Journal of Rural Community Psychology, Vol. E2, No. 1, 1999 
 
Parenting Attitudes and Behaviors
in Rural Appalachian Mothers:
A Comparison Study
 
Margaret Fish
Marshall University School of Medicine
 
Cynthia A. Stifter
Pennsylvania State University
 
 
 Abstract
 
There is a general belief, supported by considerable empirical data, that low socioeconomic status (SES) constitutes a risk for less optimal parenting. However, in a cross-SES comparison, economically-disadvantaged rural Appalachian mothers were comparable to well-educated middle-class mothers on maternal self-efficacy, significantly higher on responsive attitude to infant crying, and rated either more optimal or comparable to the middle-class sample on mother-infant interaction rated at two points during the first year.
 
Introduction

It is generally assumed that inadequate income is a risk condition for both parenting and child development. For example, Elder's (1974, 1979) studies of parenting during the depression years and more recent studies of economic hardship in the agricultural midwest (Conger et al., 1992; Lempers, Clark-Lempers, & Simons, 1989) show an association between financial hardship and less nurturant, inconsistent parenting behavior. McLoyd & Wilson (1991) summarize evidence that low socioeconomic status (SES) parents are subject to greater stress, anxiety, and depression than higher SES parents and thereby are at risk for providing less optimal parenting, stating "it is reasonable to conclude on the basis of studies directly linking negative emotional states and punitive, nonsupportive parenting that environmental and psychological distress partially account for well-established social class differences in parenting behavior" (p. 110).
 
Low levels of parental education, which often occur in conjunction with poverty and other risk factors, have also been linked to less favorable social and academic outcomes in several longitudinal studies (Dubow & Luster, 1990; Sameroff, Seifer, Barocas, Zax, & Greenspan, 1987; Werner & Smith, 1992). Of particular interest is Werner & Smith's (1992) finding that more parental education was associated with more positive parent-child interaction during infancy.
 
Beyond the notion that low SES constitutes a general risk factor, two issues merit attention. First, it is important to note that parenting and child development show a wide range of individual difference within low SES families. Thus, while low SES does constitute a risk factor, less optimal parenting and child development is not true of all low SES families. From the work of Werner and Smith (1982, 1992), Rutter (1985, 1987), and Garmezy (1993), it is evident that both risk and protective factors need to be taken into account to explain individual differences. In particular, these studies of resilience in high-risk environments reveal that basic temperamental qualities of children and their close relationships with others function as protective factors in long term adjustment.
 
Second, research which has examined parenting and child development in samples defined as high-risk, based on poverty or low SES, has tended to view all low SES groups as similarly at risk, despite having diverse characteristics. There has been insufficient attention in developmental research to cultural characteristics as potential strengths or protective factors against the stress of poverty.
 
Among low SES families, much of the research concerned with parenting and child development has examined urban minority populations. Probably the least studied group of low SES families are those living in rural areas. Yet recent data from the Children's Defense Fund (1996) show that: all 10 states with the highest rates of child poverty are rural, that 23% of rural children are poor, and that only 1 in 10 poor children fits the common stereotype of a poor child--African-American, urban, with a single mother receiving public assistance. Thus, a large, high risk population--rural, poor, White families--has been inadequately studied.
 
In this investigation, we examine parenting in a rural Appalachian sample. Rural Appalachia has been identified as a region with chronic high poverty and low levels of education (West Virginia Human Resources Association, 1988). West Virginia, the only state entirely within the Appalachian region, has the highest poverty rate in the nation for white children, double the national average (Children's Defense Fund, 1996).
 
A number of authors point out that descriptions of the rural Appalachian culture have often "blamed the victim" for chronic poverty in the region (Beaver, 1980; Lohmann, 1990; Peoples Appalachian Research Collective, 1971), without adequately considering regional economic and social conditions (Duncan, 1992) or cultural and family strengths (Beaver, 1980; Singleton, Ratliff, Carpenter, Davis, & Brunner, 1990).
 
