Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Jun 6.
Published in final edited form as: J Fam Psychol. 2012 Apr 30;26(3):400–410. doi: 10.1037/a0028330

Family Members' Unique Perspectives of the Family: Examining their Scope, Size, and Relations to Individual Adjustment

Justin Jager 1, Marc H Bornstein 1, L Putnick Diane 1, Charlene Hendricks 1
PMCID: PMC3368101  NIHMSID: NIHMS373231  PMID: 22545933

Abstract

Using the Family Assessment Device (FAD; Epstein, Baldwin, & Bishop, 1983) and incorporating the perspectives of adolescent, mother, and father, this study examined each family member's “unique perspective” or non-shared, idiosyncratic view of the family. To do so we used a modified multitrait-multimethod confirmatory factor analysis that (1) isolated for each family member's six reports of family dysfunction the non-shared variance (a combination of variance idiosyncratic to the individual and measurement error) from variance shared by one or more family members and (2) extracted common variance across each family member's set of non-shared variances. The sample included 128 families from a U.S. East Coast metropolitan area. Each family member's unique perspective generalized across his or her different reports of family dysfunction and accounted for a sizable proportion of his or her own variance in reports of family dysfunction. Additionally, after holding level of dysfunction constant across families and controlling for a family's shared variance (agreement regarding family dysfunction), each family member's unique perspective was associated with his or her own adjustment. Future applications and competing alternatives for what these “unique perspectives” reflect about the family are discussed.

Keywords: Multiple perspectives, unique perspectives, family system, adolescence, multitrait-multimethod

Different family members can experience the same event in different ways (Bartle-Haring, Kenny, & Gavazzi, 1999; Deković & Buist, 2005), as evidenced by studies indicating that different family members' answers to the same measures of the family or dyad often correlate at low to moderate levels (Caster, Inderbitzen, & Hope, 1999; Tein, Roosa, & Michaels, 1994). Measurement error contributes to the modest correlation among family member reports, but real differences in perspective contribute as well (Cook & Goldstein, 1993; Eisler, Dare, & Szmukler, 1988). Recognizing that each family member's perspective of the family is to some degree subjective, researchers interested in the family often incorporate the perspectives of multiple family members and focus on where those perspectives overlap or converge. Relative to a single individual's perspective of the family (say a mother), these “shared perspectives” are thought to yield a more reliable and objective picture of the family system (Bartle-Haring et al., 1999; Deković & Buist, 2005).

This line of research has neglected the portion of each family member's perspective of the family that is not shared by any other family member; that is, the portion of each family member's perspective that is idiosyncratic to the individual. Yet, many acknowledge the potential importance of each family member's non-shared or unique perspective to the inner-workings of the family and the adjustment of individual family members (Carlson, Cooper, & Spradling, 1991; Cook & Goldstein, 1993; Deal, 1995). In this paper, we give full attention to family members' unique perspectives of the family and consider their scope and magnitude as well as their relations to the adjustment of individual family members. More specifically, using the Family Assessment Device (FAD; Epstein, Baldwin, & Bishop, 1983), a multi-dimensional self-report measure of family functioning, and incorporating the perspectives of the mother, father, and their firstborn adolescent child, this study focuses on each family member's “unique perspective”. We focus on families with adolescent children because differences among family members' perceptions of family climate and familial relationships are thought to be elevated among families with adolescent children (Laursen & Collins, 2009; Smetana, 1989).

Shared Perspectives as Measures of the Family System

“Shared perspectives” represent the portions of the individual perspectives that generalize across, or are shared by, multiple family members. Using the perspectives of a hypothetical three-member family, the two types of shared-perspectives are illustrated graphically in Figure 1. Figure 1 is conceptual in nature and assumes perfect measurement; it is not intended to illustrate the degree to which different perspectives overlap, but merely the ways in which different perspectives can overlap. Where all perspectives overlap or converge is the family perspective, a shared perspective that is a measure of the family system. Where two perspectives overlap are also shared perspectives, but while the family perspective is a measure of the family system, these dyad perspectives are measures of family sub-systems, specifically the mother-adolescent, father-adolescent, and marital sub-systems. Note that there is a portion of each dyad perspective that does not overlap with the family perspective. These dyad-specific perspectives represent the commonality between two family members' perspectives that is specific to a particular dyad.

Figure 1.

Figure 1

Open in a new tab

Diagram of shared and unique perspectives among a three-member family

Depending on the area of the family system under examination, research to date has focused on either the family perspective (Bögels & van Melick, 2004; Schwarz, Barton-Henry, & Pruzinsky, 1985) or dyad perspectives (Bartle-Haring et al., 1999; Cook, 2001; Martin & Cole, 1993), while largely ignoring unique perspectives. In Figure 1, unique perspectives are the portion of each family member's perspective that does not overlap with any other family member's perspective. Conceptually, this is the view of the family that is idiosyncratic to each individual. Unique perspectives have gone unaddressed despite the fact that researchers have long recognized that they exist and many have even speculated on their relevance to the health of the family and its individual members (Carlson et al., 1991; Cook & Goldstein, 1993; Deal, 1995).

Directing Focus to the Unique Perspective

Incorporating the perspectives of the adolescent, mother, and father, this study has two broad aims. Our first aim, more descriptive in nature, is to examine the generalizability and magnitude of each family member's unique perspective. Our second aim is to examine how each family member's unique perspective relates to his or her own adjustment.

The generalizability and magnitude of the unique perspectives

Family health is multi-dimensional and based on a host of distinct factors, such as communication, warmth, and problem solving (Epstein et al., 1983; Feldman, Wentzel, & Gehring, 1989). The extent that a family member's unique perspective of one domain of family functioning, say communication, is similar or related to his or her unique perspective of every other domain of family functioning (e.g., they are all characterized by positivity or all characterized by negativity) is unclear. Opposed to being generalized, it is possible that each family member has multiple unique perspectives, each specific to a different domain of family functioning (the specificity principle; Bornstein, 2006). The question of generalizability is an important one because the more each family member's unique perspective generalizes across multiple domains of family functioning, likely the more it reflects something about the individual family member that bleeds across all aspects of the family system, and in cases of intervention or family therapy, the more it can be treated as a singular target of intervention.

Even if each family member's unique perspective proves to be generalized, it could still be of little theoretical or clinical use if it only accounts for a trivial amount of that person's overall perception of the family. Theoretically, this is possible and would amount to the overlap in different family members' perspectives being so sizable that, after accounting for measurement error, the shared perspectives (i.e., family- and dyad-specific-perspectives) account for nearly all of each individual's perception of the family. Therefore, in addition to generalizability, we also examine the magnitude of each family member's unique perspective.

