Maternal knowledge of child development and quality of parenting among White, African-American and Hispanic mothers

Applied Developmental Psychology 26 (2005) 149 – 170

Maternal knowledge of child development and quality of parenting among White, African-American and Hispanic mothersB Keng-Yen Huanga,*, Margaret O’Brien Caughyb, Janice L. Genevroc, Therese L. Millerc a

University of Maryland, Baltimore County, United States University of Texas Health Science Center, Houston, United States c Johns Hopkins University, United States

b

Abstract This study examined the relationship between early maternal knowledge of child development and later quality of parenting behaviors. Differences by race/ethnic group were also examined. Mother–infant dyads (N=378) participated in the study. Mothers completed the Knowledge of Infant Development Inventory (KIDI) when the infant was 2–4 months, and mother–toddler dyads were videotaped in their homes at 16–18 months. The Home Observation for Measurement of the Environment Inventory (HOME), Parent/Caregiver Involvement Scale (P/ CIS), and Nursing Child Assessment by Satellite Training (NCAST) were used to measure quality of mother– toddler interaction. A series of hierarchical multiple regression analyses was conducted controlling for confounding demographic variables. Results revealed no significant main effect of maternal correct estimation of child development on quality of parenting, but there was a significant main effect of maternal underestimation of child development on quality of parenting during a teaching task. There was also a significant interaction of maternal knowledge and race in relation to quality of parenting behavior. Implications for generalizability and interventions are discussed. D 2004 Elsevier Inc. All rights reserved. Keywords: Parenting; Maternal teaching behavior; Maternal knowledge of development; Race and ethnic differences; Maternal cognition

B This research is based on data collected from the Healthy Steps for Young Children Program, which was funded by the William T. Grant Foundation, The Amarillo Area Foundation, The Duke Endowment, The Hogg Foundation and The McLeod Foundation. Keng-Yen Huang is now at the NYU Child Study Center, New York University School of Medicine. * Corresponding author. NYU Child Study Center, New York University, School of Medicine, 215 Lexington Ave., 14FL, New York, NY 10016, United States. E-mail address: [email protected] (K.-Y. Huang).

0193-3973/$ - see front matter D 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.appdev.2004.12.001

150

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

1. Introduction Many developmental psychologists would agree that maternal cognitions play an important role in parenting and child development (Conrad, Gross, Fogg, & Ruchala, 1992; Damast, Tamis-LeMonda, & Bornstein, 1996; Fewell & Wheeden, 1998; Holden, 1995; MacPhee, 1981; Miller, 1988; Myers, 1982; Sigel & Kim, 1996; Sigel & McGillicuddy-De Lisi, 2002; Stoiber & Houghton, 1993). The study of maternal cognition has attracted increasing attention from researchers during the past two decades because it provides a framework for understanding parental actions and the processes of child development (Goodnow, 1996; Tamis-LeMonda, Shannon, & Spellmann, 2002). Among different dimensions of maternal cognition, such as maternal bbelief,Q bjudgment,Q bidea,Q bvalue,Q bexpectation,Q and bknowledgeQ (Goodnow, 1996; McGillicuddy-DeLisi & Sigel, 1995) maternal knowledge is one of the important dimensions. Maternal knowledge has been defined as bthe parent’s understanding of developmental norms and milestones, processes of child development, and familiarity with caregiving skillsQ (Benasich & Brooks-Gunn, 1996, p. 1187). Several aspects of maternal knowledge have been studied, such as knowledge of the timing of developmental milestones, mothers’ conceptions about how to parent, and what strategies to adopt during parent–child interactions (Goodnow, 2002; Stoiber & Houghton, 1993; Tamis-LeMonda et al., 2002). One central conceptual model that guides the study of maternal knowledge and parenting is the cognitive–behavioral perspective. It is hypothesized that maternal knowledge about the processes of child development influences the ways in which mothers interpret the behavior of their children and how they interact with their children (Cote & Bornstein, 2001; Goodnow, 1988; Miller, 1988). Mothers who are knowledgeable about child development are more likely to create an environment that is appropriate to their children’s development and/or more likely to interact with their children in more sensitive ways, which in turn will support their children’s social and cognitive development (Goodnow, 1988; Sigel, 1992). In contrast, mothers who have unrealistic expectations about child development or a limited repertoire of knowledge for effectively interacting with their children may use more ineffective parenting strategies, such as harsh and inconsistent discipline, thus setting the stage for less optimal child development (Dukewich, Borkowski, & Weitman, 1996). The influences of maternal knowledge on parenting have also been supported by research evidence. Mothers who judge their children’s abilities more accurately have been found to construct more appropriate learning environments and to interact with their children with better sensitivity/quality (Miller, 1988). Maternal knowledge and expectations about developmental competencies influence the type of environment that mothers provide for their infants (MacPhee, 1981; McGillicuddy-DeLisi, 1982; Stevens, 1984). Mothers who have a more accurate estimate of infants’ developmental milestones are more likely to talk, tell stories, and read to their babies. Knowledgeable and perceptive mothers structure an environment that is bmatchedQ to the infant’s level of development (Parks & Smeriglio, 1986; Stevens, 1984). Therefore, knowledgeable mothers are more likely to have children with better cognitive development (Dichtelmiller et al., 1992). Several investigators also have reported that mothers who judge their children’s abilities more accurately are more likely to interact with their children with greater sensitivity. Fry (1985) reported that adolescent mothers who have better knowledge of child development tend to interact more positively with their infants than adolescent mothers who have less developmental knowledge. Also, mothers with more knowledge of child development tend to use fewer love withdrawal and power assertive discipline strategies, and use more inductive reasoning (Smith, 2002). Mothers with distorted

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

151

and unrealistic expectations about their children are more likely to abuse or use severe discipline with their children compared to mothers with more accurate child development expectations (Belsky, 1980; Dukewich et al., 1996; Hunt & Paraskevopoulos, 1980). It is likely that knowledgeable mothers provide warm and positive environments that promote emotional growth; therefore, knowledgeable mothers are more likely to have children with better social–emotional development, including better compliance (Smith, 2002). Although numerous studies suggest that maternal knowledge is a significant component of promoting high-quality parenting, several limitations of this literature remain to be addressed. First, not all studies support the hypothesis that maternal knowledge promotes high quality of parenting. There are some studies that show no significant relation between maternal knowledge and parenting, especially for the studies that focus on predominantly White, well-educated, and middle-class mothers (Conrad et al., 1992; Myers, 1982). This suggests that the subgroup differences need to be further examined. Because maternal knowledge has been shown to vary among different cultures and subcultures within the U.S. (Bornstein et al., 1996), and because research to date has remarkable little understanding of race/ethnicity differences on maternal knowledge and their impact on parenting, exploring race/ethnic similarities, and differences in the link between maternal knowledge and parenting is extremely important. Such studies can contribute to the literature of how culture/ethnicity shapes childrearing (e.g., Cote & Bornstein, 2001) and the development of effective and culturally sensitive prevention/ intervention programs that meet the needs of a variety of minority populations. Given the diverse cultures/subcultures that shape maternal cognition, some basic assumptions of maternal knowledge about child development processes may not be universal (Bornstein et al., 1996; Zepeda & Espinosa, 1988). For example, studies have found that maternal knowledge about the development of certain childhood capabilities varied with cultural affiliation and/or socioeconomic status (Miller, 1988; Zepeda & Espinosa, 1988). Savage & Gauvain (1998) studied European-American and Latino families and found that minority mothers who were more acculturated to the U.S. were more similar to European American mothers in their beliefs about child development. Also, Zepeda & Espinosa (1988) compared four ethnic groups of low-income mothers in the U.S. (Anglo, Black, USborn Hispanic, and foreign-born Hispanic parents) and found that foreign-born Hispanic mothers were more likely to underestimate their children’s development (e.g., believe later attainment of behavioral skills, including motor and self-help skills and language abilities) than were either the U.S.-born Hispanic mothers or the Anglo and Black mothers. These findings suggest that maternal knowledge of child development in different cultural/ethnic groups may represent differences in beliefs regarding normal and acceptable behavior. The different maternal conceptualizations of normal and acceptable behavior in different groups may also lead to differences in maternal behavior (Cote & Bornstein, 2001). Therefore, the association between maternal knowledge of child development and parenting could be moderated by race/ethnicity. The second limitation of the maternal knowledge literature is that most of the supporting evidence is based on high-risk samples (e.g., teen mothers, low-income families, first-time mothers, depressed mothers, mothers of premature babies). Findings from studies based on high-risk samples may not generalize to low-risk samples. For example, studies showed that teen/younger mothers were less knowledgeable about developmental milestones of young children than older mothers (Brooks-Gunn & Furstenberg, 1986; Hammond-Ratzlaff & Fulton, 2001; Karraker, 1996). Poverty and low socioeconomic status were found to be associated with lower levels of maternal knowledge (Benasich &

