Introduction

Research over the past years has clearly demonstrated that a positive family environment is beneficial to child development. Rohner’s parental acceptance–rejection theory (PARTheory; Rohner 2004), recently renamed interpersonal acceptance–rejection theory (IPARTheory; Ali et al. 2015), assumes that high perceived parental acceptance and low perceived parental rejection are associated with positive and negative developmental outcomes respectively; negative developmental outcomes may include depression, externalizing problems and school failure (Hoeve et al. 2011; Khaleque and Rohner 2012; Putnick et al. 2014). Most parenting research has been conducted with children and mothers, or has failed to differentiate maternal and paternal roles; nevertheless, there is an increasing evidence that fathering (Parke 2002; Rohner and Veneziano 2001; Veneziano 2003) makes a unique, significant contribution to children’s adjustment and that the effects of negative parenting effect persist into adolescence (Carrasco and Rohner 2013; Dwairy 2010; Kim et al. 2006). Although there is evidence that mothers and fathers have different parenting attitudes and styles (Gamble et al. 2007; Steinberg and Skill 2002; Tacon and Caldera 2001; Winsler et al. 2005), the combined effects of parents’ styles on adjustment of adolescents has not been fully described.

Even though maternal and paternal styles are often highly related, what happens if, for instance, one parent is rejecting and the other is not? Most research comparing the effects of parental consistency and inconsistency has followed Maccoby’s approach (Maccoby and Martin 1983) which, intersecting parental responsiveness and demandingness, allows to classify parenting styles as authoritative, permissive, uninvolved, and authoritarian; so, some studies have investigated the effects on child development to have, for instance, a father authoritarian whereas the mother is permissive (Braza et al. 2013; McKinney and Renk 2008). We are not aware of any study which examined the effects of parental inconsistency from the perspective of IPARTheory and this study was intended to address this gap in the research. IPARTheory considers parental acceptance–rejection as a continuum. The acceptance extremity represents parents who express high levels of love and warmth; the rejection extremity represents parents who display hostility, aversion, disapproval and undifferentiated rejection and neglect their children or use of severe or abusive disciplinary strategies (Rohner 1986, 2004). Although there is cultural variation in parental acceptance norms (Dwairy 2010; Wu and Chao 2005), meta-analyses which have included studies carried out in several cultures have provided evidence that children’s perceptions of parental (both maternal and paternal) acceptance are associated with a specific form of psychological adjustment in children and adults, regardless of differences in gender, race, geography, language, or culture (Khaleque and Rohner 2012).

Parental rejection is a developmental risk factor and leads children to evaluate themselves and their future negatively; these negative evaluations may make them vulnerable to depression and or facilitate acquisition of socially unacceptable behaviors, such as externalizing behavior (Kim et al. 2003; Nolan et al. 2003). Parental acceptance is a protective factor which promotes trust and reciprocity between parent and child (Darling and Steinberg 1993; Maccoby and Martin 1983; Steinberg 2001). Parental warmth promotes the internalization of family rules and encourages children to adopt their parents’ values and behaviors (Grolnick and Farkas 2002).

Although most research on parenting has been mother-focused, there have been frequent calls for the systematic inclusion of fathers (Bornstein and Venuti 2013), for investigation of the direct or indirect influence of a paternal figure on child development (Lila et al. 2007; Pace et al. 2012) and for research into how adolescents’ adjustment is affected by the combined influence of maternal and paternal behavior (Di Maggio and Zappulla 2014). Recent cross-cultural studies (Deater-Deckard et al. 2011; Putnick et al. 2012) comparing nine different cultures showed that mothers and fathers self-reported different views of their own acceptance and rejection of their children; in general parental self-reports suggest that mothers behave more warmly than fathers, but this difference is very small. Consistent with these results, adolescents report that their mother is more accepting than their father (Dwairy 2010; Tulviste 2012).

