• Conceptualization and Initial Planning
  • Focused Adult ADHD Assessment Tool
  • Greater Alignment with Conners 4
  • Fairness and Cultural Sensitivity
  • Updated Normative Samples and Addition of ADHD Reference Samples
  • New and Improved Metrics to Capture Response Style Analysis
  • Screening Items for Associated Clinical Concerns
  • Updated and Expanded Coverage of ADHD-Related Content
  • Inclusion of Items Assessing Functional Impairment
• Item Development

The initial development phase of the CAARS 2 was guided by three sources of information. First, comprehensive and systematic reviews were conducted on the literature regarding (a) ADHD in adults since the publication of the Conners’ Adult ADHD Rating Scales (CAARS™; Conners et al., 1999), and (b) the CAARS itself in order to identify key improvements needed in rating scales used in the assessment of ADHD in adults. Second, market research was conducted with users of the CAARS to determine the most valued components of the original scale and to identify changes deemed the most important in terms of improving its utility. Finally, the development team worked closely with internal and external subject-matter experts, including the authors of the measure (Drs. Keith Conners, Drew Erhardt, and Elizabeth Sparrow), to determine appropriate revision goals. Drs. Erhardt and Sparrow were actively involved with the MHS team throughout the development of the CAARS 2. Dr. Conners participated in the early conceptualization and review of preliminary pilot items for the CAARS 2. Although Dr. Conners passed away in 2017, his command of the science of ADHD, deep understanding of the disorder, and experience with rating scales are all evident in the final CAARS 2

Information gained from the literature reviews, CAARS users, and subject-matter experts shaped the following revision goals for the CAARS 2:

• Retain the original CAARS’ unique identity as a focused assessment of adult ADHD.
  
  
• Create greater alignment with MHS’s parallel measure of ADHD in youth, the Conners 4th Edition (Conners 4™; Conners, 2022), to facilitate the comparison of findings in childhood/adolescence with later adult-level assessments.
  
  
• Meet standards for fairness and cultural sensitivity through more representative normative samples and careful, inclusive choices regarding the language used.
  
  
• Update normative samples and add ADHD Reference Samples.
  
  
• Revise and enhance metrics to include sophisticated tools (including validity scales) for describing response style patterns that could impact the interpretation of results.
  
  
• Add screening items for additional clinical concerns (such as suicidality, self-harm, anxiety, depression).
  
  
• Update and expand coverage of ADHD-related content.
  
  
• Add items to assess ADHD-related impairments and possible outcomes across a range of functional domains.

Conceptualization and Initial Planning

Focused Adult ADHD Assessment Tool

The CAARS 2 was developed with the intention of building on the core strengths of the original CAARS, while also addressing potential limitations identified in the literature reviews and market research. One challenge involved how to address conflicting user requests to develop a scale that was both brief and comprehensive. Consequently, an important goal of the revision was to strike a balance between adding useful content to the CAARS 2 without sacrificing the highly valued brevity of the original measure.

Greater Alignment with Conners 4

It is well documented that ADHD symptoms persist into adulthood for the majority of individuals who have been diagnosed with ADHD in childhood. Williamson and Johnston (2015) reported a persistence rate of approximately 36% to 55%, as reflected in children followed into adulthood who retained a full diagnosis. Additionally, about 30% to 70% of individuals diagnosed in childhood continued to demonstrate significant impairment (Barkley et al., 2008; Kessler et al., 2010). Given the often continuous nature of ADHD or aspects thereof across childhood, adolescence, and adulthood, having compatible rating scales to apply to the evaluation of ADHD across the lifespan would be extremely beneficial for clinicians and researchers alike. Not only would this enable examination of how reported problems in core functional areas of inattention, hyperactivity, and impulsivity change over time, but it would also facilitate tracking changes in other areas that are closely related to or associated with ADHD (e.g., executive functioning, emotional regulation). Thus, in order to realize these benefits, the Conners 4 (for use with youth) and CAARS 2 (for use with adults) were designed to be aligned with and comparable to one another to the greatest extent possible.

