846602

research-article2019
CCP0010.1177/1359104519846602Clinical Child Psychology and PsychiatryIzzo et al.

Psychometrics
Clinical Child Psychology
and Psychiatry
The Conners 3–short forms: 2019, Vol. 24(4) 791­–808
© The Author(s) 2019
Evaluating the adequacy of Article reuse guidelines:
sagepub.com/journals-permissions
brief versions to assess ADHD DOI: 10.1177/1359104519846602
https://doi.org/10.1177/1359104519846602
journals.sagepub.com/home/ccp
symptoms and related problems

Viola Angela Izzo1 , Maria Anna Donati2,

Federica Novello3, Dino Maschietto3 and
Caterina Primi1
1NEUROFARBA Department, Section of Psychology, University of Florence, Italy
2Department of Developmental and Social Psychology, Sapienza University of Rome, Italy
3AULSS 4, Italy

Abstract
The Conners’ Rating Scales are widely used to assess attention deficit/hyperactivity disorder
(ADHD) and related difficulties in children and adolescents. A short form of the scales is available,
which, along with the several advantages of brief versions, also displays good psychometric
properties. Nonetheless, no studies have confirmed them in cultural contexts different from
the original one. The present study examined the psychometric properties of the Self-Report,
Parent, and Teacher Conners 3–Short Forms in terms of reliability and validity in an Italian sample.
Analyses were performed on 591 children and adolescents, 631 parents’ ratings, and 325 teachers’
ratings. To test for discriminative validity, ADHD clinical samples of 55 youth, 63 parents, and
15 teachers were compared to gender- and age-matched groups. Findings confirmed the original
multidimensional structures and supported the Conners 3–Short Form scales as reliable and valid
tools to assess ADHD and its main comorbid conditions.

Keywords
Attention deficit/hyperactivity disorder, Conners 3, short form, brief scale, psychometric
properties

Attention deficit/hyperactivity disorder (ADHD) is a neurodevelopmental disorder whose essential

feature is a persistent pattern of inattention, hyperactivity, and impulsivity that interferes with
functioning or development (American Psychiatric Association [APA], 2013). Nonetheless, it is
often associated with other cognitive, social, and emotional impairments, including difficulties in
executive functioning, poorer school performance, adverse peer and family relationships, and
aggressive behavior (APA, 2013; Barkley, 1997; Capano, Minden, Chen, Schachar, & Ickowicz,

Corresponding author:
Viola Angela Izzo, NEUROFARBA Department, Section of Psychology, University of Florence, via di San Salvi 12 -
Padiglione 26, 50135 Firenze, Italy.
Email: [email protected]
792 Clinical Child Psychology and Psychiatry 24(4)

2008; Spencer, Biederman, & Mick, 2007; Willcutt, Doyle, Nigg, Faraone, & Pennington, 2005).
ADHD is one of the most diagnosed disorders in childhood and adolescence, whose worldwide
prevalence in school-aged children under 18 years old is estimated to be around 7% (Thomas,
Sanders, Doust, Beller, & Glasziou, 2015). Focusing on Europe, prevalence seems to be lower than
in the United States (Anderson, 1996; Timimi & Taylor, 2004). For instance, Spanish prevalence is
estimated to be around 6.8% (Catalá-López et al., 2012), prevalence in Greece is about 6.0%
(Skounti et al., 2010), German prevalence is about 4.8% (Huss, Hölling, Kurth, & Schlack, 2008),
French prevalence of ADHD is between 3.5% and 5.6% (Lecendreux, Konofal, & Faraone, 2011),
and Italian prevalence is estimated to reach 4% (http://old.iss.it/binary/adhd/cont/Registro_nazion-
ale_dell_ADHD_2007_2016.pdf).
Given the negative impact of ADHD symptoms and its associated complications, it is crucial to
dispose of valid, reliable but nevertheless brief instruments that may help clinicians identifying not
only ADHD symptoms, but also the most frequent comorbid difficulties evidenced in literature that
complicate the youth’s clinical profile. The diagnostic process should never be based only on rating
scales, as a thorough clinical assessment including observations and interviews is needed. In fact,
the clinical guidelines suggest that information should be obtained primarily from reports from
parents or guardians, teachers, and mental health clinicians involved in the child’s care
(Subcommittee on Attention-Deficit/Hyperactivity Disorder et al., 2011). However, when com-
bined with clinical expertise and other clinical tools, the Conners’ Rating Scales (Conners, 1989,
1997, 2008) may represent a remarkable option to assist clinicians in ADHD diagnosis and screen-
ing. The advantages of the Conners scales have been supported by both studies confirming the
excellent psychometric properties of the scales (e.g. Arias Martínez, Arias González, & Gómez
Sánchez, 2013; Christiansen et al., 2016; Conners, Sitarenios, Parker, & Epstein, 1998; Izzo,
Donati, & Primi, 2018a, 2019; Morales-Hidalgo, Hernández-Martínez, Vera, Voltas, & Canals,
2017) and clinical studies. In fact, research has demonstrated the effectiveness of the Conners
scales in distinguishing people with ADHD from those with no clinical diagnoses; moreover, the
Conners scales have been found to have predictive validity for the development of behavioral
problems in adolescence (e.g. Gau, Soong, Chiu, & Tsai, 2006; Izzo, Donati, & Primi, 2018b;
Lindquist, Carlsson, Persson, & Uvebrant, 2006; Mannuzza, Klein, Abikoff, & Moulton Iii, 2004;
Rucklidge, 2006).
Despite the strengths of the Conners scales, the most recent edition included about 100 items for
all three versions, filled in by parents (Conners 3–Parent; Conners 3-P), teachers (Conners 3–Teacher;
Conners 3-T), and youth (Conners 3–Self-Report; Conners 3-SR) (Conners, 2008). Thus, some alter-
native forms of the scales have been developed to dispose of shorter assessment tools: two indices
and the Conners 3 short forms. Nevertheless, the indices (i.e. the Conners 3 ADHD Index and the
Conners 3 Global Index; Arias Martínez et al., 2013; Conners, 2008; Morales-Hidalgo et al., 2017)
are structurally different from the Conners 3–long forms. In fact, the Conners 3 ADHD Index only
assesses Inattention and Hyperactivity/Impulsivity, instead of presenting a clear multidimensional
structure assessing ADHD and its most frequent comorbid conditions; and the Conners 3 Global
Index is a general measure of psychopathology, instead of focusing on ADHD.
Instead, the short forms include the same Content scales present in the long form. In fact, they
represent a restricted version of the long ones, including five to six items per each Conners 3–long
form Content scale. The Content scales composing the short form of the Self-Report version
(Conners 3–Self-Report Scale–SF, Short Form) are Inattention, Hyperactivity/Impulsivity,
Learning Problems, Defiance/Aggression, and Family Relations. The Content scales composing
the short form of the Parent version (Conners 3–Parent Scale–SF, Short Form) are Inattention,
Hyperactivity/Impulsivity, Learning Problems, Executive Functioning, Defiance/Aggression, and
Peer Relations. Eventually, the scales composing the short form of the Teacher (Conners 3–Teacher
Izzo et al. 793

Scale–SF, Short Form) version are Inattention, Hyperactivity/Impulsivity, Learning Problems/

