Psychometric properties of a new ADHD screening questionnaire: Parent report on the (potential) underlying explanation of inattention in their school-aged children

ABSTRACT

The present study examined psychometric properties of a recently developed parent report screening questionnaire, i.e., Parent ADHD Screening questionnaire: Signaling the Core explanation underlying behavioral symptoms (PASSC). The PASSC aims to measure (1) ADHD symptoms and (2) what parents view to be the main underlying explanation(s) of these symptoms. The PASSC questions 3 (potential) underlying explanations based on the triple pathway model (TPM): i.e., time, cognition and/or motivation problems. Parents of 1166 Dutch children aged 4–12 filled in the PASSC, as well as 2 questionnaires measuring time, cognition and motivation (i.e., the FTF and the SPSRQ-C). Reliability of the PASSC is good, indicated by high internal consistency of the sumscores. Principal component analyses supported the distinction between inattention and hyperactivity-impulsivity symptoms as defined in the DSM-5, and the distinction between the 3 TPM explanations given by parents for inattention, but not for hyperactivity-impulsivity symptoms. The majority of parents selected one and the same explanation for inattention problems of their child, most often being cognition (31.2%) and motivation (28.2%). PASSC validity was further supported by positive associations between the explanation sumscores for inattention symptoms and other parent questionnaires measuring the same constructs (i.e., time, cognition and motivation; convergent validity), although we found no evidence for discriminant validity. Groups (based on age group, sex and ADHD diagnosis) differed on the PASSC sumscores in the expected directions. Concluding, the PASSC is a promising tool to assess a child’s ADHD symptoms as well as the parent view on (potential) explanation(s) of inattention.

KEYWORDS:

• Attention deficit hyperactivity disorder
• parent report
• psychometric properties
• neuropsychological heterogeneity
• screening

Attention Deficit Hyperactivity Disorder (ADHD) is one of the most common developmental disorders, affecting 5% of children (Sayal et al., 2018). According to the Diagnostic and Statistical Manual of Mental Disorders (DSM-5; American Psychiatric Association [APA], 2013), ADHD is characterized by symptoms of inattention and/or hyperactivity-impulsivity. It often leads to lower quality of life and impairments in academic and social functioning (Escobar et al., 2005). ADHD, especially if untreated, is a risk for, e.g., other mental health disorders, low self-esteem, social difficulties and/or substance abuse (Nigg, 2013). Early detection of ADHD enables intervention at an earlier stage and has a positive effect on the course and severity of the disorder (Sonuga‐Barke & Halperin, 2010).

According to Hudziak et al. (2007), screening the general population is suited as first step for early detection of ADHD. Screening is often done by administrating questionnaires that (almost) literally ask an informant (e.g., parents) to rate the existence of the n = 18 DSM-5 symptom criteria for ADHD (APA, 2013). Such questionnaires are cost-effective, easy to administer and widely applicable in large groups (e.g., Pelham et al., 2005). However, they purely ask for the behavior and fail to provide insight into why an informant thinks the child shows this behavior. Understanding the underlying explanation(s) of behavioral symptoms within a heterogeneous disorder as ADHD has great potential for individualizing interventions and increasing treatment effectivity, as the DSM classification ADHD on its own is not predictive of treatment response (Insel et al., 2010). Several researchers have implied that multiple underlying explanation(s) can lead to the same behavior characteristic of ADHD (Costa Dias et al., 2013; Nigg et al., 2005). One causal model is the triple pathway model (TPM; Sonuga-Barke et al., 2010). This model focusses on the neuropsychological heterogeneity within ADHD and states that a person with ADHD can, for instance, fail to give close attention to details due to (1) time processing problems (i.e., one underestimates time it takes to complete homework), (2) cognition problems (i.e., one fails to give close attention to details), and (3) motivation problems (i.e., immediate reward of playing weighs “heavier” than the delayed, larger reward that environment is happy tomorrow when the homework is finished). These TPM explanations of ADHD are thought to be independent and are associated with distinct (parts of) neural brain networks (Durston et al., 2011; Lecei et al., 2019).

