Disadvantaged, urban fathers' health and risky health behaviors: the role of relationship status and fathering practice.
Keywords: bidirectional relationships, fathers, family structure, health, parenting
Research on disadvantaged, minority men's health has focused on the confluence of socio-economic and contextual factors, mainly low socio-economic status, incarceration, and racial discrimination, that place low-income and minority men at a greater disadvantage for living healthy lives. The health disparities among low-income, minority men are a concern not only because of the financial cost to society, but because numerous low-income, minority men are fathers. Less research has focused on how relationship status and fathering practices influence paternal health and health behaviors. Understanding the association between relationship status, fathering practices and men's health could not only lead to overall improvement in the health of disadvantaged, minority men, but could also improve family relations.
RELATIONSHIP STATUS AND HEALTH
Multiple studies have indicated that there are several health benefits to marriage. Research on the relationship between couples' status and their health suggests that married couples compared to unmarried individuals have better self-rated health (Williams & Umberson, 2004) and mental health (Gove, Hughes, & Style, 1983; Marks & Lambert, 1998). In addition, married men compared to cohabiting men are less likely to report alcohol problems (Horwitz & White, 1998). With respect to parents' marital status and their health, the pattern is similar with married parents demonstrating better mental health compared to unmarried parents (Meadows, 2009). Married parents are also less likely to abuse alcohol or participate in risk taking activities compared to unmarried parents (Umberson, 1987) and are less likely to experience higher levels of depression compared to cohabiting parents (Brown, 2000; Kurdek, 1991). The exception is a study by Ross and colleagues (Ross, Mirowsky, & Goldsteen, 1990) who found no health difference between married and unmarried parents.
There is a dearth of research that has focused on how men's relationship status (i.e., family structure) influences health of disadvantaged fathers. Meadows (2009) found that among disadvantaged fathers, those continuously married have better self-rated health and fewer problems related to binge drinking, illicit drug use, and diagnosis of depression compared to continuously single fathers. However, Meadows suggests that because the disparity between two groups does not increase over time, the results give little support to the marital resource model. In other words, the benefits of marriage do not accumulate the longer the individual stays married.
The weak association between marital status and disadvantaged father's health might be because of (1) the fluidity and the lack of stability in relationships among this population and (2) the large proportion of childbearing outside of marriage among this population. Although cohabitation is becoming more common among all educational groups, cohabitation has been a common lifestyle among the less educated for several decades (Cherlin, 2010). However, the entry into cohabitation is less clear as the process is gradual, resulting in couples "sliding" or "drifting" into cohabitation (Manning & Smock, 2005), which consequently results in an ambiguous relationship status. Second, among low-income and less-educated populations and in neighborhoods where marriages are short lived, it is common to bear children outside of marriage (Edin & Kefalas, 2005). Thus, in populations where relationship statuses are vague and marriage is short-lived, marital status may be less significant than the family processes that occur among family units. Family processes are dynamic roles that individuals inherit and coincide with specific duties and normative behaviors (Yinger, 1965). Role theory suggests that individuals occupy multiple roles and the expectations of fulfilling these roles depend on the context and situation. Fathers have multiple roles and expectations that they must fulfill in order to meet their obligations which can promote or deter their health. Building off of work by Meadows (2009), the study below investigates whether changes in family processes in the form of fathering practices deter changes in negative health and health behaviors among disadvantaged, urban fathers. Further, the study also investigates whether a bidirectional relationship exists between fathering practices and men's health.
FATHERING PRACTICES AND HEALTH
From a psychosocial developmental perspective, adults have a desire to fulfill the needs of the next generation (Erikson, 1963); however, fulfilling children's needs, on top of other obligations and responsibilities, can at times be demanding for parents. For parents with resident children, children are consistently in the immediate surroundings and directly influencing how they feel. For instance, fathers who reside with their children are more likely to experience greater life satisfaction compared to those who do not reside with their children (Eggebeen & Knoester, 2001) and are less likely to engage in risk taking behaviors compared to men who are not fathers (Umberson, 1987). Resident fathers are also more involved in day-to-day activities with their children than nonresident fathers (Cabrera et al., 2004). Related, father involvement decreases as men father additional children in the same household (Harris & Morgan, 1991) and/or across different households (i.e., multiple partner fertility) (Cooksey & Craig, 1998; Manning & Smock, 1999; Manning & Smock, 2000; Manning, Stewart, & Smock, 2003).
Traditionally fathers have been designated as the main provider of the household, and economic downturns have the potential to jeopardize their individual well-being and relations with their families. Empirical studies have highlighted that men are particularly susceptible to spillover from stressful demands (Conger et al., 1990; Repetti, 1994). For example, work-related stress has been associated with resident fathers being withdrawn and displaying negative emotions towards their children (Repetti). With respect to mothers and parenting stress, mothers who report high levels of parenting stress also report greater psychological distress compared to mothers with low levels of parenting stress (Thompson, Merritt, Keith, Murphy, & Johndrow, 1993). Thus, high levels of parenting stress, or role strain, may have a negative influence on fathers' health.
According to family systems theory (Cox & Paley, 1997; Minuchin, 1974), parents experience a dynamic dyadic relationship when sharing childrearing responsibilities, or coparenting. Coparenting is characterized as the extent to which mothers and fathers work together and support each other in the parenting role. Though it does not necessarily encompass the romantic, emotional, nor financial aspects that two parents may share, coparenting cannot exist independently of the parents' overall relationship (Feinberg, 2003). Previous research has suggested that mothers and nonresident fathers who have contentious or disengaged relationships participate in parallel parenting rather than cooperative parenting; these parents minimally support each other's parenting role (Furstenberg, 1988; Maccoby & Mnookin, 1992). Thus, a lack of consistent parenting support may be related to fathers' negative well-being.
