Abstract
Little attention has been paid to how early adolescents make attributions for their fathers’ behavior. Guided by symbolic interaction theory, we examined how adolescent gender, ethnicity, family structure, and depressive symptoms explained attributions for residential father behavior. 382 adolescents, grouped by ethnicity (European American, Mexican American) and family structure (intact, stepfamilies), reported attributions for their fathers’ positive and negative behaviors. Results indicated that for positive events girls made significantly more stable attributions, whereas boys made more unstable attributions. Mexican American adolescents tended to make more unstable attributions for positive events than European Americans, and adolescents from intact families made more stable attributions for positive events than adolescents from stepfamilies. Implications are discussed for the role of attributions in father-adolescent relationships as prime for intervention in families.
Keywords: Attributions, family relations, depressive symptoms, ethnicity
Although much of the early research on the parent-child relationship focused on mothers and their children, knowledge about the role of father involvement has rapidly advanced (Lamb, 2010). Fathers offer a unique and additive contribution to the prediction of adolescent adjustment and problem behaviors when data are assessed longitudinally or cross-sectionally from multiple reporters (Cookston & Finlay, 2006; Lamb, 2010). Whereas important work has delineated the relative contributions of quantity versus quality of father involvement (Amato, 1994), considerably less attention has been paid to children’s social constructions; namely their cognitive efforts to make sense of their relationships with their fathers (Parke et al., 2003). Although adolescents report that they feel less close, share fewer activities, and talk with their fathers less as compared to mothers (Cookston & Finlay, 2006), fathers are both directly and indirectly linked to the emergence of subsequent psychopathology among children (Flouri, 2005). Using symbolic interaction theory as a guide, we argue adolescents’ interpretations of fathering behavior must include family structure and contextual considerations (e.g., social settings, cultural features). This study seeks to address these gaps in the literature by focusing on adolescent attributions of fathering behaviors of positive and negative real-life events among Mexican American and European American intact and stepfamilies. We also link attributions to adolescents’ gender, depressive symptoms, family structure, ethnicity, and acculturation (as measured by familism).
Attributions
Attributions are the cognitive interpretations that explain the behaviors of others and, in turn, influence how behaviors are enacted. Adolescence offers a time of increased flexibility in understanding and predicting others’ mental states (Elkind, 1996). One important continuum along which attributions vary is the stability or instability of the attribution. Stable attributions explain behaviors as enduring aspects of a person’s disposition (e.g., she was mean today because she’s not a nice person) whereas unstable attributions are explanations for behavior caused by situations (e.g., she was mean because she had a bad day at work). Children’s stability attributions for the behaviors of others have been linked with both positive (Paleari, Regalia, & Fincham, 2003) and negative child behaviors (MacKinnon-Lewis, Castellino, Brody, & Fincham, 2001).
Family Relationships and Attributions
Attributions are related to family relationships bidirectionally because they both inform interpretations and are informed by experiences (Bugental, Johnston, New, & Silvester, 1998). Relationship quality and observed interactions between fathers and children have been related to more favorable child attributions of their parents’ behavior (Fincham, Beach, Arias, & Brody, 1998). Greater relationship positivity has been related to less hostile attribution patterns among children for both mother-child and father-child relationships, and better relationships promote attributions that lead to less conflict. Although Fincham and colleagues demonstrated that parental behaviors and child attributions are linked, their study relied on imagined situations and focused only on negative attributions. In this study, we consider positive and negative attributions for actual events provided by our adolescents.
Depressive Symptomology and Attributions
Adolescent attribution patterns vary with the degree of adolescent depression (Gladstone, Kaslow, Seeley, & Lewinsohn, 1997), though the direction of this association is unclear (Gladstone & Kaslow, 1995). Depressed children report that their family environments are more controlling and conflictual and are less supportive, cohesive, or communicative than nondepressed children (Kaslow, Deering, & Racusin, 1994). More conflictual family environments have been linked with higher levels of depressive symptoms (Sheeber, Hops, Alpert, Davis, & Andrews, 1997). Depressed children tend to give more negative explanations for events that happen than their less depressed peers (Gladstone & Kaslow, 1995). Depressed youth are more likely to think about the negative behaviors of others as internal to the actor, likely to occur again in similar situations (stable), and influential to behaviors in different situations (global). Depressed youth also tend to make external-unstable attributions for positive events (Nolen-Hoeksema, Girgus, & Seligman, 1992). As a result, explaining attributions demands attention to better understand depressive symptoms.
Theoretical Framework
For this analysis the theoretical framework of symbolic interaction theory was used to guide selection of control variables and analyses. Symbolic interaction theory states that social processes and experiences should include context (e.g., social settings, cultural features) as a factor that informs daily lives, interactions between adolescents and parents, and adolescent development. Thus, symbolic interaction theory helps to explain how explanations for father’s behavior emerge (Blumer, 1969). Specifically, an adolescent’s perspectives and expectations (e.g., being female, being part of a certain social class) inform her day-to-day interactions with her father, and she will interpret these moments based on her experiences in such exchanges. Furthermore, culturally dependent conceptualizations of expectations for parent behavior are important because, as symbolic interaction theorists suggest, they influence future actions and social interactions (Lal, 1995). For example, if a father does something nice for his daughter such as buy her an ice cream, she may make a stable attribution for this behavior (e.g., my father bought me an ice cream because he is a nice guy) because girls are more likely to make stable attributions for positive events. Previous research reviewed below provides some evidence of aspects and contexts of experience that will be associated with adolescent attributions.
Gender Differences in Depressive Symptomology and Attributions
The depression-attribution association varies by gender of child. Boys are more likely to attribute positive events to unstable causes and negative events to stable causes than girls (Gladstone et al., 1997). For both genders a more negative attributional style is significantly correlated with depressive symptoms, although more strongly for girls than for boys. However, the Gladstone et al. study’s focus on overall attributional style, which asks adolescents to make attributions about hypothetical situations, leaves unexplained the issue of whether similar attribution patterns would be found for adolescent attributions about their father’s actual (nonhypothetical) behavior. One goal of our study was to assess the associations among gender, depressive symptoms, and adolescent attributions for actual events with their fathers as identified by the adolescent. We focus on stable and unstable attributions because such explanations have been linked to prior family functioning (Gladstone et al., 1997; Nolen-Hoeksema et al., 1992) and expand on previous studies by focusing on attributions for specific events rather than overall attributional style. Symbolic interaction theory suggests that the experience of gender (being a girl or boy) will shape the decisions and interpretations of life experiences, including attributions of fathering behaviors and depressive symptomology (Blumer, 1969). In addition, we consider three factors that have the potential to be related to attributions for father’s behavior: the adolescent’s ethnicity, cultural experiences in the family (i.e., familism), and whether the father is a biological or stepparent.
