Perceived paternal and maternal involvement: factor structures, mean differences, and parental roles.

This study was designed to compare the factor structures and means for mothering and fathering, as retrospectively perceived by young adult children. Three dimensions of perceived parenting were examined: nurturance, reported involvement, and desired involvement. We used the existing Nurturant Fathering and Father Involvement Scales, and the newly developed parallel Nurturant Mothering and Mother Involvement Scales. Confirmatory factor analyses indicated that the originally validated factor structure of the fathering measures fit the data well and that the factor structures of paternal and maternal nurturance and involvement were isomorphic. These scales appear to index parenting functions that generalize across mothers and fathers. Mean-difference analyses indicated that fathers were significantly less involved than mothers in all of the domains surveyed except providing income, and that the patterns of means for fathers (but not for mothers) were highly consistent with the seminal work of Parsons and Bales. Implications for the study of parenting are discussed.

Keywords: fathers, mothers, parenting, involvement, nurturance, scales

**********

Research on families has gained increasing momentum in recent decades, due in large measure to broad social changes in family forms and family roles. Although scholars long have recognized that parents are among the most important contributors to child development (e.g., Maccoby, 1992; Steinberg, 2001), more research is needed regarding the differential contributions of fathers and mothers (1) to developmental outcomes (cf. Andrews, Luckey, Bolden, Whiting-Fickling, & Lind, 2004; Finley & Schwartz, 2006).

Theorists and researchers recognize that mothers and fathers play different roles in the family system (Craig, 2006: Parke, 2002, 2004). Indeed, more than 50 years ago, Parsons and Bales (1955) wrote that fathers would be expected to fulfill largely instrumental functions, such as providing income and disciplining children, whereas mothers would be expected to fulfill largely expressive functions, such as caregiving, companionship, and sharing leisure activities. Parsons and Bales spoke not only of the division of parenting roles into instrumental and expressive components, but they also stated that these roles would be distributed by parent gender (i.e., instrumental roles for fathers and expressive roles for mothers). This distinction between instrumental and expressive involvement has been shown to apply well to young adults' perceptions of father involvement, both in terms of factor structure (Finley & Schwartz, 2004) and in terms of differential endorsement of instrumental versus expressive involvement (Finley & Schwartz, 2006). In the present study, one objective was to examine the applicability of Parsons and Bales' theoretical perspective to mothers, in terms of both factor structures and mean levels of endorsement.

Testing the extent to which Parsons and Bales' (1955) perspective is applicable to both mothers and fathers requires identical measures of mother and father involvement. Finley and Schwartz (2004) developed measures of young adults' retrospective reports of father involvement and nurturance. The development of parallel measures of mother involvement and nurturance provides us with a much needed opportunity to make direct comparisons between perceived mothering and fathering. Because obtaining equivalent factor structures across parents is a prerequisite to comparing mean levels of endorsement (Little, 1997; Vandenberg & Lance, 2000), we locus first on comparability of factor structure of nurturance and involvement between mothers and fathers, and second on mean levels of endorsement.

Conceptual and Methodological Approach to Parenting ill the Present Study

Three core conceptualizations underlie our approach to the study of parental involvement. First, parental involvement is viewed as a highly differentiated construct with many different domains of a child's life in which a parent may or may not be involved (cf. Hawkins & Palkovitz, 1999; Parsons & Bales, 1955). This position stands in contrast to "general" conceptualizations that assess parental involvement as a singular construct (e.g., Gorman-Smith, Tolan, Zelli, & Huesmann, 1996; Jacob, Moser, Windle, Loeber, & Stouthamer-Loeber, 2000). Indeed, prior research has found that fathers often are described as having been differentially involved in different domains of parenting (Finley & Schwartz, 2006). Second, in contrast to time-based measures of parenting, it is argued that what is most important is not the amount of time a parent actually spends with the child, but rather the child's perception of the parent's level and quality of involvement (Pleck, 1997). Third, this approach holds that the long-term impact that the parent has on the child is a function not of the parent's actual behavior, but rather of the child's long-term perceptions of parental behavior.

In our measures of fathering and mothering, we assess nurturance in addition to involvement. Parental nurturance refers to warmth and acceptance received from parents and is quite similar to what Rohner and colleagues (e.g., Rohner, 2004; Rohner, Khaleque, & Cournoyer, 2005) have studied as perceived parental acceptance-rejection. Involvement refers to the extent to which parents participate in various aspects of their children's lives (see Day & Lamb, 2003; Tamis-LeMonda & Cabrera, 2002, for collections of reviews). Following Hawkins and Palkovitz (1999), our measures assess parental involvement in 20 domains, including providing income, caregiving, monitoring schoolwork, encouraging independence, discipline, and companionship.

Exploratory and confirmatory factor analyses using a large and ethnically diverse sample of young adults (Finley & Schwartz, 2004) previously have shown that the original father involvement measure assesses the two primary dimensions of parenting outlined by Parsons and Bales (1955)--instrumental (e.g., discipline, protection, providing income) and expressive (e.g., companionship, caregiving, sharing leisure activities). Moreover, four items patterned onto both dimensions and were placed onto a third subscale labeled "mentoring/advising." Additional validation evidence for the instrumental-expressive distinction has been provided by demonstrating (a) that fathers were rated as significantly more involved in instrumental than in expressive domains (Finley & Schwartz, 2006) and (b) that instrumental fathering functions are more heavily compromised by nonresident status than are expressive fathering functions (Schwartz & Finley, 2005a,b).

When our conceptualization is extended to mothers, however, several new issues emerge. First, it is not known whether the fathering domains identified by Hawkins and Palkovitz (1999) and by Finley and Schwartz (2004) generalize to mothers--that is, are they inherently parenting functions that are applicable to both mothers and fathers? Second, provided that Parsons and Bales (1955) accurately described the factor structure of parenting, were they also correct in terms of mean levels of perceived involvement? That is, do mothers play primarily expressive roles, and do fathers play primarily instrumental roles? Challenges to the applicability of Parsons and Bales' work to contemporary parenting include broad social changes such as the social and sexual revolutions of the 1960s and 1970s, the rise of the dual-earner family, and the dramatic increase in single-mother families. All of these changes have thrust mothers into the role of breadwinners as well as caregivers (Giele & Holst, 2004). As a result, these social-structural changes require a reevaluation of the extent to which the structures of young adults' perceptions of mothering and fathering are similar or different, as well as the ex tent to which mothers and fathers are perceived to have been differentially involved in various areas of the child's life. Only by addressing these issues can we ascertain the extent to which Parsons and Bales' (1955) work applies to the contemporary study of parenting.

