%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
For many decades the dimensional and typological approaches to parenting have mapped out the key components of parenting and explored their relations to predict unique variance using developmental criteria (Bean, Barber, & Crane, Reference Bean, Barber and Crane2006).
In general, both approaches essentially agreed that the quality of parent-child interactions would be adequately captured in a set of three interrelated dimensions, featuring parental support, behavior control and psychological control (Barber & Xia, Reference Barber, Xia, Morris, Larzelere and Harrist2013).
Parental support covers the parental attitudes of autonomy-support, that promote self-initiation, freedom of expression and intrinsic motivation (Barber, Reference Barber1996; Deci & Ryan, Reference Deci and Ryan1985), and responsiveness-warmth, related to the affective and involved ways through which parents interact with their children (Barber, Stolz, & Olsen, Reference Barber, Stolz and Olsen1995; Soenens, Duriez, Vansteenkiste, & Goossens, Reference Soenens, Duriez, Vansteenkiste and Goossens2007). Behavior control features the positive and active parental efforts intended to regulate or provide structure for the children’s behavior (Barber, Reference Barber1996; Steinberg, Reference Steinberg, Feldman and Elliot1990, Reference Steinberg, Damon and Smetana2005). Finally, psychological control characterizes the manipulative and autonomy-inhibiting parental attitudes of guilt-induction, shaming, love withdrawal and invalidation of the child’s perspective that intrudes on the child’s individuality (Barber, Reference Barber1996; Reference Barber2002; Barber & Harmon, Reference Barber, Harmon and Barber2002).
However, the broad consensus gained in the identification of the parenting dimensions contrasts with the diverse modeling approaches that have been used to examine their dimensionality. On the one hand, the research conducted on the dimensional approach to parenting (e.g., Gray & Steinberg, Reference Gray and Steinberg1999) modeled one or two of the three parental dimensions at a time, at the risk of exaggerating or misinterpreting the effect of specific dimensions, when other dimensions were not considered. On the other hand, the research based on the typological approach to parenting (e.g., Baumrind, Reference Baumrind1966) usually aggregated specific dimensions to form different parenting styles or clusters, making it impossible to isolate and to examine the unique or joint effects of specific parental dimensions on motivational outcomes (Bean et al., Reference Bean, Barber and Crane2006).
The modeling diversity has subsidized the persistence of ambiguities at both the conceptual and operational levels (Bean et al., Reference Bean, Barber and Crane2006). For instance, despite the broad consensus obtained for the linear positive effects of supportive/nurturing parenting based on developmental criteria (for a review see Ryan & La Guardia, Reference Ryan, Guardia, Qualls and Abeles2000), there is still some confusion regarding the linear, piecewise or even non-linear effects of behavior control on motivational outcomes (Soenens & Byers, Reference Soenens and Beyers2012; Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010; for a review, see Barber et al., 2005). Likewise, despite the consistent support obtained for the effects of parental psychological control on maladjustment (e.g., Barber, Reference Barber1996; Barber et al., 2005; Wang, Pomerantz, & Chen, Reference Wang, Pomerantz and Chen2007), it is still not absolutely clear how other dimensions of parental psychological control relate to ill-being and maladjustment (e.g., rejection, chaos; for a review, see Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010).
To this ambiguous findings contributed the fact that the parenting dimensions/typologies have been identified from predominantly empirically-driven approaches (e.g., psychological control; Schaefer, Reference Schaefer1965). Although the inductive approach yielded numerous important insights in the meaning and consequences of parenting, the research on parenting would benefit from top-down, or relatively more theory driven approaches (Steinberg, Reference Steinberg, Damon and Smetana2005) to bring a more conceptual and operational unity to the findings.
