Language development

Teachingby Listening: The Importance of Adult-Child Conversations to LanguageDevelopment

WHAT’SKNOWNONTHISSUBJECT: It iswell known that adult language input is important to healthy child language development. This knowledgehasmotivated advice that parents

should read to their children regularly.

WHATTHISSTUDYADDS: The results of this study provide evidence that adult-child conversations are at least as important as adult language input.Wepresent empirical evidence andoffer

theoretical reasons for this contention.

abstract OBJECTIVE: To test the independent association of adult language in- put, television viewing, and adult-child conversations on language ac- quisition among infants and toddlers.

METHODS: Twohundred seventy-five families of children aged 2 to 48 monthswhowere representative of the US censuswere enrolled in a cross-sectional study of the home language environment and child languagedevelopment (phase1). Of these, a representative sampleof 71 families continued for a longitudinal assessment over 18 months (phase2). Inthecross-sectionalsample, languagedevelopmentscores wereregressedonadultwordcount, televisionviewing,andadult-child conversations,controlling forsocioeconomicattributes. In the longitu- dinal sample, phase 2 language development scores were regressed onphase1languagedevelopment,aswellasphase1adultwordcount, televisionviewing,andadult-childconversations, controlling forsocio- economic attributes.

RESULTS: In fullyadjustedregressions, theeffectsofadultwordcount were significant when included alone but were partially mediated by adult-childconversations. Televisionviewingwhen includedalonewas significant and negative but was fully mediated by the inclusion of adult-child conversations. Adult-child conversations were significant when included alone and retained both significance and magnitude whenadultword count and television exposurewere included.

CONCLUSIONS: Television exposure is not independently associated with child language development when adult-child conversations are controlled. Adult-child conversations are robustly associated with healthy language development. Parents should be encouraged not merely to provide language input to their children through readingor storytelling, but also to engage their children in two-sided conversa- tions.Pediatrics 2009;124:342–349

CONTRIBUTORS: Frederick J. Zimmerman, PhD,a Jill Gilkerson, PhD,b Jeffrey A. Richards,MA,b Dimitri A. Christakis,MD,MPH,c

Dongxin Xu, PhD,b SharmisthaGray, PhD,b andUmit Yapanel, PhDb

aDepartment of Health Services, School of Public Health, University of California, Los Angeles, California; bDepartment of Research, LENA Foundation, Boulder, Colorado; cOutcomes Research, Children’s Hospital andRegionalMedical Center, Seattle,Washington

KEYWORDS languagedevelopment, reading, television

ABBREVIATIONS PLS—Preschool LanguageScale CI—confidence interval

www.pediatrics.org/cgi/doi/10.1542/peds.2008-2267

doi:10.1542/peds.2008-2267

Accepted for publicationNov 7, 2008

Address correspondence to Frederick J. Zimmerman, PhD, University of California, Los Angeles, Department of Health Services, School of Public Health, Los Angeles, California. E-mail: [email protected]

PEDIATRICS (ISSNNumbers: Print, 0031-4005; Online, 1098-4275).

Copyright©2009by the American Academyof Pediatrics

FINANCIALDISCLOSURE:DrsGilkerson, Xu, Gray, Yapanel, and MrRichardsare employedby the LENA Foundation anon-profit organization that developed thedata-collection product. To the extent that this articlewill raise awareness of their product, theymaybenefit from its publication. However, the specific results of the article donot create a conflict of interest for them. Drs ZimmermanandChristakis have nomaterial interest andnofinancial relationships relevant to this article to disclose.

342 ZIMMERMANet al at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

The amount of language input a child receives before age 3 is significantly and strongly associated with subse- quent language acquisition and cogni- tivedevelopment.1–8Becauseof thisre- lationship, pediatricians and others are encouraged to advise parents to provide as much language input to young children as possible, through reading, storytelling, and simple nar- rationofdailyevents.Althoughthisad- vice is undoubtedly sound, it may not place enough emphasis on the child’s role in language-based exchanges.

The framing of advice offered to par- ents may differ depending on how adult speech is understood to foster child language development. If adult speech input is presented as intrinsi- cally valuable, because it serves as a model for language that children intu- itivelycopy, thenparentscanconclude that themore adult speech the better, even if some of this adult speech comes through television or videos. Manyparentshavedrawnexactlysuch conclusions.9,10 On the other hand, if theprimary valueof adult speech is to potentiate child speech as part of a trial-and-error,experientialprocessof language acquisition, then adult speech is valuable inasmuch as it fos- ters child speech, and either adult speech or electronic stimulus that crowds out child speech may be counterproductive.