Third, in our program of research, we assess not only reports of the amount of parental involvement received, but also desires for more or less involvement. Desires for parental involvement reflect the degree to which the level of involvement reported was sufficient to meet the young adult child's perceived needs. Desired involvement is particularly important in divorced families, as it indicates the extent to which the custody arl'angements under which the person lived were or were not sufficient to provide the level of involvement that she or he wanted or needed (e.g., Fabricius, 2003; Fabricius & Braver, 2003, 2004; Fabricius & Hall, 2000; Marquardt, 2005). Indeed, Finley and Schwartz (2007) have found that desired father involvement serves as an important outcome of divorce and represents the degree to which the person expresses "missed opportunities" and "emotional longing" for a father-child relationship. Critically, however, the importance of desired parental involvement may not be limited to divorced families. As more and more mothers enter the paid labor force, and as the number of dual-earner families continues to increase (Sayer, 2005), children and adolescents may not be receiving sufficient involvement from either parent to meet their needs. Desired parenting may therefore become an increasingly important construct for both fathers and mothers in many different family forms (cf. Hawkins, Amato, & King, 2006; Lansford, Ceballo, Abbey, & Stewart, 2001).

Use of Retrospective Reports

Evaluating parental nurturance, reported involvement, and desired involvement across intact-biological, cohabiting, divorced, never-married, adoptive, step-, and adoptive stepfamilies, among other family forms, is a critically important research endeavor but presents serious methodological challenges (e.g., Finley & Schwartz, 2004; Schwartz & Finley, 2005a, 2006). One assessment method that can be used across family forms--and that can inform public policy on family issues--is to gather young adults' retrospective self-reports of parenting (cf. Fabricius, 2003; Hawkins et al., 2006; Laumann-Billings & Emery, 2000). Self-reports allow for statistically equating different father and mother figures across participants (e.g., where one person rates a biological father and another person rates a stepfather) and take into account the young person's perceptions of the parent or parent figure's involvement (cf. Finley & Schwartz, 2006; Schwartz & Finley, 2006). Moreover, there is evidence that young people's retrospective perceptions of their fathers' and mothers' nurturance and involvement are uniquely predictive of their developmental and behavioral outcomes (see Rohner & Britner, 2002: Rohner et al., 2005, for reviews).

Although the retrospective method may be vulnerable to recall bias, it does permit the gathering of overall, global ratings of parental nurturance and involvement from a more "mature" young adult perspective. As argued elsewhere (Finley & Schwartz, 2007), when young adults are no longer constrained either by their parents' or by the courts' legal control, and they may be fleer to speak their minds than are children or younger adolescents. This is an especially critical issue in divorced families. Many decisions regarding custody, for example, are based on judgments of "the best interests of the child"Ias reported by someone other than the child (Finley & Schwartz, 2007; Schwartz & Finley, 2005b). Minor children from divorced families still may be involved in the family court system and thus may not be comfortable revealing their true feelings about one or both parents. Young adults' retrospective self-reports are flee from these constraints and thus can be used to inform the legal system regarding children's views of their own "best interests" (Finley, 2002)--not for themselves, but for future generations of children of divorce.

Hypotheses

Three hypotheses were advanced. First, it was hypothesized that the factor structures of paternal nurturance and involvement reported by Finley and Schwartz (2004) would be replicated in the present study. Second, it was hypothesized that the factor structures obtained initially for fathers would represent truly parenting variables, in that the structures of perceived paternal nurturance and involvement would be metrically invariant (equivalent) across parents. For reported involvement, we expected that a three-factor solution (instrumental, expressive, and mentoring/advising) would emerge for mothers as well as fathers. For desired involvement, we expected that a two-factor solution, representing instrumental and expressive desired parenting, would emerge for both parents. Third, because the formulation advanced by Parsons and Bales (1955) applies both to factor structures and to mean levels of parental involvement (Finley & Schwartz, 2004, 2006), we also anticipated that mothers would score significantly higher than fathers on involvement in expressive domains, but that these patterns of differences would be reversed (or at least smaller) in instrumental domains.

Method

Participants

Participants in the present study were 1714 young-adult university students (75% female; mean age 19.90 years, SD 3.52, 92.1% between 18 and 29). Nearly half of all participants (49%) were freshmen, with the remainder being sophomores (20%), juniors (17%), seniors (12%), or graduate students (3%). In terms of ethnicity, 56% were Hispanic, 24% were non-Hispanic White, 15% were non-Hispanic Black, 5% were Asian, and 1% were classified as Other. The majority (80%) of data were gathered at a public university in Miami where the majority of students are Hispanic. To increase the number of non-Hispanics in the sample, additional data were gathered at another public university in Florida that serves a largely non-Hispanic White student population.

The majority (74%) of participants were born in the United States, whereas the majority of mothers (67%) and fathers (66%) were born abroad. The most common countries of origin for immigrant participants and parents were Cuba, Colombia, Nicaragua, Haiti, and Jamaica. Of those participants reporting socioeconomic status (during the participant's adolescence), 14% of participants reported annual family incomes below $30,000; 25% between $30,000 and $50,000; 36% between $50,000 and $100,000,15% between $100,000 and $150,000, and 10% above $150,000. Forty-one percent of fathers and 37% of mothers were college graduates, whereas 14% of fathers and 9% of mothers had not graduated high school.

Participants were asked to rate whomever they felt were the most important mother and father figures in their lives. The vast majority of fathers rated (91%) were biological fathers. Six percent were stepfathers, 0.5% were adoptive stepfathers, and 2% were other father figures (e.g., grandfathers, uncles). In terms of mother figures rated, 98% were biological mothers, 1% were adoptive mothers, and less than 1% rated adoptive stepmothers or other mother figures.

Measures

Nurturant mothering and fathering. Adult children's retrospective reports of maternal and paternal nurturance were measured using the Nurturant Fathering Scale (Finley & Schwartz, 2004) and using a "mother" version in which "father" was replaced with "mother." Each item is responded to on a five-point scale, with the scale endpoints varying as a function of item content. A sample item on this scale is "When you needed your father's (mother's) support, was he (she) there for you'?" Using a similar sample, Finley and Schwartz (2004) reported a Cronbach's alpha coefficient of .94 for nurturant fathering scores. Internal consistency estimates from the present sample are reported in the results section.

