Abstract
Through a lens of confirmation theory, this study examined how parental confirmation acceptance and challenge was related to emerging adults’ body image, and how these associations were mediated by social competence and self-concept. This study also examined whether the hypothesized links differed by parental and child sex. Male and female college students (N = 447) completed an online questionnaire. A multiple-group analysis generally supported the proposed mediation model, providing a somewhat complex picture of associations among key concepts in terms of which component of confirmation was associated with certain components of body image and through which mediator varied as a function of parental sex. Unlike parental sex, child sex played a minimal role in the current study. This research underscores the utility of employing a confirmation perspective in understanding the mechanisms of how family interactions are related to body image.
Body image refers to our attitudes about our own weight, body size and shape, and appearance (Cash & Pruzinsky, 1990). Negative body image, defined as concerns, problems, or disturbances involving body image, is prevalent in the United States (Fallon, Harris, & Johnson, 2014). Emerging adulthood, typically defined as 18 to 25 years of age (Arnett, 2000), is especially marked by a high risk of body dissatisfaction and maladaptive eating behaviors for women (Eisenberg, Nicklett, Roeder, & Kirz, 2011) and men (Mayo & George, 2014). Poor body image is related to a host of negative psychological outcomes including low self-esteem, low life satisfaction, and depressive symptoms in both men and women, as well as eating disorders for women and muscle dysmporphia and the use of muscle-building drugs for men (Grogan, 2016). Considering the prevalence of body image problems and their negative psychological and behavioral correlates, it has become increasingly important to identify their possible contributing factors.
Although a complex web of factors contributes to negative body image (Thompson, Heinberg, Altabe, & Tantleff-Dunn, 1999), family-of-origin experience is one of the salient influences (Kluck, 2010). Family members, especially parents, are key socialization agents for children: Through interactions with their parents, children develop views toward themselves (Cooley, 1902), including their bodies (Arroyo & Segrin, 2013). A large body of literature has examined how parent-child interactions, particularly mother-daughter interactions, are related to negative body image and eating disorders (e.g., Humphrey, 1989; Prescott & Le Poire, 2002). Although extant research is informative in understanding the general importance of parent-child interactions, there are key limitations that need to be addressed. First, even though the associations between family dynamics and body image issues are relatively well established, the mechanisms through which family factors influence body image problems are less explored. Furthermore, most extant body image literature on the role of family factors has focused on mother-daughter dyads, limiting our understanding of paternal influence as well as sons’ experiences.
Taking these limitations of extant research into consideration, this study examines how maternal and paternal communication is related to male and female college students’ body image by examining the potential mediators. A wealth of research has focused on parental messages specific to appearance and weight (e.g., appearance-related teasing and criticism) as a contributor of negative body image for their children (e.g., Schaefer & Salafia, 2014). Unlike previous research, we aim to examine how the general quality of parental communication, that is, communication that does not necessarily involve talking about appearance and weight, can also play an important role in their children’s body image. One insightful framework in examining this issue is confirmation theory (Dailey, 2010). Confirmation is a process in which one feels endorsed by others through recognition and acknowledgment (Cissna & Sieburg, 2006) and is hypothesized to contribute to children’s development in a variety of domains. In addition, two concepts relevant to parental confirmation and body image will be examined as potential mediators: social competence and self-concept (i.e., identity strength and self-esteem). Whereas previous studies have examined the role of either social competence or self-concept as a single predictor of body image, this study provides a more systematic investigation of multiple key predictors of body image in the same model. Thus, taking all sex-specific parent-child dyads (i.e., mother-daughter, mother-son, father-daughter, father-son) into consideration, this study tests a model in which parental confirmation is hypothesized to be positively associated with emerging adult children’s development of self-concept and social competence, which, in turn, are positively associated with their body image.
This study focuses on the period of emergent adulthood, which is characterized by important transitions including leaving home and developing self-identity (Nelson, Story, Larson, Neumark Sztainer, & Lytle, 2008). As they transition into a new college environment, emerging adults might turn to their memories of interactions with their parents as a blueprint for how they should see the social world, others, and the self. Although they no longer live with their parents, emerging adults are in a new phase of life, and approach new interpersonal encounters with a mental framework that has been developed over years through interactions with their parents. In this sense, past interactions with parents can further fortify or diminish emergent adults’ self-concept, even in the domain of body image.
Parental Confirmation and Body Image
Confirmation theory (Dailey, 2010) is a useful framework to understand how parents might contribute to children’s body image from a communicative perspective. Confirmation theory suggests that all human beings have a fundamental need to be validated by others, which is essential to developing a strong sense of self (e.g., Buber, 1965). Confirmation is conceptualized as having two components: acceptance and challenge (Dailey, 2010). Acceptance is broadly defined as the degree to which specific others care for and value another by showing positive regard, care, warmth, and attentiveness during interactions (Brock, Sarason, Sanghvi, & Gurung, 1998). Challenge involves pushing and challenging a person so that he or she can achieve greater potential (Dailey, 2010).
Evidence suggests that parental acceptance is related to positive body image. A number of studies examining parental behaviors similar to acceptance (e.g., warmth, care, support, and nurturance) demonstrated such an association, both concurrently (Ellis, 2002) and longitudinally (Boutelle, Eisenberg, Gregory, & Neumark-Sztainer, 2009). Compared with acceptance, the role of challenge in children’s body image is less clear. Because challenge should facilitate growth of self-efficacy, ability, skills, and self-identity (Dailey, 2008), challenge should also foster the development of children’s positive body image. At the same time, challenge, in the absence of acceptance, might lead to feeling criticized and unworthy, which can lead to the development of a negative view toward themselves, including their bodies. In fact, authoritarian parenting, which can be argued to be similar to a combination of low acceptance and high challenge, is associated with negative body image (Enten & Golan, 2009).
