Adult Children’s Problems and Successes: Implications for Intergenerational Ambivalence

Abstract

Ambivalence theory suggests that parents experience ambivalence due to conflicting desires to help children in need and to launch children into adulthood. This study examined parents’ reports of their adult children’s problems and successes and implications for ambivalence. Participants aged 40–60 years (302 men and 331 women from different families) reported on up to 3 of their adult children (N = 1,251). Men and women differentiated among children in ratings of problems, successes, and ambivalence. Men and not women reported greater ambivalence regarding children with more physical–emotional problems and less career success. Men and women reported greater ambivalence regarding children with less relationship success. Consistent with ambivalence theory, individuals feel more ambivalent regarding problematic and less successful children but men’s ambivalence appears to be more sensitive to their children’s problems and successes than women's ambivalence.

ADULT children are a significant source of support as well as strain for mothers and fathers across the life span. Indeed, ambivalence, the simultaneous experience of positive and negative feelings, is common in the parent–child tie due to conflicting desires for independence and closeness (Pillemer & Suitor, 2005). More than 50% of midlife and older parents report some ambivalence toward their grown children (Fingerman, Hay, & Birditt, 2004; Pillemer & Suitor, 2002). Although middle-aged and older parents report greater ambivalence toward offspring who have not achieved adult milestones (e.g., marriage) and older mothers report more ambivalence regarding problematic children (Fingerman, Chen, Hay, Cichy, & Lefkowitz, 2006; Pillemer, Suitor, et al., 2007), researchers have not explicitly examined whether children’s problems and achievements help explain midlife men’s and women’s feelings of ambivalence toward those children. Parents may feel greater ambivalence regarding problematic children (e.g., trouble with the law and illness) and less successful children (e.g., career success) due to their continuing dependence and violation of adult status attainment norms (Aldous, Klaus, & Klein, 1985; Pillemer, Sechrist, Steinhour, & Suitor, 2007). Problematic children provide less care to parents (Cicirelli, 1983), they cause parental distress (Greenberg & Becker, 1988; Pillemer & Suitor, 1991), they tend to drain parents’ resources (Aldous et al.), and they are parents’ least preferred support providers (Pillemer & Suitor, 2002). In contrast, successful children (e.g., doing well in their relationships and career) may increase parents’ well-being and provide more support to parents (Fingerman, Miller, Birditt, & Zarit, 2009; Ryff, Lee, Essex, & Schmutte, 1994).

Ambivalence theory provides a useful framework for understanding the complexity of the parent–child relationship and adult children’s problems and successes. Ambivalence refers to the simultaneous experience of positive and negative sentiments about the same relationship (Luescher & Pillemer, 1998). The present study defines ambivalence as reporting positive and negative relationship quality regarding the same child. Individuals experience ambivalence when there are incompatible norms or expectations that cause contradictory emotions or beliefs (Merton & Barber, 1963). Parents experience ambivalence due to competing desires to launch their children into adulthood and to support to children in need (e.g., children with problems).

It is particularly important to identify predictors of ambivalence because ambivalence is associated with greater depression, lower quality of life, and poorer health among parents (Fingerman, Pitzer, Lefkowitz, Birditt, & Mroczek, 2008; Lowenstein, 2007; Ward, 2008). The majority of work in this area has examined older mother’s reports and often reports of only one child. This study seeks to understand whether middle-aged women and men differentiate among their children in terms of their successes and problems and whether these differentiations are associated with ambivalence. This study examines (a) within-individual variations in reports of adult children’s problems, successes, and ambivalence among middle-aged women and men and (b) implications of problems and successes for intergenerational ambivalence among middle-aged women and men. This research has practical implications because it may provide insights into how to improve the parent–child relationship as parents grow older and in greater need of support from these ties.

PROBLEMS AND SUCCESSES DEFINED

Identifying problems among children often involves distinguishing between less controllable physical–emotional problems and more controllable lifestyle–behavioral problems (Pillemer & Suitor, 2002; Suitor, Pillemer, & Sechrist, 2006). Physical–emotional problems include health problems, developmental disabilities, and mental health issues. In contrast, lifestyle–behavioral problems include financial trouble, drug and alcohol abuse, and trouble with the law. Research suggests that mothers evaluate grown children with physical–emotional problems more favorably, whereas they view children with lifestyle–behavioral problems with greater ambivalence (Pillemer & Suitor, 2002; Suitor & Pillemer, 2000).

