Manual

Conners 4 Manual

Chapter 9: Relation to Conceptually Related Constructs


Relation to Conceptually Related Constructs

view all chapter tables | print this section

Relation to the Conners 3rd Edition

The Conners 4th Edition is a revised version of the Conners 3rd Edition™ (Conners 3™; Conners, 2008), which assesses symptoms of ADHD in childhood and adolescence. To provide evidence of the validity of the Conners 4 scores based on their relation to scores from a similar measure of the constructs, parents, teachers, and youth were asked to complete the Conners 4 and Conners 3. The two instruments are inherently related due to the shared nature of the items; however, many items in the Conners 4 have been revised and many are newly added (see chapter 6, Development, for detailed information about updated content from Conners 3 to Conners 4). Accordingly, it was expected that the Conners 4 and Conners 3 would have moderate to strong relations, as the items and scales are similar but distinct.

Raters (Parent N = 85, Teacher N = 58, and Self-Report N = 81) must have completed both the Conners 3 and Conners 4 within one month; the interval between administrations ranged from completion on the same day to up to 30 days later for Parent, up to 9 days later for Teacher, and up to 25 days later for Self-Report. The demographic characteristics of the rated youth are provided in appendix F in Table F.6. Details about the demographic characteristics of the parent and teacher raters are provided in appendix F in Table F.7.

Scales with conceptually similar content were paired together for this study, and these paired scales were expected to relate strongly to one another. Correlations among scales with similar content are presented in Tables 9.15 to 9.17. Correlation coefficients were corrected for range variation (i.e., due to sampling variability, the selected sample’s SD deviates from the Normative Sample’s SD; Stauffer & Mendoza, 2001). Additionally, descriptive statistics (means, medians, and standard deviations) are also provided.

As expected, for Parent, Teacher, and Self-Report, corrected correlations revealed statistically significant, moderate-to-very strong relationships between the corresponding scales (Parent corrected r = .42 to .88, Teacher corrected r = .61 to .87, Self-Report corrected r = .60 to .98; all p < .001).

It is worth noting that the correlations between Schoolwork on the Conners 4 and Learning Problems on the Conners 3 are consistently lower (corrected r = .42 for Parent, .69 for Teacher, .60 for Self-Report) than the other relationships reported. This attenuated correlation, relative to the other scales, is likely due to the shift in item content. While the Conners 3 Learning Problems scale captured content about academic difficulties, such as specific challenges in reading, writing, or math, the Conners 4 Schoolwork scale captures difficulties about school in general that students with ADHD often report, such as submitting incomplete work or misplacing assignments. Similarly, the weaker relationship between Peer Interactions on the Conners 4 and Peer Relations on the Conners 3 (corrected r = .56 for Parent, .61 for Teacher) can also be explained by the changed content focus from the Conners 3 to the Conners 4. While the Conners 3 Peer Relations scale focused on the youth’s ability to make and maintain friendships, the Conners 4 Peer Interactions scale includes additional content about how peers perceive and interact with the youth.

Overall, the high degree of association between the two measures provides evidence for the convergent validity of the Conners 4, as it aligns closely and appears to measure the same constructs as an existing, validated measure of ADHD in youth.





Relation to the Behavior Assessment System for Children 3 (BASC–3)

The Behavior Assessment System for Children, Third Edition (BASC–3; Reynolds & Kamphaus, 2015) offers a comprehensive system for identifying, evaluating, monitoring, and remediating behavioral and emotional problems in youth. Convergent and discriminant validity were investigated with the test constructs measured by the Conners 4 and the BASC–3, as both measures contain some degree of similar content related to behavioral assessment and adaptive functioning areas in youth. Scores from scales with corresponding content from the two assessments were correlated. Ratings from the Conners 4 were collected alongside ratings from six versions of the BASC–3: Parent Ratings Scales–Child (PRS–C), Parent Ratings Scales–Adolescent (PRS–A), Teacher Rating Scales–Child (TRS–C), Teacher Rating Scales–Adolescent (TRS–A), Self-Report of Personality–Child (SRP–C), and Self-Report of Personality–Adolescent (SRP–A).