Information on parenting and family characteristics in rural Appalachia is largely descriptive and presents a mixed picture. For example, a panel of native Appalachian social service and health practitioners listed among family strengths perseverance through difficult times, making do with whatever resources are available, pride, and being close-knit, protective, and caring about family and friends (Singleton et al., 1990). Beaver (1980) states, "For Appalachian rural families, familialism is still highly valued and depended upon. Reliance on cooperative networks provides a variety of psychological, social, and economic supports and outlets" (p. 33).
 
On the other hand, frequently quoted observers from the War on Poverty era characterized parenting in rural Appalachia in terms of negative stereotypes:

As children grow, adults largely cease to play with them, supervise them inappropriately, and furnish no positive models for training in setting limits on impulses, for establishing disciplinary controls, or for training in age-appropriate relating to or talking with other adults . . . adults generally notice the misbehavior of children, usually in public, only when the situation is beyond the point at which an adult could have intervened to set the training model for the child's more effective mastering of his feelings, impulses, and relationships. Instead, misbehavior is allowed to come to a head, and physical punishment is meted out swiftly and sternly as a consequence (Looff, 1971, p. 15).

Mountain children are reared impulsively, with relatively little of the conscious training found in middle class families. Discipline is meted out with no concern about how the child will react as a child. In such an adult-centered society, children's feelings are not given great weight. When they misbehave, they are punished--and it is always physical punishment, not just a scolding or being sent to their rooms (Weller, 1965, p. 65).

However, without current, methodologically sound, scientific studies on parent-child interaction and factors influencing individual differences in parenting, evidence is lacking to either verify or refute such statements. The purpose of the present study was to compare low SES rural Appalachian mothers to a previously studied sample of well-educated, middle class mothers in order to empirically determine if differences in parenting could be documented.
 
 
Methods

Participants
 
Rural Appalachian Sample
 
The 94 low SES, rural Appalachian mothers were participants in a longitudinal study of infant and caregiving environment factors related to stability and change in infant temperament from birth through 15 months (Fish, in press). These mothers were recruited from prenatal patients at a community health clinic in a rural West Virginia county. All patients who were not experiencing medically high risk pregnancies, and who planned to remain in the area until their child was 15 months old, were asked to participate in the study. Of the 121 women recruited during the last trimester of pregnancy (69% of those approached), 116 had infants who met study criteria (singleton birth, normal neonate). Over the course of the study, 17 mothers either moved (7) or declined to participate (10) at some point. Data on 5 other infant-mother pairs were excluded due to infant characteristics which might affect early development.
 
The demographic data reflect low socioeconomic status. Almost three quarters of mothers reported that their annual family income was less than $10,000, and 70% received public assistance (either food stamps or Aid to Families with Dependent Children or both). About half of the parents (55% of the mothers and 53% of fathers) were high school graduates; nearly one fourth had a ninth grade education or less. All parents were White.
 
Middle-Class Pennsylvania Sample
 
The Pennsylvania sample consisted of 67 mothers, living in or near a middle-sized university community, who were participants in a longitudinal study of the physiological and behavioral correlates of infant temperament from birth through 18 months (Fish, Stifter, & Belsky, 1991; Stifter & Braungart, 1995). Mothers were recruited on the maternity ward of a community hospital in central Pennsylvania; 33% of those approached agreed to participate. Of 101 mothers recruited, 22 moved or declined to participate at some point, and 12 were not included due to missing data or technical difficulties.
 
Income data were not collected in this sample; however, descriptive statistics on occupation and education indicate that this was a middle class group. Although occupations of both mothers and fathers ranged from professional to service/laborer, in 80% of the families, one or both spouses held a professional or semi-professional position. Further, fathers had completed an average of 4 years of college education and mothers 3 years. Parents were 98.5% White; two parents were African-American.
 
 
Procedures and Measures

Overview

In both samples, mothers completed demographic, personality, and relationship satisfaction questionnaires prior to substantive interaction with the infant, during a prenatal interview for the rural Appalachian sample and neonatally for the Pennsylvania sample. (Because some women in the rural Appalachian sample had low reading levels, for that group, the questionnaires were completed in an interview format with a research assistant reading the items, clarifying them as needed, and recording the subject's responses.) In both samples, mothers were seen with their infants for laboratory assessments, which included mother-infant interaction, twice during the first year of the infant's life (at 4 and 9 months for the rural West Virginia sample and at 5 and 10 months for the Pennsylvania sample). For both groups, the quality of mother-infant interaction was coded for sensitivity and over-controlling/intrusive behavior at each time.
 
