97 0.86 1.87 2.67 3.49 5.6% (56)
EDI57s Not responsive to praise or good things happening 2.59 0.83 1.72 2.65 3.38 8.0% (80)
EDI51s Refuses to leave the house or go to school or activities unless forced 1.81 1.04 1.85 2.89 3.48 9.4% (94)
Open in a new tab

Note. Column a displays the slope parameter (how well the item discriminates between respondents with low or high dysphoria). Columns b1- b4 display threshold values for individual responses (low threshold values indicate that the item is sensitive to low severity levels and high threshold values indicate that the item is sensitive to high severity levels).

IRT Calibrations

In order to examine the impact of the sparse cells on item parameter estimations, we performed two rounds of calibration: a) based on the five original response options, and b) based on four response options by collapsing the top two response categories of “severe” and “very severe” into one category. The discrimination and location parameter estimates for the two rounds were very similar. Therefore, we decided to retain the original five response categories to be consistent with the number of response options that were used in prior work with the autism sample (Mazefsky, Yu, et al., 2018).

The IRT parameters for the YouGov sample using the complete five response options are provided in Tables 2 and 3. The slope parameters (“a” in column one of Tables 2 and 3) measure how well the item discriminates between respondents with low or high reactivity/dysphoria, with higher values suggesting that the item is a stronger indicator of the latent construct. The range of slope parameters in this sample (2.13 to 3.62 for Reactivity and 1.81 to 3.64 for Dysphoria) was comparable to the prior work in the autism sample (1.52 to 3.66 for Reactivity and 1.22 to 3.89 for Dysphoria; Mazefsky, Yu, et al., 2018), and all above the generally accepted value of one or higher. The threshold (b) parameters indicate the location along the latent continuum where an individual is more likely to endorse the higher versus lower response category. Compared with the parameter estimates from the prior work in the autism sample (Mazefsky, Yu, et al., 2018), the discrimination parameter estimates for YouGov sample were smaller and the threshold parameter estimates were shifted to the right. This observation confirmed that the symptoms reported in the YouGov sample were milder than those in the autism sample. The difference in severity between the YouGov sample and the autism sample justify the second, separate IRT calibration on the YouGov sample presented here, even as the content, response options, and factor structure of the EDI remains the same. The YouGov parameters provide meaningful clinical thresholds (e.g., one SD above the YouGov mean) for comparison to expectations among a general population of children and adolescents.

Static Short Form Development

The bolded items in Table 2 denote the seven items selected for the Reactivity short form. Criteria evaluated for selection included: discrimination parameters, the percentage of times the item would have been selected in a simulated CAT using the YouGov sample, expected information under the standard normal distribution with a mean of zero and SD of one, and expected information under a normal distribution with a larger SD (i.e., a mean of one and SD of 1.5). Six of the seven items from the YouGov sample overlapped with the selection from the prior work in the autism sample (Mazefsky, Yu, et al., 2018). We decided to keep the same seven items from the autism sample as the static short form items for the YouGov sample to maintain consistency across populations. The scores from the complete EDI-R item bank and its short form were strongly correlated, r (976) = .95, p < .001.

Concurrent Calibrations with Established Measures

To compare the EDI-R and its short form to the established measures of related constructs, items from ARI, ERC-ER, ERC-L/N, and CBCL-Irr were calibrated concurrently with EDI-R. Similarly, items from ERC-ER, ERC-L/N, CBCL-A/D, and CBCL-W/D were calibrated concurrently with EDI-D. Overall, the EDI-R and EDI-D provided the most information in comparison to established measures of related constructs. Even with fewer items, EDI-R short form and EDI-D provided more information than the corresponding established measures of related constructs. Figures 1 and 2 display the corresponding test information curves. Information values of 10 correspond approximately to a classical test theory reliability of .90. At this threshold, the effective range of measurement for both scales was substantial: Reactivity, −0.75 to + 4 SDs, and Dysphoria, 0 to + 3.75 SDs.

Figure 1. Total test information curves for EDI Reactivity.