Executive Functioning, Defiance/Aggression, and Peer Relations.
If the Conners 3 long scales represent a remarkable option in the clinical field, the short form of
the Conners 3 scales may be an even better solution, since they allow to gather a great amount of
clinical information about school-aged children and adolescents from parent, teacher, and youth,
though presenting a restricted pool of items. In fact, the Conners 3–short form scales combine the
advantages of a brief assessment with the existence of a multifactorial structure that allows clini-
cian to delineate an accurate, precise profile: instead of providing a unique total score, the multidi-
mensionality of the scales allows to discriminate whether the child’s difficulties would mainly
affect the inattention area rather than the hyperactive/impulsive or the aggression spheres, thus
enabling to identify which domain(s) may require an intervention. Notably, the dimensions are
correlated between them, meaning that, though allowing to investigate the specific problematic
area(s), the constructs examined by the scales are related each other, in line with the clinical defini-
tion of ADHD as a neurodevelopmental disorder often associated with difficulties in the areas of
executive functioning, learning, aggression, and social relationships (APA, 2013). This structure
pattern gives support to the strength of the Conners scales as a measure aiming to assess ADHD
and its most frequently related impairments.
Unlike the previous brief versions (Conners, 1989, 1997), the short forms of the last edition of
the Conners (2008) appear to be a valuable short instrument tool to measure ADHD symptoms and
related problems as their items have been submitted to a thorough statistical procedure which
allowed to select the most effective items from the full-length scales. In fact, according to the item
discrimination analyses, the means and standard deviations of various clinical groups (i.e. ADHD,
Major Depressive Disorder, Bipolar Disorder, Anxiety Disorders, Learning Disorders, and Conduct
and Oppositional Defiant Disorders) were compared to the general population and among all those
clinical groups. If the item did not discriminate well between the target clinical group and both the
general population and the other clinical groups, the item was considered for exclusion from the
short form. Moreover, item response theory (IRT) analyses were performed; IRT models provide a
mathematical expression of the latent trait, theta (θ), which includes two parameters, namely a, that
is, the ability of the item to discriminate between people with varying levels of the trait, and b, that
is, the item location, which is the trait level required for a person to be as likely to choose a particu-
lar response category as to not respond in this category (Ostini & Nering, 2006). IRT also allows
to obtain an estimate of the amount of information provided by each item, summarized with item
information curves (IIC). Items that provided the least amount of discrimination (i.e. items with the
smallest a parameter estimates), items with unusual b parameters (i.e. no gradual movement from
one response option to the next), and items that provided the least amount of information (i.e. flat
IICs) were considered for exclusion from the final solutions. Finally, the five items with the highest
factor loadings from the exploratory analysis conducted on the long scales were regressed onto the
full-length scale score to maximize the multiple correlation between the five items and the full-
length scale score to ensure that the shortened form would account for the maximum amount of
variance in the full-length score. The listed analyses resulted in five to six items per each scale, so
the short form of the Conners 3-P had 45 items, and the short forms of both the Conners 3-T and
the Conners 3-SR had 41 items.
When examining their psychometric properties, the original short scales resulted to have ade-
quate reliability in terms of internal consistency. In fact, the mean Cronbach’s alpha for the Conners
3–Parent Scale–SF was .89, ranging from .85 to .92; the mean Cronbach’s alpha for the Conners
3–Teacher Scale–SF was .91, ranging from .87 to .94; and the mean Cronbach’s alpha for the
Conners 3–Self-Report Scale–SF was .83, ranging from .77 to .89. Moreover, the scales showed
good factorial validity, with results from the confirmatory factor analyses (CFAs) revealing that the
794 Clinical Child Psychology and Psychiatry 24(4)

six-factor model for the Conners 3-P and the five-factor models for both the Conners 3-SR and the
Conners 3-T had adequate fit—that is, comparative fit index (CFI) values were higher than .90,
root mean square error approximation (RMSEA) values were lower than .08, and all parameter
estimates and all correlations were significant for all three versions. Also, good validity in terms of
both across-informant correlations (expressed as the correlation between different informants’ rat-
ings of the same youth) was found, with mean parent to teacher correlation equal to .59 (ranging
from .50 to .66), mean parent to youth correlation equal to .57 (ranging from .50 to .66), and mean
teacher to youth correlation equal to .49 (ranging from .42 to .57). Eventually, discriminative valid-
ity, assessed with the ability to distinguish between clinical and non-clinical samples, was sup-
ported, being the means for the clinical group significantly higher than both the general population
and other clinical groups, and being sensitivity and specificity values acceptable to high for all
three versions (Conners, 2008).
The psychometric properties of the full-length forms of the Conners 3 scales have been previ-
ously tested and confirmed in cultural contexts different from the original one (Christiansen et al.,
2016; Izzo et al., 2018a, 2019). Differently, the psychometric properties of the short forms of the
Conners 3 have not been tested by any other study but the original one—in both the English ver-
sion and the Spanish version, which was validated for the Spanish-speaking population of the
United States (Conners, 2008). Thus, the excellent features characterizing the original edition have
not been confirmed by other studies yet.
With these premises, the present study aimed at assessing the psychometric properties of the
Conners 3–short form scales in the Italian version (Primi & Maschietto, 2017), as to validate them
for the Italian context. The first goal was to find evidence of the internal structure of the scales by
performing a CFA, as to probe the replicability of the original factors and give support to their
multidimensional structure, in line with the original work (Conners, 2008). However, since any
other studies but the original one have verified the structure of the short forms of the Conners 3
scales, we decided to compare three different competing models for all versions of the scales, as to
test whether the original structures were in fact the best fitting solutions. The first model replicated
the original multidimensional structure (Conners, 2008). Then, since in the original work the cor-
relations between the different dimensions of the short scales were particularly high (Conners,
2008), two more models were performed: a one-factor model that considered the Conners 3 scales
as being unidimensional, and a second-order factor model that hypothesized the existence of a
second-order factor such as ADHD, which may break down into the dimensions originally found.
We also aimed to assess the reliability of the scales in terms of internal consistency as a meas-
ure of the accuracy of the test. Moreover, this study aimed to assess the across-informant correla-
tions between the different versions (parent–teacher ratings, parent–youth rating, and teacher–youth
ratings) as an evidence of the validity of the Conners 3–short form scales. In fact, since the Self-
Report, Parent, and Teacher versions all measure similar constructs, concordance in scores across
the different informants would provide support for the validity of the Conners 3–short form scales.
However, considering that the different sources of information assess the child from different
points of view—that is, the parent delineates a profile based on how the son behaves at home, the
teacher describes the student according to what can be observed in class, and the child responds
consistent with his or her personal perceptions and insights—a certain degree of incongruence
between informants should be expected, so correlations should be moderate in size (Conners,
2008). In fact, it is not unusual that informant discrepancies occur (Achenbach, 2006; Achenbach,
McConaughy, & Howell, 1987; De Los Reyes et al., 2015). In details, we expected to find correla-
tions between informants around .36, in line with previous literature reporting the level of multi-
informant correspondence when referring to externalizing problems (De Los Reyes et al., 2015).
In addition, validity in terms of discriminative validity was examined by comparing a clinical
Izzo et al. 795

sample with a diagnosis of ADHD with a non-clinical sample, with the expectation that the scales
were able to discriminate between these two groups.

