|Home | About | Journals | Submit | Contact Us | Français|
The parent-child relationship is often characterized by ambivalence, defined as the simultaneous experience of positive and negative relationship quality. This study examines reports of intergenerational ambivalence in three developmental periods: adolescence, emerging adulthood, and young adulthood, as well as its implications for depressive symptoms over a 12-year period. Participants aged 13 to 29 (n = 255) were interviewed in 1992 and again in 2005 at ages 25 to 41 (n = 186). Results indicate that offspring's reports of intergenerational ambivalence decreased over time. Greater ambivalence towards mothers predicted increased depressive symptoms over time while greater ambivalence towards fathers predicted decreased depressive symptoms over time. These results suggest that depressive symptoms in adulthood are vulnerable to the quality of the parent-child relationship earlier in the life course.
The parent-child tie is one of the longest lasting and most emotionally intense ties across the lifespan (Fingerman, Sechrist, & Birditt, 2013). Parents and their children often report feeling simultaneously positive and negative towards one another, i.e., ambivalent (Birditt, Miller, Fingerman, & Lefkowitz, 2009). Theorists suggest that parent-child ties tend to be more ambivalent than other close ties due to conflicting societal norms of both closeness and independence (Merton & Barber, 1963; Pillemer & Suitor, 2005). In addition, ambivalent ties are thought to be more harmful than solely negative ties because these relationships tend to be unpredictable (Uchino, Holt-Lunstad, Uno, Campo, & Reblin, 2007). Intergenerational ambivalence among adults predicts greater depressive symptoms, poorer health, and lower quality of life (Fingerman, Chen, Hay, Cichy, & Lefkowitz, 2006; Fingerman, Pitzer, Lefkowitz, Birditt, & Mroczek, 2008; Silverstein & Bengston, 1991; Willson, Shuey, Elder, & Wickrama, 2006). The majority of research, however, focuses on intergenerational ambivalence in middle and older adulthood rather than ambivalence among adolescents or young adults (Fingerman et al., 2008; Kiecolt, Blieszner, & Savla, 2011; Lendon, Silverstein, & Giarrusso, 2014; Lowenstein, 2007; Ward, 2008), a period in which the issues of independence and closeness may be particularly salient developmentally.
Researchers have yet to examine whether feelings of ambivalence exist at younger ages, such as adolescence, and whether these feelings change over time among younger age groups. The present study examines reports of ambivalence towards mothers and fathers among offspring in three early developmental periods: adolescence, emerging adulthood, and young adulthood and then again after 12 years, in later developmental periods. This study is one of the few to examine parent-child ties over the transition from these younger ages to adulthood and midlife. The purpose of the present study is threefold: (a) examine whether reports of intergenerational ambivalence vary by age and developmental period, (b) investigate whether intergenerational ambivalence increases or decreases over time, and (c) examine whether intergenerational ambivalence is associated with offspring's depressive symptomatology.
Ambivalence theory and individuation theory provide useful frameworks for understanding the complexity of the parent-child relationship and the development of the relationship over time. This study assesses psychological ambivalence, which is defined as the simultaneous experience of positive and negative emotions towards the same person (Luescher & Pillemer, 1998). Psychological ambivalence originates from contradictions in motivations, emotions, and cognition (Luescher & Pillemer, 1998; Weigert, 1991). Here, ambivalence, or ambivalent relationship quality, is conceptualized as a combination of both positive and negative components of the parent-child tie (Connidis, 2015).
The parent-child tie is particularly ambivalent because parents and their children experience competing desires and social norms that encourage independence as well as closeness. Thus, ambivalence tends to be higher when parents and children are not able to maintain an appropriate balance of independence and closeness. For example, adult children and parents tend to report greater ambivalence when either parents or children have health problems (Birditt, Fingerman, & Zarit, 2010; Fingerman et al., 2006; Willson, Shuey, & Elder, 2003). Parents also tend to experience less ambivalence toward their adult children when their children have attained social roles associated with independence from the parent-child tie including marriage and employment (Fingerman et al., 2006; Pillemer, Suitor, Mock, Sabir, Pardo, & Sechrist, 2007). Intergenerational ambivalence research is often focused on the ties between parents and their adult children (i.e., children older than 18 years old). However, ambivalence may be even higher among adolescents and emerging and young adults who are in the developmental periods when they are just beginning to seek independence from parents.
Ambivalence is distinct from conflict which has commonly been studied in adolescence and young adulthood. In contrast to conflict, which is a solely negative dimension, ambivalence implies the presence of both positive and negative relationship characteristics. It is well known that parent-child ties are complex. Studying ambivalence, rather than solely positive or negative attributes, allows researchers to more critically and accurately describe familial relationships (Birditt et al., 2010; Connidis, 2015; Pillemer, Munsch, Fuller-Rowell, Riffin, & Suitor, 2012). Ambivalence studies have typically not considered a developmental perspective, i.e., how ambivalent feelings may change as people grow and develop over time.