152

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

Brooks-Gunn, 1996; Conrad, Gross, Fogg, & Ruchala, 1992; Stevens, 1984). In addition, maternal experience was related to parental knowledge. First-time mothers had less accurate knowledge of child development than experienced mothers (MacPhee, 1981). Moreover, mothers who reported higher symptoms of depression in the early postpartum period were found to have less accurate knowledge of child development (Veddovi, Kenny, Gibson, Bowen, & Starte, 2001). These findings reinforce the need for research based on samples that can be generalized to the larger population. The third limitation of the maternal knowledge literature is that most studies have been based on only one measure of parenting behavior. Few studies have looked at maternal knowledge and quality of parenting in different contexts, such as teaching and free-play. Also, less attention has been given to systematically examining what domains of parenting behaviors were more likely to be influenced by improvement of maternal knowledge and what domains of parenting behavior were less likely to be influenced. Studying multiple dimensions of parenting behaviors will provide a better understanding about the link of maternal knowledge and parenting. From an applied developmental science perspective, these questions have important implications. Knowledge gained in this area can contribute to the development of effective prevention/intervention programs that meet the needs of a variety of minority populations. For example, systematically studying the association between specific dimensions of parental knowledge and parenting behaviors (e.g., parent–child interaction) can help us understand which components of parental knowledge to target with interventions in order to see maximal improvement in the quality of mother–infant interaction. Such information can help us to design parenting programs that are implemented in a culturally sensitive manner. In the present study, we sought to resolve some of the limitations in the literature and further investigate whether the association between maternal knowledge and parenting varies with race/ ethnicity. Specifically, maternal knowledge was defined as maternal knowledge of developmental norms and milestones. Although, undoubtedly, other parental knowledge constructs are worthy of study, this dimension was selected because it has been the exclusive focus in many early intervention and parenting programs. Three different parenting behaviors were studied, including quality of home environment and quality of parent–toddler interaction in a teaching and a free-play context. In addition, this study was based on a relatively low-risk and culturally diverse sample. We examined how parental knowledge at 2– 4 months relates to parenting behaviors at 16–18 months. Three questions were addressed: (1) How does maternal knowledge vary by demographic characteristics (e.g., maternal education, age, race/ethnicity, depressive symptoms, and poverty)? (2) Does early maternal knowledge relate to later parenting behaviors? Do dimensions of maternal knowledge (e.g., correct estimation of development and incorrect estimation) relate to parenting behavior differently? (3) Are there racial/ethnic group differences in the association between maternal knowledge and parenting behavior?

2. Method 2.1. Participants Participants were drawn from The Healthy Steps for Young Children (Healthy Steps) Embedded Observational Study. The Healthy Steps Embedded Observational Study was a direct observational

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

153

study of mother–child interaction nested within the National Evaluation of Healthy Steps. Healthy Steps was initiated in September 1996 as a pediatric-based parent support program for families of newborns. Families were followed from the child’s birth until three years. Families at 15 sites across the country participated in the national evaluation, six of which featured a randomized design and nine utilized a quasi-experimental design. For more information on Healthy Steps and its evaluation, please see Caughy, Miller, Genevro, Huang, and Nautiyal (2003), Guyer, Hughart, Strobino, Jones, and Scharfstein (2000) and Minkovitz et al. (2003). Families at two of the randomized sites were asked to participate in the embedded study, one site located in the southwest and the other in the southeast. Of the 5565 families enrolled in the Healthy Steps National Evaluation, 658 families were eligible to participate in the Embedded Study. Families were excluded from the Embedded study if: (1) the child never made a visit to the pediatric practice; (2) the family withdrew from the National Evaluation; (3) the family had moved out of the study range, a 1-h travel distance by car; (4) the study child died; (5) the study child no longer lived with the biological mother; (6) the mother did not speak English; or (7) the child was born after September 15, 1998. A total of 378 of the eligible families (57%) completed at least part of the first home visit at age 16–18 months; 126 (19%) refused to participate; 72 (11%) could not be located; 61 (9%) of the children were over 18 months before they could be located and interviewed; and 23 (4%) did not complete the visit for other reasons such as language constraints. The final sample for this study included the 378 mothers who completed the 2- to 4-month telephone interview and 16–18 months observation. Preliminary analyses revealed that the mothers who completed the observation tended to be older, v 2(2, N = 657) = 10.7, p b .01; better educated, v 2(3, N = 657) = 22.6, p b .001; married, v 2(2, N = 657) = 15.3, p b .001; living with the baby’s father, v 2(1, N = 636) = 12.4, p b .001; and less likely to have received Medicaid during pregnancy, v 2(2, N = 654) = 11.0, p b .001 than the mothers who did not complete the interview. Further, significantly more White women completed the interview than did not, 62% versus 38%, v 2(3, N = 658) = 12.7, p b .01. Table 1 presents the baseline demographic characteristics of the final study sample (N = 378).

Table 1 Characteristics of study families Maternal race, education, and age Maternal race White/Non-Hispanic African-American Hispanic Maternal education bHigh school High school/GED Some college zCollege Maternal age V19 years 20–29 years z30 years

N

%

234 93 48

61.9 24.6 12.7

54 125 124 74

14.3 33.1 32.8 19.6

65 217 95

17.2 57.4 25.1

Balance represents missing observations (N = 378).