Several studies have confirmed the TPM validity in children with ADHD (De Zeeuw et al., 2012; Peijnenborgh et al., 2016). Most of these studies used neuropsychological performance tests measuring time, cognition and/or reward processing and showed that 56–71% of the children with ADHD showed a deficit in one of the 3 domains, although the exact percentages differed between studies (De Zeeuw et al., 2012; Sonuga-Barke et al., 2010). Neuropsychological tests thus have the potential to reveal the underlying explanation of ADHD symptoms in a specific child. Although such tests are often, but not necessarily, included in the diagnostic process of ADHD, this is not feasible for screening purposes (due to e.g., high costs).

When screening children, questionnaires for ADHD rely mostly on parent reports. Parent reported ADHD symptoms as well as impairment of their child by ADHD symptoms proved to be a positive predictor for, respectively, an ADHD diagnosis (Hall et al., 2019) and ADHD severity and poorer overall functioning at a 6 year follow-up (Van Lieshout et al., 2017). Next to accurate perception of children’s (problem) behavior, parents also proved to be able to differentiate between attributions or explanations of different types of behavior, e.g., they judged inattentive behavior as less controllable by the child than prosocial behavior (Freeman et al., 1997). Applying the TPM or the more general idea of different underlying explanations of ADHD symptoms to a parent questionnaire has, however, not been done before. Such a questionnaire including parents’ beliefs about the explanation(s) of their child’s behavior might increase their insight in the experienced problem behavior and increase motivation for and acceptance of interventions (Johnston & Freeman, 1997).

The aim of the current study was to evaluate the psychometric properties of a recently developed parent report ADHD screening questionnaire that identifies parental explanation(s) underlying (potential) ADHD symptoms in school-aged children. The instrument is called the “Parent ADHD Screening Questionnaire: Signaling the Core explanation underlying the behavioral symptoms (PASSC).” The validity of the PASSC was evaluated by 3 different methods (Evers et al., 2010). First, we expected principal component analyses on the PASSC items to show evidence for 2 ADHD dimensions (i.e., inattention and hyperactivity-impulsivity symptoms) and within each ADHD dimension evidence for the 3 different TPM explanations (i.e., time, cognition and motivation). Second, scores on the 3 TPM explanations were predicted to correlate positively and significantly with questionnaires measuring the same construct (e.g., PASSC time explanation sum score is associated with score on other published questionnaire measuring time), but not with questionnaires measuring another construct (i.e., convergent and discriminant validity). Third, relevant groups (i.e., based on age group, sex and ADHD diagnosis) were thought to differ on the PASSC in line with theoretical expectations (i.e., hyperactive-impulsive symptoms decline whereas attention symptoms are stable with advancing age, ADHD symptoms are more prevalent in boys than girls, and in children diagnosed with ADHD compared to typical developing children (APA, 2013)). Reliability of the PASSC was assessed by the internal consistency and was expected to be sufficient (Evers et al., 2010).

Methods

Procedure

Because the PASSC is meant for screening in the general population, parents of children in grades 1 to 8 of regular primary schools were recruited in 2 ways. First, Dutch regular primary schools were approached and 23 schools agreed to participate in the study. These schools distributed an information letter to parents and asking them to sign an informed consent and fill in an online questionnaire about their child (i.e., including the PASSC and 2 other questionnaires). Second, 2 Dutch certified (www.isoregister.nl) internet research agencies approached members of their response panels for participation. Inclusion criteria were being between ages 30–50 years and having a child attending regular primary school. The number of members invited by the internet research agencies was based on census data (e.g., age groups, levels of education and provinces within the Netherlands) to get a sample representative for the Dutch parents population. The potential respondents received an email by the research agencies about participating in the study, and a reminder within 1 week. After correctly answering a verification question, respondents were informed about the study, asked to sign an informed consent and fill in an online questionnaire about their child (the same as in the first recruitment method). In both recruitment methods, all questionnaires included were implemented online through Qualtrics Survey Software and took 15–20 minutes to complete. Parents did not receive a monetary reward for participation, although the internet research agencies gave points to respondents for an internal saving system.

The present study was approved by the Ethical Review Committee Psychology and Neuroscience (ERCPN), Maastricht University, The Netherlands (approval number ERCPN-161_02_04_2011).

Participants

In total, n = 1166 parents participated in the study and filled in the questionnaires about 1 child. Of these, n = 335 parents were recruited through schools, and n = 831 through the internet research agencies. The questionnaire was filled in by mothers in 66.2% of cases, fathers in 31.7% of cases, and/or other caregivers in 0.2% (for 1.9% this information was missing). Level of parental education (LPE) ranged from primary school (1) to university degree (8) (De Bie, 1987), on a scaling comparable to the International Standard Classification of Education (UNESCO Institute for Statistics, 2012). If the LPE differed between mother and father, the highest level of education was chosen. In total, 31.6% of parent-pairs had a low/moderate level of education (1–4) and 68.1% of parent-pairs had a high level of education (5–8; for 0.3% information missing).