Beyond individual contributions of acceptance and challenge, examining the combined effect of these two components has more theoretical importance, and, therefore, is the focus of the current study. Confirmation theory predicts that the effect of acceptance and challenge might be interactive, rather than additive. The combination of a high level of acceptance and a high level of challenge (i.e., confirmation) is hypothesized to produce more optimal child outcomes as compared with other combinations (e.g., high acceptance and low challenge). Findings on the interaction effect between these two components are somewhat mixed, with some studies supporting interaction effects (Dailey, 2008), whereas others did not (Dailey, 2010). Yet, there is a good theoretical reason to predict interaction effects between acceptance and challenge to predict body image. Children who feel both accepted and pushed to improve should feel comfortable with their bodies as well competent to take care of their bodies through proper eating and exercising. In fact, literature on parental behaviors similar to confirmation provides indirect support for this prediction. For instance, positive parenting conceptually close to confirmation (e.g., authoritative parenting) is related to children’s positive body image (Enten & Golan, 2009). Therefore, the following hypothesis is created:
Mediators of the Association Between Parental Confirmation and Body Image
Although there is relatively clear evidence of the link between family-of-origin experience and children’s development of body dissatisfaction and eating problems, the mechanisms through which family experience come to affect children’s body image are less clear. Two concepts that are most relevant to parental confirmation and body image are examined as potential mediators: social competence and self-concept. This study uses the term social competence as a general term (in a broad sense) that subsumes most other specific approaches to social competence. In addition, this study focused on two aspects of self-concept that are most relevant to the literature of confirmation and body image: self-identity (i.e., a stable, coherent, and integrated sense of self; Erikson, 1968) and self-esteem. Although some previous studies examined either social competence or self-concept as a potential predictor of body image issues, these studies did not simultaneously examine the role of both concepts in the same model. This study, thus, tests a more comprehensive model in which both potential mediators are simultaneously estimated (see Figure 1 for the hypothesized model).

Hypothesized model.
Social Competence as a Mediator
The current study builds on the model proposed by Mallinckrodt, McCreary, and Robertson (1995), which proposed that family dynamics directly and indirectly affect children’s eating disorders through social competence. Their model hypothesized that dysfunctional family environments lead to insecure childhood attachment, resulting in a deficit in the social competencies that are necessary to develop supportive relationships in adulthood. When stressful events occur, people with limited social skills cannot garner social support from interpersonal relationships, and, thus, are more vulnerable to the impact of the stressor, which might further lead to greater likelihood of psychological and physiological symptoms including eating disorders (Mallinckrodt, 2000).
In applying Mallinckrodt et al.’s (1995) model, the current study hypothesizes that social competence mediates the association between parental confirmation and body image. Because parental acceptance provides a safe and supportive environment for children to explore different ideas as well as practice skills to communicate their emotions and thoughts, acceptance should foster children’s development of social competence. Acceptance should also facilitate a positive view toward the world, which may lead to feeling comfortable and confident in social interactions. Yet, in the absence of challenge, children might not be pushed to improve their skills and attain a greater potential. Thus, acceptance alone might help children to be comfortable in social contexts, but might not encourage further growth.
Conversely, challenge alone might push children to strengthen their interpersonal ability but can potentially make them feel incompetent. Challenging parents push their children to cultivate an ability to articulate their thoughts and to provide logical reasoning behind them. These parents encourage their children to explore different ideas and experiences, which helps their children to develop a broad response repertoire when interacting with others. These parents also push children to consider other people’s perspectives, which is a key component of successful interpersonal relationships (Schröder-Abé & Schütz, 2011). In the absence of acceptance, however, challenge might be enough to induce such behaviors but might make children feel criticized and dismissed, and believe that parental challenge reflects incompetence on their part.
Beyond the main effects, high acceptance and high challenge (i.e., confirmation) should facilitate the optimal development of social competence, compared with other combinations of acceptance and challenge. Children who feel both accepted and challenged to improve should feel safe and comfortable as well as competent in social contexts. In fact, a relatively large body of parenting literature suggests that positive parenting practices (e.g., authoritative parenting style, affection expression), which can be argued to be similar to confirmation, are related to later social competence in childhood and even in adulthood (Shaffer, Burt, Obradović, Herbers, & Masten, 2009), providing indirect support that the combination of acceptance and challenge is conducive to children’s development of social competence.
This study further proposes that greater social competence, in turn, should be related to more positive body image. There are at least two related explanations for this association. First, individuals with limited social competence lack self-efficacy for desired social outcomes. They also have limited ability to build a social network to garner social support while coping with stress (Mallinckrodt et al., 1995). These individuals are likely to experience an array of psychological problems including low self-esteem, loneliness, and depression (Arroyo & Segrin, 2013; Segrin, 2001). Arguably, such negative psychological states might transfer to the domain of appearance. In fact, psychological distress (low self-esteem, depression) prospectively predicted negative body image (Paxton, Eisenberg, & Neumark-Sztainer, 2006).