To assess children’s successes, researchers have typically examined social roles, such as whether children are married or employed full time. The importance parents place on different roles varies as a function of their own values (Ryff et al., 1994) and the child’s age. Thus, in the present study, rather than examining the presence or absence of roles, we examined middle-aged parents’ ratings of how successful they believed their children were in their careers and relationships compared with other children of their age.

This study examines parents’ ratings of their children’s problems and successes simultaneously because parents may hold complex and orthogonal beliefs about each of their children’s successes and problems in a family. For example, a child with problems may experience success in some areas and less successful children do not necessarily have problems.

DIFFERENTIATIONS AMONG CHILDREN

Prior work on child problems and ambivalence has focused mostly on older mothers and often reports of one child (Pillemer & Suitor, 2002; Pillemer, Suitor, et al., 2007). Older mothers differentiate among their children in ratings of emotional closeness, the types of support they prefer to receive, and the support they provide to children (Suitor, Sechrist, & Pillemer, 2007; Suitor et al., 2006). In the present study, we examine whether middle-aged men and women differentiate among their children in ratings of problems, successes, and ambivalence. Examining within-parent differentiations allows us to assess whether characteristics of the children contribute to parents’ ratings of problems and successes rather than only parent characteristics.

We also consider whether women and men vary in the extent to which they differentiate among their children. Because mothers’ relationships with children are often closer and involve more contact, women may differentiate more among children than do men (Umberson, 1989; Ward, 2008).

IMPLICATIONS OF PROBLEMS AND SUCCESSES FOR AMBIVALENCE

Parents may feel more ambivalent regarding problematic children and less successful children due to their lack of independence, needs for support, violation of normative expectations, and elicitation of embarrassment (Pillemer & Suitor, 1991). In contrast, more successful children may provide a source of pride, well-being, and reduced ambivalence.

Some studies report no associations between problems and ambivalence (Pillemer & Suitor, 2002), and others show that problems predict greater ambivalence. Pillemer, Suitor, and colleagues (2007) found that mothers report greater ambivalence regarding children with more lifestyle–behavioral and physical–emotional problems. However, other studies suggest that parents differentiate between problems; mothers report feeling closest to children who have physical–emotional problems but the least close to children who have lifestyle–behavioral problems (e.g., substance abuse and trouble with the law; Suitor & Pillemer, 2000). It is not clear whether these findings regarding closeness also apply to feelings of ambivalence.

Furthermore, women and men may experience children’s problems and successes differently. Women’s feelings of ambivalence may be more highly associated with children’s problems and successes than mens’ feelings of ambivalence (Nelson et al., 2007). Mothers of young adult children with mental illness experience more emotional distress than do fathers (Cook, 1988). Similarly, older mothers experience more stress related to their children’s problems compared with fathers (Greenberg & Becker, 1988). Women may possess more of a success orientation to children and experience greater impact of children’s issues than men who blame children for problems (McBride & Black, 1984). Mothers may also experience a more detrimental impact due to their greater investment in the tie (Rossi & Rossi, 1990). Other studies, however, reveal no gender differences in associations between child problems and parental relationship perceptions (Greenfield & Marks, 2006).

OTHER FACTORS INFLUENCING THE PARENT–CHILD TIE

Other child characteristics may influence parents’ perceptions of their children. Ambivalence theory suggests that parents feel more ambivalent about children who have not reached adult roles indicative of independence, such as employment, education, and marriage (Fingerman et al., 2006; Pillemer & Suitor, 2002; Willson, Shuey, Elder, & Wickrama, 2006). For example, mothers feel greater ambivalence regarding children who are not married (Pillemer, Suitor et al., 2007) or who have lower education (Pillemer & Suitor, 2005). Individuals also make differentiations among their children based on children’s gender, age, employment status, and parental status (Suitor et al., 2006).