A Self-Determination Theory Perspective on Parenting
Self-Determination Theory (SDT; Deci & Ryan, Reference Deci and Ryan1985, Reference Deci and Ryan2000) provides a comprehensive account of the dynamics and mechanisms through which parenting impacts the psychosocial development of children (Skinner, Johnson, & Snyder, Reference Skinner, Johnson and Snyder2005). SDT posits that the effects of parenting on development are mediated by the satisfaction of the basic psychological needs of autonomy, competence and relatedness (Deci & Ryan, Reference Deci and Ryan1985; Ryan, Reference Ryan and Jacobs1993; Ryan, Reference Ryan1995). Autonomy is the need to experience self-endorsement, volition and choice in the initiation and regulation of personal actions (Deci & Ryan, Reference Deci and Ryan1985,), competence is the need to be effective in goal attainment and environmental interactions, in the process of mastering various challenges (Deci & Ryan, Reference Deci and Ryan1985; Schunk & Zimmerman, Reference Schunk and Zimmerman2007) and relatedness, the desire to feel emotionally connected to others within warm, supportive and caring interpersonal relationships (Baumeister & Leary, Reference Baumeister and Leary1995; Ryan, Reference Ryan1995). For SDT, the basic psychological needs are innate and universal psychological mechanisms that energize and regulate goal-directed behavior towards psychological development, integrity, and well-being, in a continual dialectical interplay between the organismic tendency towards growth and psychological development, and the degree to which parents, as primary socialization agents, support or thwart the satisfaction of needs (Deci & Ryan, Reference Deci and Ryan1985, Reference Deci and Ryan2000).
Parents promote basic need-satisfaction when they support the children´s autonomy in a volitional way (Autonomy-support-PVF; Deci & Ryan, Reference Deci and Ryan1985, Reference Deci and Ryan2000), when they interact with their children in a responsive and warm manner (Responsiveness-warmth; Barber, Reference Barber1996; Deci & Ryan, Reference Deci and Ryan1985; Ryan, Reference Ryan and Jacobs1993; Soenens et al., Reference Soenens, Duriez, Vansteenkiste and Goossens2007) or when they provide structure for their behavior ( behaviour control; Barber, Reference Barber1996; Barber, Olson, & Shaggle, Reference Barber, Olsen and Shagle1994; Grolnick & Ryan, Reference Grolnick and Ryan1989). Autonomy-supportive parents provide an optimal amount of choice for their actions, or an adequate rationale when choice is constrained, and refrain from using insidious, manipulative and invasive practices (Deci & Ryan, Reference Deci and Ryan1985, Reference Deci and Ryan2000). Responsive-warmth parents easily attune and empathize with their children’s experiences and feelings, and interact with their children in involved, affectionate and accepting ways (Soenens, Vansteenkiste, Duriez et al., Reference Soenens, Vansteenkiste, Luyckx and Goossens2006). Finally, behaviorally controlling parents make positive efforts to regulate and structure the child’s behavior (e.g., manners, study activities, and involvement with peers) through the provision of clear expectations/rules and active monitoring efforts (e.g., Barber, Reference Barber2002).
Parents may also actively thwart or block the satisfaction of psychological needs (parental psychological control; Deci & Ryan, Reference Deci and Ryan2000, Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013), when they use psychological control to regulate the psychological experiences and behavior of their children (Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010). For SDT, psychological control characterizes the internally controlling and manipulative techniques of guilt-induction, shaming and love withdrawal used to coerce the children to change their psychological (thoughts, feelings) and behavioral experiences according to their expectations (Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010).
The research based on SDT has consistently demonstrated that parental need-support and behavior control promote growth, intrinsic motivation and psychological well-being via the experience of basic need-satisfaction (Deci & Ryan, Reference Deci and Ryan1985; Ryan & Deci, Reference Ryan and Deci2000; Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013), whereas parental psychological control relates to maladjustment, ill-being and psychopathology, through the experience of basic need-frustration (Barber, Reference Barber1996; Bartholomew, Ntoumanis, Ryan, & Thøgersen-Ntoumani, Reference Bartholomew, Ntoumanis, Ryan and Thøgersen-Ntoumani2011; Cordeiro, Paixão, Lens, & Sheldon, Reference Cordeiro, Paixão, Lens and Sheldon2015; Soenens et al., Reference Soenens, Luyckx, Vansteenkiste, Duriez and Goossens2008; Soenens, Vansteenkiste, Luyten, Duriez, & Goossens., Reference Soenens, Vansteenkiste, Luyten, Duriez and Goossens2005; Verstuyf, Vansteenkiste, Soenens, Boone, & Mouratidis, Reference Verstuyf, Vansteenkiste, Soenens, Boone and Mouratidis2013) Footnote 1 .
