Peacock-chambers2017.pdf

The Influence of Parental Self-Efficacy and Perceived Control on the Home Learning Environment of Young Children Elizabeth Peacock-Chambers, MD, MSc; Justin T. Martin, MD; Kelly A. Necastro, MPH; Howard J. Cabral, PhD; Megan Bair-Merritt, MD, MSCE From the Division of General Pediatrics, Boston Medical Center (Drs Peacock-Chambers and Bair-Merritt), Boston University School of Medicine (Dr Martin), Department of Community Health Sciences (Ms Necastro), and Department of Biostatistics (Dr Cabral), Boston University School of Public Health, Boston, Mass The authors have no conflicts of interest to disclose. Address correspondence to Elizabeth Peacock-Chambers, MD, MSc, Baystate Medical Center, 759 Chestnut St, Springfield MA 01199 (e-mail: [email protected]). Received for publication June 1, 2016; accepted October 22, 2016.

ABSTRACT OBJECTIVE: To: 1) examine sociodemographic factors associated with high parental self-efficacy and perceived control, and 2) determine how self-efficacy and control relate to the home learning environment (HLE), including whether they mediate the relationship between sociodemographic characteristics and HLE, among low-income parents of young children. METHODS: Cross-sectional survey of English- and Spanishspeaking parents, 18 years of age and older, with children 15 to 36 months old, to assess parental self-efficacy, perceived control, HLE, and sociodemographic characteristics. Bivariate analysis identified sociodemographic predictors of high self-efficacy and control. Separate multivariate linear regression models were used to examine associations between selfefficacy, control, and the HLE. Formal path analysis was used to assess whether self-efficacy and control mediate the relationship between sociodemographic characteristics and HLE. RESULTS: Of 144 participants, 25% were white, 65% were immigrants, and 35% completed the survey in Spanish. US-born subjects, those who completed English surveys, or who had

higher educational levels had significantly higher mean self-efficacy and perceived control scores (P < .05). Higher self-efficacy and perceived control were associated with a positive change in HLE score in separate multivariate models (self-efficacy b ¼ .7 [95% confidence interval (CI), 0.5–0.9]; control b ¼ .5 [95% CI, 0.2–0.8]). Self-efficacy acted as a mediator such that low self-efficacy explained part of the association between parental depressive symptoms, immigrant status, and less optimal HLE (P ¼ .04 and < .001, respectively). CONCLUSIONS: High parental self-efficacy and perceived control positively influence HLEs of young children. Self-efficacy alone mediates the relationship between parental depressive symptoms, immigrant status, and less optimal early home learning.

KEYWORDS: early childhood development; home learning; parenting; poverty; self-efficacy ACADEMIC PEDIATRICS 2017;17:176–183

WHAT’S NEW

might buffer against the adverse effects of poverty on child development.5 Despite the strong association between poverty, minority status, and a less enriching home learning environment (HLE) at the population level, many low-income families do provide stimulating home environments for their children.6,7 Less is known about parent beliefs and attitudes associated with positive home environments among families who face adversity. Understanding parental perceptions of their ability to affect their young child’s development might have implications for improving the HLE. Existing theory and research propose that 2 factors, parental self-efficacy and perceived control, might critically affect a parent’s ability to promote their child’s development. Parental self-efficacy (hereafter referred to as “self-efficacy”) is defined as parent knowledge and confidence to effectively manage the tasks and situations of

Immigrant status was a strong predictor of lower parental self-efficacy among a sample of low-income parents with children aged 15 to 36 months. Parental self-efficacy mediated the relationship between depressive symptoms, immigrant status, and lower-quality home learning environments.