Parent characteristics including race, marital status, number of living children, socioeconomic status (SES) (education and income), neuroticism, and self-rated health may differentially influence the parent–child relationship (Fingerman et al., 2006; Rossi & Rossi, 1990; Suitor et al., 2006). For example, being African American, having more education, scoring higher in neuroticism, and having lower self-rated health are associated with greater ambivalence (Birditt, Rott, & Fingerman, in press; Connidis & McMullin, 2002; Fingerman et al., 2006).

Relationship characteristics including contact frequency and financial dependence also differentially predict ambivalence. Parents who report greater contact frequency and greater financial support to children often report greater ambivalence (Fingerman et al., 2006; Pillemer & Suitor, 2002). The present study included covariates for child, parent, and relationship characteristics potentially associated with ambivalence.

STUDY GOALS

This study addressed the following two questions:

1. Do women and men differentiate among their children in their reports of problems, successes, and ambivalence? We predict that individuals make differentiations among their children in ratings of their problems, successes, and ambivalence (Suitor et al., 2006, 2007), and there is greater variation within individuals regarding each of their children rather than between individuals. Because mothers are often closer to children, we predict that women differentiate among children more than men.

2. Are women’s and men’s reports of problems and successes associated with ambivalence? We predict that individuals will report greater ambivalence regarding children with more problems (lifestyle–behavioral and physical–emotional). We also predict that lifestyle–behavioral problems and ambivalence are more highly associated than physical–emotional problems and ambivalence (Pillemer & Suitor, 2002). We hypothesize that individuals feel less ambivalent about children they rate as more successful. We expect that the associations are stronger among women than men (Greenberg & Becker, 1988).

METHODS

Participants

The Family Exchanges Study included 633 individuals aged 40–60 years (302 men and 331 women from different families) who had at least 1 child aged 18 years or older. We randomly selected participants from phone lists obtained through Genesys Corporation as well as through random digit dialing from the Philadelphia Primary Metropolitan Statistical Area (five counties in southeastern Pennsylvania and four counties in New Jersey) and stratified by gender and age (40–50 and 51–60 years; Pennsylvania State Data Center, 2001). We oversampled participants living in Philadelphia county, high-density minority neighborhoods, and lower SES households, which resulted in a total of 37% minority participants (see Table 1). Participants completed hour-long computer-assisted telephone interviews and received $25. Of the total number of participants contacted, 75% of the eligible individuals participated and completed the interviews.

Table 1.

Characteristics of the Sample

Variable	Women (N = 331)	Men (N = 302)	Offspring of mothers (N = 666)	Offspring of fathers (N = 585)
Age, M (SD)	50.45(4.98)	50.76(5.01)	25.17(5.89)	24.29(5.17)
Years of education, M (SD)	14.22(2.01)	14.14(2.03)	13.87(1.84)	13.56(1.99)
Self-rated healtha, M (SD)	3.48(1.06)	3.48(1.08)	4.23(.94)	4.29(.93)
Household incomeb, M (SD)	4.37(1.41)	4.43(1.50)	—	—
Number of childrenc, M (SD)	2.79(1.34)	2.85(1.59)	1.98(1.18)	1.80 (1.13)
Number of children aged 18 years or     olderd, M (SD)	2.22(1.22)	2.15(1.25)	—	—
Ethnicity (%)
African American	30.3	35.2	30.3	35.2
European American	63.6	59.1	63.6	59.1
Hispanic	0.1	0.1	0.1	0.1
Multiracial	6.0	5.6	6.0	5.6
Marital status (%)
Married or remarried	71.3	68.2	19.6	13.1
Work status (%)
Employed full time	57.9	73.1	52.7	49.6
Parental status (%)
At least one child	1.00	1.00	28.8	26.4
One child aged 18 years or olderd	29.6	34.4	2.0	1.0
Two or more children aged 18   years or older	70.4	65.6	—	—

Notes: ^aRated 1 = poor, 2 = fair, 3 = good, 4 = very good, and 5 = excellent.

^bHousehold income 2007 rated 1 = less than $10,000, 2 = $10,001–$25,000, 3 = $25,001–$40,000, 4 = $40,001–75,000, 5 = $75,001–$100,000, and 6 = more than $100,000.

^cAmong those who had children.

^dA total of 18 (13 mother offspring and 5 father offspring) had at least 1 child aged 18 years or older.