Mother and father involvement. Adult children's reports of maternal and paternal involvement were measured using the Father Involvement Scale (Finley & Schwartz, 2004) and using a "mother" version in which "father" was replaced with "mother." The measure consists of 20 domains of parenting (e.g., intellectual development, developing responsibility, caregiving, providing income). Each domain is listed between two blanks, where the participant is instructed to provide a reported involvement rating in the left-hand blank and a desired involvement rating in the right-hand blank. The prompt for reported involvement reads, "How involved was your mother/father in the following areas of your life and development?" The prompt for desired involvement read, "What did you want your mother's/father's level of involvement to be, compared to what it actually was?" The response scale for reported involvement within each domain ranges from 1 (never involved) to 5 (always involved). The response scale for desired involvement within each domain ranges from 1 (much less involved) to 5 (much more involved), where a response of 3 corresponds to "just right."

Procedure

Participants completed the nurturance and involvement scales, along with psychosocial functioning measures (presented subsequently) that were not analyzed for this report. The order of the nurturance and involvement measures (mother first versus father first) was counterbalanced to control for order effects. Most participants came to a research laboratory where they completed the measures individually in sessions of 10 participants each. Some participants completed their measures in class. Participants completed the entire assessment battery within approximately 30 minutes. Data were gathered between September 2004 and January 2006.

Results

Diagostics for Normality and Missing Data

Our first step was to conduct descriptive analyses and to ascertain the extent of nonnormality in the distributions for the nurturance and involvement items. Lei and Lomax (2005) specify that skewness and kurtosis values of 2.3 or below are not problematic for confirmatory factor analyses and other types of structural equation models. Absolute skewness and kurtosis values for the paternal nurturance items were all below 1.3, but kurtosis values for three of the maternal nurturance items were above 3.0. Similarly, absolute skewness and kurtosis values for the reported paternal involvement items were all below 1.7, but kurtosis values for one of the reported maternal involvement items was above 6.0. Mothers were rated as highly nurturant and involved on most of the items and domains. All absolute skewness and kurtosis values for the desired maternal and paternal involvement items were below 1.3. As a result, in our confirmatory factor analysis models, we used the robust maximum likelihood (MLR) estimator, which adjusts standard errors and model fit indices to account for nonnormality in the data (Satorra & Bentler, 1994).

We next examined patterns of missing data for each of our study variables. Less than 1% of participants were missing responses on any of the study variables. Nonetheless, we evaluated the extent to which the missing data that we did have could be considered to be missing at random. Little's (1988) MCAR (missing completely at random) test indicated that data were not missing at random, [chi square] (4898) = 5748.24, p < .001. However, given that maximum likelihood estimation can reduce bias even when data are not missing completely at random (Little & Rubin, 1989; Byrne, 2001), we proceeded to utilize maximum likelihood procedures to handle missing data in our analyses.

Replicating the Structure of the Fathering Scales

Because we already have some validation evidence for the fathering scales (Finley & Schwartz, 2004), and because the various dimensions of fathering previously identified have been shown to relate differentially to comparison variables (Finley & Schwartz, 2006, 2007; Schwartz & Finley, 2005b), a finding of good fit was regarded as successfully replicating the structure of the fathering scales. Provided that these analyses were successful, the mothering scales then were analyzed. In evaluating the fit of confirmatory factor analysis models, the comparative fit index (CFI), the nonnormed fit index (NNFI), and the root mean square error of approximation (RMSEA) were used, where good fit is represented by CFI [greater than or equal to] .95, NNFI [greater than or equal to] .90, and RMSEA [less than or equal to] .05 (Tomarken & Waller, 2003), and where .90 and .08 represent the lower and upper bounds for acceptable CFI and RMSEA values, respectively (Kline, 2006; Quintana & Maxwell, 1999).

A confirmatory factor analysis indicated that the one-factor solution for nurturant fathering fit the data well, [chi square] (23) = 169.11, p < .001; CFI = .98; NNFI = .97; RMSEA = .068. Standardized pattern coefficients ranged from .67 to .91, with a mean of .77 (see Table 1). We then examined the consistency of factor structure across gender and across the three largest ethnic groups in the sample (Whites, Blacks, and Hispanics) using multigroup invariance analyses. We did this because of the ethnic diversity of the sample, because of the imbalance of women in the sample, and because of the importance of child gender (Kilmann, Carranza, & Vendemia, 2006) and ethnicity (King, Harris, & Heard, 2004) in parenting research. In the invariance analyses by gender, the sample sizes were 420 for men and 1282 for women (12 participants did not indicate their gender and were not included in invariance analyses). In the invariance analyses by ethnicity, the sample sizes were 392 for Whites, 252 for Blacks, and 952 for Hispanics. For these invariance analyses, African Americans and Caribbean Islanders were combined into a single Black group. The numbers of Asian and mixed-ethnicity participants were too small for inclusion in invariance analyses.

The null hypothesis of invariance was tested by comparing a model with all factor pattern coefficients (loadings; Thompson, 2004) free to vary across groups to a model with all factor pattern coefficients constrained equal across groups (Vandenberg & Lance, 2000). Invariance testing involves examining differences in fit between the constrained and unconstrained models in terms of the chi-square index, the CFI, and the NNFI (Cheung & Rensvold, 2002; Little, 1997). A finding of equivalent fit between the constrained and unconstrained models is taken as evidence of invariance.

Two of the following three criteria had to be satisfied for the null hypothesis of invariance to be rejected: statistically significant [delta][chi square] (Byrne, 2001), ACFI [greater than or equal to] .01 (Cheung & Rensvold, 2002), and [delta]NNFI [greater than or equal to] .02 (Vandenberg & Lance, 2000). Using these criteria, this solution was equivalent across gender, [delta][greater than or equal to] (9) = 18.16, p < .004; [delta]CFI = .001; [delta]NNFI < .001; and across ethnicity, [delta][chi square] (2(1 8) = 30.94, p < .03; [delta]CFI = .003; [delta]NNFI < .001. The Cronbach's alpha estimate for scores on this scale was .93.

The three-factor solution for reported father involvement, as used in prior work (Finley & Schwartz, 2004, 2006, 2007; Schwartz & Finley, 2005a, 2005b, 2006), provided an adequate fit to the data, [chi square](161) = 1784.44, p < .001; CFI = .91; NNFI = .90; RMSEA = .081. Standardized pattern coefficients ranged from .62 to .83 for instrumental involvement, from .56 to .81 for expressive involvement, and from .76 to .79 for mentoring/advising (see Table 2). The reported father involvement solution provided equivalent fit across gender, [delta][chi square](20) = 33.38, p < .04; [delta]CFI = .001; [delta]NNFI < .001; and across ethnicity, [delta][chi square](40) = 135.70, p < .001; [delta]CFI = .008; [delta]NNFI = .001. Cronbach's alpha estimates in the present sample were: instrumental fathering, .90; expressive fathering, .91; and mentoring/advising, .88.