COMPARED WITH THEIR more advantaged peers, children living in poverty are at greater risk of developmental delays, specifically in domains related to language, executive functioning, and social-emotional development.1,2 However, substantial evidence shows that positive parent–child relationships3 and a stimulating home environment during the first 3 years of life can significantly promote optimal brain growth, healthy development,4 and ACADEMIC PEDIATRICS Copyright ª 2016 by Academic Pediatric Association

176

Volume 17, Number 2 March 2017

ACADEMIC PEDIATRICS

PARENTAL SELF-EFFICACY AND EARLY LEARNING

parenthood.8,9 Self-efficacy, which has been primarily studied in middle class parents of school aged children, is predictive of positive parenting practices and child outcomes10 including greater parental academic involvement,5,11 fewer child behavioral problems12 and improved school readiness.13 Furthermore, higher self-efficacy is associated with more responsive caregiving practices in high-risk mothers of infants (ie, teenage mothers and those with depression).14 The relationship of self-efficacy to tasks that promote social and cognitive development in the home has not been well studied, particularly among low-income families with children younger than age 5. Parental perceived control (hereafter referred to as “control”) is defined as parent beliefs about their personal effect on a specific outcome, in this case their child’s development and learning.15 Few studies have examined the association between control and parent–child relationships16 or developmental outcomes,12 revealing important gaps in the literature. Despite the lack of data on control and child outcomes, control might be a particularly important factor to assess among families disempowered by structural and sociocultural barriers such as racism and exposure to trauma.17 Although self-efficacy as well as control might influence parent behaviors, previous research shows that they act independently as distinct belief systems.18 For example, a mother might believe in her ability to manage her toddler’s difficult behavior, but might not believe that her actions will affect how the child will behave toward others. Such belief systems also exist in the context of varying social circumstances. Personal histories of violence, inadequate access to housing and food, or lack of social support might lead to generally lower perceptions of selfefficacy and control, but it remains unclear the degree to which adverse social circumstances might differentially affect parental beliefs and behaviors that promote positive child development. We therefore conducted a cross-sectional survey in a diverse, low-income population of parents with young children (15–36 months old) to better understand which parent and child factors predict self-efficacy and control and how these perceptions relate to the HLE (the amount of developmentally stimulating activities provided in the home). We hypothesized that older parent age, more education, and fewer depressive symptoms would be predictors of higher levels of self-efficacy and control. We further hypothesized that parents with higher levels of self-efficacy and control would score higher on measures of a stimulating HLE, and that self-efficacy and control would mediate the association between parent factors and the HLE.

Children nutrition centers in the Boston area. These recruitment sites serve a diverse population of US-born and immigrant families from Central America, South America, the Caribbean, Asia, and Africa. Potential participants were excluded if they were younger than 18 years old, or if the index child had a significant medical condition that might affect their development such as autism. Participants gave verbal informed consent to participate and were compensated $10 for their time. Surveys were conducted by research assistants either in person or by phone in the language of the participant’s choice from December 2014 through May 2015. Time for survey completion ranged from approximately 30 minutes to 1 hour. The study was deemed exempt by the authors’ internal review board.

METHODS STUDY DESIGN AND PARTICIPANTS We surveyed English- and Spanish-speaking parents of healthy children aged 15 to 36 months, recruited from an urban primary care clinic and 3 Women, Infants, and

177

MEASURES PARENTAL SELF-EFFICACY We used 2 subscales (16 questions) from the Self-Efficacy for Parenting Tasks Index–Toddler Scale8 to assess self-efficacy in 2 domains that are critical to early childhood learning: empathetic responsiveness and teaching (cognitive stimulation).19 For example, one item states: “Helping my child learn colors, names of objects, etc, is not one of my strongest points.” Participants responded to statements on a 6-point Likert scale from 1 ¼ strongly disagree to 6 ¼ strongly agree (total range, 16–96). In previous studies, this tool was validated against the Maternal Efficacy Questionnaire14 and the Parenting Sense of Competence Scale.20 Developers of the tool report a Cronbach a for empathetic responsiveness questions of .71 and .73 for teaching questions.8 In this sample, the Cronbach a for the 16 items was .77 (empathetic responsiveness ¼ .70 and teaching ¼ .73). PARENTAL PERCEIVED CONTROL Six items from the Parent Opinion Survey15 were used to assess parental perception of their personal control over their child’s development, social relationships, and behavior. As an example of the measure, one item states: “I believe that the way I treat my child will strongly influence how he/she will behave toward others.” We modified this instrument by adding 2 similarly worded questions specific to parents’ perceived control over their child’s learning and cognitive development, for a total of 8 items. The final measure is not a validated instrument. Participants responded to statements on a 6-point Likert scale from 1 ¼ strongly disagree to 6 ¼ strongly agree (total range, 8–48). The Cronbach a for these 8 items was .82. HLE The 25-item Home Learning Experiences subscale from the Home-Learning Environment Profile21 was used to assess home activities and experiences that promote early learning. Items are used to assess frequency of specific school readiness activities and common family or cultural practices that support learning and development. Each item is rated on a 4-point Likert scale: 1 ¼ never, 2 ¼ once in a while, 3 ¼ fairly often, 4 ¼ often, or 0 ¼ does not apply

