Abstract
Test Description
This review focuses on the Anger Regulation and Expression Scale (ARES) which was developed by DiGiuseppe and Tafrate (2011) and published by Multi-Health Systems Inc. The ARES was designed to be a self-report measure of anger expression and regulation in youth aged 10 to 17 years and was intended to be used in screening, individual assessment, and treatment evaluation. The ARES may also be used for research purposes. The ARES includes a full-length and a short version, the ARES(S). The full-length ARES test is a 75 question self-report form. Participants are presented with statements regarding situations in which they feel anger. The test taker indicates whether they find each item representative of their own thoughts, feelings, or actions using the descriptors: never, hardly ever, sometimes, often, or always.
The development of the ARES was based on research by DiGiuseppe and Tafrate (2011) concerning the construct of anger as a clinical problem. Prior to the ARES, the authors developed the Anger Disorders Scale (ADS), an anger inventory for adults, which provides the foundation for the ARES. The ARES claims to assess more constructs that comprise the experience of anger and more types of anger expression than any other youth measure. The ARES was designed to be clinically relevant in identifying specific patterns of feeling, behavior, and thinking that can be targeted in intervention. The manual provides a review of literature on anger as a clinical disturbance, an explanation of test development, information on ARES administration and interpretation and comprehensive treatment plans based on ARES scores.
Test Administration
The ARES and ARES(S) are administered in paper-and-pencil format using a response form or online using an online form. Standardized scores from the ARES assessments allow for the comparison of an individual youth to their normative group. The ARES(S) may also be used for screening a group of children or adolescents to identify individuals who may require a more comprehensive evaluation. High scores on the ARES(S) indicate the need for further assessment.
Scales and Subscales
The ARES scales are based on the proposed criteria for anger regulation and expression disorder (ARED; DiGiuseppe & Tafrate, 2007), which is a possible addition in the future revision of the DSM. The proposed criteria for ARED include three subtypes: Predominately Subjective Type, Predominately Expressive Type, and the Combined Type. Scoring categories on the ARES fit into these subtypes. The three scoring clusters of the ARES are the Internalizing Anger Cluster, the Externalizing Anger Cluster, and the Extent of Anger Cluster. High scores on only Internalizing Anger and Extent of Anger Clusters meet criteria for ARED: Predominately Subjective Type. High scores on only the Externalizing Anger Cluster meet criteria for ARED: Predominately Expressive Type. High scores on all three Clusters meet criteria for ARED: Combined Type.
The authors approach anger from the five-domain model of emotions, which posits that understanding an emotional episode involves the Provocations domain, Cognitions domain, Arousal domain, Motives domain, and Behaviors domain (Clore & Ortony, 1991; Frijda, 1986; Izard, 1977; Lazarus, 1991; Power & Dalgleish, 2008). The ARES scales and subscales assess each of these five domains and provide an unofficial diagnosis based on the ARED criteria.
The Provocations domain concerns the stimuli that triggered the emotion and is related to the scales: Scope of Triggers and Rejection. The Cognitions domain concerns individuals’ appraisals of the provocations leading to the emotion and includes the Bitterness Scale and subscales Resentment and Suspiciousness, as well as the scales of Impulsivity and Cognitive Arousal. The Arousal domain concerns the physiological experience of anger, including anger intensity and duration. This domain is comprised of the scales of Problem Duration, Episode Duration, and Arousal with subscales Physiological Arousal and Cognitive Arousal. The Motives domain concerns what an individual hopes to achieve by reacting to the problem that triggered the emotion and includes the scales of Revenge and Bullying. The Behaviors domain concerns behaviors engaged in by the individual to express the emotion or cope with the problem. This domain includes the scales of Anger-In, Overt Aggression/Expression with subscales Physical Aggression and Verbal Aggression, Covert Aggression, and Subversion with subscales Relational Aggression and Passive Aggression.
The Internalizing Anger Cluster includes the scales of Arousal, Rejection, Anger-In, and Bitterness, focusing on the tendency to hold anger in and the private feelings associated with this. The Externalizing Anger Cluster concerns the outward expression of anger and is comprised of the following scales: Overt Aggression/Expression, Covert Aggression, Revenge, Subversion, Bullying, and Impulsivity. The Extent of Anger Cluster is made up of the Scope of Triggers, Problem Duration, and Episode Duration scales and concerns the extent of angering stimuli and anger maintenance over time.
Scoring
ARES and ARES(S) Form responses are entered into a computerized scoring program available either as scoring software or online. The program provides two report types, namely, the assessment report and the progress report. The assessment report provides detailed results from a single administration along with an individualized treatment plan. The progress report examines change over time by comparing results from up to four administrations. The ARES computerized reports provide raw scores, T-scores, percentile ranks, and confidence intervals.
Test Construction
Item Generation Procedures
Assessing anger as a clinical problem necessitated the construction of items and scales with a positively skewed distribution, as clinicians are interested in finer distinctions at the higher end of the anger continuum. Item generation was designed to measure constructs most relevant to children and adolescents in the five domain areas. Several items from the ADS were rewritten to have more suitable content and language for children and adolescents and new items were generated when needed.
The most recent ARES is in its third revision. Pilot data were collected for items on the first two publications of the ARES from more than 700 culturally diverse youth. Data from normative and clinical groups were used to finalize the factor structure of the ARES.
The total sample (both normative and clinical youth) was divided into two groups: the derivation sample and the confirmatory sample. An explanatory factor analysis was conducted on the derivation sample to construct the factor structure of the ARES. A confirmatory factor analysis was then conducted on the confirmatory sample to determine if the factor structure was appropriate. As a result of several factor analyses, an eight-factor solution was found most appropriate. Exploratory factor analyses were conducted separately for girls and boys, and for age groups of 10 to 13 years and 14 to 17 years. Results indicated that there was a high consistency across males and females and across the age groups. Analyses indicated that four of the eight factors could potentially have subscales.