Second, another reason for the association is the misguided belief that attaining socially idealized bodies will lead to successful interpersonal relationships. Because social connection is one of the most fundamental human needs (Maslow, 1955), individuals who feel socially isolated might, consciously or subconsciously, attempt to reconnect with others. When individuals with poor social competence feel isolated, they might attempt to improve interpersonal situations through other means, namely, by attaining socially idealized bodies (Budd, 2007; Gerner & Wilson, 2005). Physical attractiveness is often presented as a key ingredient of interpersonal success in our society (Abbott & Goodheart, 2011). As a result, those who particularly experience interpersonal difficulties might believe that attaining an ideal body will bring interpersonal improvement (Gerner & Wilson, 2005). Because the societal standards of thinness (for women) and muscularity (for men) are unrealistically elevated (Tiggemann, Gardiner, & Slater, 2000), the desire to attain an idealized body will likely result in body dissatisfaction and maladaptive weight control behaviors (Dobmeyer & Stein, 2003; Duggan & McCreary, 2004). Therefore, several areas of literature support the link between poor social competence and negative body image (Ferrier & Martens, 2008; Grisset & Norvell, 1992).
Although disordered eating and poor body image are two distinct concepts, Mallinchrodt et al.’s (1995) model for eating disorders can be used to predict body image. Both issues—disordered eating and poor body image—can be seen as different manifestations of compensatory reaction to a lack of social competence. In fact, at least two empirical studies with female college students provide general support for this model by explaining poor body image and disordered eating attitudes in nonclinical populations. Arroyo and Segrin (2013) found that problematic interaction patterns in families (e.g., criticism) were related to college-aged daughters’ low social competence, and low social competence was associated with psychological distress, which, in turn, was related to disordered eating attitudes. Furthermore, using family communication patterns (FCPs; Ritchie & Fitzpatrick, 1990) theory, Taniguchi and Thompson (2015) found that perception of mothers’ conversation orientation was positively associated with women’s social competence, which, in turn, was negatively related to their body dissatisfaction.
Self-Concept (Identity Strength and Self-Esteem) as a Mediator
In addition to the potential mediating role of social competence, the current study simultaneously tests whether self-concept also mediates the association between parental confirmation and body image. The link between parental acceptance and self-concept is relatively well established. Perceptions of parental acceptance, understanding, and nurturance have been consistently found to be related to greater identity strength and self-worth (Dailey, 2009; Schrodt, Ledbetter, & Ohrt, 2007). Yet, acceptance alone might be sufficient to encourage positive and strong sense of self but might not encourage children to obtain a greater potential.
Conversely, challenge alone might facilitate growth but might not be conducive to the development of positive sense of self. Challenge could provide children with opportunities to explore their potential and develop a strong sense of who they are, and who they want to become. In fact, some evidence shows a positive main effect of challenge in predicting children’s self-concept (Dailey, 2010), even after controlling for other parenting components, including acceptance (Dailey, 2008). In the absence of acceptance, however, challenge might push children to obtain a higher potential but might not encourage the development of a positive sense of self. When parents repeatedly push and challenge their children without expressing acceptance, children might feel criticized and come to believe that their parents push them because they are simply not good enough.
Beyond the main effects, the combination of high acceptance and challenge is argued to be related to the highest level of self-concept. Children who feel both accepted and pushed to improve should feel worthy and valued as well as competent. Acceptance should facilitate feeling valued and cared for, and challenge should induce behaviors to explore different ideas and experiences and improve their abilities (Dailey, 2010). Evidence generally suggests the benefit of the combination of higher levels of acceptance and challenge. Although Dailey (2010) reported no evidence of an interaction between parental acceptance and challenge in predicting adolescents’ self-concepts, siblings’ acceptance and challenge interacted in an expected manner. Dailey (2008) further showed that acceptance and challenge interacted to predict college students’ identity strength even after controlling for the individual effects of acceptance and challenge. Therefore, there is a good theoretical rationale and empirical evidence to expect an interactive effect between acceptance and challenge.
The proposed model hypothesizes that parental confirmation is related to the self-concept, which, in turn, is associated with body image. Those who lack a strong sense of identity might be more inclined to turn to body weight to compensate for the associated feelings of powerlessness (Bruch, 1982). These individuals might rely on socially subscribed standards of attractiveness in an effort to search for selfhood. In fact, those with a weak sense of identity are more likely to internalize such societal standards of attractiveness and/or engage in appearance-related social comparison (Vartanian & Dey, 2013), perhaps in an effort to define the self. In addition to the link between self-identity and body image, the association between self-esteem and body image is also well established among individuals with a wider variety of developmental stages (O’Dea, 2012), including college students (Tylka & Sabik, 2010). Low self-esteem might lead individuals to view themselves, including their bodies, more negatively. Therefore, poor self-concept (i.e., a weak self-identity and low self-esteem) might place individuals at risk of negative body image.
In line with the proposed model, at least two studies found that self-concept mediated the association between parental behaviors and body image as well as eating problems among children including college-aged ones (Cella, Iannaccone, & Cotrufo, 2014; Perry, Silvera, Neilands, Rosenvinge, & Hanssen, 2008). Based on the reasoning detailed previously, the combination of high acceptance and high challenge is argued to be related to the greatest degree of self-concept, which, in turn, will be related to the most positive body image compared with other combinations of acceptance and challenge. As such, the following hypothesis is posed:
Parental and Child Sex Differences
As some family scholars contend (Cowan, Cowan, & Kerig, 1993), family processes can be better understood by specifying the sex of the parental figure, as well as the sex of the child. It is important to consider each parent’s unique contribution to the family environment for at least two reasons. First, children’s perceptions of their interactions with fathers and those with mothers often differ, with mother often being perceived as more accepting and supportive than fathers (Boutelle et al., 2009). Second, interactions with mothers and with fathers may have distinctive influence on children’s development. The literature has produced somewhat inconsistent findings regarding the effect of parent sex on children’s outcomes in terms of body image (e.g., Ellis, 2002), social competence (Laible & Carlo, 2004), and self-concept (Dailey, 2010). Some studies highlighted the importance of mothers in children’s development in these domains (Beato-Fernández, Rodriguez-Cano, Belmonte-Llario, & Martinez-Delgado, 2004), whereas other studies demonstrated a unique contribution of fathers to these same domains (e.g., Lam, McHale, & Crouter, 2012).