Heavy television viewing in the early childhood years has been shown in previousstudies tobeassociatedwith poor development of language and reading andmath skills, although the reasons for this adverse association are still notwell understood.11–13

This analysis uses a large, unique newdata set of parent and child lan- guage use collected naturalistically in the home environment to test the independent contributions of adult language input, child language use, and television viewing on subse-

quent child language development among children who were between theagesof2and36monthsatenroll- ment. We hypothesized that adult speechwouldhaveabeneficial effect on child language development, but that this beneficial effect would be partially mediated by an increase in adult-child conversations.

METHODS

This research relies on data collected in children’s own environments: homes,playgrounds, schools, andany- where else children use or hear lan- guage.Datawerecollected for12-hour periods1dayamonthfor6months(or over 18 months in the longitudinal sample)byusinganewproductcalled LENA (Language Environment Analysis [LENA Foundation, Boulder, CO]). The hardwarecomponent isasmalldigital recordercalledadigital languagepro- cessor, which fits into a pocket on a special vestwornby thechild. Thedig- ital language processor weighs 2 oz and can hold 16 hours of digitally re- corded sound.14 The software compo- nent consists of a digital sound ana- lyzer that produces estimates of the child’s exposure to adult speech, child speech, and television during the re- cordingperiod.

Study Sample

Parents of children aged 2 to 48 months were invited to participate through advertising in local newspa- persanddirect-mail solicitation. From among thoseexpressingan interest, a sample was invited to participate, stratified on maternal education and the age of the child. Because the em- phasis of the study was on language development,bydesignallbut15ofthe children were aged 2 to 36 months. Childrenwhoseparentsreportedade- velopmental or language delay were excluded. Only children in English- speakinghouseholdswere eligible.

Of the 364 participants invited to par- ticipate, 334 (92%) participants com- pleted consent forms and were en- rolled. During the course of the original 6-month study period, 13 par- ticipants dropped out ormoved away, and final assessments could not be scheduled with an additional 8 fami- lies. Five participants were excluded becausetheirrecordingswereall�12 hours induration; 5wereexcludedbe- cause their Preschool Language Scale (PLS) data were judged by raters to have poor validity; and 28 were ex- cludedbecausetheirPLSassessments were not completed during phase 1. These exclusions left a sample of 275 (82%of thoseenrolled). Thefinal sam- ple is highly representative of the US census data with regard to maternal education.

For a randomly chosen day, parents were instructed to begin recording when their childwokeup in themorn- ing and to continue without interrup- tion until bedtime. Parentswere given the option of deleting all data on the device if they felt that anything re- corded during that day compromised their privacy. Only 1 parent chose to exercise this option. Each child con- tributed an average of 4.7 recording sessions throughout the 6-month period.

Participants’ language capacity was formally assessed by a speech lan- guage pathologist by using the Pre- school Language Scale, Fourth Edition (PLS-4).15 These assessments oc- curred throughout the 6-month study period,witheachchildassessedanav- erage of 2.3 times.

A subset of 80 families, selected to be representative of the entire sample stratifiedbyagegroups,wasrecruited toparticipate inan18-month-longcon- tinuationwiththesameprotocolasde- scribed above. This subsample pro- vides the opportunity to conduct longitudinal analyses and is referred

ARTICLES

PEDIATRICS Volume 124, Number 1, July 2009 343 at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

to as the “longitudinal sample” or “phase 2.” Complete data were avail- able for 71 families in the longitudinal sample.

The recruitment strategy and study protocolwere reviewed and approved bytheEssexinstitutionalreviewboard. Additional details on thestudy recruit- ment and data-collection procedures have beenpublishedpreviously.16

Variables

The dependent variable is the PLS-4, a well-validated measure of child lan- guage development in the preschool years.15 ThePLS-4comprehensivelyas- sessesachild’semergingcommunica- tive capacity. It identifies ability in many domains, including gesture, so- cial communication, language struc- ture, phonological awareness, and at- tention.15 Test-retest reliability ranges from .90 to .97 for thesummaryscore, and interrater reliability is .99. The age-normedsummaryscorewasused in this analysis.