The children (562 boys) about whom the questionnaire was filled in were distributed over the grades 1 to 8. Their age ranged from 4.03 to 13.08 years (M = 8.46, SD = 2.32). The majority of children had the Dutch nationality (93.5%; 3.3% Belgian nationality; 3.1% other or information missing). Since we wanted to include a sample representative of the population, children with a DSM-5 diagnosis as reported by parents, such as ADHD (4.1%), learning disability (5.3%), autism spectrum disorders (2.8%) and anxiety (1.6%) were not excluded. Children using medication (5.3%), such as antihistamines or methylphenidate, were included as well.

Instruments

The Parental ADHD Screening Questionnaire: Signaling the Core explanation underlying the behavioral symptoms (PASSC) is a screening questionnaire, which includes 18 symptom items that equal the DSM-5 symptom descriptions for ADHD (APA, 2013). For each of those 18 items parents were asked to indicate (a) to what extent their child showed the behavioral symptom and (b) to what extent each of the 3 given explanations (i.e., time, cognition, motivation problems) clarifies this behavioral symptom, both on a 5 points scale ranging from 0 (never) to 4 (very often). Additionally, per symptom, parents were forced to choose the core explanation out of the 3 possible explanations which best and/or most often explained the child’s behavioral symptom. They could also opt for the answer option “none” in case their child did not show the behavioral symptom. While constructing the instrument, we tried to avoid words or content that may be unnecessary culture-bound or offensive to groups of individuals (Evers et al., 2010). The PASSC offers 10 sum scores in total, 2 symptom sum scores and 2 × 4 explanation sum scores. For the DSM-5 ADHD symptoms sum score for inattention (items 1–9) and a sum score for hyperactivity-impulsivity (items 10–18) is calculated (max score 36). For each ADHD symptom dimension (i.e., inattention and hyperactivity-impulsivity) the following 4 explanation sum scores can be calculated: 3 sum scores that indicate to what extent parents explain symptoms by time, cognition, and motivation problems (max score 36) and 1 sum score that indicate what parents answered when forced to choose the core explanation per symptom (max score 9). Figure 1 contains an example item of the PASSC.

Figure 1. Instruction and example item PASSC

The following statements are about behavior that a child can show in everyday situations. Indicate to what extent your child has shown this behavior during the past 6 weeks by checking the box that is most applicable. You can choose between ‘never’, ‘rarely’, ‘sometimes’, ‘often’, and ‘very often’. For each statement, 3 potential explanations are offered for this behavior, namely (a) Due to time estimation or time management problems, (b) Due to cognitive function problems and (c) Due to motivation regulation problems. Time estimation or management problems refer to how well your child e.g., can estimate how long a certain task takes, how long ago an event took place, what time it is, or how much time is needed to complete a specific task. Unsettled behavior caused by time pressure is also part of this. Cognitive functions include attention, inhibition, and directing behavior, not understanding the explanation, not knowing how a task should be performed or not knowing how something works. Problems with motivation regulation concern not (willingly) wanting to carry out a task or having no or less interest(s) in something. Please indicate for each statement, to what extent this applies to your child by checking the corresponding answer. If your child ‘never’ shows the behavior, you can also choose the option ‘never’ for the explanations (a), (b) and (c). Finally, you are asked to choose one of the 3 given explanations that best and/or most often explains this behavior of your child. If your child ‘never’ shows the behavior, you choose the option ‘none’. Example situation: A child often fails to give close attention to details or makes careless mistakes in schoolwork. This is mainly because he does not understand what he has to do cognitively (e.g., he does not understand how to check his schoolwork or how to avoid being distracted), and not so much because he does not feel like executing nor is he not motivated to execute the task. There is also no time pressure in this case, there is enough time to execute the task(s). Example item: In this example situation, the coming item should be answered as follows:

Instruments used for validation

To validate the PASSC, sum scores on the PASSC were compared to scores on other, published questionnaires measuring the same constructs, described below.