Although inconsistency exists, the literature points to the possibility that mothers and fathers contribute to their children’s development differently. It is also important to take child sex into consideration because daughters and sons might respond differently to parental behaviors. Prior research indicated that, compared with sons, daughters are more strongly affected by (perceptions of) parental behaviors in terms of their body image (Crespo, Kielpikowski, Jose, & Pryor, 2010) and social competence (Hsu, 1998). It is less clear whether child sex moderates parental effects on children’s self-concept (identity strength and self-esteem). Some studies suggested parental behaviors affect sons’ and daughters’ self-concept in a different way (e.g., McKinney, Milone, & Renk, 2011), whereas other studies found no such differences by child sex (Allen, Hauser, Bell, & O’Connor, 1994). The reason that daughters might be generally more sensitive to familial communication quality than sons could be related to gender socialization. Women are often socialized to value connectedness and to feel responsible for the emotional climate of relationships (Nolen-Hoeksema & Jackson, 2001). Women might be more likely than men to attribute interpersonal difficulties, including those with parents, to the self.
Considering the paucity of prior confirmation research using sex-specific parent-child combinations, and the mixed findings of existing developmental literature, this project posits the following research question:
Method
Participants and Procedures
On receiving approval from the university’s institutional review board, undergraduate students enrolled in communication courses at a large Southwestern University were recruited to participate in this study. This was a 30-minute online survey hosted on Qualtrics. Participants were offered extra credit for their participation. Four hundred forty-seven participants (age M = 20.20 years, SD = 1.38 years, range = 18-25 years) participated in the study. The majority of the participants (71%) were women. The participants were largely Caucasian (60%), followed by Latino(a) (20%), Asian (13%), and African American (3%). Most of the participants (91%) were currently living away from home.
Measures
Unless otherwise stated, all measures used a 7-point scale ranging from 1 = strongly disagree to 7 = strongly agree. The order of the scales was randomized to minimize order effects. See online appendices for reliabilities, means, standard deviations, and correlations among the variables.
Parental confirmation
Although acceptance and challenge can be domain specific (e.g., see Dailey, McCracken, & Romo, 2011), the current study assessed the general nature of acceptance and challenge. Perceptions of parental acceptance were assessed by eight items adapted from Schaefer’s (1965) Parental Support subscale of the Child Report of Parent Behavior Inventory. Whereas Schaefer’s original items use the present tense (“My mother [father] smiles at me often”), all items were changed to reflect the past tense (“My mother[father] smiled at me often”). Perceptions of parental challenge were assessed by Dailey’s (2008) 10-item Parental Challenge Questionnaire (“My mother [father] helped me channel my negative emotions into more positive actions”). For both acceptance and challenge measures, participants were asked to respond to the same measure twice, once for mothers and another for fathers. In the case that participants had a stepparent, they were asked to choose the parental figure (e.g., biological mother or stepmother) who has been most influential in their development. The vast majority of participants reported on their biological mother (96%) and biological father (90%).
Body image
The Body-Esteem Scale for Adolescents and Adults (Mendelson, Mendelson, & White, 2001) was used to assess participants’ attitudes and feelings regarding their bodies and appearance. This scale was administered for both men and women, as the content applied to both sexes. Participants’ body image was indicated by three subscales: Appearance Satisfaction, Weight Satisfaction, and Attribution (one’s perceptions of how others evaluate his or her body and appearance). This study used a 6-point scale ranging from 1 = never to 6 = always.
Social competence
A latent factor of social competence was indicated by communication apprehension, positive relationship with others, and self-rated interpersonal competence. Communication apprehension was measured by McCroskey’s (1982) Dyad Subscale of Personal Report Communication Apprehension. Ability to establish positive and supportive relationships with others was assessed by the Positive Relationships With Others subscale from Ryff’s (1989) Scales of Psychological Well-Being. Participant perceptions of their competence in same-sex friendships were assessed by the Interpersonal Competence Questionnaire (ICQ; Buhrmester, Furman, Wittenberg, & Reis, 1988), which consists of five subscales: Initiating Relationships, Asserting Displeasure, Self-Disclosure, Providing Emotional Support and Advice, and Conflict Management. Because each subscale assessed different domains of interpersonal competence, a latent factor of the ICQ was indicated by five subscales. Participants reported how comfortable they are at handling each situation using a 5-point rating scale, ranging from 1 = poor at this to 5 = extremely good at this.
Self-concept
A latent factor of self-concept was indicated by identity strength and self-esteem. Participants’ identity strength was assessed by the 12-item Identity subscale of the Erikson Psychosocial Inventory (EPSI) Scale (Rosenthal, Gurney, & Moore, 1981). Self-esteem was measured by the 10-item Rosenberg Self-Esteem Scale (Rosenberg, 1965).
Preliminary Analyses and Data Analyses Overview
Because the data were nonnormal, maximum likelihood robust (MLR) estimators, which are robust to nonnormality of the data (Muthén & Asparouhov, 2002), were used for analyses. Rather than using the averaged scores of all items in each scale, the present study used factor scores because it reduced the number of parameters that had to be estimated in structural analysis. Moreover, using factor scores, rather than using observed scores (e.g., averaging items), is advantageous, in that, measurement error is “removed” (Hershberger, 2005). Confirmatory factor analyses (CFAs) for each measure were conducted for the purpose of creating factor scores (see the online appendix for fit indices of CFAs for each measure).