We used a two-factor solution for desired father involvement, rather than the three-factor solution that emerged for reported father involvement, for two reasons. First, the mentoring/advising subscale, which represents the conceptual and empirical overlap between instrumental and expressive parenting, did not emerge in our original analysis of desired fathering (Finley & Schwartz, 2004). Second, we have successfully used the two-factor solution for desired fathering in our prior work (Finley & Schwartz, 2007; Schwartz & Finley, 2005).

The two-factor solution for desired father involvement fit the data adequately, [chi square](166) = 1004.26, p < .001 ; CFI = .93; NNFI = .92; RMSEA = .055. This solution was equivalent across gender, [delta][chi square](20) = 74.94, p < .001; [delta]CFI = .004; [delta]NNFI < .001; but not across ethnicity, [delta][chi square](40) = 170.23, p < .001; [delta]CF1 = .012; [delta]NNFI = .006. Although the omnibus test of invariance across ethnicity suggested some noninvariance, follow-up analyses did not identify any domains whose pattern coefficients were not invariant across ethnicity. Cronbach's alpha estimates in the present sample were: desired instrumental fathering, .91; and desired expressive fathering, .91.

Scale Development for Nurturant Mothering and Mother Involvement

Nurturant mothering. A confirmatory factor analysis supported the one-factor solution taken from the fathering scales, [chi square](26) = 316.78, p < .001; CFI = .95; NNFI = .93; RMSEA = .081. Standardized pattern coefficients ranged from .60 to .85, with a mean of .72. Using the full sample, this solution fit equivalently across gender, [delta][chi square](8) = 13.28, p = .10; [delta]CFI = .002; [delta]NNFI < .001, and across ethnicity, [delta][chi square](16) = 22.79, p = .12; [delta]CFI = .001; [delta]NNFI < .001. In the present sample, the Cronbach's alpha estimate for scores on this scale was .90 (see Table 3).

We then compared the mother and father nurturance solutions against one another using an invariance test. This test was designed to evaluate the extent to which the nurturance factor structure was equivalent for mothers and fathers. This analysis indicated significant noninvariance, [delta][chi square](9) = 246.81, p < .001; [delta]CFI = .015; [delta]NNFI = .011. To explore the source of this noninvariance, we began with the fully unconstrained model and constrained one item at a time across parents (cf. Byrne, 2001). We examined the change in chi-square, CFI, and NNFI following each successive constraint. None of the items met or exceeded the criteria for rejecting the null hypothesis of invariance across parents, although the item referring to support was associated with the most evidence for noninvariance, [DELTA][chi square](I) = 28.08, [DELTA]CFI = .005: [DELTA]NNFI = .006. As a result, we concluded that the factor structure for nurturance was fairly equivalent across parents, but somewhat stronger for fathers.

Reported mother involvement. To test the extent to which the previously obtained factor structure for father involvement may in fact represent a parenting factor structure, we first imposed the father involvement factor structure onto the mother involvement data. For the reported mother involvement data, this produced an adequate fit, [chi square](163) = 1087.31, p < .001; CFI = .92; NNFI = .91 ; RMSEA = .058. Standardized pattern coefficients ranged from .39 to .72 for instrumental involvement, from .46 to .79 for expressive involvement, and from .69 to .76 for mentoring/advising (see Table 4). This solution was invariant across gender, [DELTA][chi square](2(20) = 84.99, p < .001; [DELTA]CFI = .006; [DELTA]NNFI < .001; and across ethnicity, [DELTA][chi square](40) = 135.70, p < .001; [DELTA]CFI = .008; [DELTA]NNFI = .001. Cronbach's alpha coefficients were .80 for instrumental mother involvement, .86 for expressive mother involvement, and .82 for mother mentoring/advising.

To ascertain the extent to which this factor structure was equivalent between mothers and fathers, we conducted an invariance analysis. This analysis indicated significant noninvariance between the mother and father factor solutions, [DELTA][chi square](20) = 298.43, p < .001; [DELTA]CFI = .011; [DELTA]NNF1 = .008. As a result, we began with the fully unconstrained model and constrained one domain at a time to identify the specific sources of noninvariance. However, none of the domains showed any evidence of noninvariance. As a result, we concluded that the factor structure for reported involvement was largely invariant across parents, but was slightly stronger for fathers.

Desired mother involvement. The two-factor solution used for desired lather involvement fit the desired mother involvement data adequately, [chi square](165) = 892.01, p < .001; CFI = .92; NNFI = .91; RMSEA = .051. Standardized pattern coefficients ranged from .54 to .77 for desired instrumental involvement and from .53 to .79 for desired expressive involvement (see Table 4). This solution was consistent across gender, [DELTA][chi square](20) = 75.11, p < .001; [DELTA]CFI = .006; [DELTA]NNFI = .001, and across ethnicity, [DELTA][chi square](40) = 126.43, p < .001; [DELTA]CFI = .009; [DELTA]NNFI = .003. Internal consistency estimates for the expressive and instrumental desired mothering subscales were .89 and .88, respectively.

We then tested for invariance of the desired parenting factor structure between mothers and fathers. Results indicated that the factor structure was invariant across parents, [DELTA][chi square](20) = 80.01, p < .001; [DELTA]CFI = .004; [DELTA]NNFI = .002.

Mean Differences between Reported Paternal and Maternal Involvement

To test our third hypothesis, we began by conducting a multivariate analysis of variance (MANOVA) on reported instrumental, expressive, and mentoring/advising subscale scores by parent. This analysis yielded a significant multivariate effect, Wilks' [lambda] = .65, F (3, 1683) = 300.43, p < .001, [[eta].sup.2] = .35. As shown in Table 5, exploration of univariate effects indicated that mother-father differences in reported involvement were fairly large on the expressive subscale ([[eta].sup.2] = .34) but were much smaller on the instrumental ([[eta].sup.2] = .12) and mentoring/advising ([[eta].sup.2] =. 13) subscales.

We then decomposed the subscales into their component items and conducted a follow-up MANOVA. Results were highly significant, Wilks' [lambda] = .46, F(20, 1645) = 97.23, p < .001, [[eta].sup.2] = .54. As shown in Table 5 and in Figure 1, the largest mean differences between perceived mothering and fathering emerged for the domains of emotional development, caregiving, spiritual development, companionship, and social development. These domains all pattern onto the expressive dimension of parenting. The smallest mean differences between perceived mothering and fathering emerged for the domains of providing income, career development, developing independence, and leisure/fun/play. Aside from leisure, these domains pattern onto the instrumental domain of parenting.