The unique perspective and individual adjustment

Even if every family member's unique perspective generalizes across multiple domains of family functioning and is non-trivial in size, it would still be unclear whether it captures important information about individual family members and/or the family system and is, therefore, of substantive interest to researchers and clinicians. Adolescent (Amato & Cheadle, 2000; Rothbaum & Weisz, 1994) and parent (Burke, 2003; Cummings, Keller, & Davies, 2005) adjustment (e.g., depressive affect, stress, and self-esteem) are all clearly linked to the health of the family system. As a result, there is clear reason to expect that a family's shared perspective (i.e., the family perspective) of the health of the family is related to the adjustment of individual family members. However, to the extent that a family member's unique perspective of the family is characterized by family dysfunction, that view, even if erroneous, could also influence his or her adjustment (Bögels & van Melick, 2004; Furman, Jackson, Downey, & Shears, 2003). Therefore, while controlling for a family's shared perspective, the study's second aim is to examine the relation between each family member's unique perspective and his or her own adjustment.

Capturing the Unique Perspectives Empirically

We use a modified version of a multitrait-multimethod (MTMM) Confirmatory Factor Analysis (CFA) to identify and examine family members' unique perspectives (Figure 2). A derivative of the MTMM correlation matrix (Campbell & Fiske, 1959), the MTMM-CFA approach assumes that there are measures of several constructs or “traits” with each construct being measured by several methods. The aim of a MTMM-CFA is to isolate the trait variance from the method variance, because method variance is typically sizable but of little theoretical interest. By reinterpreting shared variance across different family members as “trait” variance, and variance specific to each family member as “method” variance, the MTMM-CFA approach can be adapted to the family system. Because the FAD is a measure of the family-system, opposed to family sub-systems, we extract family perspectives, opposed to dyad-perspectives, by loading all three family members' reports on a single factor. Unlike a conventional MTMM-CFA model, we also covary different family members' reports of the same FAD subscale to capture variance associated with dyad-specific perspectives. After extracting variance shared with both other family members (i.e., family-perspective factors) and any additional variance shared with one other family member (i.e., across-reporter residual covariances), the “non-shared” variance remaining for each observed variable is a combination of variance idiosyncratic to the individual (i.e., their unique perspective) plus measurement error. We isolate variance idiosyncratic to the individual from measurement error by loading the measures of a given family member on a single factor. Because family members' unique perspectives are independent of each other, we fix covariance among different family members' unique perspectives at zero. To test for construct validity, we compare the association between the family perspectives and amount of family dysfunction to the associations between the unique perspectives and amount of family dysfunction. Relative to the family perspective, family members' unique perspectives are thought to be less tethered to reality and more subjective (Deal, 1995) and they therefore should correlate more weakly with a family's amount of dysfunction.

Figure 2.

Figure 2

Open in a new tab

Modified MTMM-CFA Model. White ovals are family perspectives; Gray ovals are unique perspectives. A = adolescent, M = mother, F = father. PS = problem solving; CM = communication; AR = affective responsiveness; RL = roles; AI = affective involvement; BC = behavioral control.

We are not the first to adapt the MTMM-CFA approach to the family system. Family researchers focused on shared-perspectives have already done so (Cole & McPherson, 1993; Cook, 2001; Martin & Cole, 1993), most commonly by using the correlated uniqueness (CU) approach (Kenny & Kashy, 1992). Though a CU approach extracts only shared-perspectives (i.e., trait factors), it separates the shared variance from the unique variance by correlating all within-reporter residual variances with one another. Because the CU approach does not extract unique perspectives (i.e., method factors), it does not allow for the examination of family members' unique perspectives and is not appropriate for this study.

Methods

Participants

Families were originally recruited through mass mailings and newspaper advertisements from a U.S. East Coast metropolitan area. Out of the 185 families providing data, analyses were limited to those 128 families providing complete family functioning (i.e., FAD) data from adolescent, mother, and father. Levels of family functioning as well as socioeconomic status did not differ between those included and those excluded from the analyses. All children were firstborns, around 14 years of age, and 49 (38%) were girls. The sample consisted of European American families who were mostly intact (92%), of middle- to upper-socioeconomic status (Hollingshead, 1975) with a mean of 55.11 (SD = 9.02, range = 30–66), and well-educated (80% and 75% of mothers and fathers, respectively, had a college degree or greater). This sample is limited to European Americans because ethnic differences in the effects of parenting (Park & Bauer, 2002), parent-adolescent relations (Hofferth, 2003), and adjustment (Jager, 2011; Twenge & Crocker, 2002) could cloud the effects of this study if ethnic groups were combined.

Procedure

Both a home visit and laboratory visit were scheduled; only children and their mothers participated in the laboratory visit. For families living significant distances from the laboratory, all visits were conducted in the home. Though data are cross-sectional, data collection spanned 2004 to 2009. Participants were compensated for their time.

Measures

Family dysfunction

We used the McMaster Family Assessment Device (FAD; Epstein et al., 1983) to measure family functioning. The FAD totals 60 items, each having a possible response range of 1 (strongly disagree) to 4 (strongly agree). The 60 items form 7 subscales: Problem Solving (6 items; e.g., “We resolve most everyday problems around the house” (reverse coded); αa = .68, αm = .78, αf = .65), Communication (9 items; e.g., “When someone is upset the others know why”; αa = .72, αm = .84, αf = .79), Affective Responsiveness (6 items, e.g., “We are reluctant to show our affection for each other”; αa = .73, αm = .70, αf = .80), Roles (11 items; e.g., “We make sure members meet their family responsibilities” (reverse coded); αa = .73, αm = .65, αf = .68), Affective Involvement (7 items; e.g., “If someone is in trouble the others become too involved”; αa = .69, αm = .73, αf = .70), Behavior Control (9 items; e.g., “You can easily get away with breaking the rules”; αa = .71, αm = .77, αf = .73), and General Functioning (12 items; e.g., “We do not get along well together”; αa = .86, αm = .87, αf = .88). The FAD and its sub-scales have strong internal consistency, adequate test-retest reliability, as well as concurrent and discriminant validity (Epstein et al., 1983; Miller, Epstein, Bishop, & Keitner, 1985). FAD's factor structure is also invariant across clinical and non-clinical samples (Kobacoff, Miller, Bishop, Epstein, & Keitner, 1990). For each family member (i.e., adolescent, mother, and father), the mean scores of the FAD sub-scales (excluding general functioning) were used. Higher values indicate greater family dysfunction. To create a measure of mean family dysfunction, the adolescent, mother, and father general functioning subscales were averaged.

Individual adjustment

Adolescent externalizing, internalizing, and global self-worth were assessed by self-report. Externalizing behaviors were assessed using the combined aggressive behavior and delinquent behavior subscales of the Youth Self-Report Inventory (YSR; Achenbach & Rescorla, 2001). Internalizing behaviors were assessed using the combined withdrawal, somatic complaints, and depression-anxiety subscales of the YSR. Each item was rated on a scale of 0 (not true) to 2 (very true or often true). The externalizing scale was the average of 30 items (α = .90), and the internalizing scale was the average of 31 items (α = 84). Global self-worth was assessed using the average of 5 items (α = .88) from the Self-Perception Profile for Adolescents (Harter, 1988). Each item is scored on a 4-point scale with a higher score reflecting a more positive view of self.