For all measurement and structural models, four common fit indices were used to assess model fit: (a) the relative chi-square (χ2/df) of 3 or less (Kline, 2004), (b) a root mean square error of approximation (RMSEA) as well as standardized root mean square residual (SRMR) of 0.08 or less, and (c) a comparative fit index (CFI) of 0.9 or greater (Browne & Cudeck, 1993; Byrne, 2012; Hu & Bentler, 1999). The relative chi-square is used here because the chi-square test is sensitive to sample size (Bentler & Bonett, 1980); with a large sample size, chi-square values are inflated (statistically significant), erroneously implying poor model fit to the data (Hoe, 2008).
All hypotheses and the research question were tested and examined within a proposed model. Structural equation modeling (SEM) was performed through Mplus 7.0 (Muthén & Muthén, 1998-2017) with the MLR estimator. The full information maximum likelihood (FIML) was used to estimate the missing data. When deciding which model to retain between a nested model and a comparison model (i.e., a more parameterized model), a chi-square difference test with the Satorra-Bentler scaled chi-square (SBχ2; Satorra & Bentler, 1988) was used. To examine whether there were differences in the hypothesized links between men and women, a multiple-group SEM was conducted. A summary of model fit for multigroup invariance can be found in Table 1.
Summary of Model Fit for Baseline Model and Multigroup Invariance.
Note. SBχ2 = Satorra-Bentler rescaled χ2; CFI = comparative fit index; TLI = Tucker-Lewis index; SRMR = standardized root mean square residual; RMSEA = root mean square error of approximation; CI = confidence interval.
Results
Baseline Models for Men and Women
A two-step procedure was used to establish a baseline model for each group before testing for multigroup invariance. The original measurement model for men demonstrated an inadequate fit, χ2(78) = 177.87, p < .01, χ2/df = 2.29, RMSEA = 0.10 (90% confidence interval [CI] = [0.08, 0.12]), CFI = 0.93, SRMR = 0.04. Post hoc model modifications were performed on the bases of the Lagrange Multiplier (LM) Test. Based on the recommendation for respecification of the model (Byrne, 2012), only one new parameter at a time was included. Ultimately, three error covariances were added in total. 1 This improved the fit to the data, χ2(75) = 128.81, p < .001, χ2/df = 1.72, RMSEA = 0.075 (90% CI = [0.05, 0.10]), CFI = 0.95, SRMR = 0.04.
Having established the overall soundness of the model, the hypothesized structural model was examined. The hypothesized structural model demonstrated a poor fit, χ2(86) = 183.76, p < .001, χ2/df = 2.14, RMSEA = 0.10 (90% CI = [0.08, 0.13]), CFI = 0.91, SRMR = 0.11. After the inspection of the LM Test, a correlation between the error terms of two latent variables (social competence and self-concept) was added. The resulting model improved the fit to the data, and was used as a baseline model for men.
The original measurement model for women demonstrated adequate fit to the data, χ2(78) = 230.73, p < .01, χ2/df = 2.96, RMSEA = 0.08 (90% CI = [0.07, 0.09]), CFI = 0.94, SRMR = 0.05. Furthermore, the original structural model produced poor fit, χ2(89) = 290.69, p < .01, χ2/df = 3.27, RMSEA = 0.10 (90% CI = [0.09, 0.11]), CFI = 0.87, SRMR = 0.11. An error correlation between social competence and self-concept was added based on the LM Test, which improved the model fit. The resulting model was used as a baseline model for women.
Configural Invariance
Having established separate baseline models that reflect a satisfactory good fit to the data for each group, a configural model was examined. No parameters were constrained to be equal across men and women. The model, therefore, represents the least constrained model possible, and it fit the data well.
Structural Invariance
Because the current project’s focus is to test whether hypothesized structural regression paths might differ depending on sex-specific parent-child dyads, an invariance test was conducted only in terms of structural paths. The configural model established earlier was compared with a structural invariance model that constrained all structural paths to be equal for men and women. The model in which all structural paths were constrained to be equal across groups fit significantly worse than the original freely estimated model, ΔSBχ2 = 55.00, df = 36, p < .05.
The LM Test indicated the improvement of fit by releasing only one constraint on a structural path from social competence to body esteem. A resulting model with one parameter freely estimated demonstrated a significantly better fit to data compared with the constrained model, ΔSBχ2 = 17.47, df = 1, p < .001. This model with one parameter freely estimated did not significantly differ from the configural model, ΔSBχ2 = 43.66, df = 35, p > .05. Therefore, this model was retained and used to test all hypotheses and answer the research question.
Examining Hypotheses and Research Question
Table 2 describes the summary of the tests of hypothesized indirect effects that are statistically significant. Figure 2 represents a final structural model.
Tests of Indirect Effects in Structural Model.
Note. B = unstandardized indirect effect; β = standardized indirect effect; CI = confidence interval; M Acc = mother acceptance; Social = social competence; BE-Att = Body esteen-attribution; Self = self-concept; BE-App = Body esteem-appearance; F Acc = father acceptance; F Chall = father challenge; M Chall = mother challenge.
p < .05. **p < .01. ***p < .001. Only statistically significant indirect effects were presented in the table.

Final structural model with standardized regression coefficients and percentage of variance explained.
A mediating role of social competence
Of six exogenous variables, only mother acceptance (β = .34, p < .01, for men; β = .31, p < .01, for women) and mother acceptance × challenge (β = .26, p < .01, for men; β = .21, p < .01, for women) were significant predictors of social competence. The latent variable of social competence, in turn, was a significant predictor of BE-Appearance (β = .24, p < .05, men only), BE-Attribution (β = .34, p < .001, for men; β = .35, p < .001, for women), but not BE-Weight.