Research into the effects of parental inconsistency in parenting style has become a research topic in its own right (Dwairy 2008; Lengua 2006): parental inconsistency may occur in two different forms (Dwairy 2010): intra-parental inconsistency—which concerns to the individual parents instability and unpredictability in their interactions with children, across time and situations—and inter-parental inconsistency, which refers to the paternal and maternal dissimilarity in parenting styles, which may be confusing for children and have a damaging effect on learning and socialization (Wenar and Kerig 2000); however the empirical basis for the claim that inter-parental inconsistency is detrimental to adolescents is still far weaker than might be assumed. To date research on parental inconsistency has consisted mainly of studies of intra- rather than inter-parental inconsistency (Fletcher et al. 1999). Few studies have investigated adolescents’ perceptions of the consistency of their mothers’ and fathers’ parenting styles and the results have been mixed. Some studies concluded that regardless of the predominant style, inconsistencies are harmful (Berkien et al. 2012; Garcia-Linares et al. 2014; Lengua and Kovacs 2005; Tildesley and Andrews 2008) whereas others found evidence that positive outcomes were linked to inconsistent parenting, suggesting that one parent can buffer or complement the effects of the other parent on adolescents’ development (Fletcher et al. 1999; McKinney and Renk 2008). Most of these studies examined parenting characteristics in terms of parenting style (Braza et al. 2013; Dwairy 2008; McKinney and Renk 2008), and only few studies (Berkien et al. 2012; Jaursch et al. 2009) examined the effects of perceived parental dissimilarity in emotional warmth or rejection; these studies suggested that inter-parental inconsistency in warmth was associated with development of internalizing problems.

Moreover, studies suggest that many variables, for example adolescent gender and age, might influence parent–adolescent relationship and moderate the association between parenting and adolescent adjustment (Holmbeck et al. 1995). Studies based on IPARTheory suggest that parental behavior changes during development; it is common for adolescents to reduce the intensity of their relationships with parents and to invest more time in less hierarchical relationships such as sibling relationships and friendships and so it is plausible that the parental influence on children’s adjustment declines as children mature (Ahmed et al. 2012; Khaleque and Rohner 2002; Steinberg and Silk 2002). Furthermore, research suggests that there are gender differences in the response to parental rejection: boys are more likely to develop externalizing behaviors, whereas girls are more likely to have internalizing problems (Hankin and Abramson 2001; Galambos et al. 2003; Tulviste 2012). As regards the parental inconsistency, research suggested that it has less impact on boys (Dwairy 2009). Many studies have also reported that, for children, divorce is positively associated with externalizing behaviors and internalizing problems and problems in social relationships and negatively associated with academic achievement (Lansford 2009). Nevertheless studies which used Rohner’s construct found no significant differences between adolescents from single-parent and two-parent families with regard to perceived parental acceptance and adjustment, depressive affect and alcohol abuse (Tulviste 2012). From this, and because of the increase in the divorce rate worldwide over the last two decades (Eurostat 2014; United Nations Statistic Division 2008), the importance to investigate whether family structure moderates the effects of rejection and inconsistent maternal-and-paternal parenting is warranted. Finally, a large body of evidence suggests that low SES has a detrimental effect on parenting style and that children belonging to families with higher SES are more likely to have access to the resources needed for development than children from lower SES families (Bornstein et al. 2003; Dodge et al. 2008; Hoff et al. 2002).

In light of this earlier research, the aim of our study was to investigate the relationship between adolescent maladjustment (anxiety–depression symptoms and aggressive behavior) and maternal and paternal acceptance–rejection, considering the direct impact of inter-parental inconsistency and the specific ways in which mother and father acceptance–rejection combine each other. Specifically, we first tested whether the index of dissimilarity between mother and father AR was an independent predictor of adolescent maladjustment, in addition to single parents’ measures of AR. Secondly, we explored the ways in which maternal and paternal acceptance–rejection combine each other to examine if the risk linked to parental rejection was buffered by only one accepting parent and increased when all two parents were rejecting. As child-reported maternal and paternal acceptance–rejection are generally highly correlated (Putnick et al. 2014), we predicted that most parents will present consistent profiles (i.e. both parents display either high or low rejection), and that few parents will show inconsistent profiles (father’s or mother’s rejection inversely perceived by their offspring). We also wanted to test whether gender, adolescent age, family structure or socioeconomic status (SES), which we considered potential influences on parent–adolescent relationships, moderated the association between parenting and adolescent adjustment (Holmbeck et al. 1995).

Method

Participants

The original sample comprised 2843 adolescents (1440 boys and 1403 girls) aged between 10 and 16 years who were attending schools located in the urban area of Naples in southern Italy. Forty adolescents refused to participate, ninety-eight questionnaires were not completed and eighty-one participants who had at least one deceased parent were excluded from the analysis. The final sample thus included consisted of 2624 adolescents (1316 boys and 1308 girls), drawn from 61 classes of seven secondary schools (1206 early adolescents, M age = 11.82 years, range 10–13, SD = 1.06) and 70 classes of seven high schools (1418 middle adolescents, M age = 15.61 years, range 14–16, SD = 1). The social composition of the sample was broadly representative of the city as a whole.