The Dutch translation of the Five to Fifteen (FTF) questionnaire was used to measure parent reports on skills and behaviors of children in various domains of development (Kadesjö et al., 2004). In the present study, 2 FTF-subscales were included measuring “Concepts of time” and “Ability to plan and organize activities”. Items are answered on a 5-point Likert scale, ranging from 1 (never) to 5 (very often). Total scores on both subscales were used as measures of time processing (max score 20) and cognition (max score 15), respectively, with higher scores indicating more problems. Psychometric properties of both scales were sufficient (Kadesjö et al., 2004; Trillingsgaard et al., 2004).

The Dutch translation of the Sensitivity to Punishment and Sensitivity to Reward Questionnaire for Children (SPSRQ-C) is developed to measure sensitivity to punishment and reward (Luman et al., 2012). In this study, only the “Reward Responsivity” subscale was included. Items are answered on a 5-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree) with a maximal score of 35. Higher scores indicate a stronger reward responsivity. Total score on this subscale was used as measure of motivation, since deviant sensitivity to reinforcement such as reward in ADHD is highly associated with motivation (Luman et al., 2005). The psychometric properties of the “Reward Responsivity” subscale were sufficient (Luman et al., 2012).

Results

Data inspection indicated that the distributions of all variables were normal (skewness and kurtosis values between −1.5 and 1.5) (Tabachnick & Fidell, 2013). Outliers were defined as a score of more than 3 standard deviations above or below the variable mean. N = 7 outliers on the PASSC symptom sum scores were excluded. Analyses were conducted using the statistical package IBM SPSS Statistics (version 24.0; SPSS, Inc., Chicago, IL), α was set at .05.

Validity

To evaluate the validity of the PASSC we conducted (1) principal components analyses (PCA), a study on (2) convergent and discriminant validity and (3) differences between relevant groups (e.g., age group, sex and ADHD diagnosis).

Principal components analyses

First, a PCA with varimax rotation (Nunnally & Bernstein, 1994) was conducted to investigate whether the 18 ADHD symptom items measured the underlying 2 constructs as theoretically assumed in the DSM-5. The Kaiser–Mayer–Olkin (KMO) coefficient was 0.95 and Bartlett’s Test of sphericity significant (χ₍₁₅₃₎² = 8880.44; p < .001), indicating that the data were suitable for data reduction. The structure of the PASSC symptom items was confirmed according to the 2 dimensions, i.e., inattention (items 1–9) and hyperactivity-impulsivity (items 10–18), accounting for 51% of the total variance. Each item revealed a loading greater than 0.4 for the expected component.

The second PCA was run with the PASSC explanation items, i.e., (a) time, (b) cognition and (c) motivation problems as explanation for each symptom (see Figure 1). The PCA on the explanations items was done separately for the inattention and hyperactivity-impulsivity symptoms. For the inattention symptoms, all explanation items revealed a loading greater than 0.4 (only item 7b had a slightly lower loading of 0.39) for the expected component. For example, explanation time for the first inattention symptom (1a) loaded on a component interpreted as Time, explanation cognition for the first inattention symptom (1b) loaded on component Cognition, etc. (Table 1). For the hyperactivity-impulsivity symptoms, the loadings of the explanation items on the 3 components were less straightforward. For 5 explanation items loadings did not pass the threshold of 0.4 on the expected component and 7 explanation items also showed a loading greater than 0.4 on another component (Table 1).

Table 1. Loadings of PASSC explanation items on each component of the PCA

		component
PASSC item		1	2	3
Inattention items		Time	Cognition	Motivation
Variance explained	57%
KMO coefficient	0.94
Bartlett’s test of sphericity	(χ(351)^2 = 20535.80; p < .001)
1(a)		0.610
1(b)			0.766
1(c)				0.685
2(a)		0.709
2(b)			0.776
2(c)				0.701
3(a)		0.700
3(b)			0.476
3(c)				0.610
4(a)		0.591	0.409
4(b)			0.714
4(c)				0.634
5(a)		0.589
5(b)			0.610
5(c)				0.628
6(a)		0.657
6(b)			0.657
6(c)			0.417	0.591
7(a)		0.729
7(b)
7(c)				0.582
8(a)		0.761
8(b)			0.795
8(c)				0.638
9(a)		0.697
9(b)			0.503
9(c)		0.431		0.617
Hyperactivity- Impulsivity items		Time	Cognition	Motivation
Variance explained	60%
KMO coefficient	0.94
Bartlett’s test of sphericity	(χ(351)^2 = 22433.52; p < .001)
10(a)		0.699
10(b)			0.736
10(c)		0.514
11(a)		0.798
11(b)		0.508	0.665
11(c)		0.634	0.420
12(a)		0.801
12(b)		0.533	0.616
12(c)		0.662
13(a)		0.660
13(b)		0.426	0.595
13(c)			0.422
14(a)		0.578		0.500
14(b)			0.737
14(c)				0.499
15(a)		0.494		0.599
15(b)			0.669
15(c)				0.686
16(a)		0.442		0.640
16(b)			0.605
16(c)				0.706
17(a)				0.578
17(b)			0.750
17(c)				0.666
18(a)		0.535		0.567
18(b)			0.690
18(c)				0.631