Maternal marital status Married Single Divorced/separated/widowed Poverty status b1.0 poverty z1.0 poverty

First-time mother First-time mother Not first-time mother

N

%

252 117 8

66.7 31.0 2.1

97 255

25.7 67.5

192 160

50.8 42.3

154

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

2.2. Procedure Trained interviewers conducted two in-home observations, one at 16–18 months and the other at 34–37 months, each lasting approximately 2 h. Each home visit consisted of a structured interview with the mother and direct observation of the mother and child in a series of videotaped tasks, including a teaching task (6 min) and a free-play task (15 min). This study utilized data from the first home observation (16–18 months) as well as data collected as part of the Healthy Steps National Evaluation when the family enrolled in the study and when children were between 2 and 4 months old. During the teaching task, the mother was asked to select one task that her child had not yet learned from a list of options (selected from the Nursing Child Assessment by Satellite Training; NCAST; Sumner & Spietz, 1994) and teach it to her child. For the free-play task, the mother was asked to play with her child on the floor as she normally would for 15 min. A standard collection of toys was provided by the interviewers, including some gender specific as well as gender-neutral types of toys. Parental behaviors during the free-play task, the teaching task, and the home observation during the interview were coded. The coding system is described in more detail in a subsequent section of this report. In addition to the home visit data, several measures were derived from a questionnaire that was completed when the child enrolled in the Healthy Steps program as a newborn and from a telephone questionnaire administered at age 2–4 months. The newborn questionnaire provided basic demographic information on the family. The telephone interview at 2–4 months included questions about the mother’s knowledge of child development and maternal depressive symptoms. 2.3. Measures 2.3.1. Maternal knowledge of child development The measure of mothers’ knowledge of child development was adapted from the Knowledge of Infant Development Inventory (KIDI; MacPhee, 1981). Thirty statements were selected from the original 75 items of the KIDI (21 items asked questions related to norms and milestones, and 9 items asked questions related to parenting). The items included in this study describe a typical infant’s behavior at a given point in time, including motor, perceptual, and cognitive milestones, as well as parenting responses regarding crying, feeding, and diapering. Mothers responded to items by agreeing or disagreeing with each statement, or reporting they were not sure. A correct score was computed for each item. For a subset of 17 statements that related to norms and milestones, if mothers disagreed with the statements, they were further asked if each statement was true of a younger or older child. An underestimate and overestimate score was computed for these 17 items. Underestimate was defined as the mother’s incorrect estimation that the child is too young to be capable of an activity/task. Overestimate was defined as the mother’s incorrect estimation that the child is capable of an activity/task that is actually too difficult for them at their present age. Three proportion scores were then computed including the percentage of correct answers, the percentage of items for which the respondent underestimated development, and the percentage of items for which the respondent overestimated development. In our study, internal consistency (Cronbach’s alpha) of the entire KIDI scale (30 items) and internal consistency of the 17 selected norms and milestones items were both .70. Consistent with previous research findings (i.e.,

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

155

MacPhee, 1981; Veddovi et al., 2001), our modified version of the KIDI was correlated with maternal education (r = .20), parity (r = .16), poverty level (r = .14), and maternal depressive symptoms (r =
.14) in the expected directions, suggesting that the modified KIDI has adequate reliability and validity. 2.3.2. Parenting behaviors Three parenting behaviors were measured. The Home Observation for Measurement of the Environment (HOME) Inventory, Infant/Toddlers Version (Bradley & Caldwell, 1984) was used to assess the home environment that parents provided. The HOME scale measures the quantity and quality of social, emotional, and cognitive support available to a young child within his or her home, including parental interaction and involvement with their child, materials and stimulation parents provide to their children, organization of the environment, and acceptance of the child. The internal consistency (Cronbach’s alpha) of the HOME total scale was .72 for our sample. 2.3.3. The parent/caregiver involvement scale The P/CIS (Farran et al., 1987) is a behavioral rating scale of caregiver interaction with young children ages 3 to 36 months. It was used to assess the quality, appropriateness, and general impression of the mother’s involvement with her child during a videotaped free play session. The P/CIS includes 11 specific aspects of parenting related to the construct of maternal sensitivity, each rated on five-point scales for Amount, Quality, and Appropriateness in the dyadic interaction. Only the quality and appropriateness scores were used because these two aspects are central to the notion of maternal sensitivity. In this study, inter-rater reliability was assessed using an intraclass correlation coefficient (ICC) and ranged from .74 to .77. 2.3.4. The teaching scale score of the nursing child assessment by satellite training (NCAST) This measure (Barnard, 1978; Sumner & Spietz, 1994) consists of 73 items organized into six subscales. Four subscales measure maternal behavior (sensitivity to cues, response to child’s distress, social–emotional growth fostering, and cognitive growth fostering) and two subscales measure child behavior (clarity of cues and responsiveness to parent). Only the maternal behavior subscales (50 items), which were combined into one maternal behavior scale, were used in the present study. Higher scores indicate a more optimal relationship between mother and child and more sensitive parenting. The ICC for inter-rater reliability of the maternal behavior scale was .76. 2.3.5. Covariates Covariates included treatment effect of Healthy Steps, maternal education, marital status, race/ ethnicity (not used as a covariate in interaction analyses), and parity (first-time mother). Family socioeconomic status, assessed at the 2- to 4-month interview, was measured using an income-to-need ratio based on the ratio of family income to family size (divided total family income by the appropriate poverty threshold for each household size based on 1997 U.S. Department of Labor poverty guidelines). Living below the federal poverty level was defined as equivalent to an income-to-needs level of less than 1.0. Maternal depressive symptoms were assessed during the telephone interview at 2–4 months using questions adapted from the Centers for Epidemiological Studies—Depression (CES-D) inventory (Radloff, 1977). A shortened version of the instrument, decreased from 20 items to 14, was developed for the Healthy Steps evaluation by assessing redundant items using data from 600 parents in an

156

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

evaluation of a child abuse prevention program in Pennsylvania. The correlation of the reduced version with the full scale was over 0.90 and 0.95 in a separate study involving pregnant drug users (D.M. Strobino, personal communication, 2000).

3. Results Characteristics of the participants are displayed in Table 1. More than half of the mothers were between age 20 and 29 at the time of the child’s birth. Sixty-seven percent of the mothers were married. For 51% of the mothers, this was their first child. Because we were also interested in examining differences in knowledge among racial/ethnic groups, we examined demographic differences by race/ ethnic group. In general, compared to White mothers, African-American and Hispanic mothers were younger, v 2(4, N = 374) = 10.92, p b .05; less well-educated, v 2(6, N = 374) = 38.64, p b .001; more likely to be living below the poverty level, v 2(2, N = 349) = 61.99, p b .001; and less likely to be married, v 2(4, N = 374) = 65.67, p b .001. 3.1. Differences in parental knowledge by demographic characteristics Average scores for the three KIDI scores and adjusted mean scores for the three racial/ethnic groups are shown in Table 2. On average, mothers correctly estimated their child’s development 56% of the time. Analysis of covariance (ANCOVA) was utilized to examine whether there were racial/ethnic group differences in the KIDI mean scores (see Table 2). Results revealed that after adjusting for demographic differences (e.g., maternal age, education, poverty, and marital status), the KIDI percent correct scores differed significantly among the three racial/ethnic groups. Bonferroni’s adjustment of post hoc analyses results indicated that compared to African-American mothers, White and Hispanic mothers had higher KIDI percent correct scores. Zero-order correlations between the three KIDI scores and demographic characteristics are displayed in Table 3. Older, married, more experienced (not first-time) and wealthier (not living in poverty) mothers, and mothers with higher education and lower depressive symptoms tended to estimate child development more accurately than their counterparts. In contrast, younger, not married, and poorer Table 2 Mean (and SD) KIDI scores for mothers from three racial/ethnic groups Racial/ethnic group Type KIDI score % items correct % items underestimate % items overestimate

Total

White

African-American

Hispanic

n = 347

n = 218

n = 82

n = 44

.56 (.11) .09 (.06) .11 (.08)

.57 (.01) .09 (.00) .10 (.01)

.51 (.01) .10 (.01) .13 (.01)

.57 (.02) .10 (.01) .10 (.01)

F(2, 337) 6.29*** 0.67 2.73y

F statistics compared the mean differences among three racial groups, adjusting for maternal age, education (2 groups: 0 = high school or less; 1 = above high school), marital status (0 = unmarried, 1 = married); and poverty level (0 = not poor, 1 = poor). The adjusted means (and SD values) are reported for three racial groups. y p b .10. *** p b .001.

