Infant Behavior & Development 33 (2010) 101–110

Contents lists available at ScienceDirect

Infant Behavior and Development

A comparison of maternal sensitivity and verbal stimulation as uniquepredictors of infant social–emotional and cognitive development

Melissa Page ∗, Mari S. Wilhelm, Wendy C. Gamble, Noel A. CardUniversity of Arizona, 1717 E. Speedway, #1101, Tucson, AZ 85719, United States

a r t i c l e i n f o

Article history:Received 14 August 2008Received in revised form15 September 2009Accepted 24 December 2009

Keywords:Maternal sensitivityVerbal stimulationSocial–emotionalCognitive

a b s t r a c t

Although maternal sensitivity has been shown to influence social–emotional develop-ment, the role of verbal stimulation on infant developmental outcomes has received lessexploration. Recent research has focused on intentional behaviors within the context ofa mother–infant interaction as a critical influence and as distinct from sensitivity. In thisinvestigation 6377 mother–infant dyads participated in a teaching task as part of the sam-ple from the Early Childhood Longitudinal Study Birth Cohort (ECLS-B). Analyses focusedin deciphering the role of maternal sensitivity and verbal stimulation as contributors to theinfant’s social–emotional (S–E) and cognitive (Cog) development. We further hypothesizedthat inclusion of infant age as a moderator of maternal behaviors would illuminate any dif-ferences between younger and older infants. Results: For the infant’s S–E development, ourhypothesis that maternal sensitivity would be a stronger predictor than verbal stimulationwas not supported; nor did we find support for our hypothesis that the association would bemoderated by age. For Cog development, only verbal stimulation had a direct positive effecton the infant’s cognitive ability; our findings for moderation showed that mothers spokemore to older infants than younger infants. Conclusion: Identification of specific maternalbehaviors associated with infant outcomes informs the child development field, and alsoprovides strategies for early intervention to assist mothers with developing or maintaininga consistent relationship that includes sensitivity and verbal stimulation.

Published by Elsevier Inc.

1. Introduction

It is well accepted that the family provides important relational contexts for infant and child development, and that posi-tive parent–child interactions are critical for optimizing developmental outcomes, including social–emotional and cognitivegrowth. Because the mother–child relationship is uniquely close and long-term and has proven to significantly influencedevelopmental outcomes, it has been under close scrutiny. Interest has often focused on the mother’s health or psychologicalmood (Carter, Garrity-Rokous, Chazan-Cohen, Little, & Briggs-Gowan, 2001; Field, 1992) in explaining poor outcomes forchildren throughout their lives. Alternatively, others have investigated the specifics of the mother’s cognitions when inter-acting with her infant (Meins, Fernyhough, Fradley, & Tuckey, 2001) and the impact of her verbal and nonverbal behaviorsin stimulating the growth and development of her child (Landry, Smith, Miller-Loncar, & Swank, 1997). One area of researchfocuses on observations of mother–child interactions and caregiving skills such as feeding, free play, teaching strategies, aswell as the mother’s behaviors to soothe the infant during stressful and non-stressful events (Beckwith & Rodning, 1996;

∗ Corresponding author. Tel.: +1 520 621 3902.E-mail address: mpage@u.arizona.edu (M. Page).

0163-6383/$ – see front matter. Published by Elsevier Inc.doi:10.1016/j.infbeh.2009.12.001

102 M. Page et al. / Infant Behavior & Development 33 (2010) 101–110

Field, Guy, & Umbel, 1985; Landry, Smith, & Swank, 2006; McElwain & Booth-LaForce, 2006). A common theme among thesestudies is the significance of maternal sensitivity and verbal interactions within the mother–child relationship.

Defining behaviors associated with maternal sensitivity, Meins (1997) purported that maternal behaviors must be appro-priate for the situation; thus her research focused on the importance of the mother’s cognitions, and in particular herinterpretation of her infant’s actions to determine the appropriateness of her behavior. Of the five sensitivity behaviorsindicated by Ainsworth, Blehar, Waters, and Wall (1978), Meins (1997) believed that behaviors that promoted social andemotional development were (1) responding promptly to a cry recognized as signaling distress; (2) comforting the infantwhen upset; and (3) being available for interaction, whereas the behaviors (4) interpreting the infant’s actions as meaningful;and (5) treating the infant as an intentional agent promoted the development of cognitive growth in the infant. Tying backinto other’s research, the first three behaviors can be seen to reflect responsiveness, whereas the latter two may be seen toreflect contingency as suggested by Landry, Swank, Assel, Smith, and Vellet (2001).

It is our hypothesis, that within any context a mother exhibits a variety of behaviors, and unpacking the meaning andimpact of each type of behavior as a unique influence on a specific infant outcome is difficult. Rather we believe that maternalbehaviors influence multiple outcomes such that maternal sensitivity may also contribute to cognitive development andverbal stimulation may also contribute to social–emotional development. The purpose of this manuscript is to explore theimpact of maternal sensitivity and verbal stimulation on two infant outcomes (i.e., social–emotional and cognitive develop-ment), but also to explore if either of the two maternal behaviors have differing impact based on infant age; thus we testedfor moderation. This study involves a secondary analysis of the Early Childhood Longitudinal Study Birth Cohort (ECLS-B)national study of parents and their infants born during 2001. Findings from this study will contribute to the child developmentliterature as well as to inform practitioners about the impact of sensitivity and verbal stimulation to infant development.

To develop our hypotheses, we begin with a review of the literature regarding the influence of maternal behaviors oninfant social–emotional and cognitive development. In addition to maternal behaviors, we explore the findings that infantage may influence maternal behaviors resulting in discontinuity across infant developmental ages.

1.1. Maternal behavior as a foundation for infant social–emotional development

Numerous studies support the notion that sensitive maternal behaviors influence the social–emotional development ofthe infant through the development of an initial relationship based on trust that is necessary for reciprocal maternal–infantinteractions (e.g., Landry et al., 2001). Two of the more commonly identified aspects of sensitive behavior include respon-siveness and contingency. Observations from an intervention study (Landry et al., 2006) showed responsive mothers wereprompt, sensitive, contingent and treated their infants as capable of intentions supporting previously defined aspects ofsensitive behavior.