178

PEACOCK-CHAMBERS ET AL

(total range, 0–100). This is a culturally and socioeconomically sensitive measure of home learning with significant correlation to teacher ratings of cognitive ability (r ¼ 0.15; P < .05) and verbal ability on the Peabody Picture Vocabulary Test-Revised (r ¼ 0.16; P < .05).20 The Cronbach a was .86 in this sample. PARENT DEPRESSIVE SYMPTOMS The Patient Health Questionnaire (PHQ)-822 was used to evaluate depressive symptoms in the study participants. The PHQ-8 includes all questions from the original PHQ-9 except for the question of suicidality. The suicidality question was removed on the basis of recommendations from our institutional review board advisors who suggested that we did not have appropriate staff to address needs of actively suicidal participants. The questionnaire is a well validated screening tool on the basis of the symptoms associated with major depressive disorder. A score of $10 has 88% sensitivity and 88% specificity for major depression in a primary care setting.22 The Cronbach a was .89 in a previous study.22 This variable was examined as a continuous as well as a dichotomous variable on the basis of established cutoffs (score <10 vs $ 10). PARENT AND CHILD CHARACTERISTICS A priori sociodemographic variables on the basis of the published literature23 included parent age (years), child age (months), child sex (male/female), and parent education (high school degree or less vs education beyond a high school degree). In addition, we assessed the following parent variables: race (white, black, American Indian, Asian, more than 1 race, or unknown), ethnicity (not Hispanic or Latino, Hispanic or Latino, or unknown), relationship to the index child (mother, father, or other), US-born versus immigrant, survey language (English or Spanish), marital status (married, single/never married, live-in partner, divorced, widowed, separated), social support (“poor,” “fair,” or “good” on the basis of a composite score of 2 questions: “Do you have someone in your life who: understands your problems?” and “could help you care for your child if you were sick?”), employed versus unemployed, and receiving public assistance versus not. Variables with respect to the index child included preterm versus full term, first child versus not, and the relationship to the daily primary caretaker (mother, father, family, babysitter, daycare/preschool, or other). The recruitment site and whether the interview was conducted in person or via phone were also recorded. STATISTICAL ANALYSES Data were analyzed using SAS version 9.3 (SAS Institute Inc, Cary, NC). Data were transferred from paper survey to REDCap and checked by a separate reviewer for data entry error. In the original power calculations we determined a necessary sample size of 152 participants on the basis of the primary outcome measure of a positive screen for child developmental delay (estimated at a prevalence of 20%24; a ¼ .05; power ¼ 0.80). Post hoc power

ACADEMIC PEDIATRICS

calculations for the multiple linear regression models between self-efficacy, control, and the more proximal outcome HLE were both >0.90. We conducted bivariate analyses to examine the associations between parent and child sociodemographic characteristics and self-efficacy, as well as control using a t test, Pearson correlation, and analysis of variance tests as appropriate for the given variable. We tested potentially related variables for collinearity on the basis of their relation in previous literature and the distribution of variables in our sample (self-efficacy and control, marital status and social support, race and ethnicity, language and country of origin) by calculation of the variance inflation factor.25 Using multivariable linear regression models, we examined independent associations between self-efficacy and control (in separate models) on changes in HLE (b equal to unit change in HLE score for every 1 unit change in self-efficacy and control score). Covariates were included in models on the basis of either empirical support or evidence of potential confounding (defined as being associated with predictor as well as outcome variables and if the variable’s removal from the model changed the b point estimate by $ 10%).26 Finally, we used a path model to test the degree to which the association between parent sociodemographic factors and HLE were mediated by self-efficacy and control. Mediators were defined as factors that are on the pathway connecting the predictor variable and the outcome variable. We assessed the interaction between the mediators (self-efficacy and control) and between parent factors and the mediators before creation of the path model. Using path modeling rather than the traditional Baron and Kinney approach has the advantage of being able to simultaneously examine relationships between multiple variables of interest in a single model. Specifically, we compared the model values of Akaike’s Information Criterion with full mediation (no direct effects) with a model that included direct effects. In other words, we compared indirect effects (effects explained by the mediator) with direct effects (effects not explained by the mediator). Because of the lack of consensus on a single index for judging overall model fit, we used 4 indices with the following cutoffs as indicators of reasonable fit: c2 likelihood ratio statistic P $ .05 (an absolute index indicative of residual variance remaining after fitting of the model), Bentler normed comparative fit index (CFI) $ 0.95,27 Tucker-Lewis non-normed index (TLI) $ 0.90,28 and root mean square error of approximation (RMSEA) # 0.08.29 We used Mplus version 7.31 (Muthen & Muthen, Los Angeles, CA) in conducting the path analyses.