Parent depressive affect, distress, and satisfaction with parenting were also assessed by self-report. Depressive affect was assessed using the Center for Epidemiological Studies Depression Scale (CES-D; Radloff, 1977), a measure of depressive symptomatology during the last week. The scale consists of the average of 20 items (αm = .89, αf = .90), each rated from 1 (rarely or none of the time) to 4 (most or all of the time). Distress was assessed using the 12-item (αm = .81, αf = .83) Parental Distress sub-scale of the short-form of the Parenting Stress Index (PSI; Abidin, 1995). The parental distress sub-scale of the PSI was designed to assess the distress a parent experiences in his or her role as a parent due to personal factors that are directly (e.g., “I feel trapped by my responsibilities as a parent”) and indirectly (e.g., “I often have the feeling that I cannot handle things very well”) related to his or her being a parent. Items were rated on a scale of 1 (strongly disagree) to 5 (strongly agree). The items were summed to create an index of distress. Satisfaction with parenting was assessed using the 5-item satisfaction sub-scale of the Self-Perception of the Parental Role scale (SPPR; MacPhee, Benson, & Bullock, 1985). Data were available for mothers only (α = .79). Each item is scored on a 4-point scale with a higher score reflecting greater satisfaction. The mean score of the satisfaction scale was used.

Social desirability

The 13-item Social Desirability Scale (SDS-SF; Reynolds, 1982) was used to assess social desirability bias. Statements like “I'm always willing to admit when I make a mistake” were rated as True or False. The SDS-SF's correlation with the full-length SDS is .93 (Reynolds, 1982). Data were available for mothers (α = .70) and fathers (α = .71) only.

Results

All analyses were conducted with Mplus Version 5.2 (Muthén & Muthén, 1998–2009), and utilized a maximum likelihood estimator that is robust to non-normality. Basic descriptive statistics are presented in Table 1. Both within- and across-reporter correlations were consistently positive. However, within-reporter correlations were stronger (more positive) than across-reporter correlations. Thus, there was more similarity in how the same reporter viewed different aspects of family functioning than there was in how different reporters viewed the same aspect of family functioning. Perceived level of family dysfunction varied by reporter. Mothers viewed the family as less dysfunctional, whereas fathers and adolescents typically reported levels of family dysfunction that were equivalent to one another and higher relative to mothers. Where differences emerged between adolescents and fathers, adolescents reported higher levels of family dysfunction – and therefore the highest overall levels.

Table 1.

intercorrelations of FAD subscales from all family members' perspectives

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 M SD
Adolescent
 1. Problem solving -- 2.09m .37
 2. Communication .68 -- 2.16m,f .35
 3. Affective responsiveness .63 .63 -- 2.14m .45
 4. Roles .67 .66 .53 -- 2.11 .32
 5. Affective involvement .45 .49 .43 .62 -- 2.14m .40
 6. Behavioral control .49 .48 .42 .57 .45 -- 1.82m .36
Mother
 7. Problem solving .31 .20 .33 .29 .25 .20 -- 1.96a,f .35
 8. Communication .37 .34 .42 .34 .26 .21 .73 -- 1.93a,f .36
 9. Affective responsiveness .40 .32 .40 .32 .28 .32 .66 .62 -- 1.83a,f .39
 10. Roles .23 .26 .25 .26 .15 .20 .59 .57 .50 -- 2.12 .28
 11. Affective involvement .22 .22 .22 .19 .23 .13 .55 .58 .52 .64 -- 1.95a,f .34
 12. Behavioral control .35 .34 .33 .34 .26 .24 .56 .58 .52 .54 .50 -- 1.58a,f .33
Father
 13. Problem solving .15 .15 .19 .27 .27 .16 .41 .18 .22 .22 .11 .14 -- 2.05a,m .31
 14. Communication .09 .27 .19 .25 .26 .11 .28 .23 .20 .14 .09 .10 .69 -- 2.05a,m .34
 15. Affective responsiveness .26 .23 .37 .32 .24 .30 .28 .23 .46 .18 .15 .13 .55 .58 -- 2.07m .46
 16. Roles .14 .09 .11 .28 .27 .12 .18 .07 .11 .21 .13 .07 .45 .40 .33 -- 2.14 .30
 17. Affective involvement .19 .07 .14 .23 .20 .05 .26 .06 .17 .16 .14 .12 .56 .53 .41 .53 -- 2.07m .36
 18. Behavioral control .18 .14 .10 .27 .23 .19 .27 .17 .20 .17 .10 .11 .57 .49 .40 .56 .53 -- 1.76m .34
Open in a new tab
m

Estimate differs from mother estimate at .05 level;

f

Estimate differs from father estimate at .05 level;

a

Estimate differs from adolescent estimate at .05 level

Note Within-reporter, across-subscale correlations are in bold; across-reporter, within-subscale correlations are underlined

Preliminary Analyses

With CFAs in general, and MTMM-CFAs in particular, misidentified solutions are common and important to avoid because, even when providing an excellent fit, misidentified solutions can still yield biased or incorrect parameter estimates (Kenny & Kashy, 1992; Marsh, 1989). Common forms of misidentification include out of range estimates, Heywood cases (negative error variances), and non-convergence (Kenny & Kashy, 1992; Marsh, 1989; Rindskopf, 1984). Regarding MTMM-CFAs, the most common cause of model misidentification is model misspecification due to extracting more factors than the data support (Rindskopf, 1984). Aside from indications of misidentification, other indicators of over-factoring are (a) poor discriminant validity (i.e., high correlations among latent factors) and (b) poor convergent validity (i.e., a sizable proportion of the loadings for a particular factor are small and non-significant). MTMM-CFA models using the FAD may be particularly susceptible to over factoring because full-factor CFA models (i.e., a model that extracts all 6 theorized factors of the FAD) often result in misidentification and/or poor model fit, especially relative to models extracting far fewer factors (Aarons, McDonald, Connelly, & Newton, 2007; Ridenour, Daley, & Reich, 1999). Given these susceptibilities, before examining our two research aims, we first identified the superior MTMM-CFA model. We did so by maximizing model fit while eliminating all indications of model misidentification and misspecification. We then examined the superior model's construct validity. We discuss the specifics of each below.

Identification of superior MTMM-CFA

Using the model in Figure 2 as a starting point, we identified the superior MTMM-CFA by varying the number of family perspectives extracted while holding the number of unique perspectives extracted constant at three (i.e., one each for adolescent, mother, and father). The number of family perspectives extracted could range from one (a single family perspective that incorporates all sub-scales of the FAD) to six (a separate family perspective for each sub-scale of the FAD). We used the following criteria to determine the superior MTMM-CFA: (a) indices and change of model fit (e.g., χ2, CFI, RMSEA) from nested models, (b) indications of model misidentification, and (c) the degree of convergent and discriminant validity. Out of all the models fitted, all those extracting three or more family perspectives displayed poor convergent validity and suffered from model misidentification, and most displayed poor discriminant validity as well. A one-family perspective model, χ2(99) = 119.85, p = .08, CFI = .971, RMSEA = .045, and a two-family perspective model, χ2(98) = 114.29, p = .13, CFI = .981, RMSEA = .036, each provided a fit that was excellent as well as superior to all models with three or more family perspectives. Both models also showed no indications of model misidentification. However, of these two models, the two-family perspective model displayed the better (and best overall) convergent validity and fit the data better, Δχ2(1) = 5.56, p < .05, and best overall. The final “superior” MTMM-CFA model – a two family perspective, three unique perspective model - and its unstandardized estimates are presented in Figure 3.