Individuation theory provides a framework to understand possible developmental changes in ambivalence. As adolescents’ age, experience puberty, and expand cognitively and socially, they develop a separate sense of individual identity, become autonomous, and take increasing responsibility for their actions (Aquilino & Supple, 2001; Blos, 1967; Holmbeck, Paikoff, & Brooks-Gunn, 1995; Smetana, 1988; Steinberg, 1999). Rather than relying on parents, individuation involves letting go of childhood dependencies and seeking more mature, less dependent relationships (Steinberg, 1999). Successful individuation leads to positive outcomes such as ego development, decision-making responsibility, higher self-esteem, and lower depressive symptoms (Grotevant & Cooper, 1986; Peterson, Bush, & Supple, 1999). Thus, ambivalence may be highest in adolescence when parents and children are navigating relationship changes and may possibly decrease over time when these changes are realized.
Ambivalence may originate early in the parent-child relationship and change over time as offspring age and achieve additional developmental milestones. This study focuses on three developmental phases: adolescence (ages 13 to 17), emerging adulthood (ages 18 to 25), and young adulthood (ages 26 to 29) and follows the offspring as they transition into emerging adulthood, young adulthood, and midlife, respectively. Few studies have examined offspring's perceptions of parents and well-being in these particular groups (Aquilino, 1997, 2006b). Due to different levels of independence, we anticipate that each age group will experience varying levels of ambivalent relationship quality with their parents.
Ambivalence may be particularly high in adolescence compared to emerging and young adulthood. Adolescence is a particularly eventful developmental period. During adolescence, individuals experience important biological and cognitive maturation processes coupled with increased interaction with peers (Bucx & van Wei, 2008; Eccles, Templeton, Barber, & Stone, 2003; Steinberg, 1999). Adolescents also gain a more important role in the family, are expected to take more responsibility for themselves and their actions, and strive for autonomy (Blos, 1967; Bucx & van Wei, 2008; Grotevant & Cooper, 1986; Steinberg, 1999). These developmental changes often affect the parent-child relationship and create an imbalance in the family as it adjusts to the changes (Steinberg, 1999). Often adolescents’ perceptions of the quality of the parent-child tie declines while conflict increases (McGue, Elkins, Walden, & Iacono, 2005). The experience of continued dependence on parents, coupled with new needs for independence may conflict with adolescents’ developmental needs, thus engendering high levels of ambivalence regarding parents during this stage.
Emerging adults may experience less ambivalence than adolescents but more ambivalence than young adults. Individuals in late adolescence and early adulthood do not usually have the tumultuous feelings of adolescents, but still do not have the freedom and independence of young adults. Many young people do not regard themselves as fully adult until they reach their later twenties, which suggests that there is a developmental life stage between adolescence and young adulthood sometimes referred to as “emerging adulthood” (Arnett, 2001b). During this period, at least among more affluent families, emerging adults are less constrained by role requirements and responsibilities and are in a stage of exploration (Aquilino, 2006a; Arnett, 2000, 2001a; Eccles et al., 2003; Rindfuss, 1991). While emerging adults are increasingly independent, they may remain somewhat dependent on parents, which may be a cause of ambivalence as they navigate this developmental stage (Arnett, 2000).
Young adults may experience less ambivalence than adolescents and emerging adults due to increased independence from parents. During the shift into young adulthood, offspring learn to take on adult responsibilities such as leaving the parental home, beginning employment and financial independence, entering marriage, or becoming a parent (Noack & Buhl, 2004). By the end of their twenties, most young people have already made significant decisions about their future including choosing career paths and entering into adult relationships (Arnett, 2000). With these life transitions, young adults become increasingly less dependent on parents and the parent-child tie may become less focal (Bucx & van Wei, 2008). Fingerman and colleagues (2006) found that adult children who invested in several roles (i.e., parent, spouse, worker) experienced less parent-child ambivalence if they were highly invested in those roles. Other research indicates that autonomy in the parent-offspring relationship may deter ambivalence as both parents and offspring have roles and outlets independent of the tie (Fingerman, 2001; Pillemer & Suitor, 2005). This developmental stage may be less conflictual as parental behavior and acquisition of adult, independent roles by offspring leading to less ambivalent feelings than previous phases.
The present study is unique because we not only examine age differences in ambivalence, but also assess whether individual reports of ambivalence change over 12 years. This longitudinal design allowed us to examine within person changes as individuals transition to older stages of life such as middle adulthood. We predicted adolescents would report the highest levels of ambivalence and that the greatest declines in ambivalence would occur among adolescents (during the transition from adolescence to young adulthood) compared to changes among the emerging adults and young adults.