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

157

Table 3 Zero-order correlations among KIDI scores, demographic factors, and parenting behavior scores 1

2

KIDI 1. KIDI % correct 2. KIDI % under 3. KIDI % over

–
.24
.59

–
.17

–

Demographic 4. Age 5. Education 6. Marital status 7. First-time mother 8. Depressive symptom 9. Poverty status 10. Child’s gender

.15 .14 .17
.15
.14
.17
.06


.02
.12
.09
.04 .17 .05 .02


.14
.08
.18 .08 .06 .13 .08

.19 .12 .13 .13


.07
.03
.03
.15


.17
.18
.16
.16

Parenting behaviors 11. HOME-total 12. P/CIS-Quality 13. P/CIS-Appropriate 14. NCAST-parent

3

4

5

6

.34 .01
.14
.31
.05

–
.21
.10
.39
.15

.33 .24 .27 .29

.36 .29 .25 .31

7

8

9

10

11

12

13

14


.08
.01
.04
.03

– .43 .39 .29

– .86 .42

– .34

–

– .39 .46
.32
.15
.27
.10

.24 .20 .19 .23

– – .02
.02 .06

–

.12 .03 .03
.05


.21
.19
.22
.14

.17 .02

– .17


.27
.31
.28
.31

–

Correlations with education (2 groups; 0 = high school or less; 1 = above high school), marital status (0 = unmarried, 1 = married), first time mother (0 = no, 1 = yes), poverty (0 = not poor, 1 = poor), and child’s gender (2 groups; 0 = female, 1 = male) are based on a Spearman’s rho. Significant levels are as follows: for .11 b j r j b .13, p b .05; for .14 b j r j b .16, p b .01; for .17 b j r j b .84, p b .001.

mothers tended to overestimate their child’s ability; they believed that their children were capable of doing activities that are actually too advanced for them. Mothers of lower education and higher depressive symptoms tended to underestimate their child’s ability—they believed that their children were not capable of doing activities that were actually appropriate for their age. 3.2. Relationship of parental knowledge to parenting behaviors Zero-order correlations showed that the KIDI percent correct and percent overestimated scores were significantly associated with all of the parenting measures, although the KIDI percent underestimated score was only associated with one parenting measure, the NCAST maternal subscale (see Table 3). Mothers with higher KIDI percent correct scores and lower KIDI percent overestimate scores also had higher scores on the HOME scale, NCAST maternal subscale, and P/CIS Quality and Appropriateness. These results indicate that the larger the number of correct estimates made by a mother, the higher the quality of home organization and mother–child interaction. Because the data were drawn from an intervention study, the association between maternal knowledge and parenting behavior might be confounded by the intervention group status (intervention vs. control group). If this is the case, statistical models need to take the intervention group status into account. We examined whether there were initial intervention group difference on KIDI scores by using ANCOVA— adjusting for maternal age, education, marital status, and family poverty level. The results showed that mothers of the intervention group did not differ from mothers of control group in child development

158

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

knowledge in the outset of the intervention. Therefore, subsequent analyses exploring the associations between KIDI scores and parenting outcomes did not control for the intervention status. A series of four multiple linear regression analyses were conducted to examine the degree to which parental knowledge (KIDI scores) were predictive of parenting behavior (HOME scores; P/CIS quality of involvement; P/CIS appropriateness; and NCAST scores for optimal mother–child relationship and sensitive parenting) after controlling for the maternal age, education, marital status, parity, poverty, and race/ethnicity. These analyses were conducted first with the total number correct scores from the KIDI as the predictive variable and again with the scores for parental underestimation of the child’s abilities as the independent variable. Because the KIDI percent overestimate score was strongly correlated with the KIDI percent correct score (r =
.59) and both variables were correlated with parenting behaviors with similar effect size but in opposite directions (as expected), the KIDI percent overestimate score was not included in the multivariate analyses. Results from the hierarchical regression model applied to KIDI total correct scores are shown in Table 4 and findings relevant to KIDI underestimation scores are presented in Table 5. The results indicated that after controlling for demographic variables, the KIDI percent correct scores were no longer significantly associated with measures of parenting behaviors (See Table 4). However, the KIDI percent underestimate score was significantly related to NCAST scores even after the association of demographic characteristics with parental behaviors were controlled (see Table 5). Mothers with scores indicating greater underestimation of children’s capabilities, as assessed by KIDI, tended to interact less sensitively with their children during the teaching context. As indicated in the regression models presented in Tables 4 and 5, maternal race/ethnicity was the major factor explaining the association between parental knowledge and parenting behavior. AfricanAmerican mothers differed significantly from White mothers in terms of all measures of parenting behaviors. Hispanic mothers differed only from White mothers on the NCAST maternal subscale score. 3.3. Race/ethnicity differences in the relationship of parental knowledge to parenting behavior Given the consistent evidence that maternal race/ethnicity explained a significant proportion of the variance of most of the parenting measures, we further explored possible interaction effects of maternal knowledge and race/ethnicity in relation to parenting behaviors. A series of hierarchical multiple regression analyses were conducted on parenting behaviors. At the first step of the analysis, maternal demographic characteristics were entered as covariates. At the second and third steps, maternal race/ ethnicity and the KIDI score were entered separately to test for main effects of each variable. At the final step, two interaction terms (KIDI score  Race/African-American, and KIDI score  Ethnicity/Hispanic) were entered simultaneously. The results of these analyses are presented in Tables 6 and 7; only significant moderator effects are shown in the tables. As expected, the relationships between the parental knowledge and parenting behaviors varied depending on the racial/ethnic groups. Specifically, race/ethnicity moderated the relation between maternal correct estimation of child development (KIDI % correct scores) and three parenting behaviors—HOME total score, P/CIS-Quality subscale, P/CIS-Appropriateness and NCAST maternal behavior scores (See Table 6). The findings presented in Table 7 show that race/ethnicity also moderated the relation between maternal percent underestimate (KIDI % under scores) of child development and two parenting outcomes—HOME total score and P/CIS-Appropriateness subscale.

Four parenting behavior scores Variables

HOME Total (n = 337) B

Control variables Age Education Marital status Parity Poverty status Race/AfricanAmerican Race/Hispanic

(Step 1) 0.09* 0.99*
1.13* 1.16**
0.25
2.88***
0.44

SE R 2 change F change 0.04 .29 0.40 0.49 0.39 0.46 0.51

19.12***

0.58

Knowledge variable (Step 2) KIDI % Correct 1.60 1.74 .00

0.85

P/CIS Quality (n = 313) B

SE R 2 change F change

0.01 0.06
0.08 0.04
0.13*
0.24***

0.00 .23 0.04 0.05 0.04 0.05 0.06


0.03

0.06


0.07

0.18 .00

12.63***

0.16

P/CIS Appropriateness (n = 313) B

SE R 2 change F change

0.01 0.10*
0.06 0.03
0.12*
0.17**

0.00 .18 0.04 0.05 0.04 0.51 0.06


0.04

0.06

0.05

0.18 .00

9.28***

NCAST Maternal Subscale (n = 289) B 0.04 1.16*
1.05
0.67
1.26y
2.66***

SE

R 2 change F change

0.05 .24 0.56 0.70 0.54 0.66 0.74

12.42***


2.93*** 0.82

0.08


0.59

2.40 .00

0.06

Age (continuous); Education (2 groups: 0 = high school or less, 1 = above high school); Marital status (0 = single/separated/divorced/widow, 1 = married); Parity (0 = not first-time mother, 1 = first-time mother), Poverty (0 = not poor, 1 = poor); Race/ethnicity (2 groups). Coefficients and Standard errors from the last step of the analyses were reported. y p b .10. * p b .05. ** p b .01. *** p b .001.