Child language exposure and use as well as electronic media (hereafter, “television”) are detected by the Info- ture software’s analysis of the sound files. By analyzing acoustical proper- ties in the file, the software estimates thenumberofwordsspokenbyadults, vocalizations by the key-child, child-

adult conversational turns, and the amount of time the childwas exposed to television. The technical details of this process have been described elsewhere.17–19

To assess fidelity of the software, a subsample of seventy 12-hour ses- sions (anage-stratifiedrandomsam- ple) were coded by human coders, and the resulting estimates were compared with those of the soft- ware. The software achieves a high degreeoffidelity incoding.Cohen’s� of interrater agreement between the machine and human coding of seg- ments into adult speech is 0.65; for television exposure it is 0.57. Among segments that human transcribers identified as adult speech, 82%were correctly identified as such by the software, with �2% erroneously coded as child speech, and 4% erro- neously coded as television. Among segments that the software identi- fied as adult speech, 68% were so identified by human transcribers. The full-fidelity matrix is reported in Table 1. Diagonal elements generally exceed70%, indicatingahighdegree of concordance between machine andhuman transcribers.Wheremis- codings occur, they rarely involve confusion of key variables in this analysis.

Ourmeasureofadult languageinput is the adult word count estimate, which measures the quantity of language in- put by any adult, not just the parent. The software cannot identify whether thespeechwasaddressed to thechild ornot,onlywhetherthechildwasnear enough for it to beheard clearly.

Ourmeasure of child speech is adult- child conversational turns, defined as the number of times that the speaker changes within a single conversation. Aconversation isdefinedasasegment of speech of any length or number of speakers, separated by not �5 sec- onds of silence (or other noise, in- cluding television).18 For example a conversation consisting of adult-child- adult-child would include 2 con- versational turns, whereas a “conver- sation”consistingonlyofadult speech wouldincludenoconversational turns. Conversational turns exhibit modest correlationwith the adultword count, with a correlation coefficient of 0.55.

Television (or other electronicmedia) exposure is most clearly identified by the softwarewhen it is a clear part of the child’s auditory environment. The software cannot identify whether the child was attending to the television while hearing it, and as such cannot distinguishbetweenviewingandback- ground exposure to television, both of whichdegradeyoungchildren’sability to attend to tasks or people.20 For the identification of television exposure, this method compares favorably to parent-report and othermeans of as- sessingchild television viewing,which is notoriously difficult tomeasure.21–24

Thisascertainmenterror introducesa conservative bias into the analyses that follow.

Television viewing was measured in number of hours per day, averaged over the observation periods. Expo- sure toadult language inputwasmea- sured as the estimated number of words spoken by adults in the child’s

TABLE1 FidelityMatrix

Machine-CodedAs: Total,%

Adult Speech,%

Child Vocalization,%

Television, %

Other, %

Sensitivity Among segments human-transcribed as Adult speech 82.0 1.9 3.9 12.2 100 Child vocalization 7.3 76.0 0.1 16.6 100 Television 7.8 0.5 70.5 21.2 100 Other 13.5 4.5 6.3 75.7 100

Specificity Human-codedas Adult speech 67.9 4.1 13.0 5.7 Child vocalization 2.5 69.8 0.2 3.2 Television 0.9 0.2 33.4 1.4 Other 28.7 25.9 53.4 89.7 Total 100 100 100 100

344 ZIMMERMANet al at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

near proximity. To facilitate interpret- ability, we rescaled this variable in thousands. Conversational turns esti- mateswere rescaled in hundreds.

StatisticalMethods

We conducted both cross-sectional and longitudinal regressions. In both sets of regressions, we regressed PLS-4 scores on the recorded mea- sures of adult word count, conversa- tional turns,andtelevision,controlling for sociodemographic variables, in- cluding the child’s age, gender, and race, themother’s and father’s educa- tion level, and household income. Fa- ther’s education was imputed where missing, and amissing-value flagwas included in the regressions. Because themainpredictorvariablesaswellas theoutcomevariablesweremeasured at more than 1 time period for many participants, we averaged the mea- sures over the number of recording sessions and included the number of sessions used to create these aver- ages as covariates. To appropriately weight the different observations for the different number of sessions that contributed to the child averages, we used analytical weights, with the number of sessions as the weighting variable.

In the cross-sectional regressions, we regressed PLS-4 scores in phase 1 on the contemporaneously recorded measures of adult word count, con- versational turns, and television. In the longitudinal regressions, we re- gressed phase 2 PLS-4 scores on phase 1 adult word count, conversa- tional turns, and television, also con- trolling for phase 1PLS-4 scores.

Although the longitudinal analysis is the stronger study design, it also has substantially fewer participants than the cross-sectional analysis. Results from both sets of regressions are reported.