Figure 1

Open in a new tab

Note. Theta (x-axis) refers to how well the item differentiates higher versus lower severity on the latent trait (emotional reactivity). Higher total information (y-axis) indicates better precision of the item in measurement along the latent trait. The test information of 10 derived from Item Response Theory on the y axis is roughly equivalent to the reliability of .90 derived from Classical Test Theory. Therefore, the curves above the horizontal line (test information of 10 to reliability of .90) indicate the section on the theta scale has reliability of .90 or above. EDI-R = Emotion Dysregulation Inventory Reactivity; .EDI-R-SF = Emotion Dysregulation Inventory Reactivity Short Form; ARI = Affective Reactivity Index; ERC- ER = Emotion Regulation Checklist – Emotion Regulation Scale; ERC-L/N = Emotion Regulation Checklist – Lability/Negativity Scale; CBCL-Irr = Child Behavior Checklist Irritability Scale.

Figure 2. Total test information curves for Factor 2 (Dysphoria).

Figure 2

Open in a new tab

Note. Theta (x-axis) refers to how well the item differentiates higher versus lower severity on the latent trait (dysphoria). Higher total information (y-axis) indicates better precision of the item in measurement along the latent trait. The test information of 10 derived from Item Response Theory on the y axis is roughly equivalent to the reliability of .90 derived from Classical Test Theory. Therefore, the curves above the horizontal line (test information of 10 to reliability of .90) indicate the section on the theta scale has reliability of .90 or above. EDI-D = Emotion Dysregulation Inventory Dysphoria; ERC-L/N = Emotion Regulation Checklist – Lability/Negativity Scale; ERC- ER = Emotion Regulation Checklist – Emotion Regulation Scale; CBCL A/D = Child Behavior Checklist Anxious/Depressed Scale; CBCL W/D = Child Behavior Checklist Withdrawn/Depressed Scale.

Validity Analyses

Correlations between the EDI scales and established measures were significant and predominantly moderate in magnitude (see Table 4). Comparison of groups with expected differences in EDI scores yielded results in the anticipated direction with large effect sizes. Participants with a lifetime history of one or more psychiatric hospitalizations had significantly higher EDI-R [d = .88, F (1, 998) = 116.55, p < .001] and EDI-D [d = .85, F (1, 998) = .81.64, p < .001] scores than those who had never been psychiatrically hospitalized. In addition, participants who had an in-home evaluation, emergency room evaluation, or police contact for a behavioral crisis in the past two months had significantly higher EDI-R [d = 1.46, F (1, 998) = 100.95, p < .001] and EDI-D [d = 1.28, F (1, 998) = 88.10, p < .001] scores than those who did not. Participants currently taking an antipsychotic, mood stabilizer, or antidepressant medication [EDI-R: d = .84, F (1, 998) = 43.52, p < .001; EDI-D: d = .95, F (1, 998) = 54.68, p < .001] or who were in therapy [EDI-R: d = .97, F (1, 998) = 116.55, p < .001; EDI-D: d = 97, F (1, 998) = 123.32, p < .001] all had significantly higher EDI-R and EDI-D scores than those who were not receiving psychiatric or psychological treatment. Finally, participants with current diagnoses of depressive or anxiety disorder diagnoses had higher EDI-R [d = .72, F (1, 995) = 50.96, p < .001] and EDI-D [d = 88, F (1, 995) = 81.67, p < .001] scores than those who never had these diagnoses or only had them in the past but not currently. Participants with current diagnoses of disruptive mood dysregulation disorder, oppositional defiant disorder, or bipolar disorder had higher EDI-R [d = 1.51, F (1, 995) = 105.19, p < .001] and EDI-D [d = 1.35, F (1, 995) = 95.55, p < .001] scores than those who never had these diagnoses or only had them in the past but not currently. See Figures 3 and 4 for a visual depiction of mean theta scores by known-groups.

Table 4.

Correlations between the EDI Scales and Established Measures

EDI Reactivity EDI Reactivity Short Form EDI Dysphoria ERC Emotion Regulation ERC Emotional Lability Affective Reactivity Index CBCL Irritability CBCL Anxious/ Depressed
EDI Reactivity Short Form .95
EDI Dysphoria .71 .63
ERC Emotion Regulation .39 .35 .47
ERC Emotional Lability .66 .64 .53 .58
Affective Reactivity Index .68 .67 .60 .43 .72
CBCL Irritability .57 .55 .46 .34 .60 .67
CBCL Anxious/ Depressed .47 .44 .51 .31 .49 .49 .58
CBCL Withdrawn/Depressed .44 .40 .55 .43 .50 .53 .62 .71
Open in a new tab

Note. All correlations were significant at p < .001. EDI = Emotion Dysregulation Inventory; ERC = Emotion Regulation Checklist; CBCL=– Child Behavior Checklist.