Although generally accepted, most findings were obtained from narrow-scoped factor analysis (for an exception, Skinner et al., Reference Skinner, Johnson and Snyder2005), making it difficult to establish firm conclusions about the validity of the parental dimensions assessed and the precise ways through which the parenting dimensions relate (e.g., interact) to predict the developmental outcomes (Barber & Xia, Reference Barber, Xia, Morris, Larzelere and Harrist2013).
The present study
In this study we will examine the factor structure of the Parenting Questionnaire Scales (PQS; Soenens, Vansteenkiste, Duriez et al., Reference Soenens, Vansteenkiste, Luyckx and Goossens2006), to provide evidence for their and construct validity. This aim is of particular importance, for two reasons. From a conceptual point of view we are among the few to use of a top-down approach, and particularly the Self-Determination Theory, to integrate the findings (See et al., 2005 for a similar approach). Methodologically we offer, for the first time, a broad factor-analysis of the full set of the scales.
As a first goal we examined the factor structure of the PQS in two Models. The first - Model1m,f (m for mothers; f for fathers), tested the PQS in a two-factor structure. The first factor measures parental need-support (Duriez, Soenens, & Vansteenkiste, Reference Duriez, Soenens and Vansteenkiste2007) in a composite score combining the scales of autonomy-support (POPS; Grolnick, Ryan, & Deci, Reference Grolnick, Ryan and Deci1991), responsiveness-warmth (CRPBI; Schaefer, Reference Schaefer1965), and (a lack of) psychological control (YSR; Barber, Reference Barber1996). The second factor assesses behavioral control in a composite score that aggregates the measures of expectations for behavior and monitoring of behavior (PRS – YSR; Barber, Reference Barber2002). The second - Model 2m,f tested the PQS in an alternative three-factor structure. The first factor measures parental need-support, in a composite score that aggregates the scales of autonomy-support and responsiveness-warmth. The second factor assesses behavior control in a composite score that combines the measures of expectations for behavior and monitoring of behavior. The third factor measures parental psychological control from the psychological control scale. In both Model 1m,f and Model 2m,f the dimensions of behavior control and psychological control are measured as two separate factors, in line with the SDT argument that behavior control, deals with “what” parents do to regulate their children’s behavior outcomes, whereas autonomy-support refers to “how” parents implement it (Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010). However, Model 1m,f and Model 2m,f diverge in the way as the Psychological Control items are codified. In fact, whereas in Model 1m,f the psychological control items were reverse-scored to measure parental need-support (see Miklikowska, Duriez, & Soenens, Reference Miklikowska, Duriez and Soenens2011 for a similar approach), in Model 2m,f they are direct-scored to assess psychological control. We expect a better fit for Model 2m,f, based on the SDT premise that the parental need-support and the psychological control dimensions of parenting have different substantive interpretations and effects (e.g., Bartholomew et al., Reference Bartholomew, Ntoumanis, Ryan and Thøgersen-Ntoumani2011; Sheldon, Abad, & Hinsch, Reference Sheldon, Abad and Hinsch2011; Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013).
In a second goal we examined the criterion-related validity of the best-fitting model, according to five SDT-based hypotheses (Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013). In a first hypothesis we expect the primary effects of parental need-support and of behavior control on need-satisfaction and of parental psychological control on need-frustration (Hypothesis 1). Second, we expect the cross-lagged effects of parental need-support on need-frustration and of psychological control on need-satisfaction (Hypothesis 2). Third, we hypothesize that parental need-support and psychological control moderate the positive effect of behavior control on need-satisfaction (Hypothesis 3a), that parental need-support moderates the positive effect of psychological control on need-frustration (Hypothesis 3b), and also that psychological control moderates the positive effect of parental need-support on need-satisfaction (Hypothesis 3c). We intended to verify whether our hypothesis are valid for the developmental stage of adolescence, the reason why we used a sample of high school students.