Within each set of regressions, we tested whether the effects of adult word count and television weremedi- ated throughconversational turns. Ac- cordingly,weconducted5regressions within each sample set: 1 each in which the3mainpredictorsweresep- arately included; 1 inwhich all 3main predictors were simultaneously in- cluded; and 1 regression of conversa- tional turns on adult word count and television.

RESULTS

Table 2 presents descriptive statistics for the cross-sectional sample. Chil- dren hear an average of some 13000 words spoken to them by adults each day. Children participate in some 400 adult-child conversational turnsaday.

There is considerable sample varia- tion around these means, and in par- ticular the variation is higher for con- versational turns (for which the coefficientofvariation is54%)thanfor adult speech (coefficient of variation: 34%).

The independent effects of all 3 main predictors on language development are significant in the cross-sectional regressions fully adjusted for social and demographic characteristics (Table 3). Each 1000-word increase in the adult word count is associated witha0.44 increase in thePLS-normed score (95% confidence interval [CI]: 0.09–0.79). On average, each hour of televisionviewingperday towhich the child is exposed is associated with a

TABLE2 Sample Descriptive Statistics

Variable Cross-Sectional Sample (N � 275)

Longitudinal Sample (N � 71)

PLS-4 standardized score (phase 1),mean (SD) 108 (13) 105 (13) PLS-4 standardized score (phase 2),mean (SD) 105 (16) Adultword count (1000s/d),mean (SD) 12.8 (4.4) 13.2 (4.7) Conversational turns (phase 1) (100s/d),mean (SD) 4.1 (2.2) 3.3 (1.9) Conversational turns (phase 2) (100s/d),mean (SD) 4.7 (1.9) Television exposure (foreground) (h/d),mean (SD) 1.0 (0.6) 1.2 (0.7) Child demographics Girl,% 51 58 Black,% 3 4 Latino, nonblack,% 8 8 Other nonwhite race/ethnicity,% 7 7 Age,mo (phase 1),mean (SD) 21.2 (10.8) 13.7 (10.0) Age,mo (phase 2) 27.9 (11.2) Household income

�$20000 annually,% 20 23 $20000–$40000 annually,% 31 35 $40000–$60000 annually,% 19 17 $60000–$100000 annually,% 19 15 �$100000 annually,% 11 10 Mother’s education

� High school,% 21 18 High school graduate,% 24 20 Somecollege,% 29 37 College degree or higher,% 26 25 Father’s education

� High school,% 20 17 High school graduate,% 16 25 Somecollege,% 21 24 College degree or higher,% 24 28 Father’s education data unknown,% 19 6 No. of observation sessions overwhichPLSwas measured,mean (SD)

2.3 (1.2) 2.4 (0.7)

No. of observation sessions overwhichpredictors weremeasured,mean (SD)

4.7 (1.3) 6.0 (1.3)

ARTICLES

PEDIATRICS Volume 124, Number 1, July 2009 345 at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

2.68 decrease in the language score (95%CI: �5.25 to 0.11). Each hundred conversational turnsperday isassoci- ated with a 1.92 increase in the lan- guage score (95%CI: (1.12–2.73).

When all 3 of themain predictors are simultaneously included in theregres- sion, adult word count and television exposure are no longer significant, suggesting complete mediation by conversational turns,whichretains its statistical significance with a slightly higher magnitude (coefficient: 2.08 [95%CI: 0.97–3.19]). In the regression of conversational turns, adult word counthasasignificantandpositiveas- sociation (coefficient 0.28 [95% CI: 0.24–0.32]), whereas television expo- surehasasignificantandnegativeas- sociation (coefficient: �0.51 [95% CI: �0.79 to 0.23]).

A similar pattern obtains in the longi- tudinal regressions, in which phase 1 PLS-4 scores are controlled (Table 4). Adult word count is a significant pre- dictorof subsequentPLS-4 scores (co- efficient: 1.34 [95% CI: 0.59–2.10]). In

this regression, unlike the cross- sectional one, the effect of adult word count is not fully mediated by conver- sational turns, but retains its signifi- canceandmuchof itsmagnitude.Tele- vision viewing is not a significant predictor of language development in any of the longitudinal regressions. Conversational turns is a robust pre- dictor in the longitudinalmodel (coef- ficient: 3.33 [95% CI: 1.58–5.07]), and retains itssignificanceandmuchof its magnitude inthefullysaturatedmodel (coefficient: 2.18 [95% CI: 0.18–4.18]). Adultwordcount predicts subsequent conversational turns (coefficient: 0.17 [95% CI: .06–.29]). Phase 1 language ability is not a significant ormeaning- fulpredictorofphase2conversational turns.