Figure 3.

Figure 3.

Open in a new tab

Mean EDI Reactivity theta scores for known-groups. Theta scores have a mean of zero and standard deviation of one. Error bars represent 95% confidence intervals. DMDD = Disruptive Mood Dysregulation Disorder. ODD = Oppositional defiant disorder.

Figure 4.

Figure 4.

Open in a new tab

Mean EDI Dysphoria theta scores for known-groups. Theta scores have a mean of zero and standard deviation of one. Error bars represent 95% confidence intervals. DMDD = Disruptive Mood Dysregulation Disorder. ODD = Oppositional defiant disorder.

In a second approach to identifying clinical significance, we performed receiver operating characteristics (ROC) analyses using the six known-group outcomes described above. The ROC analyses were performed with the both the EDI-R and the EDI-D as predictors of the six binary outcomes, e.g., history of lifetime hospitalization, yes versus no. All the areas under the curve (AUCs) for these analyses were significant at the .001 level. For the EDI-R, the AUCs ranged from 0.702 to 0.867 (SDs ranged from 0.022 to 0.038), and for the EDI-D, the range was 0.726 to 0.829 (SDs ranged from 0.022 to 0.040).

Discussion

The EDI was initially developed for use in autism spectrum disorder and this study sought to evaluate its utility in general youth samples. The results strongly supported the EDI as an efficient and precise measure of emotion dysregulation in a general population of US youth. The EDI scales were correlated with established measures of related constructs as expected, and youth who would be expected to have worse emotion dysregulation (e.g., those in psychiatric treatment, with recent crisis evaluations, or with psychiatric diagnoses) had higher EDI scores than those not in these groups. These findings help to establish the EDI as a tool validated in both neurodevelopmental and general youth populations, providing a new opportunity for transdiagnostic emotion dysregulation studies and clinical monitoring in cross-population clinics, as well as use of the EDI in clinical populations other than autism.

The most noteworthy finding was the EDI’s superior precision in comparison to measures developed for general populations and widely applied in clinical research. This result may stem from the methods used to develop and refine the EDI. For example, other measures developed following the PROMIS Scientific Standards and with IRT have similarly been more precise with fewer items than established measures (e.g., Cella et al., 2007; Pilkonis et al., 2011). It is also likely that beginning in an autism sample ultimately optimized the EDI’s ability to differentiate across a wide range of severity. An additional benefit of the EDI’s original development in autism spectrum disorder may be the exclusive use of observable manifestations of emotion dysregulation. While originally done to ensure that the EDI could be used with nonverbal youth, this may have the benefit of increasing the reliability of informant ratings, even among verbal youth with autism spectrum disorder and non-autism youth.

Nonetheless, it is worth noting that the measures used for IRT cross-calibration with the EDI measure slightly different constructs or components of emotion dysregulation, which may have also contributed to their lower information yield compared to the EDI. For example, the ERC ER scale includes several items tapping the ability to communicate one’s emotions to others, which is not a facet included in the EDI. In addition, there are important differences in the time frames of each measure. Because one of the goals of the EDI development was to provide a sensitive measure for clinical trials or treatment monitoring, it utilizes “the past seven days” as its time frame, whereas other comparison measures are focused on more stable patterns of mood and utilize the past month (CBCL) or past six months (ARI) time frames.

The results confirmed the original EDI factor structure of Reactivity and Dysphoria scales in the general US census-matched sample. As such, the EDI Reactivity and Dysphoria scales can be used separately, or in combination, but with no overall total score. The Reactivity scale includes items capturing the initial emotional intensity (e.g., strong and fast emotional reactions) as well as difficulty down-regulating emotion. This is consistent with prior work finding that emotional reactivity and emotion regulation are best explained by a single factor in general adult populations (Zelkowitz & Cole, 2016), as well as factor analytic work in youth with irritability (Vidal-Ribas et al., 2016). Further, the Reactivity scale’s content is consistent with conceptualizations of emotion dysregulation in several psychiatric populations, such as youth with severe irritability (Brotman et al., 2017) and individuals with borderline personality disorder (Carpenter & Trull, 2013), among others. The Dysphoria scale, in contrast, captures attenuated positive affect and includes items capturing sadness and low motivation. Future research is needed to investigate how EDI Dysphoria corresponds to clinical diagnoses, though the current results indicate that it captures more information than composite indices of DSM-5-informed anxiety and depression and symptoms of withdrawal and depression.