Stepfathers and Attributions
In a rare study that examined attributions among stepfamilies, stepfathers reported their children as the cause for problems more often than biological parents; however, this trend was not statistically significant. Adolescent attributions for fathering behaviors may be influenced by the stepfathers’ role in the family; fathers may be less involved in daily family activities than mothers, and stepfathers may have more difficulty engaging their stepchildren than biological fathers (Coleman & Ganong, 1997). Giles-Sims and Crosbie-Burnett (1989) found that the stepfather’s influence in family decision-making was linked with his financial contribution. Experiences with fathers and interpretations of their behaviors may vary as a function of family structure and differences in role responsibilities between the two types of fathers. Symbolic interaction theory suggests that these role differences may influence adolescent interpretations and should be accounted for as part of the context of adolescent experience. As a result, we predicted that attributions for stepfather behavior would be less favorable (e.g., more stable for negative events, less stable for positive events) than attributions for biological father behavior.
Ethnicity and Attributions
Prior work did not examine the associations between ethnicity and attributions (Fincham et al., 1998; Gladstone et al., 1997), but research that investigated depression has provided helpful guidance. Mexican American adolescents are at higher risk for depressive symptoms than European American adolescents (Chapman & Perreira, 2005; Roberts & Sobhan, 1992). Mexican American youth reported believing that parental behavior and a nonsupportive home environment could cause depression (Fornos et al., 2005). Sociocultural factors experienced in Mexican American families (e.g., familism) combined with adolescent autonomy seeking and poor family function could elevate risk for suicide among Mexican American girls (Zayas, Lester, Cabassa, & Fortuna, 2005). As the authors note, the empirical evidence for the role fathers play has been understudied. However, as fathers are both directly and indirectly linked to the emergence of subsequent psychopathology among children (Flouri, 2005), understanding the association between attributions of fathering behaviors and mental health among Mexican American youth, a particularly vulnerable group compared to European American youth, is important for elucidating family processes and creating interventions to promote healthy development. Based on symbolic interaction theory, ethnicity should influence adolescents’ interpretations of fathering behaviors through contextual experiences that shape belief systems. However, no studies to date have explored attributions among Mexican American adolescents, thus, exploratory analyses were conducted. As Mexican American adolescents may be at elevated risk for mental health problems compared to European American adolescents, we will examine whether they report higher levels of depressive symptoms than European Americans and how associations between ethnicity and attributions may vary with their depressive symptoms.
Familism, a construct that assesses family support and interdependence, is an aspect of family life that has been repeatedly linked to Mexican American families (Santisteban, Muir-Malcolm, Mitrani, & Szapocznik, 2002), and recent evidence suggests the construct is similar among diverse ethnic groups, including European Americans (Schwartz, 2007). Increased levels of perceived family obligation have been associated with higher reports of cohesion with mother and father and better academic adjustment (Fuligni, Tseng, & Lam, 1999). Adolescents who have better relationships with their parents may, in turn, offer different explanations for their parents’ behavior than adolescents who have more conflictual relationships with their parents. Cultural family values such as familism have relevance to interactions within their families, thus, we include familism as a cultural values indicator and predictor of the association between attributions and depressive symptoms. Symbolic interaction theory suggests that cultural values such as orientation to the family may influence social constructions of family experiences and we control for familism to account for variations in attributions by this cultural context.
The Present Study and Hypotheses
The present study has two main purposes. First, we contribute new information to the area of attribution work by focusing on adolescent attributions for fathering behaviors rather than on adolescent attributions of mothers or family life. In contrast to previous studies, we examined attributions from real rather than hypothetical events in which adolescents interacted with their fathers or stepfathers. In addition, we asked adolescents about both positive and negative events whereas other studies tend to focus on negative attributions. Adolescents were asked about the causal reasons for why their fathers or stepfathers were nice or were mean to them. Second, we investigated the role of gender, ethnicity and family structure in moderating the associations among depressive symptoms, adolescent gender, and attributions.
Using symbolic interaction theory as a guide, several aspects of adolescent experience that are related to interpretations of adolescents were selected for analysis. Gender and ethnicity were selected as personal characteristics of adolescents that shape their experiences. Family context was also examined with family socioeconomic status, family structure (biological or stepfather families), and culture (i.e., familism) as factors influencing adolescents’ interpretations of fathering behaviors. Previous research by Gladstone et al. (1997) suggests two hypotheses for gender: (a) Girls were hypothesized to endorse more stable attributions for positive events and boys were hypothesized to endorse more unstable attributions for positive events, and (b) it was hypothesized that adolescents who reported greater depressive symptoms would make more stable attributions for negative events and unstable attributions for positive events than nondepressed adolescents. Although it is unclear if and how ethnicity will impact attributions, Mexican American adolescents are at higher risk for mental health problems compared to European American adolescents. We hypothesized that Mexican Americans will endorse higher levels of unstable attributions for positive events and exploratory analyses will investigate how their attributions are related to depressive symptoms as compared to European Americans. Stepfathers have been suggested to have a more difficult time engaging with adolescents than biological fathers (Coleman & Ganong, 1997), thus, adolescents in stepfamilies are hypothesized to endorse more unstable positive and more stable negative attributions for stepfather behavior as compared to adolescents from intact families. To account for variance explained by the nature of an adolescent’s relationship with her or his father, relationship quality is included as a control. Greater relationship positivity (i.e., relationship quality and observed behavior) is associated with children offering most positive attributions for their fathers’ behaviors (Fincham et al., 1998). Father-adolescent relationship quality as reported by the father was included along with family income and familism to account for their unique effects.
Method
Participants
All participants were drawn from the Parent and Youth Study (see NIH grant number 5R01HD056615) that investigated adolescent adjustment and fathering (see Parke et al., 2003 for a description). Participants were 392 families in which mothers, fathers, and adolescents were interviewed. The sample included 183 girls and 199 boys in 7th grade who ranged in age from 11 to 15 years old (M = 12.40, SD = 0.54 and M = 12.52, SD = 0.59 for girls and boys, respectively). The families were of either Mexican American (N = 193) or European American (N = 199) ancestry and were from either intact two-parent biological families (N = 217) or from stepfamilies (N = 175). The fathers in the sample were slightly older than the mothers (M = 40.24, SD = 7.39 and M = 38.44, SD = 5.97 for fathers and mothers, respectively) and the stepfathers had been in the home for an average of 5.43 years (SD = 2.97). The socioeconomic status of the family, as indexed by total family income, was calculated by adding the mother’s report of her earnings, the father’s report of his earnings, and the father’s report of public assistance (e.g., TANF, food stamps), child support, or other income (e.g., rental income, family). Household incomes ranged from $8,000 to $467,500 with a mean of $67,410 (SD = 47,194; for EA, M = 86,678, SD = 54,392; for MA, M = 47,514, SD = 26,588; for stepfamilies; M = 68,362, SD = 47,490; and for intact families; M = 66,705, SD = 47,151). Although there were differences between the EA and the MA families in terms of family income, when we compared our participants from the Census tracts from which they were sampled, no differences existed, thus, our participants appear to be representative of the communities from which they were sampled.