The nature of the interaction was probed by the LOOP option in Mplus, which uses a method called the Johnson-Neyman (J-N) technique (Johnson & Neyman, 1936). This method allows for identifying the values of a moderator under which the simple slope is statistically significant. The use of the J-N technique is recommended over the traditional methods of splitting the data into categories (e.g., high vs. low) to probe interaction effects (Preacher, Curran, & Bauer, 2006; Figure 3). 2 Two different interaction figures (in which the predictor and the moderator were switched) were plotted for each significant interaction effect throughout this study to fully understand the nature of interaction effects. 3 For this interaction, the positive effect of mother acceptance on social competence was enhanced as the degree of mother challenge increased. The simple slope of social competence regressed on mother acceptance was significant for the vast majority of participants (those who scored in the top 88% on mother challenge). When mother acceptance was plotted as a moderator, the effect of mother challenge on social competence was enhanced as mother acceptance increased. However, the simple slope of social competence regressed on mother challenge was significant only for a small number of participants (i.e., those who scored the top 7% on acceptance).

Interaction effect between acceptance and challenge predicting outcomes.
To determine whether there are significant indirect effects of parental confirmation on body image through social competence, bootstrapping analyses were conducted to obtain standard error estimates, 95% CIs, and effect size estimates with 5,000 samples (Hayes, 2009). One significant indirect effect emerged from these analyses (see Table 2). Mother acceptance had a significant indirect effect on BE-Attribution through social competence for both men (β = .12, p < .05) and women (β = .11, p < .05). Indirect effects of mother acceptance
A mediating role of self-concept
Of six exogenous variables, mother acceptance (β = .50, p < .001, for men; β = .38, p < .001, for women), mother acceptance × challenge (β = .31, p < .01, for men; β = .21, p < .01, for women), and father acceptance were significant predictors of self-concept (β = .21, p < .05, for men; β = .20, p < .05, for women). Probing of the nature of the interaction between mother acceptance and mother challenge indicated that, consistent with the prediction, the effect of mother acceptance on self-concept was enhanced when it was combined with a higher level of mother challenge, and the simple slope was significant for the vast majority of participants (i.e., those who scored in the top 93% of mother acceptance). However, there was a negative effect of mother challenge on self-concept for those who scored extremely low (the bottom 6%) on mother acceptance. For these participants, the greater mother challenge they reported, the lower self-concept they reported.
The self-concept latent variable, in turn, was a significant predictor of all endogenous variables: BE-Appearance (β = .47, p < .001, for men; β = .59, p < .001, for women), BE-Attribution (β = .24, p < .01, for men; β = .30, p < .01, for women), and BE-Weight (β = .38, p < .001, for men; β = .41, p < .001, for women).
Several significant indirect effects emerged from bootstrapping analyses. The indirect effects of mother acceptance on all three subscales of body image through self-concept were significant: BE-Appearance (β = .23, p < .01, for men; β = .22, p < .01, for women), BE-Attribution (β = .12, p < .05, for men; β = .11, p < .05, for women), and BE-Weight (β = .19, p < .01, for men; β = .15, p < .01, for women). Similarly, there were significant indirect effects of father acceptance on BE-Appearance (β = .10, p < .01, for men; β = .12, p < .01, for women) and BE-Weight (B = .06, β = .08, p < .05, for both men and women) through self-concept.
Furthermore, there was a significant indirect effect of mother acceptance × mother challenge on BE-Appearance (β = .15, p < .01, for men; β = .13, p < .01, for women) and BE-Weight (β = .12, p < .05, for men; β = .09, p < .05, for women) through self-concept. The indirect effect of mother acceptance × mother challenge on BE-Attribution through self-concept only approached significance, β = .07, p = .08, for men, and β = .06, p = .08, for women. Therefore, H2b, which hypothesized a mediating role of self-concept between the interaction of acceptance and challenge and body image was partially supported. This result logically provides partial support for Hypothesis 1, which predicted an association between the interaction of acceptance and challenge and body image.
Direct effects of father acceptance × father challenge on body image
Father acceptance × father challenge significantly predicted BE-Appearance (β = .09, p < .01, for men; β = .14, p < .01, for women) and BE-Weight (β = .13, p < .05, for men; β = .17, p < .05, for women) without being mediated by hypothesized latent variables. The nature of these interaction effects was consistent with Hypothesis 1. The positive effect of father acceptance on BE-Appearance was enhanced as father challenge increased. The simple slope of father acceptance regressed on BE-Appearance was significant for those who scored high (in the top 35%) on father challenge. Furthermore, there was a negative impact of challenge in the presence of no to little acceptance. For participants who reported low (the bottom 7.6%) on father acceptance, the greater father challenge, the less appearance satisfaction they reported. In addition, the positive effect of father acceptance on BE-Weight was enhanced as the level of father challenge increased. The simple slope of BE-Weight regressed on father acceptance was significant for participants who scored high (in the top 9%) on father challenge. Regardless of the level of father acceptance (i.e., for all participants), the simple slope of weight satisfaction on father challenge was not significant. This result reinforces the partial support of Hypothesis 1.