Method
Participants
Six schools (two primary schools, two middle schools, and two high schools) were randomly
selected from the schools present in the North-Center of Italy. However, one middle school declined
to participate because it was already involved in other projects. Eventually, participants were 599
youths (53% males, age ranged from 8 to 18 years), 631 youth’s parents (49% of youth was male,
with age ranging from 6 to 18 years; 84% of the rating parents were mothers), and 409 youth’s
teachers (53% of youth was male, with age ranging from 6 to 18 years; 98% of the rating teachers
were women). Some questionnaires were excluded because the percentage of omitted responses
exceeded the allowable number reported in the original manual, which is equal to one item for all
the Content scale (Conners, 2008). Thus, the final youth sample included 591 children and adoles-
cents (54% males), attending primary (41%), middle (33%), and high schools (26%). Concerning
the parent sample, instead, all questionnaires were completed or with an acceptable number of
omitted responses, so the final sample included parent evaluations of 631 children and adolescents
(49% of youth was male) attending primary (58%), middle (30%), and high schools (12%).
Eventually, for the teacher sample, less than 80% of them returned the questionnaire completed or
with an acceptable percentage of missing answers, so the final sample included teacher ratings of
325 youth (51% males; all rating teachers were women) belonging to primary (63%), middle
(25%), and high schools (12%).
As to measure discriminative validity, a clinical sample of youth with a diagnosis of ADHD—and
their parents and teachers—was involved in this study. Participants were 65 youth attending a clinical
center in the North of Italy specialized in ADHD and other neuropsychological disorders, as well as
their parents and teachers. After the analysis of missing cases, the study included ratings from a sam-
ple of 55 youth with ADHD (93% males aged 8–18 years), though 50% of them (n = 28) had an addi-
tional diagnosis. Specifically, nine children had Oppositional Defiant Disorder (32%), nine had a
Learning Disorder (32%), six had an Anxiety Disorder (21%), two had a mood disorder (7%), one
had a Limited Cognitive Functioning (4%), and one had Motor Coordination Disorder (4%). For the
Parent version, after deleting missing cases, the final sample included 63 youth’s parent ratings (95%
of rated youth was male, with age ranging from 6 to 18 years). Eventually, for the Teacher version,
only 23% of them (n = 15) returned the protocol completed or with an acceptable percentage of miss-
ing answers (100% of rated youth was male, with age ranging from 6 to 18 years).
The clinical samples were compared to gender and age-matched non-clinical samples for the
Conners 3–Self-Report scale (n = 55, 93% males), for the Conners 3–Parent scale (n = 63; 95%
males), and for the Conners 3–Teacher scale (n = 15; 100% males). Descriptive statistics of the
samples are reported in Table 1.

Instruments
The short forms of the Self-Report, Parent, and Teacher Conners 3 Rating Scales (Italian version:
Primi & Maschietto, 2017) include items aiming to measure youth’s ADHD symptoms and their
most frequent associated complications, that is, learning problems, defiance/aggression, and diffi-
culties in social relationships. Content scales vary in number according to the specific version: the
Conners 3–Self-Report Scale–SF contains 27 items loading onto the five Content scales, the
796 Clinical Child Psychology and Psychiatry 24(4)

Table 1.  Descriptive statistics of the samples to examine the properties of all versions of the Conners 3
scales.

General sample Non-clinical sample Clinical sample

Conners Males/females 320/271 (N = 591) 51/4 (N = 55) 51/4 (N = 55)
3-SR Age (range) M = 12.26, M = 12.08, M = 12.21,
SD = 2.96 (8.00–18.91) SD = 2.50 (8.00–17.58) SD = 2.38 (7.75–17.58)
Conners Males/females 312/319 (N = 631) 60/3 (N = 63) 60/3 (N = 63)
3-P Age (range) M = 10.75, M = 10.92, M = 11.58,
SD = 3.05 (6.00–18.75) SD = 2.82 (6.16–18.33) SD = 2.82 (6.00–17.58)
Conners Males/females 165/160 (N = 325) 15/0 15/0
3-T Age (range) M = 10.47, M = 11.09, M = 11.11,
SD = 3.13 (6.00–18.75) SD = 3.02 (7.41–17.33) SD = 3.00 (7.33–17.16)

SR: Self-Report; P: Parent; T: Teacher.

Conners 3–Parent Scale–SF includes 31 items on its six Content scales, and the Conners 3–Teacher
Scale–SF includes 27 items on its five Content scales. Items are rated on a Likert-type scale rang-
ing from 0 (not true at all/never) to 3 (very much true/very frequently).

Procedure
Participants were recruited in schools; to obtain the approval of this research, a study protocol in
accordance with the criteria of the Declaration of Helsinki was reviewed and approved by each
Head Teacher and school board of different primary, middle, and high schools of North-Center
Italy. Parents and teachers were informed with a short study description and asked to provide their
informed consent; furthermore, parents were asked to provide the informed consent to allow their
children to take part in the research. Written informed consent was obtained from all participants
(and from participants’ parents in case of underage youth), and their confidentiality was ensured.
All participants completed their questionnaire through a self-administered procedure, though par-
ents completed it at home, whereas teachers and students completed it in class, during school-time.
The time needed to complete the scales was approximately 10 minutes.
Concerning the clinical sample, parents of the patients attending the center were informed with
a short study description and asked to provide their informed consent. Furthermore, they were
asked both to provide the informed consent to allow their children to take part in the research, and
to involve one of the child’s teachers and to give them an informed consent as to obtain also a
teacher rating. Written informed consent was obtained from all participants (and from participants’
parents in case of underage youth), and their confidentiality was ensured. All participants com-
pleted their questionnaire through a self-administered procedure.

Statistical analyses
A preliminary handling of missing values was performed, with values of omitted answers being
estimated through the application of a formula calculating a prorated score, as referred in the
original manual (Conners, 2008).1 To examine the internal structure of the Conners 3-SR,
Conners 3-P, and Conners 3-T short forms, CFAs employing the maximum likelihood (ML)
estimator (AMOS software) were conducted on the Content scales, with factor covariances
being analyzed; items were loaded onto their respective scales, and the scales correlated with
Izzo et al. 797

each other. Adequate fit for the model was defined according to the ratio of chi-square to its
degrees of freedom (χ2/df), the Tucker–Lewis Index (Tucker & Lewis, 1973), the CFI (Bentler,
1990), the Steiger–Lind Root Mean Square Error of Approximation Index (RMSEA; Steiger &
Lind, 1980), and the standardized root mean square residual (SRMR; Bentler, 1995). In details,
adequacy of the model was defined as TLI and CFI values equal to .90 or greater, RMSEA values
of .06 or below, and SRMR values of .08 or below (Hu & Bentler, 1999). As to compare different
CFA models to test which model had the best fit to data—that is, either the original structure, or
the unidimensional model, or the second-order factor model—the following indices were exam-
ined: the Akaike Information Criterion (AIC; Akaike, 1974), the Consistent version of AIC
(CAIC; Akaike, 1974), and the Bayesian Information Criterion (BIC; Schwarz, 1978). The AIC
is derived from information theory and is generally used when comparing non-nested or non-
hierarchical models estimated with the same data. The CAIC is a modified version of the AIC
which adjusts for sample size. The BIC is another criterion for model selection that measures the
trade-off between model fit and complexity of the model. Note that none of those criteria inform
us directly about the quality of the model; we use them when we want to compare models and
determine which of several models is the best, with smaller values suggesting a better fitting
model (Busemeyer & Diederich, 2014).
To assess reliability, internal consistency of the Content scales through the Cronbach’s
alphas was obtained. To examine across-informant agreement as a validity measure, correlation
coefficients (Pearson’s r) were calculated among the three versions. In addition, to test for dis-
criminative validity (i.e. to test how much the Conners 3 scales were able in distinguishing
between youth with and without ADHD), a receiver operating characteristic (ROC) curve using
those two groups was plotted for each Content scale of each version of the Conners 3 sepa-
rately. The area under curve (AUC), sensitivity, and specificity were calculated for each Content
scale of each version. As a “gold standard” to examine sensitivity and specificity of the scales,
participants were either diagnosed with ADHD or not based on the scores on the second edition
of the Conners (1997) and on a clinical interview. According to Marôco (2018), accuracy of
sensitivity and specificity were good when values were higher than .80, and acceptable between
.50 and .80. Moreover, t test comparisons were performed to compare the clinical and the non-
clinical sample on the scores obtained for each Content scale of the Self-Report, Parent, and
Teacher versions.