Procedure
Families from two large urban areas in different states in the southwest United States were sampled through a brief postcard survey that identified 7th grader ethnicity and house composition. A letter and brochure about the project was mailed to families one week prior to phone contact. Once interest to participate was assessed, a team of researchers was sent to the home to interview the mother, father or stepfather, and adolescent. Both parents and the adolescent in each family were interviewed separately either in the family home or in a community facility. Approval for the current study was granted by the Institutional Review Boards at Arizona State University, San Francisco State University, and the University of California at Riverside.
Measures
Attributions
In a pilot version of our study, 20 adolescents completed a version of the Kids Relationship Attribution Measure (Fincham et al., 1998) that was modified in two ways. First, we asked participants to recall actual situations involving their fathers rather than general/hypothetical events to create greater involvement. Second, we asked them to describe two situations - one that involved positive parenting behaviors and one negative. After describing each situation, adolescents answered a series of questions regarding the reasons for the fathers’ behavior. Three items specifically targeted attributions of locus, stability, and globality of the behavior. In the pilot, these modifications were easily understood and appeared to lead to meaningful responses. In the final sample, adolescents answered 24 fixed response questions about specific positive and negative events involving their fathers or stepfathers. The items and coding manual for the Parents and Youth Study Attribution Scale are available on our website: http://bss.sfsu.edu/devpsych/pays/. The interviewer asked the adolescent to think about four events that occurred in the past few years: one in which the father said something nice, one in which the father did something nice, one in which the father said something upsetting, and one in which the father did something upsetting. Questions were skipped if the adolescent could not identify an event (three adolescents could not identify positive events and 28 adolescents could not identify negative events). After identifying each event, adolescents were then asked to respond to six causal explanations for their fathers’ behaviors. With event type, the positive events (i.e., He said something nice, He did something nice) and the negative events (said something mean and did something mean) were averaged within valence of event as the responses tended to be consistent (correlations ranged from .52 to .84). For positive behaviors, there were three stable attributions (i.e., He’s a positive person, He likes to make you happy, He cares about you) and three unstable (i.e., You really deserved it, He happened to be in a good mood, Someone else told him to or wanted him to). For negative behaviors, there were three stable attributions (i.e., He’s a mean or difficult person, He’s ALWAYS down on you, He doesn’t care if something he says bothers or hurts you) and three unstable attributions (i.e., You really deserved it, He happened to be in a bad mood, It was just one of those times that he really got upset). The attributions responses were on a 5-point Likert scale coded as 1 = not at all to 5 = exactly. Within each category (positive events, negative events), ratings were averaged across the three stable items and across the three unstable items. Four attributions for fathering behaviors were identified: (a) stable attributions for positive events (α = .85 for all participants; α = .81 for EA; α = .89 for MA), (b) unstable attributions for positive events (α = .57 for all participants; α = .51 for EA; α = .58 for MA), (c) stable attributions for negative events (α = .90 for all participants; α = .89 for EA; α = .89 for MA), and (d) unstable attributions for negative events (α = .59 for all participants; α = .53 for EA; α = .64 for MA). Although the alpha estimates for the two unstable attributions are low, the association is misleading. Because we measured attributions for a specific event, the adolescent’s attributions for the event informed the latent attribution variables and, thus, were conceptualized as causal rather than effect indicators (Bollen & Lennox, 1991). Causal indicators like life events and our attribution measures tend to have low degrees of inter-item consistency because each item contributes to measurement of the causal hypothetical construct.
Overall relationship quality
Two fixed response questions assessed the overall relationship quality of adolescents and fathers/stepfathers as reported by fathers. The first question was “How well do you get along with your child?”; response options were coded as follows: (1) not well at all, (2) not too well, (3) just okay, (4) pretty well, and (5) extremely well. The second question was “What kind of relationship do you have with your child?”; response options were (1) the worst, (2) very bad, (3) not too good, (4) just okay, (5) good, (6) very good, and (7) the best. A mean score of the two items was computed to create father’s reports of adolescent-father relationship quality with higher scores reflecting a better overall relationship (α = .80; Mexican American fathers, α = .75; European American fathers, α = .84).
Depressive symptoms
Adolescents answered 8 fixed response questions from a shortened version of the Child Depression Inventory (CDI; Kovacs, 1985), which correlated .87 with the total CDI. Item responses ranged from 1 to 3 and a higher value indicated a higher level of depressive symptoms. An average score was created across the items and reliability was .66 (European Americans: girls, α = .77; boys, α = .67; Mexican Americans: girls, α = .68; boys, α = .34).
Familism
The familism scale was created from 15 fixed response items from the Mexican American Cultural Values Scale (Knight, Gonzales, Saenz, Roosa, & Updegraff, 2007), which has been validated in two independent large samples of Mexican and Mexican American youth living in a metropolitan area of the southwestern United States (Knight et al., 2010). Both Mexican American and European American adolescents were asked 50 acculturation questions, but analyses were limited to three subscales; family support, family obligations, and family members as referents due to differences found between individuals born in Mexico compared to individuals born in the United States. Sample items for family support include “Parents should teach their children that the family always comes first” and “It is important for family members to show their love and affection to one another.” Family obligations sample items include “Parents should be willing to make great sacrifices to make sure their children have a better life” and “Older kids should take care of and be role models for their younger brothers and sisters.” Sample items for family members as referents are: “When it comes to important decisions, the family should ask for advice from close relatives” and “Children should always do things to make their parents happy.” Across the three familism subscales, the alpha for European Americans was .85 and for Mexican Americans was .89. There was no significant difference between Mexican American and European American adolescents on reports of familism. Familism has been shown to maintain its operational definition even when used among diverse ethnic groups including European Americans (Schwartz, 2007).
Plan of Analysis
To explore differences among participants as a function of gender, ethnicity, and family structure on the four attribution behaviors, a Multivariate Analysis of Covariance was estimated with income as a covariate. Descriptive analyses examined attributions and depressive symptoms by gender, ethnicity and family structure. Independent samples t-tests were used to compare adolescents by gender, ethnicity, and family structure on the four attributions, depressive symptoms, and familism. Next, correlation analyses examined associations among attributions, depressive symptoms, ethnicity, family structure, income, gender, familism, and adolescent-father relationship quality.
