Disentangling Ethnicity and Context as Predictors of Parenting Within Rural African American Families

Abstract

This study will address the initial question: Are there ethnic differences in parenting that remain when contextual variables are controlled and are related to culture, focusing on two samples of rural African American families. This study is part of a series of coordinated studies presented in this special issue (Le et al., 2008). Specific attention was given to ascertaining whether these differences were explained by contextual factors, such as socioeconomic factors (i.e., parents’ educational level, homeowner-ship, family structure, and number of children in the household). Finally, this study examined whether any differences in parenting (warmth, monitoring, communication) were attributed to cultural factors, after controlling for the contributions of contextual factors. Results indicate that parental education, family size, negative life events, racial discrimination, neighborhood characteristics, and religiosity were significantly associated with various domains of parenting among rural African Americans.

As part of a series of coordinated studies presented in this special issue (Le et al., 2008), this study will address the initial research question: Are there ethnic differences in parenting that remain when contextual variables are controlled and are related to culture, focusing on rural African American families.

METHOD

Participants

Two datasets were used to examine patterns of parenting processes among African Americans residing in rural communities and small towns. One dataset included 121 families who participated in a longitudinal study of the Rural Georgia Single-Parent Family Study that examined the links among family processes, parenting, and psychosocial competence for children living in economically stressed families. The second dataset included 897 African American families who completed three waves of data as participants in the Family and Community Health Study (FACHS). The FACHS project focuses on ways in which ecological contexts, such as family processes, peer influences, and community context combine with racial socialization and discrimination forecast rural African American youths’ development and adjustment. Families were recruited from neighborhoods in Iowa and Georgia that varied on demographic characteristics.

Rural African American Single-Mother Study

African American single-mother headed families with a 7- to 16-year-old first-born child participated in the larger study (152 families). Because of the age gap in the participating children, only families that included a first-born child who was 9–14 years at the first wave of data collection were considered for the present analyses. Thus, data from 121 families who participated across all three waves of data collection were included in the present study. The children’s mean age was 11.17 years (SD =1.44) at the first wave, 11.83 years (SD =1.51) at the second wave, and 12.97 years (SD =1.46) at the third wave. The children, 60 boys and 61 girls, and their mothers completed questions between September 1997 and December 1999, and follow-up assessments were conducted approximately 12 months later.

Family and Community Health Study

FACHS was designed to identify neighborhood and family processes that contribute to African American children’s development in families living in a wide variety of community settings outside the inner city core. Each family included a child who was 10 and 11 years of age at the time of recruitment. Interviews were conducted with the target child, his/her primary caregiver, and a secondary caregiver when one was present in the home. A primary caregiver was defined as a person living in the same household as the target child and who was responsible for the majority of the child’s care. Families who completed all data collection for all three waves were used for the present analyses. Most (84%) of the primary caregivers were the target child’s biological mother, with a mean age of 37.1 years. An average of 4.5 children lived in their homes. Education ranged from less than high school (19%) to advanced graduate degrees (3%). Seventy-one percent were employed full or part-time, and slightly more than half of the families also had a secondary care-giver. Thirty-five percent were the target child’s biological father, 19% the child’s stepfather, 12% the child’s grandmother, and the remainder represented a variety of other relationships (e.g., aunt, grandfather). The median income for the study families was $20,803. There was no significant difference in income or education of the primary caregiver between the Iowa and Georgia subsamples.

Procedure

The research protocol included a two-part interview that trained African American field researchers (see Brody et al., 2001) administered during two separate sessions at each wave of data collection. The questionnaires that comprised the interviews assessed parenting processes, community and cultural processes, stressful life events, personal characteristics, coping strategies, social support, participation in religious activities, and psychological functioning. The measures were administered via computer-assisted personal interviews (CAPI), in which questions appeared in sequence on a laptop computer screen that both the interviewer and the participant could see. To eliminate any concerns about participants’ literacy, the interviewer read each question aloud to the participant and entered her response into the computer before proceeding to the next question.