The second facet of maternal sensitivity focuses on the mother’s contingent behavior during reciprocal interactions andbasic caregiving. Contingent behavior suggests that the mother acknowledges and responds to the infant’s needs within areasonable period of time after the infant signals (e.g., crying, calling, or crawling after) (Eshel, Daelman, de Mello, & Martines,2006; Landry et al., 2001; Smith, Landry, & Swank, 2006). Researchers focusing on social emotional development suggest thatmaternal contingent and responsive behaviors serve as a model for the child’s own emotional development and regulationof emotions (McElwain & Booth-LaForce, 2006) and are important for promoting social development by emphasizing therules of socialization within the interaction (Landry et al., 2006). Therefore, in this study, maternal sensitivity includes bothverbal and nonverbal behaviors that are responsive and contingent in nature.

1.2. Maternal behavior associated with infant cognitive development

The amount of maternal sensitivity, responsiveness, and contingent behavior, as well as verbal stimulation offered bythe mother to her infant, has been shown to have a positive association to both language development and overall cognitiveability (Beckwith & Rodning, 1996; Eshel et al., 2006; Field et al., 1985; Landry et al., 2006; Lemelin, Tarabulsy, & Provost,2006; Smith et al., 2006). Factors associated with the amount of verbal communication between a mother and her infantin promoting cognitive growth include maternal age, education, culture and ethnicity, sensitivity, and verbal stimulation orimitation (Eshel et al., 2006; Field, 1992; Laosa, 1980; Paavola, Kunnari, & Moilanen, 2005).

Conversational or descriptive dialogue from the mother to her infant lays the foundation for dyadic interactions and forcontributions to healthy cognitive outcomes (Landry et al., 2006), including language development. Early maternal dialoguehas been shown to be a function of the mother’s educational level, with mothers having fewer years of education beingless verbal with their infant, or not verbalizing to them until the infant began vocalizing in the latter half of the first year(Beckwith & Rodning, 1996; Berlin, Brady-Smith, & Brooks-Gunn, 2002; Landry et al., 2001, 2006).

Salient to our study are the findings of maternal–infant interactions, and specifically mothers who used verbal behaviorsthat encouraged and maintained the infant’s attention towards an object or task, or utilized a directive skill in teachingpromoted infant vocabulary and overall cognitive ability (Landry et al., 1997, 2006; Ruddy & Bornstein, 1982). Further,researchers focusing on how and what the mother says have found impacts to cognitive development and emotion under-standing (Taumoepeau & Ruffman, 2006, 2008). Verbal behaviors within the current study include directions, descriptions,and encouraging words for the infant to continue, and are representative of the types of verbal stimulation utilized duringthe teaching task.

M. Page et al. / Infant Behavior & Development 33 (2010) 101–110 103

1.3. Maternal sensitivity and verbal stimulation associated with infant age

Infant age has been shown to be associated with fluctuations in maternal behaviors during interaction with her infant (e.g.,Deater-Deckard et al., 2001; Feldman & Eidelman, 2009). Researchers focusing on maternal behaviors have suggested thatfluctuations may be a result of the emotional wellness of the mother such as the onset of depression as well as developmentalaspects of the infant themselves including their development toward becoming a social individual. For example, Isabella(1998) found maternal sensitivity from 1 to 9 months was highest at 4 months and decreased by the 9th month. Severalother studies reported a decline in maternal sensitivity, finding that risk factors such as maternal depression, unplannedpregnancy, and low social support were associated with the decline or instability of sensitivity (Campbell, Matestic, vonStauffenberg, Mohan, & Kirchner, 2007; Kemppinen, Kumpulainen, Raita-Hasu, Moilanen, & Ebeling, 2006; Mertesacker,Bade, Haverkock, & Pauli-Pott, 2004).

Infant age also appears to influence the early vocalizations by the mother as well as the content or structure of theconversation. Jones and Moss (1971) found infant vocalizations at 2 weeks of age were non-socially derived, whereas those at3 months of age were associated with social interaction with the mother. Studies observing maternal speech at 1 and 3 monthsof age found mothers engaged in more utterances, more words, and a richer vocabulary to their 3 month olds compared towhen the infant was 1-month old, and these increases are often in conjunction with more infant smiles and positive infantvocalizations (Henning, Striano, & Lieven, 2005). Mothers increased their utterances and alternated vocalizations withina conversation style as their infant became older (Clearfield & Nelson, 2006; Elias, Hayes, & Broerse, 1988). Mothers alsoincreased the use of directive speech, questions, and reports as well as the frequency of information-salient and affect-salientlanguage to 13-month olds compared to 5-month olds (Bornstein et al., 1992). Continuity of maternal verbalizations wasfound for sibling pairs when each child was 12 and 24 months old, with an inconsistent measure of verbalizations for eachchild separately at 12 and 24 months (Dunn, Plomin, & Daniels, 1986) notably due to the developmental changes for eachinfant.

2. Hypotheses

The purpose of the current study is to examine the relationships between maternal sensitivity and maternal verbalstimulation as predictors of infant outcomes of social–emotional and cognitive development, with an interest in differentialmaternal behaviors moderated by infant age. In this study, maternal behaviors observed during maternal–infant teachinginteractions provided a foundational context for exploring two infant developmental outcomes. The overall goal of thecurrent study, thus, is to determine the impact of maternal behaviors characterized by sensitivity and verbal stimulation oninfant social–emotional and cognitive development. Specifically we will test the following hypotheses:

(1a) We expected maternal behaviors associated with maternal sensitivity, as compared to maternal behaviors associatedwith verbal stimulation, will have a stronger predictive value for infant social–emotional development.