Decreases in ambivalence over time may be linked to offspring's well-being as they develop over time. Fingerman et al. (2008) found that young adult and middle-aged offspring, aged 22 to 49 years, who reported greater intergenerational ambivalence, also reported lower life satisfaction and higher depressive symptoms. Although there are no studies examining ambivalence and well-being among younger age groups, studies of younger people have shown links between positive and negative aspects of the parent-child tie and offspring's well-being. Parental support in early life is linked to better psychological well-being during late adolescence and early adulthood (Aquilino & Supple, 2001; Meadows, Brown, & Elder, 2006; Richman & Flaherty, 1986). Longitudinal studies have shown individuals who reported poor parent-child relations during childhood were at an increased risk for having depressive symptoms in adolescence (Aseltine, Gore, & Colten, 1998; Sheeber, Hops, Alpert, Davis, & Andrews, 1997) and in adulthood (Luecken & Roubinov, 2013; Richman & Flaherty, 1986). Further, longitudinal research has shown that more positive parent-child ties in adolescence and young adulthood are associated with better psychological well-being in middle age (Roberts & Bengston, 1993).
This study goes beyond existing research by examining the implications of ambivalence on depressive symptoms among adolescents, emerging adults, and young adults over a 12-year time period. We predict adolescents will experience greater declines in ambivalence and the association between ambivalence and depressive symptoms will be greater among adolescents compared to emerging or young adults. Adolescents are more dependent on their parents and do not yet possess many roles independent of the parent-child relationship. Thus, adolescents may experience greater ambivalent parent-child relationship quality and the effects of ambivalence more severely than emerging or young adults.
Research indicates that there are differences in maternal and paternal parenting characteristics and that mothers and fathers may have distinct effects on well-being. Overall, mothers tend to be more involved, provide more warmth and responsiveness, interact more frequently, and have closer relationships with their offspring than fathers (Collins & Russell, 1991; Forehand & Nousiainen, 1993; Holmbeck et al., 1995; Phares, 1999; Tsai, Telzer, & Fuligni, 2013). But there is also more mother-child conflict compared to fathers (Collins & Russell, 1991; Montemayor, Eberly, & Flannery, 1993; Steinberg & Silk, 2002), perhaps because mothers are typically more involved in the daily tasks of parenting and have more frequent interactions with children.
In general, fathers are less involved and less nurturing than mothers. Children feel less close to and spend less time with fathers (Hosley & Montemayer, 1997; Larson & Richards, 1994; Lewis & Lamb, 2003; Phares, 1999). Nevertheless, early research shows that closeness and warmth in the father-offspring relationship are correlated with children's achievement and psychosocial adjustment (Biller, 1971; Radin, 1981). Children lacking a positive paternal relationship may have an increased risk of depression (Phares, 1999). With this evidence in mind, ambivalence towards mothers and ambivalence towards fathers are examined in separate models. We hypothesize that participants would report greater ambivalence toward mothers than fathers but regardless of parent gender, greater ambivalence would be associated with greater depressive symptoms.
Several other factors have been shown to influence the parent-child tie and well-being include the offspring's gender, race, working status, marital status, and contact frequency with parents and we controlled for these factors. We controlled for offspring gender as there are contrasting studies regarding the role of offspring gender on ambivalent feelings. For instance, Willson, Shuey, and Elder (2003) found daughters reported greater ambivalence while other studies found no difference in ambivalence reports between sons and daughters (Fingerman et al., 2006; Pillemer & Suitor, 2002). Other research shows that adolescent girls experience higher levels of depressive symptoms than boys (Bracken & Reintjes, 2010; Cyranowski, Frank, Young, & Shear, 2000).
African American families may experience greater ambivalence than European American families due to increased stress including discrimination and lower socioeconomic status (Pillemer & Suitor, 2002). Depressive symptoms may also vary by race, however, the findings are inconsistent (Bracken & Reintjes, 2010; Twenge & Nolen-Hoeksema, 2002; Williams et al., 2007). Young people who have taken on more adult roles and responsibilities such as employment and marriage may report less conflict, ambivalence, and depressive symptoms than those who have not taken on those roles (Aquilino, 1997; Fingerman et al., 2006; House, Landis, & Umberson, 1998; Pillemer & Suitor, 2002; Willson et al., 2006). Furthermore, past research shows that greater contact frequency is associated with greater feelings of ambivalence, greater negative relationship quality, and less depressive symptoms (Birditt, et al., 2009; Fingerman, Hay, & Birditt, 2004; Lendon et al., 2014).