Measures

Parenting Variables

Parental monitoring

The Monitoring and Control Questionnaire (MCQ) was used to assess the extent to which caregivers monitored youth behavior. This 17-item scale was developed for the current studies and based on monitoring measures used by Patterson and Stauthamer-Loeber (1984) and by Steinberg et al. (1991). This measure assesses caregivers’ perception of their knowledge about various aspects of their children’s lives. Items are rated on a 4-point Likert scale ranging from 1 (never) to 4 (always). The internal consistency ranged from .82 to .88 across the various data assessments periods.

Parental warmth was measured using the short form of the Interaction Behavior Questionnaire (IBQ; Prinz et al., 1979) which assesses warmth and support in the parent-child relationship. The alpha coefficients across waves of data collection ranged from .79 to .85.

Family communication

We selected three indicators to assess family communication patterns. General family communication which assessed the extent to which family interaction exchanges were frequent, open, and/or argumentative, consisting of a 7-item scale, in which parents describe the context of conversations with youth, with responses ranging from 1 (“I usually do most of the talking and often I catch myself just telling my child what to do or believe”) to 4 (“We usually talk about it openly and we each share our side of the issue”). Cronbach’s alphas were >.80 across waves and datasets.

Family communication quality

The Discussion Quality Scale (Brody & Flor, 1998) was used to assess caregivers’ and youth’s perception of quality of their communication patterns, including the frequency of conversations, the relative contributions of parent and child to the conversations, and the frequency with which discussions become arguments. Cronbach’s alphas were .84 for caregivers and .79 for youths.

Open communication was assessed using a 9-item sub-scale which included statements to determine the degree to which caregivers express their views clearly and encourage youth to engage in conversations. Cronbach’s alphas for each of the measures ranged from .64 to .90 for caregivers and .58 to .71 for youth. We recognize that a coefficient estimate of .58 is low. Our decision to maintain this measure in our analysis was informed by previous research supporting its theoretical relevance to the issues examined in the current study. In addition, while much emphasis has been placed on the importance of yielding measures with high reliability, there is little guidance on what constitutes acceptable or insufficient reliability (Helms et al., 2006; Peterson, 1994).

Contextual Variables

Owning a home was assessed by one dichotomous item for whether parents owned their own home.

Mother’s and father’s education was assessed by an item asking what is the highest level of education completed for each parent, which included the following eight options: 1 =no formal schooling, 2 =some elementary school, 3 =finished elementary school, 4 =finished middle school, 5 =finished high school, 6 =some vocational or college training, 7 =finished four-year college degree, and 8 =finished graduate degree.

Single parent household was assessed through an item asking whom the youth lived with that included nine options. The options of “only my mother” and “only my father” were coded as single parent households.

Family size

We assessed family sizebased on the total number of children in the household, derived from items asking parents how many older and younger siblings reside in the household with the target child.

Neighborhood safety

A revised version of the Project on Human Development in Chicago Neighborhoods (PHDCN; see Sampson, Raudenbush, & Earls, 1997) was used to assess caregivers’ concern about how often various criminal acts occur within the community. It includes behaviors such as fighting with weapons, robbery, gang violence, and sexual assault. The Cronbach’s alphas across data collections were consistently above .80.

Neighborhood cohesion

We selected two items from the Social Cohesion and Trust Scale developed by Sampson et al. (1997), to measure neighborhood cohesion. Using a three-point Likert scale, caregivers and youths indicated the extent to which their neighbors do favors for each other, and “When a neighbor is not at home, how often do you and other neighbors watch over their property?” We produced a composite score by summing the two items, which correlated at .71.

Collective socialization

This construct was assessed by a scale adapted from the PHDCN. Respondents rated (1 = very likely, 4 =very unlikely) the extent to which eight statements described conditions in their community. The items focused on whether neighborhood residents would do something about it if children spray-painted graffiti on a building, showed disrespect to an adult, or skipped school, etc. Coefficient alpha for this scale was .90.

Neighborhood social ties

This construct was assessed by a 9-item revised version of the Social Cohesion and Trust Scale developed for the PHDCN (Sampson et al., 1997). The items focused upon the extent to which individuals in the area interact, trust and respect each other, and share values. Both primary caregivers and target children completed these items in reference to their residential neighborhood. The Cronbach’s alpha was consistently above .80.