(1b) We hypothesized that the impact of maternal sensitivity on social–emotional development will diminish for olderinfants, whereas the impact of verbal stimulation on social–emotional development will not be affected (i.e., moderated)by infant age. No main effects of age are expected.

(2a) We expected maternal behaviors associated with verbal stimulation, as compared to maternal behaviors associatedwith maternal sensitivity, will have a stronger predictive value for infant cognitive development.

(2b) We hypothesized that the impact of maternal verbal stimulation on cognitive development will increase for olderinfants, whereas infant age will not affect the relation of maternal sensitivity to cognitive development. A main effectof infant age was expected; specifically, a positive association with cognitive development.

3. Method

3.1. Measures

This study is a secondary data analysis of the Early Childhood Longitudinal Study Birth Cohort (ECLS-B). Primary datacollection for the ECLS-B dataset was administered by the research team headed by Jennifer Park. For a full review of the dataand measures collected, please refer to the website http://nces.ed.gov/ecls/Birth.asp. Self-reported data from the womenwas used for control variables. The Bayley Mental Scale and the Nursing Child Assessment Teaching Scale were utilized forthe infant outcomes and the maternal predictors. A description of the measures and the specific procedures for use in thissecondary data analysis are provided.

3.1.1. Bayley Mental ScaleThe 9-month Bayley Short Form-Research Edition (BSF-R) Mental Scale (BSID-II; Bayley, 1993) included 22 items designed

to assess infants’ cognitive development (e.g., memory, means-end behavior, exploratory competence, object permanence,expressive communication, and receptive communication). Infants were presented with tasks such as putting blocks in acup, ringing a bell, and responding to a parent’s request (e.g., peek-a-boo). For most items, infants were presented with

104 M. Page et al. / Infant Behavior & Development 33 (2010) 101–110

objects and verbal instructions, and their behavior was observed and recorded. The Bayley Mental Scale Score was utilizedas the dependent measure of infant cognitive ability (sample range = 54–112) with higher scores reflecting higher levels ofcognitive development.

3.1.2. Nursing Child Assessment Teaching ScaleThe Nursing Child Assessment Teaching Scale (NCATS) (Barnard, 1978) includes scales to measure maternal and infant

behaviors during a teaching task for use with children ranging in age from birth to 3 years. Specifically, for the current paper,three scales are derived from this measure: material sensitivity, maternal verbal stimulation, and infant social–emotionaldevelopment. The parent domain measures maternal behaviors based on the mother’s ability to interact and respond to herinfant during a teaching task. The infant scale measures overall social–emotional development based on the infant’s abilityto use social skills and emotional control to interact and respond to the parent.

As a secondary data analysis, we utilized the NCATS data gathered by the ECLS team, which included 6–18 individualswho were trained by the University of Washington (UW). The sole responsibility of these trained individuals was to code thetapes collected during the year. These individuals were blinded to other data collected from the mother and infant duringthe home interview. During the year of data collection, 2% of the videotaped interaction sessions were coded by the ECLSteam as well as by the University of Washington to establish coding reliability. The inter-rater reliability between UW andthe ECLS team resulted in a Cronbach’s alpha = .86 which met the .85 reliability criteria agreed upon at the outset of theproject. The first author for this manuscript has received training and attained reliability in the scale, thus allowing for afamiliarity with the items and the coding scheme.

3.1.2.1. Infant social–emotional development. Infants were rated on 23 observed behaviors representing the infant’s clarityof cues and their responsiveness to caregiver cues. Due to copyright permissions, the specific scales cannot be reproducedin full in this manuscript (Sumner & Spietz, 1994). The infant scales include items such as the child vocalizes or babblesafter the caregiver’s verbalization, the child looks at caregiver’s face to establish eye contact, and the child stops displayingpotent disengagement cues after caregiver’s soothing attempts. Developing social interaction skills and the ability to regulateemotions during an interaction are the basis of these scales focusing on the infant’s ability to send clear cues to the caregiverand the infant’s ability to respond to the caregiver’s attempts to communicate and interact. If an infant performs a behavior,that item is coded as “yes”. A total score is created by summing score the yes items, resulting in a total social emotional scorewith a range of 0–23.

3.1.2.2. Maternal sensitivity and verbal stimulation. Various studies have utilized one or more of the intact NCATS mater-nal scales as a predictor variable of infant outcomes (e.g., Bannerjee & Tamis-LeMonda, 2007; Gaffney, Kodadek, Meuse,& Jones, 2001; Horodynski & Gibbons, 2004). Our intent was to identify two sets of items that closely aligned with thematernal behaviors of interest to this study: maternal sensitivity and verbal stimulation. Coding of the teaching task utilizeda dichotomous (yes/no) system to specify whether mothers exhibited a series of 50 behaviors. The dichotomous, yes/nocoding system, rendered factor analysis an inappropriate data reduction strategy that did not provide meaningful factors.Therefore, we determined that items would be reviewed first for response variance and then using face validity to groupbehaviors into two scales that align with the literature on maternal sensitivity and verbal stimulation. Checking for responsevariance was important because, for many behaviors, all mothers performed or did not perform the behavior. Face valid-ity involves “reformulating test items in terms that appear relevant and plausible in the particular setting in which theywill be used” (Anastasi & Urbina, 1997, p. 118). Thus this procedure allowed us to assign the maternal behaviors fromthe NCAST checklist to one of two groups of behaviors (maternal sensitivity and verbal stimulation) based on previousliterature.

