Introduction

Gratitude has recently gained increasing attention from scholars of different disciplines, mainly positive psychology. McCullough et al. (2002) defined gratitude as a “a generalized tendency to recognize and respond with grateful emotion to the roles of other people’s benevolence in the positive experiences and outcomes that one obtains” (p. 112). Gratitude can be considered a prosocial affect because it is a response to behaviours that, in turn, motivate other behaviours (McCullough et al. 2001). Emmons and Shelton (2002) stated that gratitude is a positive universal characteristic that is present in all cultures and is integral to most world religions. Others (e.g., Froh et al. 2011) consider gratitude an emotion that impresses a sense of connection to people and communities. For these reasons, gratitude has important implications in psychological well-being (Emmons and McCullough 2003; McCullough et al. 2002; Kashdan et al. 2006; Wood et al. 2007, 2008).

This topic is relevant because there is evidence showing how gratitude can be a viable path to promoting a variety of positive outcomes among adolescents, such as academic achievement, better social and emotional functioning, and increased self-esteem (Dongping et al. 2012; Froh and Bono 2008; Froh et al. 2008, 2009, 2010; Kashdan et al. 2006; Toussaint and Friedman 2009).

Research suggests that gratitude is associated with adolescents’ self-esteem (see for example Dongping et al. 2012; Kashdan et al. 2006; Kong et al. 2015; Lin 2015). Self-esteem refers to a person’s global evaluation of his or her worth as a human being (Rosenberg 1965) and is closely related to health and well-being (Heinonen et al. 2004). Adolescents who display high levels of gratitude are more likely to acknowledge how the other people’s benevolent actions contribute to their well-being (McCullough et al. 2001). Therefore, they tend to feel that they are loved and cared for by others. Toussaint and Friedman (2009), for example, found that outpatients with high levels of gratitude were more likely to evaluate themselves positively. Gratitude can foster positive representations of the self, including high levels of self-esteem and self-worth.

Nevertheless, few studies addressed what factors can promote gratitude, and none that looked into family factors promoting gratitude. One’s family is a crucial environment for the development of a child’s self-esteem, attitudes, values, and norms (Alfieri et al. 2014a; Barni et al. 2013; Coopersmith 1967), and it is potentially crucial for learning and developing gratitude (Scabini et al. 2006). Within the family, people can experience, for the first time, a sense of gratitude directed toward their parents (Scabini 2011); during adolescence, children start to fully manifest what their parents have passed down to them (Scabini et al. 2006). To date, however, there are no studies demonstrating such relations.

There is evidence that parents play an important role in the process of developing self-esteem in their children (see for example Alfieri et al. 2014b). Coopersmith (1967) stated that self-esteem develops through respect, acceptance, and attention within close relationships, particularly emphasizing parents’ role. Hopkins and Klein (1993) highlighted how parental nurturance is positively related to high self-esteem in the family. Lim et al. (2015) also suggested that children who perceive love, acknowledgement, and understanding from parents develop self-confidence and look at themselves more positively. We focused on emotional support that, with regard to parents, has been defined as parents’ warmth, responsiveness, acceptance and involvement towards their offspring (Scabini et al. 1999). Emotional support from close relationships is negatively related to psychological distress among adolescents (Alfieri and Lanz 2015; Alfieri et al. 2016; Demaray and Malecki 2002) and it has also been found to be positively associated with self-esteem (Alfieri et al. 2014b; Bámaca et al. 2005; Plunkett et al. 2007).

Only a few studies, however, have focused on the specific role of the mother’s and father’s emotional support (Graziano et al. 2009; Noack et al. 1999). The mother and the father are role models in socializing their children through specific patterns of behaviour, and most studies focused on the maternal role highlighting how the mother-adolescent relationships are characterized by a higher level of intimacy and greater disclosure, especially of emotional aspects, and adolescents are more likely to seek maternal rather than paternal advice (Dolgin and Berndt 1997; Miller and Stubblefield 1993).

The father-adolescent relationships have been, instead, described as less focused on affective aspects and more on discipline, preparation for future life tasks and the acquisition of a social status (Collins and Russell 1991; Siegal 1987). A study by Shulman and Seiffge-Krenke (1997) highlighted how fathers have a greater influence on adolescents’ self-esteem than mothers, and promote their individuation process. Verschueren and Marcoen’s (2002) study showed there is a positive correlation between both father’s and mother’s attachment and adolescents’ self-esteem, whereas Simons et al. (2001) suggested that only the mother’s attachment is associated with children’s self-esteem.