Negative life events

A 15-item negative stressful life events measure was adapted from the Iowa Youth and Families Project (Conger & Elder, 1994). Sample questions include, “In the past 12 months, did any close friend or close relative die?” and “In the past 12 months, did you have a son or daughter involved with an unwanted pregnancy?” The Cronbach’s alpha was consistently above .80.

Racial discrimination

At each wave, the target youths completed 13 items from the Schedule of Racist Events (SRE; Landrine & Klonoff, 1996). This instrument, which has strong psychometric properties, has been used extensively in studies with African Americans. The items assess the frequency, ranging from 1 (never) to 4 (several times), with which the respondent has experienced discriminatory events during the past year. The events include racially based slurs and insults, disrespectful treatment from community members, physical threats, and unwarranted suspicion or false accusations from law enforcement officials or other authority figures. The Cronbach’s alpha for the scale exceeded .85 for each of the three waves of data collected.

Cultural Variable

Religiosity

Using a revised version of Jessor’s Value on Religion Scale (Jessor & Jessor, 1977), we developed an index to assess the extent to which the respondents relied on their belief and faith in God as a source of meaning and inspiration. Respondents were asked to rate the importance of believing in God, relying on religious teachings when you have a problem, to be able to turn to prayer when you’re facing a personal problem, and to rely on your religious beliefs as a guide for day-to-day living. The internal consistencies ranged from .78 to .81 over assessments.

RESULTS

Hierarchical regression analyses were conducted to determine the contribution of cultural and contextual factors explaining variations in parenting processes utilized by rural African American families. Consistent with the consensual model described in Le et al. (this issue), contextual variables were entered first, followed by cultural variables. Access to longitudinal data allowed us to examine the continuity or discontinuity of these patterns over time. Findings from the single-parent study are presented first, followed by results from each wave of the FACHS data. A summary of overall patterns in both datasets across waves is provided in the final section of this report. Results are presented in Tables 1–5 and include only those variables that reached significance-level of p <.05. Thus, the beta coefficients included are those that were statistically significant.

Single-Mother Study

Parental monitoring

Results revealed that mother’s education, neighborhood social ties, racial discrimination, and negative life events were significant contextual factors accounting for variation in parental monitoring. An examination of the beta coefficients shows that an increase in parental monitoring among rural African American single mothers was more evident than among those who were educated and resided in supportive neighborhoods, as well as among those who reported less exposure to racial discrimination and among those experiencing fewer negative life events. Further, one cultural factor, in particular, religiosity, emerged as a significant predictor of parental monitoring. Thus, parents who were highly religious were more likely to report increase monitoring. The combination of these variables over time accounted for 8.4% to 12.5% of the variance at Wave 1; 6.5%, 20.2%, and 24.4% of the variance at Wave 2; and 5.5%, 11%, and 19.2% of the variance at Wave 3 (see Table 1).

TABLE 1.

Standardized (β) and Unstandardized Coefficients (B) with Standard Errors (S.E.) for Within Group Analyses for Rural African Americans Single Mother Families with Outcome of Parental Monitoring (Parent’s Report)

Predictor Variables	Parental Monitoring
	β	β	β	β
	B (S.E.)	B (S.E.)	B (S.E.)	B (S.E.)
	Step 1	Step 2	Step 3	Step 4
Wave 1
Mother’s education	.31	.32
	.35 (.11)	.28 (.12)
Racial discrimination		−.23
		−.25 (.11)
Adjusted R2	.084	.125
F	8.62**	4.80*
Df	(1,82)	(1,81)
Wave 2
Neighborhood social ties	.28	.21	.20
	.32 (.12)	.24 (.11)	.23 (.11)
Religiosity			.29
			.25 (.11)
Adjusted R2	.065	.202	.244
F	6.81**	15.02***	5.48*
Df	(1,82)	(1,81)	(1,80)
Wave 3
Mother’s education	.26	.24	.26
	.27 (.11)	.25 (.11)	.27 (.10)
Negative life events		−.26	−.22
		−.29 (.12)	−.24 (.11)
Adjusted R2	.055	.110	.192
F	5.81**	6.10**	9.23**
Df	(1,82)	(1,81)	(1,80)

Note: The models include only standardized betas that are significant at p <.05. Significance levels for the final model are:

^*p <.05,

^**p <.01,

^***p <.001.