The present study examines ambivalence regarding mothers and fathers among adolescents, emerging adults, and young adults over 12 years. This study addressed the following three questions:
Participants were from the Social Relations and Health across the Lifespan study which included a regionally representative random sample of 1,703 people, ages 8 to 93 years at Wave 1 in (1992) from the greater Detroit area, and a total of 1,074 at Wave 2 (2005) (Antonucci, Birditt, & Webster, 2010; Birditt, Jackey, & Antonucci, 2009). The selected sample used in this study included participants who were adolescents (ages 13-17, n = 51), emerging adults (ages 18-25, n = 116), and young adults (ages 26-29, n = 88) at Wave 1. Of the selected sample, 76.4% (n = 188) participated in Wave 2. See Table 1 for description of the sample.
An attrition analysis used logistic regression to examine interview completion at Wave 2 (0 = drop out or deceased; 1 = completed) as a function of ambivalence scores, depressive symptoms, and all covariates at Wave 1. The analysis revealed that none of the variables significantly predicted interview completion at Wave 2.
Participants completed face-to-face interviews in their homes at Wave 1 and phone interviews in Wave 2. The interviews were approximately an hour in length and included questions regarding relationships with mother and father as well as depressive symptomology. Researchers have found that responses do not tend to vary between in-person and telephone survey methods (Herzog & Rodgers, 1988; Herzog, Rodgers, & Kulka, 1983).
The age variable included three categories: 0 (adolescents aged 13-17), 1 (emerging adults aged 18-25), 2 (young adults aged 26-29). Young adults serve as our comparison group in all analyses. Time was coded as 0 (Wave 1) or 1 (Wave 2).
Participants rated the positive and negative aspects of their relationship with their mother and father which were combined to form an ambivalence score. This is a common way to assess ambivalence in the literature (referred to as indirect measure of ambivalence) and it is associated with other direct measures of ambivalence. Positive qualities of the relationship included 5 items: “When my (mother/father) is having a hard time, I want to help (her/him),” “I feel my (mother/father) supports me, that (she/he) is there when I need (her/him),” “I enjoy being with my (mother/father),” “I feel my (mother/father) encourages me in whatever I do,” and “I feel my (mother/father) believes in me.” Items were rated on a 5-point scale (1 = strongly disagree, 2 = somewhat disagree, 3 = neither agree nor disagree, 4 = somewhat agree, or 5 = strongly agree). The items were averaged to create a positive mother and a positive father relationship quality score (Wave 1: α = .83, .92; Wave 2: α = .86, .93, respectively). Negative quality of the relationship included two items: “My (mother/father) gets on my nerves” and “My (mother/father) makes too many demands on me.” These items were also rated on a 5-point scale (1 = strongly disagree to 5 = strongly agree). The items were averaged to create a negative mother and a negative father relationship quality score (Wave 1: rs = .54, .54; Wave 2: rs = .62, .62, respectively). The small number of relationship items may underestimate the reliability of the negative quality scale (Carmines & Zeller, 1979), although similar reliability estimates are found in other research using the same or similar relationship items (Birditt et al., 2010; Birditt, Tighe, Fingerman, & Zarit, 2012; Fingerman et al., 2008; Umberson, 1992).
Ambivalent relationship quality scores were created for mothers and fathers (Fingerman et al., 2006, 2008; Willson et al., 2003, 2006) using the Griffin's Similarity and Intensity of Components formula to calculate ambivalence using positive and negative quality scales: [(positive + negative) / 2 - |positive – negative|] + 1.5 (Thompson, Zanna, & Griffin, 1995), which is similar to other ambivalence research (Fingerman et al., 2006, 2008; Willson et al., 2006). Higher scores indicate greater ambivalence. In a review of ambivalence measures, Thompson et al. (1995) showed that Griffin's measure is superior conceptually and in predictive power compared to other statistical methods for creating an indirect measure of ambivalence.
The respondents completed the 20-item Center for Epidemiologic Studies Depression Scale (CES-D) about their feelings during the past week which included items such as: “I was bothered by things that usually don’t bother me,” “I felt depressed,” and “I felt that everything I did was an effort” (Radloff, 1977). Participants rated items from 0 (rarely or none of the time), 1 (some or a little of the time), 2 (occasionally or a moderate amount of time), to 3 (most or all of the time). The items were summed to create an index and higher scores indicate more depressive symptoms (Wave 1: α = .88; Wave 2: α = .89).
We included offspring's gender, race, working status, marital status, and contact frequency as covariates. Gender was coded as 0 (men) or 1 (women). We coded race as 0 (non White) or 1 (White). Working status was coded as 0 (not working) or 1 (working full or part-time). Marital status was coded dichotomously (0 = not married, 1 = married). Participants rated their contact frequency based on the following question: “In the past 12 months how often have you seen (mother/father) in person?”: 1 (everyday), 2 (once a week), 3 (once a month), 4 (once a year), or 5 (irregularly).We reverse coded the item so that higher scores represented more contact.
