The role of parenting styles on behavior problem profiles of adolescents

International Journal of Environmental Research and Public Health Article The Role of Parenting Styles on Behavior Problem Profiles of Adolescents Bárbara Lorence 1, Victoria Hidalgo 1, Javier Pérez-Padilla 2,* and Susana Menéndez 3 1Faculty of Psychology, University of Seville, Calle Camilo JoséCela, s/n, 41018 Sevilla, Spain 2Faculty of Humanities and Education Sciences, University of Jaen, Campus Las Lagunillas, s/n, 23071 Jaén, Spain 3Faculty of Education, Psychology and Sports Sciences, University of Huelva, Campus El Carmen, Avenida de las Fuerzas Armadas, s/n, 21007 Huelva, Spain *Correspondence: [email protected]; Tel.: +34-953-212-614 Received: 11 June 2019; Accepted: 1 August 2019; Published: 2 August 2019   Abstract: Parental behavior is one of the most influential factors on the development of adolescent externalizing and internalizing behavior problems. These behavioral problems are closely related and often co-occur. The objectives of this work were: (i) to identify adolescents profiles according to their behavior problems; (ii) to explore individual, family, and social characteristics associated with these profiles; and (iii) to analyze the potential role of parenting styles in belonging to adolescents’ profiles. A total of 449 Spanish adolescents (223 from families declared at-risk and enrolled in Child Welfare Services and 226 from families from the general population) participated in this study. The analyses revealed three profiles of adolescents based on external and internal behavior problems (adjusted, external maladjustment, and internal maladjustment). Parenting styles explained the adolescents’ belonging to different profiles, in which the indulgent style was the most favorable in general terms. The distinctive role of parenting styles on two types of maladjustment profiles was confirmed. The relationship between parenting styles and adolescent adjustment is a key component that should be included in interventions according to adolescents’ behavior problem profiles. Furthermore, the results shed light on the need that family interventions are complemented with individualized interventions with adolescents that accumulate stressful life events. Keywords: parenting styles; adolescent; behavior problems; stressful life events 1. Introduction There is broad consensus that the goal of parenting is to establish positive relationships with children and adolescents within the family to ensure their development and well-being. The contemporary concept of positive parenting [ 1 ] implies that parent–child relationships should be based on affection, support, communication, stimulation, and structuring in routines, in the establishment of limits, norms, and consequences, as well as in the involvement in the daily life of children and adolescents [ 2 ]. However, positive parenting is a difficult task, especially during adolescence, when there is a tendency toward an increase in family conflict, which is due in part to developmental changes and new challenges faced by boys and girls [ 3 , 4 ]. In view of this reality, current family support policies in most countries provide parental education programs for the parents of adolescents [ 1 ]. In these interventions, parenting styles are often a central component in promoting positive parenting. This paper explores the relationship between parenting styles and adolescent adjustment profiles, with the overall aim of identifying key components of parenting styles for preventing internalizing and externalizing problems. Int. J. Environ. Res. Public Health 2019,16, 2767; doi:10.3390/ijerph16152767 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2019,16, 2767 2 of 17 1.1. Psychosocial Adjustment in Adolescence: Behavior Problems and Self-Concept Normative changes that occur during adolescence have been related to an increase in self-perceived maladjustment during this developmental stage. Psychological adjustment in adolescence is explained by the interaction of the multiple changes that take place at these ages [ 5 , 6 ]. In terms of self-development, there tends to be a drop in boys’ and girls’ assessment of themselves during early adolescence, with this gradually recovering as they reach adulthood [ 7 ]. Self-concept is an indicator of adjustment associated with behavior problems [ 8 ]. These types of problems affect many adolescents [ 9 ], although few persist with behavioral problems into adulthood [ 10 , 11 ]. There is an increase in externalizing problems, such as delinquency, substance abuse, aggressiveness, or breaking of rules [ 12 , 13 ]; and internal problems related to social isolation, withdrawal, anxiety, and depression [ 14 , 15 ]. It has been found that the prevalence of certain disorders follows different patterns in males and females. Most of the studies that make sex comparisons highlight a higher rate of internalizing problems in girls [ 16 , 17 ] and a higher rate of externalizing problems in boys [ 6 , 17 – 19 ], although not all recent studies found such a clear-cut difference along sex lines [ 20 , 21 ]. The co-occurrence, which is referred to the simultaneous presence