Parental warmth

Findings in Table 2 indicate that only one contextual factor emerged as significant predictors of parental warmth, namely negative life events. Thus, increased exposure to negative life events compromised parents’ expression of warmth toward their child. One cultural factor accounted for variation in this parenting process, namely, religiosity (β =.23 for Wave 2). Accordingly, parents who were more religious were more likely to report displaying warmth toward their children. The explanatory contribution of these two factors, negative life events and religiosity, ranged from 3.8% to 17.3% of the variance expressions of warmth among rural African American single-mothers. None of the other variables emerged as significant predictors and no significant predictors emerged for Wave 1.

TABLE 2.

Standardized (β) and Unstandardized Coefficients (B) with Standard Errors (S.E.) for Within Group Analyses for Rural African American Single-Mothers’ Dataset with outcome of Parental Warmth (Parent’s Report)

Predictor Variables	Parental Warmth
	β	β
	B (S.E.)	B (S.E.)
	Step 1	Step 2
Wave 2
Religiosity	.267
	.234 (.082)
Adjusted R2	.173
F	19.47***
Df	(1,87)
Wave 3
Negative life events	−.211	−.175
	−.214 (.083)	−.178 (.082)
Adjusted R2	.038	.094
F	6.594**	9.815**
Df	(1,142)	(1,141)

Note: The models include only standardized betas that are significant at p <.05. Significance levels for the final model are:

^*p <.05,

^**p <.01,

^***p <.001.

Family communication

Several factors emerged as significant predictors of communication patterns in rural African American single mother headed families (see Table 3). In terms of contextual factors, racial discrimination was the only significant predictor remaining in the model, accounting for variability in the frequency of communicative interactions in families. The direction of the beta coefficient suggests that parental experiences of racial discrimination facilitated more frequent conversations in these families (β =.21, p <.05). The association between racial discrimination and frequency of family communication was consistent across several waves, with βetas ranging from .20 to .36. Racial discrimination was also associated with compromised family discussion quality, such that parents who reported experiencing increased incidence of racial discriminatory were more likely to engage in argumentative exchanges with their children. This association, however, was only significant at one wave (Wave 4).

TABLE 3.

Standardized (β) and Unstandardized Coefficients (B) with Standard Errors (S.E.) for Within Group Analyses for Rural African American Single-Mothers’ Dataset with Outcome of Family Communication (Parent’s Report)

Predictor Variables	Family Communication
	β	β	β	β
	B (S.E.)	B (S.E.)	B (S.E.)	B (S.E.)
	Step 1	Step 2	Step 3	Step 4
Wave 1—open
Religiosity	.266
	.230 (.096)
Adjusted R2	.105
F	10.736**
Df	(1,82)
Wave 2—open
Religiosity	.313
	.299 (.097)
Adjusted R2	.087
F	9.425**
Df	(1,87)
Wave 1—frequency
Mom’s report religiosity	.326	.285
	.345 (.097)	.301(.098)
Child’s report religiosity		.193
		.213 (.102)
Adjusted R2	.098	.126
F	12.592***	4.366*
Df	(1,106)	(1,105)
Wave 2—frequency
Racial discrimination	.211	.248
	.191 (.095)	.224 (.085)
Mom’s report religiosity		.456
		.423 (.087)
Adjusted R2	.034	.233
F	4.058*	23.668***
Df	(1,87)	(1,86)
Wave 3—frequency
Racial discrimination	.326	.285
	.345 (.097)	.301(.098)
Child’s report religiosity		.193
		.213 (.102)
Adjusted R2	.098	.126
F	12.592***	4.366*
Df	(1,106)	(1,105)
Wave 4—rguing
Racial discrimination	.321
	.340 (.084)
Adjusted R2	.097
F	16/355***
Df	(1,142)

Note: The models include only standardized betas that are significant at p <.05. Significance levels for the final model are:

^*p <.05,

^**p <.01,

^***p <.001.