A series of post hoc analyses also considered several additional time-varying covariates associated with parent-child ties and depressive symptoms including cohabitation with mother or father, education, or life problems in both Waves 1 and 2. Cohabitation was coded as 0 = not cohabitating or 1 = cohabitating separately for mother and father. Participants reported the number of years of education achieved ranging from 0 (no years of completed schooling) to 17 (completion of undergraduate degree or higher). The respondents completed an assessment of life problems. They were asked how many occurred in the past year. Examples of items include: “Has there been any change in your health, such as an illness or injury?,” “Has there been any trouble between you and your relatives?,” “Have you lost your job?,” and “Have you been involved in a crime?” Participants rated items as 0 (no) or 1 (yes). The items were summed with higher scores indicating more life problems.
To examine the research questions, multilevel models using SAS PROC MIXED were estimated. To assess whether ambivalence varied by age and over time, models included age, time, and the interaction between age and time as predictors with participant gender, race, working status, marital status, and contact frequency as covariates. Models were estimated separately for mother and father ambivalence. These models included two levels: upper level variables, i.e., participant characteristics (e.g., race, gender, Wave 1 age) and lower level variables, i.e., characteristics that varied within participant by wave (e.g., ambivalence scores, marital status, working status, contact frequency).
The models examining links between ambivalence and depressive symptoms included ambivalence in Wave 1, age, and time as well as all possible 2-way interactions as predictors to test whether depressive symptoms vary over time as a function of ambivalence in Wave 1 and whether the associations varied by age group. We also tested the 3-way age X time X ambivalence interaction to examine whether links between ambivalence in Wave 1 and depressive symptoms varied by both time and age but these interactions were removed in the final analyses due to non significance. These models had two levels: upper level variables included ambivalence at Wave 1, race, gender, and Wave 1 age and lower level variables included depressive symptoms, marital status, working status, and contact frequency. Interaction terms significant at the p < .10 level were interpreted due to well-documented difficulties in detecting significant interactions in non-experimental studies and survey research (McClelland & Judd, 1993).
There were less than 9% missing data for the selected sample with regard to ambivalence scores, depressive symptoms, and covariates. Multilevel models are ideal because participants with missing data are not removed from the analyses, and the models are not affected by an unbalanced design (i.e., in which some participants provided more waves of data than others).
We present the results in three sections. First, we describe the data. Next, we assess whether ambivalent feelings varied by age group and varied over time. Lastly, we examine associations between ambivalence and depressive symptoms over time.
To understand ambivalence scores in terms of the composition of positive and negative relationship quality, we examined means and standard deviations of positive and negative quality by the quartiles of ambivalence. Individuals in the lowest quartile of ambivalence scores (mother: n = 115, father: n = 90) had high positive scores (mother: M = 4.72, SD = .81; father: M = 4.74, SD = .73) and low negative scores (mother: M = 1.42, SD = .83; father: M = 1.13, SD = .72). Participants in the highest quartile of ambivalence scores (mother: n = 103; father: n = 88) had high positive scores (mother: M = 4.69, SD = .33; father: M = 4.31, SD = .61) and high negative scores (mother: M = 4.38, SD = .38; father: M = 4.02, SD = .64). Thus, low ambivalence was associated with high positive and low negative feelings. High ambivalence was associated with high positivity and high negativity.
Across age groups, participants reported greater ambivalence regarding mothers than fathers in Wave 1 (Mother: M = 3.60, SD = 1.71; Father: M = 3.09, SD = 1.76; t(207) = 3.45, p < .01) and Wave 2 (Mother: M = 2.87 , SD = 1.73 ; Father: M = 2.14, SD = 1.56; t(122) = , p < .01). See Table 1 for descriptives.
Analyses examining ambivalence regarding mothers and fathers in Wave 1 revealed significant main effects of age predicting father but not mother ambivalence (see Table 2). Adolescents reported greater ambivalence towards fathers than emerging adults (B = 1.04, SE = .53, t = 1.94, p < .10) and young adults (B = 1.14, SE = .55, t = 2.05, p < .05; see Table 1 for means). This is partially consistent with our hypothesis that adolescents would report greater ambivalence than emerging or young adults.
Analyses investigating whether mother ambivalence changed over time revealed a significant main effect of time (see Table 2). Ambivalence towards mothers decreased over time (B = −.64, SE = .30, t = −2.21, p < .05). These findings are consistent with our hypothesis that ambivalence would decline over time. However, there was no age effect as previously predicted.