Building on transactional developmental theory (Bell, 1968; Sameroff, 1975), where children are said to be influenced in a reciprocal manner by the relationships and con text that surround them, fathers own development may be influenced in such manner as well. There are a couple of studies that have investigated the bidirectional relationship between paternal parenting and behavior among socio-economically disadvantaged, urban families. For example, Nepomnyaschy (2007) tested the reciprocal relationship between father-child contact and child support payment using cross-lagged regression models. Results indicated a positive reciprocal relationship between contact and informal child support payment. Carlson, McLanahan, and Brooks-Gunn (2008) also estimated cross-lagged effects to assess the relationship between coparenting support and non-resident father involvement using structural effect modeling techniques. Carlson and colleagues found coparenting support to strongly predict nonresident fathers' future involvement; however, nonresident father involvement weakly (but significantly) predicted future coparenting support. In terms of the current study, it may not only be that father's parenting practices influence their own health but that their health influences their parenting practices. For instance, healthy men may be better able to cope with parenting demands; whereas men with declining health may lessen parental involvement. If this is the case, then parenting programs and social service agencies are encouraged to address the health needs of fathers in order to facilitate greater parenting practices among men.
Earlier research on marital status and health has primarily focused on populations where marriage is common, relationships are stable, and childbearing usually occurs after marriage. Further, previous research on father parenting practices has primarily examined the association between father involvement and child development (e.g., Coley & Medeiros, 2007; Flouri & Buchanan, 2004), with little focus on the association between paternal parenting practices and fathers own health and even less attention on whether a bidirectional relationship exists between paternal parenting practices and paternal health. The current study attempts to gain a better idea of how changes in relationship status and family processes in the form of fathering practices enhance or lessen disadvantaged, urban fathers' self-rated health, mental health (i.e., depression), and drug use. The panel data provides relationship status and health outcomes at the four available time points. It is hypothesized that cohabiting and single fathers compared to married fathers will be associated with worse health and health behavior outcomes. In addition, the data provides information on several family processes in the form of fathers' parenting practices. This allows the opportunity to assess the extent to which fathers' parenting practices mediate the relationship between relationship status and fathers' health outcomes. It is hypothesized that fathers' parenting practices will fully mediate the association between relationship status and health outcomes. Last, the availability of the longitudinal data provides the opportunity to assess whether a bidirectional relationship exists between parenting practices and health outcomes. Based on previous research that has found a bidirectional relationship between paternal parenting and behavior among socio-economically disadvantaged, urban father (Carlson, McLanahan, & Brooks-Gunn, 2008; Nepomnyaschy, 2007), it is hypothesized that a bidirectional relationship exists between father parenting practices and health outcomes in the current sample of fathers.
Data are drawn from the Fragile Families and Child Well-Being Study (FFCW), a longitudinal study that examines the conditions and capabilities of a disadvantaged urban cohort of 4,898 parents of newborn children. Approximately three quarters of the parents are unmarried at the time of the child's birth. The full sample is representative of nonmarital births in cities where population exceeds 200,000. Mothers and fathers were interviewed separately soon after birth of their child (i.e., year 0), and again at 1,3, and 5 years from the child's birth; however, the analyses below focused solely on information received from fathers with the exception of child characteristics which were based on mother reports. For information on sample and design of the study please see Reichman, Teitler, Garfinkel, and McLanahan (2001).
Taking advantage of the repeated measures in multiple waves of data, the data for this study was organized as a pooled time series in order to study how changes in relationship status and father parenting practices influences changes in health. In pooled time series data, the space and time dimensions of the data are used. The spatial dimensions of the data relate to the cross-sectional units of observations (e.g., individuals such as fathers); the temporal dimension of the data relates to the periodic observations of a set of variables (e.g., father involvement) characterizing the cross-sections units (e.g., fathers) over a specific time span (e.g., 4 different waves of data) (Yaffee, 2005). Observations for each father in the sample are represented by a separate record for each wave of data collection. In other words, the same respondent contributes to more than one record in the data; this results in the total number of observations being the number of fathers times the number of waves of data (i.e., 4) (Johnson, 1995).
Fixed effects models were then used to estimate the effect of relationship status and fathering practices on health outcomes. In a pooled time series with fixed-effects estimators the models are based on variation within-individuals in the sample over time (rather than between-individuals). When between-individual variation is eliminated, sampling variability is also eliminated; in addition, larger standard errors can result compared to methods that include between- and within-individual variation. However, eliminating between-individual variation results in models with a lower likelihood of unbiased estimates because unmeasured characteristics that could contaminate or bias the estimates have been eliminated. Although there is no direct measurement of time invariant characteristics in the models, the fixed effect estimator does control for all stable characteristics, such as race, that could influence the independent and dependent variable (Johnson, 1995). Further, with this data arrangement, all of the dependent, independent, and control variables essentially represent change scores. For example, the coefficient for "father involvement" will reflect either a positive or negative change in involvement over time.
The multivariate regression model estimates were performed in several steps. The first two steps involved using pooled time series fixed-effects models to estimate the association between relationship status, fathering practices, and health outcome of interest. Specifically, step one included regressing the health outcome of interest (e.g., poor health) on relationship status variables and the covariates. Step two included adding fathering practice variables to the regression equation and assessing whether mediation occurred using methods outlined by Baron and Kenny (1986) and Sobel and Leinhardt (1982).
Step three focused on assessing whether a bidirectional relationship exists between father parenting practices and health outcomes of interest. In these regression models, cross-lagged regression models were estimated for fathering practices that significantly predicted health outcomes from step two. For instance, father involvement in wave 4 was regressed on father involvement in wave 2, the health predictor of interest (e.g., depression) in wave 2, the relationship status variables and other covariates. Then depression in wave 4 was regressed on depression in wave 2, father involvement in wave 2, the relationship status variables, and the other covariates. Because fathering practices was not measured at the baseline survey (i.e. year 0), data from year 2 and year 4 were used to estimate the cross-lagged regression models. Ordinary least-square (OLS) regressions were used in to estimate the regression models when the outcome of interest was a continuous variable and logistic regression models were used to estimate regression models when outcome of interest was a dichotomous variable.