Reviewing all 50 items for response variance, we eliminated 15 items on which fewer than 5% of the mothers eitherdemonstrated or failed to demonstrate the behavior. Of the 35 remaining variables, a face validity analysis found that 6items from the response to distress subscale represented both responsive and contingent behaviors after the infant displayeddistress (e.g., averting gaze, back arching, or crying) with Cronbach’s alpha reliability of .76. Using the established literatureon sensitive and contingent behaviors and the study by McElwain and Booth-LaForce (2006) that suggested mothers displaygreater sensitivity during infant distress; we concluded that the use of these items as a measure of maternal sensitivityreplicated maternal sensitive behaviors as defined in the literature review. Within the sample of 8489 infants, 75% exhibiteddistress during the teaching task. Unless an infant shows distress or disengages during the teaching task, maternal behaviorsare not observed, and mothers receive a yes score for all 11 items in the response to distress subscale. Within our sample,25% of the mothers had a perfect score for this subscale resulting in a skewed distribution for this variable. Thus, in order tofully explore the impact of maternal sensitivity, and to capitalize on both contingent and sensitive behaviors in response todistress, we reduced our sample from 8489 to 6377 utilizing only mother–infant dyads that exhibited distress.

Verbal stimulation within this study focused on the verbal statements from the mother that encouraged the infant orthat described the teaching task or provided feedback to assist the infant towards success. From the remaining 29 variables,12 verbal items were selected meeting the study definition. Cronbach’s alpha reliability for the verbal stimulation scale was.70. The maternal behaviors for the two scales are listed in Table 1.

M. Page et al. / Infant Behavior & Development 33 (2010) 101–110 105

Table 1Percentage of mothers performing maternal behaviors comprising sensitivity (S) and verbal stimulation (VS).

Maternal behavior No Yes

S1. Caregiver (CG) stops teaching at potent disengagement cues. 9.9 90.1S2. CG makes positive, sympathetic, or soothing verbalization. 36.3 63.7S3. CG changes voice volume to softer or higher pitch, does not yell. 19.4 80.6S4. CG rearranges the child’s position or task materials. 40.1 59.9S5. CG makes soothing nonverbal response, e.g. pat, touch, rock, caress, kiss. 78.8 21.2S6. CG diverts the child’s attention by playing games, introduces a new toy. 95.3 4.7VS1. CG praises child’s successes or partial successes. 61.5 38.5VS2. CG praises child’s efforts or behaviors broadly at least once during the episode. 86.5 13.5VS3. CG makes cheerleading type statements to the child during the teaching interaction. 60.3 39.7VS4. CG describes perceptual quality of the task materials to the child. 89.8 10.2VS5. CG uses at least two different sentences or phrases to describe the task to the child. 8.5 91.5VS6. CG uses explanatory verbal style more than imperative style in teaching the child. 41.7 58.3VS7. CG’s directions are stated in clear, unambiguous language. 56.7 43.3VS8. CG uses verbal description and modeling simultaneously in teaching any part of the task. 24.1 75.9VS9. CG verbally praises child after child has performed better or more successfully. 77.3 22.7VS10. CG responds to the child’s vocalizations with a verbal response. 27.4 72.6VS11. CG uses both verbal and nonverbal instruction in teaching the child. 4.8 95.2VS12. CG uses the teaching loop at least once (alerting, instruction, performance, feedback). 35.7 64.3

3.2. Participants

The sample for the current study was comprised of a subset of the mother–infant dyads from the ECLS-B cohort. The infantswere born in 2001. The 10,688 participating dyads came from all 50 states and comprised the nationally representativesample for the Early Childhood Longitudinal Study Birth Cohort (ECLS-B). A secondary data analysis was conducted utilizingthe public use file for the 9-month old data collection, which included infants ranging in age from 8 to 12 months. The sub-sample for the current study included those biological mothers and infants who were evaluated with the completed NursingChild Assessment Teaching Scale (NCATS) and the infants’ were assessed with the Bayley Mental Scale. This sub-sampleconsisted of 8489 mother–infant dyads with complete data on these two scales; however, our sample was further reducedto those dyads in which the infant displayed distress during the teaching task (N = 6377) as noted in the above descriptionof the NCATS measure.

The sample of 6377 mother–infant dyads consisted of 51.9% male infants (N = 3311), and the mean age of the infantswas 10.25 months (SD = 1.33). The mothers’ age ranged from less than 20 to over 40; the mean age was 26 years. Thematernal sample was ethnically diverse: 47.1% non-Hispanic White, 17.3% Hispanic, 16.2% African-American, 12.6% Asian,4.0% American Indian, and 2.9% multi-ethnic, non-Hispanic. Mother’s education ranged from less than high school (25.8%),high school or equivalent (21.1%), vocational (2.2%), some college (24.3%), to bachelor’s degree or higher (26.6%).

3.3. Predisposing characteristics for infant development

Several characteristics were included in our analysis as control variables to disentangle the main effects from thesepotential predisposing characteristics. These include maternal age (Berlin et al., 2002; Landry et al., 2001; Parks & Arndt,1990; Roberts, Bornstein, Slater, & Barrett, 1999; Stevens, 1984), maternal education (Eshel et al., 2006; Laosa, 1980; Roe, Roe,Drivas, & Bronstein, 1990), ethnicity (Bretherton, 1992; Feeney & Noller, 1996; Laosa, 1980; Meins, 1997; Smith et al., 2006),and infant sex (Carter et al., 2001; Clearfield & Nelson, 2006; Grace, Evindar, & Stewart, 2003; Hinde & Stevenson-Hinde,1987; Milgrom, Westley, & Gemmill, 2004).

3.4. Analysis

Table 2 provides the means, standard deviations, and possible ranges of the predictor and outcome variables. Table 3shows the correlations among the study variables. Two hierarchical regression analyses were used to test the hypotheses.

Table 2Mean, standard deviations, and range for predictor and outcome variables.

Mean (SD) Range

PredictorsMaternal sensitivity 3.20 (1.36) 0–6Verbal stimulation 6.26 (2.39) 0–12

Infant outcomesNCAST score 15.95 (2.53) 0–23Bayley Scale score 76.16 (7.00) 0–178

N = 6377.

106 M. Page et al. / Infant Behavior & Development 33 (2010) 101–110

Table 3Correlations among control, predictor, and outcome variables.

1. 2. 3. 4. 5. 6. 7. 8.