RESULTS Overall, the sample was racially and ethnically diverse with one-quarter of participants who identified as white and more than half born outside of the United States (Table 1). The mean child age was 25 months and mean parent age was 30.9 years. More than half of the sample had a high school degree or less, nearly half (44%) were

ACADEMIC PEDIATRICS

PARENTAL SELF-EFFICACY AND EARLY LEARNING

Table 1. Participant Demographic Characteristics (N ¼ 144)

Table 3. Association Between Parent Demographic Characteristics, Self-Efficacy, and Perceived Control (N ¼ 144)

Characteristic

Value

Child Age, months Born premature (<37 weeks) Male sex First child Childcare provider Mother Father Family Babysitter Daycare Other Parent Age, years Caregiver relationship to child Mother Father Other Race White Black or African American American Indian/Alaska Native/Asian More than 1 race Unknown/not reported Ethnicity Not Hispanic or Latino Hispanic or Latino Unknown/not reported Born in United States and territories Survey completed in English Marital status Married Single/never married Live-in partner Divorced Widowed Separated Level of support “good” High school degree or less Unemployed Receiving public assistance Depressive symptom score continuous (PHQ-8) Depressive Symptom Score < 10

25.0 (6.7) 14.9 55.6 38.2 51.4 6.9 12.5 5.6 22.2 1.4 30.9 (6.7) 85.4 12.5 2.1 25.0 36.1 3.5 4.9 29.9 43.1 52.8 4.2 34.7 65.3 36.1 30.6 25.0 2.8 0.0 5.6 90.3 61.8 43.8 81.9 3.8 (4.4) 87.5

PHQ indicates Patient Health Questionnaire. Data are presented as mean (SD) or %.

unemployed, and 82% received public assistance. Distribution of self-efficacy, control, and HLE scores are presented in Table 2. Tests of collinearity between self-efficacy and control, as well as language and immigrant status were negative whereas race and ethnicity, and marital status and social support exhibited some collinearity. Table 2. Distribution of Parental Self-Efficacy, Perceived Control, and HLE Scores (N ¼ 144)

Self-efficacy Perceived control HLE score

Mean (SD)

Median

Range

80.6 (9.5) 37.7 (7.8) 71.6 (13.4)

82 39 73

48–96 16–48 32–96

HLE indicates home learning environment. Possible range for self-efficacy: 16–96; perceived control: 8–48; and HLE: 0–100.

Parent Characteristic Race White Black or African American American Indian Asian More than 1 race Unknown/not reported Ethnicity Non-Hispanic or Latino Hispanic/Latino Unknown/not reported Survey language English Spanish US-born Immigrant Marital status Married Single Live-in partner Divorced Separated Education High school degree or less More than high school Public assistance Yes No

Mean Self-Efficacy

P

Mean Perceived Control

P

.30

.22

81.7 81.7

37.8 38.5

67.0 80.8 82.4 78.3

40.0 42.8 41.0 35.6 .27

81.9 70.4 82.2

.03 39.5 36.1 39.3 <.0001

.002 82.4 77.3 84.7 78.4

<.0001

39.9 33.7 41.8 35.6

<.0001

.02 80.8 82.2 80.4 83.8 70.3

.31 38.0 38.9 36.7 40.5 33.3

.04 79.4

.03 36.6

82.6

39.5 .20

80.1 82.8

0.15

.12 37.3 39.9

Correlation Coefficient Depressive symptoms

179

Correlation Coefficient .07

0.004

.96

Possible range for parental self-efficacy: 16–96; parental perceived control: 8–48.