Participants answered detailed questions regarding up to three of their children aged 18 years and older. Individuals reported each child’s gender, age, marital status, work status, parental status, contact frequency, financial support problems, successes, and relationship quality (N = 1,251). Participants had an average of 2.61 (SD = 1.30) children aged 18 years or older (range = 1–10). Of these children, 79.3% were aged 18–29 years. The majority of participants (88%) had three or fewer children aged 18 years or older. Most children (88.2%) were biological, 1.3% were adopted, 9.7% were stepchildren, and the remaining 0.9% were not classified. A total of 48.2% of mother’s offspring and 47.5% of father’s offspring were daughters. See Table 1 for a sample description.

Measures

Problems.—

Individuals reported whether their children experienced a series of problems, which we grouped into two scales: physical–emotional problems and lifestyle–behavioral problems (Pillemer, Suitor, et al., 2007). Physical–emotional problems included two lifelong problems (having a developmental delay or disability and a physical disability) and two problems in the past 2 years (a serious health problem or injury and a serious emotional or psychological problem). Lifestyle–behavioral problems included four problems in the past 2 years: a drinking or drug problem, financial problems, trouble with the law or police, and divorce or other serious relationship problems. We adapted this measure from the National Survey of Midlife Development in the United States (MIDUS; Greenfield & Marks, 2006). Due to skew in distribution, we summed the number of problems and truncated both scores into 0, 1, and 2 or more problems. Each child received a score for both types of problems.

Successes.—

We adapted items from previous work (Ryff et al., 1994) to assess parents’ perceptions of children’s successes with two items (Fingerman et al., 2009). Respondents indicated how successful they believed that each child is compared with others of the same age in two domains, career (education, work, and career) and relationships (romantic relationships and family life), on a 5-point scale (1 = less successful, 2 = somewhat less successful, 3 = about the same as other people of his/her age, 4 = somewhat more successful, and 5 = more successful than other people of his/her age). Because participants were asked across the domains with only two items rather than each domain separately, there were little missing data on these items.

Child characteristics.—

Individuals reported their child’s demographics and social roles. Demographics included gender (0 = daughter and 1 = son) and years of education (highest grade in school or year of college completed). We dichotomized age to distinguish the period of emerging adulthood (0 = ages 18–29) from young to middle adulthood (1 = ages 30–46; Arnett, 2000). Social roles included the child’s marital status (0 = not married and 1 = married or remarried), parental status (0 = no children and 1 = one or more children), and work status (0 = not employed full time and 1 = employed full time).

Participant characteristics.—

Participant characteristics included demographics, health, and neuroticism. Demographics incorporated gender (0 = women and 1 = men), race (0 = non-White and 1 = White), marital status (0 = not married and 1 = married or remarried), number of living children, education (highest grade or year of college completed), and income rated from 1 (<$10,000) to 6 (>$100,000). Individuals rated their physical health from 1 (excellent) to 5 (poor). We reverse coded the item so that higher scores represent better health.

Neuroticism included four items from the MIDUS. Parents indicated how well each of four adjectives (moody, worrying, nervous, and calm) described themselves from 1 (a lot) to 4 (not at all). We reverse coded all items, with the exception of calm, so that higher scores represent greater neuroticism.

Relationship characteristics.—

We also considered participant reports of contact frequency and the financial dependence of the child. Participants rated how often they had seen each child in the past 12 months from 1 (daily) to 8 (less than once a year or never). We reverse coded the item so that higher scores represent greater contact. Participants reported how often they provided financial support to each child (e.g., giving or loaning money, helping the child purchase goods, services, insurance, and education) from 1 (daily) to 8 (less than once a year or never). We reverse coded the item so that higher scores represent greater financial support.