1. Infant age –2. Infant sex −.01 –3. Maternal age −.02* −.01 –4. Maternal ethnicity −.01 −.01 .12** –5. Maternal education −.03* .01 .44** .20** –6. Maternal sensitivity .02 .02 .07** .02 .13** –7. Verbal stimulation .09** −.03* .13** .21** .25** .17** –8. Bayley Mental Scale .70** −.05** −.04** .02 .01 .04** .16** –9. NCAST infant score .07** −.05** .01 .01 .05** .24** .24** .10**

Note: infant sex was coded 0 = girls and 1 = boys. Maternal ethnicity was coded 0 = non-White ethnic minority and 1 = non-Hispanic White.* p < .05.

** p < .01.

Step one of both regression analyses included the control variables; maternal age and education were entered as continuousvariables, whereas all minority groups were collapsed into a single group coded 0 for non-white minorities and 1 for non-Hispanic white. Infant sex was coded 0 for female and 1 for male. At Step 2 of each regression analysis, the three predictorvariables (sensitivity, verbal stimulation, and infant age) were entered. These variables were centered prior to entry tominimize colinearity with the interactions tested in the next step (Aiken & West, 1991). Finally, in step 3, the two interactions(products of centered maternal behavior and centered infant age variable) were entered to test for age moderation of theassociations of maternal behaviors with infant outcomes.

4. Results

Results are presented for the regression analyses associated with our hypotheses. SPSS (SPSS, Inc., 2005, Version 14.0)was used for all data analyses.

4.1. Regression analyses

A series of regression analyses were conducted to address the hypotheses. The first hypothesis evaluated whether mater-nal behaviors associated with maternal sensitivity, as compared to maternal behaviors associated with verbal stimulation,acted as a stronger predictor of infant social–emotional development (1a). We also evaluated age moderation, expecting theeffect of maternal sensitivity on social–emotional development to be stronger for younger than older infants, whereas wedid not expect infant age to moderate the effect of verbal stimulation on social–emotional development (1b). In contrast,for our second hypothesis, we expected that maternal behaviors associated with verbal stimulation, as compared to mater-nal behaviors associated with maternal sensitivity, would more strongly predict infant cognitive development (2a). As formoderation, we expected maternal verbal stimulation to more strongly predict cognitive development among older thanyounger infants, but did not expect infant age to moderate the link between maternal sensitivity and cognitive development(2b). For each analysis, control variables were entered at step 1, the two maternal behaviors and infant age were centeredaccording to Aiken and West (1991) and entered at step 2, and the two moderators were created using the centered variablesand entered at step 3.

4.1.1. Maternal behaviors predicting social–emotional developmentThe result of the first hierarchical regression predicting infant’s social–emotional development (summarized in Table 4)

indicated that both maternal sensitivity (ˇ = .208, p < .001) and verbal stimulation (ˇ = .219, p < .001) were significant pre-dictors. Age did not moderate these associations (i.e., the interaction terms of Step 3 were not significant). The reducedmodel without the moderators was significant and explained 10.4% of the child’s social–emotional score, F(7, 6364) = 104.71,p < .001. Results of these regression analyses confirmed our hypothesized positive relation between maternal sensitivity andsocial–emotional development, but this effect was not stronger than that of verbal stimulation nor did infant age moderatethis effect. Thus, results failed to support either hypothesis 1a or 1b.

4.1.2. Maternal behaviors predicting cognitive developmentThe second hypothesis addressed the impact of maternal behaviors on the infant’s cognitive development (2a) as a

function of the infant’s age (2b). Results of the regression analysis evaluating this hypothesis are summarized in Table 5.Verbal stimulation (ˇ = .073, p < .001), but not maternal sensitivity (ˇ = .015, p > .05), significantly predicted infants’ cognitivescore, supporting hypothesis 2a. Also, as expected, age was strongly positively related to cognitive scores (ˇ = .762, p < .001)Age moderation of the effect of sensitivity was not significant (ˇ = .006, p > .05), but age did significantly moderate the relationverbal stimulation with cognitive scores (ˇ = .027, p < .001), thus supporting hypothesis 2b. The model explained 60% of thevariance in the infant’s cognitive score (F(9, 6362) = 1079.56, p < .001), with infant age contributing the most substantial betaweight for the full model. To follow-up the significant interaction, Fig. 1 shows the graphical representation for verbal

M. Page et al. / Infant Behavior & Development 33 (2010) 101–110 107

Table 4Regression analysis for infant’s NCAST social–emotional score.

Variable B SE B ˇ �R2

Step 1 .005***

Infant sex −.222 .063 −.044***

Maternal age .000 .028 .000Maternal ethnicity −.053 .067 −.010Maternal education .090 .023 .057***

Step 2 .098***

Sensitivity .392 .023 .208***

Verbal stimulation .233 .014 .219***

Infant age .084 .022 .045***

Step 3 .000Age × sensitivity −.031 .017 −.025Age × verbal stimulation .000 .010 −.001

Note: infant sex was coded 0 = girls and 1 = boys.Maternal ethnicity was coded 0 = non-White ethnic minority and 1 = non-Hispanic White.

*** p < .001.

Table 5Regression analysis for Bayley Mental Scale score.

Variable B SE B ˇ �R2

Step 1 .009***

Infant sex −.839 .179 −.058***

Maternal age −.290 .080 −.051***

Maternal ethnicity .704 .192 .048***

Maternal education .194 .066 .043**

Step 2 .595***

Maternal sensitivity .079 .043 .015Verbal stimulation .209 .026 .069***

Infant age 4.08 .042 .762***

Step 3 .001**

Age × sensitivity .022 .033 .006Age × verbal stimulation .062 .019 .027**

Note: infant sex was coded 0 = girls and 1 = boys.Maternal ethnicity was coded 0 = non-White ethnic minority and 1 = non-Hispanic White.

** p < .01.*** p < .001.

Fig. 1. Graphical representation of cognitive development from verbal stimulation as moderated by infant age (younger, middle, and older infants).