Figure 1 and Table 5 suggest the presence of several "break points" in the acrossdomain patterns of means for mothers and for fathers. Such break points would allow us to cluster those domains that were rated similarly from those that were rated higher or lower (Finley & Schwartz, 2006). For mothers, break points emerged between leisure/fun/play and school/homework, F(1, 1685) = 19.67, p < .001, [[eta].sup.2] = .01; between independence and competence, F(1, 1676) = 41.71, p < .001, [[eta].sup.2] = .02; between ethical/moral development and being protective, F(1,1684) = 27.48, p < .001, [[eta].sup.2] = .02; and between being protective and caregiving, F(1, 1685) = 32.41, p < .001, [[eta].sup.2] = .02. For fathers, break points emerged between spiritual and emotional development, F(1, 1685) = 58.11, p < .001, [[eta].sup.2] = .03; between school/homework and sharing activities, F(1, 1683) = 22.34, p < .001, [[eta].sup.2] = .01 ; between developing responsibility and being protective, F(1, 1684) = 58.23, p < .001, [[eta].sup.2] = .03; and between being protective and providing income, F(1, 1684) = 64.91, p < .001, [[eta].sup.2] = .04.

[FIGURE 1 OMITTED]

Discussion

The purpose of the present study was threefold. We first sought to replicate the factor structure of the nurturance and involvement scales for fathers (Finley & Schwartz, 2004). Second, we sought to determine the extent to which these scales were unique to fathers or represented truly parenting scales applicable to mothers as well as fathers. Third, we sought to determine the applicability of the Parsons and Bales theoretical formulation to both fathers (Finley & Schwartz, 2006) and to mothers--both in terms of factor structures and in terms of mean levels of endorsement.

The present findings can be summarized as follows. First, all of the nurturant fathering and father involvement scales have been closely replicated in the new sample and yielded Cronbach's alphas that were virtually identical to those reported earlier (Finley & Schwartz, 2004). Second, and perhaps most importantly, the mothering and fathering scales appear to be characterized by isomorphic factor structures. This finding implies that these scales tap into parenting functions that are structurally equivalent for mothers and for fathers. Although mothers and fathers clearly play different roles within the family system (Parke, 2004), the array of functions in which mothers and fathers can engage appears to be isomorphic. As a result, the list of domains compiled by Hawkins and Palkovitz (1999), with additional domains added by Finley and Schwartz (2004), appears to be appropriate for assessing involvement from mothers, from fathers, or from both.

Thus, the instrumental-expressive distinction, with a mentoring/advising factor representing the overlap between the instrumental and expressive dimensions (as originally reported by Finley & Schwartz, 2004, for fathers), is applicable to both mothers and fathers. The factor structures appear to strongly correspond to Parsons and Bales' (1955) division of parenting roles into instrumental and expressive domain clusters. Intriguingly, the goal in compiling the original set of domains was to most fully capture the set of functions in which fathers could be involved in their children's lives, without any consideration of Parsons and Bales' theoretical framework. This makes the strong consistency of the factor structures with Parsons and Bales all the more remarkable.

Third, regarding mean differences between parents, and across parenting functions within each parent, our findings suggest that mothers are rated as more involved than fathers in all domains studied, with the exception of providing income. Mothers were rated as "often" or "always" involved (means of 3.66 or higher on a 1-5 scale) on all of the domains studied, whereas fathers were rated as "sometimes involved" (means at or below 3.50) in 9 of the 20 domains surveyed. Moreover, whereas means for reported father involvement follow the pattern that would be expected given Parsons and Bales (1955)--seven of the nine lowest-rated domains for fathers were in the expressive domain--no similar pattern emerged for mothers. Nonetheless, it should be noted that, as shown in Figure 1, and consistent with Parsons and Bales, providing income was associated with the highest mean level of endorsement for fathers, whereas caregiving was associated with the highest mean level of endorsement for mothers (see also Christiansen 8,: Palkovitz, 2001).

The finding that mothers were rated as heavily involved in nearly all domains is not surprising, given that mothers have generally been found to be more involved than fathers in most studies (e.g., Craig, 2006). However, mean differences between reported maternal and paternal involvement were generally smaller for instrumental domains than for expressive domains. This provides some degree of support for Parsons and Bales' contention that fathers are more involved in instrumental roles than in expressive roles (cf. Finley & Schwartz, 2006). However, as shown in Figure 1 and in Table 5, our results do not suggest that the opposite is true for mothers. In fact, the three domains in which mothers were rated lowest were all expressive: physical development, sharing activities, and leisure/fun/play. What this implies is that neither parent is fulfilling these roles, which is not surprising given the ever-increasing amount of time that both men and women spend in work-related activities, including commuting (Sayer, 2005).

The present findings thus strongly suggest the need to draw separate sets of conclusions about factor structures versus about mean levels of endorsement. As reported in Finley and Schwartz (2006), the distribution of father involvement across domains, at least as perceived by young adult children, is highly consistent with what Parsons and Bales hypothesized more than 50 years ago. Separate analyses of the factor structures of reported and desired father involvement (Finley & Schwartz, 2004) yielded a very similar expressive-instrumental distinction. The present results are thus supportive of both of these sets of prior findings. However, for mothers, only factor structures--and not patterns of mean differences--were consistent with Parsons and Bales. Because mothers are often the primary caregivers for children and adolescents regardless of family form (Craig, 2006), the instrumental/expressive distinction is less descriptive of mothers than of fathers. However, and in deference to Parsons and Bales (1955), if one looks at the most highly endorsed parenting functions for fathers (providing income) and for mothers (caregiving), the results are fully supportive of Parsons and Bales' (1955) theoretical formulation.

Limitations

The limitations of the present study include using a sample comprised of college students and drawn from a heavily immigrant population. First, college students are a concern because they may represent a somewhat advantaged segment of the total population. University student samples may underrepresent individuals with financial, cognitive, social, and emotional handicaps. Additionally, in both the present sample and in the previous dataset, only a very small number of participants' parents had never been married to one another. As never-married parents represent an increasing share of the population, it is important to include them in studies on parenting. Second, the overrepresentation of immigrants and children of immigrants is simultaneously a strength and a limitation. The primary strength involves providing opportunities to ascertain whether or not findings previously obtained in largely White populations replicate in immigrant populations. For example, Finley and Schwartz (2007) found that desires for paternal involvement held by young adults from intact and divorced immigrant families were highly similar to those studied in native-born families in other parts of the United States. Such findings are important because immigrants and immigrant families will become increasingly prominent in the United States (and other Western countries) as the 21st century progresses (Day, 1996; Huntington, 2004). The limitation inherent in using a primarily immigrant sample is that non-Hispanic Whites and African Americans are underrepresented, relative to their shares of the total population. It should be noted, however, that the structures of scores generated each of the measures were consistent across ethnicity and gender. As a result, the scales are likely applicable to a wide range of ethnic groups.