of both internalizing and externalizing problems, has also come under close scrutiny in adolescent studies [9,22]. The previous research goals tend to focus exclusively on internalizing or externalizing problems, without taking into account the development of co-occurring symptoms for some children. The studies of the etiology or treatment of internalizing and externalizing symptoms in adolescents should account for co-occurrence (e.g., identifying existing adolescent adjustment profiles). In addition to normative changes, several studies report a strong association between the accumulation of stressful life events in adolescents and both internalizing and externalizing problems, regardless of sex [ 23 – 26 ]. The accumulative risk hypothesis understands that experiencing stressful circumstances has a negative impact on people’s development, so that the greater the number of stressful events, the greater the presence of clinical problems [ 27 ]. Studies with adolescents living in at-risk families enrolled in Child Welfare Services (CWS) found complicated risk trajectories in these youngsters, finding that they were particularly susceptible to developing this type of behavioral problem [ 28 – 30 ]. That high level of stress may be more harmful to the development of boys and girls in these surroundings than to the adults themselves due, amongst other reasons, to adolescents’ lack of psychosocial maturity to cope with these adversities [ 28 – 30 ]. Self-concept studies have shown that those adolescents who have experienced a greater number of significant negative experiences, particularly in areas which are important to them, present more difficulties and fluctuations in self-concept [7]. 1.2. Parenting and Psychosocial Adjustment in Adolescence Parental behavior is one of the most influential factors in terms of the development of externalizing and internalizing behaviors of adolescents. A great amount of research has focused on which parenting styles are best at promoting the growth and development of adolescents [31,32]. The typological perspective of parenting is the most widely used, and includes a constellation of parental dimensions such as affection, punishment, dialogue, or indifference [ 31 , 32 ]. This perspective allows a multidimensional approach on parenting [ 33 ]. Traditionally, the relations between parenting styles and child adjustment have been studied from the classical two-dimensional orthogonal model of parental socialization. Based on these two dimensions (acceptance/involvement and strictness/imposition), Musitu and Garc í a [ 34 ] identified four parental socialization styles: authoritative (characterized by the use of high strictness/imposition and high acceptance/involvement); neglectful (low strictness/imposition and low acceptance/involvement); indulgent (low strictness/imposition and high acceptance/involvement); and authoritarian (high strictness/imposition and low acceptance/involvement). This typology emerges from a contextual [ 31 ] and situational [35] approach to parenting. The relation between parenting styles and adolescent adjustment has a great deal of empirical evidence. In general, findings from previous studies have shown better adolescent outcomes with authoritative parents and poorer with neglectful parents; while the authoritarian and indulgent Int. J. Environ. Res. Public Health 2019,16, 2767 3 of 17 styles have occupied an intermediate position [ 36 – 39 ]. In general, the available empirical data show adolescents from authoritative families perform better academically [ 40 ] are more optimistic [ 41 ], make better use of adaptive strategies [ 36 ], have less drug use [ 38 ], higher self-concept [ 42 ], more behavioral regulation [ 43 ] and are more resilient [ 44 ] than adolescents from families with other parenting styles. Additionally, this relationship remains invariant with regard to the sex of the participants in previous studies. [36,45,46]. Although in general terms the positive effects of the authoritative style and the negative effects of the neglectful style have been widely documented, some studies have shown variations in the effects of the different styles depending on the type of behavioral problem analyzed. In relation to externalizing problems, the review by Marcone, Affuso, and Borrone [ 43 ] showed that the most positive style for promoting external adolescent adjustment is the authoritative, and the most harmful is the authoritarian style. This same review did not find such clear-cut results in relation to the indulgent (sometimes associated with positive effects) and neglectful styles (sometimes associated with negative effect styles). These authors concluded that parental control has been linked to beneficial effects if it is exerted as supervision, behavioral control, or guide; in contrast, psychological, coercive, or restrictive control is associated with externalizing behaviors. There is less research into internalizing problems. The systematic review by Rose, Roman, Mwaba and Ismail [ 47 ] of the 2012–2015 period only found five studies focusing specifically on parenting styles and their relation to internalization (articles with children over the age of 