Characteristics of the Samples
 
Prenatal/neonatal parent measures. SES/demographic and work variables assessed included: mother's and father's ages, years of education, and occupations; maternal parity; and marital status and number of years married. We also asked mothers if they currently smoked and if they planned to breast or bottle feed their infant.
 
Mother personality indicators assessed the broad traits of positive affectivity (extraversion) and negative affectivity (neuroticism) with well standardized scales. These traits have been related to individual differences in parenting in a number of studies (e.g., Belsky, Fish, & Isabella, 1991; Fish, Stifter, & Belsky, 1993; Mangelsdorf, Gunnar, Kestenbaum, Lang, & Andreas, 1990), and psychological distress is often associated with low SES (McLoyd & Wilson, 1991). Measures used were: the hostility, anxiety, and depression scales from the neuroticism dimension and the warmth and positive emotions scales from the extraversion dimension of the NEO-AC Personality Inventory (Costa & McCrae, 1985; McCrae & Costa, 1983).
 
Marital or relationship satisfaction (those women living with a partner were treated in analyses the same as those who were married) was measured with the Braiker and Kelley (1979) index of relationship sentiments and activities. This is a 25-item measure of two positive (love and maintenance) and two negative (ambivalence and conflict) aspects of relationships. These measures have previously predicted differences in mother interactional behavior (Fish et al., 1993).
 
Measures of Parenting
 
Attitudes about parenting. Maternal self-efficacy was measured at 4-5 months with an 18-item scale covering concepts such as feelings of competence as a parent, support sought and experienced in the parenting role, and ability to meet this infant's individual needs (Fish, et al., 1991). Internal consistency was .75 in the rural Appalachian sample and .82 in the suburban Pennsylvania sample. Mothers also completed the Crockenberg and Smith (1982) 10-item scale of parental responsiveness, which has been shown to predict subsequent infant irritability and observed mother responsiveness (Crockenberg & Smith, 1982; Crockenberg & McCluskey, 1986). A 5-item scale dealing specifically with responding to crying was used in the comparison. Internal consistency (coefficient alpha) for this scale was .67 for the rural Appalachian sample and .76 for the Pennsylvania sample.
 
Parenting behavior. Parenting behavior was measured by rating the quality of mother interaction from videotapes of infants interacting with their mothers in an unstructured free play situation in project playrooms at two points in the first year - 4 and 9 months in the rural Appalachian sample and 5 and 10 months in the Pennsylvania sample. For both groups, mothers were seated on a blanket on the floor, provided with a basket of toys, and asked to play with their infant "the way you would at home."
 
Behaviors rated from the videotapes were maternal sensitivity and over-controlling/intrusive behavior.  Differences in rated sensitivity and appropriate responsiveness of mother interactional behavior have been linked to infant-mother attachment, theorized to be an indicator of the quality of the infant-parent relationship (e.g., Ainsworth, Blehar, Water, & Wall, 1978; Belsky, Rovine, & Taylor, 1984; Grossmann, Grossmann, Spangler, Suess, & Unzner, 1985; Smith & Pederson, 1988) and to children's early cognitive and verbal abilities (Bornstein & Tamis-LeMonda, 1989; Hart & Risley, 1992; Olson, Bates, & Bayles, 1984).
 
In the free play situation, behavioral examples of sensitivity included contingent responses to the infant's affect and actions, positive verbal encouragement of the infant's efforts, providing assistance to allow the infant to achieve his/her goals, and responding to the infant's focus by either sustaining interest in one toy/activity or appropriately changing to another. Mother behavior rated as over-controlling/intrusive was overly stimulating, poorly-timed, out of synchrony with the baby's pace and signals, and often appeared to address the mother's agenda at the expense of the infant's.