In unadjusted models (Table 3), survey language and country of origin exhibited the strongest association of any parent characteristic with self-efficacy and control. English-speaking participants and those born in the United States had significantly higher self-efficacy and control scores than immigrants or Spanish speakers (P < .005). However, ethnic identification as non-Hispanic or Latino was only associated with higher perceived control (P ¼ .03), not self-efficacy. Additionally, participants who completed post-high school education had higher mean self-efficacy and control scores (P < .05). Participants with a greater number of depressive symptoms had lower mean self-efficacy scores (P ¼ .07) and those who identified as “separated” from their partners scored nearly 10 points lower, on average, on the self-efficacy scale than married, single, or divorced participants (P ¼ .02). Control scores were not significantly different on the basis of marital status or depressive symptoms. In the first adjusted linear regression model (Table 4), a higher level of self-efficacy was associated with significantly higher HLE scores (unstandardized b ¼ .7 [95% confidence interval, 0.5–0.9; P < .001; R2 ¼ 0.26),

PEACOCK-CHAMBERS ET AL

180

ACADEMIC PEDIATRICS

Table 4. Separate Multivariable Linear Regression Models: PSE and PPC as Predictors of the HLE (N ¼ 144) Variable Self-efficacy* Unstandardized Standardized Perceived control† Unstandardized Standardized

Adjusted b (95% CI)

Adjusted R2

P

0.7 (0.5–0.9) 6.6 (4.7–8.6)

0.26

<.0001

0.5 (0.2–0.8) 4.0 (1.6–6.2)

0.12

.0005

PSE indicates parental self-efficacy; PPC, parental perception of control; HLE, home learning environment; CI, confidence interval; and HS, high school. *PSE adjusted for child age, child sex, parent age, parent education (HS degree or less vs not), parent depression (continuous), and interview type (in person vs phone). †PPC adjusted for child age, child sex, parent age, parent education, parent depression, primary language, and interview type.

controlling for covariates. In the second adjusted model used to examine the association between control and HLE, higher control scores were also associated with significantly higher HLE scores (unstandardized b ¼ .5 [95% confidence interval, 0.2–0.8]; P < .001; R2 ¼ 0.12). In the path model we examined the degree to which self-efficacy and control fully or partially mediated the association between the 4 significant parent variables (education, depressive symptoms, country of origin, and language) and HLE in a combined model. In this model (Fig, a), depressive symptoms had a significant indirect effect on HLE through self-efficacy (P ¼ .04) but no significant direct effect on HLE. US-born versus immigrant status also had a significant indirect effect on HLE through self-efficacy (P < .01), but no significant direct effect on HLE. In other words, the model did not detect a direct effect between parent variables and HLE after controlling for mediation by parental self-efficacy. Control was not significantly associated with HLE in this model, which included self-efficacy. Overall, the model had a poor fit (c2 ¼ 32.0; P < .01; CFI ¼ 0.77; TLI ¼ 0.18; RMSEA ¼ 0.19). A refined model (Fig, b), including only the indirect effects of depressive symptoms, country of origin, and language on HLE through self-efficacy, met criteria for a model with reasonable fit (c2 ¼ 0.91; P ¼ .82; CFI ¼ 1.00; TLI ¼ 1.11; RMSEA < 0.01), again with significant indirect effects of depressive symptoms and US-born versus immigrant status on HLE through self-efficacy.

DISCUSSION This study showed an association between parent characteristics, including US-born status and higher educational attainment, and higher levels of self-efficacy and perceived control. When adjusting for covariates in separate models, we found high self-efficacy and control to be associated with a more enriching HLE. In a combined model of key parent characteristics predictive of the HLE, we found that self-efficacy alone acted as a mediator of the association between fewer parental depressive symptoms, US-born status, and a more optimal HLE.