In light of the theoretic review, the present study aims to investigate the relation between parental emotional support and self-esteem through the mediation of gratitude. In particular, we hypothesised that: h1) both the mother’s and the fathers’ emotional support promote gratitude; h2) gratitude plays a mediation role between the mother’s and father’s support and self-esteem.

Method

Participants

The sample included 279 15–19 year-old Italian adolescents (M = 16.83, SD = .85; 46.9% male). 81.5% were living with both parents, 10.7% with one of them, 5.5.% with a parent and a stepmother/father, and 2.2% indicated “other.” Participants were recruited in eight different high schools: four scientific/humanities (50.0%), two technical (25.0%), and two professional schools (25.0%) in northern Italy. Five of these schools are located in urban areas and three in suburban ones.

Procedures

Participants were administered an anonymous self-report questionnaire during school time. No participant showed any sign of uneasiness while completing the questionnaire and none of them was indicated by teachers as having behavioural difficulties. Participants were informed about confidentiality and use of the data, and that participation was voluntary and unrewarded. Underage students’ parents (at least one of them) provided written consent before the administration of the questionnaire.

A total of 337 questionnaires were distributed and 296 were returned (retention rate of 87.8%). Attrition was due to failure to get parents’ consent to participate in the study. Furtherly, 17 incomplete (i.e., large sections of the questionnaire were left blank) or grossly unreliable (i.e., the same answer was indicated throughout the entire questionnaire) questionnaires were excluded. 54.2% of the included questionnaires had no missing data, 30.0% reported one missing item, 9.5% had two missing items and 6.3% reported three or more missing items. Full information maximum likelihood (FIML) has been used to handle missing data found in the other questionnaires.

Measures

Emotional support

We measured emotional support using Scabini and Cigoli’s (1992) Parental Support Scale, which measures adolescents’ perceptions of the degree of emotional support received from their parents. The 18 items (e.g., “My mother/father doesn’t understand me”) are measured on a 5-point Likert scale (1 = Never to 5 = Always). Adolescents were asked to answer in regards to both their father and mother. We tested relational measurement invariance in order to know whether the measurement of emotional support was equivalent in the relationship with the father and with the mother. We applied the procedure proposed by Tagliabue and Lanz (2014), in which item residuals were correlated across the two relationship types (i.e. item 1 residual in mother-child relationship was correlated to item 1 residual in father-child relationship). Then, for testing metric invariance, we constrained to be equal the factor loadings of each item across the two relationship types. Cut-off points of ΔCFI < .01 and of ΔRMSEA < .015 were used to compare models (Chen 2007). Findings showed that relational metric invariance was verified (ΔCFI = .002 and ΔRMSEA = .000). We also tested for structural invariance, constraining to be equal the intercepts of each item across the two relationship types, but we did not find confirmation of that level of measurement invariance (ΔCFI = .013 and ΔRMSEA = .003), thus only constraints about metric relational invariance were included in the mediation model. Reliability was computed in the context of the confirmatory factor analysis (CFA), for each relationship, using McDonald’s ω (McDonald 1999; Peters 2014). Excellent reliability was found for both paternal emotional support (ω = .93) and maternal emotional support (ω = .92).

Gratitude

We measured gratitude using the Italian adaptation of McCullough et al.’s (2002) Gratitude Questionnaire-6, which measures the degree to which a person is grateful about life and other people. The six items (e.g. “I have so much in life to be thankful for”) are measured on a 7-point Likert scale (1 = not at all to 7 = extremely). Results of CFAs support the mono-factorial structure of the scale (χ2 (6) = 11.812, p = .066, CFI = .988, RMSEA = .059 [.000–.109]), although some residuals have been allowed to correlate. Reliability was acceptable (ω = .73).

Self-esteem

We measured self-esteem using the Italian validation (Prezza et al. 1997) of the Rosenberg’s (1965) self-esteem scale, which includes ten items measured on a 5-point Likert scale (1 = Not very true for me to 5 = Very true for me) (e.g., “On the whole, I am satisfied with myself”). Results of CFAs support the mono-factorial structure of the scale, χ2 (30) = 85.892, p < .05, CFI = .960, RMSEA = .082 (.062–.102), although some residuals have been allowed to correlate. Reliability was good (ω = .88).