Method
Participants
We sampled 371 Portuguese high school students (grade 10: n = 101 [27.2%], grade 11: n = 148 [39.9%], grade 12 n = 122 [32.9%]), of both sexes (Male: n = 171 [46.1%], female n = 200 [53.9%]), aged between 16 and 23 years old (M = 18; SD = 1.309)2. Students attended scientific-humanistic (n = 153 [41.2%]) and technical-vocational courses (n = 218 [58.8%]) in public (n = 182 [49.1%]) and private schools (n = 189 [50.9%]). Passive informed consent was obtained from the parents of younger students. All subjects volunteered for the study and completed the questionnaires without missing responses. No credits were granted for participating in the study.
Procedure
Prior to data collection the researchers obtained the mandatory permissions from the General Directorate for Innovation and Curricular Development and from the school principals. Next, the questionnaires were group-administered in the classroom, during regular class hours. The primary researcher read aloud the instructions of the PQS “The following statements deal with the way in which your father/mother behaves towards you; indicate to what degree you agree with these statements by circling one of the numbers”, and of the Balanced Measure of Psychological Need Scale (Sheldon & Hilpert, Reference Sheldon and Hilpert2012) ‘‘Please read each of the following statements carefully, thinking about how true it is for you”. Aspects such as the voluntary participation and confidentiality of the data were secured in the instructions. Students took around 20 minutes to complete the questionnaires.
Measures
Perceived Parenting
The 76-item Parenting Questionnaire Scales was used to measure perceived parenting. The PQS is not a questionnaire per se, but a composite instrument composed of five scales: the 7-item autonomy-support scale, retrieved from the Perceptions of Parents Scales (POPS; Grolnick et al., Reference Grolnick, Ryan and Deci1991, e.g., “My mother/father is usually willing to consider things from my point of view), the 7-item responsiveness-warmth scale, included in the Child Report of Parent Behavior Inventory (CRPBI; Schaefer, Reference Schaefer1965, e.g., “My father/mother makes me feel better after I discuss my worries with him/her”), the 8-item psychological control scale, integrated in the Youth Self-Report (PCS-YSR; Barber, Reference Barber1996; e.g., “My mother/father changes the subject whenever I have something to say), the 8-item expectations for behavior scale (e.g., “My mother/father believes that children should not be able to do anything they want”) and the 8-item Monitoring of Behavior Scale (e.g., “My mother/father asks me questions about how I am behaving outside the home”), included in the Parental Regulation Scale – Youth Self-Report (PRS-YSR; Barber, Reference Barber2002). The PQS was rated separately for mothers (N = 38) and fathers (N = 38), on a 6-point Likert-type scale, ranging from 1 (Totally Agree) to 5 (Totally disagree). The Cronbach’s alphas reported for maternal and paternal ratings ranged from .67 to .70 for autonomy-support (Grolnick et al., Reference Grolnick, Ryan and Deci1991), .88 to .92 for responsiveness-warmth (Soenens et al., Reference Soenens, Vansteenkiste, Luyten, Duriez and Goossens2005), .82 to .80 for psychological control (Barber, Reference Barber1996), and .83 to .82 for behavior control (Soenens, Vansteenkiste, Luyckx et al., Reference Soenens, Vansteenkiste, Luyckx and Goossens2006).
Translation
We translated the Parenting Questionnaire Scales into Portuguese using the back-translation procedure (Hambleton, Reference Hambleton2001). A professional interpreter collaborating with the fluent English-speaking researchers translated the PQS from English to Portuguese. Next, an independent interpreter translated the scales back into English. Both original and back-translated versions were checked for accuracy, and the discrepancies resolved through consensus. The readability and unambiguous understanding of the PQS items was further examined in a pilot study (N = 11 Portuguese high school students), resulting in the wording and syntax modifications of three items (items 5, 29, and 14).