The analysis sample was based on data that was pooled from four waves of panel data. The total sample size analyzed was 19,592 records from 4,898 different individuals represented in the sample. Multiple imputation techniques were used to impute missing data for all variables using STATA's ICE command. With multiple imputation all relevant cases are kept in the analysis for accurate parameter estimates and the standard errors are corrected for the amount of missing information. This is in comparison to list wise deletion, which results in a sample that no longer represents the population (Graham & Schafer, 1999; Graham, 2009). The multiple imputation technique created five imputed data sets. To perform multivariate regression models, all models employed the five mirror data set with the use of STATA's MICOMBINE. STATA's MICOMBINE command in conjunction with the XTREG and XTLOGIT command was used to estimate pooled time series fixed-effects models; while STATA'S MICOMBINE command in conjunction with REG and LOGISTIC command was used to estimate the cross-lagged regression models.
Dependent variables. Three indicators of health were used to measure fathers' wellbeing over time. Fathers reported on their overall health in the past year on a 1 = poor to 5 = excellent scale. Items were reverse coded with higher values indicating poor health. Diagnosis of major depression is dichotomous measure created by the FFCW project based on fathers' affirmative responses to items from the Composite International Diagnostic Interview Short Form (CIDI-SF) in the past year (Kessler et al., 2003). Last, fathers reported whether they engaged in drug use, such as marijuana, cocaine, or heroin, or used drugs without a doctor's prescription in larger amounts than prescribed in the past year. Due to the relatively low prevalence of drug use, drug use was scored as a group, adding 1 to the overall score, regardless of the number of specific drugs used. Thus, a dichotomous measure was created based on fathers' affirmative responses.
Self-rated health was measured at all 4 time points; however, depression was measured at 3 time points (i.e., year 1,3, and 5) and drug use was measured at 2 time points (i.e., year 3 and 5). Models were conducted with the available time points and also with imputed data for the waves in which data was not collected. The results are robust with the imputed data when compared to the non-imputed data; therefore, for reasons of consistency, the results from the imputed data are presented.
Independent variables. The independent variables comprised of relationship status and fathering characteristics. Relationship Status. Fathers were asked whether they were married (1 = married; 0 = not married), cohabiting (1 = cohabiting; 0 = not cohabiting), or single (1 = single; 0 = not single) to the focal child's biological mother at all four data collection points. Three mutually exclusive dichotomous variables were created representing fathers' relationship status at each data collection point. Within the context of pooled time series analysis, the measures of relationship status reflect shifts into and out of the above statuses. Fathering Characteristics. Three indicators of family dynamics were used to measure fathering characteristics at the three available data collection points (i.e., Year 1 - 3). Regarding father involvement, fathers reported on how many days a week (0 to 7) they engaged in 4 items capturing caretaking activities with the focal child (e.g., sing songs; read stories; tell stories; play inside with toys). At each data collection point, items were averaged to create a father involvement measure where a higher score indicated greater engagement ([[alpha].sub.1] = .78, [[alpha].sub.2] = .80, [[alpha].sub.3] = .77). Role strain was measured based on father reports of 4 items describing their role as parents on a 1 = strongly agree to 4 = strongly disagree scale (e.g., I feel trapped by my responsibilities as a parent to the focal child). Items were reverse coded and summed to represent greater role strain ([[alpha].sub.1] = .58, [[alpha].sub.2] = .62,
[[alpha].sub.3] = .65). Based on 5 items, fathers responded to the extent to which they felt supported in the parenting role by the other parent using a 1 = always to 3 = never scale (e.g., "How often does [the focal child's mother] support you in the way you want to raise your child?"). Items were reverse coded and summed to represent higher coparenting support ([[alpha].sub.1] = .67, [[alpha].sub.2] = .66, [[alpha].sub.3] = .76).
Covariates, The following variables are included as covariates in the model because they reflect selection factors that could lead men into different relationship statuses, as well as having different fathering experiences. By adding these covariates to the model, a better estimate of the effects of family structure and fathering on fathers' health will be attained, as well as the effects of health (and family structure) on fathering practices. Covariates included the child's age (in months) and health (1 = excellent health to 5 = poor health) as reported by the mother. Fathers reported on their age (in years), employment (number of weeks employed in the past year), and whether or not they experienced economic hardship in the past year (1 = yes; 0 = no). Father's household income was categorized into three federal poverty thresholds [poor (< 100% FPL; reference), working poor (100-299% FPL), not poor ([greater than or equal to]300% FPL)]. Fathers also reported on the number of non-biological and biological children that resided in the household with him, along with the number of non-biological children that did not reside with him. Father's health insurance coverage was captured as private, public (reference), and no insurance.
The descriptive information regarding all the variables in the pooled time series data set is presented in Table 1. For sake of simplicity the results displayed are based on the non-imputed data; however, the results are similar to the imputed data. On average across all four waves of data, fathers were healthy with the majority of fathers averaging very good health (score of 2), 12 percent experienced a clinical diagnosis of depression, and 11 percent engaged in drug use. When averaged across all four waves, thirty percent of the fathers were cohabiting, 35 percent of the fathers were married, and 35 percent were single. Further, fathers were engaged in fathering activities slightly more than 3 days a week, experienced an average role strain of 2 (out of 4), and perceived to receive high coparenting support (averaged 2.8 out of 3) when pooled across all four waves of data collection. Additionally, fathers averaged across the four data collection points 40 hours a week of employment, 47 percent of fathers experienced economic hardship, and 27 percent had household incomes in the < 100 percent FPL range. Fourteen percent of fathers lacked health insurance coverage across the four waves of data.
Table 2 displays cross-sectional descriptive characteristics of the main variables. In general, there was little fluctuation in father's health over the four time points. However, the relationship status of fathers did fluctuate over time, especially among cohabiting and single fathers. While cohabitating unions with the focal child's mother decreased over time (44% at Year 0 to 16 percent at Year 5), single fatherhood increased over time from 28 percent at Year 0 to 45 percent at Year 5. The greatest cross-sectional differences among fathering practices was observed among father involvement. On average, fathers were involved with their children approximately 4 days a week in Year 1; by Year 5, fathers were involved with their children approximately 3 days a week. Role strain and coparenting support remained fairly steady throughout the 4 year period.
