Psychometric properties of the parent’s versions of the SDQ and the PANAS-X in a community sample of Portuguese parents

Abstract

This study examined the psychometric properties of the parent’s versions of the Strengths and Difficulties Questionnaire (SDQ) and the Positive and Negative Affect Schedule–Extended Form (PANAS-X) in a community sample of Portuguese parents. A total of 1100 SDQ and PANAS-X were collected from an online sample of 761 parents, whose ages ranged from 23 to 65 years (M = 42, SD = 5). Confirmatory factor analyses provided evidence of the internal factor structure of both the SDQ and the PANAS-X and invariance of the factor structure across age and gender groups, with the exception of the SDQ, which failed to provide evidence of invariance between genders. Internal reliability and discriminant validity were confirmed for both measures, although convergent validity was only confirmed for the PANAS-X. Concurrent validity was also confirmed by comparing the results from the SDQ dimensions and the PANAS-X broad dimensions subscales.

Keywords

Child development psychosocial adjustment conduct problems mood-related symptoms screening measures

Introduction

The Strengths and Difficulties Questionnaire (SDQ) is a widely used behavioral screening measure that covers different symptomatology in children and youth (Goodman, 1997). It is highly regarded for its sensitivity in detecting behavioral and emotional problems, and some studies have provided evidence of the screening accuracy of the SDQ in comparison to other well-established measures such as the Child Behavior Checklist (CBCL; Warnick et al., 2008). The Positive and Negative Affect Schedule (PANAS) purports to assess positive and negative affect dimensions that characterize temperamental traits associated with anxiety and depression, and also can be used as part of a screening assessment of different anxiety and mood symptomatology (Watson et al., 1988). Their brevity, ease of administration, and adaptation studies conducted in several countries and in different languages are among the strengths of both of these measures. However, there are few and sometimes contradictory findings regarding the internal structures of both measures (e.g., Goodman et al., 2010).

The SDQ

The SDQ is a highly regarded screening measure of emotional and behavioral problems in children and youth (Goodman, 1997). It has been widely used in research as well as in child mental health practice due to its usefulness in identifying problematic behavior, emotional problems, and social difficulties at home and at school. There is both theoretical and empirical evidence of the existence of five dimensions namely Conduct Problems, Hyperactivity/Inattention, Emotional Symptoms, Peer Relationship Problems, and Prosocial Behaviors although the creator of the scale has later provided further evidence of three dimensions in which both conduct problems and hyperactivity/inattention comprise an Externalizing dimension, emotional symptoms and peer relationships problems comprise an Internalizing dimension, while the third dimension consists of only positive, prosocial behaviors. The authors recommended that the five separate dimensions can be used for screening of mental health or behavioral problems while the three dimensions can be used for assessing rates of problems in low-risk community samples (Goodman et al., 2010).

The SDQ is available in over 80 languages and dialects and has three main versions—one for parents (children 2–4 years old or children 4–17 years old), one for teachers (children 2–4 years old or children 4–17 years old), and one self-report for children/adolescents (11–17 years old). However, in 2015 normative data were only available in 10 countries, excluding Portugal (Youth in Mind, 2015). When compared with another well-established child mental health screening measures (e.g., the CBCL, Achenbach & Ruffle, 2000), the SDQ has shown to be highly reliable in discriminating between clinical and community samples. Some studies have found that the SDQ was at least as good as the CBCL in screening for emotional and behavioral problems, and some have reported that the few differences between the two measures seemed to favor the SDQ (Warnick et al., 2008).

There is little evidence of the psychometric properties of the SDQ with Portuguese samples. A search at the website “Youth in Mind” (www.sdqinfo.com), which collects information about studies using the SDQ worldwide, returned only one result for Portugal (Marzocchi et al., 2004), although a Portuguese translation is available on the website (Fleitlich et al., 2005). The self-report Portuguese version of the SDQ has shown contradictory findings, with some studies reporting difficulties in confirming the five-factor model of the scale and weak construct validity (Pechorro et al., 2011). Other studies have confirmed the five-factor structure in the three versions (self-report, teacher and parent) but have found some reliability issues and have failed to show evidence of its discriminant validity (Azevedo et al., 2013; Marzocchi et al., 2004).

The PANAS

The PANAS is a widely used screening measure for mood-related symptomatology, namely depression, and anxiety (Watson et al., 1988). Contrary to the SDQ, it is not focused on behavioral symptomatology. Furthermore, it was initially developed with adults as a self-report measure but has since then evolved to be used with children and youth. In fact, it is now regarded as a useful screening measure for affective problems in children (e.g., Lonigan et al., 2003). The PANAS consists of 20 items divided into broad categories of Positive Affect and Negative Affect and asks respondents to recall the extent to which they felt each emotion during a specified period. Clark and Watson (1991) proposed that the PANAS scales measure both positive and negative affect which in differing proportions underlie both depression and anxiety. Clarke and Watson’s tripartite model of anxiety and depression posited that people with depression and anxiety shared high negative affect, although people with depression would score low on positive affect while people with anxiety would score high on positive affect. Several studies have corroborated the tripartite model providing evidence of the correlation between PANAS scores and emotional and internalizing problems (Anthony et al., 2002; Mikolajewski et al., 2013).