12 years and which had a medical focus). These authors found that the lowest levels of internalizing problems were found in adolescents with authoritative parents, and the uninvolved parenting style (neglectful) was associated to the highest level of internalizing problems. In turn, they also found that hostile and punitive parenting was identified as a risk factor for internalization problems in several studies. Despite these results, the systematic review performed by Yap, Pilkington, Ryan, and Jorm [ 48 ] showed that the empirical evidence of the relation between authoritarian, authoritative, and indulgent styles and internalizing problems is weak and remains equivocal. The authors concluded that more empirical evidence is needed about the role of parenting styles in belonging to internalizing problems, because other significant factors have been identified for these types of behavior problems (temperamental response styles and emotion regulation strategies) [49]. The association between parenting and behavior problems seems to vary when co-occurrence of internalizing and externalizing problems is controlled. Likewise, Lorence et al. [ 50 ] found that the association between parenting and externalizing problems remains invariant once this co-occurrence is controlled, but not for internalizing problems. These authors found that the effect of parenting styles on internal adjustment disappeared while the influence of other variables, as the emotional impact of stressful life events accumulation, is maintained. Finally, the role of a parenting style varies according to the socio-cultural context [ 51 – 55 ]. Some studies in European or Latin American countries support the indulgent style as the optimum parenting style [ 56 – 61 ]. Parenting practices based on acceptance, support, and reasoned communication promote adolescent adjustment regardless of the cultural context, although according to Garc í a and Gracia [ 59 ], this is particularly beneficial in collectivist cultures, where cooperation and egalitarian relations are emphasized [ 58 , 62 ]. However, findings about the effects of parental control are more controversial when comparing different cultures [ 55 , 59 , 63 , 64 ] and specific social environments [ 65 – 67 ]. For all of the above, the results on parenting should be understood in the socio-cultural context (country) in which these are studied. 1.3. Objectives This research is proposed to add knowledge to the relationships between mothers’ parenting styles and adolescent adjustment in Spain. Specifically, the objectives of this study were the following: •To identify profiles of adolescents according to their behavioral problems. •To explore the individual, family, and social characteristics of these profiles. Int. J. Environ. Res. Public Health 2019,16, 2767 4 of 17 •To analyze the potential role of parenting styles in belonging to adolescents’ profiles. 2. Methods 2.1. Participants A total of 449 adolescents (211 females and 238 males; M=13.65 years, SD =1.89) from southern Spain participated in this research. The sample was split into two groups: 223 adolescents living in families declared at-risk and enrolled in Child Welfare Services (CWS), and 226 adolescents living in families not receiving intervention (community adolescents). Both groups were similar in age (t(447) =−0.250, p=0.803) and sex (χ2(1, N=449) =3.95, p=0.208). In regard to the families of the participants, the average number of children under 18 in these homes was 2.11 (SD =0.95), and 29.9% were single-parent families (100% single-mother families). The fathers’ mean age was 44.94 years (SD =6.60), the majority had completed primary education (31.3%) or had finished secondary education (29.9%), and 89% of the fathers were working. Regarding the mothers, the mean age was 42.15 years (SD =6.86), and 21.7% had not completed primary education, 37.6% had completed primary education, 25.5% had finished secondary education, and 15.2% had a university degree. As for employment, 62.4% of the mothers were working at the time. 2.2. Measures Socio-demographic profile: We developed a questionnaire to collect socio-demographic information about the adolescent (age, sex, and educational level), family (structure of family) and parents (age, sex, educational level, and employment situation). A short version of the Inventory of Stressful Life Events (ISLE, [ 68 ]) was administered to adolescents. Fifteen items from the original 29 negative or potentially problematic events were included in this study (situations not directly controllable by the adolescent have been chosen), covering stressful experiences at individual (“severe illness”, “sexual abuse”), family (“family member’s addiction”, “move”, “parental divorce”, “parental conflicts”, “new couple”, “economic hardship”, “death of family members”, “illness of family members”, “a family member moves/leaves home), and peer-related levels (“isolation”, “couple breakup”, “couple betrayal”, “fight with friends”). The occurrence of these kind of events in the previous five years was rated using a dichotomous scale (0 =no, 1 =yes). The calculation of the accumulated score of SLEs refers to the total number of