To test whether adolescents who reported more depressive symptoms tended to make more stable attributions for negative events and unstable attributions for positive events than nondepressed adolescents, four separate hierarchical regression analyses were conducted for: (a) stable attributions for positive events, (b) unstable attributions for positive events, (c) stable attributions for negative events, and (d) unstable attributions for negative events. Family structure, income, ethnicity, depressive symptoms, gender, familism, and relationship quality were entered in the first block as control variables; the two-way interactions for depressive symptoms by gender, depressive symptoms by ethnicity, and ethnicity by gender were entered at block 2; and a three-way ethnicity by depressive symptoms by gender interaction was entered at block 3. Centered scores were utilized for continuous predictors and in creating the interaction terms.
Results
Descriptive Statistics
Multivariate Analysis of Covariance indicated significant differences for gender (Pillai’s Trace = .04, F = 3.64, df = 4, p < .01), family structure (Pillai’s Trace = .03, F = 2.59, df = 4, p < .05), and ethnicity (Pillai’s Trace = .07, F = 6.53, df = 4, p < .001). The means, standard deviations, and t-tests for attributions and depressive symptoms are provided in Table 1. Girls made more stable attributions for positive events (M = 4.36) than boys (M = 4.20), t(365) = −2.23, p < .05. Boys did not, however, make more unstable attributions for positive events (M = 2.77) than girls (M = 2.75), t(365) = 0.31, ns. For negative events, boys made more stable attributions (M = 1.69) than girls (M = 1.46), t(365) = 3.11, p < .01, and more unstable attributions (M = 2.61 versus M = 2.40, respectively), t (365) = 2.70, p <.01. Excepting the absence of a gender difference for unstable attributions for positive events, these results appear to offer conditional support for the conclusion that the attribution styles of boys and girls differ. Boys are less likely to make stable attributions for positive events and more likely to make stable attributions for negative events. They are also more likely than girls to make unstable attributions for negative events. There was no significant difference between girls and boys on depressive symptoms, t(391) = 1.06, ns. Adolescents from stepfamilies reported significantly more depressive symptoms than children from biological intact families, t(388) = −3.02, p < .01, and adolescents from intact homes were more likely to make stable attributions for fathers’ positive behavior than were the adolescents in stepfather homes, t(388) = 2.30, p < .05. Mexican American adolescents reported more depressive symptoms than European Americans, t(383) = −2.17, p < .05, and Mexican American adolescents were much more likely to endorse unstable attributions for positive events, t(382) = −5.20, p < .001, and stable attributions for negative events, t(357) = −3.36, p < .01.
Table 1.
Descriptive Statistics and T-tests of Adolescent Attributions, Depressive Symptoms, and Familism
| Adolescents | Boys | Girls | European American |
Mexican American |
Intact Families |
Stepfamilies | ||||
|---|---|---|---|---|---|---|---|---|---|---|
|
|
||||||||||
| M/SD | M/SD | M/SD | t | M/SD | M/SD | t | M/SD | M/SD | t | |
| Stable attributions for positive events |
4.28/0.69 | 4.20/0.72 | 4.36/0.66 | −2.23* | 4.28/0.63 | 4.29/0.76 | −0.13 | 4.36/0.60 | 4.19/0.78 | 2.30* |
| Unstable attributions for positive events |
2.76/0.60 | 2.77/0.61 | 2.75/0.60 | 0.31 | 2.61/0.56 | 2.91/0.61 | −5.20*** | 2.81/0.61 | 2.69/0.59 | 1.92 |
| Stable attributions for negative events |
1.57/0.72 | 1.69/0.80 | 1.46/0.62 | 3.11** | 1.45/0.63 | 1.70/0.78 | −3.36** | 1.52/0.64 | 1.64/0.81 | −1.46 |
| Unstable attributions for negative events |
2.50/0.76 | 2.61/0.73 | 2.40/0.78 | 2.70** | 2.48/0.73 | 2.53/0.79 | −0.65 | 2.49/0.75 | 2.51/0.77 | −0.25 |
| Depressive symptoms | 1.42/0.32 | 1.44/0.29 | 1.40/0.35 | 1.06 | 1.38/0.34 | 1.45/0.30 | −2.17* | 1.37/0.29 | 1.47/0.35 | −3.02** |
| Familism | 4.42/0.49 | 4.41/0.51 | 4.42/0.47 | −0.29 | 4.40/0.45 | 4.43/0.53 | −0.67 | 4.47/0.42 | 4.35/0.56 | 2.18* |
p < .05,
p < .001.
Correlation Analyses
Correlation analyses indicated that, among the attribution variables, five of the six were significantly related to each other, p < .01, and between attributions and depressive symptoms three of the four associations were significant (Table 2). Both being Mexican American and having a lower income were correlated with unstable attributions for positive events, stable attributions for negative events, and depressive symptoms. Higher rates of familism correlated with stable and unstable attributions for positive events whereas lower rates of familism correlated with stable attributions for negative events and depressive symptoms. Stable attributions for positive events correlated with being from an intact family and being female whereas stable and unstable attributions for negative events were correlated with being male. Parent-adolescent relationship quality negatively correlated with depressive symptoms and presence of a stepfather in the home. Despite many significant correlations, two variables – familism and unstable attributions for negative events – showed few significant associations across the study variables.
Table 2.
Correlations of Positive and Negative Attributions and Depressive Symptoms
| Constructs | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 |
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1. Stable attributions for positive events |
- | .23*** | −.46*** | .01 | −.24*** | .01 | −.12* | −.02 | .11* | .36*** | .30*** |
| 2. Unstable attributions for positive events |
- | .20*** | .24*** | .08 | .26*** | −.10 | −.20*** | −.02 | .13* | −.02 | |
| 3. Stable attributions for negative events |
- | .31*** | .33*** | .18** | .08 | −.14* | −.16** | −.19*** | .24** | ||
| 4. Unstable attributions for negative events |
- | .23*** | .03 | .01 | −.02 | −.14** | .03 | −.07 | |||
| 5. Depressive symptoms | - | .11* | .16** | −.12* | −.05 | −.12* | −.22** | ||||
| 6. Ethnicitya | - | .00 | −.42*** | .01 | .03 | −.03 | |||||
| 7. Presence of stepfather in the homeb | - | .02 | .01 | −.11* | −.24** | ||||||
| 8. Income | - | .00 | −.03 | .11* | |||||||
| 9. Genderc | - | .02 | .02 | ||||||||
| 10. Familism | - | .07 | |||||||||
| 11. Adolescent-father relationship quality | - |
European American = 0, Mexican American = 1.
Intact family = 0, Stepfamily = 1.