108 M. Page et al. / Infant Behavior & Development 33 (2010) 101–110

stimulation suggesting distinct differences for each of three age ranges. Age groupings (older, middle, younger) were basedon the mean (middle = 10.25 months) plus or minus one standard deviation (older = 11.58 months and younger = 8.92 months,respectively). Although the differences in this figure are not obvious given the strong effect of age, the association betweenverbal stimulation and cognitive development is most strongly positive among the older infants (top line) than among theyounger infants (bottom line). In sum, infant scores for cognitive development were predicted by maternal verbal stimulationand infant age, with infant age also moderating the effect of maternal stimulation such that stronger positive relations areevident among older than younger infants.

5. Discussion

Prior research has shown the importance of maternal behaviors for infant development. One unique contribution ofthe present study was to group individual maternal behaviors into two broad domains of maternal sensitivity and verbalstimulation that were hypothesized to have a simultaneous influence on infant developmental outcomes. Typically, maternalverbal behaviors are thought to promote cognitive development, but our findings indicated that behaviors comprising theverbal stimulation scale also contributed to the infant’s social–emotional development. Thus, the current study adds to theliterature by showing that the proposed maternal behaviors defined as sensitivity and verbal stimulation exhibited duringmaternal-child teaching interactions are positively associated with infant social–emotional and cognitive development. Ourstudy focused on infants displaying distress during the teaching task in order to observe maternal behaviors (sensitivity)associated with response to that distress. Further, the items within our sensitivity scale represent a subset of contingencyitems in the NCAST teaching task that identify maternal responsiveness to the infant. Specific to this study, items for sensitivebehaviors included two verbal and several nonverbal actions to redirect her distressed infant back to the teaching task (seeTable 1 for item descriptions).

Our findings for the first hypothesis concerning the impact of maternal behaviors on social–emotional development didnot confirm our hypotheses that sensitivity would have a stronger impact than verbal stimulation. Rather, the regressioncoefficients suggested that both maternal sensitivity and verbal stimulation have unique but comparable effects in predictingsocial–emotional development at nine months.

While the results support existing literature that maternal sensitivity promotes social–emotional development, the addi-tion of verbal stimulation specifically including items designed to teach and encourage an infant to complete a task mayfurther act to promote social–emotional development. Thus, these findings suggest an importance to increase maternalawareness of all types of verbal stimulation to one’s infant from birth.

We were also interested in exploring if parenting behaviors matter more at certain ages, specifically that of sensitivity andverbal stimulation. Whereas prior research has shown both continuity and discontinuity of sensitivity over time (Campbellet al., 2007; Isabella, 1998; Kemppinen et al., 2006; Mertesacker et al., 2004), our study did not find maternal sensitivitymoderated by infant age. The most plausible explanation could be our use of cross-sectional data; thus we did not have datashowing a prior or post-measure of sensitivity to detect the decrease as infants matured.

Our second hypothesis focused on infants’ cognitive development as measured by the Bayley Mental Scale. We specificallyhypothesized that maternal verbal stimulation would act as a key predictor in cognitive gains, particularly in the older groupof infants. We found verbal stimulation did act as a significant predictor of cognitive development as hypothesized andmaternal sensitivity did not predict cognitive development in this sample. Again, our interest in the influence of maternalbehaviors at incremental ages, the moderation by age showed an increase in the effect of verbal stimulation on cognitivedevelopment with increasing age. As expected, infants’ age had a strong main-effect association with cognitive developmentwithin the model. Although these findings are important, it is important to bear in mind that the cross-sectional samplelimits our ability to explore within-dyad changes; but based on the literature, it seems likely that the increase in verbalstimulation could be attributed to a mother’s greater likelihood to speak more to her older infant (Bornstein et al., 1992;Clearfield & Nelson, 2006).

The maternal verbal stimulation variable was created from the context of a teaching interaction, thus many of the itemsare reflective of the mother’s ability to teach her child a task using verbally based indicators (i.e., use of 2 different sentenceswas performed by 91.5% of mothers, 75.9% verbally described or modeled the task, and 95.2% used verbal and nonverbalmethods to teach their child). Future research is needed to observe other types of caregiving interactions to determine ifmothers who scored higher in verbal stimulation consistently speak more to their infants throughout the day and throughvarious types of interactions. Further, use of longitudinal data could substantiate an increase in greater verbal stimulation asthe infant matures. Despite these limitations, the use of a national dataset provided adequate demographic representation,thus strengthening the findings from this study and allowing for generalizations to be made about maternal behaviors andchild outcomes as observed within the dyadic teaching task.

In addition to the positive contributions of this study to the extant literature, several limitations merit consideration. First,our use of the data was based on secondary data analysis, thus our measures of maternal behaviors were limited to only thoseobserved during a maternal–child interaction during the 9-month data collection. Though the NCAST teaching scale has beenused in many studies, we do not expect our findings to align with those reported previously given our construction of twonew maternal behavior scales from the existing items. Because the NCAST scale uses dichotomous responses limiting theability to alter the data, we created two clusters of maternal behaviors based on face validity and variance of item responses,and established an alpha reliability for each scale. During initial analysis of the scales, it was necessary to reduce our sample

M. Page et al. / Infant Behavior & Development 33 (2010) 101–110 109

to those infants who disengaged from the teaching task in order to limit the bias from mothers with perfect sensitivity, whichcaused the sensitivity measure to appear in a cubic (i.e., sideways ‘S’) shape. Removing mothers whose infants showed nodistress resulted in a linear line with greater sensitivity positively correlating with greater social–emotional development.Despite the necessity of this removal, it also limits the generalizability of the findings of the current study.

Second, we believe other maternal behaviors may influence infant outcomes, but our use of the ECLS-B data limited ourability to test additional behaviors beyond those observed during the teaching interaction. Future research should includeadditional contextual factors as control variables such as the child’s participation in day care, maternal depression, or numberof children in the home as included in studies by Belsky (1984) and Abidin (1992) to identify influences affecting parentingbehaviors.