Procedure

In accordance with Italian law and the ethical principles of the Italian Association of Psychology (AIP), we obtained permission to contact families in schools from university Institutional Review Board (IRB). After the school superintendent, school principals and teachers had granted their approval researchers described the study to adolescents in their classes, and provided them with a sealed envelope to take home to their parents. This envelope contained a letter explaining the purpose of the study, what it would involve and a consent form. After parents had provided written consent, the questionnaires were administered in the classroom, during ordinary class sessions. The questionnaires took about 30 min to complete. It was emphasized to the adolescents that their data would be anonymous and treated confidentially and that their participation was voluntary. All data were collected in Spring 2013.

Measures

Adolescent Maladjustment

Emotional and behavioral problems were measured using the Youth Self Report (YSR; Achenbach 1991) for ages 4–18 years. Adolescents reported whether they had experienced symptoms of anxiety or depression (e.g. nervousness, feeling highly strung, feeling tense) and behaved aggressively (e.g. bragging, boasting) in the past 6 months using a three-point scale ranging (0 = not true, 1 = somewhat or sometimes true, or 2 = very true or often true.). Internal consistency was good (α = .83 for anxiety–depression symptoms; α = .87 for aggressive behavior).

Parental Acceptance–Rejection

Perceived parental acceptance and rejection was measured with the Parental Acceptance–Rejection Questionnaire—child version (PARQ) (PARQ_child, Rohner and Khaleque 2005). Participants responded to two versions of the PARQ to provide their assessments of maternal (PARQ Mother) and paternal (PARQ Father) acceptance–rejection. The PARQ is a 24-item self-report instrument used to assess respondents’ perceptions of maternal or paternal warmth, affection, hostility, aggression, indifference, neglect and undifferentiated rejection (example items ‘My [mother/father] makes me feel wanted and needed’; ‘My [mother/father] goes out of [her/his] way to hurt my feelings’). We analyzed overall scores for perceived acceptance–rejection; high scores indicate rejection and low scores indicate acceptance. Participants indicated how well each statement described the parent’s behavior using a four-point Likert scale (4 = almost always true to 1 = almost never true). Both maternal and paternal versions had high reliability (α = .90 and α = .89 respectively).

Demographic Variables

All demographic data were obtained directly from the adolescent participants. The students reported their age, the highest level of education reached by their parents, parental occupational prestige and the family structure (intact family; non-intact family). A composite SES variable was created from reported parental educational attainment (from 1 = finished only some primary classes or did not go to school to 5 = finished university or higher) and occupational prestige (for current job; from 1 = has never worked outside the home for pay to 10 = professional). We calculated the SES index by applying principal component analysis (PCA) to the four SES constituent items: mother’s education, father’s education, mother’s occupational status, and father’s occupational status. PCA model is appropriate for formative measurement because the direction of causality is from the variables to the SES index and not vice versa (Caro and Cortés 2012; Diamantopoulos et al. 2008).

Adolescents were asked if they lived with both parents (intact family) or with only one parent (non-intact family); 217 (8.26 %) participants reported that they lived in a non-intact family.

Data Analysis

A repeated measures ANOVA with parental gender as the within-subjects factor and adolescent gender and stage of adolescence as the between-subjects factors was carried out to compare perceived mother and father AR. Intraclass correlation coefficients (ICCs) (Kenny et al. 2006) related to mother and father scores were used to obtain a continuous score of parental inconsistency in order to test, through a regression analysis, the direct effect and independent effect of inconsistency on adolescent maladjustment. The ICC ranged from 1 to −1: an intraclass correlation of 1 indicates perfect similarity, that is a perfect match between the two variables across all observations. Ward’s Hierarchical Cluster Analysis Method (Gordon 1996) was used to build family typologies based on differences in parental AR, so that clusters were produced by maximizing intragroup similarities and intergroup differences. A subsequent multivariate analysis of covariance (MANCOVA) was performed in order to compare different family typologies in relation to adolescent maladjustment considering the possible confounding effects of gender, age, SES, and family structure.