It was expected that the Conners 4 and BASC–3 scales that purport to measure similar content, such as the Conners 4 Hyperactivity and the BASC–3 Hyperactivity Scale, or the Conners 4 Anxious Thoughts scale and the BASC–3 Anxiety scale, would have moderately sized relationships, as there should be some degree of overlap between the constructs but still some unique content captured by each scale. Additionally, weak to moderate correlations were expected for the BASC–3 Learning Problems scale and the Conners 4 Schoolwork scale. While they both measure difficulties in academic settings, the BASC–3 scale has item content more closely related to learning disabilities and challenges, whereas the Conners 4 Impairment & Functional Outcome scales are geared toward difficulties at school in general that are common for youth with ADHD. On the Self-Report, negative correlations between the Conners 4 Peer Interactions and Family Life scales and the BASC–3 Interpersonal Relations and Relations with Parents scales, respectively, were expected due to the opposing directions of response (i.e., higher scores for these BASC–3 scales indicate fewer difficulties, while high scores for the Conners 4 scales represent greater difficulties); however, these scales were still expected to display a moderate correlation, as individuals are reporting on similar settings and behaviors in each scale.

Respondents (N = 71 for Parent, N = 120 for Teacher, and N = 77 for Self-Report) were administered the Conners 4 and then completed the BASC–3 within 14 days. Details about the youth being rated in this study can be found in appendix F in Table F.8, and details about the parent and teacher raters are presented in Table F.9.

Tables 9.18 to 9.20 show the results of the analyses and descriptive statistics for the comparison of Conners 4 and BASC–3 scales in terms of correlations and mean differences. Correlation coefficients were corrected for range variation (i.e., due to sampling variability, the selected sample’s SD deviates from the Normative Sample’s SD; Stauffer & Mendoza, 2001). For Parent, Teacher, and Self-Report, corrected correlations revealed a statistically significant, moderate-to-very strong relationship (Parent corrected r = .60 to .96, Teacher corrected r = .57 to .81, Self-Report corrected r = .62 to .83; all p < .001) between the corresponding scales of the two tests. The expected level of association between the two measures provides evidence for the convergent validity of the Conners 4, as it aligns closely and appears to measure the same constructs as an existing, validated test of behavioral and emotional problems in children and adolescents. However, the degree of association varies by scale and rater form (ranging from 8% to 68% unshared variance), supporting the use of these measures as complementary, but not redundant, tests.



Click to expand


Relation to the Comprehensive Executive Function Inventory (CEFI)

For evidence that the Conners 4 accurately measures the construct of executive functioning, parents, teachers, and youth were asked to complete both the Conners 4 and the Comprehensive Executive Function Inventory (CEFI®; Naglieri & Goldstein, 2013). The CEFI is a 100-item measure of behaviors associated with executive function for youth aged 5 through 18 years, with higher standard scores reflecting greater ability or functioning (M = 100, SD = 15; note that these scales are scored in the opposite direction of the Conners 4 T-scores such that higher scores indicate greater ability, rather than deficit). The CEFI is composed of items related to Attention, Emotion Regulation, Flexibility, Inhibitory Control, Initiation, Organization, Planning, Self-Monitoring, and Working Memory, as well as a total score that measures executive functioning more broadly (Full Scale score). Executive function deficits are a core feature of ADHD, and scores on the Conners 4 were expected to show significant negative (due to the opposing scoring directions) correlations with corresponding content areas measured by the CEFI. It was expected that scores from the Conners 4 Inattention/Executive Dysfunction scale would show the strongest relations with scores from the CEFI Attention, Organization, Initiation, Planning, and Working Memory scales. Scores from the Conners 4 Impulsivity scale were expected to correspond with scores from the Inhibitory Control scale on the CEFI, and scores from the Conners 4 Emotional Dysregulation score were expected to correspond with scores from the Emotion Regulation scale on the CEFI. Note that DSM ADHD Inattentive Symptoms and DSM ADHD Hyperactive/Impulsive Symptom Scales are not included in these analyses, as the items are fully subsumed by the corresponding Content Scales that are reported. Instead, the DSM Total ADHD Symptoms scale was correlated with the CEFI Full Scale score to provide an overall association between the mixture of ADHD features across the Conners 4 and the full scope of executive functioning subscales measured by the CEFI.

Raters (N = 137 for Parent, N = 122 for Teacher, and N = 54 for Self-Report) were administered the Conners 4 and were then asked to complete the CEFI within four weeks. Details about the youth being rated can be found in appendix F in Table F.10, and details about the parent and teacher raters is presented in Table F.11.