Boys = 0, Girls = 1.
p < .05,
p < .01,
p < .001.
Predicting Stable and Unstable Attributions for Positive Events
Regression models for stable and unstable attributions for positive events were significant (Table 3). For stable attributions, significant main effects were found such that the adolescents with higher levels of familism tended to hold more stable attributions for positive events, β = 0.31, t = 6.74, p < .001, and higher levels of relationship quality was positively associated with stable attributions, β = 0.26, t = 5.40, p < .001. No significant interaction effects were observed.
Table 3.
Hierarchical Regression Analysis Predicting Stable and Unstable Attributions for Positive Events
| Stable attributions for positive events |
Unstable attributions for positive events |
|||||||
|---|---|---|---|---|---|---|---|---|
| Variable | B | SE B | β | t | B | SE B | β | t |
| Step R2 | .23*** | .10*** | ||||||
| Family structure | −0.01 | 0.06 | −0.01 | −0.11 | −0.12 | 0.06 | −0.10 | −1.88 |
| Income | 0.00 | 0.00 | −0.06 | −1.15 | 0.00 | 0.00 | −0.10 | −1.80 |
| Ethnicity | 0.00 | 0.07 | 0.00 | 0.04 | 0.24 | 0.07 | 0.20 | 3.76*** |
| DSa | −0.32 | 0.10 | −0.15 | −3.23** | 0.12 | 0.10 | 0.07 | 1.29 |
| Gender | 0.12 | 0.06 | 0.09 | 2.02* | −0.02 | 0.06 | −0.01 | −0.26 |
| Familism | 0.43 | 0.06 | 0.31 | 6.76*** | 0.15 | 0.06 | 0.12 | 2.45* |
| Relationship quality | 0.10 | 0.02 | 0.25 | 5.33*** | −0.01 | 0.02 | −0.02 | −0.37 |
| Step R2 | .25 | .14** | ||||||
| Family structure | −0.02 | 0.06 | −0.02 | −0.32 | −0.14 | 0.06 | −0.11 | −2.23* |
| Income | 0.00 | 0.00 | −0.04 | −0.84 | 0.00 | 0.00 | −0.07 | −1.29 |
| Ethnicity | 0.08 | 0.20 | 0.06 | 0.39 | 0.17 | 0.19 | 0.14 | 0.91 |
| DS | 0.75 | 0.44 | 0.35 | 1.72 | 1.73 | 0.41 | 0.92 | 4.24*** |
| Gender | 0.19 | 0.20 | 0.14 | 0.98 | −0.09 | 0.18 | −0.08 | −0.51 |
| Familism | 0.43 | 0.06 | 0.31 | 6.72*** | 0.14 | 0.06 | 0.11 | 2.33* |
| Relationship quality | 0.10 | 0.02 | 0.25 | 5.38*** | −0.01 | 0.02 | −0.02 | −0.31 |
| DS × Gender | −0.38 | 0.20 | −0.30 | −1.89 | −0.63 | 0.19 | −0.57 | −3.36** |
| DS × Ethnicity | −0.32 | 0.20 | −0.23 | −1.64 | −0.41 | 0.18 | −0.33 | −2.22* |
| Ethnicity × Gender | −0.05 | 0.13 | −0.08 | −0.38 | 0.05 | 0.12 | 0.09 | 0.43 |
| Step R2 | .25 | .14 | ||||||
| Family structure | −0.02 | 0.06 | −0.01 | −0.30 | −0.14 | 0.06 | −0.11 | −2.27* |
| Income | 0.00 | 0.00 | −0.04 | −0.86 | 0.00 | 0.00 | −0.07 | −1.25 |
| Ethnicity | 0.07 | 0.20 | 0.05 | 0.35 | 0.19 | 0.19 | 0.15 | 0.10 |
| DS | 0.20 | 1.02 | 0.09 | 0.20 | 2.82 | 0.96 | 1.51 | 2.94** |
| Gender | 0.19 | 0.20 | 0.14 | 0.98 | −0.09 | 0.18 | −0.08 | −0.51 |
| Familism | 0.43 | 0.06 | 0.31 | 6.74*** | 0.14 | 0.06 | 0.11 | 2.24* |
| Relationship quality | 0.10 | 0.02 | 0.26 | 5.40*** | −0.01 | 0.02 | −0.02 | −0.36 |
| DS × Gender | −0.04 | 0.60 | −0.03 | −0.07 | −1.30 | 0.57 | −1.17 | −2.30* |
| DS × Ethnicity | 0.07 | 0.69 | 0.05 | 0.10 | −1.19 | 0.65 | −0.95 | −1.84 |
| Ethnicity × Gender | −0.04 | 0.13 | −0.07 | −0.35 | 0.04 | 0.12 | 0.08 | 0.38 |
| Ethnicity × DS | ||||||||
| X Gender | −0.24 | 0.40 | −0.28 | −0.59 | 0.48 | 0.38 | 0.65 | 1.26 |
Depressive symptoms.
p < .05,
p < .01,
p < .001.
Adolescents from stepfather families were less likely to make unstable attributions for positive events when compared with adolescents from biological father families (Table 3). Adolescents with either higher levels of depressive symptoms, β = 1.51, t = 2.94, p < .01, or higher levels of familism, β = 0.11, t = 2.24, p < .05, tended to report higher levels of unstable attributions for positive events. A significant interaction effect was observed for the depressive symptoms by gender interaction, β = −1.17, t = −2.30, p < .05 (Figure 1). This figure indicates that among boys with high levels of depressive symptoms unstable attributions for positive events are highest and unstable attributions are low at lower levels of depressive symptoms. Among girls, levels of unstable attributions for positive events are low among all levels of reported depressive symptoms.
Figure 1.
The association between depressive symptoms and gender in predicting unstable attributions for positive events of fathering behaviors among adolescents.
Predicting Stable and Unstable Attributions for Negative Events
Regression models for stable and unstable attributions for negative events were significant (Table 4). Stable attributions of negative events were more common among adolescents with more depressive symptoms or with lower levels of familism or relationship quality. There were two interaction effects for stable attributions for negative events: A two-way depressive symptoms by gender interaction, β = −1.21, t = −2.42, p < .05, and a three-way ethnicity by depressive symptoms by gender interaction, β = 1.17, t = 2.30, p < .05. Because our three-way interaction included variables from the two-way interaction, we focus here on just the three-way interaction (Figure 2). Although Mexican American and European American adolescents report similar patterns of stable attributions for negative events, Mexican American girls and boys report higher mean levels of attributions than European American girls and boys with Mexican American girls exhibiting greater difference between those with low and those with high depressive symptoms. For all four groups, adolescents with low levels of depressive symptoms tended to hold low levels of stable attributions for negative events and at higher levels of depressive symptoms a stronger endorsement of stable attributions is reported.