Finally, our reliance on cross-sectional data precludes conclusions of direction of influence. Although we have framedour hypotheses on the assumption that maternal behaviors influence infant outcomes (e.g., Ainsworth et al., 1978; Meins,1997), we must recognize that infants also influence their parents’ behaviors (e.g., Bell, 1968; Belsky, 1984). Additionalanalyses using subsequent waves of data are needed to evaluate the strength and direction of influence within maternal-childinteractions.

6. Conclusion

The importance of maternal behaviors to infant outcomes begins at birth and continues throughout the child’s life.Maternal behaviors can be observed as the mother responds to her infant’s needs through sensitivity and verbal stimula-tion. Although data exist observing the vocalizations between mother and infant in the first few months, few studies havefocused on the specific influences of maternal verbalizations on social–emotional development, as well as to cognitive devel-opment. Our findings demonstrate that both maternal sensitivity and verbal stimulation predict increases in the infant’ssocial–emotional outcome; however, only verbal stimulation contributes to cognitive development, and was slightly moreinfluential when infants were older.

Campaigns and education materials focusing on maternal sensitivity have benefited maternal–child relationships, andreading programs have emphasized the benefit of reading to young children. However, in addition to these campaigns,educational materials and encouragement from hospital staff can impart the importance of maternal verbal stimulation toinfants, regardless of their age or vocabulary. Simple conversations can include what the mother is doing, naming objectsand places around them, and just talking to the infant about any subject the mother thinks about. Interventions supportingthe increase in sensitive behaviors have shown promise in influencing maternal–child interactions, suggesting by extensionthat interventions that model the role of verbal stimulation may have the potential to alter the mother’s view or comfort levelregarding the impact of her communication with her infant thus promoting language development such as words, syntax,taking turns during a dialogue, and cognition through the exploration and naming of objects in the infant’s environment.

References

Abidin, R. R. (1992). The determinants of parenting behavior. Journal of Clinical Child Psychology, 21, 407–412.Aiken, L. S., & West, S. G. (1991). Multiple regression: Testing and interpreting interactions. Newbury Park, CA: Sage.Ainsworth, M. D. S., Blehar, M. C., Waters, E., & Wall, S. (1978). Patterns of attachment: A psychological study of the strange situation. Hillsdale, NJ: Erlbaum.Anastasi, A., & Urbina, S. (1997). Validity: Basic concepts. In Pychological testing. Upper Saddle River, NJ: Prentice Hall.Bannerjee, P. N., & Tamis-LeMonda, C. S. (2007). Infants’ persistence and mothers’ teaching as predictors of toddlers’ cognitive development. Infant Behavior

and Development, 30, 479–491.Barnard, K. (1978). Nursing child assessment satellite training: Learning resource manual. Seattle: University of Washington, School of Nursing.Bayley, N. (1993). Bayley scales of infant development, second edition manual. San Antonio, TX: The Psychological Corporation.Beckwith, L., & Rodning, C. (1996). Dyadic processes between mothers and preterm infants: Development at ages 2 to 5 years. Infant Mental Health Journal,

17, 322–333.Bell, R. Q. (1968). A reinterpretation of the direction of effects in studies of socialization. Psychological Review, 75, 81–95.Belsky, J. (1984). The determinants of parenting: A process model. Child Development, 55, 83–96.Berlin, L. J., Brady-Smith, C., & Brooks-Gunn, J. (2002). Links between childbearing age and observed maternal behaviors with 14-month-olds in the Early

Head Start Research and Evaluation Project. Infant Mental Health Journal, 23(1–2), 104–129.Bornstein, M. H., Tal, J., Rahn, C., Galperín, C. Z., Pêcheux, M. G., Lamour, M., et al. (1992). Functional analysis of the contents of maternal speech to infants

of 5 and 13 months in four cultures: Argentina, France, Japan, and the United States. Developmental Psychology, 28, 593–603.Bretherton, I. (1992). The origins of attachment theory: John Bowlby and Mary Ainsworth. Developmental Psychology, 28, 759–775.Campbell, S. B., Matestic, P., von Stauffenberg, C., Mohan, R., & Kirchner, T. (2007). Trajectories of maternal depressive symptoms, maternal sensitivity, and

children’s functioning at school entry. Developmental Psychology, 43, 1202–1215.Carter, A. S., Garrity-Rokous, F. E., Chazan-Cohen, R., Little, C., & Briggs-Gowan, M. J. (2001). Maternal depression and comorbidity: Predicting early parenting,

attachment security, and toddler social–emotional problems and competencies. Journal of the American Academy of Child and Adolescent Psychiatry, 40,18–33.

Clearfield, M. W., & Nelson, N. M. (2006). Sex differences in mothers’ speech and play behavior with 6-, 9-, and 14-month old infants. Sex Roles, 54(1/2),127–137.

Deater-Deckard, K., Pike, A., Petrill, S. A., Cutting, A. L., Hughes, C., & O’Connor, T. G. (2001). Nonshared environmental processes in social–emotionaldevelopment: An observational study of identical twin differences in the preschool period. Developmental Science, 49(2), F1–F6.

Dunn, J. F., Plomin, R., & Daniels, D. (1986). Consistency and change in mothers’ behaviors toward young siblings. Child Development, 57, 348–356.Elias, G., Hayes, A., & Broerse, J. (1988). Aspects of structure and content of maternal talk with infants. Journal of Child Psychology and Psychiatry, 29, 523–531.Eshel, N., Daelman, B., de Mello, M. C., & Martines, J. (2006). Responsive parenting: Interventions and outcomes. Bulletin of the World Health Organization,

84, 991–998.Feeney, J., & Noller, P. (1996). Attachments in infancy and beyond. In Adult attachment. Thousand Oaks: Sage Publications., pp. 1–21

110 M. Page et al. / Infant Behavior & Development 33 (2010) 101–110

Feldman, R., & Eidelman, A. I. (2009). Biological and environmental initial conditions shape the trajectories of cognitive and social–emotional developmentacross the first years of life. Developmental Science, 12(1), 194–200.