N = 1166. Item loadings higher than 0.4 are included in the table. Double loadings were allocated on the basis of content and their strength in component loading. The loading of each item allocated to the expected component is highlighted in bold.

In sum, it can be concluded that the explanation items asking parents to indicate their view on the underlying explanation for the behavioral symptom of their child correspond well to the theoretically proposed 3 TPM explanations for the inattention, but not the hyperactivity-impulsivity symptoms. Therefore, in the following analyses on the PASSC explanation subscales (i.e., time, cognition, motivation problems), we focus only on the inattention items.

Convergent and discriminant validity

Bivariate correlation analyses within the PASSC showed strongly positive associations between the inattention symptom sum score and explanation sum scores for the inattention symptoms (Table 2). When forcing parents to indicate one core explanation underlying the indicated inattention symptoms most parents (72.0%) chose one and the same core explanation across inattention symptoms. In that case, parents chose the same explanation for the majority of inattention symptoms (e.g., minimal 5 out of 9 symptoms). Cognition problems were most often chosen (31.2%), followed by motivation (28.3%), none (6.9%) and time problems (5.7%). In 10.3% of cases parents equally often chose 2 explanations for the attention problems of their child and the third explanation was not chosen. Cognition and motivation problems were the most often occurring combination of explanations (6.8%). In 5.6% of cases all 3 explanations were equally often given by parents for inattention items. In 11.9% of cases, information on core explanation was missing.

Table 2. Scores on and correlations between PASSC sum scores and questionnaires measuring the same/another construct (i.e., convergent/discriminant validity)

	Mean N	SD (range)	PASSC inattention symptoms	PASSC inattention symptoms explanation Time	PASSC inattention symptoms explanation Cognition	PASSC inattention symptoms explanation Motivation
PASSC inattention symptoms	11.56 N = 1163	6.36 (0–35)	-
PASSC hyperactivity/ impulsivity symptoms	9.51 N = 1163	6.39 (0–33)	.669***
PASSC inattention symptoms (a) explanation Time	7.18 N = 1163	5.83 (0–27)	.665***	-
PASSC inattention symptoms (b) explanation Cognition	11.26 N = 1163	7.44 (0–36)	.910***	.659***	-
PASSC inattention symptoms (c) explanation Motivation	10.06 N = 1163	6.39 (0–31)	.781***	.645***	.719***	-
PASSC inattention core explanation Time	0.73 N = 1143	1.35 (0–9)	.092**	.352***	.079**	.064*
PASSC inattention core explanation Cognition	2.51 N = 1143	2.58 (0–9)	.560***	.308***	.649**	.253***
PASSC inattention core explanation Motivation	2.19 N = 1143	2.36 (0–9)	.192***	.115***	.087**	.448***
PASSC inattention core explanation None	2.31 N = 1143	2.57 (0–8)	−.584***	−.443***	−.564***	−.586***
FTF time	6.22 N = 1122	2.29 (0–13)	.295***	.202***	.304***	.201***
FTF planning	4.12 N = 945	2.66 (0–12)	.550***	.364***	.508***	.446***
SPSRQ-C reward	13.54 N = 973	4.08 (0–26)	.204***	.154***	.196***	.201***

PASSC = Parental ADHD Screening Questionnaire: Signaling the Core explanation underlying the behavioral symptoms; FTF = Five to Fifteen questionnaire; SPSRQ-C = Sensitivity to Punishment and Sensitivity to Reward Questionnaire for Children. Correlation coefficients up to 0.30 were considered low, from 0.31 to 0.50 were considered moderate and those exceeding 0.50 were considered strong (Cohen et al., 2013). ***p < .001; **p < .010; *p < .050.