DISCUSSION

The analyses presented here collec- tively make a strong case for the im- portance of adult-child conversations to early child language development. The number of conversational turns

that children have with adult care- givers is robustly and positively as- sociated with scores on a well- validated measure of child language development in a variety of model specifications.

The longitudinalresultsreportedhere, although emerging from a smaller sample than the cross-sectional re- sults, are particularly compelling, be- cause theycontrol for thechild’sbase- line languagedevelopment.Thismodel teases apart the separate effects of a child’sownabilityandachild’sconver- sationswithothersonsubsequent lan- guagedevelopment.

These results are consistent with a number of causal interpretations. It might be that children who are more advancedwith their language abilities are better at initiating or prolonging conversations. It could also be that there is a common but unobserved variable that causes certain children to have high language scores and to engageinalotofconversation.Thesig- nificance of conversational turns in

TABLE3 Cross-Sectional Regression of PLS-4 Scores onAdultWordCount, Conversational Turns, and Television Exposure, and of Conversational Turns onAdultWordCount and Television Exposure

Dependent Variable PLS-4 (Phase 1), Coefficient (95%CI)

PLS-4 (Phase 1), Coefficient (95%CI)

PLS-4 (Phase 1), Coefficient (95%CI)

PLS-4 (Phase 1), Coefficient (95%CI)

CT (Phase 1), Coefficient (95%CI)

Predictors Adultword count (1000 s/d) 0.44 (0.09 to0.79)a �0.16 (�0.63 to0.30) 0.28 (0.24 to0.32)b

Television exposure (foreground) (h/d) �2.68 (�5.25 to0.11)a �1.4 (�3.97 to1.14) �0.51 (�0.79 to0.23)b

Conversational turns (100 s/d) 1.92 (1.12 to2.73)a 2.08 (0.97 to3.19)a

R2 0.23 0.23 0.28 0.28 0.68

Resultsadjustedforchild’sage,gender, race/ethnicity,mother’sandfather’seducation,household income,andnumberofrecordingsessions(N�275).CT indicatesconversational turns. aP � .05. bP � .01.

TABLE4 Longitudinal Regression of Phase 2PLS-4 Scores onPhase 1PLS-4, AdultWordCount, Conversational Turns, and Television Exposure, and of Conversational Turns onAdultWordCount and Television Exposure

Dependent Variable PLS-4 (Phase 2), Coefficient (95%CI)

PLS-4 (Phase 2), Coefficient (95%CI)

PLS-4 (Phase 2), Coefficient (95%CI)

PLS-4 (Phase 2), Coefficient (95%CI)

CT (Phase 2), Coefficient (95%CI)

Predictors Phase 1PLS-4 0.46 (0.24 to0.68)a 0.46 (0.21 to0.70)a 0.32 (0.10 to0.54)a 0.38 (0.16 to0.61)a 0.03 (�0.01 to0.06) Adultword count (1000s/d) 1.34 (0.59 to2.10)a 0.82 (0.05 to1.60)b 0.17 (0.06 to0.29)a

TV exposure (foreground) (h/d) 1.45 (�4.71 to7.61) �0.38 (�5.90 to5.14) 0.14 (�0.66 to0.94) Conversational turns (100s/d) 3.33 (1.58 to5.07)a 2.18 (0.18 to4.18)b

R2 0.68 0.60 0.69 0.71 0.56

Results adjusted for child’s age, gender, race/ethnicity,mother’s and father’s education, household income, andnumber of recording sessions (N � 71). aP � .01. bP � .05.

346 ZIMMERMANet al at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

predicting phase 2 language scores evenwhen phase 1 language ability is controlled ina longitudinal regression militatesagainst such interpretations, asdoes thenonsignificanceofphase1 language ability in predicting phase 2 conversational turns.