Data Analyses

Model goodness-of-fit

Model goodness-of-fit was checked using several indices simultaneously (Bollen 1989). The first index was χ2: a model fits the data well when χ2 is not significant. However, the chi-square statistic is sensitive to sample size; therefore, we adopted additional fit indices that were less sensitive to sample size: CFI, and RMSEA. The CFI values equal to or above .95 were considered good and above .90 acceptable (Hu and Bentler 1997, 1999). Finally, RMSEA values equal to or smaller than .05 were considered satisfactory and smaller than .08 acceptable (Hu and Bentler 1997, 1999).

Mediation analyses

We used Mplus 7 (Muthén and Muthén 1998–2010) to test a mediational model in which gratitude mediates the link between paternal and maternal emotional support and self-esteem. Bootstrap analyses were used to assess direct and indirect effects in the mediation model (Lau and Cheung 2012; Shrout and Bolger 2002). Bootstrapping statistic allows examining the statistical significance of both direct and indirect effects. Significance of mediation is found when 95% bias-corrected confidence intervals (95% CI) of the indirect effects do not contain zero (MacKinnon et al. 2004). In the present study, parameter estimates were based on 1000 bootstrap samples.

Results

Table 1 shows the means, standard deviations, Cronbach’s αs, and correlations among variables. Model fit indices of the mediational model showed good fit, χ2 (1260) = 2120.746, p < .001, CFI = .906, RMSEA = .049 (.046–.053), and explained 16.1% of variance in self-esteem (R2 = .161) (Fig. 1). Direct effects of paternal emotional support (b = .068, SE = 0.097, p > .05) and maternal emotional support (b = .189, SE = 0.106, p > .05) on self-esteem were not significant, whereas the direct effect of gratitude on self-esteem was significant (b = .259, SE = 0.128, p < .05). Furthermore, paternal emotional support was not significantly related to gratitude (b = .179, SE = 0.109, p > .05), whereas maternal emotional support was (b = .316, SE = 0.134, p < .05). We computed unstandardized indirect effects for each of 1000 bootstrapped samples and a 95% bias-corrected confidence interval was used (Table 2). The initial total effect for maternal emotional support was 0.27 and the mediation effect of gratitude explain 30% of the total effect. Findings show that gratitude mediates the relationship between maternal, but not paternal, emotional support and self-esteem.

Table 1 Mean, standard deviation, Pearson r correlation coefficient and Cronbach’s alpha of variables
Full size table
Fig. 1
figure 1

Model

Full size image
Table 2 Indirect effects of paternal and maternal emotional support on self-esteem through gratitude (unstandardized coefficient)
Full size table

Discussion

In this study we aimed to contribute to the literature on factors that can promote gratitude by considering the role of emotional support. This is relevant because there is evidence that there is a strong positive relation between gratitude and positive outcomes, including well-being and self-esteem, in all phases of the life cycle, and particularly in adolescence.

The first result that emerges from our study is that there is a strong relation between the maternal emotional support and gratitude; on the contrary, no statistically significant relation was found between paternal emotional support and gratitude. These results confirms the importance of the maternal figure in the support of gratitude and how the maternal role is related to emotional aspects (Dolgin and Berndt 1997; Miller and Stubblefield 1993), in line with the definition of gratitude that was provided. These results, therefore, support the idea that when adolescents perceive that their mothers provide them care, closeness, and affection, they will be more “equipped” and more inclined to being grateful not only to their own family but, more generally, to life. Notably, despite a recent ferment among scholars on the importance of the father for many aspects of children’s adjustment and well-being, no evidence that the paternal emotional support is related to gratitude was found in our study (Cabrera and Tamis-LeMonda 2014). Cultural differences, however, may play a role given that in Italy the maternal role is still strongly felt as the delegate to the care of the children and to the task of transmitting emotional and relational aspects (Alfieri et al. 2014b; Scabini et al. 2006).

A second result is the existence of a strong relation, already found in literature, between gratitude and self-esteem (Dongping et al. 2012; Kashdan et al. 2006; Kong et al. 2015; Lin 2015). In our study, no statistically significant relation was found between emotional support and self-esteem, because the relation was totally mediated by gratitude. This is an innovative result given, until now, there were no studies investigating the factors that support it. In this sense, the present study, although in correlational and descriptive terms, could indicate how parental emotional support should pass from a process of recognition and gratitude to being able to produce well-being. Further longitudinal studies are needed to provide support for this claim.

A reflection should be made regarding the relation between support and gratitude. Froh et al. (2009) proposed a model in which the relational fulfilments act as mediators between gratitude and physical symptoms in a sample of pre-adolescents. This model resulted in adequate fit indices, whereas the alternate model (i.e., in which it is gratitude that plays the mediator role) did not. In our study, rather, we have tested the opposite relation, that is, gratitude mediating the relation between support and self-esteem. Sun et al. (2014), as well, proposed a model in which social support acts as mediator between gratitude and scholastic well-being, producing satisfying indices of fit. There can be several reasons explaining difference in results of our study: firstly, support is understood differently. In our case, it is a perception of support that the adolescents believe to have obtained from their parents. In the research of Froh et al., on the other hand, support refers to the support that the preadolescents feel they can give to their friends, family members, etc. A second aspect is that Froh et al.’s (2009) study included aspects in the relational fulfilment that are quite diverse—family satisfaction, friend satisfaction, family support and friend support—thus combining two different actors (familial and friends), and two different constructs (emotional support and satisfaction); utilizing as a synthetic indicator the average of the four items, possible differences may have been “flattened”. Similarly, Sun et al. (2014) used the average of three indicators as a supporting index: “significant others,” “family,” and “friend”, thus incurring in the same construct flattening issue. A second reason explaining the different results obtained by the research of Froh et al. (2009) may relate to the fact that the authors did not find gratitude was a mediator between relational fulfilment and physical symptoms, but only between familial variables and well-being outcomes.

It is also worth Italian parents are particularly present and attentive all throughout the transition to adulthood (Crocetti and Meeus 2014; Scabini et al. 2006), with a clear unbalance towards the affective quality at the expense of the ethical-normative ones (Alfieri et al. 2014b). A prime example is of the average age of leaving home, though now a widespread phenomenon in the countries of southern Europe: in Italy 62.3% of young people between 18 and 35 years of age live with their parents (Istat 2011). Given this and the other two studies mentioned have a cross-sectional nature, it is not currently possible to temporal and casual nature of the relationship between these variables, and future research could cast light on this.

Limitations

The results of this study needs to be considered in light of some limitations. The sample size was small and not representative of the Italian population. The sample size did not allow us to conduct further analyses specific to subgroups (e.g., between adolescents that live with both parents and those who only live with one parent). Furthermore, the cross-sectional nature of this study did not allow us to monitor possible changes over the course of different life stages. It is worth noting, in addition, that this study used a self-report instrument, which gives us only the respondents’ views and is susceptible to various degrees of social desirability. Finally, research (see for example McCullough et al. 2002; Wood et al. 2008) suggests that neuroticism or an emotional state form the base of gratitude. It is possible that the relations identified in the present study reflect the effects of a single underlying factor (such as neuroticism or emotional state). Even if this was the case, however, the relations between the variables found in this study would still be valid, and should only be incorporated in a larger model that has, at its base, a factor that explains them in their complexity.

Tthe present study also has several strengths. Firstly, our results support the idea of the relevance of the family in our society, not only for the well-being of the family itself, but of the society overall. This is because the sense of gratitude experienced by adolescents can be generalized beyond one’s own family to other people, and can also enhance their well-being decreases, thus contributing to reduce the social and economic burden of disorders such as depression or anxiety. Future research, however, should better investigate the reasons behind the lack of relationship between parental emotional support and self-esteem. Furthermore, it would be interesting to investigate the relations between gratitude and generativity because that “is something that is received and given, something that others have passed on to us and that we, in turn, will pass on, after having given it our own imprint” (Scabini et al. 2006, p. 7). Finally, we must underline that we have taken into consideration a population—that of Italian adolescents—or which there has been no research to date on gratitude and that, as previously discussed, presents some peculiarities linked to the family’s importance.

In conclusion, the present study highlights that (a) only mother’s emotional support promote gratitude; (b) gratitude plays a role of mediation between the mother’s support and self-esteem. Results shown in the present research should be confirmed and expanded by future research. In order to confirm the results of the present study, authors suggest to use a multi-method approach. It would be interesting to look more deeply into the relation between maternal emotional support and gratitude from a qualitative point of view. The qualitative in-depth analysis could clarify why only maternal emotional support sustains gratitude. In addition, we suggest, as preciously highlighted in this paper, a longitudinal research design. This latter could explain the causal effects between variables that were examined in an explorative way in the present research. The present study could be expanded considering gratitude as sustained by some other familial variables such as e.g., quality of the family relationship, children’s parental legitimacy, and monitoring. To understand which factors promote gratitude is of extreme importance since gratitude has effects on health variables. One last consideration regards the possibility to test the model presented in this paper also with adults or young adults to understand if it works also in different phases of the life cycle.