Correlations between the PASSC explanations scores for inattention (i.e., time, cognition, and motivation problems) and their corresponding questionnaires as indication of convergent validity were all significant in the expected direction (Table 2). Correlations ranged from low (i.e., PASSC explanation Time - FTF Time, and PASSC explanation Motivation - SPSRQC Reward) to high (i.e., PASSC explanation cognition- FTF Planning). However, the PASSC explanation scores for inattention also correlated significantly with the questionnaires measuring the other constructs (i.e., discriminant validity), ranging from low (i.e., PASSC explanation time-SPSR-C Reward) to moderate (i.e., PASSC explanation motivation -FTF planning).

Group differences on PASSC sum scores

To study age and sex differences on the PASSC, univariate general linear models were used with school grade (as indication of age; 8 levels) and sex (2 levels) as fixed between-subject factors. For the PASSC inattention symptoms, the main effect of sex (F_{(1, 1147)} = 19.16; p < .001; ƞ_p² = .016; boys> girls) was significant, but the main effect of school grade (F_{(7, 1147)} = 0.95; p = .466; ƞ_p² = .006) and the interaction sex x school grade (F_{(7, 1147)} = 0.58; p = .774; ƞ_p² = .004) were not significant. For the PASSC hyperactivity-impulsivity symptoms, main effects of sex (F_{(1, 1149)} = 24.88; p < .001; ƞ_p² = .021; boys> girls) and school grade (F_{(7, 1149)} = 6.90; p < .001; ƞ_p² = .040; grade 1 > 2, 3, 4, 5, 6 > 7, 8) were significant. The interaction sex x school grade (F_{(7, 1149)} = 0.58; p = .776; ƞ_p² = .003) was not significant. The univariate general linear models for the explanation sum scores for inattention symptoms (i.e., time, cognition and motivation problems) revealed the same results as for the PASSC inattention symptoms. Namely, only the main effect of sex was significant for each explanation. Parents explained the PASSC inattention symptoms of boys more often by time, cognition and motivation problems than the PASSC inattention symptoms of girls.

A one-way ANOVA with ADHD diagnosis as independent variable (i.e., an ADHD diagnosis reported by parents was compared to children with another DSM (but not ADHD) diagnosis and to children without any DSM diagnosis; 3 levels) was conducted for the PASSC sum scores. For both the PASSC inattention (F_{(2, 1160)} = 60.69; p < .001) and the PASSC hyperactivity-impulsivity (F_{(2, 1162)} = 43.28; p < .001) symptoms, children with ADHD scored higher than children with another DSM diagnosis and than children with no diagnosis. Children with a by parents reported DSM diagnosis other than ADHD scored higher than children without a diagnosis on the inattention, but not on the hyperactivity-impulsivity symptoms. Parents of children with ADHD more often explained their inattention symptoms by time (F_{(2, 1163).} = 50.56; p < .001) and cognition (F_{(2, 1163)} = 49.90;p < .001) problems than parents of children with another DSM diagnosis, who reported more often time and cognition explanations than parents of children without a diagnosis. Additionally, parents explained the inattention symptoms of children with ADHD more often by motivation (F_{(2, 1162)} = 14.80; p < .001) problems compared to parents of children with no diagnosis, but the motivation explanation did not differ from parents of children with another DSM diagnosis.

Reliability

Internal consistency

The PASSC had high internal consistency reliability coefficients for the inattention symptoms and the explanation sum scores for inattention (Cronbach’s alpha = 0.88 for PASSC inattention symptom sum score, 0.90 for explanation time sum score, 0.91 for explanation cognition sumscore and .89 for explanation motivation sum score). Internal consistency for the PASSC hyperactivity-impulsivity symptoms was also high (Cronbach’s alpha = 0.86). All items contributed to the corresponding scale (corrected item-total correlations >.50). The deletion of any of the items within the PASSC sum scores did not increase Cronbach’s alpha.

Discussion

ADHD is a neuropsychological heterogeneous disorder (Costa Dias et al., 2013; Nigg, 2013). Research has shown for instance, that individuals with ADHD are affected to varying degrees by dissociable causes, such as problems in time processing, cognition and/or motivation (Sonuga-Barke et al., 2010). Here, we present a recently developed screening questionnaire (the PASSC) meant to measure parents view on the presence of ADHD symptoms in their child and what parents view to be the main underlying explanation(s) of these symptoms. The instrument is constructed based on the DSM-5 symptom criteria set for ADHD (APA, 2013) and the Triple Pathway Model (TPM; Sonuga-Barke et al., 2010). This paper describes the reliability and validity of the PASSC.