Multivariate Regression Models
The first set of regressions models estimated the effect of relationship status and fathering characteristics on poor health, controlling for socio-demographic family characteristics using fixed effects estimator (Table 3). In column 1, model 1, changes in relationship status did not predict changes in poor health. In column 2, model 2, relationship status continued to be non-significant, and father involvement and co-parenting support significantly predicted better health over time (i.e., improvements in poor health). Fathers who increased their involvement and those who felt greater parenting support over time were less likely to experience negative changes in their poor health over time. Socio-economic factors also strongly predicted poor health. Fathers who experienced an increase in economic hardship were more likely to also have experienced increases in poor health. And fathers who experienced increases in their socio-economic status (as measured by 100-299% of FPL and [greater than or equal to]300% of FPL) were less likely to have experienced worsening of their poor health.
When focusing on the association between relationship status and depression, fathers who became single fathers compared to fathers who married were associated with an increased likelihood of becoming clinically depressed (model 1, column 3). The coefficient for the effect of single fatherhood on depression decreased from .67 to .36, and became non-significant, once parenting practices were included in model 2 (column 4). All three parenting practices significantly predicted depression. Fathers who increased their involvement and those who felt increased support in their parenting over time were less likely to become depressed. However, fathers whose role strain increased over time were more likely to become depressed. Further, the Sobel test assessed whether the decline in the coefficient for single parenthood predicting depression was significantly related to parenting practices. Results indicated that father involvement (z = 2.80, p < .01) and co-parenting support (z = 1.98, p < .05) fully mediated the positive association between single parenthood and depression. Role strain was not found to be a significant mediator.
In the final two sets of models, changes in relationship status did not predict changes in drug use in model 1 (column 5) nor in model 2 (column 6). When fathering practices were added to the regression model (column 6, model 2), role strain predicted drug use. Fathers whose role strain increased over time were more likely to begin using drugs. Similar to the models predicting poor health, economic hardship was a strong predictor of drug use in both models. Fathers who experienced increases in economic hardship over time were more likely to begin using drug.
Table 4 displays the cross-lagged regression estimates for models predicting fathering practices and health. Results indicated that there is a bidirectional relationship between all three fathering practices and fathers experiencing depression (models 1, 3 and 5). Fathers who were depressed decreased their involvement (model 1, column 1), experienced increased levels of role strain (model 3, column 1), and decreased levels of co-parenting support (model 5, column 1). Further, fathers with higher levels of father involvement (model 1, column 2) and of co-parenting support (model 5, column 2) decreased their chances of being depressed over time; while, fathers who experienced higher levels of role strain increased their chances of being depressed over time (model 3, column 2).
The purpose of study was to use longitudinal data to examine how relationship status and family processes in the form of father parenting practices influence disadvantaged, urban fathers' health and risky health behaviors. In doing so, pooled time series fixed effects models indicated that relationship status has no influence on self-rated health or drug use; however, single parenthood was associated with experiencing depression. Specifically, the association between single parenthood and depression is fully mediated by parenting practices, such as father involvement and co-parenting support. Cross-lagged regression estimates indicated that a bidirectional relationship exists between father parenting practices and paternal depression. An inverse bidirectional relationship exists between a) father involvement and depression and b) co-parenting support and depression. A positive bidirectional relationship exists between role strain and depression. Several socioeconomic factors were a strong predictor of poor health and drug use. More details pertaining to the association between relationship status and health, the bidirectional relationship between fathering practices and depression, and the association between socio-economic status and health, as well as the contributions and limitations of this study, are discussed below.
Contrary to the hypothesis, relationship status did not predict fathers' poor health or drug use. Further, only in one model did father involvement and co-parenting support fully mediate the relationship between single parenthood and depression. Thus, the time period of 5 years may not be a long enough window to observe health improvements (or declines) associated with relationship status if they take longer to manifest. At the same time, relationship status may play less of a role in the health status of socioeconomically disadvantaged men considering the fluidity and lack of stability in relationships in less educated populations (Cherlin, 2010) and the large proportion of nonmarital childbearing in communities where marriages are short-lived (Edin & Kefalas, 2005). The FFCW sample mirrors the above description as seventy-five percent of parents in this sample were unmarried at the birth of the focal child in the survey (Reichman et al, 2001) and family instability is common (see Beck, Cooper, McLanahan, & Brooks Gunn, 2010; Bzostek & Beck, 2010). Just as others have found a weak association between relationship status and health among this population (i.e., Meadows, 2009), the present study did as well; in addition, the findings from the present study suggest that the proximal measures of family processes and contextual factors contribute to disadvantaged, urban fathers' health and risky health behaviors. The majority of prior research that has found a positive association between marital status and health has primarily focused on mainstream populations in which marriages are both more common and more stable and childbearing usually occurs after marriage.
Bidirectional Relationship between Father Parenting Practices and Health
According to transactional developmental theory (Bell, 1968; Sameroff, 1975), individual development is influenced in a reciprocal manner by the relationships and the context that surround them. Estimates from cross-lagged regression models indicated that a bidirectional relationship exists between parenting measures and depression. A decrease in father involvement, or less physical engagement with a child on a daily basis, was an important predictor of depression among fathers. In addition, experiencing depression was associated with decrease levels of father-child engagement. The same inverse bidirectional relationship described above was found between co-parenting support and depression. The association between fathering and mental health complements Erikson's theory of generativity (Erikson, 1963). Fathers who engage in increasing levels of childrearing and who experience increasing levels of co-parenting support over time may receive a sense of satisfaction and confidence from the positive interactions. The satisfaction and confidence may result in fathers to be more inclined to take care of themselves in order to ensure the continuation of life satisfaction and confidence. The mental upkeep, in turn, facilitates greater father involvement and supportive parenting relationships. This is an important finding as some fathers are accessible to their children but interact with their children minimally.