The PANAS has been used in various studies about affectivity and mood-related problems in Portugal, although little research has been devoted to investigate its psychometric properties and its validity as a screening measure. Furthermore, only the PANAS with the broad positive and negative affect dimensions has been studied in Portugal. Three psychometric studies have been published using Portuguese samples, and all three studies were conducted by the same research team (Galinha et al., 2013; Galinha & Pais Ribeiro, 2005). Using the equivalent process to the original development studies conducted by Watson et al. (1988), the original 60 items were reduced to a final 20-item Portuguese version of the Positive Affect and Negative Affect (Galinha & Pais Ribeiro, 2005). In the final version, six out of the 10 items in the Positive Affect dimension were the same, and seven out of 10 in the Negative Affect dimension were the same. Furthermore, both the internal consistency and inter-factor correlations were equivalent to the original version. Further studies corroborated the factorial stability of the PANAS and temporal invariance (Galinha et al., 2013). However, these studies were conducted with adults, mostly using students or young adult samples. To our knowledge, the extended version of the PANAS has never been used for research in Portugal, or psychometrically assessed with children and adolescents.

The present study

Given the identified psychometric problems, the aim of this study was to provide a more complete picture of a child’s socioemotional development through the combined use of two highly regarded measures of child’s adjustment—The SDQ and the PANAS-X—that could be useful both in clinical and research work to assess both problematic and positive behavioral indicators and subjective well-being. Thus, we aimed to evaluate the psychometric properties of the parents’ version of both the SDQ and the PANAS-X, and investigate the contribution of both measures together for a complete assessment of child psychosocial adjustment through the analyses of reliability, sensibility, construct, and criterion validities in a community sample of Portuguese parents.

Materials and methods

Participants and procedures

The online survey was named “Parenting: Capabilities and Difficulties,” and the study aims were described as exploring the strengths, difficulties, and challenges that Portuguese parents may face in raising their children, and how these may relate to the well-being of their children. A total of 1100 completed questionnaires were collected, and from these, 18% were completed by fathers and 82% were completed by mothers. Parents’ ages ranged from 23 to 65 (M = 42; SD = 5). Children’s ages ranged from 6 to 18 years (M = 11, SD = 3), and gender groups were balanced. Parents were recruited through contacts with schools and parents’ associations. Schools and parents’ associations were sent an email describing the study aims and instructions on how to participate in the study. Further information could be obtained by contacting the main researcher. The questionnaires were completed and hosted on an online survey platform. No compensation was offered to participants. Informed consent was obtained from all individual participants.

Measures

The Strengths and Difficulties Questionnaire

Parents were asked to complete the SDQ (Goodman, 1997). The SDQ is a brief behavioral screening measure, composed of 25 items divided into four difficulties subscales with five items each: Emotional Symptoms, Conduct Problems, Hyperactivity/Inattention, and Peer Relationship Problems, and one positive subscale: Prosocial Behavior. Items are measured on a 3-point Likert-type scale, from 0 (not true) to 2 (certainly true). Higher scores on the four difficulties subscales reflect greater difficulties, whereas higher scores on the positive subscale reflect greater prosocial behaviors, although no cutoff values on the SDQ are currently available in Portugal.

The Positive and Negative Affect Schedule

Parents were asked to complete the PANAS-X (Watson & Clark, 1994) regarding their children. PANAS-X is a 60-item adjective checklist designed to measure positive and negative affect. It includes 10 items for General Negative Affect and 10 items for General Positive Affect. Items are measured on a 5-point Likert-type scale, from 1 (very slightly or not at all) to 5 (extremely). Given the lack of validation studies, we decided to test the original PANAS-X scales and asked parents to describe their children over the last few weeks. Higher scores on the negative affect subscale reflect greater negative affect, whereas higher scores on the positive affect subscale reflect greater positive affect.