Analyses assessing ambivalent feelings regarding fathers revealed no main effect of time but a significant age X time interaction (see Table 2). In order to investigate the interaction, we examined the effects of time for each age group with separate multilevel models. Similar to the models examining ambivalence toward mothers, father ambivalence decreased over time among adolescents (transitioning to young adulthood) (B = −3.81, SE = 1.09, t = −3.49, p < .01) and young adults (transitioning to middle age) (B = −.45, SE = .14, t = −3.21, p < .01) but not among emerging adults (transitioning to young adulthood) (B = −.26, SE = .34, t = −.76, p = .45). Examination of the covariates indicates that non-married offspring reported more ambivalence towards their fathers. These findings are partially consistent with our hypothesis that ambivalence would show declines among adolescents but also show that ambivalence declines among young adults.
Models examined the effects of Wave 1 mother and father ambivalence separately on depressive symptoms over time (see Table 3). The model predicting ambivalence regarding mothers revealed no main effect of mother ambivalence on depressive symptoms but there was a significant interaction between Wave 1 mother ambivalence X age. To investigate the interactions with age, we estimated separate models examining the effects of mother ambivalence by age group. Models revealed that adolescents and emerging adults who reported greater ambivalence towards mothers in Wave 1 reported greater depressive symptoms (B = 1.69, SE = .71, t = 2.38, p < .05; B = 1.76, SE = .44, t = 4.01, p < .01, respectively). These findings are partially consistent with our hypothesis that ambivalence in Wave 1 predicts higher depressive symptoms among adolescents but it also demonstrates that ambivalence predicts depressive symptoms in emerging adults.
Next, models assessed whether ambivalence towards fathers in Wave 1 predicted depressive symptoms over time. There was no main effect of father ambivalence but there was a significant Wave 1 father ambivalence X time interaction (see Table 3). The interaction showed that greater ambivalence towards fathers in Wave 1 is associated with a significant decline in depressive symptoms over time (b = −4.19, p < .05; see Figure 1). These findings are not consistent with our hypothesis that ambivalence would predict higher depressive symptoms over time. Furthermore, this association existed across all age groups rather than only among adolescents as we had predicted.
We conducted a series of post hoc analyses to examine whether the findings varied by the gender of the offspring. There was no evidence that ambivalence or the ambivalence-depressive symptoms link varied by gender of offspring.
Additionally, we conducted post hoc analyses to rule out concurrent associations between ambivalence and well-being. We estimated models with Wave 1 and Wave 2 ambivalence predicting depressive symptoms in Wave 2 controlling for Wave 1 depressive symptoms. Models including both Wave 1 and Wave 2 mother ambivalence as predictors showed that Wave 1 mother ambivalence predicted greater depressive symptoms in Wave 2 among emerging adults compared to young adults (B = 2.26, SE = 1.02, t = 2.21, p < .05). There was no effect of Wave 2 ambivalence for Wave 2 depressive symptoms. Thus, it appears that previous feelings of ambivalence in the mother tie have a greater impact on depressive symptoms than current feelings.
Models with both Wave 1 and Wave 2 father ambivalence revealed no significant effects of ambivalence on depressive symptoms. However, models with only current levels of ambivalence (omitting the lagged effect) showed that greater ambivalence predicted greater depressive symptoms (B = .82, SE = .30, t = 2.77, p < .01). Thus, contrary to the findings with mothers, it appears that present feelings regarding fathers have a greater impact on depressive symptoms than feelings experienced 12 years previously.
Furthermore, a series of post hoc analyses were conducted to assess the stability of the findings after controlling for additional factors that may influence ambivalence and depressive symptoms. These included cohabitation with mother or father, education level, and number of life problems. Each covariate was entered into the models separately. Findings remained the same for both mother and father ambivalence models except when the covariate number of life problems was entered. In the mother model, time was no longer significant. In the father model, the age X time interaction was no longer significant.
Models predicting depressive symptoms were re-estimated controlling for cohabitation with mother or father, education level, and number of life problems in separate models and the pattern of findings remained the same. In the mother models, the covariates cohabiting with mother and the life problems covariates were significant (B = 4.31, SE = 1.87, t = 2.31, p < .05; B = 1.72, SE = .29, t = 6.01, p < .01, respectively). In the father models, the covariate cohabitating with father was significant (B = 5.67, SE = 1.95, t = 2.91, p < .01).
Unlike previous research, the present study examined intergenerational ambivalence and its associations with depressive symptoms among young age groups: adolescents, emerging adults, and young adults. This study showed that ambivalence towards fathers is particularly high among adolescents, ambivalence decreases over time, and ambivalence has important implications for depressive symptoms as adolescents, emerging adults, and young adults grow older. These findings move the field forward by showing that ambivalent feelings exist during adolescence. Further, ambivalent feelings towards mothers predict increased depressive symptoms over time.