Results

Preliminary Analyses

Direct effect of acceptance rejection and inter-parental inconsistency on adolescent’s maladjustment.

The ANOVA repeated measure revealed main effects of parental gender, F(1, 2620) = 36.82, p < .001; adolescent gender, F(1, 2620) = 17.31, p < .001 and stage of adolescence, F(1, 2620) = 40.63, p < .001. Comparison of mean AR scores showed that fathers were more rejecting than mothers (fathers: M = 1.57, SE = .01; mothers: M = 1.52, SE = .01). Adolescent boys reported experiencing more parental rejection than girls (boys: M = 1.58, SE = .01; girls: M = 1.51, SE = .01), and that middle adolescents reported more parental rejection than early adolescents (middle adolescents: M = 1.59, SE = .01; early adolescents: M = 1.49, SE = .01). There were no interaction effects. The zero-order coefficient of correlation between maternal AR and paternal AR was significant and positive (r = .61, p < .001).

After computing the ICCs, a Pearson’s correlation and hierarchical regression models were conducted. Correlation analysis showed significant negative correlations between the inter-parental inconsistency and adolescent maladjustment (anxiety–depression symptoms: r = −.22, p < .001, aggressive behavior: r = −.25, p < .001). Two hierarchical multiple regression analyses were then carried out. Perceived maternal and paternal acceptance–rejection were entered at Step 1 and inconsistent parenting ICC at Step 2. Hierarchical multiple regressions showed that all predictors made independent contributions to adolescent maladjustment. Respectively, perceived maternal and paternal acceptance–rejection explained 9 % of the observed variance, with inconsistent parenting ICC contributing an additional 1 % (Table 1) of anxiety–depression symptoms; perceived maternal and paternal acceptance–rejection explained 10 % of the observed variance, with inconsistent parenting ICC contributing an additional 1 % of aggressive behavior (Table 1).

Table 1 Results of regression analyses of perceived maternal and paternal acceptance–rejection and Inconsistent parenting ICC, associated with anxiety–depression symptoms and aggressive behavior
Full size table

Adolescent Maladjustment and Different Parenting Clusters

Six clusters representing family typologies were then identified through visualization techniques (dendrogram, agglomeration scheme, euclidian distances, number of cases per cluster) of Hierarchical Cluster Analysis, This solution produced an increase of .14 in the coefficient of agglomeration and represented a good trade-off between parsimony (a minimum number of groups) and group homogeneity. Figure 1 shows standardized scores (Z-scores) for maternal and paternal AR for all the clusters.

Fig. 1
figure 1

Standardized perceived parenting mean scores for mothers and fathers on the parental acceptance–rejection (high scores mean high rejection). Note Cluster 1 = consistent parenting in high rejection, Cluster 2 = inconsistent parenting with rejecting mother, Cluster 3 = inconsistent parenting with rejecting father, Cluster 4 = consistent parenting in moderate rejection, Cluster 5 = consistent parenting in moderate acceptance, Cluster 6 = consistent parenting in high acceptance

Full size image

The six clusters were as follows, C1: consistent high rejection, both mother and father were rated highly rejecting (n = 146; 5.57 %); C2: inconsistent parenting with maternal rejection, mother was rated highly rejecting and father rated no more than moderately rejecting (n = 80; 3.05 %); C3: inconsistent parenting with paternal rejection, father highly rejecting and mother rated no more than moderately rejecting (n = 40; 1.52 %); C4: consistent moderate rejection, both mother and father were rated moderately rejecting (n = 570; 21.72 %), C5: consistent moderate acceptance, both mother and father were rated moderately accepting (n = 1025; 39.06 %) and C6: consistent high acceptance, both mother and father were rated highly accepting (n = 763–29.08 %).

Finally, a 6 × 2 × 2 × 2 multivariate analysis of covariance (six parenting clusters × males vs. females × early-adolescents vs. middle-adolescents × intact vs. non-intact families) was conducted with anxiety–depression symptoms and aggressive behavior as dependent variables and SES as a covariate to test the hypothesis that consistent parental acceptance was associated with better adjustment than inconsistent parental rejection or consistent parental rejection. Data on AR, anxiety–depression symptoms and aggressive behavior were standardized as Z-scores for this analysis.