Results
Preliminary item-level statistics showing the endorsement rates for each of the Likert-type catego-
ries were performed (Table 2). Overall, items showed a good level of variability in the answers—
suggesting a normal distribution, in line with values of both skewness and kurtosis—except for
some items showing very low variability. However, these results are in line with the content of the
involved items, since the critical items mostly belonged to the Defiance/Aggression scale (i.e. “I
bully or threaten other people” and “He/She starts fights with other people”).

Structural validity
Concerning the Conners 3–Self-Report Scale–SF, a model replicating its original structure was
performed. When comparing this model to both a one-factor model and a second-order factor
model, goodness-of-fit indicators suggested that the original model had the best fit to data (Table
3). Each item loaded strongly and significantly on its hypothesized factor (Table 4), and the cor-
relations between the five factors were all significant, ranging from .31 to .81 (Table 5).
Table 2.  Item analysis showing skewness, kurtosis, and the response rates (%) for each of the Likert-type categories. 798

Conners 3-SR Conners 3-P Conners 3-T

  Item Skewness Kurtosis Likert-type categories Item Skewness Kurtosis Likert-type categories Item Skewness Kurtosis Likert-type categories
(SE = .10) (SE = .20) (SE = .10) (SE = .19) (SE = .14) (SE = .27)
  0 1 2 3 0 1 2 3 0 1 2 3

Inattention 3 0.98 0.14 50.6 32.3 13.5 3.6 17 0.50 –0.29 28.7 46.4 20.0 4.9 3 1.12 .47 52.6 31.4 11.4 4.6
5 0.84 –0.02 42.5 38.1 14.2 5.2 27 1.14 0.49 53.6 30.3 11.3 4.9 10 1.28 .77 56.3 28.3 9.2 6.2
14 1.03 0.39 48.7 35.5 11.5 4.2 30 1.11 0.58 49.8 35.0 10.0 5.2 26 0.52 –.50 28.9 42.5 20.3 8.3
16 0.31 –0.68 21.5 42.1 26.2 10.2 34 0.91 0.33 43.3 41.0 11.7 4.0 36 0.95 –.17 49.5 28.6 13.8 8.1
18 0.96 0.28 48.2 36.4 12.2 3.2 41 1.15 0.82 53.1 34.7 9.4 2.9 39 0.93 –.10 46.5 32.6 13.2 7.7
27 1.01 0.21 47.7 34.3 12.0 5.9  
Hyperactivity/ 1 0.77 –0.29 42.5 35.4 16.6 5.6 3 1.52 1.48 65.8 21.7 9.5 3.0 1 2.04 3.80 75.1 17.5 5.8 1.5
Impulsivity 4 0.93 –0.30 50.4 26.4 14.6 8.6 5 1.83 2.70 72.4 18.4 7.4 1.8 8 2.56 6.30 81.8 11.7 4.9 1.5
15 1.42 1.11 61.4 24.2 9.0 5.4 7 1.96 3.44 74.6 18.4 5.9 1.1 22 2.58 6.25 81.8 11.1 4.9 2.2
29 1.09 0.31 52.3 30.3 12.0 5.4 13 3.34 11.58 88.7 8.1 2.9 0.3 24 1.91 3.29 70.8 20.9 5.5 2.8
31 0.41 –0.90 29.1 36.0 21.7 13.2 24 0.71 –0.60 43.3 30.1 18.4 8.2 30 1.53 1.60 67.4 22.2 8.9 1.5
  28 1.19 0.90 55.8 33.0 9.0 2.2 32 2.03 3.83 75.7 18.2 5.2 0.9
Learning 9 1.42 1.33 61.9 26.1 9.0 3.0 8 1.59 1.93 65.0 24.2 7.4 3.4  
Problems 13 1.08 0.13 54.3 26.6 13.5 5.6 10 0.90 0.24 45.0 39.1 12.4 3.5  
33 1.59 2.06 65.8 24.9 7.1 2.2 25 1.90 3.23 71.6 20.4 5.9 2.1  
36 1.16 0.50 51.3 32.3 9.1 7.3 36 1.03 0.08 54.0 27.6 14.4 4.0  
38 0.77 –0.75 48.2 23.5 16.4 11.8 39 1.87 3.06 72.4 19.7 6.3 1.6  
Executive 1 2.18 4.67 78.6 17.3 3.6 0.5 14 1.26 .79 56.6 28.6 10.2 4.6
Functioninga   15 1.51 1.64 63.1 25.4 7.9 3.6 16 1.35 1.06 59.1 27.1 9.2 4.6
  20 1.06 0.71 48.5 38.5 9.5 3.5 18 1.25 1.00 58.8 30.5 8.9 1.8
  32 1.41 1.41 62.6 26.9 8.4 2.1 23 0.22 –.79 12.6 43.4 27.4 16.6
  35 0.71 –0.30 38.8 38.4 17.0 5.9 28 1.07 .12 53.5 27.1 13.2 6.2
  35 1.06 .35 52.0 32.0 12.3 3.7
Defiance/ 8 4.13 18.92 91.5 6.4 1.7 .3 14 6.37 47.50 96.0 3.2 0.6 0.2 5 6.04 40.10 95.1 3.4 0.6 0.9
Aggression 10 3.04 9.06 85.8 8.5 3.7 2.0 19 3.86 15.42 92.7 6.7 0.6 0 7 2.42 5.87 81.8 14.8 3.1 3.0
17 2.04 3.31 75.8 13.7 7.1 3.4 21 4.86 27.13 93.8 5.1 1.0 0.2 12 2.02 9.25 85.5 9.5 3.7 1.2
30 1.33 1.19 56.0 31.6 8.0 4.4 23 7.43 67.38 96.8 2.7 0.3 0.2 15 5.94 41.99 95.1 4.0 0.6 0.3
32 4.41 21.51 91.7 5.9 1.5 .8 26 1.22 1.27 58.5 34.7 5.7 1.1 17 3.72 14.27 91.7 7.1 1.2 0
39 3.86 17.02 88.2 9.6 .8 1.4  
Relationsb 19 1.96 3.23 71.4 19.3 5.2 4.1 4 1.88 3.10 69.9 21.6 5.5 3.0 11 1.59 1.77 67.4 21.2 8.9 2.5
22 0.90 –0.11 44.3 34.7 13.2 7.8 6 3.03 10.31 857 11.9 1.9 0.5 20 2.18 4.65 75.7 17.8 4.6 1.8
26 1.84 3.20 69.4 23.4 5.1 2.2 18 2.81 7.42 86.8 10.5 2.7 0 27 1.84 2.79 69.2 20.9 5.8 4.0
34 1.35 1.06 59.9 26.6 9.6 3.9 38 4.08 18.11 91.1 6.2 2.1 0.6 29 2.17 4.63 74.8 18.8 4.3 2.2
37 1.92 3.16 69.7 21.3 4.7 4.2 43 2.22 4.71 76.4 17.0 4.8 1.9 37 2.10 4.49 74.2 20.3 4.0 1.5

SR: Self-Report; P: Parent; T: Teacher.