The findings also suggest that an increase in role strain increases the likelihood of experiencing depression and that experiencing depression is related to experiencing increases in role strain. The analyses in this paper focused on a developmental time period when children are highly dependent on there parents as they are in their early childhood stage of development. The majority of the fathers became single by year 5 of the survey, and over time 36 percent of the fathers have children with another woman that is not the focal child's mother and 39 percent of the fathers have step-children (i.e., nonbiological children) residing with him (analyses not shown). Although the measure of role strain in the study focused on the father's perception of his role in relation to the focal child in the study, it may be difficult for fathers to isolate parenting stress he receives from the focal child and focal child's mother, from the parenting stress he receives from a) his other biological child(ren) and their mother(s) from other relationships and b) his step-children. Further, it may be difficult for fathers to isolate parenting stress from work-related stress. Thus, these conflicting demands may have a spillover effect and can contribute to greater role strain, or parenting stress, as measured in this study which appears to be contributing negatively to father's mental health. In turn, experiencing negative mental health facilitated increases in role strain. In order to better understand how fathers cope with multiple parenting roles across households, while maintaining work-related responsibilities, qualitative studies that focus on men's parenting needs are necessary.
The results suggest that fathering has broader implications than child development. Parent education programs that focus on building stronger father-child relationships through father-and-child related activities may not only benefit the child, but the father's sense of well-being and competence as well. The results suggest that if parent education programs want to facilitate greater parenting practices among fathers, they must address the mental health needs of fathers.
Experiencing economic hardship was a consistent predictor of poor health and engagement in drug use, while greater socio-economic status was predictive of better health. A life course perspective suggests that the health of individuals may be affected negatively when developmental milestones, such as marriage and childbearing, occur at non-normatives stages (Elder, 1985). For men who are financially unstable, the birth of a child (or additional children to be responsible for) may mean a restructuring of financial responsibilities, along with an increase in child care responsibilities. Thus, the unanticipated juggling of financial and caretaking responsibilities may enhance drug use and a decline in self-reported health among fathers who were not prepared for fathering responsibilities.
To capture the complexity of father parenting practices, fathers' survey reports of his own fathering behaviors over time were used. This is a contribution to the literature as an abundance of research on fathering continues to be based on mother reports. However, limitations in the measures and analysis do arise. The measures of the independent variables and the majority of the covariates, along with the dependent variables, were based on fathers' reports, which may introduce bias in the results. Although mothers, compared to fathers, have a tendency to report lower levels of father involvement (Coley & Morris, 2002; Mikelson, 2008), prior research on disadvantaged fathers found father reports of father involvement to be reliable and valid (Hernandez & Coley, 2007). Further, the measures of self-rated health and depression have been used in many epidemiological studies and validated on adults (Kessler et al., 2003; Miilunpalo, Vuori, Oja, Pasanen, & Urponen, 1997). The self-rated health assessment is a valid health status indicator in adults and a strong predictor of the number of physician contacts per year, physical fitness, and mortality (Miilunpalo et al., 1997).
Aside from the limitations in the measures, the sample presents limitations as well. The data described in this study is derived from socio-economically disadvantaged, urban fathers, whose partners gave birth in urban hospitals. The fathers in the sample have young children which make them susceptible to a unique set of stressors. Thus, the fathers (and their families) in the current sample may differ systematically from other fathers and their families, limiting the generalizability of the findings. Related, the focus placed on socio-economically disadvantaged, urban fathers in their mid-twenties, results in the sample to be quite restricted. The restricted sample may display less variation in health status and health behaviors compared to the entire U.S. population. In turn, the findings could be conservative due to the limited variation.
In the pooled time series models, the data was pooled from the four existing data collection time points in order to take advantage of the longitudinal data and analyses were conducted with a fixed effects estimator. An advantage of the using a fixed effect estimator on data in a pooled time-series format is that specification errors due to unmeasured characteristics can be eliminated and thus stronger inferences about the findings can be concluded compared to alternative methods. Additionally, this type of modeling technique estimates coefficients based on a restricted number of time varying characteristics which results in the coefficient estimates to only be based on variation within individuals over time. Thus, the downfall of this type of modeling technique is that if there is less variation within individuals (compared to between individuals as in alternate methods), the technique could be considered less efficient (Johnson, 1995).
Current marriage promotion and family policies focus on the "status," or the position a person holds within a family structure. The results from this study demonstrate that single parenthood is pertinent to disadvantaged, urban fathers' mental health; but in general, relationship status is not as influential on disadvantaged, urban fathers' health as the association between parenting behaviors and health. Specifically, the results illustrate how proximal parenting behaviors associated with fathers' role in the family unit influences paternal health and how health can influence parenting behaviors. If the end goal of family policies and parent education programs is to improve family relations, policies and programs need to emphasize positive engagement, regardless of family structure. In addition, parent education programs need to provide support and resources to disadvantage urban fathers who would benefit from mental health assistance. Such a change in the emphasis of family policies and parent education programs may improve family relations and fathers' mental health.
The author thanks Professor Paul Amato for methodological assistance and comments on the preliminary draft.
Baron, R.M., & Kenny, D.A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations. Journal of Personality and Social Psychology, 51 (6), 1173-1182.
Beck, A.N., Cooper, C.E., McLanahan, S., & Brooks Gunn, J. (2010). Partnership transitions and maternal parenting. Journal of Marriage and Family, 72(2), 219-233.
Bell, R.Q. (1968). A reinterpretation of the direction of effects in studies of socialization. Psychological Review, 75, 81-95.
Brown, S.L. (2000). The effect of union type on psychological well-being: Depression among cohabitors versus marrieds. Journal of Health and Social Behavior, 41, 241-255.
Bzostek, S., & Beck, A. (2010). Family instability and young children's physical health. Unpublished Working Paper WP08-11-FF.
Cabrera, N., Shannon, J., Vogel, C., Tamis-Lemonda, C., Ryan, R., Brooks-Gunn, J., et al. (2004). Low-income fathers' involvement in their toddlers' lives: Biological fathers from the Early Head Start Research and Evaluation Study. Fathering: A Journal of Theory, Research, and Practice About Men as Fathers, 2(1), 5-30.
Cherlin, A. (2010). Demographic trends in the United States: A review of research in the 2000s. Journal of Marriage and Family, 72(3), 403-419.
Coley, R.L., & Medeiros, B .L. (2007). Reciprocal longitudinal relations between nonresident father involvement and adolescent delinquency. Child Development, 78, 132-147.