Basic Psychological needs
We used the Portuguese adaptation of the Balanced Measure of Psychological Needs Scale (BMPN; Sheldon & Hilpert, Reference Sheldon and Hilpert2012; Portuguese version: Cordeiro et al., Reference Cordeiro, Paixão, Lens and Sheldon2015). The BMPN is an 18-item self-report questionnaire measuring basic psychological need-satisfaction and need-frustration in six three-item scales. Three positively-worded scales measure the satisfaction of autonomy (“My choices are based on my true interests and values”), competence (“I am successful at completing difficult tasks and projects”) and relatedness needs (“I feel a sense of contact with people who care for me, and whom I care for”). In addition, three negatively-worded scales measure the frustration of autonomy (e.g. “I do things against my will”), competence (“I do stupid things that make me feel incompetent”), and relatedness needs (“I feel unappreciated by one or more important people”). All items are rated on a 5-point Likert scale, ranging from 1 = no agreement to 5 = much agreement. The internal consistency of the scales reported for the Portuguese version of the BMPN was of .84, .79, and .82 for autonomy, competence, and relatedness satisfaction, and of .85, .82, and .77 for autonomy, competence, and relatedness frustration, respectively. In this study, we modeled the six BMPN scales in a structure of two second-order factors measuring the general experience of basic need-satisfaction and of basic need-frustration (α = .82, .85, respectively; Cordeiro et al., Reference Cordeiro, Paixão, Lens and Sheldon2015).
Plan of analysis
We examined the internal structure of the Parenting Questionnaire Scales using AMOS (V.20, SPSS Inc, Chicago, IL). In the first step we computed successive confirmatory factor analyses (CFA, Byrne, Reference Byrne2010) with ML estimation, to test the fit of Model 1m,f and Model 2m,f to the PQS data. Goodness-of-fit was judged from multiple fit indices, namely the Chi square (χ Footnote 2 ) statistics, the Standardized Root Mean Square Residual (RMR), the Comparative Fit Index (CFI) and the Root Mean Squared Error of Approximation (RMSEA). The combined cut-off values of .09 for SRMR, .06 for RMSEA, p (RMSEA ≤ 0.05) and .90, or above, for CFI, showed an acceptable fit (Hu & Bentler, Reference Hu and Bentler1999). The lowest values obtained in the Akaike’s Information Criteria (AIC; Byrne, Reference Byrne2010) indicated the preferred model. Further, we used the Standardized Factor Loadings and the Modification Indices to modify the best-fitting model. We excluded from further analysis all items presenting poor factor loadings (λi ≥ 0.5; λFootnote 2 ij ≥ 0.25; Maroco, Reference Maroco2010) or high cross-loadings (MI > 9). In the second step we performed an Exploratory Factor Analyses (EFA; McIver & Carmines, Reference McIver and Carmines1981), in principal components (PCA), and varimax orthogonal rotation to the data, in order to corroborate the best-fitting model found in CFA. In the third step we used the STATS Tool Package (Gaskin, 2012) to examine the Convergent (AVEi ≥ 0.5; Fornell & Larcker, Reference Fornell and Larcker1981) and Discriminant Validities (MSV < AVE; ASV < AVE; Fornell & Larcker, Reference Fornell and Larcker1981) of the factors. Finally, in the fourth step we performed successive Linear Regression Analyses to examine the main/cross-lagged effects between the variables. In addition, we used hierarquical regression analyses to examine for possible moderation effects. In the hierarquical regression procedure we examined the slope of the relationship between the predictors and the outcome variables, at low (one SD below the mean) and high (one SD above the mean) levels of moderator.
Preliminary Results
Confirmatory Factor Analysis
Table 1 summarizes the estimates of goodness-of-fit and model quality of Model 1m,f and Model 2m,f. Figure 1 provides a graphical representation of the respecified models. Initial CFAs yield an important misfit for the two models tested, nevertheless favoring the Model2m,f across all the fit indices considered. We modified the Model 2m,f. to improved goodness-of-fit. Firstly, we dropped from further analyses 14 items with high cross-loadings on non-intended factors (MI > 9; Maroco, Reference Maroco2010) and 12 items with poor loadings on the respective factor (λ ij ≥ 0.5). Secondly, we aggregated several scales presenting high empirical correlations (Maroco, Reference Maroco2010). Specifically, we combined the scales of expectations of behavior and monitoring of behavior (Model 2m r = .89, p < .001; Model 2f, r = .84, p < .001) to measure behavior control and the scales autonomy-support and responsiveness-warmth (Model2m, r = .85, p < .001; Model 2f, r = .95, p < .001) to measure parental need-support. In a final procedure we correlated the measurement errors of the items 1; 31, 31; 20, 5; 20. The resulting respecified Model (now designated Model 2mr,fr) is organized in a solution of 15 items and three factors, measuring parental need-support, behavior control and parental psychological control. Subsequent CFA results show the improved fit of Model 2mr,fr. In addition, the lowest AIC scores indicate that Model 2mr,fr is preferred to interpret the factorial structure of the PQS.