For each interval of free play (15-seconds in the rural Appalachian sample and 30-seconds in the Pennsylvania sample), both sensitivity and over-controlling/intrusive behavior were rated on 4-point scales. Reliabilities (kappa) for sensitivity at each age were .62 and .66 for the rural Appalachian sample and .61 and .68 for the Pennsylvania sample. Reliabilities for over-control/intrusive behavior were .75 and .73 for the West Virginia group and .71 and .72 for the Pennsylvania sample. At each age, scores across all intervals were summed to create total scores for sensitivity and over-control.
 
Data Transformation
 
Differences noted above in length of free play sessions and in the intervals coded necessitated mathematical transformations before comparisons could be made. The duration of the free play sessions ranged from 5 minutes (both times for the Pennsylvania mothers and infants) to 8 minutes (second time point for the rural West Virginia mothers and infants). Coding intervals were 15 seconds in one sample and 30 seconds in the other. Appropriate formulas were constructed, using proportions, to make the totals for all free play codes equal to what would have been obtained for a 7-minute sample rated each 15-seconds, which was the configuration for the rural Appalachian data at 4 months. Thus, for example, the 5-month Pennsylvania data were multiplied by a factor of 2.8. In other words, a Pennsylvania mother receiving a sensitivity score of 20 scored for 10 30-second intervals would have received a sensitivity score of 56 had she been rated for 28 15-second intervals.
 
 

Results
 
Our first analyses compared the demographic, personality, relationship satisfaction, and maternal attitude measures with either t-tests or Chi Square analyses. Results are reported in Tables 1, 2, and 3.
 
Sample Characteristics
 
Group comparisons of rural West Virginia and suburban Pennsylvania mothers on demographic data are presented in Table 1. In addition to highly significant differences in education, it also can be seen that the rural Appalachian women were younger1 (mean age 23.21 vs. 29.45), but more likely to have other children. The Pennsylvania mothers were somewhat more likely to be married or living with a partner and considerably more likely to be employed outside the home.
 

Table 1
 
Sample Comparisons on Demographics
 
 

Variable

 

WV Mean (SD)

 

PA Mean (SD)

 

t-test

 

mother age

 

23.21 (4.88)

 

29.51 (4.59)

 

8.27****

 

yrs education

 

10.95 (1.72)

 

15.02 (2.11)

 

12.87****

 

 

 

WV Percent

 

PA Percent

 

Chi Square

 

multiparous

 

71.3%

 

55.2%

 

4.40*

 

spouse/partner

 

77.7%

 

89.6%

 

3.85*

 

working mother

 

24.5%

 

54.0%

 

14.19***

 
* = p < .05, *** = p < .001, *** = p < .0001

 
Comparisons of personality/emotional outlook and relationship appraisals are found in Table 2. On personality traits, presumed to be influenced by developmental history, and in terms of satisfaction with spousal/partner relationships, all of the differences found reflect more negative views on the part of the rural Appalachian women. In particular, they were significantly higher on neuroticism (the anxiety, depression, and hostility scales) and lower on extraversion (warmth and positive emotions) and reported significantly more conflict and ambivalence regarding their relationship with their partner than women in the middle class Pennsylvania sample. These results support the prevalent view in the literature that economically-disadvantaged mothers have a more negative outlook than higher SES mothers.
 

Table 2
 
Comparisons on Personality, Relationship Satisfaction
 
  

Variable

  

WV Mean (SD)

 

PA Mean (SD)

 

t-test

 

neuroticism

 

74.71 (13.30)

 

63.77 (12.40) 

 

5.27****

 

extraversion

 

59.77 (5.97)

 

64.23 (7.47)

 

4.03****

 

relationship positive

 

111.66 (12.92)

 

112.03 (15.61)

 

ns

 

relationship negative

 

35.33 (15.82)

 

30.06 (10.33)

 

2.33*

 
* = p < .05, ***** = p < .0001
 


 
 
Maternal Attitudes and Behavior
 
Comparing measures which reflected maternal attitudes and behavior by SES group produced mixed results. As can be seen in Table 3, the West Virginia mothers were much more likely to report that they smoked and much less likely to say that they planned to breast feed their infant. However, on the response to crying scale, the rural Appalachian mothers' reported that they would respond to a crying baby significantly sooner, and there were no differences between the groups on maternal self-efficacy.
 