Likert-type categories: 0 = not true at all; 1 = just a little true; 2 = pretty much true; 3 = very much true.
aExecutive Functioning scale for the Conners 3-P scale and Learning Problems/Executive Functioning scale for the Conners 3-T scale.
Clinical Child Psychology and Psychiatry 24(4)

bPeer Relations for both the Conners 3-P and the Conners 3-T scales, and Family Relations for the Conners 3-SR scale.
Izzo et al. 799

The same procedure was conducted for the Conners 3–Parent Scale–SF, so a model testing the
original structure was compared to both a one-factor model and a second-order factor model.
Nonetheless, results showed that the original model had the best fit (Table 3). Factor loadings were
all greater than .30 (p < .001; Table 4), and the correlations between the six factors were all signifi-
cant, ranging from .13 to .93 (Table 5).
Eventually, to test for the structural validity of the Conners 3–Teacher Scale–SF, three CFAs on
a model replicating its original structure, on a one-factor model, and on a second-order factor
model were performed. Once again, results supported the original model as having the best fit to
data (Table 3). Each item loaded strongly and significantly on its hypothesized factor (Table 4), and
the correlations between the five factors were all significant, ranging from .37 to .89 (Table 5).

Reliability
Concerning reliability, we aimed to measure the internal consistency of the content scales. For
what concerns the Conners 3–Self-Report Scale–SF, Cronbach’s alphas showed an acceptable
internal consistency with alpha’s values ranging from .63 to .79 (Table 6). For the Conners 3–Parent
Scale–SF, the Content scales showed acceptable to high levels of reliability (ranging from .72 to
.89), except for the Defiance/Aggression scale (α = .47). Finally, for the Conners 3–Teacher Scale–
SF, Cronbach’s alphas were high, with values ranging from .84 to .94.

Validity
As a measure of validity in terms of across-informant correlations, correlation coefficients were
calculated among the three versions. We performed across-informant correlations both using the
total summed scores (Table 7), and estimating the correlations among latent variables in a series of
CFA models—one for each Content scale of the Conners 3 scales (see Supplementary Appendix 1).
In both cases, the correlations were found to be all significant and moderate in size, indicating that
there was a good deal of consistency between different informants’ ratings of the same youth
across the Conners 3–short form scores. Thus, the correlations were not high enough to suggest
redundancy in collecting all three assessments.
As to measure discriminative validity, we compared the clinical sample with a non-clinical sam-
ple. Results for the Conners 3–Self-Report–SF showed that the scale had good discriminative valid-
ity. In fact, sensitivity and specificity values were acceptable, with sensitivity values ranging between
.55 and .80, and specificity values ranging between .53 and .71. Moreover, children with a diagnosis
of ADHD scored themselves as having more problems than the non-clinical sample on all the Content
scales, except for the Family Relations scale, which did not present a significant difference between
groups. The Cohen’s d effect sizes were large (Cohen, 1988), with values ranging from .77 to 1.24,
except for the Family Relations scale value, which was small (d = .22) (Table 8).
Concerning the Conners 3–Parent Scale–SF, results revealed an excellent discriminative valid-
ity, with sensitivity values ranging between .70 and .87, and specificity values ranging between .65
and .87. Furthermore, parents of the clinical group assessed their children as having more problems
than parents of the non-clinical sample. The effect sizes were large (Cohen, 1988), ranging from
1.01 to 1.95 (Table 8).
In addition, sensitivity and specificity values for the Conners 3–Teacher Scale–SF were accepta-
ble, with sensitivity values ranging between .51 and .73, and specificity values ranging between .66
and .80 (Table 8). Moreover, results showed a significant difference between the clinical and the non-
clinical samples only for the Inattention (t(28) = –2.70; p < .05; d = .99) and the Learning Problems/
Executive Functioning Content scales (t(28) = –2.38; p < .05; d = .87), with teachers giving higher
 800

Table 3.  Goodness-of-fit statistics for the confirmatory-factor-analysis models for the Conners 3 scales–short forms.

Scale Model χ2(df) p χ2/(df) TLI CFI RMSEA SRMR AIC BIC CAIC
Conners 3-SR Original model 650.87 (314) <.001 2.07 90 .91 .04 (.04–.05) .03 778.87 1059.31 1123.31
One-factor model 2212.47 (350) <.001 6.32 .50 .51 .10 (.09–.10) .09 2268.47 2391.16 2419.16
Second-order model 809.19 (323) <.001 2.51 .86 .87 .05 (.05–.06) .05 919.19 1160.19 1215.19
Conners 3-P Original model 1132.82 (419) <.001 2.70 .89 .90 .05 (.05–.06) .03 1286.82 1629.27 1706.27
One-factor model 2551.55 (434) <.001 5.88 .68 .70 .09 (.09–.11) .04 2675.55 2951.28 3013.28
Second-order model 1880.45 (433) <.001 4.34 .78 .80 .07 (.07–.08) .07 2006.65 2286.83 2349.83
Conners 3-T Original model 992.27 (314) <.001 3.16 .90 .91 .08 (.08–.09) .03 1120.27 1362.43 1426.43
One-factor model 4640.45 (350) <.001 13.26 .41 .41 .20 (.19–.20) .15 4696.45 4802.40 4830.40
Second-order model 1386.89 (323) <.001 4.29 .84 .85 .10 (.10–.11) .13 1496.89 1705.00 1760.00

TLI: Tucker–Lewis Index; CFI: comparative fit index; RMSEA: root mean square error of approximation; SRMR: standardized root mean square residual; AIC: Akaike informa-
tion criterion; BIC: Bayesian information criterion; CAIC: Consistent version of AIC; SR: Self-Report.
Clinical Child Psychology and Psychiatry 24(4)
Izzo et al. 801

Table 4.  Factor loadings for all the three versions of the scales.

Factor Conners 3-SR Conners 3-P Conners 3-T

Items Loadings Items Loadings Items Loadings
Inattention 3 .59 17 .81 3 .87
5 .50 27 .87 10 .91
14 .57 30 .65 26 .82
16 .58 34 .61 36 .89
18 .73 41 .89 39 .89
27 .78  
Hyperactivity/Impulsivity 1 .41 3 .74 1 .88
4 .51 5 .65 8 .83
15 .49 7 .40 22 .85
29 .65 13 .62 24 .80
31 .64 24 .75 30 .64
  28 .76 32 .74
Learning Problems 9 .54 8 .50  
13 .32 10 .61  
33 .63 25 .62  
36 .47 36 .73  
38 .71 39 .40  
Executive Functioninga 1 .78 14 .73
  15 .45 16 .85
  20 .81 18 .77
  32 .54 23 .42
  35 .45 28 .70
  35 .83
Defiance/Aggression 8 .55 14 .59 5 .79
10 .60 19 .31 7 .72
17 .65 21 .39 12 .68
   
30 .44 23 .51 15 .79
32 .53 26 .37 17 .70
39 .40  
Relationsb 19 .62 4 .66 11 .80
22 .69 6 .65 20 .86
26 .56 18 .62 27 .84
34 .61 38 .50 29 .87
37 .79 43 .62 37 .90

SR: Self-Report; P: Parent; T: Teacher.

aExecutive Functioning scale for the Conners 3-P scale and Learning Problems/Executive Functioning scale for the

Conners 3-T scale.

bPeer Relations for both the Conners 3-P and the Conners 3-T scales, and Family Relations for the Conners 3-SR scale.

scores to those students belonging to the clinical sample rather than to the non-clinical group.
However, given the small sample dimension, an evaluation of the effect sizes for the other three
Content scales could be highly informative as an indicator of discriminative validity. As a result, the
effect size for the Defiance/Aggression scale was small (d = .18), whereas the effect size for both the
Hyperactivity/Impulsivity and Peer Relations scales was medium (respectively d = .55 and d = .58).
802 Clinical Child Psychology and Psychiatry 24(4)

Table 5.  Factor correlations for the Conners 3 scales.