stressful life events experienced by the subject. Parental Socialization Scale for Adolescents (ESPA29, [ 34 ]). This instrument assesses parental socialization strategies in 29 situations of everyday family life: 13 compliance situations (e.g., “a teacher calls your mother and tells her that you are behaving well in class”) and 16 non-compliance situations (e.g., “you arrive home late”). In each of the compliance situations, the adolescent had to rate the parenting practices of Affection (“my mother shows affection”) and Indifference (“my mother seems indifferent”), and in each of the non-compliance situations, the adolescent had to rate the parenting reactions on Dialogue (“my mother talks to me”), Detachment (“my mother doesn’t tell me anything”), Verbal Scolding (“my mother scolds me”), Physical Punishment (“my mother spanks me”), and Revoking Privileges (“my mother takes something away from me”). ESPA29 consists of 116 items ranging from 1 (never) to 4 (always), providing us with two parenting dimensions: (1) the Acceptance/Involvement dimension, the result of averaging the responses on Affection, Dialogue, Indifference, and Detachment (the last two subscales are reverse coded); and (2) the Strictness/Imposition dimension, by averaging the responses on Verbal Scolding, Physical Punishment, and Revoking Privileges. According to the ESPA29 original authors [ 34 ], after dichotomizing the total sample on acceptance/involvement and strictness/imposition, and examining the two parenting dimensions simultaneously, four parenting styles emerged: Authoritative (above the 50th percentile on both dimensions), Authoritarian (above the 50th percentile on strictness/imposition but below the 50th percentile on acceptance/involvement), Indulgent (above the 50th percentile on acceptance/involvement Int. J. Environ. Res. Public Health 2019,16, 2767 5 of 17 but below the 50th percentile on strictness/imposition) and Neglectful (below the 50th percentile in both dimensions). This study is focused exclusively on mothers’ parenting styles (main caregiver according to adolescents). Youth Self-Report (YSR, [ 18 ]) assesses behavior problems in adolescence. YSR consists of 112 items with response options ranging from 0 (not true) to 2 (very true), which measure eight behavior problems: Withdrawn (e.g., “I would rather be alone than with others”), Somatic complaints (e.g., “I feel dizzy or lightheaded”), Anxiety and depression (e.g., “I cry a lot”), Social problems (e.g., “I’m too dependent on adults”), Thought problems (e.g., “I see things that other people think aren’t there”), Attention problems (e.g., “I feel confused or in a fog”), Aggressive behavior (e.g., “I destroy things belonging to others”), and Rule-breaking behaviors (e.g., “I steal from places other than home”). The first three subscales are related to Internalizing problems, whereas the last two correspond to Externalizing problems. Cronbach’s alphas in this study were α =0.79 for internalizing problems and α =0.86 for externalizing problems. Multidimensional Self-Concept Scale (AF5, [ 69 ]). This scale considers that self-concept has different, but related, components. This instrument consists of 30 items with a five-point response scale (ranging from 1 “complete disagreement” to 5 “complete agreement”), which is designed to measure five self-concept dimensions: Academic, Social, Emotional, Family, and Physical. Cronbach’s alpha in this study for each dimension was: Academic, 0.82, Social, 0.70, Emotional, 0.77, Family, 0.80, and Physical, 0.78. 2.3. Procedure Adolescents from at-risk families were selected by CWS professionals according to the following inclusion criteria: (1) adolescents between 11–17 years old, without a diagnosed mental health condition; and (2) living in a family being assessed by CWS professionals at a medium level of psychosocial risk. Community adolescents were selected based on a random sample stratified by conglomerates, bearing in mind (in addition to age and sex) the type of school. Fourteen schools were chosen randomly in southern Spain, and the selection of participants was also random between students per academic course of each school. Adolescents from both samples were living in the same neighborhoods. All the subjects participated in the study voluntarily, after signing an informed consent form in accordance with the Declaration of Helsinki. The aims of the research project were explained, and all the participants were assured that their anonymity would be protected. Ethics approval was obtained from the ethics committee of the Andalusian Health Services (code 22/0509). No monetary incentives were offered. 