Table 3
 
Comparisons on Maternal Attitudes, Behavior
 
  

Variable

 

WV Mean (SD)

 

PA Mean (SD)

 

t-test

 

responsive attitude toward crying

 

17.07 (2.23)

 

15.23 (2.89)

 

4.36****

 

maternal self-efficacy

 

82.55 (10.93)

 

80.43 (11.84)

 

ns

 

 

 

WV percentage

 

PA percentage

 

Chi Square

 

mother smokes

 

34.0%

 

4.5%

 

20.10****

 

breast feeding

 

43.6%

 

83.6%

 

26.09****

 
**** = p < .0001

 
To compare mother interactional behavior rated in the free play situations, we carried out two repeated measures analyses of variance. The repeated measures ANOVA on sensitivity with group as a between-subjects factor showed significant effects of group [F(1,159) = 48.91, p < .0001], time [F(1,159) = 6.54, p < .05], and the group by time interaction [F(1,159) = 60.48, p < .0001]. The repeated measures ANOVA on over-controlling/intrusive behavior with group as a between-subjects factor also showed significant effects of group [F(1,159) = 26.95, p < .0001], time [F(1,159) = 10.25, p < .01], and the group by time interaction [F(1,159) = 41.68, p < .0001].
 
Planned follow-up comparisons revealed that at the earlier time, low SES rural Appalachian mothers were significantly more sensitive (t = 8.72, 99.40 df, p < .0001, M's 51.82 and 34.77, SD's 8.55 and 14.29) and less over-controlling/intrusive (t = 6.76, 84.41 df, p < .0001, M's 4.35 and 15.88, SD's 5.77 and 13.08), whereas at the later time there were no differences between the groups.2 Within groups, across time the rural Appalachian mothers significantly decreased in sensitivity (t = 4.97, 93 df, p < .0001) and increased in over-controlling/intrusive behavior (t = 10.40, 93 df, p < .0001). In contrast, the middle-class Pennsylvania mothers increased in sensitivity (t = 5.57, 66 df, p < .0001) and did not change significantly in over-controlling/ intrusive behavior.

The results of these comparisons are shown in Figure 1.


 
To summarize: the rural Appalachian mothers were clearly distinguished by poverty and low levels of education and also by more negative personality traits and more feelings of conflict and ambivalence about their intimate relationships. In keeping with other studies linking education level with behavior, rural West Virginia mothers were less likely to breast feed and more likely to smoke, even during pregnancy. Overall, we found little evidence that low SES and a negative outlook carried over to other maternal attitudes we measured or to interactional behavior. Low SES rural Appalachian women did not have lower maternal self-efficacy than the middle-class Pennsylvania mothers. Further, they were higher in responsive attitude toward crying and were rated as more sensitive and less over-controlling/intrusive with younger infants and as having similar behavior with older infants to the middle class group. However, across the first year, rural Appalachian mothers, as a group, declined in sensitivity and increased in intrusiveness, whereas middle-class Pennsylvania mothers increased in sensitivity and maintained the same level of intrusiveness.
 
 

Discussion
 
 
The results of this study suggest we should not assume low SES has a negative impact on parenting without considering the broader social context (Bronfenbrenner, 1979). Despite more negative responses to personality and marital satisfaction measures, and despite having much lower levels of income and education, the rural Appalachian mothers did not show less optimal maternal behavior than the middle class mothers, and in fact, were rated as significantly more sensitive and less intrusive with younger infants.
 
Consistent with the large body of literature on the effects of living in poverty, the low income West Virginia mothers did have more negative feelings about their lives and relationships than well educated, middle-class mothers. But, unlike the families studied by Elder (1974, 1979), Conger et al. (1992), and Lempers et al., we did not find negative parenting, at least as compared to our middle class sample. Rural Appalachian mothers were comparable to well-educated middle-class mothers on maternal self-efficacy, significantly higher on responsive attitude to infant crying, and rated either more optimal or comparable to the middle-class sample on mother-infant interaction.
 