Despite these limitations, the present research yields important insights into how children perceive their fathers and mothers, as well as into the roles that fathers and mothers enact within the family system. Perhaps of equal importance is the development of isomorphic scales for perceived fathering and mothering. These scales easily can be used to explore family issues and questions that require measurement of children's perceptions of their fathers and mothers. Moreover, the scales can be used across the entire spectrum of family forms found today. In sum, the present findings and scales have the potential to make substantial contributions to the empirical, applied, and family policy literatures.

We are grateful to a number of undergraduate and graduate students, including Tara Sheehan. Leif Elliott, Crystal Langlois, Mike Mira, Dax Rodriguez, and Santresa Stewart. We are also grateful to Ron Mullis for his help with off-site data collection.

References

Andrews, A.B., Luckey, I., Bolden. E., Whiting-Fickling, J., & Lind, K.A. (2004). Public perceptions about father involvement: Results of a statewide household survey. Journal of Family Issues, 25, 603-633.

Byrne, B.M. (2001). Structural equation modeling with AMOS: Basic concepts, applications, and programming. Mahwah, NJ: Lawrence Erlbaum Associates.

Cheung, G.W., & Rensvold, R.B. (2002). Evaluating goodness-of-fit indexes for testing measurement invariance. Structural Equation Modeling, 9, 233-255.

Christiansen, S., & Palkovitz, R. (2001). Why the "good provider" role still matters: Providing as a form of paternal involvement. Journal of Family Issues, 22, 84-106.

Craig, L. (2006). Does father care mean fathers share? A comparison of how mothers and fathers in intact families spend time with children. Gender and Society, 20, 259-281.

Day, J.C. (1996). Population projections of the United States by age, sex, race, and Hispanic origin: 1995 to 2050 (Current Population Report P25-1130). Washington, DC: U.S. Census Bureau.

Day, R.D., & Lamb, M.E. (Eds.). (2003). Conceptualizing and measuring father involvement. Mahwah, NJ: Lawrence Erlbaum Associates.

Fabricius, W.V. (2003). Listening to children of divorce: New findings on living arrangements, college support, and relocation that diverge from Wallerstein, Lewis, and Blakeslee. Family Relations, 52, 385-396.

Fabricius, W.V., & Braver, S.L. (2003). Non-child-support expenditures on children by nonresidential divorced fathers: Results of a study. Family Court Review, 41, 321-336.

Fabricius, W.V., & Braver, S.L. (2004). Expenditures on children and visitation time: A reply to Garfinkel, McLanahan, and Wallerstein. Family Court Review, 42, 350-362.

Fabricus, W.V., & Hall, J.A. (2000). Young adults' perspectives on divorce: Living arrangements. Family and Conciliation Courts Review, 38, 446-461.

Finley, G.E. (2002). The best interest of the child and the eye of the beholder. [Review of C. Panter-Brick & M.T. Smith (Eds.), Abandoned Children]. Contemporary Psychology, 47, 629-631.

Finley, G.E., & Schwartz, S.J. (2004). The Father Involvement and Nurturant Fathering Scales: Retrospective measures for adolescent and adult children. Educational and Psychological Measurement, 64, 143-164.

Finley, G.E., & Schwartz, S.J. (2006). Parsons and Bales revisited: Young adult children's characterization of the fathering role. Psychology of Men and Masculinity, 7, 42-55.

Finley, G.E., & Schwartz, S.J. (2007). Father involvement and long-term young adult outcomes: The roles of divorce and gender. Family Court Review, 45, 571-587.

Giele, J.Z., & Hoist, E. (2004). New life patterns and changing gender roles. Advances in Life Course Research, 8, 3-22. Gorman-Smith, D., Tolan, P.H., Zelli, A., & Huesmann, L.R. (1996). The relation of family functioning to violence among inner-city minority youths. Journal of Family Psychology, 10, 115-129.

Hawkins, A.J., & Palkovitz, R. (1999). Beyond ticks and clicks: The need for more diverse and broader conceptualizations and measures of father involvement. The Journal of Men's Studies, 8, 11-32.

Hawkins, D.N., Amato, P.R., & King, V. (2006). Parent-adolescent involvement: The relative influence of parent gender and residence. Journal of Marriage and Family, 68, 125-136.

Huntington, S.P. (2004). The Hispanic challenge. Foreign Policy, 141, 30-45.

Jacob, T., Moser, R.P., Windle, M., Loeber, R., & Stouthamer-Loeber, M. (2000). A new measure of parenting practices involving preadolescent- and adolescent-aged children. Behavior Modification, 24, 611-634.

Kilmann, P.R., Carranza, L.V., & Vendemia. J.M.C. (2006). Recollections of childhood experiences and attachment patterns for college women of intact versus non-intact families. Journal of Adolescence, 29, 89-102.

King, V., Harris, K.M., & Heard, H.E. (2004). Racial and ethnic diversity in nonresident fathers' involvement. Journal of Marriage and Family, 66, 1-21.

Kline, R.B. (2006). Principles and practice of structural equation modeling (2nd ed.). New York: Guilford.

Lansford, J.E., Ceballo. R., Abbey, A., & Stewart, A.J. (2001). Does family structure matter? A comparison of adoptive, two-parent biological, single-mother, stepfather, and stepmother households. Journal of Marriage and Family, 63, 840-851.

Laumann-Billings, L., & Emery, R.E. (2000). Distress among young adults from divorced families. Journal of Family Psychology, 14, 671-687.

Lei, M., & Lomax, R.G. (2005). The effect of varying degrees of nonnormality in structural equation modeling. Structural Equation Modeling, 12, 1-27.

Little, R.J.A. (1988). A test of missing completely at random for multivariate data with missing values. Journal of the American Statistical Association, 83, 1198-1202.

Little, R.J.A., & Rubin, D.B. (1989). The analysis of social science data with missing values. Sociological Methods and Research, 18, 292-326.

Little, T.D. (1997). Mean and covariance structures (MACS) analyses of cross-cultural data: Practical and theoretical issues. Multivariate Behavioral Research, 32, 53-76.

Maccoby, E.E. (1992). The role of parents in the socialization of children: An historical overview. Developmental Psychology, 28, 1006-1017.

Marquardt, E. (2005). Between two worlds: The inner lives of children of divorce. New York: Crown.

Parke, R.D. (2002). Fathers and families. In M.H. Bornstein (Ed.), Handbook of parenting: Vol. 3. Being and becoming a parent (pp. 27-73). Mahwah, NJ: Lawrence Erlbaum Associates.

Parke, R.D. (2004). Fathers, families, and the future: A plethora of possible predictions. Merrill-Pahner Quarterly, 50, 456-470.