Data analysis

The data were analyzed using the Predictive Analysis Software (PASW) Statistics version 21 and the Analysis of Moment Structures (AMOS) version 21. PASW was used to calculate descriptive statistics and correlations between variables. AMOS was used to conduct confirmatory factor analyses (CFA) and multi-group confirmatory factor analyses (MGCFA) of the SDQ and the PANAS-X. CFA models were evaluated with χ², and an alpha level of .05 was used to determine statistical significance, using Maximum Likelihood as estimator. There were no missing data on either measure. Model fit was assessed using the chi-square test, the comparative fit index (CFI), the goodness of fit index (GFI), the Tucker Lewis index (TLI), and the root mean square error of approximation (RMSEA). Values above .90 on the CFI, the GFI, and the TLI, and below .08 on the RMSEA are indicators of an acceptable model fit. Values above .95 on the CFI, the GFI, and the TLI, and below .05 on the RMSEA are indicators of good model fit (Hair et al., 2006). Model invariance between groups (gender and age groups) regarding baseline model (i.e., unconstrained) and resulting model (fixed factor loadings, structural means, and structural covariances) was calculated using ΔCFI (⩽ .01) (Byrne, 2010).

Results

Construct validity of the SDQ

Factorial validity

Following the proposed factor structure of the SDQ, a CFA was performed. Multivariate normality was confirmed for all but two items (Items 12 and 22), which had both Skewness (|Sk| > 3) and Kurtosis (|Ku| > 7) problems (Table 1). However, to fully assess the construct validity of the Portuguese version of the SDQ, these two items were not deleted. All factors were allowed to correlate. The first-order five-factor model had an acceptable fit, χ²(263) = 1125.079, p < .001, CFI = .860, GFI = .920, TLI = .830, RMSEA = .055, 90% CI [.051, .058], AIC = 1249.079, BIC = 1559.269. All items significantly loaded onto their subscale [Supplementary Table 1]. Internal consistency reliabilities (Cronbach’s alpha) ranged from low to acceptable: α = .64 for Emotional Symptoms; α = .56 for Conduct Problems; α = .79 for Hyperactivity/Inattention; α = .60 for Peer Relationship Problems; and α = .70 for Prosocial Behavior.

Table 1.

Goodness of fit statistics for the measurement models and factorial invariance for gender and age groups for the three-factor model for the SDQ.

Description	Gender
Description	χ²	df	GFI	CFI	TLI RMSEA	RMSEA 95% CI
2. Multiple-group model
2a. Unconstrained	1069.609*	522	.871	.818	.809 .042	[.038, .045]
2b. FL	11141.277*	544	.864	.802	.789 .043	[.039, .046]
2c. FL + SM	1205.709*	569	.864	.789	.781 .043	[.040, .046]
2d. FL + SM + SC	1219.464*	575	.862	.796	.683 .043	[.040, .047]
2e. FL + SM + SC + MR	1575.104*	611	.831	.680	.679 .051	[.048, .054]
	Age groups
	χ²	df	GFI	CFI	TLI RMSEA	RMSEA 95% CI
2. Multiple-group model
2a. Unconstrained	1587.650*	534	.895	.817	.815 .042	[.040, .045]
2b. FL	1634.722*	556	.892	.813	.815 .042	[.040, .044]
2c. FL + SM	1832.371*	581	.890	.783	.813 .044	[.042, .047]
2d. FL + SM + SC	1859.099*	587	.809	.779	.806 .044	[.042, .047]
2e. FL + SM + SC + MR	2002.674*	617	.882	.759	.798 .045	[.043, .047]

Note. SDQ = Strengths and Difficulties Questionnaire; GFI = goodness of fit index; CFI = comparative fit index; TLI = Tucker Lewis index; RMSEA = root mean square error of approximation; FL = factor loadings; SM = structural means; SC = structural covariances; MR = measurement residuals.

p < .001.

A second CFA was performed to assess the three-factor model, consisting of Internalizing Problems (emotional symptoms & peer relationship problems), Externalizing Problems (conduct problems & hyperactivity/inattention), and Prosocial Behavior. All factors were allowed to correlate. The first-order three-factor model had an acceptable fit, χ²(265) = 1178.017, p < .001, CFI = .842, GFI = .919, TLI = .831, RMSEA = .056, 90% CI [.053, .059], AIC = 1298.017, BIC = 1598.201. All items significantly loaded onto their subscale. Internal consistency reliabilities (Cronbach’s alpha) were all acceptable; Internalizing: α = .72, Externalizing: α = .79, Prosocial Behavior: α = .70.

Overall, although model fit for the three-factor solution and for the five-factor solution were similar, we chose to report only the former solution for the remainder of the analysis as it has been suggested its better adequacy in assessing rates of problems in low-risk community samples (Goodman et al., 2010).¹ The correlations among the subscales were all significant. The correlation between Internalizing and Externalizing was .379, the correlation between Prosocial Behavior and Externalizing was –.328, and the correlation between Prosocial Behavior and Internalizing was –.244 (all ps < .001).