Difference by sex composition of parent-child dyad
Research Question 1 asked how, if at all, the hypothesized associations vary depending on the sex composition of parent-child dyad. The results of the multigroup analyses revealed no significant child sex difference for each hypothesized link except for one path between social competence and BE-Appearance, which did not involve parental confirmation. This suggests that how acceptance and challenge were associated (or not associated) with children’s outcomes did not differ by children’s sex. There were, however, at least three key differences between mothers’ and fathers’ contributions in predicting hypothesized mediators and body image as detailed above. First, mothers’ behaviors (acceptance and acceptance
Supplemental Analysis: An Alternative Model
The model in the current study is based on the causal inferences that parental confirmation influences social competence and self-concept, which in turn, influences body image. However, the cross-sectional nature of this study does not allow us to establish the hypothesized causality. One can argue that parental confirmation might affect body image, and those who feel good about their bodies might subsequently feel better about social situations and interpersonal relationships. This can further facilitate social interactions that can result in the development of social competence. Similarly, those who feel good about their bodies might subsequently feel good about themselves and establish a strong (and positive) sense of self. To explore this possibility, an alternative model was run in which the direction of paths from the hypothesized mediators (i.e., social competence and self-concept) to endogenous variables (BE-Appearance, BE-Attribution, and BE-Weight) was switched.
The fit of the alternative model is Akaike information criterion (AIC) = 2,650.20, χ2(88) = 378.40, p < .001, χ2/df = 4.30, RMSEA = 0.18 (90% CI = [0.16, 0.20]), CFI = 0.72, SRMR = 0.15, for men, and AIC = 5,924.39, χ2(91) = 669.85, p < .001, χ2/df = 7.36, RMSEA = 0.16 (90% CI = [0.15, 0.17]), CFI = 0.62, SRMR = 0.13, for women. Path coefficients for the alternative model are available in the online appendix. The AIC of the hypothesized model was 2,474.57 and 5,505.94 for men and women, respectively. Based on the AIC, which is recommended to use when comparing two or more unnested modes (Byrne, 2012), the results support that the hypothesized model is a better fitting model.
Discussion
Building on confirmation theory (Dailey, 2010), the current study sought to examine the effects of parental confirmation on children’s body image through social competence and self-concept. Furthermore, this study tested how the hypothesized associations differ depending on the sex composition of parent-child dyads. The results generally supported the hypothesized model, yet provided a somewhat complex picture of associations among key variables: How each component of confirmation (acceptance and challenge) was related to certain components of body image and through which mediator varied depending on parental sex. Unlike parental sex, child sex played a minimal role in the current study. This suggests that the model operates similarly for both sons and daughters.
Parental Confirmation
Mother acceptance, and not father acceptance, was positively related to social competence as well as self-concept, suggesting that warm and supporting maternal behaviors are conducive to the children’s development of ability to relate to others as well as how they see themselves. The data further showed a negative association between parental challenge and social competence, self-concept, and body image in the presence of little or no acceptance. Such a compromised effect of high challenge in the absence of acceptance should not be particularly surprising: Challenge does not always have to be enjoyable or pleasant experiences (Dailey, 2008). After all, challenging experiences provide opportunity for growth and development in skills, but do not necessarily result in the advancement of abilities (Dailey, 2008). The negative result of challenge in the absence of acceptance might be because children interpret challenging messages differently depending on the degree to which they experience warmth and support (Rathunde, 1996). Without feeling loved and supported, children might interpret challenging messages as criticism motivated by parental frustration and disapproval.
There are some notable differences between mothers and fathers regarding their roles in their children’s social competence and self-concept. Mother acceptance was associated with social competence as well as self-concept. Although mother challenge did not have main effects on any key variables, it played an important role by enhancing the positive effect of mother acceptance on social competence and self-concept. This parallels previous confirmation research demonstrating these interaction effects (Dailey, 2010). Unlike mothers, neither father acceptance nor father challenge was associated with social competence. The reason for this parental sex difference might be because mothers, compared with fathers, spend more time communicating with their children (Milkie, Raley, & Bianchi, 2009), allowing mothers more opportunities to influence their children’s social competence. However, it would be a mistake to conclude that father behaviors are not important to children’s social competence. The reason why neither father acceptance nor father challenge was related to social competence might be due to the high correlations between mother acceptance and father acceptance and between mother challenge and father challenge.
Another key difference between mothers and fathers involves how parental confirmation was related to body image. Whereas the association between mother confirmation and body image was mediated by social competence and self-concept, father confirmation (interaction between father acceptance and father challenge) directly contributed to the variance in body image without being mediated by social competence and self-concept. These results suggest that having fathers who are both highly accepting and challenging might contribute to a positive body image. However, having fathers who are highly challenging without showing warmth and support might be associated with the development of a poor body image.
It remains unclear as to why the interaction between father acceptance and challenge was related directly to body image without being mediated by the hypothesized mediators. A somewhat similar pattern was reported in previous research (Taniguchi & Thompson, 2015), which found that mother communication patterns were associated with college-aged daughters’ body image through social competence, whereas father communication patterns were directly associated with body image without being mediated by social competence. There seems to be something unique about father communication in how it is related to body image. There might be other mediators that were not captured by the current model. Future research might benefit from examining such potential mediators (e.g., self-efficacy) in an effort to identify the mechanisms of the link between father communication and body image.
Regardless of parental sex differences, one general conclusion that can be drawn from the findings is that recollections of parental communication experienced over years are related to emerging adult children’s body image. Yet, why does parental communication matter to emerging adult children, the majority of whom had already left home as is the case with this study? Interestingly, the majority of emerging adults report that not being “deeply tied to parents emotionally” is not necessarily a prerequisite to be considered an adult (Arnett, 2001, p. 137), which suggests a continuing bond with parents. Therefore, it is likely that parental confirmation experienced during formative years continues to be an important source of evaluation and formation of self for individuals who are transitioning into adulthood. This speculation is in line with previous research that has demonstrated how parents remain a source of self-evaluation and body image evaluation for emerging adult children (Arroyo & Segrin, 2013; Ellis, 2002).