Fairness and Cultural Sensitivity

According to the Standards for Educational and Psychological Testing (American Educational Research Association [AERA], American Psychological Association [APA], & National Council on Measurement in Education [NCME], 2014), fairness is a fundamental validity issue in test development and must be addressed throughout all stages of a test’s development and use. Fairness is defined as the “responsiveness of a test to individual characteristics and testing contexts so that the test scores will yield valid interpretations for intended uses” (AERA, APA, & NCME, 2014, p. 50). A fair test not only provides impartial treatment of all test takers during testing, but also provides unbiased measurement as reflected by (a) a lack or absence of measurement bias, (b) accessibility to constructs being measured, and (c) the validity of individual test score interpretations for the test’s intended uses. In other words, a fair test does not put some individuals at an advantage or a disadvantage because of characteristics irrelevant to the intended construct being measured. A goal of the CAARS 2 was to meet these standards for fairness and cultural sensitivity throughout the entire development process.

Updated Normative Samples and Addition of ADHD Reference Samples

Normative Sample Updates. The norms of the original CAARS gradually became outdated due to the dramatic demographic changes in the North American population since its publication in 1999. The diminishing representativeness of the Normative Sample became particularly notable for historically marginalized groups given the substantial increases in their numbers as a percentage of the U.S. and Canadian populations (e.g., Hispanic individuals comprised 14.6% of the U.S. population in 2018 compared to 8.9% in 1990; U.S. Census Bureau, 1992, 2018). Therefore, new norms were needed to more accurately reflect the recent demographic makeup of the U.S. and Canada. Additionally, the original CAARS included four age bands (18–29 years, 30–39 years, 40–49 years, 50+ years). However, the oldest of those bands (ages 50+) was too broad to adequately capture changes as individuals age. Therefore, the CAARS 2 aimed to narrow the age bands generally with a particular focus on further stratifying the over-50 age group into multiple categories.

ADHD Reference Samples. The CAARS 2 conceptualization also included adding normative data from adults diagnosed with ADHD to improve the assessment of (a) symptom severity and (b) the degree of similarity to individuals diagnosed with the disorder. Therefore, in addition to updating the general population norms, plans were developed to collect normative data for an ADHD Reference Sample.

New and Improved Metrics to Capture Response Style Analysis

A rater’s response style can impact the interpretation of results on a rating scale. The validity of symptom reports is widely recognized to be a serious concern when evaluating adults for ADHD (e.g., Fuermaier et al., 2016; Harp et al., 2011; Jasinski et al., 2011; Musso & Gouvier, 2014; Quinn, 2003; Suhr et al., 2008; Sullivan et al., 2007; Tucha et al., 2015). Moreover, as invaluable as rating scales are to gathering assessment data, they can be compromised by factors such as being distracted, unmotivated, inattentive, or careless when completing the measure, or intentionally misrepresenting symptoms to acquire a diagnosis for secondary gains. All of these factors can affect the validity of the scores. Efforts to address these threats can include corroborating findings across multiple informants, as well as embedding response style analyses within rating scales to help evaluate for various forms of invalidity in an informant’s response style. The original CAARS included an Inconsistency Index designed to capture inconsistent response patterns that can reflect factors such as carelessness, random responding, or low motivation. However, research findings (e.g., Harrison et al., 2007; Quinn, 2003) and market survey research indicated that the CAARS would be greatly improved by the inclusion of additional and more sophisticated response style metrics. Thus, an important goal of the CAARS 2 was to incorporate a sophisticated, multi-faceted, state-of-the-art Response Style Analysis.