Invariance of the factor structure

Two multi-group factor analyses were performed to assess the invariance of the three-factor model structure between gender groups (boys and girls) and age groups (6–10 years and 11–18 years). The criterion for dividing the age groups was based on the applications of the SDQ; normative data are available for these age groups and also for gender groups. Model fit statistics are shown in Table 1. Regarding the invariance for gender groups, when comparing the model with factor loadings constrained (Model 2b) with the unconstrained model (Model 2a), results confirmed the non-invariance of the factor loadings, ΔCFI = .002. When comparing the model with mean structure constrained (Model 2c) with the unconstrained model (Model 2a), results confirmed the invariance of structural mean, ΔCFI = .03. When comparing the model with structural covariances constrained (Model 2d) with the unconstrained model (Model 2a), results confirmed the invariance of the structural covariances, ΔCFI = .02. Regarding the invariance for age groups, when comparing the model with factor loadings constrained (Model 2b) with the unconstrained model (Model 2a), results confirmed the invariance of the factor loadings, ΔCFI = .004. When comparing the model with structural means constrained (Model 2c) with the unconstrained model (Model 2a), results confirmed the invariance of mean structure, ΔCFI = .03. When comparing the model with structural covariances constrained (Model 2d) with the unconstrained model (Model 2a), the results confirmed the invariance of the structural covariances, ΔCFI = .04 [Supplementary Table 2].

Reliability, and convergent and discriminant validities

To assess the reliability of each factor in the three-factor model, all three-factor revealed acceptable reliability but low convergent validity. Discriminant validity was assessed by calculating the square root of average extracted mean which must be higher than inter-factor correlations. Discriminant validity was confirmed for all comparisons (Table 2).

Table 2.

Reliability, convergent validity, and discriminant validity for the SDQ.

SDQ dimensions	Three-factor model
SDQ dimensions	R ²	CR	AVE
Internalizing (I)	.043	.716	.204
Externalizing (E)	.160	.772	.265
Prosocial Behaviors (P)	.076	.710	.337
	ρ² (Fij)		Discriminant validity
Internalizing (I)—Externalizing (E)	.144		<AVE I; <AVE E^a
Externalizing (E)—Prosocial Behaviors (P)	.108		<AVE E; <AVE P^a
Internalizing (I)—Prosocial Behaviors (P)	.060		<AVE I; <AVE P^a

Note. SDQ = Strengths and Difficulties Questionnaire; CR = composite reliability; AVE = average variance extracted.

Evidence of discriminant validity.

Construct validity of the PANAS-X

Factorial validity

Following the proposed factor structure of the PANAS-X, a CFA was performed. Multivariate normality was confirmed for all but one item (Item 31 “scared”), which had Skewness (|Sk| > 3) and Kurtosis (|Ku| > 7) problems. However, to fully assess the construct validity of the Portuguese version of the PANAS-X, this item was not deleted. Both factors were allowed to correlate. The first-order two-factor model had an acceptable fit, χ²(158) = 828.137, p < .001, CFI = .911, GFI = .926, TLI = .893, RMSEA = .062, 90% CI [.058, .066]. All items significantly loaded onto their subscale [Supplementary Table 3]. All items significantly loaded onto their subscale. Internal consistency reliabilities (Cronbach’s alpha) were α = .857 for Positive Affect and α = .802 for Negative Affect. The correlation between the two broad dimensions (Positive and Negative Affect) in the two-factor model was weak but significant, r = –.276, p < .001.

Invariance of the factor structure

Two multi-group factor analyses were performed to assess the invariance of the two-factor model structure between genders (boys and girls) and age groups (6–10 years and 11–18 years). Model fit statistics are shown in Table 3. Regarding the invariance for gender groups, when comparing the model with factor loadings constrained (Model 2b) with the unconstrained model (Model 2a), results confirmed the invariance of the factor loadings, ΔCFI = .001. When comparing the model with mean structure constrained (Model 2c) with the unconstrained model (Model 2a), results confirmed the invariance of mean structure, ΔCFI = .003. When comparing the model with structural covariances constrained (Model 2d) with the unconstrained model (Model 2a), results confirmed the invariance of structural covariances, ΔCFI = .003. Regarding the invariance for age groups, when comparing the model with factor loadings constrained (Model 2b) with the unconstrained model (Model 2a), results confirmed the invariance of the factor loadings, ΔCFI = .001. When comparing the model with structural means constrained (Model 2c) with the unconstrained model (Model 2a), results confirmed the invariance of mean structure, ΔCFI = .03. When comparing the model with structural covariances constrained (Model 2d) with the unconstrained model (Model 2a), the results confirmed the invariance of structural covariances, ΔCFI = .03 [Supplementary Table 4].

Table 3.

Goodness of fit statistics for the measurement models and factorial invariance for gender and age groups for the two-factor model for the PANAS-X.