Standardization Sample
The normative sample consisted of 800 self-report ratings out of more than 1,300 completed ARES forms. The normative sample had equivalent age and gender representation, with 50 females and 50 males at each age from 10 through 17 years old. Data were collected between 2006 and 2009 throughout the United States in private practice, school, and agency settings. The race/ethnicity of the normative sample was African American (16.9%), Hispanic/Latino (14.5%), White (61.8%), and Multiracial/Other (6.9%). In terms of geographic region distribution, 42.6% of the data came from the Northeast, 36.6% from the South, 9.8% from the West, 9.1% from the Midwest, and 1.9% were missing/other.
In the norming procedure, researchers explored if there were any age or gender effects in the data to determine the proper norm groups. Gender was found to have a small, but statistically significant effect on the Externalizing Anger cluster. Similarly, age was found to have a small, but significant effect on the total score.
Reliability
Internal consistency and test-retest were used to measure the reliability of ARES scores. Internal consistency was measured by calculating Cronbach’s alpha across age and gender groups. The total score alpha ranged from .97 to .99 and Cluster score values varied from .87 to .97, showing an excellent level of reliability. The Internalizing Anger, Externalizing Anger, and Extent of Anger scale and subscale values ranged from .77 to .96, .68 to .95, and .65 to .95, respectively.
Test-retest reliability was assessed with a sample of 58 individuals. The participants were retested within a 2-to-4-week period and test-retest values ranged from .58 to .92. All test-retest reliability estimates were found significant for the ARES T-scores (p < .001). However, there was a restricted range issue for some scales that was solved by test-retest reliability adjustment by using Cohen, Cohen, West, and Aiken’s (2003) correlation correction methods for restricted range cases.
Validity
The ARES examined both forms of construct validity: discriminant and convergent. Discriminative validity describes an instrument’s ability to differentiate a related group of individuals. Convergent validity determines if the results of an instrument are appropriately similar to the results of other similar measures. The scores of the ARES also were tested for generalizability across race/ethnic groups.
The discriminative validity of the ARES was tested via analyses of covariance (ANCOVAs), multivariate analyses of covariance (MANCOVAs), and discriminant function analyses (DFAs). The ANCOVA method was used to analyze the total score considering the scores on the ARES as dependent variables. Following the ANCOVA, MANCOVAs were conducted to examine the relationship of dependent variables such as the cluster scores, the scales, and the subscales with gender, race/ethnicity and age variables. The target groups were identified as the conduct disorder and oppositional defiant disorder groups because these groups are prevalently related with problematic anger reactions. The researchers planned three different comparisons before the analysis: (a) target clinical groups versus the general population group, (b) target clinical groups versus other clinical groups, and (c) target clinical groups versus each other. Results of every analysis indicated significant main effect of group membership.
A sample of youth from the general population that completed the ARES also completed at least one other test, the Conners Comprehensive Behavior Rating Scales (Conners CBRS) and/or the Jesness Inventory–Revised (JI-R). The correlation between the total score of the ARES and the Conners CBRS was very strong with correlation coefficients between .64 and .70. The selected JI-R scores correlated with ARES scores ranged from moderate to strong with correlation coefficients between .42 and .52 for the total score.
Commentary and Recommendations
Overall, the ARES appears to accomplish its intended purpose of measuring the expression and regulation of anger in youth aged 10 to 17 years. The ARES claims to assess more constructs that comprise the experience of anger and more types of anger expression than any other youth measure. Extensive background information that the authors provided concerning development of this test supports this claim. The ARES approach to testing children and adolescents by developing new items and rewording language to make items more situationally appropriate is one of its strengths, making it possibly the superior testing instrument available at this time. Another strength of the ARES is that it can be administered in a short amount of time (ARES—25 min, ARES[S]—5 min). Finally, while the ARES has a complex and integrative system of clusters, scales, and subscales, the authors clearly explain how they developed this model. The amount of supporting and explanatory information in the manual is immense and allows the user to gain a thorough understanding of the testing process.
The ARES does have some weaknesses, which are acknowledged within the manual. The test assumes that all participants are proficient in English and also states that it is unable to capture people who are disoriented, psychotic, or severely neurologically or intellectually impaired. The ARES manual does not provide recommendations or alternatives for these situations. Another potential weakness of the ARES is that it assumes that all participants are honest in their responses to the items. This weakness is discussed in depth and the manual points out extreme score responses that may be clues to such behavior. Although this is a weakness with the ARES and any other self-report measure, a strength of the ARES is that it addresses this topic thoroughly and indicates possible solutions for such an occurrence.
In terms of reliability and validity, the ARES has some weaknesses that should be considered. Even though the authors provided sufficient information on some types of reliability and validity for the ARES, they did not take into account some reliability and validity types. Although the internal consistency and test-retest reliability measurements were conducted, some other types of reliabilities such as interrater reliability were not considered. Moreover, for validity of the ARES the authors merely provided evidence of construct validity (discriminant and convergent), but they did not include any other types of validity such as content validity, criterion-related validity, and consequential validity. Although the ARES has some weaknesses regarding insufficient reporting of reliability and validity, reporting the validity and reliability scores with effect sizes is a strength for the ARES.
In relation to other testing measures currently available in this area, the ARES appears to be a good measure of anger as a clinical problem for children and adolescents. Although the ARES is not ideal for all situations, it would be a valuable tool to use in conjunction with information gathered from observations and other measures of anger in youth for assessment, treatment planning, and progress monitoring.