The steps identified to produce an enhanced Response Style Analysis included expanding the validity scales by updating the Inconsistency Index, drafting new items to detect biased responding, and embedding an ADHD Symptom Validity Test into the CAARS 2. It is well-established that self-reported ADHD symptoms can be easy to simulate (Quinn, 2003; Suhr et al., 2011; Tucha et al., 2015). Two methods have been shown to be effective in analog studies at differentiating between feigned and honest responses (e.g., Berry & Nelson, 2010; Harrison et al., 2007; Sollman et al., 2010; Suhr et al., 2011). One approach embeds unlikely presentations or improbable symptoms of ADHD into the scale. Doing so involves generating items that seem to be relevant to ADHD but, in fact, are not characteristic of the disorder (e.g., “One minute is the longest I can focus on anything” and “I am unable to control my behavior”). Individuals motivated to feign ADHD symptoms are prone to endorse such items at high rates, which will distinguish them from individuals with genuine ADHD who are unlikely to endorse such items (Harrison & Armstrong, 2016). A second approach is to include real but rare symptoms of ADHD (Fuermaier et al., 2016; Suhr et al., 2011). Because high endorsement of these infrequent ADHD symptoms is unlikely to occur even among individuals with the disorder, they can help identify raters who may exagerate their symptoms or adopt an overly negative response style.

A Negative Impression Index would also be included to detect biased responding aimed at deliberately creating a negative impression. A Negative Impression Index includes items that describe general problems that are not necessarily related to ADHD (e.g., “I am impossible to please”; Hopwood et al., 2007; Rios & Morey, 2013). Individuals who are feigning, to create an overall negative impression of themselves or to present themselves in the worst possible light, are more likely to endorse these negative impression, or “fake bad,” items (Hopwood et al., 2007) leading to elevated scores on this validity index.

To meet the development goal of having an enhanced Response Style Analysis, items were written to detect various forms of biased responding. Other potential response style metrics, including the number of omitted items and the pace of responding, were explored to provide a more multifaceted Response Style Analysis.

Screening Items for Associated Clinical Concerns

Adult ADHD often presents with co-occurring disorders or associated areas of clinical concern. Co-occurring psychiatric disorders have been found in 80% or more of adults with ADHD, with mood and anxiety disorders being particularly common (Barkley, 2015; Katzman et al., 2017; Sobanski, 2006). Not only is complex ADHD (the co-occurrence of ADHD with other disorders or clinical issues) the rule rather than the exception, but it is also associated with more severe impairment and complicates both accurate diagnosis and treatment planning. Moreover, compared to the general population, adults with ADHD have higher rates of suicidal ideation, suicide attempts, completed suicides, and various forms of non-suicidal self-injury (e.g., cutting, burning; Allely, 2014; Chronis-Tuscano et al., 2010; Huang et al., 2018; Hurtig et al., 2012; Impey & Heun, 2012; Ljung et al., 2014; Nock, 2010; O’Connor et al., 2013; Septier, 2019; Simioni et al., 2018; see Associated Impairments and Functional Outcomes in chapter 2, Background, for more details). Additionally, as described in Comorbidity in chapter 2, anxiety and depression are among the most common disorders accompanying ADHD in adults. Given the numerous important implications of the aforementioned co-occurrences and associated concerns for diagnosis, prognosis, treatment planning, and harm reduction, it was determined that the CAARS 2 should include items to screen for suicidality, self-harm, anxiety, and depression.

Updated and Expanded Coverage of ADHD-Related Content

The original CAARS included both factor-derived sub-scales (viz., Inattention/Memory Problems, Hyperactivity/Restlessness, Impulsivity/Emotional Lability, and Problems with Self-Concept) and ADHD Symptom subscales based on DSM criteria (viz., Inattentive Symptoms, Hyperactive-Impulsive Symptoms, and Total ADHD Symptoms). The CAARS also incorporated a 12-item ADHD Index, comprised of the statistically-determined best set of items for distinguishing adults with ADHD from adults from the general population. The revision sought to retain these key components from the CAARS, with updates based on new analyses of a revised item pool and current DSM criteria.