Coley, R.L., & Morris, J.E. (2002). Comparing father and mother reports of father involvement among low-income minority families. Journal of Marriage and Family, 64(4), 982-997
Conger, R.D., Elder, G.H., Jr., Lorenz, F.O., Conger, K.J., Simons, R.L., Whitbeck, L.B., et al. (1990). Linking economic hardship to marital quality and instability. Journal of Marriage and the Family, 52, 643-656.
Cooksey, E.C., & Craig, P.H. (1998). Parenting from a distance: The effects of paternal characteristics on contact between nonresidential fathers and their children. Demography, 35, 187200.
Cox, M., & Paley, B. (1997). Families as systems. Annual Review of Psychology, 48,243-267.
Edin, K., & Kefalas, M. (2005). Promises I can keep: Why poor women put motherhood before marriage. Berkeley, CA: University of California Press.
Eggebeen, D., & Knoester, C. (2001). Does fatherhood matter for men? Journal of Marriage and Family, 63(2), 381-393.
Elder, G.H. (1985). Life course dynamics: Trajectories and transitions 1968-1980. New York: Cornell University Press.
Erikson, E.H. (1963). Childhood and society. New York: Norton.
Fagan, J. (2009). Relationship quality and changes in depressive symptoms among urban, married African Americans, Hispanics, and Whites. Family Relations, 58(3), 259-274.
Feinberg, M.E. (2003). The internal structure and ecological context of coparenting: A framework for research and intervention. Parenting, 3, 95-131.
Flouri, E., & Buchanan, A. (2004). Early father's and mother's involvement and child's later educational outcomes. British Journal of Educational Psychology, 74, 141-153.
Frech, A., & Williams, K. (2007). Depression and the psychological benefits of entering marriage. Journal of Health and Social Behavior, 48, 149-163.
Furstenberg, F.F., Jr. (1988). Marital disruptions, child custody, and visitation. In A.J. Kahn & S.B. Kamerman (Eds.), Child support: From debt collection to social policy (pp. 277-305). Newbury Park, CA: Sage.
Gove, W., Hughes, M., & Style, C. (1983). Does marriage have positive effects on the psychological well-being of the individual? Journal of Health and Social Behavior, 24, 122-131.
Graham, J.W., & Schafer, J.L. (1999). On the performance of multiple imputation for multivariate data with small sample size. In R.H. Hoyle (Ed.), Statistical strategies for small sample research (pp. 1-27). Thousand Oaks, CA: Sage.
Graham, J.W. (2009). Missing data analysis: Making it work in the real world. Annual Review of Psychology, 60,549-576.
Harris, K., & Morgan, S. (1991). Fathers, sons, and daughters: Differential paternal involvement in parenting. Journal of Marriage and the Family, 53, 531-544.
Hernandez, D.C., & Coley, R.L. (2007). Measuring father involvement within low-income families: Who is a reliable and valid reporter? Parenting: Science and Practice, 7(1), 69-97.
Horwitz, A., & White, H. (1998). The relationship of cohabitation and mental health: A study of a young adult cohort. Journal of Marriage and the Family, 60, 505-514.
Johnson, D. (1995). Alternative methods for the quantitative analysis of panel data in family research:
Pooled time-series models. Journal of Marriage and the Family, 57(4), 1065-1077.
Kessler, R.C., Berglund, P., Demler, O., Jin, R., Koretz, D., Merikangas, K.R., et al. (2003). The epidemiology of major depressive disorder: Results from the National Comorbidity Survey Replication (NCS-R). Journal of the American Medical Association, 289(23), 3095-3105.
Kurdek, L. (1991). The relations between reported well-being and divorce history, availability of a proximate adult, and gender. Journal of Marriage and the Family, 53, 71-78.
Maccoby, E.E., & Mnookin, R.H. (1992). Dividing the child: Social and legal dilemmas of custody: Harvard University Press.
Manning, W., & Smock, P. (2005). Measuring and modeling cohabitation: New perspectives from qualitative data. Journal of Marriage and Family, 67(4), 989-1002.
Manning, W.D., & Smock, P.J. (1999). New families and nonresident father-child visitation. Social Forces, 78, 87-116.
Manning, W.D., & Smock, P.J. (2000). "Swapping" families: Serial parenting and economic support for children. Journal of Marriage and the Family, 62, 111-122.
Manning, W.D., Stewart, S.D., & Smock, P.J. (2003). The complexity of fathers' parenting responsibilities and involvement with nonresident children. Journal of Family Issues, 24,645-667.
Marks, N.F., & Lambert, J.D. (1998). Marital status continuity and change among young and midlife adults--Longitudinal effects on psychological well-being. Journal of Family Issues, 19, 652-686.
Meadows, S.O. (2009). Family structure and fathers' well-being: Trajectories of mental health and self-rated health. Journal of Health and Social Behavior, 50(2), 115-131.
Mulunpalo, S., Vuori, I., Oja, P., Pasanen, M., & Urponen, H. (1997). Self-rated health status as a health measure: The predictive value of self-reported health status on the use of physician services and on mortality in the working-age population. Journal of Clinical Epidemiology, 50, 517-528.
Mikelson, K.S. (2008). He said, she said: Comparing mother and father reports of father involvement. Journal of Marriage and Family, 70(3), 613-624.
Minuchin, S. (1974). Families and family therapy. Cambridge, MA: Harvard University Press.
Reichman, N.E., Teitler, J.O., Garfinkel, I., & McLanahan, S.S. (2001). Fragile families: Sample and design. Children and Youth Services Review, 23(4-5), 303-326.
Repetti, R.L. (1994). Short-term and long-term processes linking job stressors to father-child interaction. Social Development, 3, 1-15.
Ross, C., Mirowsky, J., & Goldsteen, K. (1990). The impact of the family on health: The decade in review. Journal of Marriage and the Family, 52, 1059-1078.
Sameroff, A. (1975). Transactional models in early social relations. Human Development, 18, 6579.
Sobel, M.E., & Leinhardt, S. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. Sociological Methodology, 13,290-312.