2.4. Data Analyses Statistical analyses were conducted by using IBM SPSS 20 software for Windows [ 70 ]. For the univariate analyses, the descriptive statistics were presented through the means, standard deviations, and the minimum and maximum values (Min–Max) of the quantitative variables, thereby describing their central tendency, variability, and range. For the bivariate analyses, t-tests for two independent samples were computed, using Student’s t-statistic. For the multivariate analyses, a series of cluster analyses were computed in order to describe the joint variability of the sample with respect to behavior problems. Cluster analysis is a multivariate classification technique that identifies groups of individuals defined by similarities on multiple dimensions, so that members of the resulting groups are as similar as possible to others within their group (high within-group homogeneity) and as different as possible to those in the other groups (low between-group homogeneity) [ 33 ]. Prior to clustering, all the selected measures were z-standardized to equate the variables. Although this statistical technique is very robust if assumptions for parametric statistics are violated, following P é rez [ 71 ], both the presence of influential extreme cases and the existence of linearity problems were examined before computing the analysis. Initial groupings were obtained using hierarchical cluster analysis with squared Euclidean distance, Int. J. Environ. Res. Public Health 2019,16, 2767 6 of 17 and the nearest-neighbor method was used as the measure of linkage. The best solution (number of clusters) was determined by examining both the agglomeration schedule and the dendogram. Then, the centroids of these initial clusters were submitted to an iterative clustering procedure (K-means cluster analysis) to refine final cluster membership. Once the cluster dimensions had been determined through confirmatory K-means clustering, the quality of the solution was checked by analyzing atypical subjects, calculating the t-distribution of the sample, comparing standardized residuals for all the variables selected, and checking for outliers. To analyze the differences between clusters, post hoc ANOVAs were computed with the selected psychosocial variables. Finally, multinomial logistic models were performed, taking into consideration deviance distribution to calculate goodness-of-fit, as well as the rate of correct classification of the observed and predicted subjects of the resulting models. Nagelkerke’s pseudo-R 2 statistic was used to assess the resulting models’ degree of explanation. After creating the models and satisfactory confirming its viability, the meaning and direction of the coefficients using the Wald statistic and odds ratios (OR) were examined. 3. Results 3.1. Characteristics of Adolescents: Parenting Styles, Stressful Life Events, Behavior Problems, and Self-Concept The results of the descriptive analyses of the individual, family, and social aspects evaluated in the study are presented in Table 1. It shows that parenting styles were homogeneously distributed, with the Neglectful style being the most frequently reported and Indulgent being the least. On the other hand, adjustment problems obtained a similar average with moderate variability in their scores. In addition, the analyses focusing on the dimension of self-concept reflected that the dimension referring to the family presented the highest average score, but also the lowest academic score. To conclude the descriptive analyses, adolescents had experienced an accumulation of four SLE over the previous five years. Table 1. Descriptions of parenting styles, behavior problems, self-concept, and stressful life events. Percentage Parenting styles Authoritative 28.73% Indulgent 17.15% Authoritarian 22.49% Neglectful 30.73% M SD Minimum Maximum Stressful life events (SLE) Accumulation SLE 4.19 2.52 0 14 Behavior problems Internalizing 0.41 0.22 0 1.26 Externalizing 0.41 0.26 0 1.44 Self-Concept Family 4.37 0.68 1.33 5 Social 4.09 0.65 1.67 5 Emotional 3.53 0.77 1 5 Academic 3.31 0.96 1 5 Physical 3.34 0.79 1.17 5 Int. J. Environ. Res. Public Health 2019,16, 2767 7 of 17 3.2. Two-Step Cluster Analysis: Adjustment Profiles of Adolescents To accomplish the first aim of this study—to identify adolescent profiles according to externalizing and internalizing problems—two-step cluster analysis was computed. The box and whiskers plots were examined for each variable, and the calculation of the Mahalanobis distance did not reveal the existence of either univariate or multivariate extreme cases, respectively. The bivariate correlation indices between the internal and external adjustment problems (r=0.39) did not exceed the value r=0.80, indicating that there were no collinearity problems. After computing a hierarchical agglomerative cluster analysis, and repeating the procedure with different clustering methods (between-groups linkage, furthest neighbor, and Ward’s method), the visual examination of each dendogram revealed the existence of three initial clusters. To confirm this solution, a K-means cluster analysis was performed requesting the definition of three groups. Iteration history showed that convergence (and thus the absence of center changes in each cluster) was achieved in the 10th iteration. Table 2contains the results of the means in each cluster of behavior problems and an analysis of variance between the conglomerates. Furthermore, Figure 1shows the centers of the clusters resulting from the analysis. Table 2. Descriptive and ANOVA between the conglomerates according to internalizing and externalizing problems. AG (49.67%) MEPG (22.05%) MIPG (26.95%) ANOVA M(SD)M(SD)M(SD)F(df) Bonferroni Internalizing