There was a main effect of gender, Wilks’s λ = .99, F(2, 2574) = 5.60, p < .01; stage of adolescence, Wilks’s λ = .99, F(2, 2574) = 4.85, p < .01; family structure, Wilks’s λ = .99, F(2, 2574) = 12.97, p < .001 and parenting cluster, Wilks’s λ = .94, F(10, 5148) = 17.13, p < .001. There was no effect associated with SES as a covariate, Wilks’s λ = .98, F(2, 2574) = 3.05, p > .05. Exploration of the univariate effects (Table 2) showed that boys reported more aggressive behavior than girls and early adolescents reported more anxiety–depression symptoms than middle adolescents. More specifically an interaction between adolescent gender and stage of adolescence, Wilks’s λ = .99, F(2, 2574) = 6.11, p < .01, showed that anxiety–depression symptoms (F(1,2575) = 10.78, p < .01) were stable across early and middle adolescence in girls (early: M = .54, SE = .11; middle: M = .56, SE = .08) and decreased during adolescence in boys (early: M = .75, SE = .12; middle: M = .10, SE = .09). Adolescents with separated or divorced parents reported more aggressive behavior and more anxiety–depression than adolescents living with cohabiting parents.

Table 2 Mean and standard errors for measures of anxiety–depression symptoms and aggressive behavior by gender, stage of adolescence, and family structure
Full size table

Means and standard errors for anxiety–depression symptoms and aggressive behavior in the parenting clusters are reported in Fig. 2. There was a main effect of parenting cluster on anxiety–depression symptoms, F(5, 2575) = 25.84, p < .001, and post hoc Bonferroni tests showed that adolescents who experienced consistent rejection (C1), inconsistent parenting with maternal rejection (C2) or inconsistent parenting with paternal rejection (C3) reported more anxiety–depression symptoms that adolescent who experienced consistent moderate rejection (C4). Adolescents who experienced consistent moderate acceptance (C5) reported an intermediate level of anxiety–depression symptoms whilst those who experienced consistent acceptance (C6) reported a low level of anxiety–depression symptoms. There was also a main effect of parenting cluster on aggressive behavior, F(5, 2575) = 21.31, p < .001, and post hoc Bonferroni tests showed that C1 adolescents reported most aggressive behavior, C2, C3 and C4 adolescents reported medium levels of aggressive behavior and C5 and C6 adolescents reported the least aggressive behavior. There was an interaction between parenting cluster and family structure with respect to anxiety–depression symptoms, Wilks’s λ = .99, F(10, 5148) = 3.09, p < .001, F(5, 2575) = 4.16, p < .001. Adolescents with divorced or separated parents reported more anxiety–depression symptoms than adolescents with cohabiting parents (post hoc Bonferroni tests, C1: p < .001; C2: p < .05; C3: p < .05). Adolescents in C4, C5 and C6 reported similar levels of anxiety–depression symptoms regardless of family structure (Fig. 3).

Fig. 2
figure 2

Means and standard error of anxiety–depression symptoms and aggressive behavior by different parenting clusters. Note estimated marginal means, covariate: SES. Adjustment for multiple comparisons: Bonferroni, a > b>c > d. Cluster 1 = consistent parenting in high rejection. Cluster 2 = inconsistent parenting with rejecting mother. Cluster 3 = inconsistent parenting with rejecting father. Cluster 4 = consistent parenting in moderate rejection. Cluster 5 = consistent parenting in moderate acceptance. Cluster 6 = consistent parenting in high acceptance

Full size image
Fig. 3
figure 3

Means and standard error of anxiety–depression symptoms by different parenting clusters and family structure. Note estimated marginal means, covariate: SES. Cluster 1 = consistent parenting in high rejection, Cluster 2 = inconsistent parenting with rejecting mother, Cluster 3 = inconsistent parenting with rejecting father, Cluster 4 = consistent parenting in moderate rejection, Cluster 5 = consistent parenting in moderate acceptance, Cluster 6 = consistent parenting in high acceptance

Full size image

Discussion

We used IPARTheory to structure an investigation of the relationship between adolescent maladjustment (measured as anxiety–depression symptoms and aggressive behavior) and perceived maternal and paternal acceptance–rejection in a sample of over 2500 Italian adolescents. Our main objectives were (1) to determine the direct and independent impact of inter-parental inconsistency and (2) to test whether the way in which mother and father acceptance–rejection combined each other was related to adolescent maladjustment. We also investigated whether this relation was moderated by adolescent gender, stage of adolescence, family structure or SES.