Content Scale Updates. As discussed in chapter 2, Background, adults with ADHD exhibit a wide range of problems with inattention and executive functioning (EF). Although the existing CAARS subscales of Inattention/Memory Problems and DSM Inattentive Symptoms included items related to difficulties in these areas, the revision aimed to more thoroughly cover the broader domain of EF deficits that are typically present in individuals with ADHD. Determining if there was an empirical basis for distinguishing an EF-focused scale from the original Inattention/Memory Problems scale required generating new items related to difficulties in planning, organization, time management, and other aspects of EF not previously addressed. Emotional dysregulation also emerged as an important aspect of ADHD in adults. It can manifest as affective lability, emotional over-reactivity, and the inability to inhibit emotional expression (Adler et al., 2017). Although the original CAARS assessed emotional dysregulation to some extent through the Impulsivity/Emotional Lability subscale, a revision goal was to assess this important domain more thoroughly. Consequently, additional items reflecting emotional dysregulation were added to determine if a separate Content Scale might be supported (i.e., dividing the original Impulsivity/Emotional Lability scale into two distinct domains). The remaining CAARS subscales (Hyperactivity/Restlessness and Problems with Self-Concept) were retained, but additional items with clear and accessible phrasing were incorporated.

DSM Symptom Scale Updates. The DSM Symptom Scales were updated to reflect changes in the DSM and to map more directly onto DSM-5¹ criteria.

ADHD Index Updates. The CAARS 2 aimed to retain the ADHD Index as a metric of distinguishing individuals with ADHD from those in the general population, but with updated items based on results from more sophisticated statistical techniques than those available at the time the original CAARS was developed.

Inclusion of Items Assessing Functional Impairment

A DSM diagnosis of ADHD requires there to be “clear evidence that the symptoms interfere with, or reduce the quality of, social, academic, or occupational functioning” (APA, 2013, p. 60). Moreover, as described in Associated Impairments and Functional Outcomes in chapter 2, Background, there is a considerable body of research that shows having ADHD as an adult elevates one’s risk for serious and widespread impairments. Given the empirical and conceptual basis for explicit assessment of impairment as an essential element of assessing for ADHD (Siekowski, 2017), an important goal in revising the CAARS was to include items that enable the standardized evaluation of several functional impairments and negative outcomes associated with adult ADHD. Thus, items assessing difficulties in functioning and adverse outcomes related to both specific tasks (e.g., money management, driving) and broad domains (e.g., work, school, relationships) were included in the CAARS 2.

Item Development

The conceptualization and specific goals described above for the CAARS revision required drafting a large bank of potential items. The CAARS 2 development team (including the authors) created a varied pool of items. In order to ensure the fairness and broad accessibility of content, this item development process was guided by the scholarly literature on culturally competent assessment along with consultation and item reviews provided by cross-cultural assessment experts. Items identified as containing colloquial phrases or terms, or that were otherwise deemed to be potentially inaccessible or unfair to diverse or marginalized populations, were either dropped or revised to be culturally appropriate. In addition, the development of the CAARS 2 followed best practice recommendations with respect to inclusivity and sensitivity (The Market Research Society, 2016). For example, on all CAARS 2 forms, there is an optional “Gender” field (as opposed to sex at birth) provided as part of the basic demographics, along with a response option of “Other” with a “Please specify” textbox. This response option was added to ensure that the CAARS 2 is gender-inclusive. Furthermore, pronouns such as the singular “they/them” are used throughout the rating scale as well as in the report (American Psychiatric Association, 2021). All items were subsequently reviewed by external subject-matter experts to ensure that critical adult-ADHD content areas were represented and adequately covered.

At the end of this first phase (Conceptualization, Initial Planning, and Item Development), the item pool consisted of 274 items for Self-Report and 271 items for Observer.