Conners 3-SR Inattention Hyperactivity/ Learning Problems Defiance/ Family

Impulsivity Aggression Relations
Inattention —  
Hyperactivity/ .71 —  
Impulsivity
Learning Problems .81 .57 —  
Defiance/Aggression .44 .59 .40 —  
Family Relations .35 .36 .31 .34 —

Conners 3-P Inattention Hyperactivity/ Learning Executive Defiance/ Peer

Impulsivity Problems Functioning Aggression Relations
Inattention —  
Hyperactivity/ .60 —  
Impulsivity
Learning Problems .88 .48 —  
Executive Functioning .93 .56 .90 —  
Defiance/Aggression .41 .51 .38 .42 —  
Peer Relations .28 .13 .32 .31 .39 —

Conners 3-T Inattention Hyperactivity/ Learning problems/ Defiance/ Peer

impulsivity executive functioning aggression relations
Inattention —  
Hyperactivity/ .55 —  
Impulsivity
Learning Problems/ .89 .37 —  
Executive Functioning
Defiance/Aggression .49 .72 .40 —  
Peer Relations .55 .41 .55 .50 —

SR: Self-Report; P: Parent; T: Teacher.

All correlations were significant (p < .001).

Table 6.  Cronbach’s alpha values for the Conners 3–short forms.

Scale Conners 3–Self-Report Conners 3–Parent Conners 3–Teacher

Inattention .79 .89 .94
Hyperactivity/Impulsivity .67 .82 .93
Learning Problems .63 .72 —
Executive Functioning — .76 —
Learning Problems/Executive — — .87
Functioning
Defiance/Aggression .65 .47 .84
Peer Relations — .72 .93
Family Relations .76 — —
Izzo et al. 803

Table 7.  Across-informant correlations.

Scale Parent–Self-Report Teacher–Self-Report Parent–Teacher

Inattention r = .39, p = .001 r = .26, p = .001 r = .39, p = .001
Hyperactivity/Impulsivity r = .28, p = .001 r = .24, p = .001 r = .29, p = .001
Learning Problemsa r = .52, p = .001 r = .43, p = .001 r = .47, p = .001
Executive Functioninga — — r = .32, p = .001
Defiance/Aggression r = .10, p = .037 r = .18, p = .007 r = .18, p = .001
Peer Relations — — r = .29, p = .001

SR: Self-Report; T: Teacher; P: Parent.

The Family Relations Content scale is not reported, as it is present only in the SR version.
aConners 3-T Learning Problems/Executive Functioning scale is correlated with the Conners 3-P Learning Problems and

the Executive Functioning scales and with the Conners 3-SR Learning Problems scale.

Discussion
Since their first edition, the Conners’ Rating Scales have represented one of the most used instru-
ments for helping clinicians in the assessment of ADHD in children and youth thanks to their
several positive aspects, including the multi-informant assessment, the excellent psychometric
properties in terms of reliability and validity, and their multidimensional structure, which allow
clinicians to depict a remarkably detailed representation of the child/adolescent. Moreover, one of
the most relevant advantages of the scales is the existence of both full-length and short forms, so
that clinicians can choose the form that best fits the administration conditions.
The short forms of the Conners 3 scales appear notably profitable, given that they provide reli-
able and valid clinical information when the administration of the full-length forms is not allowed
or practical, for example, due to limited time or when repeated measurements over time are
required. Moreover, when facing a restricted pool of items, participants tend to maximize the
response rates and reduce the omitted or false responses, thus improving the validity and the truth-
fulness of the results.
The Conners short forms have several advantages; nonetheless, any other studies but the origi-
nal one has tested their properties. Thus, the present study examined the psychometric properties
of the Conners 3–short forms in their Italian version, aiming to validate the Italian version of the
questionnaires.
Concerning their structural validity, results confirmed the original structure for all three ver-
sions (Conners 3–Self-Report–SF, Conners 3–Parent–SF, and Conners 3–Teacher–SF). In fact,
even when comparing the original model to both a one-factor and a second-order factor models,
results confirmed that the original one best fitted to data, thus providing solidity to the Conners 3
factorial structure. This finding is in line with the theoretical, clinical definition of ADHD accord-
ing to the Diagnostic and Statistical Manual of Mental Disorders (DSM; APA, 2013), which speci-
fies the existence of two separate dimensions—that is, inattention and hyperactivity/impulsivity.
Moreover, the Conners 3 scales’ structure was both theoretical and psychometrically supported as
being multidimensional, since they assess a series of aspects such as learning problems that, though
being frequently comorbid to ADHD, are not strictly included in the clinical framework of ADHD.
Some more concern may have been arisen when assessing the goodness of fit of the second-order
factor model. In fact, the existence of a unique, higher-level factor embracing all dimensions of
ADHD and comorbid difficulties could be hypothesized. Nonetheless, results did not support this
model. Therefore, the study provided support for the short forms as presenting a multifactorial
structure, which includes the same dimensions of the full-length form in its original version
 804

Table 8.  Discriminative validity for the Conners 3 short forms.

Scale AUC Sensitivity Specificity Non-clinical sample Clinical sample t (df) p Cohen’s d

  M SD M SD
Self-Report Inattention .82 .80 .66 5.22 4.03 10.64 5.15 –6.15 (108) <.001 1.24
Hyperactivity/ .77 .70 .71 4.18 2.63 7.58 3.57 –5.68 (108) <.001 1.08
Impulsivity
Learning Problems .75 .73 .66 3.86 3.10 6.95 3.58 –4.84 (108) <.001 0.92
Defiance/Aggression .68 .55 .66 1.33 1.59 3.38 3.42 –4.04 (108) <.001 0.77
Family Relations .54 .64 .53 2.76 2.66 3.44 3.37 –1.16 (108) .247 0.22
Parent Inattention .90 .87 .81 4.59 3.49 11.13 3.23 –10.92 (124) <.001 1.95
Hyperactivity/ .83 .76 .72 3.60 3.92 10.16 5.52 –7.68 (124) <.001 1.37
Impulsivity
Learning Problems .86 .75 .81 3.24 2.87 7.89 3.07 –8.78 (124) <.001 1.57
Executive Functioning .90 .84 .87 3.11 2.37 9.03 3.78 –10.53 (124) <.001 1.88
Defiance/Aggression .80 .86 .65 .67 1.22 2.48 2.22 –5.67 (124) <.001 1.01
Peer Relations .77 .70 .75 1.13 1.86 4.35 3.97 –5.84 (124) <.001 1.04
Teacher Inattention .76 .73 .66 6.13 5.34 10.93 4.35 –2.70 (28) .012 0.99
Hyperactivity/ .73 .73 .73 4.33 5.92 7.27 4.76 –1.50 (28) .146 0.55
Impulsivity
Learning Problems/ .75 .73 .80 5.53 4.97 9.53 4.22 –2.38 (28) .025 0.87
Executive Functioning
Defiance/Aggression .58 .51 .73 1.53 2.20 1.93 2.19 –.500 (28) .621 0.18
Peer Relations .66 .67 .60 2.20 2.60 4.00 3.51 –1.60 (28) .121 0.58

AUC: area under the curve.