In summary our results showed that: (1) mothers were perceived as less rejecting than fathers although the difference was statistically small; (2) adolescents’ ratings of maternal and paternal acceptance–rejection were highly correlated; (3) both paternal and maternal rejection were related to adolescent aggression and anxiety–depression symptoms; (4) inter-parental inconsistency had a negative effect on adolescent adjustment, over and above the effects of maternal and parental rejection; (5) the family structure (intact vs. non-intact) moderated the relationship between parental acceptance–rejection typology and anxiety–depression symptoms.

As regards the first objective, consistent with previous research, our results showed an independent effect of inter-parental inconsistency (Dwairy 2010), in addition to paternal and maternal acceptance–rejection (Rohner 2004; Rohner and Britner 2002). According to literature, parental rejection increases the adolescent adjustment involving several different mechanisms: since adolescents tend to base their self-image on how they think they are perceived by their parents or significant others, perceived parental rejection often results in adolescents developing negative thoughts and feelings about themselves, their future, the world and their relationships (Rohner 1986, 1999), which in turn make adolescents vulnerable to depression (Burkhouse et al. 2012; Schwartz et al. 2012) and also increase the probability that they will develop socially unacceptable behaviors, such as externalizing behaviors, both through acquiring their parents’ aggressive behavioral style (Bandura 1977; Lansford 2014) and through acquiring the attributional style and cognitive processes that typically mediate aggressive behavior (Bacchini et al. 2011; Crick and Dodge 1994). Moreover, results of previous studies support the association between inter-parental inconsistency and psychological maladjustment; specifically, when adolescents experience different treatments by different agents, such as father and mother, most probably they will develop a feeling of injustice which will interfere with the processes of attachment and identification, and cause ambivalence towards the self (Dwairy 2010).

With respect to the second objective of our study, we considered the ways in which maternal and paternal acceptance–rejection combine each other, examining if the risk linked to parental rejection was buffered by only one acceptance parent and increased when all two parents were rejecting. We used cluster analysis to uncover patterns in combinations of maternal and paternal acceptance–rejection and were able to define six distinct parental acceptance–rejection clusters. In four clusters the father and mother were perceived similarly (both highly rejecting, moderately rejecting, average rejecting and accepting, and highly accepting) and in two cluster the father and mother were perceived differently (one parent as perceived as highly rejecting whereas the other was perceived as average rejecting).

In the light of previous finding it is notable that we found that a substantial minority of parental couples (approximately 4 %) differed with respect to acceptance–rejection. This may be because previous studies (see Braza et al. 2013; Dwairy 2008; McKinney and Renk 2008) of inter-parental inconsistency investigated other aspects of parenting style (e.g. authoritative vs. authoritarian parenting). Parenting styles and cultural factors amplify differences between the paternal and maternal role. For instance, in most cultures the social norm is for mothers to be more permissive and warmer towards their children whilst fathers are stricter and authoritarian (Bornstein 2012; Conrade and Ho 2001). Consistent with our result, recent cross-cultural studies have reported correlations between perceived maternal and paternal acceptance which range from .74 to .80 (Putnick et al. 2014). It is not surprising that adolescents perceive that their mother and father behave similarly as the members of a couple tend to choose each other on the basis of similarities, a process referred to as assortative mating (Simons and Conger 2007). A second possible explanation is that studies which reported a higher frequency of inter-parental inconsistency used different analytical strategies, based on the dividing samples on the basis of the percentile distribution, maximizing in this way the belonging to one group or another (Garcia and Gracia 2009). What implications are there for child development when there are big differences in perceptions of maternal and paternal behavior? Does the risk represented by a rejecting parent prevail over the protective influence of the accepting parent?