Simulations of computerized adaptive testing were used to identify the most informative items for a short form for EDI Reactivity, and these simulations converged on the same candidate items across autism and US-census matched samples. The consistency in both the structure and precision of emotion dysregulation items across samples provides further support for the notion of emotion dysregulation as a transdiagnostic, dimensional construct. Indeed, there is burgeoning interest in transdiagnostic emotion regulation-focused treatments and service delivery models (Loevaas et al., 2019; Shaffer et al., 2019; Volkaert, Wante, Vervoort, & Braet, 2018).

One of the most commonly applied transdiagnostic frameworks for psychiatric research is the National Institute of Mental Health’s Research Domain Criteria (RDoC) project (Insel, 2014). Although emotion regulation (or dysregulation) was not included in RDoC, it has been argued that emotion regulation should be considered as a potential new domain in the RDoC matrix because it may help to synthesize understanding of functional impairment arising from the current five domains (Fernandez, Jazaieri, & Gross, 2016). Calls to study emotion dysregulation across multiple diagnostic groups with dimensional measures (Aldao, 2016) suggest that the EDI, and availability of its general US norms, may be particularly fruitful in this line of research.

The EDI may also be useful in population-focused studies, such as studies of youth with disruptive mood dysregulation disorder, depression, bipolar disorder, oppositional defiant disorder, or other disorders that are associated with emotion dysregulation. Importantly, the YouGov parameters generated from this work provide meaningful clinical thresholds (e.g., one SD above the YouGov mean) capturing expectations among a general population of children and adolescents for use as clinical cut-off score to support screening efforts or interpretation in research studies. The EDI could be applied in clinical trials as an outcome measure, in mechanistic studies to identify predictors of psychopathology development, or to compare emotion dysregulation between different disorders. It would be of interest to see how the EDI scales correspond to measures focused on strategy-oriented models of emotion regulation in order to identify treatment targets.

Several aspects of the EDI development should be considered when interpreting these findings. First, because the EDI was developed to be applicable for youth who may be nonverbal, it does not include items commonly assessed when someone is able to communicate verbally. For example, the EDI does not measure emotion regulation strategies, worrying, rumination, or suicidal thoughts. In addition, the cognitive interviews to refine the item wording were only completed with parents of children with autism spectrum disorder, and it cannot be ruled out that parents of typically-developing children or children with other diagnoses may interpret the item wording differently. These limitations notwithstanding, the psychometric properties of the EDI in this sample were strong and suggest it has utility beyond autism. Finally, because the current sample was a representative of the general US population and was not a clinically-referred sample, only a minority had parent-reported psychiatric diagnoses. Therefore, it would be beneficial to evaluate the EDI and its psychometric properties in other clinical populations in the future.

In sum, the EDI, with its original autism spectrum disorder norms and scoring conversions, is already being utilized for universal screening, treatment monitoring, clinical trials, and other research in autism, neurodevelopmental, and intellectual disability populations. With the new availability of IRT-based scoring conversions and norms for a general population of youth, the benefits of the EDI’s precision and efficiency can be extended to research and clinical care with other clinical samples and community populations. Investigators and clinicians will have the opportunity to utilize a single, brief measure of emotion dysregulation across various populations, with the freedom to apply the scoring system based on either the autism spectrum disorder or general US sample calibrations. It may be that certain studies benefit from use of the parameters and normative data from the YouGov sample, such as population-based studies or studies of psychopathology development, whereas studies of clinical populations may benefit from the enhanced range of the threshold parameters and distributional data from the autism sample for more severe manifestations of emotion dysregulation.