Our findings regarding predictors of self-efficacy and control are consistent with previous studies; however, we also identified predictors not previously reported. Parental education, marital status, and depressive symptoms are supported as predictors of self-efficacy in the extant literature.14,15,17,23 We found that immigrant status and completing the survey in Spanish were among the strongest predictors of low self-efficacy and control in this predominantly low-income sample. Few studies have examined the association between immigrant status and self-efficacy or control with a comparison group of USborn parents.30 Furthermore, studies that did examine parenting among immigrant populations did not measure parental self-efficacy in domains specific to cognitive or social–emotional development of children. An example question of general parental self-efficacy would be: “I am able to take care of my child’s needs.”15 Other studies focused on outcomes such as discipline and parenting styles (eg, authoritative, authoritarian, or permissive), rather than promotion of home learning activities.11,31 Our findings in this sample of low-income parents are consistent with the previous study by Machida et al,15 who reported an association between self-efficacy and the quality of the HLE among parents of children in Head Start (mean age, 55 months) in California and Arizona. However, our results differ in the examination of this association as it relates to parent factors, such as ethnicity and country of origin. Machida et al reported no difference in this association according to ethnicity (Anglo American vs Mexican American). The study did not explore differences according to country of origin (59% of Mexican American mothers in their sample were born in Mexico) and excluded African American, Native American, and biracial children from analyses of ethnic group comparisons. The sites of recruitment for our study serve immigrants from numerous countries, which allowed for exploration of associations according to immigrant status and language. Both studies reported a lack of association between control and HLE in a combined path model with self-efficacy. These consistent findings support the conclusion that perceived control did not maintain an independent association with the HLE when controlling for self-efficacy. In additional analyses (data not shown), we found no evidence of interaction between these 2 factors. The reason for this consistent finding remains unclear and requires further study in high as well as low-income populations. Self-efficacy is a key construct of Social Cognitive Theory32 and we show that the “risk” imposed on the HLE by certain parent characteristics (ie, depression and immigrant status) is partially explained by low self-efficacy. In other words, a risk factor such as depression or immigrant status must first change a person’s confidence in their ability as a parent to affect the quality of the HLE. The implications of such a finding further suggest that without affecting a person’s confidence in their ability to execute tasks that promote their child’s development, we are unlikely to improve the quality of the HLE among high-risk families.

ACADEMIC PEDIATRICS

PARENTAL SELF-EFFICACY AND EARLY LEARNING

181

A Depressive symptoms

Parental Self-efficacy

-.34*

.59**

-0.1

6.2** USA vs. Foreign born

-2.1 3.0

4.6**

English vs. Spanish

Home Learning Environment .17

Parental Perceived Control

3.6**

Education

1.6 X2 (5, N=144) 32.0, p < 0.01 CFI = 0.77 TLI = 0.18 RMSEA = 0.19

B Depressive symptoms -.33* USA vs. Foreign born

4.98**

Parental Self-efficacy

.60**

Home Learning Environment

3.00

English vs. Spanish X2 (3, N=144) 0.91, p = 0.82 CFI = 1.00 TLI = 1.11 RMSEA < 0.01 Figure. (a) Path analysis model on parental self-efficacy and perceived control as mediators of the relationship between sociodemographic factors and home learning environment. *P < .05; **P < .01. (b) Refined path analysis model on parental self-efficacy and perceived control as mediators of the relationship between sociodemographic factors and home learning environment. *P < .05; **P < .01.

A number of studies have explored the association between immigrant status and the quality of the HLE and reported that immigrant families engage in less booksharing,7 less verbal responsiveness, and ultimately have lower rates of school readiness.6 Few studies, however, examined the mechanism underlying these associations representing an area in need of additional study. The “home learning gap” has been attributed mostly to limited knowledge of typical early childhood development among immigrant families33 but Social Cognitive Theory suggests that the acquisition of knowledge alone is unlikely to change a person’s behavior.32 Developing programs for immigrants that focus on improving self-efficacy in domains related to promotion of child cognitive development might be particularly useful for parents with high expectations for

their child’s educational achievement, but limited knowledge and confidence in their own ability to promote their young child’s development. IMPLICATIONS FOR PRIMARY CARE A number of parent- or patient-centered programs have been proven to successfully change patient or parent selfefficacy in primary care settings. Motivational interviewing is 1 example of an alternative communication strategy designed to intentionally change patient self-efficacy related to health behaviors.34 Parenting interventions, including Video Interaction Project, Incredible Years, and Triple P, also show greater improvements in parental self-efficacy compared with standard well-child care.35 Dissemination of these programs or identification of the key components