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

Table 4 Summary of regression analysis of parenting behaviors at 16- to 18-month assessment on parental knowledge (% correct) at 2- to 4-month assessment

159

160

Four parenting behavior scores Variables

HOME Total (n = 337) B

Control variables (Step 1) Age 0.09* Education 0.99* Marital status
1.13* Parity 1.16** Poverty status
0.25 Race/African
2.88** American Race/Hispanic
0.44

P/CIS Quality (n = 313)

SE R 2 change F change 0.04 .29 0.40 0.49 0.39 0.46 0.51

19.12***

0.57

Knowledge variable (Step 2) KIDI % Under 0.64 3.24 .00

.04

B

SE R 2 change F change

0.01 0.06
0.08 0.04
0.13*
0.24***

0.00 .23 0.04 0.05 0.04 0.05 0.06


0.03

0.06

0.16

0.35 .00

12.63***

0.20

P/CIS Appropriateness (n = 313) B

SE R 2 change F change

0.01 0.10*
0.06 0.03
0.12*
0.17**

0.00 .18 0.04 0.05 0.04 0.05 0.06


0.04

0.06

0.17

0.36 .00

9.28***

NCAST Maternal Subscale (n = 289) B 0.04 1.16*
1.05
0.67
1.26y
2.66***

SE

R 2 change F change

0.05 .24 0.56 0.70 0.54 0.66 0.74

12.42***


2.93*** 0.82

0.23


9.07*

4.34 .01

4.36*

Age (continuous); Education (2 groups: 0 = high school or less, 1 = above high school); Marital status (0 = single/separated/divorced/widow, 1 = married); Parity (0 = not first-time mother, 1 = first-time mother); Poverty (0 = not poor, 1 = poor); Race/ethnicity (2 groups). Coefficients and Standard errors from the last step of the analyses were reported. y p b .10. * p b .05. ** p b .01. *** p b .001.

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

Table 5 Summary of regression analysis of parenting behaviors at 16- to 18-month assessment on parental knowledge (% underestimate) at 2- to 4-month assessment

Four parenting behavior scores Variables

HOME total (n = 339) B

Control variable (Step 1) Age 0.07y Education 1.25** Marital status
2.10*** Parity 1.50*** Poverty status
0.93*

SE R 2 change F change 0.04 .22 0.43 0.48 0.41 0.48

Race/ethnicity (Step 2) Race/African
3.07*** 0.50 .08 American Race/Hispanic
0.78 0.58 Knowledge (Step 3) KIDI % correct 1.96 Interaction effect (Step4) KIDI % correct  10.46** Black KIDI % correct  13.56* Hispanic

18.31***

PCIS Quality (n = 317) B 0.00 0.08y
0.15** 0.06
0.19***

SE R 2 change F change 0.00 .17 0.05 0.05 0.04 0.05

18.80***
0.24*** 0.05 .05
0.03

0.06

1.72 .00

1.30


0.07

0.19 .00

3.86 .02

4.98**


0.14

.25 .03

5.32

1.60**

Coefficients and Standard Errors from the last step of the analyses were reported. y p b .10. * p b .05. ** p b .01. *** p b .001.

12.70***

PCIS Appropriateness (n = 317) B 0.00 0.12**
0.11* 0.05
0.16**

10.15***
0.18*
0.04

SE R 2 change F change 0.00 .15 0.05 0.05 0.04 0.05

10.53***

0.06 .03

5.51**

0.06

NCAST Maternal subscale (n = 292) B 0.03 1.58**
1.76**
0.20
2.03**

SE

R 2 change F change

0.05 .18 0.58 0.65 0.55 0.64

12.19***


2.59*** 0.69 .05

8.78***


2.37**

0.80

0.13

0.04

0.19 .00

0.04


0.32

2.38 .00

5.10**

0.25

.19 .02

3.18*


17.19**

5.28 .04


23.53**

7.27

1.47*

.02

7.58***

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

Table 6 Summary of regression analyses showing race/ethnicity as a moderator in the relation between parental knowledge (% correct) and parenting behaviors

161

162

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

Table 7 Summary of regression analyses showing race/ethnicity as a moderator in the relation between parental knowledge (% underestimate) and parenting behaviors Two parenting behavior scores Variables

HOME total (n = 342)

PCIS Appropriateness (n = 318)

B

SE

R change

F change

B

SE

R 2 change

F change

Control variable (Step 1) Age Education Marital status Parity Poverty status

0.07y 1.25**
2.10*** 1.50***
0.93*

0.04 0.43 0.48 0.41 0.47

.22

18.48***

0.00 0.12**
0.11* 0.05
0.16**

0.00 0.05 0.05 0.04 0.05

.15

10.57***

Race/ethnicity (Step 2) Race/African-American Race/Hispanic


3.07***
0.78

0.50 0.57

.08

18.97***


0.18***
0.04

0.06 0.06

.03

5.52**

0.27

3.25

.00

0.01

0.19

0.36

.00

0.27

4.78 29.91**

7.17 11.14

.02

3.61*

1.77* 2.48*

0.79 1.23

.02

3.56*

Knowledge (Step 3) KIDI % under Interaction effect (Step 4) KIDI % under  Black KIDI % under  Hispanic

2

Coefficients and Standard errors from the last step of the analyses were reported. y pb.10. * pb.05. ** pb.01. *** pb.001.

Fig. 1 displays the significant interaction patterns for KIDI percent correct score and the four parenting behavior scores and Fig. 2 displays the significant interaction patterns for KIDI percent underestimate score and the two significant measures of parenting behaviors. There are several patterns apparent in Fig. 1. First, among White mothers, knowledge of child development appeared to be related to only one domain of parenting behavior. For White mothers, greater maternal knowledge of child development (higher KIDI % correct scores) was related to improved quality of the mother–child interaction during the teaching task as measured by the NCAST (items assessing sensitivity to cues, responsivity to distress, and fostering cognitive and socioemotional growth). Child development knowledge was not associated with improved home environment or the quality of interaction during free-play as measured by the HOME and P/CIS scales, respectively. For Hispanic mothers, knowledge of child development was associated with all four measures of parenting behaviors. In the case of the HOME scale and the P/CIS, this association was in the expected direction; that is, a higher accurate level of child development knowledge was associated with a higher score on the HOME scale measure of parental interaction, materials and stimulation parents provide, organization of the home and child acceptance, the P/CIS measure of quality of involvement; and the P/CIS measure of appropriateness in the parent–child interaction. However, the association between child development knowledge and the NCAST scores for optimal mother–child relationship and sensitive parenting was counterintuitive. Higher accuracy of child development knowledge was associated with lower scores on the NCAST for Hispanic mothers.