Acknowledgments

This work was supported by NICHD under grant R01 HD079512 (CM); International Society for Autism Research/Slifka Foundation under the Ritvo-Slifka Award for Innovation in Autism Research (CM). The authors have no conflicts of interest to report. We are grateful to YouGov for their support in data collection, as well as the families who participated. Finally, we would also like to thank Nate Dodds for his assistance in creating the figures. The data that support the findings of this study are in the process of being openly available in the National Database for Autism Research.

References

  1. Achenbach TM, & Ruffle TM (2000). The Child Behavior Checklist and related forms for assessing behavioral/emotional problems and competencies. Pediatrics in Review, 21(8), 265–271. doi: 10.1542/pir.21-8-265 [DOI] [PubMed] [Google Scholar]
  2. Aldao A (2016). Introduction to the special issue: Emotion regulation as a transdiagnostic process. Cognitive Therapy and Research, 40(3), 257–261. 10.1007/s10608-016-9764-2 [DOI] [Google Scholar]
  3. Aldao A, Nolen-Hoeksema S, & Schweizer S (2010). Emotion-regulation strategies across psychopathology: A meta-analytic review. Clinical Psychology Review, 30(2), 217–237. 10.1016/j.cpr.2009.11.004 [DOI] [PubMed] [Google Scholar]
  4. Baio J, Wiggins L, Christensen DL, Maenner MJ, Daniels J, Warren Z, … &, Dowling NF. (2018). Prevalence of autism spectrum disorder among children aged 8 years - autism and developmental disabilities monitoring network, 11 sites, United States, 2014. MMWR Surveillance Summary, 67, 1–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Brotman MA, Kircanski K, Stringaris A, Pine DS, &Leibenluft E. (2017). Irritability in youths : A translational model. The American Journal of Psychiatry, 174(6), 520–532. 10.1176/appi.ajp.2016.16070839 [DOI] [PubMed] [Google Scholar]
  6. Cai L, Thissen D, & du Toit S (2011). IRTPRO for Windows [computer software]. Lincolnwood, IL: Scientific Software International. [Google Scholar]
  7. Carpenter RW, & Trull TJ (2013). Components of emotion dysregulation in borderline personality disorder: A review. Current Psychiatry Reports, 15(1), 335. 10.1007/s11920-012-0335-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cella D, Yount S, Rothrock N, Gershon R, Cook K, Reeve B, … Rose M (2007). The Patient-Reported Outcomes Measurement Information System (PROMIS): Progress of an NIH roadmap cooperative group during its first two years. Medical Care, 45(5 Suppl 1), S3–S11. doi: 10.1097/01.mlr.0000258615.42478.55 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Choi SW (2009). Firestar: Computerized Adaptive Testing (CAT) simulation program for polytomous IRT models. Applied Psychological Measurement, 33(8), 644–645. doi: 10.1177/0146621608329892 [DOI] [Google Scholar]
  10. Cohen J (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence Earlbaum Associates. [Google Scholar]
  11. Compas BE, Jaser SS, Bettis AH, Watson KH, Gruhn MA, Dunbar JP, … Thigpen JC (2017). Coping, emotion regulation, and psychopathology in childhood and adolescence: A meta-analysis and narrative review. Psychological Bulletin, 143(9), 939–991. 10.1037/bul0000110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Conner CM, White SW, Beck KB, Golt J, Smith IC, & Mazefsky CA (2018). Improving emotion regulation ability in autism: The Emotion Awareness and Skills Enhancement (EASE) Program . Autism, 23(5), 1273–1287.. doi: 10.1177/1362361318810709 [DOI] [PubMed] [Google Scholar]
  13. Evans SC, Bonadio FT, Bearman SK, Ugueto AM, Chorpita BF, & Weisz JR (2019), Assessing the irritable and defiant dimensions of youth oppositional behavior using CBCL and YSR items, Journal of Clinical Child & Adolescent Psychology, Advance online publication. doi: 10.1080/15374416.2019.1622119 [DOI] [PubMed] [Google Scholar]