We also examined the contribution of cultural factors in predicting family communication patterns and found that only one cultural related variable remained in the model, mother’s and child’s report of religiosity, each associated with more open and frequent conversations in rural single-mother families. Religiosity was a consistent predictor across each wave. Further, racial discrimination remained in the final models at Waves 2, 3, and 4, predicting communication frequency and family discussion quality, more conversation and more arguments (see Table 3). None of the other demographic or neighborhood contextual factors emerged as significant predictors. The combination of the variables remaining in the final models, however, predicted from 8.7% to 23.3% of the variance in family communication patterns within rural African American single-mother households. Taken together, these findings suggest the relative importance of racial discrimination and religiosity in predicting multiple domains of parenting among rural African American single-mothers. The extent to which these patterns can be replicated across rural African American families of varying family compositions will be presented in the following section.

Family and Community Health Study (FACHS)

Parental monitoring

Several contextual factors emerged as significant predictors of parental monitoring (see Table 4). In terms of demographic factors, home ownership and family size were positively associated with parental monitoring. Collective socialization, residing in a community where neighbors monitor each other’s children, as well as residing in an unsafe and disorganized community evoked increased parental monitoring (βs =.21, .15 and .12, respectively). Racial discrimination was associated with reduced parental monitoring (β = −.11, p <.05). One cultural contextual variable, religiosity (β =.17) was associated with increased parental monitoring. The combination of these factors accounted for 6.8% of the variance at Wave 1, 4.3% at Wave 2, and 7.6% at Wave 3 in the variability of parental monitoring among rural African American families participating in the FACH study.

TABLE 4.

Standardized (β) and Unstandardized Coefficients (B) with Standard Errors (S.E.) for Within Group Analyses for Rural African American Family and Community Health Study Dataset with Outcome of Parental Monitoring (Parent’s Report)

Predictor Variables	Parental Monitoring
	β	β	β	β
	B (S.E.)	B (S.E.)	B (S.E.)	B (S.E.)
	Step 1	Step 2	Step 3	Step 4
Wave 1
Home ownership	.15	.12	.10	.094
	.17 (.054)	.14 (.54)	.12 (.54)	.111 (.054)
Neighborhood safety		.17	.16	.151
		.171 (.048)	.16 (.05)	.153 (.46)
Racial discrimination			−.114	−.110
			−.111 (.045)	−.107 (.044)
Adjusted R2	.019	.045	.055	.068
F	9.89**	13.44***	6.12*	7.34**
Df	1,461	(1,460)	(1,459)	(1,458)
Wave 2
Neighborhood safety	.137	.123	.121
	.132 (.042)	.119 (.042)	.117 (.042)
Negative life events		−.089	−.091
		−.088 (0.44)	−.090 (.043)
Adjusted R2	.019	.023	.043
F	9.76**	4.05*	11.64***
Df	(1,512)	(1,511)	(1,510)
Wave 3
Family size	.098	.083	.074
	.106 (.050)	.091 (.049)	.081 (.048)
Collective socialization		.211	.201
		.229 (.049)	.218 (.048)
Religiosity			.168
			.169 (.045)
Adjusted R2	.007	.050	.076
F	4.54*	21.93***	14.34***
Df	(1,469)	(1,468)	(1,467)

Note: The models include only standardized betas that are significant at p <.05. Significance levels for the final model are:

^*p <.05,

^**p <.01,

^***p <.001.

Warmth

Factors predicting parental warmth included one demographic factor, family size, and several contextual factors, namely, collective socialization, neighborhood cohesion, neighborhood social ties, neighborhood safety, negative life events, and youth reports of racial discrimination (see Table 5). An examination of the beta coefficients revealed that family size, negative life events, racial discrimination, and lack of social times with neighbors reduced expressions of parental warmth toward their children. Residing in an unsafe neighborhood, and one in which other adults monitor youth behaviors (collective socialization), as well as nested in a closely knitted community (cohesion) were associated with increased expressions of parental warmth. The combination of these variables explained from 3.4% to 10.8% of the variance in parental warmth among rural African Americans.