At least two explanations for these findings seem plausible. First, unlike families who have experienced a change in SES due to a depression or other economic reversal, poverty is not new to rural Appalachian families (Peoples Appalachian Research Collective, 1971). Further, many families are poor. In Lincoln County, West Virginia, the study site, 45% of children live in poverty (1997, West Virginia Kids Count Fund). Perhaps chronic and widespread poverty has less effect on parenting than sudden reversals, even though it may relate to more negative appraisals of self and partner.
 
In addition, since low SES can hardly be viewed as advantageous, these results suggest to us that some characteristics of the rural Appalachian mothers, either high social support, which has been documented in this sample (Fish, in press), or other cultural characteristics which we did not measure, act as protective factors against the stresses of low SES--"buffering" parenting behavior.
 
Although we did not measure general social support in the middle-class Pennsylvania sample, we think it likely that, as residents of a university community with mostly professional or semi-professional occupations, many of them are geographically mobile and lack the close and frequent contact with parents and extended family reported by the rural Appalachian women. In the West Virginia group, 91% had multiple family members living nearby, and the majority visit parents, partner's parents, and other family members once a week or more. We also observed informally that motherhood appears to be a very central role for women in the rural Appalachian culture and one for which even young mothers are well prepared in the extended family.
 
Our comparisons of sample characteristics revealed only two differences which might be interpreted as "advantages" for the rural West Virginia mothers. First, a higher proportion of the rural Appalachian group were multiparous, and therefore, experienced mothers. However, when we conducted a post hoc ANOVA with 4-month sensitivity as the dependent variable and group and parity as factors, there was no main effect of parity and no interaction with group. Thus, it seems unlikely that more experience with infants explains higher sensitivity at that age in the rural sample. The West Virginia women were also less likely to be employed outside the home, either before or after the baby was born, and some research (Stifter, Coulehan, & Fish, 1993) suggests that separation anxiety in working mothers may negatively affect interaction with infants. A post hoc comparison did show that in the Pennsylvania sample only, working mothers tended to be more over-controlling/intrusive at the later observation (t = 1.92, p = .06; M 14.66, sd 10.66) than mothers not employed outside the home (M 10.04, SD 7.90). However, working and nonworking mothers did not differ on interactional behaviors in the rural Appalachian group.
 
It should be noted that our findings are not inconsistent with the observations of Weller (1965) and Looff (1971) that rural Appalachian parents are infant-centered, but become less involved and more authoritarian as children grow. Specifically, we found that rural Appalachian mothers were more sensitive and less over-controlling/intrusive with 4-month old infants than 9-month-olds. However, additional study, from a cultural-ecological perspective (Ogbu, 1981) is needed of the interactions of parents in this population with toddlers and older children. The conclusion of Weller and Looff that the child-centered attitudes and behavior of rural Appalachian families are confined to the period before children begin to develop motoric skills and seek autonomy, has not been tested empirically, nor examined in sociocultural context (Chao, 1994).
 
This investigation has likely raised more questions than it has answered. In general, the lack of longitudinal study of parent-child interaction in differing SES and cultural groups limits understanding of how characteristics of the child, including age/stage of development, interact with other factors such as cultural group and specifics of socioeconomic status. For example, how does SES interact with child age and sex? Is parental stress higher or lower in a region of relatively widespread and chronic poverty such as rural Appalachia than it is for higher SES parents who unexpectedly become unemployed? These and other issues should be addressed in additional study of low SES rural families.
 
 
References
 
 
Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment. Hillsdale, NJ: Erlbaum.
 
Beaver, P. D. (1980). The mountain family: Kinship and society. In The Appalachian woman: Images and essence. (Abstracts of forums sponsored by the Council on Appalachian Women Inc.).
 
Belsky, J., Fish, M., & Isabella, R. (1991). Continuity and discontinuity in infant negative and positive emotionality: Family antecedents and attachment consequences. Developmental Psychology, 27, 421-431.
 
Belsky, J., Rovine, M., & Taylor, D. G. (1984). The Pennsylvania Infant and Family Development Project, III: The origins of individual differences in infant-mother attachment: Maternal and infant contributions. Child Development, 55, 718-728.
 