Figure 3. Superior two-shared factor, five-unique factor MTMM-CFA.

Figure 3

Open in a new tab

White ovals are shared perspectives; gray ovals are unique perspectives. A = adolescent; M = mother; F = father. PS = problem solving; CM = communication; AR = affective responsiveness; RL = roles; AI = affective involvement; BC = behavioral control. All estimates are unstandardized except for the associations between factors. All solid lines are significant at the .05 level or higher. Numbers listed in the square/observed variables are residual variances; all are significant at .05 level. Model fit: χ2(98) = 114.292, p = .12, CFI = .985, RMSEA = .036 (.000 | .061).

The study's sample size (family N = 128), which is large relative to other studies using the correlated uniqueness approach to examine the family's shared perspective (e.g., Cole & McPherson, 1993; Cook, 2001; Martin & Cole, 1993), should be sufficiently large to yield valid and reliable estimates given the complexity of the model in Figure 3 (91 estimated parameters). Simulations indicate that CFA models with a sample size to parameter (S/P) ratio as small as 1.25, which is smaller than the “superior” model's S/P ratio of 1.41, yield parameter estimates that are equivalent to models with S/P ratios as large as 20 (Jackson, 2001, 2003). Additional simulations indicate that CFA models with a ratio of indicators to (latent) factors (I/F) of 4.0, which is the I/F ratio of the “superior” model (i.e., 20 indicators to 5 latent factors), yield valid and reliable parameter estimates even when the sample size is as small as 100 (Gagné & Hancock, 2006). Finally, Marsh and Bailey (1991), whose simulation-study focused on MTMM-CFA models, found that, provided there are no indications of model misidentification, parameter estimates from simulations based on sample sizes similar to this study's sample size are equivalent to those found from simulations based on much larger sample sizes.

Description of two-dimensional family perspective

The family perspective from the superior MTMM-CFA had a two-dimensional factor structure: We termed one factor Family-Interaction because all of the sub-scales it loaded on (i.e., problem solving, communication, and affective responsiveness) pertain to perceptions of how effectively the family as a unit interacts. We termed the other factor Family-Structure because all the sub-scales it loaded on (i.e., roles, affective involvement, and behavioral control) pertain to structure and dependability within the family. Using adolescent reports only, Ridenour et al. (1999) found a similar factor structure when analyzing the FAD. As expected, though the family factors were empirically distinct, Δχ2(1) = 5.56, p < .05, they did covary with one another (r = .79, p < .001), indicating that families high (dysfunctional) on Family Interaction were likely to be high (dysfunctional) on Family Structure.

Construct validity of superior MTMM-CFA model

To test construct validity we included the across-family-member average of the general functioning sub-scale of the FAD (hereafter referred to as mean family dysfunction) and correlated it with all five perspective factors (i.e., the two family-perspective factors and the three unique-perspective factors). We fixed the family perspective and unique perspective-factor loadings to match those of the superior model to ensure that the factors we correlated with mean family dysfunction exactly matched those of the superior model. All family- and unique perspectives correlated positively with mean family dysfunction (not tabled), indicating the higher a family's mean level of dysfunction, the higher its family perspective (i.e., the more it is characterized by dysfunction) and the higher each family member's unique perspective (i.e., the more each is characterized by dysfunction). Based on chi-square difference tests, we examined whether the correlations differed in magnitude. As expected, the extent that family members agree that a family is dysfunctional is more closely tied to mean family dysfunction than is the extent that each family member uniquely views a family as dysfunctional. Compared to the relation between mean family dysfunction and the “Family Interaction” family perspective (r = .67, p < .001), the relation between mean family dysfunction and the adolescent unique perspective (r = .35, p < .01) was weaker, Δχ2(1) = 25.76, p < .001. The same was true for relations between mean family dysfunction and the mother unique perspective (r = .35, p < .01), Δχ2(1) = 14.31, p < .001, and the father unique perspective (r = .44, p < .01), Δχ2(1) = 8.27, p < .01. Likewise, compared to the relation between mean family dysfunction and the “Family Structure” family perspective (r = .59, p < .001), the relations between mean family dysfunction and the adolescent unique perspective, Δχ2(1) = 8.27, p < .01, the mother unique perspective, Δχ2(1) = 8.27, p < .01, and the father unique perspective, Δχ2(1) = 8.27, p < .01, were all weaker. The relation between unique perspective and mean family dysfunction did not vary by family member, Δχ2(2) = .68, p = .42.

Aim 1: Examining the Generalizability and Magnitude of the Unique Perspectives

Whether a unique-perspective factor emerged for each family member was used to determine whether a given family member's unique perspective generalizes across multiple domains of family functioning. After all, if an individual's unique perspective is generalized, then each family member's “non-shared” variances (i.e., variance remaining in FAD sub-scales after family perspective and dyad-specific perspective variance is accounted for) should all share common variance and should all load onto a single factor. To determine the magnitude of the unique perspectives, we compared the sizes of the standardized factor loadings of the family and unique perspectives.

As expected, after accounting for 3-way (family) and 2-way (dyad-specific) agreement among three family members (adolescent, mother, and father) across 6 subscales of family dysfunction, the remaining “non-shared” variance in each family member's six reports of family dysfunction factored together within family member. Each unique-perspective factor had clear convergent validity (i.e., aside from the adolescent loading for affective involvement, all loadings were significant). The percentages of variance explained by perspective (family and unique) and family member (adolescent, mother, and father) are presented in Table 2. For each family member, a substantial amount of the variance in reports of family dysfunction was captured by the unique perspective. For example, averaging across the six FAD items, the unique perspective captured 35.0% of the adolescent variance, 41.0% of the mother variance, and 38.8% of the father variance. For fathers, the unique perspective captured a greater proportion of variance than the family perspectives; this was true for each of the six FAD items as well as the overall average. For adolescents and mothers, which perspective captured more variance varied by the item, though when averaging across all the items the unique perspective captured more variance.

Table 2.