problems 0.26 (0.11) 0.45 (0.19) 0.66 (0.17) 285.95 *** (2,440) 1–2 *** 1–3 *** 2–3 *** Externalizing problems 0.24 (0.13) 0.77 (0.19) 0.41 (0.17) 406.09 *** (2,440) 1–2 *** 1–3 *** 2–3 *** *** p<0.001; AG: Adjusted Group, MEPG: Maladjustment with Externalizing Problems Group, MIPG: Maladjustment with Internalizing Problems Group. Int. J. Environ. Res. Public Health 2018, 15, x FOR PEER REVIEW 8 of 17 Figure 1. Conglomerate centers offered in typified scores. To complete the analysis, we computed bivariate tests including the three resulting clusters and some individual, family, and social variables (see Table 3). First, we observed that adolescents were not significantly distributed according to sex (χ2 = 3.41, p = 0.182). As Table 3 shows, adolescents from the AG tended to be community and experience authoritative and indulgent parenting styles, high levels of self-concept and a low accumulation of SLE, and did not perceive authoritarian and neglectful styles. In turn, the MEPG consisted mainly of at-risk adolescents supported by CWS with low self-concept (barring the social one) and a greater accumulation of SLEs. Regarding parenting styles, the adolescents were characterized as experiencing a neglectful parenting style if they did not perceive authoritative and indulgent styles. Lastly, the MIPG cluster was not defined by adolescent background and parenting styles; however, their social, emotional, and physical self-concept tended to be low compared to the other groups, and was characterized by an accumulation of SLEs similar to the MEPG and superior to the AG. Table 3. Descriptive and bivariate analyses including the three resulting clusters and different individual, family, and social variables. AG (49.67%) MEPG (22.05%) MIPG (26.95%) Crosstabs %(ras) %(ras) %(ras) χ2 (df) V Cramer Background Community 30.70% (4.2) 7.67% (−3.8) 12.64% (−1.2) 20.97 *** (2, 441) 0.22 *** Child Welfare Services 19.64% (−4.2) 14.67% (3.8) 14.67% (1.2) Parenting styles Authoritative (yes/no) 17.08%/33.71% (2.1/−2.1) 3.87%/17.99% (−2.8/2.8) 8.20%/19.13% (0.2/−0.2) 8.29 * (2, 437) 0.14 * Indulgent (yes/no) 11.85%/38.95% (3.4/−3.4) 1.59%/20.27% (−2.9/2.9) 3.87%/23.46 (−1.1/1.1) 13.18 ** (2, 437) 0.17 ** Authoritarian (yes/no) 9.11%/41.68% (−2.6/2.6) 6.38%/15.49% (1.6/−1.6) 7.52%/19.82% (1.4/−1.4) 6.60 * (2, 437) 0.12 * Neglectful (yes/no) 12.76%/38.04% (−2.5/2.5) 10.02%/11.85% (3.7/−3.7) 7.74%/19.58% (−0.6/0.6) 13.96 ** (2, 437) 0.18 ** ANOVA M (SD) M (SD) M (SD) F(df) Bonferroni Age 13.40 (1.89) 14.25 (1.84) 13.66 (1.92) 5.41 (2, 440) 1–2 * 1–3 2–3 -1.5 -1 -0.5 0 0.5 1 1.5 AG MEPG MIPG Internalizing behaviours Externalizing behaviours Figure 1. Conglomerate centers offered in typified scores. As Table 2shows, the three groups were formed differentially depending on both the internal and external maladjustment dimensions. The first cluster (AG; Adjusted Group) differed significantly from the other two, indicating lower scores of internalizing and externalizing problems. The second cluster (MEPG; Maladjustment with Externalizing Problems Group) was characterized by medium Int. J. Environ. Res. Public Health 2019,16, 2767 8 of 17 scores in internalizing problems and higher scores in externalizing problems. The third cluster (MIPG; Maladjustment with Internalizing Problems Group) revealed high scores of internalizing problems, and medium scores of externalizing problems (see Figure 1). To complete the analysis, we computed bivariate tests including the three resulting clusters and some individual, family, and social variables (see Table 3). First, we observed that adolescents were not significantly distributed according to sex ( χ2 =3.41, p=0.182). As Table 3shows, adolescents from the AG tended to be community and experience authoritative and indulgent parenting styles, high levels of self-concept and a low accumulation of SLE, and did not perceive authoritarian and neglectful styles. In turn, the MEPG consisted mainly of at-risk adolescents supported by CWS with low self-concept (barring the social one) and a greater accumulation of SLEs. Regarding parenting styles, the adolescents were characterized as experiencing a neglectful parenting style if they did not perceive authoritative and indulgent styles. Lastly, the MIPG cluster was not defined by adolescent background and parenting styles; however, their social, emotional, and physical self-concept tended to be low compared to the other groups, and was characterized by an accumulation of SLEs similar to the MEPG and superior to the AG. 3.3. Parenting Styles and Group Membership Multinomial logistic regression analyses were performed to analyze the potential role of parenting styles in adolescent’s belonging to each cluster. The resulting models estimates the factors associated with the probability that an adolescent from the maladjusted groups will become part of the AG, that is, the one characterized by better indicators in terms of internalizing and externalizing problems. The first important indicator that allows us to ascertain the correspondence of the models with the data is the goodness-of-fit, which is measured through the deviance (Authoritative Model, χ2(46) =43.86, p=0.562 ; Indulgent