Finally, it could be that the child-adult conversationsare themselvescausing language development. Such a causal conclusion would be consistent with several strands recently introduced in the empirical and theoretical litera- tures on child language acquisition. Parents are most efficient at promot- ing child language development when they calibrate their own speech to be just challenging enough for the child, neither so simplistic that the child learnsnothing fromtheparent’smodel, nor so sophisticated that the child is bewildered. This just-challenging zone hasbeen termedthe“zoneofproximal development.”25–27 Because maintain- ing adult speech in the zone of proxi- mal development depends on the adult’s being in touch with the child’s rapidly changingabilities, frequent ex- posure to child language through adult-child conversations may help keep the adult’s own speech in the zone of proximal development.28 Child language development has also been shown to benefit from active correc- tion of errors by adult speakers.29

More conversations mean more op- portunities for mistakes and, there- fore, corrections. In addition, more conversations may afford the child more opportunity to practice and con- solidate newly acquired language. Fi- nally, more conversations may indi- catemoreadult responsiveness to the child’s communication in general.

These results have meaningful impli- cations. Much of the advice given to parents has focused on the value of reading to children as a way of facili- tating adult language input. Indeed, reading-promotion programs on their

own have been shown to produce im- provements in children’s home liter- acy environment at very low cost.30–32

Yet the most effective reading is dia- logic reading,which involves explicitly soliciting language use by the child.33–36 More generally, parents should be taught that although adult speech is valuable, an equally impor- tant goal should be to get kids talking asmuchaspossible.

Television viewing before age 2 has previously been associated with de- lays in language development and poor reading skills.11,13,37 Television may adversely affect language devel- opment by limiting opportunities for parental language input38 or by limit- ing opportunities for child speech. Re- cent research has shown that al- though preschoolers can learn vocabulary fromtelevision,whenused without parental interaction, televi- sion isan ineffectivemediumfor incul- cating language skills among infants and toddlers.20,39–41 When conversa- tional turnsare included in theregres- sions, television drops from signifi- cance, suggesting that the adverse effects of television exposure, if any, would operate by reducing opportuni- ties for adult-child interactions.

This research leaves many important questions unanswered. We were not able to separately identify the effects of child language use and conversa- tional turns, which were highly corre- lated with each other (correlation co- efficient:0.84). Itwillbeuseful infuture research to teaseout the independent contributionof these2similar factors. Itwasbeyondthescopeof thisstudyto identify all the factors that promote adult-child conversations, but this is clearly a high priority for future re- search. The novel technology used in this research clearly has tremendous promise but is not yet able to distin- guishmore nuanced, but undoubtedly very important,elementsofchild-adult

interactions, includingemotional tone, questions and responses, whether adult speech is directed to the key child or merely overheard by him or her, and other qualitative differences among dyads. Recent research sug- gests that children are able to learn vocabulary by overhearing adult speech,andthis intriguingfinding, too, suggests important new avenues for thisresearch.42Moreworkneedstobe performed to refine the machine- based measurement of television ex- posure, an important component of the early childhood environment.10,13

CONCLUSIONS

This research provides strong, albeit not absolute, evidence fromnaturalis- tic observations that adult-child con- versationsareanessentialcomponent of child language development. Parents should continue to be encouraged to providespeechinputtotheirchildrenby talkingtothem,readingthembooks,and bytellingthemstories.At thesametime, it should bemade clear to parents that an important goal of this talk is to elicit talk from the child. Reading and story- telling should be punctuated by ques- tions and exchanges, and it may be appropriate to counsel parents to en- courage parent-child conversations. Parents should strive to read and talk with children and not merely to them. Parent-child interactions are bestwhen theyarea two-waystreet.

ACKNOWLEDGMENTS

Dr Zimmerman’s time on this project was supported in part by National In- stitute of Mental Health grant 5K01MH64461-5. Data collection and cleaning were paid for by The LENA Foundation.

We gratefully acknowledge Terrance Paul for conceivingof the LENAsystem and for personally funding and direct- ing its development aswell as the de- velopment of the LENA Natural Lan- guageCorpus.

ARTICLES

PEDIATRICS Volume 124, Number 1, July 2009 347 at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

REFERENCES

1. HartB,RisleyTR.MeaningfulDifferences in theEverydayExperienceofYoungAmericanChildren. Baltimore,MD: P.H. Brookes; 1995

2. Huttenlocher J. Early vocabulary growth: relation to language input and gender. Dev Psychol. 1991;27(2):236–248

3. Huttenlocher J. Language input and language growth.PrevMed. 1998;27(2):195–199

4. Pan BA, Rowe ML, Singer JD, Snow CE. Maternal correlates of growth in toddler vocabulary production in low-income families.Child Dev. 2005;76(4):763–782

5. Tamis-LemondaCS, BornsteinMH, Kahana-KalmanR, Baumwell L, Cyphers L. Predicting variation in thetimingof languagemilestones in thesecondyear:aneventshistoryapproach.JChildLang. 1998;25(3):675–700