Clinical Child Psychology and Psychiatry 24(4)
Izzo et al. 805

(Conners, 2008). This aspect is particularly relevant for clinicians. In fact, though the diagnostic
process and decisions about interventions should be based not only on rating scales, but on a thor-
ough clinical assessment based on clinical expertise and knowledge, the Conners 3–short forms
scales may support clinicians in acquiring detailed information about specific, different clinical
areas, thus drafting an accurate profile that would indicate the specific domain(s) of intervention.
Regarding the internal consistency of the scales, analyses showed acceptable to high Cronbach’s
alphas in our sample for all the three versions, except for the Defiance/Aggression scale of the
Conners 3–Parent–SF scale, which showed questionable internal consistency. However, our results
were lower than the original, American ones (Conners, 2008). Moreover, values were lower than
the Spanish version of the Conners 3–Parent–SF and the Conners 3–Self-Report–SF scales, whose
Cronbach’s alphas were .84 (ranging from .73 to .89) and .78 (ranging from .75 to .86) respectively
(Conners, 2008).
Furthermore, the across-informant correlations showed that ratings of the same youth from the
three sources of information were moderately consistent, supporting the Conners 3–short forms as
valid, but not redundant, measures. In fact, considering that the three versions (Conners 3–Self-
Report–SF, Conners 3–Parent–SF, and Conners 3–Teacher–SF) aimed to measure similar con-
structs, the significant correlations give support to the validity of the scales. However, in line with
both our hypothesis and previous studies (Achenbach, 2006; Achenbach et al., 1987; De Los Reyes
et al., 2015), correlations were moderate in size, since the different informants experience the
child’s behavior from different points of view. This degree of incongruence supports the impor-
tance of having multi-informant versions, since they allow to gather information that, taken
together, instead of overlapping, let clinicians have a global framework of the child.
In addition, the short forms of the Conners 3 displayed good discriminative validity, with
acceptable to high values of both sensitivity and specificity. In fact, they were able to detecting
significant differences between the ADHD clinical sample and the non-clinical sample. However,
some exceptions need to be discussed. For instance, the Family Relations scale of the Self-
Report version did not display a statistically significant difference between the two groups.
Nonetheless, if we look at the item formulation of this scale (i.e. “My parents expect too much
from me” and “My parents are too strict with me”), we can see that they describe some difficul-
ties when relating with parents that are not distinctive of children and adolescents with ADHD.
Moreover, because of the very small sample size, some Content scales of the Teacher version did
not show a statistically significant difference between the two groups. However, effect sizes for
those non-significant differences were good (i.e. moderate for the Hyperactivity/Impulsivity and
Peer Relations scales, and small only for the Defiance/Aggression scale), indicating that, regard-
less of the sample size, a good discriminative validity could be assumed also for the Conners
3–Teacher–SF scale.
Although some relevant strengths characterize our study, as the big sample size of the non-
clinical group and the recruitment of a clinical-matched sample, the present study does have some
limitations. First, the clinical group—mostly referring to the Conners 3-T version—should be
enlarged to better examine specificity and sensitivity of the scales. Furthermore, test–retest relia-
bility should be examined—aiming to test the stability of the scores in different administrations—
and convergent and divergent validity should be assessed by comparing the Conners 3 scales with
other measures of childhood psychopathology. Moreover, some more studies better examining the
Defiance/Aggression Content scale of all three versions should be performed. In fact, the items of
this scale displayed both the lowest variability in the answers and the lowest internal consistency.
Eventually, measurement invariance—that is, same factor loadings between the different popula-
tions, same intercepts, and same error of measurement—should be tested for both age and gender.
Future studies should be conducted to investigate these aspects.
806 Clinical Child Psychology and Psychiatry 24(4)

To conclude, our results confirmed the Conners 3 short forms as being an effective instrument
for the detection of ADHD symptoms and its main comorbid problems in children and adoles-
cents. These results have an impact on the clinical field, as the Conners 3 scales–short form
resulted to be a reliable measure that may support clinician decisions, contribute to the definition
of severity of the disorder, evaluate the presence of symptoms co-occurring with ADHD, and
allow to check the evolutionary trajectories. The short timing required for compilation also makes
it a tool that can be administered in close follow-ups as a measure of change and as an indicator
of the effectiveness of ongoing treatment. Eventually, the presence of specific versions for the
different informants (youth, parents, and teachers) may help clinicians in evaluating the patient’s
functioning in different life spans.

Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.

Note
(Obtained raw score for scale) × (Total # of items on scale)
1. Prorated score= .
Total # of items on scale with responses
ORCID iD
Viola Angela Izzo https://orcid.org/0000-0003-4794-7296

Supplemental material
Supplemental material for this article is available online.

References
Achenbach, T. M. (2006). As others see us: Clinical and research implications of cross-informant correlations
for psychopathology. Current Directions in Psychological Science, 15, 94–98.
Achenbach, T. M., McConaughy, S. H., & Howell, C. T. (1987). Child/adolescent behavioral and emotional
problems: Implications of cross-informant correlations for situational specificity. Psychological Bulletin,
101, 213–232.
Akaike, H. (1974). A new look at the statistical model identification. IEEE Transactions on Automatic
Control, 19, 716–723.
American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.).
Washington, DC: American Psychiatric Association.
Anderson, J. C. (1996). Is childhood hyperactivity the product of western culture? The Lancet, 348(9020), 73–74.
Arias Martínez, B., Arias González, V. B., & Gómez Sánchez, L. E. (2013). Calibration of Conners ADHD
index with Rasch model. Universitas Psychologica, 12, 957–970.
Barkley, R. A. (1997). Behavioral inhibition, sustained attention, and executive function: Constructing a uni-
fied theory of ADHD. Psychological Bulletin, 121, 65–94.
Bentler, P. M. (1990). Comparative fit indexes in structural models. Psychological Bulletin, 107, 238–246.
Bentler, P. M. (1995). EQS structural equations program manual. Encino, CA: Multivariate Software.
Busemeyer, J. R., & Diederich, A. (2014). Estimation and testing of computational psychological models.
In P. W. Glimcher & E. Fehr (Eds.), Neuroeconomics: Decision making and the brain (pp. 49–61).
Cambridge, MA: Academic Press.
Capano, L., Minden, D., Chen, S. X., Schachar, R. J., & Ickowicz, A. (2008). Mathematical learning disorder
in school-age children with attention-deficit hyperactivity disorder. The Canadian Journal of Psychiatry,
53, 392–399.
Catalá-López, F., Peiró, S., Ridao, M., Sanfélix-Gimeno, G., Gènova-Maleras, R., & Catalá, M. A. (2012).
Prevalence of attention deficit hyperactivity disorder among children and adolescents in Spain: A sys-
tematic review and meta-analysis of epidemiological studies. BMC Psychiatry, 12, Article 168.
Izzo et al. 807