Description	Gender
Description	χ²	df	GFI	CFI	TLI RMSEA	RMSEA 90% CI
2. Multiple-group model
2a. Unconstrained	733.029*	316	.871	.896	.875 .047	[.042, .051]
2b. FL	754.062*	334	.864	.895	.881 .046	[.041, .050]
2c. FL + SM	785.418*	354	.862	.893	.885 .045	[.041, .049]
2c. FL + SM + SC	786.748*	357	.862	.893	.886 .045	[.040, .049]
2d. FL + SM + SC + MR	846.435*	388	.831	.886	.888 .044	[.040, .048]
	Age groups
	χ²	df	GFI	CFI	TLI RMSEA	RMSEA 90% CI
2. Multiple-group model
2a. Unconstrained	979.166*	316	.916	.911	.898 .044	[.041, .047]
2b. FL	1006.508*	334	.913	.910	.897 .043	[.040, .046]
2c. FL + SM	1211.814*	354	.911	.885	.895 .047	[.044, .050]
2c. FL + SM + SC	1237.806*	357	.911	.882	.888 .047	[.045, .050]
2d. FL + SM + SC + MR	1302.333*	388	.904	.877	.888 .046	[.044, .049]

Note. PANAS-X = Positive and Negative Affect Schedule; GFI = goodness of fit index; CFI = comparative fit index; TLI = Tucker Lewis index; RMSEA = root mean square error of approximation; FL = factor loadings; SM = structural means; SC = structural covariances; MR = measurement residuals.

p < .001.

Reliability, and convergent and discriminant validities

To assess the reliability of each factor of both the two-factor and the four-factor models, composite reliabilities were calculated. All factors had acceptable reliabilities (composite reliability [CR] > .7) and convergent validities (average variance extracted [AVE] > .5). Discriminant validity was assessed by calculating the square root of average extracted mean, which must be higher than inter-factor correlations (Fornell & Larcker, 1981). Discriminant validity was confirmed for all but one comparison in the four-factor model (Table 4).

Table 4.

Reliability, convergent validity, and discriminant validity for the PANAS-X.

PANAS-X dimensions	Three-factor model
PANAS-X dimensions	R ²	CR	AVE
Negative Affect (NA)	.342	.810	.307
Positive Affect (PA)	.631	.860	.391
	ρ² (Fij)		Discriminant validity
Negative Affect (NA)—Positive Affect (PA)	.076		<AVE NA; <AVE PA^a

Note. PANAS-X = Positive and Negative Affect Schedule; CR = composite reliability; AVE = average variance extracted.

Evidence of discriminant validity.

Criterion validity

Concurrent validity

To assess the concurrent validity of the SDQ and its subscales with the PANAS-X and its subscales, bivariate correlations among the subscales were calculated (Table 5). Regarding the SDQ broad Internalizing and Externalizing and the PANAS-X general Positive and Negative Affect, moderate and high correlations were found, and particularly a correlation of .572 between Internalizing and Negative Affect. In addition, the SDQ Prosocial Behaviors subscale had a moderate positive correlation with Positive Affect and negative correlations with Negative Affect. In sum, concurrent validity between the SDQ and its subscales and the PANAS-X and its subscales were demonstrated.

Table 5.

Correlations among the SDQ and PANAS-X and its subscales.

SDQ	PANAS-X
SDQ	Negative affect	Positive affect
Externalizing	.472*	−.326*
Internalizing	.572*	−.332*
Prosocial behaviors	−.201*	.327*

Note. PANAS-X = Positive and Negative Affect Schedule; SDQ = Strengths and Difficulties Questionnaire.

p < .001.

Discussion

The main aims of this study were to assess the psychometric properties and construct and criterion validities of the parents’ version of the Strengths and Difficulties Questionnaire and the Positive and Negative Affect Schedule, and to examine the potential contribution of both measures together for a more complete assessment of child psychological adjustment that encompasses problematic as well as positive behavioral indicators. We will proceed with the discussion of the findings separately for each measure.

The SDQ

The findings regarding the psychometric properties of the SDQ were generally positive. The three-factor structure, which is often used in research on child psychosocial adjustment outcomes showed acceptable internal consistencies for the three subscales. Likewise, both the five-factor and the three-factor structures of the SDQ were confirmed in Italian community samples by Di Riso et al. (2010) with the three-factor model revealing a better fit to the data. However, the authors reported some low factor loadings on some items and generally low Cronbach’s alphas. Another Italian study (Li et al., 2017) was able to confirm both the five-factor and the three-factor model with acceptable Cronbach’s alphas, but the overall model fit was low with CFI fit indexes below .90. One review of 48 studies assessing the psychometric properties of the parents and teachers’ version of the SDQ highlighted divergent results across studies, with some replicating the five-factor structure and others the three-factor structure, and only a few able to replicate both structures (Stone et al., 2010).