The reliability of the PASSC was estimated on the basis of inter-item covariances. The Cronbach’s alphas for all PASSC sum scores were high. Reliability reflects the degree to which test score variance is due to true score variance and this is good for the PASSC (Evers et al., 2010). We used 3 different methods to study the validity of the PASSC sum scores, i.e., examining (1) dimensionality of the item scores, (2) convergent and discriminant validity and (3) differences between relevant groups.

With respect to the dimensionality of the PASSC items, we found confirmation of a bi-component model for the 18 DSM-5 ADHD symptoms in a general population of children rated by parents. Thus, it seems valid to construct 2 sumscores that measure inattention and hyperactivity-impulsivity symptoms, respectively. Previous studies have also shown that inattention and hyperactivity-impulsivity often go together, but are distinct clusters of ADHD symptoms (Ullebø et al., 2012). Furthermore, the PCA results supported the idea that parents can indicate the 3 TPM categories time, cognition, and motivation problems as distinct (potential) explanations of inattention in their child (Sonuga-Barke et al., 2010). However, we did not find support for the 3 TPM categories as distinct explanations for hyperactivity-impulsivity symptoms as reported by parents. Future research should reveal whether parents find it hard(er) to determine the explanation(s) of hyperactive-impulsive behavior, or ascribe this externalizing behavior to other explanations than included in the TPM. The idea of different underlying explanations of ADHD symptoms can only be used for the inattention symptoms in the PASSC, and therefore, the remaining discussion will focus on the attention sum scores.

In the majority of children parents indicated one and the same (i.e., time, cognition or motivation problems) core explanation underlying inattention. This is in line with the TPM model assuming independent explanations of ADHD problems, as well as with previous studies classifying children with ADHD into TPM subgroups (De Zeeuw et al., 2013; Sonuga-Barke et al., 2010). Cognition (31.2%) and motivation (28.2%) problems were the most often indicated explanations, followed by time (5.7%) problems. Other studies showed somewhat different distributions over the 3 TPM explanations (De Zeeuw et al., 2013, p. 21%, 2%, 12%, respectively; Sonuga-Barke et al., 2010, pp. 6%, 20%, 25%, respectively), but these studies differed with ours in a number of aspects. Both previous studies defined subgroups based on performance tests versus a questionnaire in our study and both used ADHD samples versus the general population in our study.

With respect to convergent and discriminant validity, we found positive associations between the PASSC sum scores and other validated parent report questionnaires measuring time, cognition and motivation (Kadesjö et al., 2004; Luman et al., 2012). However, we expected a priori a more specific association pattern, for instance, that the PASSC explanation motivation sum score would correlate strongly with a questionnaire measuring reward sensitivity, but not or far less strongly with a questionnaire measuring comprehension of time concepts. Meanwhile, we found significant positive correlations between the PASSC sum scores and all other questionnaires. The correlations of the PASSC sum scores tended to be higher with the FTF scale measuring planning compared to the FTF time and SPSRQ-C reward sensitivity scales. This might have been a result of larger overlap in item content between the PASSC inattention items and the FTF planning items (APA, 2013; Kadesjö et al., 2004). More generally, the a-specific pattern of correlations might be explained by the halo effect (Nisbett & Wilson, 1977). According to Stone et al. (2010), halo effects on questionnaires may occur “when one class of behavior influences the perception, and thus the rating, of other behaviors”. Previous research has repeatedly shown that ratings of e.g., ADHD done by parents have significantly been influenced by these halo effects (Jackson & King, 2004). Therefore, screening questionnaires should be followed by an assessment including multiple sources such as interviews and performance tests.