Bornstein, M. H., & Tamis-LeMonda, C. S. (1989). Maternal responsiveness and cognitive development in children. In M. H. Bornstein (Ed.), Maternal responsiveness: Characteristics and consequences: New directions for child development (pp. 49-61). San Francisco: Jossey-Bass.
 
Braiker, H., & Kelley, H. (1979). Conflict in the development of close relationships. In R. Burgess & T. Huston (Eds.), Social exchange and developing relationships (pp. 79-102). New York: Academic.
 
Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.
 
Chao, R. K. (1994). Beyond parent control and authoritarian parenting style: Understanding Chinese parenting through the cultural notion of training. Child Development, 65, 1111-1119.
 
Children's Defense Fund. (1996). The state of America's children: Yearbook 1996. Washington, D.C.
 
Conger, R. D., Conger, K. J., Elder, G. H., Jr., Lorenz, F. O., Simons, R. L., & Whitbeck, L. B. (1992). A family process model of economic hardship and adjustment of early adolescent boys. Child Development, 63, 526-541.
 
Costa, P., & McCrae, R. (1985). The NEO Personality Inventory. Odessa, Florida: Psychological Assessment Resources, Inc.
 
Crockenberg, S., & McCluskey, K. (1986). Change in maternal behavior during the baby's first year of life. Child Development, 57, 746-753.
 
Crockenberg, S. B., & Smith, P. (1982). Antecedents of mother-infant interaction and infant irritability in the first three months of life. Infant Behavior and Development, 5, 105-119.
 
Dubow, E. F., & Luster, T. (1990). Adjustment of children born to teenage mothers: The contribution of risk and protective factors. Journal of Marriage and the Family, 52, 393-404.
 
Duncan, Cynthia M. (1992). Persistent poverty in Appalachia: Scarce work and rigid stratification. In C. M. Duncan (Ed.), Rural poverty in America (pp. 111-133). New York: Auburn House.
 
Elder, G. H., Jr. (1974). Children of the Great Depression. Chicago: University of Chicago Press.
 
Elder, G. H., Jr. (1979). Historical patterns in life patterns and personality. In P. B. Baltes & O. G, Brim, Jr. (Eds.), Life-span development and behavior, (vol. 2). New York: Academic Press.
 
Fish, M. (in press). Negative emotionality and positive/social behavior in rural Appalachian infants: Prediction from caregiver and infant characteristics. Infant Behavior and Development.
 
Fish, M., Stifter, C. A., & Belsky, J. (1991). Conditions of continuity and discontinuity in infant negative emotionality: Newborn to 5 months. Child Development, 62, 1525-1537.
 
Fish, M., Stifter, C. A., & Belsky, J. (1993). Early patterns of mother-infant dyadic interaction: Infant, mother, and family demographic antecedents. Infant Behavior and Development, 16, 1-18.
 
Garmezy, N. (1993). Children in poverty: Resilience despite risk. Psychiatry, 56, 127-136.
 
Grossmann, K., Grossmann, K. E., Spangler, G., Suess, G., & Unzner, L. (1985). Maternal sensitivity and newborns' orientation responses as related to quality of attachment in northern Germany. In I. Bretherton & E. Waters (Eds.), Growing points of attachment theory and research (pp. 233-256). Monographs of the Society for Research in Child Development, 50 (serial no. 209).
 
Hart, B., & Risley, T. R. (1992). American parenting of language-learning children: Persisting differences in family-child interactions observed in natural home environments. Developmental Psychology, 28, 1096-1105.
 
Lempers, J. D., Clark-Lempers, D., & Simons, R. L. (1989). Economic hardship, parenting, and distress in adolescence. Child Development, 60, 25-39.
 
Lohmann, R. A. (1990). Four perspectives on Appalachian culture and poverty. Journal of the Appalachian Studies Association, 2, 76-91.
 
Looff, D. H. (1971). Appalachia's children: The challenge of mental health. Lexington, KY: The University Press of Kentucky.
 
Mangelsdorf, S., Gunnar, M. Kestenbaum, R., Lang, S., & Andreas, D. (1990). Infant proneness-to-distress temperament, maternal personality, and infant-mother attachment: Associations and goodness of fit. Child Development, 61, 820-831.
 