  14. Fernandez KC, Jazaieri H, & Gross JJ (2016). Emotion regulation: A transdiagnostic perspective on a new RDoC domain. Cognitive Therapy and Research, 40(3), 426–440. 10.1007/s10608-016-9772-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Gratz KL, Weiss NH, & Tull MT (2015). Examining emotion regulation as an outcome, mechanism, or target of psychological treatments. Current Opinion in Psychology, 3, 85–90. 10.1016/j.copsyc.2015.02.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Gross JJ (2015). Emotion regulation : Current status and future prospects. Psychological Inquiry, 26(1), 1–26.. 10.1080/1047840X.2014.940781 [DOI] [Google Scholar]
  17. Gross JJ, & Thompson RA (2007). Emotion regulation: Conceptual foundations. In Gross JJ (Ed.), Handbook of emotion regulation (pp. 3–24). New York, NY: The Guildford Press. [Google Scholar]
  18. IAN (Interactive Autism Network). (2017). Linking the autism community and researchers. Retrieved September 8, 2017, from https://www.ianresearch.org/.
  19. Insel TR (2014). The NIMH Research Domain Criteria (RDoC) Project: Precision medicine for psychiatry. American Journal of Psychiatry, 171, 395–397, doi: 10.1176/appi.ajp2014.14020138 [DOI] [PubMed] [Google Scholar]
  20. Ip KI, Jester JM, Sameroff A, & Olson SL (2019). Linking Research Domain Criteria ( RDoC ) constructs to developmental psychopathology : The role of self-regulation and emotion knowledge in the development of internalizing and externalizing growth trajectories from ages 3 to 10. Development and Psychopathology, 1–18. 10.1017/S0954579418001323 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Joshi G, Wozniak J, Fitzgerald M, Faraone S, Fried R, Galdo M, … Biederman J (2018). High risk for severe emotional dysregulation in psychiatrically referred youth with autism spectrum disorder : A controlled study. Journal of Autism and Developmental Disorders, 48(9), 3101–3115. 10.1007/s10803-018-3542-9 [DOI] [PubMed] [Google Scholar]
  22. Kelly JR, Clarke G, Cryan JF, & Dinan TG (2018). Dimensional thinking in psychiatry in the era of the Research Domain Criteria ( RDoC ). Irish Journal of Psychological Medicine, 35(2), 89–94. 10.1017/ipm.2017.7 [DOI] [PubMed] [Google Scholar]
  23. Lakens D (2013). Calculating and reporting effect sizes to facilitate cumulative science: a practical primer for t-tests and ANOVAs. Frontiers in Psychology, 4, 863. doi: 10.3389/fpsyg.2013.00863 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Leibenluft E (2011). Severe mood dysregulation, irritability, and the diagnostic boundaries of bipolar disorder in youths. The American Journal of Psychiatry, 168(2), 129–142. doi: 10.1176/appi.ajp.2010.10050766 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Loevaas MES, Sund AM, Lydersen S, Neumer S, Martinsen K, Holen S,, … Reinfjell T. (2019). Does the transdiagnostic EMOTION intervention improve emotion regulation skills in children ? Journal of Child and Family Studies, 28(3), 805–813. 10.1007/s10826-018-01324-1 [DOI] [Google Scholar]
  26. Mazefsky CA, Day TN, Siegel M, White SW, Yu L, & Pilkonis PA (2018). Development of the Emotion Dysregulation Inventory: A PROMIS®ing Method for Creating Sensitive and Unbiased Questionnaires for Autism Spectrum Disorder. Journal of Autism and Developmental Disorders, 48(11), 3736–3746. 10.1007/s10803-016-2907-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Mazefsky CA, Yu L, White SW, Siegel M, & Pilkonis PA (2018). The emotion dysregulation inventory: Psychometric properties and item response theory calibration in an autism spectrum disorder sample. Autism Research, 11(6), 928–941. 10.1002/aur.1947 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Muthén LK, & Muthén BO (2007). Mplus [computer software]. Los Angeles, CA: Muthén & Muthén. [Google Scholar]
  29. Pilkonis PA, Choi SW, Reise SP, Stover AM, Riley WT, & Cella D (2011). Item banks for measuring emotional distress from the patient-reported outcomes measurement information system (PROMIS®): Depression, anxiety, and anger. Assessment, 18(3), 263–283. 10.1177/1073191111411667 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. PROMIS Standards Committee. (2013). PROMIS® Instrument Development and Validation Scientific Standards, Version 2.0, (revised May 2013). Retrieved from http://www.healthmeasures.net/images/PROMIS/PROMISStandards_Vers2.0_Final.pdf