Mediating Role of Social Competence and Self-Concept
Despite a high correlation between the social competence and self-concept factors (r = .61 for men and r = .70 for women), both social competence and self-concept functioned as mediators that explained the link between parental confirmation and body image, demonstrating unique contributions made by each mediator. Components of mothers’ confirmation were related to social competence, which, in turn, predicted positive body image. This is consistent with the idea that dysfunctional family dynamics can hinder the development of social competencies that are necessary to develop and maintain meaningful relationships, and poor social competence can increase the vulnerability to body image issues (Mallinckrodt et al., 1995). The current study provides the first evidence that a similar mechanism is operating among men as well.
In addition to social competence, self-concept also played a mediating role in the current model. Components of confirmation from both mothers and fathers were related to self-concept, which, in turn, was positively related to all three domains of body image. This finding suggests that highly accepting and challenging parents facilitate their children’s development of a strong and positive sense of self, and these children tend to possess a positive body image. This result parallels previous research showing that problems with self-concept (e.g., weak identity, poor self-esteem) often co-occur with body image problems (e.g., Vartanian & Dey, 2013). Individuals with poor self-concept might be more vulnerable to sociocultural messages promoting thinness or muscularity. These individuals might be more likely to internalize socially subscribed standards of attractiveness (Vartanian & Dey, 2013), which predicts body dissatisfaction (Thompson & Stice, 2001).
Overall, compared with social competence, self-concept might be a more stable or consistent predictor of body image. Social competence was associated with only one (for women) or two (for men) of the subscales of body image, whereas self-concept was associated with all three subscales of body image. This might be because self-concept is a broader concept compared with social competence, which is a narrower construct specific to social contexts. Due to the global nature of the self-concept construct, it might have a greater capacity to predict outcomes in a wider variety of areas, including body image.
Practical Implications
Regardless of parental sex, confirming parental behaviors seem most conducive to the development of children’s positive body image, even though the specific mechanism might vary by parental sex. Therefore, parents could be advised to employ confirming messages that validate children’s self-expressions and acknowledge them as unique, worthy, and capable individuals. It seems particularly important for parents to be aware of the potential disadvantage of employing challenging behaviors without displaying warmth and support. Such parental behaviors could potentially lead to negative child outcomes (negative self-concept and body image) regardless of the parents’ intention.
Practitioners might consider incorporating approaches to improving social competence and self-concept into intervention and prevention programs for body image issues. Although family interactions already experienced in the past cannot be altered, social competence and self-concept could be still altered. As individuals leave home when entering college, some of them experience loneliness, isolation, and low support availability (Larose & Boivin, 1998), making them more susceptible to developing a negative body image and eating problems (Budd, 2007). Helping high school children build social competence prior to making the transition into college might have preventive value. Finally, the current results support the utility of intervention and prevention program focusing on the improvement of self-concept (see O’Dea, 2004, for review).
Limitations and Future Directions
The current study offers some important limitations and future directions. First, parental confirmation in this study was assessed by children’s self-report. Because children’s perceptions of family interactions might have a greater impact on their current sociopsychological situations than the “objective” reality (Feeney & Noller, 2013), the use of self-report data is still meaningful. However, considering most participants live apart from parents, such self-report measures might be affected by memory bias. Measuring perceptions of parental confirmation through daily diary entries written by children who currently live with their parents would be valuable, as such measurement would be less vulnerable to memory bias. Future research might benefit from collecting observational data of actual parent-child interactions as such data provide insight into how each component of confirmation is operative between parents and children.
Second, even though the current study’s model suggests a certain causal direction, causal claims cannot be made due to the correlational nature of the data. Although the current study showed that the original model fits better than the alternative model in which body image was modeled as a predictor of social competence and self-concept, this is not strong enough evidence to claim causality. Future research would benefit from conducting longitudinal research to establish a stronger basis for causality. For instance, tracking family interaction for a period of years would be useful for examining how changes in parental confirmation can predict changes in self-concept and body image. Furthermore, it would also be beneficial to conduct an intervention type of research in which parents in an experimental condition receive training to communicate in a confirming manner to their children. Researchers can then examine whether children’s self-concept as well as their body image differ as a function of whether their parents were assigned to an experimental or a control condition.
Third, the current sample was collected from undergraduate students, the majority of whom are from families with middle to high socioeconomic status (SES; Stuart, 2010). High-SES backgrounds might partially explain why most participants reported high levels of acceptance and challenge. Considerable evidence exists that parenting style varies as a function of SES (Hoff, Laursen, & Tardif, 2002). Future research might benefit from using a more socioeconomically diverse sample, which would allow for more normally distributed data for acceptance and challenge. With greater variability in acceptance and challenge, such future research might be able to detect more salient effects of parental confirmation on child outcomes.
Conclusion
Building on confirmation theory (Dailey, 2010), this study has revealed important mechanisms linking parental communication with emerging adults’ body image: social competence and self-concept. This study has also provided insight into the unique roles of mothers and fathers, highlighting the importance of differentiating parental sex when examining parental contribution to their children’s body image. Although multiple disciplines should continue to identify factors influencing body image, this study underscores the value of applying a communication perspective, namely, confirmation theory, to researching body image. Our findings address a need for additional research on parental confirmation and emerging adult children’s body image that employs longitudinal designs to establish a stronger basis for causality.
Supplemental Material
Online_appendices – Supplemental material for Parental Confirmation and Emerging Adult Children’s Body Image: Self-Concept and Social Competence as Mediators
Supplemental material, Online_appendices for Parental Confirmation and Emerging Adult Children’s Body Image: Self-Concept and Social Competence as Mediators by Emiko Taniguchi and René M. Dailey in Communication Research
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Notes
Author Biographies
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