Ambivalence.—

We combined assessments of positive and negative perceptions of the relationship to generate ambivalence scores for each child (Fingerman et al., 2006; Willson, Shuey, & Elder, 2003). This indirect approach is associated with other measures of ambivalence (Willson et al., 2003) and may be more effective than directly asking participants about their mixed feelings (Pillemer & Suitor, 2002). Participants may have difficulty directly reporting their mixed feelings (Luescher & Pillemer, 1998). Thus, this measure reflects ambivalent (mixed) relationship quality rather than ambivalent (mixed) emotions. The positive component included the mean of two items (“How much does he/she make you feel loved and cared for” and “How much does he/she understand you”). The negative component included the mean of two items (“How much does he/she criticize you” and “How much does he/she make demands on you”). Participants rated items from 1 (not at all) to 5 (a great deal). As in other studies of ambivalence (e.g., Fingerman et al., 2008; Willson et al., 2006), we used Griffin’s Similarity and Intensity of Components formula to calculate ambivalence scores: [(positive + negative)/2 − |positive − negative|] + 1.5 (Thompson, Zanna, & Griffin, 1995). Higher scores reflect greater ambivalence.

Analysis strategy.—

Analyses involved descriptive statistics and multilevel models. We calculated means and correlations among the variables to describe women’s and men’s reports. Because the data included individual’s reports regarding multiple children, we used multilevel modeling to address the issue of dependencies in the data (SAS PROC MIXED; Littell, Milliken, Stroup, & Wolfinger, 1996; Singer, 1998). Multilevel models examined whether individuals differentiated among their children and whether problems and successes predicted ambivalence. Models included a random effect allowing for correlated errors within participants. Models included two levels: higher level participant characteristics (e.g., gender and race) and lower level child characteristics (e.g., problems and successes).

Blank multilevel models without predictors assessed how much of the variance in individuals’ ratings of their children’s problems, successes, and feelings of ambivalence was due to between-person variance versus within-person variance (Singer & Willett, 2003). We estimated models for each outcome (physical–emotional problems, lifestyle–behavioral problems, career success, relationship success, and ambivalence). We then calculated intraclass correlations by dividing each covariance parameter by the total variance. To test whether women and men varied in how much they differentiated among children, we compared models in which there were separate between-person variance components estimated for women and men with those in which one variance component was estimated across women and men. We then examined whether there was a significant difference between the fit of the models by subtracting the −2 log likelihood estimations of two models and examining the difference on a chi-square distribution with a degree of freedom equaling the change in number of parameters (df = 1).

Next, we estimated multilevel models to examine whether ambivalence varied by parents’ reports of their children’s problems and successes. We estimated two models predicting ambivalence for women and men separately. Predictors included physical–emotional problems, lifestyle–behavioral problems, career success, and relationship success. Covariates included child characteristics, participant characteristics, and relationship characteristics.

To examine whether models that included problems and successes provided improved fit over models with only the covariates, we statistically compared the goodness of fit of the models (Singer & Willett, 2003). The goodness-of-fit comparison involves subtracting the −2 log likelihood estimations of two models and examining the difference on a chi-square distribution with a degree of freedom equaling the change in number of parameters (df = 4).

Results

We present the results in three sections. First, we provide a description of individuals’ reports of their children’s problems, successes, and ambivalence. Next, we examine between- and within-person reports of children’s problems, successes, and ambivalence. Finally, we examine associations among children’s problems, successes, and ambivalence.

Description of Individual’s Reports Regarding Children

Individuals reported that each child experienced an average of 0.22 (SD = 0.50) physical–emotional problems and an average of 0.47 (SD = 0.69) lifestyle–behavioral problems. In other words, middle-aged parents reported that approximately 18.5% of offspring had physical or emotional problems and 35.2% had one or more lifestyle-behavioral problems. Physical–emotional problems and lifestyle–behavioral problems were positively correlated (r = .21). The most frequently reported problems within each category included mental health problems (8%) and financial trouble (26%). On average, parents rated their children as slightly above their age-mates in career success (M = 3.50, SD = 1.17) and as similar to peers in relationship success (M = 3.15, SD = 1.06). Modest correlations were evident for ratings of success and problems. Children with greater career successes had greater relationship success (r = .38). Children with more problems were rated as less successful in careers and relationships (physical–emotional problems rs = −.22 and −.13 and lifestyle–behavioral problems rs = −.33 and −.25, respectively).

An examination of gender differences in ratings of problems, successes, and ambivalence revealed that women rated their children as having more relationship success (M = 3.22, SD = 1.10) than did men (M = 3.08, SD = 1.00; t = 2.41, p < .05). Problems, career success, and ambivalence did not vary by parental gender.