This study was the first, to our knowledge, to examine intergenerational ambivalence among younger offspring. Adolescents reported greater ambivalence than emerging adults and young adults towards fathers, but not mothers. This finding is consistent with previous research indicating increased positive parent-child relationship quality over time from adolescence to young adulthood (Aquilino, 1997) and decreased intergenerational ambivalence and negative relationship quality with age among middle-age and older adults (Birditt et al., 2010; Rossi & Rossi, 1990; Willson et al., 2006). Thus, the trajectory of decreasing ambivalence toward parents may begin in adolescence. As adolescents attempt to reconcile their need for independence with their continued parental dependence, they will likely experience high levels of ambivalence. However, as this group ages and experiences new life transitions, they will be less dependent on their parents and forge a less ambivalent parent-child relationship.
As predicted, offspring experienced more ambivalence toward mothers than fathers, regardless of age group. Offspring typically report being closer to mothers than fathers and the mother-child relationship is frequently characterized as a close, but often conflictual, tie (Fingerman, 2001; Pillemer et al., 2007; Umberson, 1992; Ward, 2008). Mothers and offspring have more contact compared to the father-child relationship (Buist, Deković, Meeus, & van Aken, 2002). Greater contact frequency in the mother tie may lead to increased ambivalence (Lendon et al., 2014) and fewer age differences in ambivalence regarding mothers.
This study is unique in that it examines whether there are increases or decreases in ambivalence during the transition to adulthood. The majority of the research regarding ambivalence has been cross- sectional. By examining individuals aged 13 to 29 over 12 years we were able to examine changes over time in intergenerational ambivalence across three young age groups. Indeed, ambivalence may be a normal developmental phenomenon as the young person reaches towards adulthood compared to an indication of maladaptation, as previously assessed among middle-aged and older adults.
This study showed that ambivalence regarding mothers decreased over time among all age groups. Mothers are typically more involved with offspring over the life course, and stay in more frequent contact and have closer relationships with offspring compared to fathers. This could explain the continuing higher rates of ambivalence towards mothers (Lye, 1996; Umberson, 1992).
Ambivalence towards fathers decreased over time among adolescents and young adults, but not among emerging adults. As individuation theory suggests, adolescence is often a tumultuous period for the parent-child relationship and is characterized by simultaneously higher levels of conflict, contact, and closeness than other developmental stages (Arnett, 2001a; Noack & Buhl, 2004; Steinberg, 1999). Decreases in ambivalence may be due to decreases in conflicts between independence and closeness experienced as adolescents shift into young adulthood (Blos, 1967; Bucx & van Wei, 2008). In this transition to young adulthood, offspring often fulfill normative expectations held by their parents, gain increased independence, and decrease contact (Aquilino, 1997).
In contrast to our prediction, ambivalence also declined over time among young adults. Previous research indicates that young adult and midlife adult children experience ambivalence towards their parents (Fingerman et al., 2008). However, young adults may become more similar to their parents as they transition to middle adulthood and acquire new adult roles and responsibilities (Nydegger, 1991). Additionally, there is less contact between parents and young or middle-aged adult children, which may relieve tensions that previously characterized the parent-offspring tie (Akiyama, Antonucci, Takahashi, & Langfahl, 2003).
Interestingly, there was no change in ambivalence towards fathers among emerging adults over time. Emerging adulthood is a period when individuals seek independence, either by moving out of the parental home or entering school, but they simultaneously rely on their parents for emotional, instrumental, or financial support (Aquilino, 2006a; Arnett 2000, 2001a). Aquilino (1997) suggests that parents must re-evaluate their expectations for their offspring and their own role as a parent during this time of transition. As offspring are adjusting to this experimental period, they may be more accepting and open to change. Because emerging adults have often not fully transitioned to complete independence or interdependence with their parents, the parent-emerging adult relationship may show little significant change in ambivalence.
This study revealed the significance of intergenerational ambivalence for depressive symptoms of offspring. Adolescents and emerging adults who reported greater ambivalence regarding their mothers reported greater depressive symptoms over 12 years. Adolescents are in more frequent contact with their mothers compared to older age groups (Holmbeck et al., 1995; Phares, 1999). And while mothers and their children often have a close relationship, this relationship can be fraught with conflicts over offspring independence and responsibilities (Pillemer et al., 2007). Greater contact and conflict between mother and child may have important implications for depressive symptoms. Similarly, emerging adults may still maintain closeness to mothers that resembles adolescence since they rely on their parents for support (Arnett 2000, 2001a). In these formative developmental stages, offspring who have conflicting feelings and feel unsupported by their mothers may be particularly negatively affected, thus significantly influencing their psychological health in later life (Hankin, Abramson, Moffitt, McGee, Silva, & Angell, 1998; Richman & Flaherty, 1986). Young adults may not be as negatively affected by these conflicted feelings as they may often have more adult-like and interdependent relationships with their mothers.