182

PEACOCK-CHAMBERS ET AL

of interventions that maximize support of self-efficacy among low-income families and incorporation into regular pediatric services could improve access to these evidencebased strategies among a wider population. LIMITATIONS This study has several limitations as a preliminary study of self-efficacy and perceived control in low-income parents of young children. With respect to our measures, HLE was measured by self-report and therefore subject to bias. For the measure of control the tool was not validated because we added 2 questions related to control over child learning. For the measure of self-efficacy we used 2 subscales of a larger questionnaire, although these subscales have been evaluated and used separately in other studies. The cross-sectional study design does not provide information on self-efficacy and control as dynamic variables over time and limits inferences of causality. For example, we cannot prove the directionality of the association between depressive symptoms and self-efficacy, limiting the interpretation of the path model. Longitudinal studies would help clarify causality in a path model. The relatively small sample size and socioeconomically homogenous sample limits generalizability to a more nationally representative population. We also did not collect data on country of origin, exposure to trauma, substance use, or incarceration, which are important areas of future study. Additional exploration of immigrant self-efficacy will require validation of measures in immigrant populations. Our measures might underestimate self-efficacy in immigrant populations, or might be influenced by other factors, such as emotional distress or general response styles (eg, reflective, overconfident, or defensive).

CONCLUSION High parental self-efficacy and perceived control are associated with positive and stimulating HLEs among vulnerable families with young children. Self-efficacy acts as a mediator of the association between certain parent characteristics (immigrant status and depressive symptoms) and quality of the HLE. Future studies are needed to understand the dynamic nature of parental self-efficacy in the first 3 years of a child’s life, and explore whether changes in the delivery of routine pediatric care can effectively affect the self-efficacy of vulnerable families with young children.

ACKNOWLEDGMENTS Financial disclosure: This work was supported by the Academic Pediatrics Association Bright Futures Young Investigators Award (U04MC07853), and Health Resources and Services Administration T32 HP10028 (IO#9940000730).

REFERENCES 1. Farah MJ, Shera DM, Savage JH, et al. Childhood poverty: specific associations with neurocognitive development. Brain Res. 2006; 1110:166–174. 2. Evans GW, Kim P. Childhood poverty, chronic stress, self-regulation, and coping. Child Dev Perspect. 2013;1:43–48.