Between- and Within-Person Reports of Problems and Successes

The next analyses tested the extent to which individuals differentiated among their children and whether women made greater differentiations among children than men. Table 2 includes the intraclass correlations representing the proportion of variance between and within women and men. Consistent with the first hypothesis, we found greater within-person variance in how parents viewed their children than between-person variance. Inconsistent with the second hypothesis, women and men made similar differentiations among their children. There were no significant differences in the fit of models in which separate between-person variance components were estimated for women and men and the models in which only one between-person variance component was estimated (Table 2).

Table 2.

Proportion of Between- and Within-Person Variance in Women’s and Men’s’ Ratings of Child Problems, Successes, and Ambivalence and a Comparison of Models with Common Variance and Separate Between-Person Variance Components

	Proportion of variance between women	Proportion of variance within women	Proportion of variance between men	Proportion of variance within men	Model with common between variance for women and men, −2 log likelihood	Model with separate between variance components for women and men, −2 log likelihood	Difference between −2 log likelihoods (critical value = 3.84)
Physical–emotional problems	.09	.91	.00	1.00	1,816.2	1,813.6	2.6
Lifestyle–behavioral problems	.18	.82	.26	.74	2,588.3	2,585.9	2.4
Career success	.16	.84	.21	.79	3,878.9	3,878.1	0.8
Relationship success	.19	.81	.10	.90	3,582.4	3,579.6	2.8
Ambivalence	.42	.58	.42	.58	3,939.7	3,939.7	0

Associations Among Problems, Successes, and Ambivalence

Table 3 includes the multilevel models predicting ambivalence as a function of the covariates in the first column, followed by a multilevel model predicting ambivalence as a function of problems, successes, and covariates in the second column. The models with problems and successes as predictors of ambivalence provided a significantly better fit than those with only the covariates as predictors of ambivalence (Table 3).

Table 3.

Multilevel Models Examining Women’s and Men’s Ratings of Ambivalence as a Function of Child Problems and Successes

	Ambivalence				Ambivalence
	Women		Men		Women		Men
	B	SEB	B	SEB	B	SEB	B	SEB
Physical problems					0.08	0.09	0.20*	0.10
Lifestyle problems					0.08	0.07	0.06	0.07
Career success					−0.03	0.05	−0.09*	0.05
Relationship success					−0.14**	0.05	−0.14**	0.05
Participant characteristics
Married	−0.08	0.16	0.01	0.16	−0.07	0.15	0.04	0.16
Education	0.08*	0.03	0.09**	0.04	0.07*	0.03	0.09*	0.04
Number of children	−0.02	0.04	0.02	0.04	−0.03	0.04	0.01	0.04
Ethnicity (White)	0.13	0.14	0.15	0.16	0.08	0.13	0.13	0.15
Income	−0.01	0.06	0.01	0.06	−0.02	0.06	−0.01	0.06
Neuroticism	0.24**	0.07	0.28**	0.08	0.21**	0.06	0.24**	0.08
Self-rated health	−0.12*	0.06	−0.09	0.06	−0.09	0.06	−0.10	0.06
Child characteristics
Gender (son)	−0.22**	0.09	−0.38**	0.09	−0.22**	0.09	−0.33**	0.09
Age (30–46 years)	0.09	0.13	−0.26	0.15	0.01	0.13	−0.30*	0.14
Married	−0.40**	0.13	−0.38*	0.16	−0.23	0.14	−0.18	0.17
Education	−0.07**	0.03	−0.07**	0.03	−0.05	0.03	−0.04	0.03
Employed	−0.11	0.10	0.04	0.10	−0.11	0.10	0.05	0.10
Have children	−0.08	0.12	0.15	0.12	−0.12	0.13	0.08	0.12
Relationship characteristics
Contact frequency	0.05*	0.02	0.03	0.03	0.05*	0.02	0.05	0.03
Financial support	0.07**	0.03	0.06*	0.03	0.07**	0.03	0.06*	0.03
−2 Log likelihood	1,855.2		1,662.9		1,811.6		1,611.8
Change in log likelihood					43.6*		51.1*

Note: *p < .05; **p < .01.