Christiansen, H., Hirsch, O., Drechsler, R., Wanderer, S., Knospe, E. L., Günther, T., & Lidzba, K. (2016).
German validation of the Conners 3U+00AE rating scales for parents, teachers, and children. Zeitschrift
für Kinder-und Jugendpsychiatrie und Psychotherapie, 44, 139–147.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Lawrence
Erlbaum.
Conners, C. K. (1989). Manual for Conners’ Rating Scales. North Tonawanda, NY: Multi-Health Systems.
Conners, C. K. (1997). Conners’ Rating Scales–Revised: Technical manual. North Tonawanda, NY: Multi-
Health Systems.
Conners, C. K. (2008). Conners 3rd Edition. Manual. Toronto, Ontario, Canada: Multi-Health Systems.
Conners, C. K., Sitarenios, G., Parker, J. D., & Epstein, J. N. (1998). The revised Conners’ Parent Rating Scale
(CPRS-R): Factor structure, reliability, and criterion validity. Journal of Abnormal Child Psychology,
26, 257–268.
De Los Reyes, A., Augenstein, T. M., Wang, M., Thomas, S. A., Drabick, D. A., Burgers, D. E., & Rabinowitz,
J. (2015). The validity of the multi-informant approach to assessing child and adolescent mental health.
Psychological Bulletin, 141, 858–900.
Gau, S. S. F., Soong, W. T., Chiu, Y. N., & Tsai, W. C. (2006). Psychometric properties of the Chinese
version of the Conners’ Parent and Teacher Rating Scales-Revised: Short form. Journal of Attention
Disorders, 9, 648–659.
Hu, L. T., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55.
Huss, M., Hölling, H., Kurth, B. M., & Schlack, R. (2008). How often are German children and adolescents
diagnosed with ADHD? Prevalence based on the judgment of health care professionals: Results of the
German Health and Examination Survey (KiGGS). European Child & Adolescent Psychiatry, 17, 52–58.
Izzo, V. A., Donati, M. A., & Primi, C. (2018a). Conners 3–Self-Report Scale: An empirical support to the
dimensionality of the content scales. Clinical Child Psychology and Psychiatry, 23, 556–566.
Izzo, V. A., Donati, M. A., & Primi, C. (2018b). Dipendenze comportamentali e Disturbo da Deficit di
Attenzione/Iperattività (ADHD) negli adolescenti. Fattori di rischio specifici e aspecifici [Behavioral
Addictions and Attention Deficit/Hyperactivity Disorder (ADHD) in adolescents. Specific and aspecific
risk factors]. Psicologia Clinica Dello Sviluppo, 22, 535–558.
Izzo, V. A., Donati, M. A., & Primi, C. (2019). Assessing ADHD through the multi-informant approach: The
contribution of the Conners’ 3 scales. Journal of Attention Disorders, 23, 641–650.
Lecendreux, M., Konofal, E., & Faraone, S. V. (2011). Prevalence of attention deficit hyperactivity disorder
and associated features among children in France. Journal of Attention Disorders, 15, 516–524.
Lindquist, B., Carlsson, G., Persson, E. K., & Uvebrant, P. (2006). Behavioural problems and autism in chil-
dren with hydrocephalus. European Child & Adolescent Psychiatry, 15, 214–219.
Mannuzza, S., Klein, R. G., Abikoff, H., & Moulton, J. L., III. (2004). Significance of childhood conduct
problems to later development of conduct disorder among children with ADHD: A prospective follow-
up study. Journal of Abnormal Child Psychology, 32, 565–573.
Marôco, J. (2018). Análise Estatística com o SPSS Statistics (7ª edição) [Statistical Analisis with SPSS
Statistics (7th edition)]. ReportNumber, Lda.
Morales-Hidalgo, P., Hernández-Martínez, C., Vera, M., Voltas, N., & Canals, J. (2017). Psychometric
properties of the Conners-3 and Conners Early Childhood Indexes in a Spanish school population.
International Journal of Clinical and Health Psychology, 17, 85–96.
Ostini, R., & Nering, M. L. (2006). Polytomous item response theory models (No. 144). Thousand Oaks, CA:
Sage.
Primi, C., & Maschietto, D. (2017). Conners 3TM. Standardizzazione italiana. Firenze, Italy: Giunti O.S.
Psychometrics.
Rucklidge, J. J. (2006). Impact of ADHD on the neurocognitive functioning of adolescents with bipolar dis-
order. Biological Psychiatry, 60, 921–928.
Schwarz, G. (1978). Estimating the dimension of a model. The Annals of Statistics, 6, 461–464.
Skounti, M., Giannoukas, S., Dimitriou, E., Nikolopoulou, S., Linardakis, E., & Philalithis, A. (2010).
Prevalence of attention deficit hyperactivity disorder in schoolchildren in Athens, Greece. Association
808 Clinical Child Psychology and Psychiatry 24(4)

of ADHD subtypes with social and academic impairment. ADHD Attention Deficit and Hyperactivity
Disorders, 2, 127–132.
Spencer, T. J., Biederman, J., & Mick, E. (2007). Attention-deficit/hyperactivity disorder: Diagnosis, lifes-
pan, comorbidities, and neurobiology. Journal of Pediatric Psychology, 32, 631–642.
Steiger, J. H., & Lind, J. C. (1980, May). Statistically based tests for the number of common factors. In Annual
meeting of the Psychometric Society, Iowa City, IA.
Subcommittee on Attention-Deficit/Hyperactivity Disorder, Steering Committee on Quality Improvement
and Management, Wolraich, M., Brown, L., Brown, R. T., DuPaul, G., & . . . Visser, S. (2011). ADHD:
Clinical practice guideline for the diagnosis, evaluation, and treatment of attention-deficit/hyperactivity
disorder in children and adolescents. Pediatrics, 128, 1007–1022.
Thomas, R., Sanders, S., Doust, J., Beller, E., & Glasziou, P. (2015). Prevalence of attention-deficit/hyperac-
tivity disorder: A systematic review and meta-analysis. Pediatrics, 135, e994–e1001.
Timimi, S., & Taylor, E. (2004). ADHD is best understood as a cultural construct. The British Journal of
Psychiatry, 184(1), 8–9.
Tucker, L. R., & Lewis, C. (1973). A reliability coefficient for maximum likelihood factor analysis.
Psychometrika, 38, 1–10.
Willcutt, E. G., Doyle, A. E., Nigg, J. T., Faraone, S. V., & Pennington, B. F. (2005). Validity of the execu-
tive function theory of attention-deficit/hyperactivity disorder: A meta-analytic review. Biological
Psychiatry, 57, 1336–1346.

Author biographies
Viola Angela Izzo is a PhD student working with Caterina Primi on the adaptation and development of scales
measuring impulsivity. Her research interests mainly regard impulsivity, examined both in ADHD and in
Impulse Control Disorders in Parkinson’s Diseases.
Maria Anna Donati is a Lecturer in Psychometrics at the Department of Developmental and Social Psychology,
University of Rome. Her research interests mainly concern the psychology of behavioural addictions in ado-
lescents. Inside this framework, she studies the factors involved in pathological behaviour, including the
adaptation and construction of assessment instruments.
Federica Novello is a Clinical Psychologist specialized in services for childhood, adolescence and adulthood.
She won a one-year scholarship to focus on ADHD. She is specializing in Cognitive Psychotherapy.
Dino Maschietto is a Child Neuropsychiatrist. He is member of the technical scientific committee of the AIFA
(Italian Drug Agency) for the ADHD register, and the Coordinator of the Reference Centers for ADHD in the
region of Veneto. He is the author of several publications on ADHD.
Caterina Primi is an Associate Professor in Psychometrics, and she is responsible for the lab Psychometrics at
the NEUROFARBA Department. She is currently conducting research into a number of topics related to test
development and adaptation.