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

39

b 3.6

38

3.5

P/CIS-Quality

HOME-Total

a

37 36 35 34

3.4 3.3 3.2 3.1 3

33

2.9

32

-1SD

+1SD

-1SD

KIDI % Correct White

African American

White

Hispanic

d

3.6 3.55 3.5 3.45 3.4 3.35 3.3 3.25 3.2 3.15 3.1

+1 SD KIDI % Correct

c

African American

Hispanic

36 35 34

NCAST

P/CIS-Appropriatness

163

33 32 31 30 29 28

-1SD

+1 SD

-1SD

KIDI % Correct White

African American

+1SD KIDI % Correct

Hispanic

White

African American

Hispanic

Fig. 1. Race/ethnicity as a moderator in the relation between accuracy of parental knowledge (% correct) and parenting behavior: (a) KIDI (% correct) and HOME data; (b) KIDI (% correct) and P/CIS-Quality; (c) KIDI (% correct) and P/CISAppropriateness; and (d) KIDI (% correct) and NCAST.

For African-American mothers, accuracy of child development knowledge was positively associated with the HOME scale scores. Similar to Hispanic mothers, African-American mothers also demonstrated an association between child development knowledge and NCAST scores that was counterintuitive, with greater accuracy of knowledge associated with less optimal parent–child interaction and sensitive parenting during the teaching task. Fig. 2 shows the pattern of relationships between child development knowledge and parenting behaviors for the two scales in which race/ethnicity was found to moderate the relationships between parental underestimates of the child’s abilities and observed behavior. For White mothers, there was no relationship between maternal underestimation of child development and the quality of parenting. For Hispanic mothers, greater underestimation of child development was associated with higher scores on the HOME scale and higher quality of mother–child interaction as measured by the P/CISAppropriateness. For African American mothers, there was no relationship of underestimation of the child’s abilities and HOME scores, but greater underestimation of child abilities was associated with higher P/CIS scores for appropriateness of the mother–child interaction. Because the association between KIDI percent correct scores and NCAST scores was counterintuitive for Hispanic and African-American mothers, we further explored the characteristics

164

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

a

40

HOME-Total

39 38 37 36 35 34 33 -1SD

+1SD

KIDI % Underestimate White

P/CIS-Appropriateness

b

African American

Hispanic

3.6 3.55 3.5 3.45 3.4 3.35 3.3 3.25 3.2 3.15 3.1

White

-1SD +1 SD KIDI % Underestimate Hispanic African American

Fig. 2. Race/ethnicity as a moderator in the relation between inaccuracy of parental knowledge (% underestimate) and parenting behavior: (a) KIDI (% underestimate) and HOME; and (b) KIDI (% underestimate) and P/CIS-Appropriateness.

of the specific teaching tasks these mothers utilized during the NCAST segment. Bivariate correlations between KIDI percent correct estimation scores and characteristics of the teaching task (difficulty of task chosen, length of task, and 50 behaviors representing sensitivity to cues, responsivity to distress, and fostering of cognitive and socioemotional growth, observed during the NCAST task) were calculated separately for Hispanic and African-American mothers. The results suggested that Hispanic mothers with more accurate knowledge of child development tended to spend more time teaching their children the task, r(37) = .30, p = .07, a weak to moderate effect size, although this relationship only approached the level of significance. They also were less likely to caress, hug, or kiss their child, or make cheerleading-type statements during teaching or when their child was distressed, r values ranging from
.31 to
.46, p b .05. In addition, these mothers were less likely to let their children explore the task material before giving them instruction [r (37) =
.34, p b .05], and were less likely to signal that the child completed the task when the teaching was complete [r (37) =
.53, p b .001]. Although only approaching the level of significance, African-American mothers with higher correct knowledge of child development tended to choose more difficult tasks to teach their children, r (73) = .20, p = .09, were less likely to yell at their child (r (74) =
.28, p b .05), did not allow nontask manipulation of the task material, r (74) =
.26, p b .05, and were somewhat more task oriented

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

165

during the teaching (e.g., focused child’s attention on the task), r (73) = .22, p = .06, and did not divert child’s attention to other task when they were in distress, r (74) =
.21, p = .06). This pattern of relationships differed from that observed for Hispanic mothers. As also indicated in Fig. 2, the associations between KIDI underestimate score and parenting behaviors for Hispanic mothers (e.g., P/CIS-Appropriateness score and the HOME scale) were counterintuitive. Additional analyses were conducted to further examine the characteristics of the behaviors these mothers utilized during the P/CIS and HOME observation segments. Bivariate correlations between the KIDI percent underestimate score and maternal behaviors (e.g., 11 behaviors assessed in the P/CIS and the 6 HOME subscale scores) were computed. The results revealed that Hispanic mothers who had higher percent of underestimation scores tended to be more involved with their children (e.g., structured child’s play, encouraged developmental advance) during the interaction, r (44) = .33, p b .05.

4. Discussion We examined how mothers’ knowledge of child development was related to specific parenting behaviors in a relatively low-risk sample. After adjusting for the effects of demographic characteristics, there were no significant relationships between the proportion of correct knowledge of child development and parenting behaviors. However, there was a significant association between the proportion of underestimation of child development and parenting behavior during a teaching task. We found that the association between maternal knowledge of child development and parenting behaviors was moderated by race/ethnicity. For Hispanic and African-American mothers, greater knowledge regarding child development was associated with higher quality home environments. For Hispanic mothers only, greater knowledge of child development was associated with higher quality of observed mother–child interaction during a free-play context. Interestingly, since it would seem that less knowledge would lead to lower quality parenting, Hispanic mothers who had higher underestimating scores (i.e., less knowledge of child development) also provided a better quality of home environment and interaction during play. For White mothers, maternal knowledge was only associated with quality of mother–child interaction during teaching. These results only partially support the hypothesis that improvement of maternal knowledge of child development will improve parenting. Our findings of nonsignificant associations between correct knowledge of child development and parenting behavior are consistent with the findings that based on a primarily middle-class White sample (Conrad et al., 1992; Myers, 1982). However, these findings are inconsistent with the majority of the previous research findings based on high-risk samples (Benasich & Brooks-Gunn, 1996; Hunt & Paraskevopoulos, 1980; Parks & Smeriglio, 1986; Stevens, 1984). Possible reasons for the discrepancy between our findings and those of previous research may be that maternal knowledge and parenting was measured contemporaneously in previous studies; that previous studies focused on samples from one racial group or high-risk groups; and/or that different measurement tools were used. Our findings suggested that studies based on high-risk populations (e.g., teen mothers, depressed mothers) or a single race/ethnic group may limit generalizability to other populations. In the present study, we found a significant negative association between maternal underestimation of child ability and quality of mother–child interaction during teaching. Mothers who underestimated their child’s ability interacted with their child less sensitively during teaching. As Hunt and Paraskevopoulos