As we hypothesized, children’s problems were associated with greater ambivalence. Men reported greater ambivalence regarding children with more physical–emotional problems. Somewhat surprisingly, there were no associations between lifestyle–behavioral problems and ambivalence and no associations between problems and ambivalence among women. As expected, women and men reported greater ambivalence regarding children with less relationship success. However, only men and not women reported greater ambivalence about children with less career success.

Ambivalence also varied by participant, child, and relationship characteristics. Participants reported greater ambivalence when they had higher levels of neuroticism and more education and gave more financial support to offspring. Participants also reported greater ambivalence regarding daughters than sons. Women who reported more contact frequency with children reported greater ambivalence, and men reported greater ambivalence regarding younger offspring. Because individuals’ opinions of children may vary depending on children’s age, we also examined interactions among age of child, problems, and successes as predictors of ambivalence and found no significant interactions.

As an alternative way of examining the implications of child problems and successes for ambivalence, we explored whether children’s problems and successes were differentially associated with the positive and negative components of ambivalence. According to ambivalence theory, parents may perceive problematic and less successful children as more positive and more negative. We estimated multilevel models with problems, successes, and covariates as the predictors and positive and negative components of ambivalence as the outcomes. Women reported greater negative perceptions of children with lifestyle–behavioral problems (B = .10, SE = 0.05, p < .05), and men reported greater negative perceptions of children with physical–emotional problems (B = .14, SE = 0.07, p < .05) and less relationship success (B = −.10, SE = 0.03, p < .01). Women and men reported more positive perceptions regarding children with more career success (B = .10, SE = 0.03, p < .01 and B = .15, SE = 0.03, p < .01, respectively) and relationship success (B = .19, SE = 0.03, p < .01 and B = .12, SE = 0.03, p < .01, respectively).

DISCUSSION

The present study examined men’s and women’s reports of their children’s problems and successes and whether problems and successes predict ambivalence. This study showed, not unexpectedly, that parents differentiated among their children in their ratings of problems, successes, and ambivalence. We also build on and extend findings of previous studies that either only examined older women’s perceptions of ambivalence or examined ambivalence toward a single child. Including parents’ opinions of up to three children, we found that parents’ perceptions of a child’s problems and successes were related to ambivalent feelings and that the specific patterns of associations differed between men and women. Thus, ambivalence toward grown children is a common experience for men and women, and these feelings reflect in varying degrees children’s lack of success and/or problems due to more controllable and less controllable factors.

Do Women and Men Differentiate Among Children?

We found greater variation in parents’ ratings of their different children’s problems, successes, and ambivalence than between individuals. We reiterate that our sample did not include mothers and fathers from the same family, but rather the men and women were not related. Given this sampling approach, the findings show that there is greater variation within a parent regarding each of his or her own children than between different unrelated parents regarding their children.

Women and men made similar differentiations in their ratings of children. We had hypothesized that women would make greater differentiations among children than men due to their greater emotional closeness and emotional intensity with children (Rossi & Rossi, 1990). More recent literature has found a lack of variation between mothers and fathers within the same families in several key variables, including ambivalence and tensions (Birditt, Miller, Fingerman, & Lefkowitz, 2009; Fingerman et al., 2006). There may be fewer gender differences in more recent cohorts as women and men take on more similar roles. The issues in adulthood may be more gender neutral (as compared with adolescence) and may elicit fewer gender differences. For example, gender intensification theory suggests that children experience an intensification of gender roles during adolescence, which coincides with greater sex-typed parental socialization (Hill & Lynch, 1983).

Do Children’s Problems and Successes Relate to Perceptions of Ambivalence?

Ambivalence theory suggests that parents experience competing desires to launch their children into an independent adulthood, yet conversely to help children in need of support (e.g., children with problems). Thus, we predicted that women and men would feel more ambivalent about children with more problems but found only partial support for this hypothesis. Ambivalence did not vary by lifestyle–behavioral problems, and men (but not women) reported greater ambivalence regarding children who had physical–emotional problems. Fathers may have a more difficult time accepting children with developmental delays, which may manifest as ambivalence regarding young adult children (Price-Bonham & Addison, 1978). It is possible that men experience more long-term ambivalence about children with disabilities than do women.