Scores on the Conners 4 and CEFI were compared via correlational analyses, and corrections were applied to the correlations to account for range variation (i.e., due to sampling variability, the selected sample’s SD deviates from the Normative Sample’s SD; Stauffer & Mendoza, 2001). These results, along with descriptive statistics for each scale, can be found in Tables 9.21 to 9.23. For all raters, corrected correlations revealed moderate to strong moderate relationships (p < .001) between the corresponding scales of the two tests. In particular, the Conners 4 Total ADHD Symptoms scale was strongly related to the CEFI Full Scale (corrected r: Parent = -.63, Teacher = -.70, Self-Report = -.68). These results provide evidence for the valid measurement of these concepts on the Conners 4 by demonstrating hypothesized and significant relationships to measures of similar constructs on another established measure.





Relation to the Weiss Functional Impairment Rating Scales (WFIRS)

The Conners 4 contains three Impairment & Functional Outcome Scales designed to capture the frequency and severity of the functional impairments commonly exhibited by youth diagnosed with ADHD, including impairments within the domains of Schoolwork, Peer Interactions, and Family Life (see chapter 1, Introduction). The association between scores on the Conners 4 Impairment & Functional Outcome Scales and scores from an established measure of functional impairment were explored to consider how well the Conners 4 Impairment & Functional Outcome Scales converge on the construct of functional impairment.

To establish evidence to demonstrate the convergent validity of the Impairment & Functional Outcome Scales that comprise the Conners 4 Parent and Self-Report, the Weiss Functional Impairment Rating Scale™ Self-Report (WFIRS-S™) and the Weiss Functional Impairment Rating Scale Parent (WFIRS-P™) were administered (see Weiss et al., 2018, for an overview of the WFIRS-S and WFIRS-P; note that a Teacher form is not available for the WFIRS, and thus the Conners 4 Teacher was not applicable to this study). Data for the WFIRS were co-collected with the Conners 4 during the same administration. The WFIRS-S and WFIRS-P were selected because they represent comprehensive indices of functional impairment that demonstrate strong psychometric properties and have been used frequently in ADHD populations (Weiss et al., 2018). Moreover, each form contains subscales that explore the same functional domains as the Conners 4 Impairment & Functional Outcome Scales. Most relevant to the Conners 4 Impairment & Functional Outcome Scales, the WFIRS examines impairment in the family, school, and social domains. For each WFIRS form, the degree of impairment is rated using a 4-point scale ranging from 0, “Never or not at all,” to 3, “Very often or very much.” Respondents may also select “Not Applicable.

Given that functional impairment is not commonly reported by respondents from non-clinical samples, and given that much of the research that has been completed to validate the WFIRS-S and WFIRS-P has been conducted among individuals diagnosed with ADHD, the Conners 4 ADHD Reference Samples were used to explore convergence between the Conners 4 Impairment & Functional Outcome Scales and the WFIRS (see chapter 7, Standardization, for a detailed description of the ADHD Reference Samples). Specifically, bivariate correlations were used to evaluate the extent to which mean raw scores on the Conners 4 Self-Report Impairment & Functional Outcome Scales associated with mean raw scores in the corresponding WFIRS-S domains, and the extent to which mean raw scores on the Conners 4 Parent Impairment & Functional Outcome Scales correlated with mean raw scores in the corresponding WFIRS-P domains. The age range of the WFIRS-S is 12 to 18 years; therefore, only youth from the ADHD Reference Sample aged 12 and up were included in the bivariate correlations between the WFIRS-S and the Conners 4 Self-Report. Raw scores were used in these analyses as the WFIRS does not provide standardized scores at this time.

For parent raters, correlations revealed statistically significant, strong-to-very strong relationships (r = .64 to .86; p < .001.) For the youth self-report, moderate to strong relationships were found (r =.52 to 62; p < .001). Thus, there is evidence to demonstrate that scores from the Conners 4 Impairment & Functional Outcome Scales tap into similar constructs as the well-established WFIRS-S and WFIRS-P, and that the Conners 4 Impairment & Functional Outcome Scales converge on the intended constructs.


Click to expand

Table 9.24. Correlations between Conners 4 Impairment & Functional Outcome Scales and WFIRS Domains

Conners 4 Scale

Parent

Self-Report

WFIRS-P Domain

WFIRS-S Domain

Schoolwork

School Learning

.64

School

.62

Peer Interactions

Social Activities

.85

Social 

.59

Family Life

Family

.86

Family

.52

Note. Parent N = 558-560; Self-Report N = 140. All correlations significant, p < .001. Guidelines for interpreting |r|: very weak < .20, weak = .20 to .39, moderate = .40 to .59, strong = .60 to .79, very strong ≥ .80.


< Back Next >