Percentage of variance in FAD subscales explained, by perspective and family member

Family Unique Error
Adolescent
 Problem solving 25.1% 45.8% 29.1%
 Communication 10.7% 64.8% 24.5%
 Affective responsiveness 25.6% 28.2% 46.2%
 Roles 41.7% 36.7% 21.6%
 Affective involvement 34.2% 12.2% 53.6%
 Behavioral control 15.3% 22.1% 62.7%
  Average 25.4% 35.0% 39.6%
Mother
 Problem solving 39.7% 33.9% 26.4%
 Communication 31.3% 38.4% 30.3%
 Affective responsiveness 32.6% 25.0% 42.5%
 Roles 7.7% 55.4% 37.0%
 Affective involvement 5.1% 60.1% 34.9%
 Behavioral control 18.2% 33.1% 48.7%
  Average 22.4% 41.0% 36.6%
Father
 Problem solving 17.9% 51.9% 30.2%
 Communication 13.4% 53.5% 33.2%
 Affective responsiveness 16.5% 30.3% 53.3%
 Roles 16.2% 24.7% 59.1%
 Affective involvement 15.0% 39.4% 45.6%
 Behavioral control 17.3% 33.4% 49.4%
  Average 16.0% 38.8% 45.1%
Open in a new tab

Finally, although the dyad-specific perspectives were not a focus of the study, out of the 18 across-reporter residual covariances, only 4 were significant (Figure 3). However, even when significant the across-reporter residual covariances were small in magnitude when compared to the loadings of the shared and unique factors. Taken together, these findings suggest that where there is across-family member agreement in family dysfunction, that agreement, by and large, is not idiosyncratic to a specific dyad but instead generalizes across all family members.

Aim 2: Relations Between Unique Perspectives and Individual Adjustment

To achieve our second aim, we included each indicator of individual adjustment as an observed variable and regressed it simultaneously on all family perspective and unique-perspective factors. Due to multicollinearity among the two family factors, we constrained their regression coefficients to be equal to one another. Doing so yielded more reliable estimates and did not impact model fit. Here we limit our focus to how the family perspective and a family member's unique perspective each uniquely relates to his or her own reports of adjustment (Table 3). To ensure that the family and unique factors exactly match those of the superior model, we again fixed the shared and unique factor loadings to match those of the superior model. We conducted two sets of analyses. The first used the observed measures of individual adjustment (Observed Model, Table 3); the second used residuals of individual adjustment measures after adjusting for mean family dysfunction (Residual Model, Table 3). We conducted this second set of analyses to remove amount of family dysfunction as a potential confound.

Table 3.

Family- and unique-perspective predicting self-reported individual adjustment

Observed Model
 Residual Model
Family-perspective1 Family member's own unique-perspective Family-perspective1 Family member's own unique-perspective
Adolescent
 Externalizing .07 .45*** −.07 .29**
 Internalizing −.04 .32*** −.03 .26**
 self-worth −.05 −.48*** .07 −.38**
Mother
 Depressive affect .14** .08 .09 .03
 Stress .15** .40*** −.02 .27**
 Satisfaction w/parenting −.08 −.32*** .02 −.22*
Father
 Depressive affect .08 .21* −.06 .15
 Stress .13** .62** −.01 .51***
Open in a new tab

Notes: All estimates standardized. Estimates from “Observed Model” based on observed measures of individual adjustment; Estimates from “Residual Model” based on residuals of individual adjustment after adjusting for mean family dysfunction.

1

For each indicator of adjustment, how it relates to Family Interaction and Family Structure was constrained to be equal because preliminary analyses indicated that how it related to Family Structure and Family Interaction did not differ.

*

p < .05,

**

p < .01,

***

p < .001

Focusing first on the analyses based on the observed measures of adjustment (Observed Model, Table 3), adolescent adjustment was negatively related to the adolescent unique perspective but unrelated to the family perspective. That is, adolescents with higher unique-perspective factor scores (i.e., relative to other adolescents their unique perspective is characterized by higher dysfunction) were also higher on externalizing and internalizing, and lower on self-worth. For mothers and fathers, their own adjustment was also negatively related to their own unique perspective. Mothers with higher unique-perspective factor scores were also higher on stress and lower on parental satisfaction, and fathers with higher unique-perspective factor scores were also higher on depressive affect and stress. Unlike adolescent adjustment, mother and father adjustment were also related to the family perspective: Families with higher family-perspective factor scores had mothers who were higher on depressive affect and stress, and had fathers who were higher on stress.

After holding mean family dysfunction constant across families, relations between unique perspectives and individual adjustment remained (Residual Model, Table 3). Thus, relations among family members' unique perspectives and individual adjustment are not spurious and simply due to the fact that the unique perspectives themselves are related to higher mean family dysfunction (as we already established when testing for construct validity). The same was not true for relations between family perspective and individual adjustment, which after adjusting for mean family dysfunction were reduced to non-significance. A likely explanation for this finding is that relative to the unique perspectives, the family perspective and mean family dysfunction are more closely associated and therefore empirically redundant. Finally, analyses controlling for parent social desirability did not alter the relations between the perspectives (family and unique) and adjustment of parents (results not tabled).

Discussion

Incorporating the perspectives of adolescent, mother, and father, this study examined the scope and magnitude of each family member's unique perspective of family functioning as well as how each family member's unique perspective related to his or her own adjustment. To do so we used a modified MTMM-CFA model that (a) isolated, for each family member's six reports of family dysfunction, the “non-shared” variance from variance shared by both other family members (i.e., family perspective) and variance shared by one other family member (i.e., dyad-specific perspective), and (b) extracted common variance across each family member's set of “non-shared” variances. Results indicated that each family member's unique perspective generalized across his or her different reports of family dysfunction and accounted for a sizable proportion of the family member's variance in reports of family dysfunction. Additionally, each family member's unique perspective was related to his or her own adjustment, and in most cases the size of the relation exceeded the relation between a family's family perspective and his or her own adjustment.

What are Family Members' Unique Perceptions of the Family?

To date family research has largely ignored family members' unique perspectives, probably due to either their presumed lack of utility or the fact that they are difficult to empirically isolate and, therefore, examine. We found that family members' unique perspectives are generalized and account for sizable proportions of variance in family members' overall perceptions of the family. But does this mean that they are useful? The answer depends on what the common variance across family member's “non-shared” variances actually reflects. One possibility is that family members' unique perspectives capture intra-familial disagreement. Such disagreements could reflect family member differences in (a) perceived amount of family dysfunction (e.g., how often family members yell at other family members), (b) perceived quality of family dysfunction (e.g., members perceive how often family members yell at other family members similarly, but disagree as to whether it is an appropriate amount), or (c) both. Disagreements in perceived quality could be the outgrowth of family member differences in expectations and cognitions regarding the family and familial relationships (Baldwin, 1992), which are particularly common among families with adolescent children (Laursen & Collins; 2009; Paikoff & Brooks-Gunn, 1991; Smetana, 1989). Another possibility is that family members' unique perspectives reflect family incohesion or the inverse of family cohesion, which is the extent to which family members are bonded together and committed to one another (Bollen & Hoyle, 1990; Moos, 1974).