Table 4.
Hierarchical Regression Analysis Predicting Stable and Unstable Attributions for Negative Events
| Stable attributions for negative events |
Unstable attributions for negative events |
|||||||
|---|---|---|---|---|---|---|---|---|
| Variable | B | SE B | β | t | B | SE B | β | t |
| Step R2 | .20*** | .07*** | ||||||
| Family structure | −0.02 | 0.07 | −0.02 | −0.34 | −0.03 | 0.08 | −0.02 | −0.38 |
| Income | 0.00 | 0.00 | −0.03 | −0.62 | 0.00 | 0.00 | 0.02 | 0.42 |
| Ethnicity | 0.19 | 0.08 | 0.13 | 2.58* | 0.03 | 0.09 | 0.02 | 0.29 |
| DSa | 0.56 | 0.11 | 0.25 | 4.99*** | 0.54 | 0.13 | 0.23 | 4.27*** |
| Gender | −0.21 | 0.07 | −0.15 | −3.10** | −0.20 | 0.08 | −0.13 | −2.57* |
| Familism | −0.22 | 0.07 | −0.15 | −3.09** | 0.08 | 0.08 | 0.05 | 0.98 |
| Relationship quality | −0.07 | 0.02 | −0.17 | −3.35** | −0.01 | 0.02 | −0.02 | −0.40 |
| Step R2 | .21 | .08 | ||||||
| Family structure | −0.02 | 0.07 | −0.01 | −0.28 | −0.03 | 0.08 | −0.02 | −0.38 |
| Income | 0.00 | 0.00 | −0.04 | −0.77 | 0.00 | 0.00 | 0.02 | 0.36 |
| Ethnicity | 0.26 | 0.22 | 0.18 | 1.20 | 0.12 | 0.25 | 0.08 | 0.47 |
| DS | 0.26 | 0.48 | 0.12 | 0.55 | 0.59 | 0.55 | 0.25 | 1.07 |
| Gender | −0.12 | 0.22 | −0.08 | −0.56 | −0.10 | 0.25 | −0.06 | −0.39 |
| Familism | −0.22 | 0.07 | −0.15 | −3.08** | 0.08 | 0.08 | 0.05 | 0.99 |
| Relationship quality | −0.07 | 0.02 | −0.17 | −3.45** | −0.01 | 0.02 | −0.03 | −0.47 |
| DS × Gender | −0.17 | 0.22 | −0.13 | −0.77 | −0.23 | 0.025 | −0.16 | −0.90 |
| DS × Ethnicity | 0.39 | 0.22 | 0.27 | 1.80 | 0.22 | 0.25 | 0.14 | 0.90 |
| Ethnicity × Gender | −0.06 | 0.14 | −0.09 | −0.41 | −0.07 | 0.16 | −0.10 | −0.43 |
| Step R2 | .22* | .09* | ||||||
| Family structure | −0.03 | 0.07 | −0.02 | −0.38 | −0.04 | 0.08 | −0.03 | −0.47 |
| Income | 0.00 | 0.00 | −0.04 | −0.70 | 0.00 | 0.00 | 0.03 | 0.43 |
| Ethnicity | 0.31 | 0.22 | −0.04 | −0.70 | 0.16 | 0.25 | 0.11 | 0.65 |
| DS | 2.60 | 1.12 | 1.16 | 2.31* | 3.09 | 1.28 | 1.30 | 2.41* |
| Gender | −0.11 | 0.22 | −0.08 | −0.52 | −0.09 | 0.25 | −0.06 | −0.35 |
| Familism | −0.23 | 0.07 | −0.16 | −3.24** | 0.07 | 0.08 | 0.04 | 0.85 |
| Relationship quality | −0.07 | 0.02 | −0.18 | −3.58*** | −0.01 | 0.02 | −0.03 | −0.58 |
| DS × Gender | −1.62 | 0.67 | −1.21 | −2.42* | −1.78 | 0.86 | −0.99 | −1.81 |
| DS × Ethnicity | −1.27 | 0.76 | −0.85 | −1.68 | −1.56 | 0.86 | −0.99 | −1.81 |
| Ethnicity × Gender | −0.08 | 0.14 | −0.12 | −0.56 | −0.09 | 0.16 | −0.13 | −0.56 |
| Ethnicity × DS | ||||||||
| X Gender | 1.02 | 0.45 | 1.17 | 2.30* | 1.10 | 0.51 | 1.18 | 2.16* |
Depressive symptoms.
p < .05,
p < .01,
p < .001.
Figure 2.
The association between depressive symptoms, ethnicity, and gender in predicting stable attributions for fathering behaviors during negative events among adolescents.
Adolescents with higher levels of depressive symptoms tended to more strongly endorse unstable attributions for negative events (Table 4). There was one significant three-way interaction effect for ethnicity by depressive symptoms by gender, β = 1.18, t = 2.16, p < .05 (Figure 3). The figure indicates that depressive symptoms and negative unstable events are most strongly related in European American boys and Mexican American girls with Mexican American girls exhibiting great difference between those with low and those with high depressive symptoms, although the patterns are similar for all groups.
Figure 3.
The association between depressive symptoms, ethnicity, and gender in predicting unstable attributions for fathering behaviors during negative events among adolescents.
Discussion
The importance of fathers for adolescent adjustment and mental health is increasingly recognized, however, we know very little about how adolescents socially construct and make sense of their relationships with their parents. The current study replicated and extended earlier work on attributions to demonstrate how explanations for father behavior are related to features of adolescents and families. Informed by tenets of symbolic interaction theory, social processes occur within the context of family life and unique elements within families should account for different explanations. Adolescent attributions of fathering behaviors were associated with depressive symptoms, gender, and ethnicity. Adolescents with higher levels of depressive symptoms endorsed more unstable attributions for fathers’ positive events, viewing their fathers’ positive behaviors as less stable than nondepressed youth. As adolescents begin seeking more autonomy and depressive symptoms begin to emerge more saliently during this age, family relationships may take on new meaning and interpretations of behaviors may alter future interaction patterns.
In our study we failed to replicate the oft-reported gender difference for depressive symptoms, however, our sample was an early-adolescent one and may have been sampled prior to the emergence of differences. Gender differences in attributions (Gladstone et al., 1997) were observed. Girls more strongly endorsed stable positive fathering behaviors and boys more strongly endorsed stable negative and unstable negative fathering behaviors. In other words, girls were more likely than boys to attribute the father’s behavior during the good times to stable causes and to explain his behavior during negative events as less stable. Although we can only speculate on the reasons for these seemingly systemic differences between girls’ and boys’ attributional styles, girls may interpret or may be socialized to interpret other peoples’ behavior more positively than boys. In a sample of adolescents from Spain, girls also tended to be more optimistic than boys and this may have accounted for their more positive attributions (Docampo Chiaromonte, 2002). Another possible explanation is that attributions for parent behavior are dependent on the gender of the adolescent and which parent’s behavior (mother or father) is in question. Perhaps boys make different attributions for their mother’s behaviors than they make for their father’s and likewise for girls. As the larger study from which these data are drawn focused on fathering behaviors, we are unable to assess whether a gender of parent by gender of adolescent interaction exists.