  31. Reeve BB, Hays RD, Bjorner JB, Cook KF, Crane PK, Teresi JA,…PROMIS Cooperative Group. (2007). Psychometric evaluation and calibration of health-related quality of life item banks: Plans for the Patient-Reported Outcomes Measurement Information System (PROMIS). Medical Care, 45 (5 Supplement 1), S22–31. [DOI] [PubMed] [Google Scholar]
  32. Reise SP, & Yu J (1990). Parameter recovery in the graded response model using MULTILOG. Journal of Educational Measurement, 27(2), 133–144. 10.1111/j.1745-3984.1990.tb00738.x [DOI] [Google Scholar]
  33. Shaffer RC, Wink LK, Ruberg J, Pittenger A, Adams R, Sorter M, … Erickson CA (2019). Emotion regulation intensive outpatient programming : Development, feasibility, and acceptability. Journal of Autism and Developmental Disorders, 49(2), 495–508. 10.1007/s10803-018-3727-2 [DOI] [PubMed] [Google Scholar]
  34. Shields A, & Cicchetti D (1997). Emotion regulation among school-age children: The development and validation of a new criterion Q-sort scale. Developmental Psychology, 33(6), 906–916. doi: 10.1037/0012-1649.33.6.906 [DOI] [PubMed] [Google Scholar]
  35. Siegel M, Smith KA, Mazefsky C, Gabriels RL, Erickson C, Kaplan D… Santangelo SL for the Autism and Developmental Disorders Inpatient Research Collaborative (ADDIRC). (2015). The Autism Inpatient Collection: Methods and preliminary sample description. Molecular Autism, 6, 61–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Sloan E, Hall K, Moulding R, Bryce S, Mildred H, & Staiger PK (2017). Emotion regulation as a transdiagnostic treatment construct across anxiety, depression, substance, eating and borderline personality disorders : A systematic review. Clinical Psychology Review, 57, 141–163. 10.1016/j.cpr.2017.09.002 [DOI] [PubMed] [Google Scholar]
  37. Stringaris A, Goodman R, Ferdinando S, Razdan V, Muhrer E, Leibenluft E, & Brotman MA (2012). The Affective Reactivity Index: A concise irritability scale for clinical and research settings. The Journal of Child Psychology and Psychiatry, 53(11), 1109–1117. 10.1111/j.1469-7610.2012.02561.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Towbin KE, Pradella A, Gorrindo T, Pine DS, & Leibenluft E (2005). Autism spectrum traits in children with mood and anxiety disorders. Journal of Child and Adolescent Psychopharmacology, 15(3), 452–464. 10.1089/cap.2005.15.452 [DOI] [PubMed] [Google Scholar]
  39. United States Census Bureau (2016). American Community Survey. Retrieved from https://www.census.gov/programs-surveys/acs/news/data-releases.html?#
  40. Vidal-Ribas P, Brotman MA, Valdivieso I, Leibenluft E, & Stringaris A (2016). The status of irritability in psychiatry: A conceptual and quantitative review. Journal of the American Academy of Child and Adolescent Psychiatry, 55(7), 556–570. 10.1016/j.jaac.2016.04.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Volkaert B, Wante L, Vervoort L, & Braet C (2018). ‘ Boost Camp ‘, a universal school-based transdiagnostic prevention program targeting adolescent emotion regulation ; evaluating the effectiveness by a clustered RCT : a protocol paper. BMC Public Health, 18(1), 904. doi: 10.1186/s12889-018-5754-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Weiss JA, Thomson K, & Chan L (2014). A systematic literature review of emotion regulation measurement in individuals with autism spectrum disorder. Autism Research, 7(6), 629–648. 10.1002/aur.1426 [DOI] [PubMed] [Google Scholar]
  43. Youngstrom EA (2013). A primer on receiver operating characteristic analysis and diagnostic efficiency statistics for pediatric psychology: We are ready to ROC. Journal of Pediatric Psychology, 39, 204–221. 10.1093/jpepsy/jst062 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Zelkowitz RL, & Cole DA (2016). Measures of emotion reactivity and emotion regulation: Convergent and discriminant validity. Personality and Individual Differences, 102, 123–132. 10.1016/j.paid.2016.06.045 [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES