Quality Assessment of the Studies

An adapted list of quality assessment criteria (Kmet et al., 2004) were used to evaluate the quality of the studies (see Table 3 for the quality scores). Studies were evaluated on a total of 9 criteria and were given 0–2 points (0 = no; 1 = partial; 2 = yes) on each. The first criterion was whether the research questions/objective was sufficiently described. Studies were given 2 points if they had a clear section in the end of the introduction that presented a clear research question/aim and information about the target; 1 point was given if this information was not clearly stated or if this information was gained elsewhere in the manuscript. The second criterion was whether the study design was evident and appropriate (2 points). Studies were given 2 points if the study design enabled testing of the proposed research question. Studies were given partial points (1 point) if the design did not enable test of the research questions (e.g., research question specifies a mediation, but the study had a cross-sectional design). The third criterion involved the description of the participants. Studies were given 2 points if they reported descriptive information on child’s age and sex, parents’ age and sex, and family background (e.g., SES). One point was given if only partial information was reported. For the fourth criterion, the appropriateness of the sample size was evaluated. N = 781 was used as the cutoff to get 2 points, which should be enough to detect a weak correlation (0.10) with a power of 0.80 (alpha = 0.05). To get 1 point, n = 82 was used as the cutoff, which should be enough to detect a moderately strong correlation (0.30) with a power of 0.80 (alpha = 0.05). The fifth criterion involved the analytical methods. Studies were given 2 points if they reported on model fit and missing data, and if the statistical analyses were appropriate for the test of the research questions. If studies did not report on this or used inappropriate analyses, they were given either 1 or 0 points. The sixth criterion was whether studies reported some estimate of variance for the main results. Two points were given if there was a standard deviation or confidence interval; 1 point was given if studies presented only standardized coefficients; 0 if they did not include any indicators of variance. For the seventh criterion, studies were evaluated on the inclusion of demographic controls. Studies were given 2 points if they accounted for at least one demographic variable (i.e., race/gender/age), either as a central control variable or a moderator of study effects. If studies did not include any demographic variables, they received 0 points. The eighth criterion involved whether the results were reported in sufficient detail. Studies received 2 points if the results section and associated tables and figures offered enough details to enable readers to discern all results related to the research questions or study aims. For the ninth criterion, studies were given points based on whether their conclusions supported the results. Two points were given if study conclusions aligned with the research questions and study aims without extending beyond the bounds of study limitations (e.g., inappropriately implying or explicitly drawing causal inferences).

Table 3 Quality assessment of the studies included in the review

As can be seen in Table 3, most studies received a high score (Mpoints = 15; maximum points were 18). Points ranged from 11 to 17, with most studies receiving either 14 (28%) or 17 (28%) points. On criteria 6–9, most studies received maximum points. On criteria 1 and 2, about 60% of the studies received maximum points and on criteria 3 and 5, about 50% received maximum points. The reasons for lower points on these criteria were mostly due to the use of cross-sectional design although research questions specified a longitudinal relation, missing information about the participants or research questions, or lack of information on missing data. Criteria 4 (i.e., appropriate sample size) included the highest number of studies with only partial points. According to the specified cutoffs described above (2 points = at least 781 participants; 1 point = at least 82 participants), most studies (69%) used too small samples to be able to detect weak correlations. Overall, then, although the majority of the studies received high scores on quality, lower scores were due to designs not appropriate for the research questions, small samples, and limited information on participants.

What Role Does PSE have in Associations with Parenting and Child Behaviors?Associations Among PSE and Parenting

A total of 19 studies have examined the associations between PSE and parent behaviors (see Table 4 for results). Across all 19 studies, independent of age of the sample, PSE and parenting were significantly associated. Eleven studies examined the cross-sectional associations at one time point only and found consistent evidence of an association between PSE and a range of positive parenting behaviors. Higher PSE have been associated with, for example, more positive parenting practices in samples in the United States (Glatz & Trifan, 2019; Glatz et al., 2017; Kiang et al., 2021) and Canada (Costigan & Koryzma, 2011), and more open and positive parent–child communication in Italy (Bandura et al., 2011), the United States (Glatz et al., 2018) and Taiwan (Chang et al., 2015). In addition, there is some evidence that higher PSE is associated with lower levels of negative parenting practices, such as parental rejection and aggression, in samples in the Philippines (Daganzo et al., 2014; Garcia & Alampay, 2012) and higher levels of parental monitoring in the United States and Taiwan (Chang et al., 2015; Malm et al., 2017). One recent study by Buchanan and colleagues (2022) found that parents with higher levels of PSE among mothers of children in early adolescence showed lower physiological responses to stress (i.e., more moderate skin conductance and a smaller increase in cortisol).

Table 4 Results on associations among PSE, parent, and child behaviors

Two longitudinal studies have examined PSE as a predictor of parenting (Egberts et al., 2015; Steca et al., 2011). In a Belgian sample (children ages 11 or 12), PSE predicted higher levels of warmth and support, and lower levels of parental reactivity two years later (Egberts et al., 2015). Further, in an Italian sample (children ages 13 or 14) (Steca et al., 2011), PSE predicted higher levels of parent–child positive and open communication three years later (Egberts et al., 2015). These two studies used longitudinal data (PSE at the first time point and parenting at the second time point), but they did not control for parenting at the first time point, and, thus, the results do not inform about increases or decreases in parenting over time.

Six additional longitudinal studies among parents of adolescents in the United States (Dumka et al., 2010; Glatz & Buchanan, 2015b; Lippold et al., 2019; Morrisey & Gondoli, 2012) and parents of children in school-age collected in Belgium (Slagt et al., 2012; van Eldik et al., 2017) have examined the possibility that PSE might both predict and be predicted by parenting. Five of these studies measured PSE and parenting at two or more time points and used cross-lagged models or alternative models to examine the reciprocal relationships between parenting and PSE (Dumka et al., 2010; Glatz & Buchanan, 2015b; Lippold et al., 2019; Morrisey & Gondoli, 2012; van Eldik et al., 2017). The sixth study (Slagt et al., 2012), measured PSE at the first and third time points, and parenting at the second time point (i.e., parenting was used as a predictor of changes in PSE, but PSE was not used as a predictor of changes in parenting). Regarding findings, four of these studies—covering both school-age and adolescence—showed support for a reciprocal relation: higher levels of PSE predicted higher levels or increases in positive parenting (high support and low inept discipline, mindful parenting, and promotive parenting [parents’ practices that aim to cultivate children’s skills, talents, and interests and to prevent negative adjustment, Furstenberg et al., 1999) or parental well-being. Higher levels of positive parenting or parental well-being also predicted increases in PSE (Glatz & Buchanan, 2015b; Lippold et al., 2019; Slagt et al., 2012; van Eldik et al., 2017). The two additional studies involved parents of adolescents (11–14 years, Dumka et al., 2010) and parents of late school-age children/early adolescents (9–11 years, Morrissey & Gondoli, 2012). These studies showed a unidirectional association: PSE predicted changes in mothers’ parenting (democratic style, parental control), but mothers’ parenting did not predict changes in PSE.

Associations Between PSE and Child Behaviors

In total, 17 studies have examined the association between PSE and child behaviors (mostly commonly externalizing behaviors) and showed evidence of a significant association between these. The longitudinal studies on this association, however, showed not only that these are associated, but also that the association might go one way.

Ten of the studies were cross-sectional and examined the association between PSE and child outcomes at one time point. These studies included both school-aged and adolescent-aged children, and all studies reported a significant association between PSE and child outcomes except for one (Garcia & Alampay, 2012). Results showed that parents with higher levels of PSE had children with fewer internalizing and externalizing behaviors in comparison to parents with lower levels of PSE. For example, Chang and colleagues (2015) found that higher PSE in Taiwanese parents about substance use was associated with lower levels of tobacco use and alcohol drinking. Mahabee-Gittens and colleagues (2011) found that U.S. parents who had higher PSE had children with fewer intentions to smoke. Other studies have found higher PSE to be associated with lower child aggression, violence, and bullying in U.S. samples (Malm et al., 2017) and Italian samples (Steca et al., 2011). Finally, there is some evidence that higher PSE is associated with better child psychological adjustment (Costigan & Koryzma, 2011), and lower risk for loneliness and depression in school-age children and adolescents in Finland (Junttila & Vauras, 2014) and Italy (Steca et al., 2011).

Three longitudinal studies have examined associations between PSE and child behaviors over time. These studies have either examined PSE as a predictor of changes in child behaviors (Bornstein et al., 2017; Dumka et al., 2010) or child behaviors as a predictor of changes in PSE (Glatz & Buchanan, 2015a). None of these studies, however, have examined bi-directional links between PSE and child behaviors. The two studies examining PSE as a predictor of changes in child behaviors showed different results. Dumka and colleagues (2010) found a significant link between PSE and changes in conduct problems among parents of children somewhat later in the adolescent period (11–14 years at the starting point) in the United States. Bornstein and colleagues (2017), on the other hand, did not find support for a link between PSE and changes in child externalizing behaviors (age 8 at the starting point) in nine different countries (China, Colombia, Italy, Jordan, Kenya, Philippines, Sweden, Thailand, and the United States). The only study that examined child behaviors as a predictor of changes in PSE (Glatz & Buchanan, 2015a) found that internalizing behaviors predicted initial levels, but not changes, in PSE.

Four longitudinal studies—all conducted in either the United States or Belgium—tested for bidirectional links between PSE and child externalizing behaviors over time (Glatz & Buchanan, 2015b; Morrissey & Gondoli, 2012; Slagt et al., 2012; van Eldik et al., 2017). All four studies measured PSE and child behaviors with two or more time points and used cross-lagged or alternative models to examine reciprocal relations. These studies covered samples of children from six to 12 years at the starting point, which, as they are longitudinal, together cover a large part of the school-age and adolescent period. All bidirectional studies provided evidence that difficult child and adolescent behaviors predicted decreases in PSE, but none showed evidence of an effect of PSE on changes in child behaviors.

Indirect Effects of PSE on Child Behaviors, via Parent Behaviors

In addition to tests of direct effects among PSE, parent, and child behaviors presented above, three identified studies tested an indirect effect of PSE on child behaviors, via parenting behaviors. These have presented mixed results, with some showing empirical support and some not. One longitudinal study on U.S. parents of children ages 11 or 12 at the starting point found evidence of such mediation. Specifically, Glatz and Buchanan (2015b) found that higher PSE predicted more promotive parenting, which in turn predicted lower levels of externalizing behavior, but only among mothers. Among fathers, the link between PSE and parenting was not significant. Yet other studies have not found support for an indirect effect. Slagt and colleagues (2012) conducted a longitudinal study in Belgium on parents of children 6 to 10 years at the starting point, and they did not find inept discipline (i.e., criticism, anger, Prinzie et al., 2007) or supportive parenting to be significant mediators between PSE and later child externalizing behaviors. Similarly, a longitudinal study by Dumka and colleagues (2010) conducted in the United States, found that the effect of maternal PSE on adolescents’ externalizing behaviors (11–14 years at the starting point) was not mediated by changes in maternal control practices.

What Additional Parent, Child, and Socio-contextual Factors are Associated with PSE?Parent Characteristics as Predictors of PSE

Twenty-three (66%) of the studies examined additional correlates of PSE. Parent characteristics were the most frequently examined factor among studies on parents of both school-age children and adolescents (n = 18, 51%). Parents’ age, ethnicity, and gender were significant predictors of PSE, although there were some mixed findings. Older parental age was found to be associated with higher levels of PSE in a sample including mostly mothers (Carless et al., 2015). Regarding parents’ gender, the few existing studies show mixed results. One study found that Belgian mothers of children ages 5–11 reported lower levels of PSE than Belgian fathers (de Haan et al., 2009), whereas another study suggested that Taiwanese mothers of adolescents have higher mean levels of PSE than Taiwanese fathers (Chang et al., 2015).

With respect to racial/ethnic identity, one cross-sectional study (Glatz & Trifan, 2019) and one longitudinal study (Glatz & Buchanan, 2015a) found that parents of adolescents who identified as African American reported higher initial levels of PSE. In the study by Glatz and Buchanan (2015a), however, parents identifying as European American changed less in their PSE over time in comparison to parents who identified as African American. Another study found no difference between African American and European American parents of children ages nine to 16 (Mahabee-Gittens et al., 2011). In the context of immigration, cross-sectional studies from the United States found that acculturation was positively associated with PSE, whereas acculturation conflict between parents and adolescent-aged children (10–18 years) was negatively associated with PSE (Costigan & Koryzma, 2011; Kiang et al., 2017).

Other studies have examined associations between parents’ psychological well-being, adjustment, and values in shaping PSE. Most of these studies were cross-sectional, and the results presented here are associations at one time point only. Several studies conducted in Europe and Asia indicate that higher levels of parent depression, anxiety, stress, and loneliness are associated with lower levels of PSE among parents of children in both school-age and adolescence (Carless et al., 2015; Junttila et al., 2007; Suzuki, 2010). Two studies (conducted in Belgium and the United States) assessed associations between parents’ personality and PSE, finding that higher levels of extraversion, agreeableness, emotional stability, autonomy, dominance, self-control, and independence were associated with higher levels of PSE; whereas higher levels of apprehension and anxiety were associated with lower levels of PSE (de Haan et al., 2009; Henney, 2016). There is also some evidence that parents’ attitudes, expectations, and aspirations as they relate to their child are linked to PSE. For instance, U.S. parents’ higher expectations of adolescents’ risk-taking behavior were associated with lower PSE (Glatz & Buchanan, 2015a). Another study on French parents of school-age children found that higher aspirations and expectations among parents for their child were associated with lower levels of PSE (Tazouti & Jarlégan, 2019).

Child Characteristics as Predictors of PSE

In seven studies (20%), child demographic characteristics and personality traits were assessed as predictors of PSE. In studies conducted in Asia (Japan and Taiwan) and in the United States, parents of school-age children and adolescents have shown to report higher levels of PSE for girls than for boys, (Chang et al., 2015; Glatz & Buchanan, 2022; Holloway et al., 2016). There is also some evidence that early pubertal changes make parents feel less efficacious before and during the transition to adolescence (Glatz & Buchanan, 2015a), perhaps because these changes signal to parents about increased independence. In terms of child personality, one longitudinal study found that higher levels of child extraversion and conscientiousness were associated with higher levels of PSE, whereas child benevolence, emotional stability, and imagination were not significantly associated with PSE (Egberts et al., 2015).

Socio-contextual and Relational Factors as Predictors of PSE

Eight studies (23%) have examined factors within a larger socio-contextual and relational domain as predictors of PSE. These include relational aspects (e.g., parent–child, co-parenting) and family aspects (family dysfunction, SES), which are not related specifically to either the parent or the child. Parent–child communication quality and coparenting quality were both found to predict PSE, both concurrently (Latham et al., 2018) and longitudinally (Glatz & Buchanan, 2015a). Parent–child conflict was found to be negatively associated with smartphone-specific PSE in a cross-sectional study on a sample including a majority of mothers of adolescents in Hong Kong (Wong & Lee, 2017). In terms of family-level characteristics, family dysfunction was found to be negatively associated with PSE in mothers of adolescents in Australia (Carless et al., 2015), whereas higher socio-economic status and household income were both found to be positively associated with PSE in parents of school-age children and adolescents in the United States, Asia, and Europe (Glatz et al., Holloway et al., 2016; Tazouti & Jarlégan, 2019).

What Moderates the Associations Among PSE, Parenting, and Child Behaviors?