ACADEMIC PEDIATRICS 3. Walker SP, Wachs TD, Grantham-McGregor S, et al. Inequality in early childhood: risk and protective factors for early child development. Lancet. 2011;378:1325–1338. 4. Landry SH, Smith KE, Swank PR, et al. The effects of a responsive parenting intervention on parent-child interactions during shared book reading. Dev Psychol. 2012;48:969–986. 5. Engle PL, Black MM. The effect of poverty on child development and educational outcomes. Ann N Y Acad Sci. 2008;1136:243–256. 6. Han WJ, Lee R, Waldfogel J. School readiness among children of immigrants in the US: evidence from a large national birth cohort study. Child Youth Serv Rev. 2012;34:771–782. 7. Festa N, Loftus PD, Cullen MR, et al. Disparities in early exposure to book sharing within immigrant families. Pediatrics. 2014;134: e162–e168. 8. Coleman PK, Karraker KH. Maternal self-efficacy beliefs, competence, in parenting, and toddlers’ behavior and developmental status. Infant Mental Health J. 2003;24:236–248. 9. Bandura A. Self-efficacy: toward a unifying theory of behavioral change. Psychol Rev. 1977;84:191–215. 10. Sanders MR, Woolley ML. The relationship between maternal selfefficacy and parenting practices: implications for parent training. Child Care Health Dev. 2005;31:65–73. 11. Pelletier J, Brent JM. Parent participation in children’s school readiness: the effects of parental self-efficacy, cultural diversity and teacher strategies. Int J Early Childhood. 2002;34:45–60. 12. Jones TL, Prinz RJ. Potential roles of parental self-efficacy in parent and child adjustment: a review. Clin Psychol Rev. 2005;25:341–363. 13. Theodille V. The Effects of Parental Self-Efficacy on the Parental Involvement and on Pre-School Aged Children’s Readiness for the Transition to Kindergarten. Auburn, Alabama: Human Development and Family Studies, Auburn University; 2013. 14. Teti DM, Gelfand DM. Behavioral competence among mothers of infants in the first year: the mediational role of maternal self-efficacy. Child Dev. 1991;62:918–929. 15. Machida S, Taylor AR, Kim J. The role of maternal beliefs in predicting home learning activities in Head Start families. Fam Relat. 2002; 51:176–184. 16. Guzell JR, Vernon-Feagans L. Parental perceived control over caregiving and its relationship to parent-infant interaction. Child Dev. 2004;75:134–146. 17. Coleman PK, Karraker KH. Self-efficacy and parenting quality: findings and future applications. Dev Rev. 1997;18:47–85. 18. Wells-Parker E, Miller DI, Topping JS. Development of control-ofoutcome scales and self-efficacy scales for women in four life roles. J Pers Assess. 1990;54:564–575. 19. Bradley RH, Caldwell BM. The relation of infants’ home environments to achievement test performance in first grade: a follow-up study. Child Dev. 1984;55:803–809. 20. Johnston C, Mash EJ. A measure of parenting satisfaction and efficacy. J Clin Child Psychol. 1989;18:167–175. 21. Heath RW, Levin PF, Tibbetts KA. “Development of Home Learning Environment Profile.” Coming Home to Pre-School: The Sociocultural Context of Early Education. Norwood, NJ: Ablex; 1993. 22. Kroenke K, Spitzer RL, Williams JB. The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16:606–613. 23. Sevigny PR, Loutzenhiser L. Predictors of parenting self-efficacy in mothers and fathers of toddlers. Child Care Health Dev. 2010;36: 179–189. 24. Grossman DS, Mendelsohn AL, Tunik MG, et al. Screening for developmental delay in high-risk users of an urban pediatric emergency department. Pediatr Emerg Care. 2010;26:793–797. 25. Belsley DA, Kuh E, Welsch RE. “Detecting and Assessing Collinearity.” Regression Diagnostics: Identifying Influential Data and Sources of Collinearity. New York: John Wiley & Sons; 1980. 26. Weng HY, Hsueh YH, Messam LL, et al. Methods of covariate selection: directed acyclic graphs and the change-in-estimate procedure. Am J Epidemiol. 2009;169:1182–1190. 27. Bentler PM. Comparative fit indexes in structural models. Psychol Bull. 1990;107:238–246.

ACADEMIC PEDIATRICS

PARENTAL SELF-EFFICACY AND EARLY LEARNING

28. Tucker LR, Lewis C. A reliability coefficient for maximum likelihood factor analysis. Psychometrika. 1973;38:1–10. 29. Steiger JH. Structural model evaluation and modification: an interval estimation approach. Multivariate Behav Res. 1990;25:173–180. 30. Izzo C, Weiss C, Shanahan T, et al. Parental self-efficacy and social support as predictors of parenting practices and children’s socioemotional adjustment in Mexican immigrant families. J Prev Intervent Commun. 2008;20:197–213. 31. Dumka LE, Gonzales NA, Wheeler LA, et al. Parenting self-efficacy and parenting practices over time in Mexican American families. J Fam Psychol. 2010;24:522–531.

32. Bandura A. Human agency in social cognitive theory. Am Psychol. 1989;44:1175–1184. 33. Bornstein MH, Cote LR. “Who is sitting across from me?” Immigrant mothers’ knowledge of parenting and children’s development. Pediatrics. 2004;114:e557–e564. 34. Ridner SL, Ostapchuk M, Cloud RN, et al. Using motivational interviewing for smoking cessation in primary care. South Med J. 2014; 107:314–319. 35. Cates CB, Weisleder A, Mendelsohn AL. Mitigating the effects of family poverty on early child development through parenting interventions in primary care. Acad Pediatr. 2016;16(3 suppl):S112–S120.

183