166

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

(1980) suggested, underestimation of a child’s ability might be associated with different parental expectations, and this might result in different parenting behaviors and different child outcomes. They also suggested that parents who underestimate their children’s abilities might provide little stimulation to evoke their child’s learning potential. Our finding further supports Hunt and Paraskevopoulos’ finding that different dimensions of maternal knowledge (correct estimation and underestimation) predict parenting behaviors differently. To our knowledge, our study is the first to explore this issue during infancy. Because of the general lack of evidence in this area, future research should further explore this link. Multivariate analyses revealed that the associations between maternal knowledge of child development and parenting behaviors varied by the observational contexts and by racial/ethnic groups. Surprisingly, we found that Hispanic mothers who had higher underestimated scores provided a better quality of home environment and of mother–child interaction during play. In addition, we found that among African-American and Hispanic mothers, greater correct knowledge of child development was associated with lower quality of mother–child interactions during the teaching tasks (a highly structured interaction context). There are at least two possible explanations for the unexpected direction of these findings. It is possible that different mechanisms may account for the different associations between knowledge and parental behavior among African-American, Hispanic and White subgroups. It may be that African-American and Hispanic mothers with higher correct or underestimated knowledge of child development have different expectations of or goals regarding their children’s capabilities and what they as parents can do to serve as their child’s teacher in different contexts. Two findings may support this hypothesis. First, follow-up analyses of specific behaviors indicated that for Hispanic mothers, greater knowledge of child development was associated with spending more time in teaching and expressing less positive affect (e.g., caresses, hugs, kiss, making cheerleading-type statements); and greater underestimations of children’s abilities was associated with higher involvement with their child in the form of more structured child’s play and greater encouragement of developmental advance, for example, during the interaction. It may be that Hispanic mothers with greater knowledge of child development have higher expectations or goals about what their child could or should learn during a teaching task. Therefore, they were more serious about their children’s learning opportunity; and behaved less playful during the teaching situation. Second, greater knowledge of child development was associated with choosing more difficult tasks in teaching and focusing the child’s attention more on the task for African-American mothers (e.g., not allowing non-task manipulation of the task material, and not shifting child’s attention to other tasks even though the child was distressed by the task, explicit focus on the task). These results suggest that other parental cognitive factors, such as cultural values, parental goals about educational practices, or beliefs about the parental role in child development, may underlie or interact with parental knowledge, leading to differences in parental teaching behavior (Sigel & McGillicuddy-De Lisi, 2002). However, these results should be interpreted cautiously as some of the associations did not achieve the customary level of statistical significance. Another possible explanation for the observed relationships between maternal knowledge of child development and parenting behavior may be that parenting measures are not culturally appropriate for African-American and/or Hispanic population. Although we chose widely used and appropriate observational instruments and coding schemes for this study, there was limited

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

167

documentation available on whether the instruments were racially/ethnically sensitive and appropriate. For example, the validity of the NCAST for different racial groups has never been carefully examined. Studies (including the present study) have found that after controlling for demographic factors, Hispanic mothers have significantly lower NCAST parent subscale scores compared to White mothers (Sumner & Spietz, 1994), suggesting that the NCAST definition or expression of bhigh-quality mother–child interactionQ may differ among different racial/ethnic groups. As a result, the NCAST may not capture the representative bsensitive parentingQ concept for varying ethnic groups. In addition, Julian, McKenry, and McKelvey’s (1994) found that White mothers used praise significantly more than Hispanic mothers and put less emphasis on obeying during mother–child interaction. Their findings suggest that praising and following a child’s lead may not be the characteristics of bsensitive/high-quality parentingQ for some non-White mothers. Such evidence suggests the possibility that the NCAST may not be a culturally sensitive or appropriate instrument. Because few studies have explored the appropriateness of using NCAST or P/CIS among different racial/ethnic groups, more research should be conducted to further explore validity issues, such as testing whether bsensitivityQ based on White samples results in similar child outcomes in minority samples. Also, future research should explore the reasons that may contribute to the negative relationships between parental knowledge and quality of interaction among different racial groups, and examine how maternal knowledge of child development interacts with maternal goals of childrearing among different racial/ethnic groups. When interpreting these results, it is important to note the potential bias inherent in the sample and measures. First, as reported earlier, mothers who did not participate in the home visit were more likely to be younger, single, less well-educated, and poorer than mothers who participated. As previous research indicates, these high-risk mothers may have lower knowledge of child development and may show less sensitivity during mother–child interactions. Thus, the associations between maternal knowledge and parenting behaviors that we found are conservative. In addition, we measured only one domain of maternal knowledge of child development—the number of developmental milestones a mother could correctly or incorrectly estimate (or underestimate). This is only one of several important maternal knowledge components that may influence parenting behavior. Other domains of maternal knowledge such as knowledge about developmental processes and effective parenting strategies need to be systematically investigated in the future. Moreover, the measures of maternal knowledge and parenting behavior were one and half year apart. This time-lag difference might weaken the association between maternal knowledge and parenting behavior. Finally, our study was limited to exploring parenting behaviors in terms of the quality of the mother–child interaction and home environment. The associations between maternal knowledge of child development and other parenting behaviors, such as use of discipline strategies, may not be the same. Therefore, future research should also explore other domains of parenting behaviors on which parental knowledge may have an impact. This study tested the generalizability of a developmental model of knowledge leading to four different parenting behaviors among three racial/ethnic groups. The findings lend some insight into racial/ethnic variations in parenting. Several implications can be drawn from this study. First, the

168

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

present findings suggest that providing mothers with accurate information about child development in prevention/intervention programs is not a bquick fixQ to improve parenting behavior. Maternal knowledge of child development may interact with other domains of maternal cognition (e.g., maternal childrearing beliefs, goals of parenting) in influencing parenting behavior, and this interaction may vary by race/ethnicity. Therefore, when designing prevention/intervention programs, practitioners must consider the differences among racial/ethnic groups in terms of knowledge of child development and beliefs about supportive parenting practices. Prevention/intervention programs need to teach mothers both accurate knowledge of development and effective ways of parenting that are consistent with their beliefs. The findings of the present study can offer important questions for future research. Further work is needed to explore the association between different dimensions of parental knowledge about child development and other indices of parenting behaviors. Also, further work is needed to explore how maternal knowledge of child development interacts with maternal childrearing beliefs (including the role of parents in helping children to learn skills, what skills parents should teach their children before they enter school, and positive or effective strategies to be used in childrearing) and parenting behavior among different race/ethnic groups. References Barnard, K. E. (1978). Nursing child assessment satellite training: Learning resource manual. Seattle7 University of Washington. Belsky, J. (1980). Child maltreatment: An ecological integration. American Psychologist, 35, 320 – 335. Benasich, A. A., & Brooks-Gunn, J. (1996). Maternal attitudes and knowledge of child-rearing: Associations with family and child outcomes. Child Development, 67, 1186 – 1205. Bornstein, M. H., Tamis-LeMonda, C. S., Pascual, L., Haynes, O. M., Painter, K., Galperin, C., et al. (1996). Ideas about parenting in Argentina, France, and the United States. International Journal of Behavioural Development, 19, 347 – 367. Bradley, R. H., & Caldwell, B. M. (1984). 174 children: A study of the relationship between home environment and cognitive development during the first 5 years. In A. W. Gottfried (Ed.), Home environment and early cognitive development (pp. 5 – 56). Orlando7 Academic Press. Brooks-Gunn, J., & Furstenberg, F. F. (1986). The children of adolescent mothers: Physical, academic, and psychological outcomes. Developmental Review, 6, 781 – 790. Caughy, M. O., Miller, T. L., Genevro, J. L., Huang, K., & Nautiyal, C. (2003). The effects of Healthy Steps on disciplines strategies of parents with toddlers. Journal of Applied Developmental Psychology, 24, 517 – 534. Conrad, B., Gross, D., Fogg, L., & Ruchala, P. (1992). Maternal confidence, knowledge, and quality of mother-toddler interactions: A preliminary study. Infant Mental Health Journal, 13, 353 – 362. Cote, L. R., & Bornstein, M. H. (2001). Social and didactic parenting behaviors and beliefs among Japanese American and South American mothers of infants. Infancy, 1, 363 – 374. Damast, A. M., Tamis-LeMonda, C. S., & Bornstein, M. H. (1996). Mother–child play: Sequential interactions and the relation between maternal beliefs and behaviors. Child Development, 67, 1752 – 1766. Dichtelmiller, M., Meisels, S. J., Plunkett, J. W., Bozynski, M. E., Claflin, C., & Mangelsdorf, S. C. (1992). The relationship of parental knowledge to the development of extremely low birth weight infants. Journal of Early Intervention, 16, 210 – 220. Dukewich, T. L., Borkowski, J. G., & Whitman, T. L. (1996). Adolescent mothers and child abuse potential: An evaluation of risk factors. Child Abuse and Neglect, 20, 1031 – 1047. Farran, D. C., Kasari, C., Upder, P., Jarber, L., Huntington, G., & Comfort, M. (1987). Rating mother–child interactions in handicapped and at-risk infants. In D. Tamir (Ed.), Stimulation and intervention in infant development (pp. 297 – 312). London7 Freund. Fewell, R. R., & Wheeden, C. A. (1998). A pilot study of intervention with adolescent mothers and their children: A preliminary examination of child outcomes. Topics in Early Childhood Special Education, 18, 18 – 26.