Parsons, T., & Bales, R.F. (1955). Family, socialization, and interaction process. Glencoe. IL: Free Press.

Pleck, J.H. (1997). Paternal involvement: Levels, sources, and consequences. In M.E. Lamb (Ed.), The role of the father in child development (pp. 66-103). New York: John Wiley and Sons.

Quintana, S .M., & Maxwell, S.E. (1999). Implications of recent developments in structural equation modeling for counseling psychology. Counseling Psychologist, 27, 485-527.

Rohner, R.P. (2004). The parental "acceptance-rejection" syndrome: Universal correlates of perceived rejection. American Psychologist, 59, 830-840.

Rohuer, R.P., & Britner, P.A. (2002). Worldwide mental health correlates of parental acceptance-rejection: Review of cross-cultural and intracultural evidence. Cross-Cultural Research, 36, 16-47.

Rohner, R.P., Khaleque, A., & Cournoyer, D.E. (2005). Parental acceptance-rejection: Theory, methods, cross-cultural evidence, and implications. Ethos, 33, 299-334.

Satorra, A., & Bentler, P.M. (1994). Corrections to test statistics and standard errors in covariance structure analysis. In A. Von Eye & C.C. Clogg (Eds.), Latent variable analysis: Applications for developmental research (pp. 399-419). Thousand Oaks, CA: Sage.

Sayer, L.C. (2005). Gender, time, and inequality: Trends in women's and men's paid work, unpaid work, and free time. Social Forces, 84, 285-303.

Schwartz, S.J., & Finley, G.E. (2005a). Father involvement in intact and divorced families: Ethnic differences in retrospective reports. Journal of Marriage and Family, 67, 207-215.

Schwartz, S.J., & Finley, G.E. (2005b). Divorce-related variables as predictors of young adults' retrospective fathering reports. Journal of Divorce and Remarriage, 44(1-2), 145-163.

Schwartz, S.J., & Finley, G.E. (2006). Father involvement, nurturant fathering, and young adult psychosocial functioning: Differences among adoptive, adoptive stepfather, and non-adoptive stepfather families. Journal of Family Issues, 27, 712-731.

Steinberg, L. (2001). We know some things: Parent-adolescent relationships in retrospect and prospect. Journal of Research on Adolescence, 11, 1-19.

Tamis-LeMonda, C.S., & Cabrera, N. (2002). Handbook of father involvement: Multidisciplinary perspectives. Mahwah, NJ: Lawrence Erlbaum Associates.

Thompson, B. (2004). Exploratory and confirmatory factor analysis: Concepts and applications. Washington, DC: American Psychological Association.

Tomarken, A.J., & Waller, N.G. (2003). Potential problems with "well fitting" models. Journal of Abnormal Psychology. 112, 578-598.

Vandenberg, R.J., & Lance, C.E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3, 4-70.

(1) Given the wide array of family forms and of potential mother and father figures in children's lives (e.g., biological parents, stepparents, adoptive parents, grandparents), the terms "mother" and "father" are used here to refer to the individuals regarded as the primary mother and father figures in the child's life.

Gordon E. Finley and Sandra D. Mira, Department of Psychology, Florida International University; Seth J. Schwartz, Center for Family Studies, Department of Psychiatry and Behavioral Sciences, Leonard M. Miller School of Medicine, University of Miami.

Correspondence concerning this article should be addressed to Gordon E. Finley, Department of Psychology, Florida International University, University Park Campus, Miami. FL 33199. Electronic mail: [e-mail]

Table 1

Confirmatory Factor Analysis of the Nurturant Fathering Scale

Item Factor Pattern                          Factor      [[eta].sup.2]
                                             Pattern
                                           Coefficient

How much do you think your father
enjoyed being a father?                        .81            .66

When you needed your father's support,
was he there for you?                          .87            .78

Did your father have enough energy to
meet your needs?                               .77            .59

Did you feel that you could confide
in your father?                                .71            .50

Was your father available to spend time
with you in activities?                        .77            .59

How emotionally close were you to your
father?                                        .78            .61

When you were a teenager, how well did
you get along with your father?                .73            .53

Overall, how would you rate your father?       .91            .83

As you go through your day, how much of
a psychological presence does your
father have in your daily thoughts
and feelings?                                  .67            .45

Note. N = 1714.

Note. For this overall model, [chi square] (23) = 169.11,
p < .001; CFI = .98; NNFI =.97; RMSEA = .068.

Table 2

Confirmatory Factor Analysis of the Father Involvement Scale
(see notes on page 72)

Item                           Instrumental   Expressive

                                  Reported involvement

Developing responsibility          .83
Ethical/moral development          .77
Career development                 .75
Developing independence            .71
Being protective                   .71
School/homework                    .69
Discipline                         .67
Providing income                   .62
Companionship                                    .81
Caregiving                                       .80
Sharing activities/interests                     .77
Emotional development                            .77
Social development                               .76
Leisure/fun/play                                 .71
Physical development                             .66
Spiritual development                            .56
Mentoring
Developing competence
Advising
Intellectual development

                                  Desired involvement

Developing responsibility          .81
Developing competence              .78
Ethical/moral development          .77
Advising                           .76
Career development                 .71
Developing independence            .68
Being protective                   .67
Providing income                   .62
School/homework                    .62
Discipline                         .59
Mentoring                                        .79
Caregiving                                       .77
Companionship                                    .76
Intellectual development                         .74
Social development                               .71
Emotional development                            .71
Sharing activities/interests                     .70
Leisure/fun/play                                 .69
Physical development                             .64
Spiritual development                            .55

Item                            Mentoring/    [[eta].sup.2]
                                 advising

                                     Reported involvement

Developing responsibility                        .68
Ethical/moral development                        .59
Career development                               .56
Developing independence                          .50
Being protective                                 .50
School/homework                                  .47
Discipline                                       .44
Providing income                                 .38
Companionship                                    .65
Caregiving                                       .64
Sharing activities/interests                     .59
Emotional development                            .60
Social development                               .58
Leisure/fun/play                                 .51
Physical development                             .44
Spiritual development                            .31
Mentoring                          .79           .63
Developing competence              .77           .59
Advising                           .77           .60
Intellectual development           .76           .58

                                   Desired involvement

Developing responsibility                        .66
Developing competence                            .61
Ethical/moral development                        .59
Advising                                         .58
Career development                               .50
Developing independence                          .46
Being protective                                 .45
Providing income                                 .38
School/homework                                  .38
Discipline                                       .35
Mentoring                                        .62
Caregiving                                       .59
Companionship                                    .58
Intellectual development                         .55
Social development                               .50
Emotional development                            .50
Sharing activities/interests                     .49
Leisure/fun/play                                 .48
Physical development                             .41
Spiritual development                            .30

Notes by table 2, page 71.