Our results represents a step forward in understanding the potential buffering effect of having at least one parent who is moderately accepting. We found few meaningful differences in the level of self-reported maladjustment (high) between adolescents who reported one rejecting parent and those who reported that both parents were rejecting. Adolescents who reported consistent moderate rejection or consistent moderate acceptance reported a medium level of maladjustment, whilst adolescents who reported consistent high acceptance reported less maladjustment. These findings highlight the harmful effects of having even one parent who is perceived as rejecting, regardless of the gender of that parent; however, it is also important to note that in our results, high rejecting parents matched up only with moderately accepting ones. Thus, we cannot assume that one-parent acceptance is not enough to buffer the psychological risk represented by the rejecting spouse, but future studies, using more sensitive instruments or different analytic strategies, can address this gap, examining the circumstances and the acceptance degree which enable a parent to scratch the negative effects of the rejecting parent on adolescent maladjustment. In the area of child abuse, for example, some authors suggest that mothers’ fail to protect their children from the abuse through a merely passive collusion with the abusing parent, but this is an under-researched problem yet (Ford 2006). However, we cannot exclude the possibility that dissimilarity between paternal and maternal style might also have a negative effect on learning and socialization processes and thence on general adjustment in adolescence (Wenar and Kerig 2000). Moreover, our results underline the importance of considering the influence of fathers and mothers separately, wherever possible.

The last research question was whether gender, age, SES or family structure moderated the relation between parental acceptance–rejection cluster and internalizing and externalizing problems. We detected only one interaction, which involved family structure. We found that when the parents were dissimilar with respect to acceptance–rejection or were both highly rejecting, adolescents in non-intact families reported more anxiety–depression symptoms than those in intact families. In other words, the negative effects of living in a non-intact family were amplified if both parents were perceived as rejecting or if at least one parent was perceived as highly rejecting. This finding suggests a new way of considering the effects of divorce and separation. If we were to consider only the main effect, we might erroneously attribute the negative effects of family structure to parental separation; however when we take into account the interaction between family structure and parental inconsistency in acceptance–rejection we see that detrimental effects of non-intact family are magnified in families where at least one parent is rejecting, but when separated parents are perceived as accepting or mildly rejecting their children are at no greater risk of maladjustment than their peers belonging to intact families. Parent–child conflict may increase following divorce or separation and family cohesion decreases (Lansford 2009; Short 2002); this results in the children being exposed to more interpersonal stress and leads to an increase in symptoms of anxiety and depression.

In summary, the findings reported here buttress a large and growing body of evidence that consistent moderately and or highly accepting parenting is the most beneficial way of parenting adolescents and that parental rejection is a risk factor for adolescent adjustment, especially in families where parents are divorced or separated. The optimal home environment for adolescent development is therefore a family in which two accepting parents participate actively in their adolescents’ lives; in the minority of homes where there is inter-parental inconsistency in acceptance, adolescents do not benefit from having one moderately accepting parent. These results have a wide range of implications for parenting practices and suggest that all professionals who provide guidance to parents or teachers should emphasize the importance of acceptance and consistency to creating a good developmental environment. Special attention should be paid to adolescents who live in non-intact families, because inter-parental inconsistency poses a greater risk to psychosocial development in this group.

Some important limitations of this study should be noted when considering the findings. First, the cross-sectional design of the study does not permit causal inferences. Longitudinal studies suggest that the association between adolescents’ behavioral problems and parental rejection is bidirectional (Cohen and Brook 1995; Lila et al. 2007), but our data set does not enable us to establish whether rejection is a cause or consequence of behavioral problems. Following the model of youth-and-parent driven processes in parenting practices, longitudinal data or transactional models could be used in future works to examine which parenting and adolescent characteristics are mutually influential (e.g. see Maccoby 1992; Reiss and Price 1996; Rothbart and Bates 1998). In other words, the problem behaviors of adolescents may elicit parental rejection. For example, Kerr and Stattin (2003) demonstrated that parents’ behaviors were reactions to problem behaviors exhibited by their child.

Another limitation is related to the use of self-report measures, and, although adolescent reports of parenting style (particularly acceptance–rejection) are often a more useful measure than self-reports of parenting style (Greco and Morris 2002; Hale et al. 2006; Khaleque 2012; Khaleque and Rohner 2002), it would be interesting to compare the relationships obtained on the basis of children’s and parents’ reports of parental acceptance. IPARTheory research suggests that analyses based on children’s and parents’ reports of parenting style generally produce similar conclusions and that where there are differences, children’s reports of parenting are more reliable than parents’ self-reports (Rohner et al. 2009). Use of other methods (e.g. observations, interviews, etc.) would provide a more objective index of parenting. In view of the rapid changes in the family structure which have taken place over the past four decades in Italy and worldwide (Deater-Deckard et al. 2011) future research might consider the impact of different family structures (traditional, mixed-culture, divorced, remarried, foster, adoptive and same-sex families) on adolescent development.