The discriminant validity of the three subscales was confirmed by the analysis of the Portuguese SDQ scores. Thus, there was evidence that the SDQ items are good indicators of internalizing, externalizing, and prosocial behaviors, as well as of the validity of the measurement scales (Fornell & Larcker, 1981). Nonetheless, it should be also acknowledged that two items showed significant violations to multivariate normality: Item 12 “Often fights with other children or bullies them,” and Item 22 “Often lies or cheats.” It may have been the case that parents were not aware of their children problematic behaviors, such as fighting, or that these are not prevalent in low-risk community samples. We suggest that future studies should further investigate any problems in the normal distribution of these two items in both clinical and community samples. The convergent validity of the SDQ assessed through the AVE was generally low for each of the three subscales in contrast with the moderate composite reliabilities found. Other authors have suggested that the AVE is a very strict and conservative measure of convergent validity (Malhotra & Dash, 2011) and suggested that measurement of convergent validity can rely solely on adequate composite reliability. Nevertheless, we recommend further assessment of convergent validity of the SDQ three-factor model.

The model invariance observed between age groups provided evidence of the construct validity of the SDQ for different ages, corroborating the validity of the SDQ to screen problem behaviors in a wide age range (from 6 years to 18 years old). However, our analyses indicated variations in the SDQ factor structure for boys and girls. This result is consistent with other reported differences between boys and girls in SDQ scores (e.g., Marzocchi et al., 2004), with boys scoring higher than girls on Externalizing behaviors and girls scoring higher than boys on Internalizing behaviors. Nevertheless, the absence of model invariance raises some concerns over the construct validity of the SDQ subscales considering that the factor loading between the two groups significantly differed. Thus, future studies should further investigate whether gender differences could be observed in self-report SDQ and/or teacher’s reports on the SDQ as there might be cultural differences in parental and youth perceptions of gender acceptable behavior.

The PANAS

Regarding the PANAS-X, the internal consistency of both negative and positive affect was high. Furthermore, both convergent and discriminant validities were observed. Model invariance was confirmed for both gender groups and age groups in the two-factor model of the broad affect dimensions, which has been reported elsewhere for children’s self-report scores (Lonigan et al., 1999, 2003). Nevertheless, to the best of our knowledge, this study was the first to use the PANAS-X to gather parental perspectives of children’ affective traits. Our findings provided further evidence of the stability of the PANAS-X construct validity and factor structure. It is noteworthy that this stability was observed in a sample that until now has been underresearched. While some authors have been developing newer and shorter version of the PANAS, namely the PANAS-C (Laurent et al., 1999) and the PANAS-C-P (Ebesutani et al., 2011), it is promising that the original PANAS-X might also be useful for children and youth because the scale already has been established with adults.

Implications and directions for future research

Both the SDQ and the PANAS-X are well-established measures of child and youth psychosocial difficulties. Neither measure was intended to diagnose psychopathology per se, but each was conceived as a screening measure that would be sensitive to problematic behaviors or emotional difficulties. As such, both the SDQ and the PANAS-X perform well and are extremely useful in identifying early signs of problems for children and youth. Nevertheless, researchers need to be particularly aware of the cultural differences that may affect the internal structures of these measures, especially for the SDQ given the difficulties in confirming the invariance of its internal structure with respect to gender, and thus be mindful of the need to assess the psychometric properties of these measures before analyzing their data. A vital strength of both the SDQ and the PANAS-X is the assessment of positive competences and features that other child adjustment measures do not contemplate. In fact, focusing on strengths as well as on difficulties may also highlight positive child developmental areas that can be informative when developing intervention programs for at-risk children and youth, and crucially reinforce the positive competences already available to the young person.

Limitations

The current study was not without limitations. Regarding the sample in this study, the characteristics of the parents (i.e., highly educated, with medium-high income, and mostly mothers) and of the recruitment procedures (i.e., online and through convenience sampling) may have influenced the results from this study. To the extent to which the results could be generalized to the Portuguese population cannot be determined. Future studies should aim to recruit more diverse samples. Furthermore, the present study only examined the factor structure, internal consistency, and concurrent validity of the parents’ version of the SDQ and the PANAS-X, and future studies should also evaluate test–retest reliability and predictive validity of these measures. Finally, this study relied solely on parent’s perspectives, and studies with teacher and child’s perspectives are needed to further validate the results found in this study.

Conclusion

This study’s findings provided further evidence of the construct and criterion validities of the SDQ and the PANAS from the parent’s perspective. Although some shortcomings of the SDQ were identified, namely some factorial instability across genders and two problematic items, the scale showed an overall good fit to the data. Regarding the PANAS, the findings were mostly positive and a good fit to the data was also found. We documented evidence of adequate reliability, factorial, convergent, and discriminant validities in a large sample of Portuguese parents. This evidence is compelling given, on the one hand, the scarcity of validation studies on these measures in different countries, and on the other hand, the need to develop reliable cross-cultural measures of children’s behavioral and emotional adjustment and social competences. Normative data on the SDQ is available for only 10 countries, whereas no normative data on the PANAS-X currently exists. Further efforts to develop norms and cutoff scores to identify problematic behavior and emotional difficulties in children and youth are needed.