McLoyd, V. C., & Wilson, L. (1991). The strain of living poor: Parenting, social support, and child mental health. In A. C. Huston (Ed.), Children in poverty: Child development and public policy (pp. 105-135). Cambridge: Cambridge University Press.
 
McCrae, R. R., & Costa, P. T. (1983). Joint factors in self-reports and ratings: Neuroticism, extraversion, and openness to experience. Personality and Individual Differences, 4, 245-255.
 
Ogbu, J. U. (1981). Origins of human competence: A cultural-ecological perspective. Child Development, 52, 413-429.
 
Olson, S. L., Bates, J. E., & Bayles, K. (1984). Mother-infant interaction and the development of individual differences in children's cognitive competence. Developmental Psychology, 20, 166-179.
 
Peoples Appalachian Research Collective. (1971). Appalachia's people, problems, alternatives: An introductory social science reader. Morgantown, WV.
 
Rutter, M., 1985. Resilience in the face of adversity: Protective factors and resistance to psychiatric disorder. British Journal of Psychiatry, 147, 598-611.
 
Rutter, M., 1987. Psychological resilience and protective mechanisms. American Journal of Orthopsychiatry, 57, 316-331.
 
Sameroff, A. J., Seifer, R., Barocas, R., Zax, M., & Greenspan, S. (1987). Intelligence quotient scores of 4-year-old children: Social-environmental risk factors. Pediatrics, 79, 343-350.
 
Singleton, M., Ratliff, K., Carpenter, D., Davis, R. R., & Brunner, W. (1990). Family strengths as viewed by native Appalachians practicing in the helping systems. Proceedings of the UK conference on Appalachia. Lexington, KY: The Appalachian Center, University of Kentucky.
 
Smith, P. B., & Pederson, D. R. (1988). Maternal sensitivity and patterns of infant-mother attachment. Child Development, 59, 1097-1101.
 
Stifter, C. A., Coulehan, C. M., & Fish, M. (1993). Linking employment to attachment: The mediating effects of maternal separation anxiety and interactive behavior. Child Development, 64, 1451-1460.
 
Weller, J. E. (1965). Yesterday's people: Life in contemporary Appalachia. Louisville, KY: University of Kentucky Press.
 
Werner, E. E. (1995). Resilience in development. Current Directions in Psychological Science, 4, 81-85.
 
Werner, E. E., & Smith, R. S. (1982). Vulnerable but invincible: A longitudinal study of resilient children and youth. New York: McGraw-Hill.
 
Werner, E. E., & Smith, R. S. (1992). Overcoming the Odds: High risk children from birth to adulthood. Ithaca, NY: Cornell University Press.
 
West Virginia Human Resources Association. (1988). Children in crisis: State at risk.
 
West Virginia Kids Count Fund. (1997). Brain development and the education of children from birth to age three: West Virginia Kids Count databook. Charleston, WV.
 
 
 
Footnotes
 
 
1This was not a sample with large numbers of adolescent mothers. Only 12% were 18 or younger.
 
2Because the period of free play was longer for the rural Appalachian sample, we also considered the possibility of a "warm-up" effect. To verify that the rural Appalachian mothers were not rated more sensitive and less intrusive at the earlier age because they had more time to get used to the free play situation, we compared only the first 5 minutes of their interaction with the Pennsylvania mothers' 5-minute sessions, using t-tests, and obtained the same pattern of results with the same level of significance.
 
 
 
This study was supported by grant MCJ-540615 from the Maternal and Child Health Bureau (Title V, Social Security Act), Health Resources and Services Administration, Department of Health and Human Services and by grant MH48395 from the National Institute of Mental Health, Department of Health and Human Services to the first author and grant MH 44324 from the National Institute of Mental Health and a grant from the Pennsylvania State University Biomedical Research Support Grant Program awarded to the second author.
 
The authors wish to thank the mothers and infants for their participation and to acknowledge the cooperation of the Lincoln Primary Care Center.
  
Corresponding author: Margaret Fish, Dept. of Family & Community Health, Marshall University School of Medicine, Huntington, WV 25755; phone (304) 691-1185; e mail fishm@marshall.edu