Regarding differences between relevant groups in PASSC sum scores, we first studied the influence of demographic variables (i.e., age and sex). We found that, according to the parents, boys scored on average higher on inattention and hyperactivity-impulsivity symptoms than girls. This difference is in line with the sex-specificity theory of ADHD, affecting more boys than girls both in clinical samples and in the general population (Biederman et al., 2002). Furthermore, we examined the relation between age, in our study defined as school grade, and PASSC sum scores. Scores on the 2 ADHD symptom dimensions differed between school grades. PASSC inattention sum scores were stable over different school grades, whereas hyperactivity-impulsivity -sum scores decreased over grades. The age-related decrease in hyperactivity-impulsivity symptoms and the stability of inattention symptoms over age are in line with the literature (Larsson et al., 2011). Like the PASSC inattention sum score, there were no differences in how parents explained inattention by time, cognition and/or motivation problems between children in different grades. Furthermore, we compared scores on the PASSC of children with an ADHD diagnosis to children with another DSM-5 diagnosis than ADHD and to children without any DSM-5 diagnosis, as reported by parents. Results indicated that on the 2 ADHD symptom dimensions (i.e., inattention and hyperactivity-impulsivity), children with ADHD scored higher than children with another DSM-5 diagnosis, who scored higher on their turn than children without a diagnosis. Likewise, parents of children with ADHD more often explained inattention by time, cognition, and motivation problems than parents of children with another DSM-5 diagnosis, who again reported higher scores on the explanation sum scores than parents of children without a diagnosis. This is in line with reports claiming that e.g., problems inattention and/or cognition are not specific for ADHD (Bloemen et al., 2018; Nigg et al., 2005).

Concluding, the PASSC is a promising screening tool to assess a child’s ADHD symptoms as well as the parent view on (potential) explanations of inattention with sufficient psychometric properties.

Limitations and future research

The current study is limited by the use of parent report measures only, which can be subject to biases and subjectivity. Future research should extend the current work to multiple methods and multiple informants, i.e., correlate the PASSC to additional measures, such as other reports (e.g., teacher, self-reports (if possible)) and performance tests measuring time, cognition and motivation. These additional measures are necessary to further examine the convergent and discriminant validity of the PASSC. Such an extended study might answer open questions, such as, how comparable are the given TPM explanations for inattention symptoms by parents on the PASSC to TPM explanations based on performance tests, and how comparable are given TPM explanations for inattention symptoms by parents on the PASSC between a general and a clinical ADHD population.

A future follow-up study in which parents are interviewed to explain, e.g., how they interpreted the PASSC questions and concepts like time estimation and management, cognition and motivation problems, is needed. This qualitative information will increase insight into their understanding of the used concepts and might require adaptations to the instructions of the questionnaire, making them more detailed and may be adding more concrete examples before administering the PASSC. Additionally, examining the impact of giving concrete examples for the 3 explanations on how parents fill in the PASSC has great value. This might lead to future adaptations in the PASSC making the number of given examples equal over the explanation items. It would be interesting as well to explore whether parents can differentiate within the cognitive explanation between different cognitive functions such as attention and inhibition as clarification for the behavioral symptoms. Moreover, this follow-up study might explore and compare the option to ask parents to attribute the main explanation on the ADHD dimension level, i.e., for all inattention symptoms, instead of on the symptom level, given that the majority of parents indicated the same explanation across attention items. This might be more efficient, which is important for screening purposes and is in line with other questionnaires, asking parents to indicate the daily impairment of problem behavior on a scale instead of item level (Kadesjö et al., 2004).

The PASSC is as a screening questionnaire that identifies children who might have ADHD based on parent reports. This result gives rise to further evaluation and treatment, but should not be interpreted on its own. Future longitudinal studies can give information about the predictive validity of the PASSC, i.e., how well predict PASSC sum scores an ADHD diagnosis and/or the underlying core problem determined by neuropsychological test performance.

Consent to participate

Informed consent was obtained from all individual participants included in the study.

Ethics approval

Approval was obtained from the ethics committee of Maastricht University, Faculty of Psychology and Neurosciences. The procedures used in this study adhere to the tenets of the Declaration of Helsinki.

Acknowledgments

The authors want to thank all participating schools, teachers and parents for their cooperation, the internet research agency Flycatcher, and in particular April Boessen, for the pleasant and successful collaboration, and Robin Barenbrug, Michelle Jacobs, Silke Kellens, Caro Paffen, Christine Resch, Jeanine Rongen, Denise Stefens, Anke van Treek, Luca Visse and Amber Weijers for their invaluable contribution to recruitment and testing.

Disclosure statement

The authors declare that they have no conflict of interest.

Additional information

Funding

This work was supported by the Abbas Foundation [in Dutch: “Stichting Abbas Fonds”], which aims to stimulate the psychometric quality of psychological instruments.