Although the studies in this review were conducted in 14 different countries, there were very few studies in which comparisons were performed between participants based on race, culture, or nationality. As exceptions, four studies compared findings in samples from two or more cultures. In two studies, Kiang and colleagues (Kiang et al., 2017, 2021) showed some differences in associations among Asian American parents and Latinx parents of adolescents. The first study (Kiang et al., 2017) showed that among Asian American participants, but not Latinx participants, acculturation conflicts predicted lower PSE, especially when parents felt less efficacious in transmitting heritage messages to their children (low levels of cultural-specific PSE). In the second study (Kiang et al., 2021), for Latinx parents, the negative correlation between grade and PSE was weaker when parents were high on involvement. In a third study by Mahabee-Gittens and colleagues (2011), the authors tested if there were racial differences in the effects of PSE on youth smoking intentions in a sample of majority mothers (ages 9–16), finding associations in their models were similar between African American and Caucasian families. A fourth study (Bornstein et al., 2017) was the only study testing differences between parents of school-age children living in different countries, and they showed that PSE was not a significant predictor of child externalizing behavior in any of the countries (China, Colombia, Italy, Jordan, Kenya, Philippines, Sweden, Thailand, and the United States).

In terms of gender, out of the 35 studies included in the review, eight included a sample of only mothers. The rest (n = 27) included both mothers and fathers, but with a general underrepresentation of the number of fathers in the sample. A few of these studies have examined gender differences in the associations involving PSE. These studies have presented mixed findings, with slightly more studies showing a non-significant effect. Babskie and colleagues (2017) found that higher PSE regarding alcohol and antisocial peers for mothers, but not fathers, was associated with less youth drinking and delinquency. Further, Glatz and Buchanan (2015b) found PSE to be more predictive of promotive parenting for U.S. mothers than for U.S. fathers, and promotive parenting mediated the association between PSE and child externalizing behaviors for mothers but not for fathers. These two studies were conducted on parents of adolescents between 11 and 18 years of age. Four other studies—covering a somewhat younger developmental period (5–12 years of age) than the two above-mentioned studies—did not find parent gender differences in the associations between PSE and parenting (Daganzo et al., 2014; de Haan et al., 2009; Egberts et al., 2015), or between PSE and child externalizing (van Eldik et al., 2017).

How Does PSE Change Over the Developmental Period?

Some studies have examined potential changes in PSE over time—either by presenting a correlation between age and PSE, by examining the means of PSE in different age groups (cross-sectional data), or by reporting on PSE at multiple times (longitudinal data). Most of these studies suggest that PSE decreases over the school-age and adolescent period.

Seven studies reported on correlations between children’s age and PSE. Three of them—covering both school-age and adolescence (Carless et al., 2015; de Haan et al., 2009; Wong & Lee, 2017)—reported a non-significant correlation. Four studies (Egberts et al., 2015; Glatz & Trifan, 2019; Glatz et al., 2018; Kiang et al., 2017) reported a significant negative correlation suggesting that PSE was lower for parents of older children than for parents of younger children. One additional cross-sectional study on American parents of adolescents (Babskie et al., 2017) tested and showed evidence that parents of older adolescents reported lower levels of PSE than parents of younger adolescents.

Eight longitudinal studies on parents of school-age children (Morrissey & Gondoli, 2012; Slagt et al., 2012; van Eldik et al., 2017) and/or adolescents (Chang et al., 2015; Glatz & Buchanan, 2015a, 2015b; Lippold et al., 2019; Steca et al., 2011) reported the means of PSE at multiple time points. Most of these studies reported lower means over time (Chang et al., 2015; Glatz & Buchanan, 2015a, 2015b; Slagt et al., 2012; Steca et al., 2011; van Eldik et al., 2017), suggesting that PSE decreases as children become older. Two studies conducted in the USA showed slightly different results. First, one study on parents of school-age children showed somewhat stable means in PSE over time (Morrissey & Gondoli, 2012), and one study covering both school-age and adolescence showed higher means in mothers’ PSE over time (Lippold et al., 2019). The study by Lippold and colleagues used data from an intervention study, which may be why means of PSE increased over time. Although several studies reported PSE means at multiple time points, only one study tested for significance in changes of PSE. Specifically, Glatz and Buchanan (2015a) used Latent Growth Curve Modeling to test for a significant slope in PSE among parents in the United States (Glatz & Buchanan, 2015a); a test that supported a significant decrease in PSE from 11 or 12 years of age to 14 or 15 years of age.