If family members' unique perspectives of the family reflect intra-familial disagreement or signal family incohesion, then they capture useful and potentially clinically relevant information about the family and are a worthy topic of research. After all, although common during adolescence, adolescent-parent and mother-father disagreements are not necessarily inconsequential. Depending on their intensity, duration, and whether or not they are resolved, adolescent-parent disagreements can negatively impact adolescent and parent adjustment (Laursen & Collins, 2009; Patterson, Capaldi, & Bank, 1991) and parent-child relationships (Belsky, Jaffe, Hsieh, & Silva, 2001) over both the short- and long-term. Likewise, family incohesion can negatively impact both adolescent and parent adjustment (Farrell & Barnes, 1993). Beyond their impact on the adjustment and relationships of family members, both intra-familial disagreement and family incohesion can also reduce the effectiveness of therapeutic intervention. For example, feelings of cohesion are a key predictor of therapeutic success within family therapy as well as general group therapy (Burlingame, Fuhriman, & Johnson, 2001; Henggeler, Melton, & Smith, 1992). Additionally, a key goal of Functional Family Therapy (FFT; Sexton & Alexander, 2003), which views the family as the unit of intervention and explicitly recognizes the subjective nature of family members' perceptions of the family, is for family members to develop a “shared-family focus” regarding their family and its problems. Indeed, this “shared-family focus” is crucial to keeping family members motivated and engaged in the therapeutic process (Sexton & Alexander, 2003) and both intra-familial disagreement and family incohesion regarding family dysfunction could serve as roadblocks to its development.

Instead of capturing intra-familial disagreement or family incohesion, another possibility is that family members' unique perspectives capture reporter-specific measurement error - a source of variance that is not of substantive interest to family researchers. For example, relations among a given family member's “non-shared” variance (i.e., unique perspective) and his or her reports of adjustment could simply reflect a systematic tendency to offer replies that will be viewed favorably by others. However this explanation is unlikely because we found that relations between each parent's unique perspective and his or her own adjustment held after controlling for social desirability bias. Common variance across a given family member's “non-shared” variances could also reflect the influence of mundane daily hassles or triumphs (e.g., getting a speeding ticket or getting an “A” on a final) that, despite being unrelated to the construct measured, could still systematically influence participants' answers to questions inquiring about it on the day they provided data. However, this explanation also seems unlikely because family members' unique perspectives accounted for a higher proportion of variance in the reports of family dysfunction (at least one-third of each family member) than such daily hassles could realistically have accounted for on their own. Although we have reason to believe that family members' unique perspectives are not merely a reflection of reporter-specific measure error, future research utilizing the reports of “outside” raters could provide more conclusive proof.

Family Members' Unique Perspectives and Family Member Adjustment

Each family member's unique perspective was associated with his or her own adjustment, even when holding level of family dysfunction constant across families. Additionally, the relation between a family's family perspective and family member adjustment was comparatively weaker and in the case of adolescent adjustment non-significant, suggesting that among an adolescent's overall perspective of the family, it is his or her unique and non-shared perspective of the family that primarily relates to his or her adjustment. Therefore, provided the degree of family dysfunction is not too severe, it may be that the impact of family dysfunction on adolescent adjustment is lessened when family members are all in agreement regarding that dysfunction. The same does not appear to hold for mother and father; when not controlling for levels of family dysfunction, the family perspective was also related to mother and father adjustment. We also found that the family's across-family-member average of family dysfunction was positively related to each family member's unique perspective of family dysfunction, indicating that family member unique perspectives, in addition to being useful indicators of individual adjustment, may also be useful indicators of the family climate. Given the correlational nature of this study, whether the relations among family members' perspectives (both family and unique), individual adjustment, and family climate are causal is unclear, and if causal the direction of effect is unknown. Additional research utilizing longitudinal data and (quasi)experimental methods will help to determine causality and order of effects.

Limitations

This study has several important limitations in addition to its lack of “outsider” reports of individual adjustment and family dysfunction. First, no measure for adolescent social desirability bias was available, rendering the influence of social desirability bias on the unique perspectives of adolescents unclear. Second, because the study utilized a normative sample of mostly middle to upper-class European American families, how our findings generalize to clinical families or families of other ethnic and socioeconomic backgrounds is unclear. However, because the FAD is a sensitive measure of family functioning among both clinical and non-clinical samples (Miller et al., 1985), and displays measurement invariance across clinical and nonclinical samples (Kobakoff et al., 1990), there is reason to believe that our findings will generalize to clinical populations. Finally, though the study's sample size (family N = 128) was sufficient to yield reliable and unbiased estimates, it may not yield the power necessary to detect small effects, increasing the likelihood of Type II errors. Therefore, caution should be used when interpreting null findings.

Future Directions

In addition to incorporating “outsider” reports of family dysfunction, future research should examine whether unique perspectives of family dysfunction are symptomatic of clinical problems within the family. Because we found that family members' unique perspectives were positively related to overall family dysfunction, and existing research indicates that clinical families report higher overall family dysfunction (Miller et al., 1985), there is reason to believe that unique perspectives of family dysfunction are, in fact, elevated among clinical families. However, future research incorporating both clinical and non-clinical families is necessary to determine whether this is actually the case. Additionally, future research should examine predictors of individual differences in family members' unique perspectives. Among families with adolescent children, characteristics of the adolescent (gender, pubertal timing), the parents (gender, age, parenting style), and the family itself (family size, structure, SES, ethnicity) are all closely associated with perceptions of family conflict and disagreement (Laursen & Collins, 2009; Paikoff & Brooks-Gunn, 1991). Future research should explore whether these same factors are also associated with individual differences in family members' unique perspectives of family dysfunction. Finally, although this study focused on perspectives of the family, the analytical approach used for this study could be adapted to dyads. In doing so, it is possible to model the unique perspective of both dyad members as well as their shared dyad-perspective.

Acknowledgments

This research was supported by the Intramural Research Program of the NIH, NICHD.