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

169

Fry, P. S. (1985). Relations between teenagers’ age, knowledge, expectations and maternal behavior. British Journal of Developmental Psychology, 3, 47 – 55. Goodnow, J. J. (1988). Parents’ ideas, actions, and feelings: Models and methods from developmental and social psychology. Child Development, 59, 286 – 320. Goodnow, J. J. (1996). From household practices to parent’s idea about work and interpersonal relationships. In S. Harkness, & C. M. Super (Eds.), Parents’ cultural belief systems: Their origins, expressions, and consequences (pp. 333 – 357). New York7 The Guilford Press. Goodnow, J. J. (2002). Parents’ knowledge and expectations: Using what we know. In M. H. Bornstein (Ed.), Handbook of parenting: Vol. 3. Being and becoming a parent (2nd ed., pp. 439 – 460). Mahwah, NJ7 Erlbaum. Guyer, B., Hughart, N., Strobino, D., Jones, A. S., & Scharfstein, D. (2000). Assessing the impact of pediatric-based, developmental services on infants, families and clinicians: Challenges to evaluating the Healthy Steps Program. Pediatrics, 105, E33. Hammond-Ratzlaff, A., & Fulton, A. (2001). Knowledge gained by mothers enrolled in a home visitation program. Adolescence, 36, 435 – 442. Holden, G. W. (1995). Parental attitudes toward childrearing. In M. H. Bornstein (Ed.), Handbook of parenting: Vol. 3. Being and becoming a parent (2nd ed., pp. 359 – 392). Mahwah, NJ7 Erlbaum. Hunt, J. M., & Paraskevopoulos, J. (1980). Children’s psychological development as a function of the inaccuracy of their mothers’ knowledge of their abilities. Journal of Genetic Psychology, 136, 285 – 298. Julian, T. W., McKenry, P. C., & McKelvey, M. W. (1994). Cultural variations in parenting: Perceptions of Caucasian, AfricanAmerican, Hispanic, and Asian-American parents. Family Relations: Interdisciplinary Journal of Applied Family Studies, 43, 30 – 37. Karraker, K. H. (1996). Adolescent mothers’ knowledge of child development and expectations for their own infants. Journal of Youth and Adolescence, 25, 651 – 666. MacPhee, D. (1981). Knowledge of infant development inventory manual. Chapel Hill, NC7 Department of Psychology, University of North Carolina. McGillicuddy-DeLisi, A. V. (1982). Parental beliefs about developmental processes. Human Development, 25, 192 – 200. McGillicuddy-DeLisi, A. V., & Sigel, I. E. (1995). Parental beliefs. In M. H. Bornstein (Ed.), Handbook of parenting: Vol. 3. Status and social conditions of parenting (2nd ed., pp. 333 – 358). Hillsdale, NJ7 Lawrence Erlbaum Associates. Miller, S. A. (1988). Parents’ beliefs about their children’s cognitive development. Child Development, 59, 259 – 285. Minkovitz, C. S., Hughart, N., Strobino, D., Scharfstein, D., Grason, H., Hou, W., et al. (2003). A practice-based intervention to enhance quality of care in the first 3 years of life: The Healthy Steps for young children program. JAMA, 290, 3081 – 3091. Myers, B. J. (1982). Early intervention using Brazelton training with middle-class mothers and fathers of newborns. Child Development, 53, 462 – 471. Parks, P. L., & Smeriglio, V. L. (1986). Relationships among parenting knowledge, quality of stimulation in the home and infant development. Family Relations, 35, 411 – 416. Radloff, L. S. (1977). The CES-D Scale: A self-report depression scale for research in the general population. Applied Psychological Measurement, 1, 385 – 401. Savage, S. L., & Gauvain, M. (1998). Parental beliefs and children’s everyday planning in European-American and Latino families. Journal of Applied Developmental Psychology, 19, 319 – 340. Sigel, I. E. (1992). The belief–behavior connection: A resolvable dilemma? In I. E. Sigel, A. V. McGillicuddy-DeLisi, & J. J. Goodnow (Eds.), Parental belief systems: The psychological consequences for children (2nd ed., pp. 433 – 456). Hillsdale, NJ7 Erlbaum. Sigel, I. E., & Kim, M. (1996). The answer depends on the question. In S. Harkness, & C. M. Super (Eds.), Parents’ cultural belief systems: Their origins, expressions, and consequences (pp. 83 – 120). New York7 The Guilford Press. Sigel, I. E., & McGillicuddy-De Lisi, A. V. (2002). Parent beliefs are cognitions: The dynamic belief systems model. In M. H. Bornstein (Ed.), Handbook of parenting: Vol. 3. Being and becoming a parent (2nd ed., pp. 485 – 508). Mahwah, NJ7 Erlbaum. Smith, T. K. (2002). The relationship between knowledge, attributions and behavior in adolescent mothers: Implications for child outcomes. Dissertation Abstracts International. B. The Sciences and Engineering, 62(11-B), 5412. Stevens, J. H. (1984). Child development knowledge and parenting skills. Family Relations, 33, 237 – 244. Stoiber, K. C., & Houghton, T. G. (1993). The relationship of adolescent mothers’ expectations, knowledge, and beliefs to their young children’s coping behavior. Infant Mental Health Journal, 14, 61 – 79.

170

K.-Y. Huang et al. / Applied Developmental Psychology 26 (2005) 149–170

Sumner, G., & Spietz, A. (1994). NCAST: Caregiver/parent–child interaction teaching manual. Seattle, WA7 NCAST Publications. Tamis-LeMonda, C. S., Shannon, J., & Spellmann, M. (2002). Low-income adolescent mothers’ knowledge about domains of child development. Infant Mental Health Journal, 23, 88 – 103. Veddovi, M., Kenny, D. T., Gibson, F., Bowen, J., & Starte, D. (2001). The relationship between depressive symptoms following premature birth, mothers’ coping style, and knowledge of infant development. Journal of Reproductive and Infant Psychology, 17, 313 – 323. Zepeda, M., & Espinosa, M. (1988). Parental knowledge of children’s behavioral capabilities: A study of low income parents. Hispanic Journal of Behavioral Sciences, 10, 149 – 159.