Global Fit Indices for the Hypothesized Models
Note: X2 qui-square; CFI = comparative fit index; RMSEA = root-mean-square error of approximation; RMR = root-mean-square residual; AIC = Akaike information criterion
Standardized estimates for M2 (maternal and paternal).
Model 2
Exploratory factor analysis (EFA)
Table 2 presents the reliability estimates for the PQS items and scales. The Exploratory Factor Analyses extracted a solution of three components with eigenvalues greater than 1, explaining 46, 20% (Model 2mr) and 57, 28% (Model 2fr) of the total variance of the data, respectively (Tiensley & Tiensley, Reference Tiensley and Tiensley1987). The three scales demonstrated good internal consistency, for both the maternal and paternal ratings, with all items loading significantly on the intended factor (λ ij ≥ .50; λ2 ij ≥ .25). Together, the EFA results corroborate the 3-factor solution found in CFA.
P-BMPN Reliability and Validity Estimates for the Reduced Solution of Portuguese version of the Parenting Questionnaire Items and Scales
Note 1: λ ij = standardized factor score weights; λ ij 2 individual-item reliability.
Note 2: Values between brackets correspond to Model 2fr estimates
Correlations
Table 3 summarizes the means, standard deviations and correlations obtained for the three factors. The examination of the correlation matrix shows a modest negative correlation (VIF < 5) between parental need-support and psychological control (Model 2mr r = –.34, p < .001 Model 2fr r = –.43, p < .001), suggesting that the factors measure two distinct parental dimensions. In addition, behavior control does not correlate to either parental need-support (Model 2mr r = .19, p = .72) nor to psychological control (Model 2mr r = .09, p = .79; M2fr r = –.07, p = .79), thus suggesting that behavior control is orthogonal to both the supportive and thwarting dimensions of parenting (see Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010 for a discussion). Remarkably, the high correlation observed between the dimensions of paternal behavior control and need-support Model 2fr (r = .56, p < .001) suggests that, for the Portuguese context, the paternal enforcement of behavior control is more positively-valued as an expression of need-support than the maternal.
Response rate (N), Means, Standard Deviations (SD), and Correlations of the Study Variables
Note: PNS = Parental Need-support; PPC = Parental Psychological Control; BC = Behaviour Control; AF = Autonomy Satisfaction; CF = Competence Satisfaction; RS = Relatedness Satisfaction; AF = Autonomy Frustration; CF = Competence Frustration; RF = Relatedness Frustration; NS = Need-satisfaction; NF = Need-frustration. Values between brackets correspond to Model 2fr estimate.
*p < .05 **p < .01.
Validity
Table 4 summarizes the validity estimates for the three PQS factors. We found Convergent validity (CV) for all the questionnaire factors except for the measures of maternal need-support and behavior control. The CV threats found are explained by the modest correlations verified between the dimensions of monitoring of behavior and autonomy-support in the respective second-order factors of need-support and behavior control. The additional CV threat detected for the factor paternal psychological control is justified by the modest loadings of the items in the factor. No threats to the discriminant validity of the scales were verified, thus suggesting the divergent validity of the three factors.
Factor Correlation, Reliability and Validity Estimates for the Models Tested
Note: R 2 = factor square correlations; Convergent validity (AVEi ≥ 0.5); Composite reliability (CR ≥ 0.7), Discriminant Validity (R2) Average Variance Extracted (AVE), Maximum Shared Squared Variance (MSV), and Average Shared Squared Variance (ASV).
Primary results
Table 5 portrays the results of linear regression analyses. We examined the effects of behavior control, parental need-support and psychological control on the adolescents´ experience of basic need-satisfaction and of need-frustration (Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013).
Regression Analysis Predicting Basic Need-Satisfaction and Need-Frustration
Note: *p < .05; **p < .01.