References

  1. Aarons GA, McDonald EJ, Connelly CD, Newton RR. Assessment of family functioning in Caucasian and Hispanic Americans: Reliability, validity, and factor structure of the Family Assessment Device. Family Process. 2007;46(46):557–569. doi: 10.1111/j.1545-5300.2007.00232.x. doi: 10.1111/j.1545-5300.2007.00232.x. [DOI] [PubMed] [Google Scholar]
  2. Abidin RR. The Parenting Stress Index Professional Manual. Psychological Assessment Services; Odesa, FL: 1995. [Google Scholar]
  3. Achenbach TM, Rescorla LA. Manual for the ASEBA School-Age Forms & Profiles. University of Vermont, Research Center for Children, Youth, & Families; Burlington, VT: 2001. [Google Scholar]
  4. Amato PR, Cheadle JE. Parental divorce, marital conflict, and children's behavior problems: A comparison of adopted and biological children. Social Forces. 2000;86(3):1139–1161. doi: 10.1353/sof.0.0025. [Google Scholar]
  5. Baldwin MW. Relational schemas and the processing of social information. Psychological Bulletin. 1992;112(3):461–484. doi: 10.1037/0033-2909.112.3.461. [Google Scholar]
  6. Bartle-Haring S, Kenny D, Gavazzi S. Multiple perspectives on family differentiation: Analyses by multitrait multimethod matrix and triadic social relations models. Journal of Marriage and the Family. 1999;61(2):491–503. http://www.jstor.org/stable/353764. [Google Scholar]
  7. Belsky J, Jaffe S, Hsieh K, Silva P. Child-rearing antecedents of intergenerational relations in young adulthood: A prospective study. Developmental Psychology. 2001;37(6):801–814. doi: 10.1037//0012-1649.37.6.801. doi: 10.1037/0012-1649.37.6.801. [DOI] [PubMed] [Google Scholar]
  8. Bögels SM, van Melick M. The relationship between child-report, parent self-report, and partner report of perceived parental rearing behaviors and anxiety in children and parents. Personality and Individual Differences. 2004;37(8):1583–1596. doi: 10.1016/j.paid.2004.02.014. [Google Scholar]
  9. Bollen KA, Hoyle RH. Perceived cohesion: A conceptual and empirical examination. Social Forces. 1990;69(2):479–504. doi: 10.1093/sf/69.2.479. [Google Scholar]
  10. Bornstein MH. Parenting science and practice. In: Renninger KA, Sigel IE, Damon W, editors. Handbook of Child Psychology: Vol 4. Child psychology in practice. 6e Wiley; Hoboken, NJ: 2006. pp. 893–949. [Google Scholar]
  11. Burke L. The impact of maternal depression on familial relationships. International Review of Psychiatry. 2003;15(3):243–255. doi: 10.1080/0954026031000136866. doi: 10.1080/0954026031000136866. [DOI] [PubMed] [Google Scholar]
  12. Burlingame GM, Fuhriman A, Johnson JE. Cohesion in group psychotherapy. Psychotherapy. 2001;38(4):373–379. doi: 10.1037/0033-3204.38.4.373. [Google Scholar]
  13. Campbell DT, Fiske DW. Convergent and discriminant validation by the multitrait-multimethod matrix. Psychological Bulletin. 1959;56(2):81–105. doi: 10.1037/h0046016. [PubMed] [Google Scholar]
  14. Carlson CI, Cooper CR, Spradling VY. Developmental implications of shared versus distinct perceptions of the family in early adolescence. New Directions for Child and Adolescent Development. 1991;51:13–32. doi: 10.1002/cd.23219915103. [Google Scholar]
  15. Caster JB, Inderbitzen HM, Hope D. Relationship between youth and parent perceptions of family environment and social anxiety. Journal of Anxiety Disorders. 1999;13(3):237–251. doi: 10.1016/s0887-6185(99)00002-x. doi: 10.1016/S0887-6185(99)00002-X. [DOI] [PubMed] [Google Scholar]
  16. Cole DA, McPherson AE. Relation of family subsystems to adolescent depression: Implementing a new family assessment strategy. Journal of Family Psychology. 1993;7(1):119–133. doi: 10.1037/0893-3200.7.1.119. [Google Scholar]
  17. Cook WL. Interpersonal influence in family systems: A social relations model analysis. Child Development. 2001;72(4):1179–1197. doi: 10.1111/1467-8624.00341. doi: 10.1111/1467-8624.00341. [DOI] [PubMed] [Google Scholar]
  18. Cook WL, Goldstein MJ. Multiple perspectives on family relationships: A latent variables model. Child Development. 1993;64(5):1377–1388. doi: 10.1111/j.1467-8624.1993.tb02958.x. doi: 10.1111/j.1467-8624.1993.tb02958.x. [DOI] [PubMed] [Google Scholar]
  19. Cummings EM, Keller PS, Davies PT. Towards a family process model of maternal and paternal depressive symptoms: Exploring multiple relations with child and family functioning. Journal of Child Psychology and Psychiatry. 2005;46(5):479–489. doi: 10.1111/j.1469-7610.2004.00368.x. doi: 10.1111/j.1469-7610.2004.00368.x. [DOI] [PubMed] [Google Scholar]
  20. Deal JE. Utilizing data from multiple family members: A within-family approach. Journal of Marriage and Family. 1995;57(4):1109–1121. doi: 10.2307/353426. [Google Scholar]
  21. Deković M, Buist KL. Multiple perspectives within the family: Family relationship patterns. Journal of Family Issues. 2005;26(4):467–490. doi: 10.1177/0192513X04272617. [Google Scholar]
  22. Eisler I, Dare C, Szmukler GI. What's happened to family interaction research? An historical account and a family systems viewpoint. Journal of Marital and Family Therapy. 1988;14(1):45–65. doi: 10.1111/j.1752-0606.1988.tb01649.x. [Google Scholar]
  23. Epstein NB, Baldwin LM, Bishop DS. The McMaster Family Assessment Device. Journal of Marital and Family Therapy. 1983;9(2):171–180. doi: 10.1111/j.1752-0606.1983.tb01497.x. [Google Scholar]
  24. Farrell MP, Barnes GM. Family systems and social support: A test of the effects of cohesion and adaptability on the functioning of parents and adolescents. Journal of Marriage and the Family. 1993;55(1):119–132. doi: 10.2307/352963. [Google Scholar]
  25. Feldman SS, Wentzel KR, Gehring TM. A comparison of the views of mothers, fathers, and pre-adolescents about family cohesion and power. Journal of Family Psychology. 1989;3(1):39–60. doi: 10.1037/h0080523. [Google Scholar]
  26. Furman R, Jackson RL, Downey EP, Shears J. Social constructivist practice with youth. Child and Adolescent Social Work Journal. 2003;20(4):263–275. doi: 10.1023/A:1024504925017. [Google Scholar]
  27. Gagné P, Hancock GR. Measurement model quality, sample size, and solution propriety in confirmatory factor models. Multivariate Behavioral Research. 2006;41(1):65–83. doi: 10.1207/s15327906mbr4101_5. doi: 10.1207/s15327906mbr4101_5. [DOI] [PubMed] [Google Scholar]
  28. Harter S. The self-perception profile for adolescents. University of Denver; Denver, CO: 1988. Unpublished manual. [Google Scholar]
  29. Henggeler SW, Melton GB, Smith LA. Family preservation using multisystemic therapy: An effective alternative to incarcerating serious juvenile offenders. Journal of Consulting and Clinical Psychology. 1992;60(6):953–961. doi: 10.1037//0022-006x.60.6.953. doi: 10.1037/0022-006X.60.6.953. [DOI] [PubMed] [Google Scholar]