Research suggests adolescents are cognitively and developmentally vulnerable to depression (Jacobs, Reinecke, Gollan, & Kane, 2008). In fact, 75% of adults with major depressive disorder report experiencing their first depressive episode during childhood or adolescence (Kim-Cohen, Caspi, Moffitt, Harrington, Milne, & Poulton, 2003). Kessler and colleagues (2005) report the lifetime prevalence rate for depressive disorders in adolescents as 7%. Feelings of ambivalence may exacerbate negative well-being, such as depressive symptoms, in offspring. During this time of significant developmental change, children who possess ambivalent ties towards their parents may become more vulnerable to increased depressive symptoms.
Surprisingly, greater ambivalence towards fathers significantly predicted decreased depressive symptoms, regardless of age. Offspring who report high ambivalence towards their fathers in Wave 1 may reduce contact with their father over time in order to avoid the potentially conflictual or harmful relationships thus leading to decreased depressive symptoms. Father ambivalence may affect offspring of all ages because offspring are typically not as close to their fathers compared to their mothers (McKinney, Donnelly, & Renk, 2008; Steinberg & Silk, 2002). This lack of closeness may be perceived as lack of support from fathers, which in turn may negatively affect offspring as they age and lead to lower well-being. Compared to mothers, father acceptance is less available and occurs less often. Thus, offspring, regardless of age, may be more eager to seek the approval of their father (Forehand & Nousiainen, 1993). Offspring may feel distressed when their expectations concerning their relationship with or approval from their fathers are not met.
Post hoc analyses were conducted in order to understand the influence of additional covariates on offspring's reports of ambivalence towards parents and the implications of ambivalence on depressive symptoms. It appears the number of life problems and cohabitation with fathers has a significant impact on offspring's ambivalent feelings towards mothers and fathers and depressive symptoms in later life. These life problems, such as decreases in health, divorce, or money troubles, are most likely a source of tension in the parent-child relationship. Additionally, research indicates that failure of adult children to achieve expected milestones results in intergenerational ambivalence (Fingerman et al., 2006; Pillemer & Suitor, 2002). Indeed, Descartes (2006) found that both parents and adult children experience ambivalence when the adult child's dependence on the parent is prolonged past the normative developmental stage.
There are limitations to this study that should be addressed in future research. The negative relationship quality score had relatively low internal consistency reliability, most likely due to the small number of negative items. Thus, future studies should consider a scale with a greater number of negative items, which would likely improve internal consistency reliability. In addition, much of the literature suggests that negative feelings overwhelm ‘ambivalence’ and variability in negative feelings may actually account for much of the variability in ambivalence (Connidis, 2015; Fingerman et al., 2013; Gilligan, Suitor, Feld, & Pillemer, 2015).
Further, the concept of emerging adulthood is not a universal developmental period and is unique to certain cultures, especially industrialized countries that postpone the adoption of adult roles and responsibilities until past the late teenage years (Arnett, 2001a). Emerging adulthood may not be distinct from young adulthood for many individuals, especially among lower or working class individuals.
This study did not investigate parents’ perceptions of their relationship with their offspring or parents’ well-being. Similar to Lendon et al. (2014) and Fingerman et al. (2008), future studies should consider parents’ and offspring's reports of relationship quality over time to determine the possible bidirectional effects of ambivalence. Indeed, ambivalence in the parent-child tie may also have significant effects on parents’ well-being (Fingerman et al., 2008; Lowenstein, 2007). Furthermore, coping strategies may moderate the association between ambivalence and well-being. Birditt, Rott, and Fingerman (2009) found that addressing parent-child conflicts using constructive strategies, such as trying to understand each other, is associated with greater affective solidarity and less ambivalence in the relationship. The lack of constructive strategies may account for the links between ambivalence and depressive symptoms. Lastly, these results are limited by the availability of only two waves of data. The 12-year gap between the waves may not gather information on developmental changes occurring within that time span.
The present study expands the ambivalence and parent-child literature by examining intergenerational ambivalence among adolescents, emerging adults, and young adults over time. Our findings reveal that age of offspring is important when experiencing parental ambivalence. Furthermore, offspring's ambivalence and the effect of ambivalence on depressive symptoms differed by parent gender. Ambivalence clearly has important implications for well-being. We believe that the current study emphasizes the significance of intergenerational ambivalence and indicates that ambivalence is important to examine among younger as well as older age groups.
This study was supported by grants from the National Institute on Aging (NIA), AG13490, AG030569, and AG030569-02S2, Social Relations, Aging, and Health (Toni C. Antonucci, Principal Investigator).
Lauren A. Tighe, Life Course Development Program, Institute for Social Research, University of Michigan.
Kira S. Birditt, Life Course Development Program, Institute for Social Research, University of Michigan.
Toni C. Antonucci, Life Course Development Program, Institute for Social Research, University of Michigan.