Main effects
The results support the main effects of parental need-support on general need-satisfaction. This effect was verified for both the first-order dimensions of autonomy-support, Model 2mr, β = .25, t(1,369) = 4.88, p < .001; Model 2fr, β = .14, t(1,369) = 2.78, p < .01, and responsiveness-warmth: Model 2mr, β = .19, t(1,369) = 3.76, p < .001; Model 2fr, β = .13, t(1,369) = 2.59, p < .05. However, when autonomy-support and responsiveness-warmth were entered together in a hierarchical regression procedure, the effect of responsiveness-warmth was reduced to non-significance, Model 2mr, β = .05, t(1,369) = .66, p = .52; Model 2fr, β = –.01, t(1,369) = – .06, p = .95, suggesting suppression effects. The results also support the main effect of behavior control on basic need-satisfaction, Model 2mr: β = .14, t(1,369) = 2,80, p < .05; Model 2fr: β = .08, t(1,369) = 1,78, p = .08, and, particularly, on competence satisfaction: Model 2mr: β = .16 t(1,369) = 3,16, p < .01; Model 2fr: β = .05, t(1,369) = .88, p = .38. However, behavior control was not a significant predictor of need-frustration, Model 2mr: β = –.03, t(1,369) = –.53, p = .60; Model 2fr β = –.05, t(1,369) = –.78, p = 43. Importantly, against what is predicted by SDT (Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013), the findings did not provide support for the main effect of psychological control on need-frustration, Model 2mr β = .03, t(1.369) = .56, p = .57; Model 2fr, β = .03, t(1.369) = .52, p = .60. Overall, the results provided partial support for the first hypothesis of our study.
Cross-lagged effects
We further examined the predicted cross-lagged effects between the variables (Vansteenkiste & Ryan, Reference Vansteenkiste and Ryan2013). We found that maternal psychological control had a significant negative effect on basic need-satisfaction, Model 2mr: β = –.14, t(1, 369) = –1, 36, p < .05), while parental need-support does not have a significant effect on need-frustration, Model 2mr: β = –.10, t(1, 369) = –.61, p > .05, Model 2fr: β = .04, t(1, 369) = –1.94, p > .05.
Moderation effects
In subsequent analyses we explored the existence of possible moderation effects between the variables. The results of hierarquical regression analysis showed that both parental need-support and psychological control moderated the positive effect of behavior control on general need-satisfaction, with parental need-support amplifying this effect and psychological control buffering it. We also found that maternal psychological control buffered the positive effect of need-support on need-satisfaction, Model 2mr: F(3, 367) = 11.62, p < .001; Model 2fr: F(3, 367) = 3.49, p < .05). The asymmetrical cross-lagged and moderation effects found provided partial support the second and third (3a, 3c) hypotheses of our study.
Discussion
The aims of this study were twofold: (1) to perform a broad-band factor-analytic study of dimensionality and construct validity of the Parenting Questionnaire Scales in Portuguese sample of high school students, and (2) to examine the construct validity of the parenting dimensions, with reports of convergent, discriminant and criterion-related validity. The findings are interpreted on the basis of Self-Determination Theory.
The combined results of exploratory and confirmatory factor analyses, along with the lowest AIC estimates obtained for Models 2mr,fr showed that the internal structure of the Parenting Questionnaire Scales is best-represented by a solution of 15-items and 3 factors that distinguish the dimensions of parental need-support, parental psychological control and behavior control.
The improved construct validity of the three-factor solution was further demonstrated at the correlation matrix. As expected, parental need-support and parental psychological control were moderately correlated, signifying that the constructs lie within two distinct motivational continua. In addition, behavior control does not significantly relate to both parental need-support and psychological control, suggesting that the parental efforts to regulate children’s behavior, based on reasonable expectations and adequate monitoring of behavior, are independent of the supportive or thwarting ways through which these efforts are communicated (Soenens & Vansteenkiste, Reference Soenens and Vansteenkiste2010).
The 3-factor model demonstrated adequate discriminant validity, but further adjustments are necessary to improve the Convergent Validity of some scales. The criterion-related validity of the three factors was also demonstrated. In particular, it was found that the Portuguese high school students experience basic need-satisfaction when they perceive their parents as highly need-supportive (particularly autonomy-supportive) and/or behaviorally controlling, but they do not necessarily experience basic need-frustration when they their parents display psychological control (or psychologically controlling) attitudes. Instead, the perception of parental psychological control, and particularly of maternal psychological control, is related to the experience of low need-satisfaction, whereas neither parental need-support nor behavior control relate to the experience of low need-frustration.