Model, χ2 (44) =39.17, p=0.678; Authoritarian Model, χ2 (48) =42.74, p=0.688 ; and Neglectful Model, χ2 (46) =38.87, p=0.763), concluding that the models were adequate. Final model likelihood ratio tests ( − 2LL) were statistically different from the initial ones for all models (Authoritative Model, χ2 (4) =43.54, p<0.001; Indulgent Model, χ2 (4) =46.41, p<0.001; Authoritarian Model, χ2(4) =38.10 , p<0.001 ; and Neglectful Model, χ2 (4) =50.64, p<0.001). Thus, we proceeded to interpret the models by examining the OR scores, which allowed us to establish whether each selected variable constituted, on its own, an element that increased the probability of belonging to the reference group. The factors associated with adjustment problems, taking the AG as reference, are presented in Table 4. Specifically, the Authoritative Model explained 10.86% of the variance of the scores, and correctly predicted 53.09% of the subjects belonging to its reference group. As Table 4shows, the analysis of the OR values of the MEPG showed that the presence of the authoritative style increased the probability that an adolescent would become part of this AG by 77%. This relationship was not observed in the MPIG. Likewise, for each increase in an SLE, the probability that an adolescent belong to the MEPG or MIPG increased by 32% and 22%, respectively. The Indulgent Model explained 11.54% of the variance and correctly predicted 52.17% of the subjects belonging to its reference group. As Table 5shows, the analysis of OR values showed that the presence of the indulgent style increased the probability that an adolescent from the MEPG would become part of the AG by 72%, and that an adolescent from the MIPG would become part of the AG by 43%. In turn, for each increase in an SLE, the probability that an adolescent belonged to the MEPG or MIPG, increased by 31% and 21%, respectively. Int. J. Environ. Res. Public Health 2019,16, 2767 9 of 17 Table 3. Descriptive and bivariate analyses including the three resulting clusters and different individual, family, and social variables. AG (49.67%) MEPG (22.05%) MIPG (26.95%) Crosstabs %(ras) %(ras) %(ras)χ2(df) V Cramer Background Community 30.70% (4.2) 7.67% (−3.8) 12.64% (−1.2) 20.97 *** (2, 441) 0.22 *** Child Welfare Services 19.64% (−4.2) 14.67% (3.8) 14.67% (1.2) Parenting styles Authoritative (yes/no) 17.08%/33.71% (2.1/−2.1) 3.87%/17.99% (−2.8/2.8) 8.20%/19.13% (0.2/−0.2) 8.29 * (2, 437) 0.14 * Indulgent (yes/no) 11.85%/38.95% (3.4/−3.4) 1.59%/20.27% (−2.9/2.9) 3.87%/23.46 (−1.1/1.1) 13.18 ** (2, 437) 0.17 ** Authoritarian (yes/no) 9.11%/41.68% (−2.6/2.6) 6.38%/15.49% (1.6/−1.6) 7.52%/19.82% (1.4/−1.4) 6.60 * (2, 437) 0.12 * Neglectful (yes/no) 12.76%/38.04% (−2.5/2.5) 10.02%/11.85% (3.7/−3.7) 7.74%/19.58% (−0.6/0.6) 13.96 ** (2, 437) 0.18 ** ANOVA M (SD) M (SD) M (SD) F(df) Bonferroni Age 13.40 (1.89) 14.25 (1.84) 13.66 (1.92) 5.41 (2, 440) 1–2 * 1–3 2–3 Stressful life events Accumulation SLE 3.52 (2.23) 5.25 (2.50) 4.62 (2.68) 19.75 *** (2, 438) 1–2 *** 1–3 *** 2–3 Self-Concept Family 4.58 (0.46) 3.99 (0.84) 4.34 (0.70) 30.57 *** (2, 438) 1–2 *** 1–3 ** 2–3 *** Social 4.16 (0.55) 4.19 (0.73) 3.86 (0.72) 10.66 *** (2, 439) 1–2 1–3 *** 2–3 *** Emotional 3.79 (0.62) 3.47 (0.76) 3.07 (0.80) 40.51 *** (2, 434) 1–2 ** 1–3 *** 2–3 *** Academic 3.58 (0.84) 2.61 (0.99) 3.38 (0.88) 39.48 *** (2, 431) 1–2 *** 1–3 2–3 *** Physical 3.45 (0.74) 3.31 (0.87) 3.17 (0.77) 5.12 ** (2, 438) 1–2 *** 1–3 * 2–3 r as =adjusted standardized residuals; *p<0.05, **p<0.01, ***p<0.001; AG: Adjusted Group, MEPG: Maladjustment with Externalizing Problems Group, MIPG: Maladjustment with Internalizing Problem Group. The Authoritarian Model explained 9.56% of the variance and correctly predicted 51.49% of the subjects belonging to its reference group. As Table 6shows, the analysis of the OR values showed that the presence of the authoritarian style increased the probability that an adolescent from the AG would become part of the MEPG or MIPG by 58% and 53%, respectively, although the p-values did not report statistical significance. In addition, for each increase in SLE, the MEPG and MIPG increased by 31% and 21%, respectively. Int. J. Environ. Res. Public Health 2019,16, 2767 16 of 17 42. Barber, B.K.; Chadwick, B.A.; Oerter, R. Parental Behaviors and Adolescent Self-concept in the United States and Germany. J. Marriage Fam. 1992,54, 128–141. [CrossRef] 43. Marcone, R.; Affuso, G.; Borrone, A. Parenting styles and children’s internalizing-externalizing behavior: The mediating role of behavioral regulation. Curr. Psychol. 2017, 1–12. [CrossRef] 44. Kritzas, N.; Grobler, A.A. The relationship between perceived parenting styles and resilience during adolescence. J. Child. Adolesc. Ment. Health 2005,17, 1–12. [CrossRef] [PubMed] 45. Barton, A.L.; Kirtley, M.S. Gender differences in the relationships among parenting styles and college student mental health. J. Am. Coll. Health 2012,60, 21–26. [CrossRef] [PubMed] 46. Gracia, E.; Fuentes, M.C.; Garcia, F.; Lila, M. Perceived neighborhood violence, parenting styles, and developmental outcomes among Spanish adolescents. J. Commun. Psychol. 