In support of previous research (Nolen-Hoeksema et al., 1992), adolescents with depressive symptoms made more stable attributions for negative events and more unstable attributions for positive events than nondepressed adolescents, but gender and ethnicity moderated these differences. Adolescents with depressive symptoms may be more sensitive to the negative behaviors of others and offer more explanations for these behaviors, stable or unstable. Observed interactions of fathers and adolescents would help determine if fathers are acting more negatively towards adolescents or if adolescents perceive their fathers’ behaviors more negatively. Father-adolescent cohesion, or lack thereof, may also influence the interactions between adolescents and their fathers (Kaslow et al., 1994). Adolescents who do not get along with their fathers may make more hostile attributions for fathering behaviors that, in turn, may increase their conflict and lead to further depressive symptoms.
We also observed differences in attributions based on family structure; there were fewer stable attributions for the stepfather’s positive events than for intact fathers. Perhaps adolescent-stepfather relationships are disrupted because the stepfathers have more negative views of their children which may lead adolescents to make more negative attributions of their fathers’ behaviors. Stepfathers report their children as the cause for problems more often than biological parents (Stratton, 2003), and negative father-child interactions predict more negative child attributions (MacKinnon-Lewis et al., 2001). The relationship between stepfathers and adolescents may be bidirectional and stepfathers may act more negatively towards their stepchildren because attempts to connect with the adolescents are interpreted negatively. We attempted to account for the contributions to father-adolescent relationship quality; however, it is possible that unmeasured aspects of the stepfather child relationship (e.g., relationship with biological father) may affect attributions for the behaviors of stepfathers.
In replication of previous studies (Roberts & Sobhan, 1992), Mexican Americans reported higher levels of depressive symptoms than European Americans. In addition, they reported higher levels of unstable attributions for positive events and stable attributions for negative events. Familism, perceived obligation and respect among family members (Knight et al, 2007), may provide some explanation for our results. As Mexican Americans become increasingly acculturated and familism decreases, family conflict may increase which may account for the observed patterns of attributions. Mexican American families tend to be closer and interactions between parents and children may occur more often or have more meaning to Mexican American adolescents than when European American adolescents interact with their parents. Although we did not measure acculturation change in the current study, familism was significant in all regression models warranting further investigation into acculturation as it relates to the values and attributions of youth.
Limitations of Current Research and Future Directions
The results of this study offer new information about the predictors of adolescent attributions for their father’s positive and negative behaviors; however, there are a few limitations. First, we assessed only the stable-unstable dimension, which is important, especially in terms of the depressive symptoms of the adolescent (Gladstone & Kaslow, 1995; Nolen-Hoeksema et al. 1992), but it is likely that other dimensions may influence father-adolescent relationships. Second, the current study was limited by a focus on depressive symptoms and the CDI measure had low reliability for Mexican American boys. Further study on the validity of this scale for these youth is needed. Other social cognitive processes have been linked to externalizing behaviors and conduct disorders (Lansford et al., 2006), and it will be important to explore whether such associations exist for attributions of father behavior. Third, although the associations between study variables may be small, we believe our results provide evidence that depressive symptoms may be helpful in predicting attributions of fathering behaviors which may help to understand adolescent-father relationships. Future studies should replicate these results with clinically depressed adolescents and with older adolescents who may exhibit more signs of depressive symptoms. Finally, because our data are cross-sectional, potential bidirectional influences between these factors are unclear.
The current study has implications for advancing knowledge related to family interventions targeted towards changing belief systems within families. The attributions that adolescents make for their father’s behavior appear to be related to their levels of depressive symptoms, their gender, their ethnicity, and whether their father is a biological parent, and these associations hold after socioeconomic indicators, familism, and father-child relationship were covaried. Such results provide evidence that conflict and positive exchanges between parents and adolescents take on other meanings depending upon characteristics of the child and father. As attribution patterns for adolescents tend to differ systematically across behaviors and groups, these results suggest that interventions offering a cognitive focus may need to examine how to communicate messages to adolescents depending on the adolescents’ possible tendency to evaluate behavior in a particular way. By focusing attention squarely on how adolescents are likely to think, it is possible to begin to further explore the links between how they might behave.
Contributor Information
Andrea K. Finlay, The Pennsylvania State University, 110 South Henderson Building, University Park, PA 16802; Work number: 650-493-5000/1-2-22434; Fax: 650-617-2736; akf134@psu.edu
Jeffrey T. Cookston, San Francisco State University, 1600 Holloway Ave, San Francisco, CA 94132; Work number: 415-405-2133; Fax 415-338-2398; cookston@sfsu.edu
Delia S. Saenz, Arizona State University, PO Box 877805, Tempe, AZ 85287-7805; Work number: 480-965-1383; Fax: 480-965-0785; Delia.Saenz@asu.edu
Melinda E. Baham, Chandler-Gilbert Community College, 2626 East Pecos Road, Chandler, AZ 85225-2499; Work number: 480-726-4076; Fax: 480-857-5155; melinda.baham@cgcmail.maricopa.edu
Ross D. Parke, Department of Psychology, University of California, Riverside, Riverside, CA 92506; Work number: 951-827-6306; Fax: 949-464-9375; ross.parke@ucr.edu
William Fabricius, Arizona State University, 950 S. McAllister, Room 237 | P.O. Box 871104, Tempe, AZ 85287-1104; Work number: 480 965-9372; Fax: 480 965-8544; william.fabricius@asu.edu.
Sanford Braver, Arizona State University, 950 S. McAllister, Room 237 | P.O. Box 871104, Tempe, AZ 85287-1104; Work number: 602-330-1312; Fax: 760-448-5398; sanford.braver@asu.edu.