We were somewhat surprised that children’s problems did not predict greater ambivalence among women given past research. Pillemer, Suitor, and colleagues (2007) found that mothers rated children with more problems with more ambivalence. There were several variations between our studies. Pillemer, Suitor, and colleagues examined mothers aged 65–75 years and four types of problems (serious illness or injury, serious mental or emotional problem, problems with drinking or drugs, and problems with the law) and used a direct assessment of ambivalence. Older mother’s feelings of ambivalence may be more highly associated with their offspring’s problems than younger mother’s feelings. Indeed, as mothers age, they may become more ambivalent about problematic children due to their own age-related increased needs for support. We examined problems ranging in severity, which may not be as highly linked with ambivalence. Finally, direct assessments of ambivalence may reflect more negative sentiments because participants report how torn they feel rather than both positive and negative qualities separately. These studies also vary in their geographic locations; however, there is little variation in the parent–child tie between U.S. regions (Sechrist, Suitor, Henderson, Cline, & Steinhour, 2007).

We also found support for the hypothesis that individuals feel more ambivalent about less successful children. Women and men reported more ambivalence about children with less relationship success. Similarly, parents favor unmarried children less than those who are married (Brackbill, Kitch, & Noffsinger, 1988). Relationship success also involved the parent’s rating of children’s success in romantic relationships and family life, which may subsume the parent–child tie and have direct links with ambivalence. Men reported more ambivalence about children who were less successful in their careers. Men may experience higher expectations for their children’s career success than do women. Men may have had children who were less successful than women; however, men and women reported similar levels of career success.

The examination of the positive and negative components of ambivalence allowed for a nuanced examination of parents’ perceptions. Parents perceived children with problems more negatively but no less positively. In contrast, parents perceived their more successful children as more positive but not necessarily as less negative. These findings reveal ambivalence by showing that positive and negative perceptions of relationships are not mutually exclusive.

Future Directions

There are several future research directions to consider. Because of the cross-sectional design, the direction of the links among ambivalence, problems, and successes is unclear and most likely bidirectional. Future research should examine these linkages over time. Future studies should examine mothers’ and fathers’ reports of problems and successes in the same families as well as the offspring’s perspective. Indeed, parents underestimate children’s problems (Suitor & Pillemer, 2000). Children’s problems may contribute to lower parental well-being because of the overall strain of those problems on the relationship, and future studies should consider these links (Greenfield & Marks, 2006). Despite few regional differences in the parent–child tie (e.g., Sechrist et al., 2007), future work should consider these associations nationally.

This study has clinical and gerontological implications. Ambivalence predicts lower quality of life, more depressive symptoms, and lower self-rated health (Fingerman et al., 2008; Lowenstein, 2007). Clinicians should be aware of the factors associated with greater ambivalence and the gender differences. Indeed, men with problematic and less successful children may be particularly vulnerable. Clinicians should consider the social context of adults as they age and how their family members may have positive as well as negative influences. In addition, parents and children who approach problems with constructive rather than avoidant or destructive strategies report lower ambivalence (Birditt et al., in press). Thus, considering how parents and children cope with children’s problems, and lack of successes may help reduce the impact of problems and successes on ambivalence. As parents grow older and are more in need of support, problematic and less successful children may become even more detrimental to parents.

Because men and women reported on multiple children, this study allowed for an examination of parental perceptions as a function of children’s characteristics over and above parent characteristics. Including men in the study revealed that men’s feelings of ambivalence were more sensitive to their young adult’s problems and successes than women. Middle-aged men may have higher expectations for their adult children than women. This study also revealed that although parents may feel more negative about children with problems, they feel no less positive regarding those children. Likewise, parents generally reported feeling more positive and no less negative regarding their successful children. Parents’ feelings about their children are multifaceted and may have important implications for well-being and support exchange across adulthood, especially among men.

FUNDING

This study was supported by grant R01 AG027769, “The Psychology of Intergenerational Transfers” from the National Institute of Aging. K.S.B. is also supported by a K99 grant from the National Institute of Aging (K99 AG029879, “Promoting Well Being Across Adulthood: The Role of Conflict Avoidance”).