2012 ,40, 1004–1021. [CrossRef] 47. Rose, J.; Roman, N.; Mwaba, K.; Ismail, K. The relationship between parenting and internalizing behaviours of children: A systematic review. Early Child. Dev. Care 2018,188, 1468–1486. [CrossRef] 48. Yap, M.B.; Pilkington, P.D.; Ryan, S.M.; Jorm, A.F. Parental factors associated with depression and anxiety in young people: A systematic review and meta-analysis. J. Affect. Disord. 2014 ,156, 8–23. [CrossRef] [PubMed] 49. Betts, J.; Gullone, E.; Allen, S. An examination of emotion regulation, temperament, and parenting style as potential predictors of adolescent depression risk status: A correlational study. Br. J. Dev. Psychol. 2009 , 27, 473–485. [CrossRef] [PubMed] 50. Lorence, B.; Hidalgo, M.V.; Dekovic, M. Adolescent adjustment in at-risk families: The role of parental socialization and psychosocial stress. Salud Ment. 2013,36, 49–57. [CrossRef] 51. Bornstein, M.H. Form and function: Implications for studies of culture and human development. Cult. Psychol. 1995,1, 123–137. [CrossRef] 52. Chao, R. Extending research on the consequences of parenting styles for Chinese Americans and European Americans. Child. Dev. 2001,72, 1832–1843. [CrossRef] [PubMed] 53. Deater-Deckard, K.; Lansford, J.E.; Malone, P.S.; Alampay, L.P.; Sorbring, E.; Bacchini, D.; Bombi, A.S.; Bornstein, M.H.; Chang, L.; Di Giunta, L.; et al. The association between parental warmth and control in thirteen cultural groups. J. Fam. Psychol. 2011,25, 790–794. [CrossRef] [PubMed] 54. Dwairy, M.; Achoui, M. Adolescents-family connectedness: A first cross-cultural research on parenting and psychological adjustment of children. J. Child. Fam. Stud. 2010,19, 8–15. [CrossRef] 55. Lansford, J.E.; Deater-Deckard, K.; Dodge, K.A.; Bates, J.E.; Pettit, G.S. Ethnic differences in the link between discipline and later adolescent externalizing behaviors. J. Child. Psychol. Psychiatry 2004 ,45, 801–812. [CrossRef] [PubMed] 56. Garaigordobil, M.; Aliri, J. Parental socialization styles, parents’ educational level and sexist attitudes in adolescence. Span. J. Psychol. 2012,15, 592–603. [CrossRef] [PubMed] 57. DiMaggio, R.; Zappulla, C. Mothering, fathering, and Italian adolescents’ problem behaviors and life satisfaction: Dimensional and typological approach. J. Child. Fam. Stud. 2013,23, 567–580. [CrossRef] 58. Garc í a, F.; Gracia, E. Is always authoritative the optimum parenting style? Evidence from Spanish families. Adolesc 2009,44, 101–131. [CrossRef] 59. Garc í a, F.; Gracia, E. The indulgent parenting style and developmental outcomes in South European and Latin American Countries. In Parenting Across Cultures: Childrearing, Motherhood and Fatherhood in Non-Western Cultures; Selin, H., Ed.; Springer: Dordrecht, The Netherlands, 2014; Volume 7, pp. 419–434. ISBN 9789400775022. 60. Lorence, B.; Hidalgo, M.V.; Men é ndez, S. Parenting style and adolescent adjustment in contexts at psychosocial risk: Evidence from Spanish families. In Parenting: Cultural Influences and Impact on Childhood Health and Well-Being; Garc í a, F., Ed.; NOVA publishers: New York, NY, USA, 2015; pp. 77–92. ISBN 978-1-63482-493-4. 61. Garcia, O.C.; Serra, E. Raising children with poor school performance: Parenting styles and short- and long-term consequences for adolescent and adult development. Int. J. Environ. Res. Public Health 2019 , 16, 1089. [CrossRef] [PubMed] 62. Musitu, G.; Garc í a, J.F. Consequences of families socialization in the Spanish culture. Psychol. Spain 2005 , 9, 34–40. 63. Aucoin, K.J.; Frick, P.J.; Bodin, S.D. Corporal punishment and child adjustment. J. Appl. Dev. Psychol. 2006 , 27, 527–541. [CrossRef] Int. J. Environ. Res. Public Health 2019,16, 2767 17 of 17 64. Nunes, C.; Bodden, D.; Lemos, I.; Lorence, B.; Jim é nez, L. Parenting practices and quality of life in Dutch and Portuguese adolescents: A cross-cultural study. Rev. Psicodidácica 2014,19, 327–346. [CrossRef] 65. Leventhal, T.; Brooks-Gunn, J. The neighborhoods they live in: The effects of neighborhood residence on child and adult outcomes. Psychol. Bull. 2000,126, 309–336. [CrossRef] [PubMed] 66. Schonberg, M.A.; Shaw, D.S. Do the pre-dictors of child conduct problems vary by high- and low-levels of socioeconomic and neighborhood risk? Clin. Child. Fam. Psychol. 2007,10, 101–136. [CrossRef] [PubMed] 67. Gracia, E.; Fuentes, M.C.; Garc í a, F. Neighborhood risk, parental socialization styles and adolescent conduct problems. Psychosoc. Interv. 2010,19, 265–278. [CrossRef] 68. Oliva, A.; Jim é nez, J.; Parra, A.; S á nchez-Queija, I. Acontecimientos vitales estresantes, resiliencia y ajuste adolescente [Stressful life events, resilience and adolescent adjustment]. Revista de Psicopatolog í a y Psicología Clínica 2008,13, 53–62. [CrossRef] 69. Garc í a, F.; Musitu, G. AF5: Autoconcepto Forma 5 [AF5: Self-concept form 5, 3rd ed.; TEA: Madrid, Spain, 2009; ISBN 8415262981. 70. IBM Corp. Released. IBM SPSS Statistics for Windows 20.0 Version; IBM Corp.: Armonk, NY, USA, 2011. 71. P é rez, C. T é cnicas de An á lisis Multivariante de Datos. Aplicaciones con SPSS Multivariate Data Analysis Techniques. Applications with SPSS; Pearson-Prentice Hall: Madrid, Spain, 2004; ISBN 9788420541044. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).