  30. Hofferth SL. Race/ethnic differences in father involvement in two-parent families: Culture, context, or economy? Journal of Family Issues. 2003;24(2):185–216. doi: 10.1177/0192513X02250087. [Google Scholar]
  31. Hollingshead AB. Four factor index of social status. Yale University; New Haven, CT: 1975. [Google Scholar]
  32. Jackson DL. Sample size and the number of parameter estimates in maximum likelihood confirmatory factor analysis: A Monte Carlo investigation. Structural Equation Modeling. 2001;8(2):205–223. doi: 10.1207/S15328007SEM0802_3. [Google Scholar]
  33. Jackson DL. Revisiting sample size and number of parameter estimates: Some support for the N:q hypothesis. Structural Equation Modeling. 2003;10(1):128–141. doi: 10.1207/S15328007SEM1001_6. [Google Scholar]
  34. Jager J. A developmental shift in Black-White differences in depressive affect across adolescence and early adulthood: The influence of early adult social roles and socio-economic status. International Journal of Behavioral Development. 2011;35(5):457–469. doi: 10.1177/0165025411417504. doi: 10.1177/0165025411417504. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kenny DA, Kashy DA. Analysis of the multitrait-multimethod matrix by confirmatory factor analysis. Psychological Bulletin. 1992;112(1):165–172. doi: 10.1037/0033-2909.112.1.165. [Google Scholar]
  36. Kobacoff RI, Miller IW, Bishop DS, Epstein NB, Keitner GI. A psychometric study of the McMaster Family Assessment Device in psychiatric, medical, and donclinical samples. Journal of Family Psychology. 1990;3(4):431–439. doi: 10.1037/h0080547. [Google Scholar]
  37. Laursen B, Collins WA. Parent-child relationships during adolescence. In: Lerner RM, Steinberg L, editors. Handbook of adolescent psychology, Vol. 2: Contextual influences on adolescent development. 3rd ed. Wiley; Hoboken, NJ, US: 2009. pp. 3–42. [Google Scholar]
  38. MacPhee D, Benson JB, Bullock D. Self-Perception of the Parental Role. Educational Testing Services; Princeton, NJ: 1985. [Google Scholar]
  39. Marsh HW. Confirmatory factor analyses of multitrait-multimethod data: Many problems and a few solutions. Applied Psychological Measurement. 1989;13(4):335–361. doi: 10.1177/014662168901300402. [Google Scholar]
  40. Marsh HW, Bailey M. Confirmatory factor analysis of multitrait-multimethod data: A comparison of alternative models. Applied Psychological Measurement. 1991;15(1):47–70. doi: 10.1177/014662169101500106. [Google Scholar]
  41. Marshall AJ, Harper-Jacques S. Depression and family relationships: Ideas for healing. Journal of Family Nursing. 2008;14(1):56–73. doi: 10.1177/1074840707312717. doi: 10.1177/1074840707312717. [DOI] [PubMed] [Google Scholar]
  42. Martin JM, Cole DA. Adaptability and cohesion of didactic relationships in families with developmentally disabled children. Journal of Family Psychology. 1993;7(2):186–196. doi: 10.1037/0893-3200.7.2.186. [Google Scholar]
  43. Miller IW, Epstein NB, Bishop DS, Keitner GI. The McMaster Family Assessment Device: Reliability and validity. Journal of Marital and Family Therapy. 1985;11(4):345–356. doi: 10.1111/j.1752-0606.1985.tb00028.x. [Google Scholar]
  44. Moos R. The Social Climate Scales: An overview. Consulting Psychologists Press; Palo Alto, CA: 1974. [Google Scholar]
  45. Muthén LK, Muthén BO. Mplus user's guide. Muthén & Muthén; Los Angeles: 1998–2009. [Google Scholar]
  46. Paikoff RL, Brooks-Gunn J. Do parent-child relationships change during puberty? Psychological Bulletin. 1991;110(1):47–66. doi: 10.1037/0033-2909.110.1.47. doi: 10.1037/0033-2909.110.1.47. [DOI] [PubMed] [Google Scholar]
  47. Park H, Bauer S. Parenting practices, ethnicity, socioeconomic status and academic achievement in adolescents. School Psychology International. 2002;23(1):386–395. doi: 10.1177/0143034302234002. [Google Scholar]
  48. Patterson GR, Capaldi D, Bank L. An early starter model for predicting delinquency. In: Pepler DJ, Rubin KH, editors. The development and treatment of childhood aggression. Lawrence Erlbaum Associates, Inc; Hillsdale, NJ: 1991. pp. 139–168. [Google Scholar]
  49. Radloff LS. The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement. 1977;1(3):385–401. doi: 10.1177/014662167700100306. [Google Scholar]
  50. Reynolds WM. Development of reliable and valid short forms of the Marlowe-Crowne Social Desirability Scale. Journal of Clinical Psychology. 1982;38(1):119–125. doi: 10.1002/1097-4679(198201)38:1<119‷AID-JCLP2270380118>3.0.CO;2-I. [Google Scholar]
  51. Ridenour TA, Daley J, Reich W. Factor analyses of the Family Assessment Device. Family Process. 1999;38(4):497–510. doi: 10.1111/j.1545-5300.1999.00497.x. doi: 10.1111/j.1545-5300.1999.00497.x. [DOI] [PubMed] [Google Scholar]
  52. Rindskopf D. Using phantom and imaginary latent variables to parameterize constraints in linear structural models. Psychometrika. 1984;49(1):37–47. doi: 10.1007/BF02294204. [Google Scholar]
  53. Rothbaum F, Weisz JR. Parental caregiving and child externalizing behavior in nonclinical samples: A meta-analysis. Psychological Bulletin. 1994;116(1):55–74. doi: 10.1037/0033-2909.116.1.55. doi: 10.1037/0033-2909.116.1.55. [DOI] [PubMed] [Google Scholar]
  54. Schwarz JC, Barton-Henry ML, Pruzinsky T. Assessing child-rearing behaviors: A comparison of ratings made by mother, father, child, and sibling on the CRPBI. Child Development. 1985;56(2):462–479. doi: 10.2307/1129734. [PubMed] [Google Scholar]
  55. Sexton T, Alexander J. Functional Family Therapy: A mature clinical model for working with at-risk adolescents and their families. In: Sexton T, Weeks G, Robbins M, editors. Handbook of Clinical Therapy. Brunner-Routledge; New York, NY: 2003. pp. 323–348. [Google Scholar]
  56. Smetana JG. Adolescents' and parents' reasoning about actual family conflict. Child Development. 1989;60(5):1052–1067. doi: 10.1111/j.1467-8624.1989.tb03536.x. http://psycnet.apa.org/psycinfo/1990-07263-001. [DOI] [PubMed] [Google Scholar]
  57. Tein J, Roosa MW, Michaels M. Agreement between parent and child reports on parental behaviors. Journal of Marriage and Family. 1994;56(2):341–355. doi: 10.2307/353104. [Google Scholar]
  58. Twenge J, Crocker J. Race and self-esteem: Meta-analysis comparing whites, blacks, Hispanics, Asians, and American Indians and comment on Gray-Little and Hafdahl (2000) Psychological Bulletin. 2002;128(3):371–408. doi: 10.1037/0033-2909.128.3.371. doi: 10.1037/0033-2909.128.3.371. [DOI] [PubMed] [Google Scholar]

RESOURCES