N = 1714. For the reported father involvement model, [chi square](161)
= 1784.44, p < .001 ; CFI = .91 ; NNFI = .90; RMSEA = .081. For the
desired father involvement model, [chi square](166) = 1004.26, p <
.001; CFI = .93; NNFI = .92; RMSEA = .055.

Table 3

Confirmatory Factor Analysis of the Nurturant Mothering Scale

Item                                         Factor      [[eta].sup.2]
                                             pattern
                                           coefficient

How much do you think your mother
enjoyed being a mother?                        .73            .53

When you needed your mother's support,
was she there for you?                         .84            .71

Did your mother have enough energy
to meet your needs?                            .67            .45

Did you feel that you could confide
in your mother?                                .68            .46

Was your mother available to spend time
with you in activities?                        .65            .42

How emotionally close were you to your
mother?                                        .77            .59

When you were a teenager, how well did
you get along with your mother?                .68            .46

Overall, how would you rate your mother?       .85            .72

As you go through your day, how much
of a psychological presence does your
mother have in your daily thoughts and
feelings?                                      .60            .36

Note. N = 1714.

Note. For the full model, [chi square] (26) = 316.78,
p < .001; CFI = .95; NNFI = .93; RMSEA = .081.

Table 4

Confirmatory Factor Analysis of the Mother Involvement Scale
(see notes on page 75)

Item                           Instrumental    Expressive

                                  Reported involvement

Developing responsibility          .72
Ethical/moral development          .66
Career development                 .69
Developing independence            .57
Being protective                   .51
School/homework                    .61
Discipline                         .52
Providing income                   .39
Companionship                                      .79
Caregiving                                         .64
Sharing activities/interests                       .74
Emotional development                              .78
Social development                                 .72
Leisure/fun/play                                   .67
Physical development                               .58
Spiritual development                              .46
Mentoring
Developing competence
Advising
Intellectual development

                                  Desired involvement

Developing responsibility          .73
Developing competence              .72
Ethical/Moral development          .77
Advising                           .72
Career development                 .64
Developing independence            .56
Being protective                   .61
Providing income                   .63
School/homework                    .59
Discipline                         .54
Mentoring                                          .73
Caregiving                                         .79
Companionship                                      .75
Intellectual development                           .72
Social development                                 .63
Emotional development                              .74
Sharing activities/interests                       .64
Leisure/fun/play                                   .58
Physical development                               .57
Spiritual development                              .53

Item                            Mentoring/    [[eta].sup.2]
                                 advising

                                    Reported involvement

Developing responsibility                          .52
Ethical/moral development                          .44
Career development                                 .48
Developing independence                            .32
Being protective                                   .26
School/homework                                    .37
Discipline                                         .27
Providing income                                   .15
Companionship                                      .62
Caregiving                                         .41
Sharing activities/interests                       .55
Emotional development                              .61
Social development                                 .52
Leisure/fun/play                                   .45
Physical development                               .34
Spiritual development                              .21
Mentoring                          .76             .58
Developing competence              .69             .48
Advising                           .76             .58
Intellectual development           .69             .48

                                   Desired involvement

Developing responsibility                          .53
Developing competence                              .52
Ethical/Moral development                          .59
Advising                                           .52
Career development                                 .41
Developing independence                            .31
Being protective                                   .37
Providing income                                   .40
School/homework                                    .35
Discipline                                         .29
Mentoring                                          .53
Caregiving                                         .62
Companionship                                      .56
Intellectual development                           .52
Social development                                 .40
Emotional development                              .55
Sharing activities/interests                       .41
Leisure/fun/play                                   .34
Physical development                               .32
Spiritual development                              .28

Notes by Table 4, page 74.

N = 1714. For the reported mother involvement model, [chi square](163)
= 1087.31, p < .001 ; CFI = .92; NNFI = .91 ; RMSEA = .058. For the
desired mother involvement model, [chi square](165) = 892.01, p <
.001; CFI = .92; NNFI = .91; RMSEA = .051.

Table 5

Mean Differences in Reported Involvement Subscales
and Items by Parent

Domain/Subscale        Fathers        Mothers
Subscales              (M, SD)        (M, SD)

Expressive           3.23 (l.02)    3.97 (0.76)
Instrumental         3.80 (0.97)    4.19 (0.66)
Mentoring/Advising   3.62 (1.09)    4.09 (0.83)

Domains
Emotional            3.11 (l.21)    4.17 (0.98)
Caregiving           3.63 (l.33)    4.66 (0.74)
Spiritual            2.86 (l.43)    3.86 (l.24)
Companionship        3.34 (l.31)    4.17 (l.04)
Social               3.12 (l.24)    3.88 (l.05)
Ethical/Moral        3.81 (1.25)    4.44 (0.87)
School/Homework      3.15 (l.37)    3.82 (1.21)
Advising             3.73 (l.25)    4.30 (0.92)
Physical             3.12 (l.31)    3.66 (1.18)
Responsibility       3.91 (1.26)    4.37 (0.92)
Being Protective     4.11 (l.21)    4.56 (0.79)
Intellectual         3.62 (l.24)    4.10 (l.00)
Mentoring            3.45 (l.32)    3.90 (1.13)
Competence           3.67 (l.26)    4.07 (1.04)
Sharing Activities   3.31 (1.24)    3.67 (1.11)
Discipline           3.89 (1.28)    4.28 (0.97)
Leisure/Fun/Play     3.38 (l.28)    3.68 (1.11)
Independence         3.62 (1.32)    3.93 (1.15)
Career               3.56 (l.36)    3.85 (1.15)
Providing Income     4.33 (1.15)    4.27 (1.08)

Domain/Subscale        F Ratio     [[eta].sup.2]
Subscales

Expressive             853.10 *         .34
Instrumental           237.93 *         .12
Mentoring/Advising     257.43 *         .13

Domains
Emotional             1037.95 *         .38
Caregiving             898.89 *         .35
Spiritual              651.83 *         .28
Companionship          532.31 *         .24
Social                 503.56 *         .23
Ethical/Moral          370.55 *         .18
School/Homework        293.73 *         .15
Advising               260.34 *         .14
Physical               195.88 *         .11
Responsibility         182.68 *         .10
Being Protective       186.27 *         .10
Intellectual           169.43 *         .09
Mentoring              140.51 *         .08
Competence             141.06 *         .08
Sharing Activities     103.55 *         .06
Discipline             104.47 *         .06
Leisure/Fun/Play        73.58 *         .04
Independence            71.73 *         .04
Career                  57.66 *         .03
Providing Income         2.35           .00

p < .001