Supplemental Material

Supplementary_Tables – Supplemental material for Psychometric properties of the parent’s versions of the SDQ and the PANAS-X in a community sample of Portuguese parents

Supplemental material, Supplementary_Tables for Psychometric properties of the parent’s versions of the SDQ and the PANAS-X in a community sample of Portuguese parents by Pedro Alexandre Costa, Fiona Tasker, Catarina Ramos and Isabel Leal in Clinical Child Psychology and Psychiatry

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Ethics approval

This study was approved by the Ethics Committee of ISPA—Instituto Universitário.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Fundação para a Ciência e a Tecnologia (FCT) under Grant number SFRH/BPD/99752/2014, awarded to the first author.

Informed consent

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

ORCID iD

Pedro Alexandre Costa

Supplemental material

Supplemental material for this article is available online.

Notes

Author biographies

Pedro Alexandre Costa is a researcher at the William James Center for Research, ISPA, Portugal. He holds a PhD in psychology from the Universidade da Beira Interior. Costa has specialized in child development and family dynamics in different family types, namely families through adoption, donor insemination, and LGBT+ parented families.

Fiona Tasker, PhD, is a reader in psychology at the Department of Psychological Sciences, Birkbeck University of London. Tasker has specialized in family psychology and systemic family therapy and is on the editorial boards of several journals. She has published widely on the psychological implications of both nontraditional and new family forms for parents and children, in heterosexual and LGBTQ-parented families.

Catarina Ramos is a researcher at the William James Center for Research at ISPA and at Lab-PSI, CiiEM, IUEM– Instituto Universitário Egas Moniz, Portugal. She has holds a PhD in health psychology and has specialized in psychometric analysis and psycho-oncology.

Isabel Leal is a psychologist and psychotherapist. She has a PhD in psychology from the Catholic University of Louvain and she is currently the coordinator of the Translational Group of the William James Center for Research, ISPA, Portugal.

References

Achenbach

T. M.

Ruffle

T. M.

(2000). The Child Behavior Checklist and related forms for assessing behavioral/emotional problems and competencies. Pediatrics in Review, 21, 265–271. https://doi.org/10.1542/pir.21-8-265

Anthony

J. L.

Lonigan

C. J.

Hooe

E. S.

Phillips

B. M.

(2002). An affect-based, hierarchical model of temperament and its relations with internalizing symptomatology. Journal of Clinical Child & Adolescent Psychology, 31, 480–490. https://doi.org/10.1207/S15374424JCCP3104_7

Azevedo

A. F.

Seabra-Santos

M. J.

Gaspar

M. F.

Homem

T. C.

(2013). The incredible years basic parent training for Portuguese preschoolers with AD/HD behaviors: Does it make a difference? Child and Youth Care Forum, 42, 403–424. https://doi.org/10.1007/s10566-013-9207-0

Byrne

B. M.

(2010). Structural equation modeling with AMOS: Basic concepts, applications, and programming (2nd ed.). Routledge.

Clark

L. A.

Watson

(1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100, 316–336. https://doi.org/10.1037/0021-843x.100.3.316

Di Riso

Salcuni

Chessa

Raudino

Lis

Altoè

. (2010). The Strengths and Difficulties Questionnaire (SDQ). Early evidence of its reliability and validity in a community sample of Italian children. Personality and Individual Differences, 49, 570–575. https://doi.org/10.1016/j.paid.2010.05.005

Ebesutani

Okamura

Higa-McMillan

Chorpita

B. F.

(2011). A psychometric analysis of the Positive and Negative Affect Schedule for Children-Parent Version in a school sample. Psychological Assessment, 23, 406–416. https://doi.org/10.1037/a0022057

Fleitlich

Loureiro

Fonseca

Gaspar

(2005). Questionário de Capacidades e Dificuldades (SDQ-Por). www.sdqinfo.com

Fornell

Larcker

D. F.

(1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18, 39–50. https://doi.org/10.2307/3151312

10.

Galinha

I. C.

Pais Ribeiro

(2005). Contribuição para o estudo de uma versão portuguesa da Positive Negative Affect Schedule (PANAS): II estudo psicométrico [Contribution for the study of the Portuguese version of the PANAS: II psychometric study]. Análise Psicológica, 23, 219–227.

11.