6. Shonkoff JP, PhillipsD.FromNeurons toNeighborhoods: TheScienceof EarlyChildhoodDevelop- ment. Washington, DC: National AcademyPress; 2000

7. Arterberry ME, Bornstein MH, Midgett C, Putnick DL, Bornstein MH. Early attention and literacy experiences predict adaptive communication. First Lang. 2007;27(2):175

8. BornsteinMH,HaynesOM. Vocabulary competence in early childhood:measurement, latent con- struct, andpredictive validity.Child Dev. 1998;69(3):654–671

9. Rideout VJ, Vandewater EA, Wartella EA. Zero to Six: Electronic Media in the Lives of Infants, Toddlers, andPreschoolers. Menlo Park, CA: Kaiser Family Foundation; 2003

10. Zimmerman FJ, Christakis DA,Meltzoff AN. Television andDVD/video viewing in children younger than 2 years.ArchPediatr AdolescMed. 2007;161(5):473–479

11. ZimmermanFJ,ChristakisDA.Children’stelevisionviewingandcognitiveoutcomes:alongitudinal analysis of national data.ArchPediatr AdolescMed. 2005;159(7):619–625

12. Zimmerman FJ, Christakis DA. Associations between content types of earlymedia exposure and subsequent attentional problems.Pediatrics. 2007;120(5):986–992

13. Zimmerman FJ, Christakis DA, Meltzoff AN. Associations between media viewing and language development in childrenunder age 2 years. J Pediatr. 2007;151(4):364–368

14. FordM,BaerCT,XuD,YapanelU,GraySS.TheLENALanguageEnvironmentAnalysisSystem:Audio Specifications.Boulder, CO: Infoture; October 2007

15. Zimmerman IL, Steiner VG, Pond RE. Preschool Language Scale. 4th ed. San Antonio, TX: Psycho- logical Corporation; 2002

16. GilkersonJ,RichardsJA.The InfotureNatural LanguageStudy.Boulder, CO: Infoture, Inc; October 2007

17. FordM,BaerCT. The LENA languageenvironmentanalysis system: audio specifications. Available at: www.infoture.org/TechReport.aspx/Audio�Specifications/ITR-03-2. Accessed November 25, 2008

18. GraySS, BaerCT, XuD, YapanelU. The LENA LanguageEnvironmentAnalysis System: The Infoture TimeSegment (ITS) File.Boulder, CO: Infoture, Inc; October 2007

19. YapanelU, GraySS, XuD.Reliability of the LENA LanguageEnvironmentAnalysis System inYoung Children’s Natural HomeEnvironment.Boulder, CO: Infoture, Inc; October 2007

20. AndersonDR, Pempek TA. Television and very young children.AmBehavSci. 2005;48(5):505–522

21. Anderson DR, Field DE, Collins PA, Lorch EP, Nathan JG. Estimates of young children’s timewith television: amethodological comparison of parent reportswith time-lapse video homeobserva- tion.Child Dev. 1985;56(5):1345–1357

22. Borzekowski D, Robinson T. Viewing the viewers: ten video cases of children’s television viewing behaviors. JBroadcast ElectrMedia. 1999;43(4):506–528

23. Robinson JL, Winiewicz DD, Fuerch JH, Roemmich JN, Epstein LH. Relationship between parental estimate and an objectivemeasure of child televisionwatching. Int J Behav Nutr Phys Act. 2006; 3:43

24. Vandewater EA, Lee S.Measuring children’smedia use in the digital age: issues and challenges. AmBehavSci. 2009;52(8):1152–1176

25. TaumoepeauM,RuffmanT.Steppingstonestoothers’minds:maternaltalkrelatestochildmental state language and emotion understanding at 15, 24, and 33 months. Child Dev. 2008;79(2): 284–302

26. FernyhoughC. The dialogicmind: a dialogic approach to the highermental functions.New Ideas Psychol. 1996;14(1):47–62

27. Meins E, Fernyhough C, Wainwright R, Das Gupta M, Fradley E, Tuckey M. Maternal mind- mindedness and attachment security as predictors of theory ofmind understanding. Child Dev. 2002;73(6):1715–1726

348 ZIMMERMANet al at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

28. Tamis-LeMondaCS, BornsteinMH, Baumwell L.Maternal responsiveness and children’s achieve- ment of languagemilestones.Child Dev. 2001;72(3):748–767