Importantly, the effect of parental psychological control on (low) need-satisfaction was not affected by the degree to which parents are simultaneously perceived as need-supportive and behavior controlling. On the contrary, the positive effects of parental need-support on basic need-satisfaction are significantly buffered by the experience of psychological control.
Finally, we found that parental need-Support and maternal psychological control moderated the positive effect of behavior control on the adolescents´ experience of basic need-satisfaction, with parental Need-support amplifying this effect and psychological control buffering it.
Overall, the results indicate that the most optimal pattern of adolescent need-satisfaction is attained when parents combine need-supportive, behavior controlling, and (the lack of) parental psychological control attitudes. Put it in a different way, lower need-satisfaction is experienced when the parents exhibit psychological controlling attitudes, irrespectively on how much need-supportive and behavior controlling they simultaneously are.
One should note, however, that the predictive effects were always stronger, if not only significant, for the maternal data. This particular finding underlines not only the importance of examining the maternal and paternal data separately (for a similar approach, see Barber et al., Reference Barber, Stolz and Olsen1995), but also the need to examine the differential impact of parental and maternal variables in development.
Future studies could examine whether different parenting profiles, resulting from different combinations of the three parental dimensions, predict unique variance on motivational outcomes (e.g., psychosocial identity; Erikson, Reference Erikson1968; Marcia, Reference Marcia and Adelson1980) and whether this relation is mediated by the experience of basic need-satisfaction and frustration. For instance, it could be examined whether the need-supportive or psychologically controlling behaviors are predominantly triggered by particular emotional states of the parents or by specific features of the child’s behavior. To this point, one could hypothesize that parents tend to be mainly need-supportive when they feel more relaxed or when the child displays appropriate behavior, and more psychological controlling when they feel more anxious or when the child is misbehaving (B. Soenens, personal communication, March 3, 2014).
Our research has several limitations. Firstly we conducted a single cross-sectional study based on self-report measures. This methodology prevented us from drawing firm conclusions about the distinctiveness of the three factors insofar as the differences found may also reflect methodological artifacts, such as the positive or negative way as items are worded. To overcome this problem, future research should combine adolescent and parental self-reports, or use more objective criteria, such as the physiological correlates associated to need-satisfaction and to need-frustration. Additionally, more prospective longitudinal studies should be undertaken to more completely address the way in which these effects develop over time.
Secondly, in this study we relied on a relatively homogeneous and well-educated sample of Portuguese high school adolescents. Now, the cultural and sampling specificity may not only exemplify two confounding variables to be controlled for, but they can themselves represent alternative explanations for the results. More cross-cultural validation studies are required to exclude the alternative hypothesis that the salience of the associations between the variables reflects the cultural a cultural bias rather than the real nature and dynamics of the constructs.
The use of a normative sample also restricted the variance of our data. In fact, we verified that the scores of parental psychological control were all below the scale mid-point, which, for many authors, indicate the absence of the construct under analysis (e.g., YSI; Young, Klosko, & Weishaar, Reference Young, Klosko and Weishaar2003). Therefore, we may have not have measured the full experience of psychological control, leaving unchecked the hypothesis that high psychological control scores would have a significant impact on need-frustration. To overcome this limitation, future research should rely on a combination of normative and clinical samples.
Finally, we based our conclusions on a shortened version of the Parenting Questionnaire Scales, and, as we know the findings obtained with shorter scales are less valid than those obtained with longer scales (Smith, McCarthy, & Anderson, Reference Smith, McCarthy and Anderson2000). Therefore it is necessary to cross-validate our findings in independent samples, if we want to generate extended evidence for the psychometric quality and predictive capacity of the 3-factor solution.
This study provided initial validation for the substantive distinction between the need-supportive, psychological control and behaviorally controlling dimensions of parenting, adding new questions on the antecedents, dimensionality and relations between perceived parenting and experienced need-satisfaction and/or frustration in adolescence.
This research was fully supported by the first author’s scholarship from the Foundation for Science and Technology - Portugal.
Appendix A:
Parenting Questionnaire Items – English version (PQS)
Appendix B:
Parenting Questionnaire Items – Portuguese version