Galinha

I. C.

Pereira

C. R.

Esteves

(2013). Confirmatory factor analysis and temporal invariance of the Positive and Negative Affect Schedule (PANAS). Psicologia: Reflexão e Crítica, 26, 671–679. https://doi.org/10.1590/S0102-79722013000400007

12.

Goodman

Lamping

D. L.

Ploubidis

G. B.

(2010). When to use broader internalising and externalising subscales instead of the hypothesised five subscales on the Strengths and Difficulties Questionnaire (SDQ): Data from British parents, teachers and children. Journal of Abnormal Child Psychology, 38, 1179–1191. https://doi.org/10.1007/s10802-010-9434-x

13.

Goodman

(1997). The Strengths and Difficulties Questionnaire: A research note. Journal of Child Psychology and Psychiatry, 38, 581–586. https://doi.org/10.1111/j.1469-7610.1997.tb01545.x

14.

Hair

Black

Babin

Anderson

Tatham

(2006). Multivariate data analysis (6th ed.). Pearson; Prentice Hall.

15.

Laurent

Catanzaro

S. J.

Joiner

T. E.

Jr. Rudolph

K. D.

Potter

K. I.

Lambert

Osborne

Gathright

(1999). A measure of positive and negative affect for children: Scale development and preliminary validation. Psychological Assessment, 11, 326–338. https://doi.org/10.1037/1040-3590.11.3.326

16.

J. B.

Delvecchio

Di Riso

Lis

Salcuni

(2017). Early evidence of the Italian Parent-Report Strengths and Difficulties Questionnaire (SDQ-P). Child Psychiatry & Human Development, 48, 335–345. https://doi.org/10.1007/s10578-016-0646-3

17.

Lonigan

C. J.

Hooe

E. S.

David

C. F.

Kistner

J. A.

(1999). Positive and negative affectivity in children: Confirmatory factor analysis of a two-factor model and its relation to symptoms of anxiety and depression. Journal of Consulting and Clinical Psychology, 67, 374–386. https://doi.org/10.1037/0022-006x.67.3.374

18.

Lonigan

C. J.

Phillips

B. M.

Hooe

E. S.

(2003). Relations of positive and negative affectivity to anxiety and depression in children: Evidence from a latent variable longitudinal study. Journal of Consulting and Clinical Psychology, 71, 465–481. https://doi.org/10.1037/0022-006x.71.3.465

19.

Malhotra

N. K.

Dash

(2011). Marketing research: An applied orientation (6th ed.). Pearson.

20.

Marzocchi

G. M.

Capron

Di Pietro

Duran Tauleria

Duyme

Frigerio

Gaspar

M. F.

Hamilton

Pithon

Simões

Thérond

(2004). The use of the Strengths and Difficulties Questionnaire (SDQ) in Southern European countries. European Child and Adolescent Psychiatry, 132(Suppl. 2), 40–46. https://doi.org/10.1007/s00787-004-2007-1

21.

Mikolajewski

A. J.

Allan

N. P.

Hart

S. A.

Lonigan

C. J.

Taylor

(2013). Negative affect shares genetic and environmental influences with symptoms of childhood internalizing and externalizing disorders. Journal of Abnormal Child Psychology, 4, 411–423. https://doi.org/10.1007/s10802-012-9681-0

22.

Pechorro

Poiares

Vieira

(2011). Propriedades psicométricas do Questionário de Capacidades e de Dificuldades na versão portuguesa de auto-resposta [Psychometric properties of the Portuguese self-report version of the Strenghts and Difficulties Questionnaire]. Revista de Psiquiatria Consiliar e de Ligação, 16–19, 99–109.

23.

Stone

L. L.

Otten

Engels

R. C.

Vermulst

A. A.

Janssens

J. M.

(2010). Psychometric properties of the parent and teacher versions of the Strengths and Difficulties Questionnaire for 4- to 12-year-olds: A review. Clinical Child and Family Psychology Review, 13, 254–274. https://doi.org/10.1007/s10567-010-0071-2

24.

Warnick

E. M.

Bracken

M. B.

Kasl

(2008). Screening efficiency of the Child Behavior Checklist and Strengths and Difficulties Questionnaire: A systematic review. Child and Adolescent Mental Health, 13, 140–147. https://doi.org/10.1111/j.1475-3588.2007.00461.x

25.

Watson

Clark

L. A.

(1994). The PANAS-X: Manual for the positive and negative affect schedule-expanded form. The University of Iowa.

26.

Watson

Clark

L. A.

Tellegen

(1988). Development and validation of brief measures of positive and negative affect: The PANAS scales. Journal of Personality and Social Psychology, 54, 1063–1370. https://doi.org/10.1037/0022-3514.54.6.1063

27.

Youth in Mind. (2015). Information for researchers and professionals about the Strengths & Difficulties Questionnaires. http://www.sdqinfo.com

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.36 MB