TABLE 5.

Standardized (β) & Unstandardized Coefficients (B) with Standard Errors (S.E.) for Within Group Analyses for Rural African Americans Family and Community Health Study Dataset with Parental Warmth

Predictor Variables	Parental Warmth
	β	β	β
	B (S.E.)	B (S.E.)	B (S.E.)
	Step 1	Step 2	Step 3
Wave 1 PC-warmth
Family size		−.114	−.105
		−.129 (.052)	−.119 (.052)
Negative life events		.097	−.100
		−.097 (.046)	−.100 (.045)
Adjusted R2	.101	.108	.052
F	6.039**	4.434*	17.662***
Df	(1,461)	(1,460)	(1,459)
Wave 1 C-Warmth
Family size	−.109	−.097	−.100
	−.114 (.049)	−.101 (.048)	−.105 (.048)
Racial discrimination		−.174	−.166
		−.156 (.041)	−.149 (.041)
Collective socialization			.150
			.138 (.041)
Adjusted R2	.010	.038	.058
F	5.531**	14.464**	11.035***
Df	(1,461)	(1,460)	(1,459)
Wave 2 PC-warmth	.149	.141
	.147 (.043)	.140 (.043)
Neighborhood cohesion
Adjusted R2	.020	.045
F	11.559***	14.513***
Df	(1,512)	(1,511)
Wave 2 C-warmth
Collective socialization	.163	.158	.183
	.154 (.041)	.149 (.041)	.173 (.042)
Negative life events		−.093	−.102
		−.090 (.042)	−.099 (.042)
Neighborhood social ties			−.104
			−.100 (.043)
Adjusted R2	.025	.031	.040
F	13.934**	4.564**	5.421***
Df	(1,5121)	(1,511)	(1,510)
Wave 2 C-warmth
Collective socialization	.161
	.163 (.044)
Adjusted R2	.024
F	13.695**
Df	(1,512)
Wave 3 PC-warmth
Neighborhood safety	.170	.165
	.183 (.049)	.177 (.045)
Negative life events	−.098
	−.096 (.045)
Neighborhood social ties
Adjusted R2	.027	.035
F	14.017	4.695
Df	(1,469)	(1,468)
Wave 3 C-warmth
Neighborhood social ties	.178	.156
	.175 (.045)	.154 (.045)
Collective socialization		.132
		.128 (.044)
Adjusted R2	.030	.045
F	15.350	8.358
Df	(1,469)	(1,468)
Wave 3 C-warmth
Neighborhood social ties	.137	.118	.157
	.136 (.046)	.117 (.046)	.156 (.050)
Collective socialization		.116	.130
		.113 (.045)	.128 (.046)
Neighborhood cohesion			−.102
			−.103 (051)
Adjusted R2	.017	.028	.034
F	8.900**	6.278	4.096
Df	(1,468)	(1,467)	(1,466)

Note: The models include only standardized betas that are significant at p <.05. Significance levels for the final model are:

^*p <.05,

^**p <.01,

^***p <.001. PC =Primary caregiver’s report; C =Child’s report.

SUMMARY

In sum, the two datasets of African American parents rearing their children in rural, often resource-scarce communities, allowed us to disentangle the contributions of contextual and cultural factors to explain variability in several parenting outcomes across multiple years of family and child development. Consistent across both datasets and data points is the centrality of specific contextual processes, in particular neighborhood, negative life events, and racial discrimination, in predicting variability in parenting processes for children reared in single-mother households, as well as those reared in two parents families or mother–grandmother dyadic families (FACHS). Neighborhood context, residing in a supportive community, or one that is unsafe and disorganized, were consistently associated with increased parental monitoring and elevated warmth. Racial discrimination and negative life events compromised the use of effective parenting, as they were often associated with less monitoring, reduced warmth, and more frequent and argumentative conversations in families. For single-mothers, education was positively associated with parental monitoring. Family size, however, was an important predictor of parental monitoring and warmth among families participating in the FACH study. Finally, our data consistently support the contribution of religiosity in predicting each domain of parenting among rural African Americans, with similar levels of influence across dataset and across waves.