Field, T. (1992). Infants of depressed mothers. Development and Psychopathology, 2, 49–66.Field, T., Guy, L., & Umbel, V. (1985). Infants’ responses to mothers’ imitative behaviors. Infant Mental Health Journal, 6, 40–44.Gaffney, K. F., Kodadek, M. P., Meuse, M. T., & Jones, G. B. (2001). Assessing infant health promotion. Clinical Nursing Research, 10(2), 102–116.Grace, S. L., Evindar, A., & Stewart, D. E. (2003). The effect of postpartum depression on child cognitive development and behavior: A review and critical

analysis of the literature. Archives of Women’s Mental Health, 6, 263–274.Henning, A., Striano, T., & Lieven, E. V. M. (2005). Maternal speech to infants at 1 and 3 months of age. Infant Behavior and Development, 28, 519–536.Hinde, R. A., & Stevenson-Hinde, J. (1987). Implication of a relationships approach for the study of gender differences. Infant Mental Health Journal, 8,

221–236.Horodynski, M. A., & Gibbons, C. (2004). Rural low-income mothers’ interactions with their young children. Pediatric Nursing, 30, 299–306.Isabella, R. (1998). Origins of attachment: The role of context, duration, frequency of observation, and infant age in measuring maternal behavior. Journal

of Social and Personal Relationships, 15, 538–554.Jones, S., & Moss, H. A. (1971). Age, state, and maternal behavior associated with infant vocalizations. Child Development, 42, 1039–1051.Kemppinen, K., Kumpulainen, K., Raita-Hasu, J., Moilanen, I., & Ebeling, H. (2006). The continuity of maternal sensitivity from infancy to toddler age. Journal

of Reproductive and Infant Psychology, 24(3), 199–212.Landry, S. H., Smith, K. E., Miller-Loncar, C. L., & Swank, P. R. (1997). Predicting cognitive-language and social growth curves from early maternal behaviors

in children at varying degrees of biological risk. Developmental Psychology, 33, 1040–1053.Landry, S. H., Smith, K. E., & Swank, P. R. (2006). Responsive parenting: Establishing early foundations for social, communication, and independent problem-

solving skills. Developmental Psychology, 42, 627–642.Landry, S. H., Swank, P. R., Assel, M. A., Smith, K. E., & Vellet, S. (2001). Does early responsive parenting have a special importance for children’s development

or is consistency across early childhood necessary? Developmental Psychology, 37, 387–403.Laosa, L. M. (1980). Maternal teaching strategies in Chicano and Anglo-American families: The influence of culture and education on maternal behavior.

Child Development, 51, 759–765.Lemelin, J. P., Tarabulsy, G. M., & Provost, M. A. (2006). Predicting preschool cognitive development from infant temperament, maternal sensitivity, and

psychosocial risk. Merrill-Palmer Quarterly, 52(4), 779–806.McElwain, N. L., & Booth-LaForce, C. (2006). Maternal sensitivity to infant distress and nondistress as predictors of infant–mother attachment security.

Journal of Family Psychology, 20, 247–255.Meins, E. (1997). Security of attachment and the social development of cognition. East Sussex, UK: Psychology Press Ltd.Meins, E., Fernyhough, C., Fradley, E., & Tuckey, M. (2001). Rethinking maternal sensitivity: Mothers’ comments on infants’ mental processes predict security

of attachment at 12 months. Journal of Child Psychology & Psychiatry & Allied Disciplines, 42, 637–648.Mertesacker, B., Bade, U., Haverkock, A., & Pauli-Pott, U. (2004). Predicting maternal reactivity/sensitivity: The role of infant emotionality, maternal

depressiveness/anxiety, and social support. Infant Mental Health Journal, 25, 47–61.Milgrom, J., Westley, D. T., & Gemmill, A. W. (2004). The mediating role of maternal responsiveness in some longer term effects of postnatal depression on

infant development. Infant Behavior & Development, 27, 443–454.Paavola, L., Kunnari, S., & Moilanen, I. (2005). Maternal responsiveness and infant intentional communication: Implications for the early communicative

and linguistic development. Child: Care, Health, & Development, 31, 727–735.Parks, P. L., & Arndt, E. K. (1990). Difference between adolescent and adult mothers of infants. Journal of Adolescent Health Care, 11, 248–253.Roberts, E., Bornstein, M. H., Slater, A. M., & Barrett, J. (1999). Early cognitive development and parental education. Infant and Child Development, 8, 49–62.Roe, K. V., Roe, A., Drivas, A., & Bronstein, R. (1990). A curvilinear relationship between maternal vocal stimulation and three-month-olds’ cognitive

processing: A cross-culturation phenomenon. Infant Mental Health Journal, 11, 175–189.Ruddy, M. G., & Bornstein, M. H. (1982). Cognitive correlates of infant attention and maternal stimulation over the first year of life. Child Development, 53,

183–188.Smith, K. E., Landry, S. H., & Swank, P. R. (2006). The role of early maternal responsiveness in supporting school-aged cognitive development for children

who vary in birth status. Pediatrics, 117, 1608–1617.Stevens, J. H., Jr. (1984). Child development knowledge and parenting skills. Family Relations, 33, 237–244.Sumner, G., & Spietz, A. (Eds.). (1994). NCAST caregiver/parent–child interaction teaching manual. Seattle: NCAST Publications, University of Washington,

School of Nursing.Taumoepeau, M., & Ruffman, T. (2008). Stepping stones to others’ minds: Maternal talk relates to child mental state language and emotion understanding

at 15, 24, and 33 months. Child Development, 79(2), 284–302.Taumoepeau, M., & Ruffman, T. (2006). Mother and infant talk about mental states relates to desire language and emotion understanding. Child Development,

77, 465–481.