29. ChouinardMM,Clark EV. Adult reformulationsof child errors asnegative evidence. JChild Lang. 2003;30(3):637–669

30. MendelsohnAL,MogilnerLN,DreyerBP,etal. The impactofaclinic-based literacy interventionon languagedevelopment in inner-city preschool children.Pediatrics. 2001;107(1):130–134

31. WeitzmanCC,RoyL,WallsT, TomlinR.Moreevidence forreachoutandread:ahome-basedstudy. Pediatrics. 2004;113(5):1248–1253

32. Sharif I,RieberS,OzuahPO,ReiberS.Exposuretoreachoutandreadandvocabularyoutcomes in inner city preschoolers. JNatlMedAssoc. 2002;94(3):171–177

33. Arnold DH. Accelerating language development through picture book reading: replication and extension to a videotape training format. J EducPsychol. 1994;86(2):235–243

34. SénéchalM. The differential effect of storybook reading on preschoolers’ acquisition of expres- sive and receptive vocabulary. J Child Lang. 1997;24(1):123–138

35. Hargrave AC, SénéchalM. A book reading interventionwith preschool childrenwhohave limited vocabularies: the benefits of regular reading anddialogic reading. Early Child ResQ. 2000;15(1): 75–90

36. Whitehurst GJ, Lonigan CJ. Child development and emergent literacy. Child Dev. 1998;69(3): 848–872

37. WrightJC,HustonAC,MurphyKC,etal.Therelationsofearlytelevisionviewingtoschoolreadiness and vocabulary of children from low-income families: the earlywindowproject. Child Dev. 2001; 72(5):1347–1366

38. TanimuraM, OkumaK, KyoshimaK. Television viewing, reduced parental utterance, and delayed speech development in infants and young children. Arch Pediatr Adolesc Med. 2007;161(6): 618–619

39. Krcmar M, Grela B, Lin K. Can toddlers learn vocabulary from television? An experimental ap- proach.Media Psychol. 2007;10(1):41–63

40. Kuhl PK, TsaoFM, LiuHM. Foreign-languageexperience in infancy: effectsof short-termexposure and social interaction onphonetic learning.ProcNatl AcadSci U SA. 2003;100(15):9096–9101

41. BarrR,MuentenerP, Garcia A. Age-related changes indeferred imitation from televisionby 6- to 18-month-olds.Dev Sci. 2007;10(6):910–921

42. Akhtar N, Jipson J, Callanan MA. Learning words through overhearing. Child Dev. 2001;72(2): 416–430

ARTICLES

PEDIATRICS Volume 124, Number 1, July 2009 349 at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2008-2267 2009;124;342Pediatrics

Dongxin Xu, Sharmistha Gray and Umit Yapanel Frederick J. Zimmerman, Jill Gilkerson, Jeffrey A. Richards, Dimitri A. Christakis,

Language Development Teaching by Listening: The Importance of Adult-Child Conversations to

Services Updated Information &

http://pediatrics.aappublications.org/content/124/1/342 including high resolution figures, can be found at:

References http://pediatrics.aappublications.org/content/124/1/342#BIBL This article cites 32 articles, 4 of which you can access for free at:

Subspecialty Collections

milestones_sub http://www.aappublications.org/cgi/collection/growth:development_ Growth/Development Milestones al_issues_sub http://www.aappublications.org/cgi/collection/development:behavior Developmental/Behavioral Pediatrics following collection(s): This article, along with others on similar topics, appears in the

Permissions & Licensing

http://www.aappublications.org/site/misc/Permissions.xhtml in its entirety can be found online at: Information about reproducing this article in parts (figures, tables) or

Reprints http://www.aappublications.org/site/misc/reprints.xhtml Information about ordering reprints can be found online:

at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from

DOI: 10.1542/peds.2008-2267 2009;124;342Pediatrics

Dongxin Xu, Sharmistha Gray and Umit Yapanel Frederick J. Zimmerman, Jill Gilkerson, Jeffrey A. Richards, Dimitri A. Christakis,

Language Development Teaching by Listening: The Importance of Adult-Child Conversations to

http://pediatrics.aappublications.org/content/124/1/342 located on the World Wide Web at:

The online version of this article, along with updated information and services, is

1073-0397. ISSN:60007. Copyright © 2009 by the American Academy of Pediatrics. All rights reserved. Print

the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, has been published continuously since 1948. Pediatrics is owned, published, and trademarked by Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it

at UNIV OF CALIFORNIA DAVIS on September 30, 2019www.aappublications.org/newsDownloaded from