References
- Amato PR. Father-child relations, mother-child relations, and offspring psychological well-being in early adulthood. Journal of Marriage and the Family. 1994;56:1031–1042. [Google Scholar]
- Blumer H. Symbolic interactionism: Perspectives and method. Prentice Hall; Englewood Cliffs, NJ: 1969. [Google Scholar]
- Bollen K, Lennox R. Conventional wisdom on measurement: A structural equation perspective. Psychological Bulletin. 1991;110:305–314. [Google Scholar]
- Bugental DB, Johnston C, New M, Silvester J. Measuring parental attributions: Conceptual and methodological issues. Journal of Family Psychology. 1998;12:459–480. [Google Scholar]
- Chapman MV, Perreira KM. The well-being of immigrant Latino youth: A framework to inform practice. Families in Society. 2005;86:104–111. [Google Scholar]
- Coleman M, Ganong LH. Stepfamilies from the stepfamily’s perspective. Marriage and Family Review. 1997;26:107–121. [Google Scholar]
- Cookston JT, Finlay AK. Paternal involvement and adolescent adjustment: Longitudinal findings from Add Health. Fathering. 2006;4:137–158. [Google Scholar]
- Docampo Chiaromonte MM. Diferencias de género en las explicaciones causales de adolescentes. Psicothema. 2002;14:572–576. [Google Scholar]
- Elkind D. Inhelder and Piaget on adolescence and adulthood: A postmodern appraisal. Psychological Science. 1996;7:216–220. [Google Scholar]
- Fincham FD, Beach SRH, Arias H, Brody GH. Children’s attributions in the family: The children’s relationship attribution measure. Journal of Family Psychology. 1998;12:481–493. [Google Scholar]
- Flouri E. Fathering and child outcomes. Wiley & Sons; New York: 2005. [Google Scholar]
- Fornos LB, Mika VS, Bayles B, Serrano AC, Jimenez RL, Villarreal R. A qualitative study of Mexican American adolescents and depression. Journal of School Health. 2005;75:162–170. [PubMed] [Google Scholar]
- Fuligni AJ, Tseng V, Lam M. Attitudes toward family obligations among American adolescents with Asian, Latin American, and European backgrounds. Child Development. 1999;70:1030–1044. [Google Scholar]
- Giles-Sims J, Crosbie-Burnett M. Adolescent power in stepfather families: A test of normative-resource theory. Journal of Marriage and the Family. 1989;51:1065–1078. [Google Scholar]
- Gladstone T, Kaslow N. Depression and attributions in children and adolescents: A meta-analytic review. Journal of Abnormal Child Psychology. 1995;23:597–606. doi: 10.1007/BF01447664. [DOI] [PubMed] [Google Scholar]
- Gladstone TRG, Kaslow NJ, Seeley JR, Lewinsohn PM. Sex differences, attributional style, and depressive symptoms among adolescents. Journal of Abnormal Child Psychology. 1997;25:297–305. doi: 10.1023/a:1025712419436. [DOI] [PubMed] [Google Scholar]
- Kaslow NJ, Deering CG, Racusin GR. Depressed children and their families. Clinical Psychology Review. 1994;14:39–59. [Google Scholar]
- Knight GP, Jacobson RP, Gonzales NA, Roosa MW, Saenz DS. An evaluation of the psychological research on acculturation and enculturation processes among recently immigrating populations. In: Dalla RL, DeFrain J, Johnson J, Abbot D, editors. Strengths and challenges of new immigrant families: Implications for research, policy, education, and service. Lexington Publishers; Lanham, MD: 2009. [Google Scholar]
- Knight GP, Gonzales NA, Saenz DS, Bonds DD, Germán M, Deardorff J, et al. The Mexican American Cultural Values Scale for adolescents and adults. The Journal of Early Adolescence. 2010;30:444–481. doi: 10.1177/0272431609338178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Kovacs M. The Children’s Depression Inventory (CDI) Psychopharmacology Bulletin. 1985;21:995–999. [PubMed] [Google Scholar]
- Lal BB. Symbolic Interaction Theory. American Behavioral Scientist. 1995;38(1):421–441. [Google Scholar]
- Lamb ME, editor. The role of the father in child development. 5th ed Wiley; New York: 2010. [Google Scholar]
- Lansford J, Malone P, Stevens K, Dodge K, Bates J, Pettit G. Developmental trajectories of externalizing and internalizing behaviors: Factors underlying resilience in physically abused children. Development and Psychopathology. 2006;18:35–55. doi: 10.1017/S0954579406060032. [DOI] [PMC free article] [PubMed] [Google Scholar]
- MacKinnon-Lewis C, Castellino D, Brody G, Fincham F. A longitudinal examination of the associations between fathers’ and children’s attributions and negative interactions. Social Development. 2001;10:473–487. [Google Scholar]
- Nolen-Hoeksema S, Girgus JS, Seligman MEP. Predictors and consequences of childhood depressive symptoms: A 5-year longitudinal study. Journal of Abnormal Psychology. 1992;101:405–422. doi: 10.1037//0021-843x.101.3.405. [DOI] [PubMed] [Google Scholar]
- Paleari F, Regalia C, Fincham F. Adolescents’ willingness to forgive their parents: An empirical model. Parenting: Science and Practice. 2003;3:155–174. [Google Scholar]
- Parke RD, Coltrane S, Powers J, Adams M, Fabricius WV, Braver S, et al. Measurement of father involvement in Mexican-American families. In: Day R, Lamb M, editors. Conceptualizing and measuring father involvement. Erlbaum; Mahwah, NJ: 2003. pp. 17–38. [Google Scholar]
- Roberts RE, Sonhan M. Symptoms of depression in adolescence: A comparison of Anglo, African, and Hispanic Americans. Journal of Youth & Adolescence. 1992;21:639–651. doi: 10.1007/BF01538736. [DOI] [PubMed] [Google Scholar]
- Santisteban DA, Muir-Malcolm JA, Mitrani VB, Szapocznik J. Integrating the study of ethnic culture and family psychology intervention science. In: Liddle HA, Levant RF, Santisteban DA, Bray JH, editors. Family psychology: Science-based interventions. American Psychological Association; Washington, DC: 2002. pp. 331–352. [Google Scholar]
- Schwartz S. The applicability of familism to diverse ethnic groups: A preliminary study. Journal of Social Psychology. 2007;147:101–118. doi: 10.3200/SOCP.147.2.101-118. [DOI] [PubMed] [Google Scholar]
- Sheeber L, Hops H, Alpert A, Davis B, Andrews J. Family support and conflict: Prospective relations to adolescent depression. Journal of Abnormal Child Psychology. 1997;25:333–334. doi: 10.1023/a:1025768504415. [DOI] [PubMed] [Google Scholar]
- Stratton P. Causal attributions during therapy II: Reconstituted families and parental blaming. Journal of Family Therapy. 2003;25:161–180. [Google Scholar]
- Zayas LH, Lester RJ, Cabassa LJ, Fortuna LR. Why do so many Latina teens attempt suicide? A conceptual model for research. American Journal of Orthopsychiatry. 2005;75:275–287. doi: 10.1037/0002-9432.75.2.275. [DOI] [PubMed] [Google Scholar]