Thompson, R., Merritt, K., Keith, B., Murphy, L., & Johndrow, D. (1993). The role of maternal stress and family functioning in maternal distress and mother-reported and child-reported psychological adjustment of nonreferred children. Journal of Clinical Child & Adolescent Psychology, 22(1), 78-84.
Umberson, D. (1987). Family status and health behaviors- social-control as a dimension of social integration. Journal of Health and Social Behavior, 28, 306-319.
Umberson, D., Williams, K., Powers, D.A., Liu, H., & Needham, B. (2006). You make me sick: Marital quality and health over the life course. Journal of Health and Social Behavior, 47, 1-16.
Williams, K., & Umberson, D. (2004). Marital status, marital transitions, and health: A gendered life course perspective. Journal of Health and Social Behavior, 45, 81-98.
Yaffee, R.A. (2005). A primer for panel data analysis, from http://www.nyu.edu/its/statistics/Docs/pda.pdf
Yinger, M.J. (1965). Toward a field theory of behavior: Personality and social structure. New York: McGraw-Hill.
Correspondence concerning this article should be sent to the author, Department of Health & Human Performance, 3855 Holman Street, Garrison Gymnasium Room 104, Houston, TX 77204. Email: email@example.com
DAPHNE C. HERNANDEZ, PHD, MSED *
* University of Houston.
Table 1 Proportion or Mean (Standard Deviation) of Variables in Pooled Dataset M SD Health Outcomes Poor Health 2.07 0.98 Depression 0.12 -- Drugs 0.11 -- Relationship Status Married 0.35 -- Cohabitation 0.30 -- Single 0.35 -- Fathering Practices Father Involvement 3.36 2.08 Role Strain 2.05 0.68 Coparenting Support 2.77 0.33 Covariates Child Age (months) 26.83 23.42 Child Poor Health 1.52 0.78 Father Age (years) 30.19 7.46 Father is Employed (# of weeks) 40.58 17.88 Father experienced Economic Hardship 0.47 -- < 100% of FPL 0.27 -- 100 - 299% of FPL 0.42 -- [less than or equal to] 300% of FPL 0.31 -- # of Non-biological Children Residing with Father 0.39 0.85 # of Biological Children Residing with Father 0.85 1.11 # of Biological Children Not Residing with Father 0.44 0.84 Father has Private Insurance 0.45 -- Father has Public Insurance 0.41 -- Father has No Insurance 0.14 -- Note. Descriptives based on non-imputed pooled data. FPL = Federal Poverty Line Table 2 Proportion or Mean (Standard Deviation) of Main Variables at Each Year of Data Collection Year 0 Year 1 Year 3 or M SD % or M SD % or M SD Health Outcomes Poor Health 2.02 0.94 2.10 1.01 2.01 0.96 Missing Values 7 696 873 Depression n/a 0.10 -- 0.14 -- Missing Values 698 870 Drugs 0.14 -- 0.07 -- 0.12 -- Missing Values 11 703 864 Relationship Status Married 0.28 -- 0.36 -- 0.39 -- Missing Values 0 702 870 Cohabitation 0.44 -- 0.35 -- 0.24 -- Missing Values 0 702 870 Single 0.28 -- 0.31 -- 0.37 -- Missing Values 0 702 870 Fathering Practices Father Involvement n/a 3.75 2.09 3.53 2.09 Missing Values 728 868 Role Strain n/a 2.06 0.67 2.07 0.67 Missing Values 1112 1111 Co-parenting Support n/a 2.79 0.33 2.80 0.29 Missing Values 1585 1148 Year 5 % or M SD Health Outcomes Poor Health 2.16 0.99 Missing Values 1048 Depression 0.11 -- Missing Values 1047 Drugs 0.11 -- Missing Values 1038 Relationship Status Married 0.38 -- Missing Values 1046 Cohabitation 0.16 -- Missing Values 1046 Single 0.45 -- Missing Values 1046 Fathering Practices Father Involvement 2.95 1.93 Missing Values 1039 Role Strain 2.02 0.69 Missing Values 1337 Co-parenting Support 2.72 0.34 Missing Values 1092 Note. Descriptives based on non-imputed data. n/a = Survey did not ask respondents about this topic in year 0; therefore, data not available. Table 3 Pooled Time Series Fixed Effects Regression Estimates Predicting Father's Health and Risk Taking Behaviors Poor Health Model 1 Model 2 B (SE) B (SE) Relationship Status Married (reference) -- -- Cohabitation -0.06 (0.07) -0.06 (0.07) Single 0.04 (0.08) -0.02 (0.08) Fathering Practices Father Involvement -- -0.03 (0.10) ** Role Strain -- 0.05 (0.03) Co-parenting Support -- -0.14 (0.05) * Covariates Child Age (months) 0.00 (0.00) 0.00 (0.00) Child Health 0.00 (0.02) 0.00 (0.02) Father Age (years) 0.02 (0.03) 0.02 (0.02) Father is Employed (# of weeks) -0.00 (0.00) * -0.00 (0.00) Father experienced Economic Hardship 0.14(0.03) *** 0.13(0.03) *** < 100% of FPL (reference) -- -- 100 - 299% of FPL -0.09 (0.04) * -0.09 (0.04) * [greater than or equal to] -0.15(0.06) ** -0.15(0.06) ** 300% of FPL # of Non-biological Children Residing with Father -0.02 (0.03) -.02(0.03) # of Biological Children Residing with Father -0.02 (0.02) -.03(0.02) # of Biological Children Not Residing with Father -0.02 (0.03) -.02(0.03) Father has Private Insurance -0.05 (0.05) -0.04 (0.05) Father has Public Insurance 0.01 (0.05) 0.01 (0.05) Father has No Insurance -- -- (reference) Constant 1.65 (0.69) * 2.10 (0.70) ** Depression Model 1 Model 2 B (SE) B (SE) Relationship Status Married (reference) -- -- Cohabitation 0.02 (0.35) -0.06 (0.36) Single 0.67 (0.34) * 0.36 (0.36) Fathering Practices Father Involvement -- -0.14 (0.05) ** Role Strain -- 0.39 (0.14) ** Co-parenting Support -- -0.52 (0.25) * Covariates Child Age (months) -0.02 (0.01) -0.02 (0.01) Child Health -0.05 (0.11) -0.05 (0.11) Father Age (years) 0.07 (0.13) 0.09 (0.13) Father is Employed (# of weeks) -0.01 (0.00) * -0.01 (0.00) * Father experienced Economic Hardship 0.33(0.15) * 0.29 (0.16) < 100% of FPL (reference) -- -- 100 - 299% of FPL -0.16 (0.19) -0.18 (0.20) [greater than or equal to] 0.18 (0.27) 0.21 (0.27) 300% of FPL # of Non-biological Children Residing with Father 0.29 (0.16) * 0.30 (0.16) # of Biological Children Residing with Father 0.16 (0.09) 0.14 (0.09) # of Biological Children Not Residing with Father 0.11 (0.12) 0.10 (0.12) Father has Private Insurance -0.22 (0.24) -0.21 (0.23) Father has Public Insurance -0.05 (0.21) -0.05 (0.21) Father has No Insurance -- -- (reference) Constant -- -- Drugs Model 1 Model 2 B (SE) B (SE) Relationship Status Married (reference) -- -- Cohabitation 0.28 (0.40) 0.27 (0.40) Single -0.05 (0.41) -0.14 (0.41) Fathering Practices Father Involvement -- -0.05 (0.05) Role Strain -- 0.40 (0.16) ** Co-parenting Support -- -0.17 (0.26) Covariates Child Age (months) -0.00 (0.01) -0.00 (0.01) Child Health -0.08 (0.12) -0.10 (0.12) Father Age (years) 0.19 (0.15) 0.17 (0.14) Father is Employed (# of weeks) -0.01 (0.00) * -0.01 (0.00) * Father experienced Economic Hardship 0.69 (0.17) *** 0.65 (0.18) *** < 100% of FPL (reference) -- -- 100 - 299% of FPL 0.20 (0.22) 0.17 (0.23) [greater than or equal to] 0.47 (0.28) 0.47 (0.29) 300% of FPL # of Non-biological Children Residing with Father 0.01 (0.14) -0.01 (0.14) # of Biological Children Residing with Father -0.02 (0.11) -0.04 (0.11) # of Biological Children Not Residing with Father -0.09 (0.15) -0.10 (0.16) Father has Private Insurance 0.22 (0.23) 0.21 (0.24) Father has Public Insurance 0.30 (0.21) 0.25 (0.22) Father has No Insurance -- -- (reference) Constant -- -- Note: FPL = Federal Poverty Line. *** p<.001; ** p<.01; * p<.05. Table 4 Cross-lagged Regression Estimates for Models Predicting Fathering Practices and Father Health Model 1 Father Depression (t4) Involvement (t4) B (SE) B (SE) Fathering Practices Father Involvement (t2) 0.35 (0.02) *** -0.06 (0.03) * Role Strain (t2) -- -- Co-parenting Support (t2) -- -- Health and Risk Taking Behaviors Poor Health (t2) -- -- Depression (t2) -0.30 (0.13) * 1.81 (0.16) *** Drugs (t2) -- -- Constant 2.11 (0.28) *** -2.12 (0.58) *** Model 2 Father Poor Health (t4) Involvement (t4) B (SE) B (SE) Fathering Practices Father Involvement (t2) 0.35 (0.01) *** -0.00 (0.01) Role Strain (t2) -- -- Co-parenting Support (t2) -- -- Health and Risk Taking Behaviors Poor Health (t2) -0.03 (0.04) 0.37 (0.02) *** Depression (t2) -- -- Drugs (t2) -- -- Constant 2.12 (0.28) *** 0.97 (0.13) *** Model 3 Role Strain 0 Depression (t4) B (SE) B (SE) Fathering Practices Father Involvement (t2) -- -- Role Strain (t2) 0.54 (0.02) *** 0.46 (0.09) *** Co-parenting Support (t2) -- -- Health and Risk Taking Behaviors Poor Health (t2) -- -- Depression (t2) 0.11 (0.05) * 1.79 (0.17) *** Drugs (t2) -- -- Constant 0.91 (0.13) *** -3.32 (0.56) *** Model 4 Role Strain (t4) Drugs (t4) B (SE) B (SE) Fathering Practices Father Involvement (t2) -- -- Role Strain (t2) 0.54 (0.02) *** 0.16 (0.10) Co-parenting Support (t2) -- -- Health and Risk Taking Behaviors Poor Health (t2) -- -- Depression (t2) -- -- Drugs (t2) 0.10 (0.06) 1.80 (0.15) *** Constant 0.97 (0.13) *** -2.53 (0.57) *** Model 5 Co-parenting Depression (t4) Support (t4) B (SE) B (SE) Fathering Practices Father Involvement (t2) -- -- Role Strain (t2) -- -- Co-parenting Support (t2) 0.33 (0.03) *** -0.76 (0.30) * Health and Risk Taking Behaviors Poor Health (t2) -- -- Depression (t2) -0.07 (0.03) ** 1.73 (0.15) *** Drugs (t2) -- -- Constant 1.81 (0.10) *** -0.21 (1.12) Model 6 Co-parenting Poor Health (t4) Support (t4) B (SE) B (SE) Fathering Practices Father Involvement (t2) -- -- Role Strain (t2) -- -- Co-parenting Support (t2) 0.34 (0.03) *** -0.22 (0.06) ** Health and Risk Taking Behaviors Poor Health (t2) -0.01 (0.01) 0.36 (0.02) *** Depression (t2) -- -- Drugs (t2) -- -- Constant 1.81 (0.10) *** 1.59 (0.20) *** Note: All models include relationship status variables and the covariates listed in Table 3. (t4) = Time 4; (t2) = Time 2. *** p<.001; ** p<.01; * p<.05.
|Printer friendly Cite/link Email Feedback|
|Author:||Hernandez, Daphne C.|
|Date:||Sep 22, 2012|
|Previous Article:||Patterns of father self-evaluations among Mexican and European American men and links to adolescent adjustment.|
|Next Article:||Understanding fatherhood and father involvement in South Africa: insights from surveys and population cohorts.|