Sc

CJa

b

a

ARRAA

KPSMLE

nct2fiflCaieeD&ii(w

1

h0

Early Childhood Research Quarterly 29 (2014) 333–344

Contents lists available at ScienceDirect

Early Childhood Research Quarterly

upporting family conversations and children’s STEM learning in ahildren’s museum

atherine A. Hadena,∗, Erin A. Janta, Philip C. Hoffmana, Maria Marcusa,acqueline R. Geddesa, Suzanne Gaskinsb

Loyola University Chicago, Department of Psychology, 1032 W. Sheridan Road, Chicago, IL 60660, United StatesNortheastern Illinois University, Department of Psychology, 5500 North St. Louis Avenue, Chicago, IL 60625, United States

r t i c l e i n f o

rticle history:eceived 23 August 2013eceived in revised form 28 March 2014ccepted 8 April 2014vailable online 21 April 2014

eywords:

a b s t r a c t

This study tested the effectiveness of a facilitated educational program in a museum for promoting familyconversations and children’s learning about STEM. A sample of 130 families (71 European-American;33 African-American; and 26 Hispanic-American) with children M age = 6.42 years were observed ina building construction exhibit. Prior to building, families were randomly assigned to conditions thatvaried in terms of the instructions about a key engineering principle and elaborative question-askingthey received. Conversation instruction resulted in adults’ asking double the number of Wh-questions

arent–child conversationTEM talkuseum research

earningthnicity

compared to families who did not receive the instruction. The building instruction was important inpromoting increases in adults’ STEM-related talk during the building activity, as well as in the children’sSTEM talk when prompted for information about what they had learned. The effects of the instructions didnot vary by families’ ethnic background. Implications for facilitating family conversations and children’slearning related to STEM are discussed.

© 2014 Elsevier Inc. All rights reserved.

The education of U.S. students in science, technology, engi-eering, and mathematics (STEM) subjects and fields has receivedonsiderable attention in recent years. There is the growing sensehat addressing the “STEM pipeline problem” (Sanders, 2009, p.2) – the fact that the numbers of individuals pursuing STEMelds is not sufficient to meet demand – requires bolstering both

ormal in-school educational opportunities and informal STEMearning experiences in non-school settings (National Researchouncil, 2007, 2009). In particular, researchers and educatorsre being called on to answer questions about what and hownformal learning experiences can foster interest in and knowl-dge of STEM (Callanan & Oakes, 1992; Crowley, Callanan, Jipson,t al., 2001; Crowley, Callanan, Tenenbaum, & Allen, 2001; Falk &ierking, 2002; Gelman & Brenneman, 2004; NRC, 2009; Palmquist

Crowley, 2007). One important characteristic of informal learnings that it is frequently social. Consistent with notions of scaffold-

ng (Wood, Bruner, & Ross, 1976) drawn from sociocultural theoryBerk, 2001; Gauvain, 2000; Rogoff, 1990; Vygotsky, 1978), theays adult caregivers and children behave and talk together while

∗ Corresponding author at: Department of Psychology, Loyola University Chicago,032 W. Sheridan Road, Chicago, IL 60660, United States. Tel.: +1 7735088226.

E-mail address: chaden@luc.edu (C.A. Haden).

ttp://dx.doi.org/10.1016/j.ecresq.2014.04.004885-2006/© 2014 Elsevier Inc. All rights reserved.

visiting museums, aquariums, zoos, and the like can enhance chil-dren’s STEM learning (see Haden, 2010; Leinhardt, Crowley, &Knutson, 2002; NRC, 2009; for reviews). Family interactions in suchinformal learning environments when children are young are fur-ther linked with children’s STEM achievement when they enterschool and beyond (Duncan et al., 2007; Tenenbaum, Snow, Roach,& Kurland, 2005). From the perspective of the pipeline problemthen, family conversations in museum exhibits designed for STEMlearning may be important in building a foundation for children’sfuture STEM educational and career pursuits (NRC, 2009; Uttal et al.,2013).

Observational studies of family interactions in museums showthat children’s conversations with their caregivers during hands-onlearning activities are related to the quality of their engage-ment in exhibits (Crowley & Callanan, 1998; Crowley, Callanan,Jipson, et al., 2001; Crowley, Callanan, Tenenbaum, et al., 2001;Gleason & Schauble, 1999; NRC, 2009). Moreover, research link-ing parent–child conversations to children’s understanding andremembering of personally experienced events in general (Haden,Ornstein, Eckerman, & Didow, 2001; Hedrick, San Souci, Haden, &

Ornstein, 2009; McGuigan & Salmon, 2004, 2006; Tessler & Nelson,1994), and STEM learning experiences in museums specifically(Borun, Chambers, Dristas, & Johnson, 1997; Callanan & Jipson,2001; Crowley, Callanan, Jipson, et al., 2001; Crowley, Callanan,

3 Resea

T2PecwsmSCtaiippholsa

teeieHtiiChb&pdgaosiwrsmf

deGoaiis(&smncmr2b

34 C.A. Haden et al. / Early Childhood

enenbaum, et al., 2001; Crowley & Jacobs, 2002; Ellenbogen,002; Falk & Dierking, 1992; Gleason & Schauble, 1999; NRC, 2009;almquist & Crowley, 2007; Rigney & Callanan, 2011; Tenenbaumt al., 2005; Valle & Callanan, 2006) points to particular forms ofonversation that may be especially important for learning. Otherork further documents that families from diverse ethnic and

ocioeconomic backgrounds engage in types of talk that can pro-ote children’s early science understanding and skills (NRC, 2009;

iegel, Esterly, Callanan, Wright, & Navarro, 2007; Tenenbaum &allanan, 2008). This descriptive work sets the stage for this studyhat involves an experimental design, allowing causal statementsbout the influence of conversation on STEM learning. More specif-cally, the key manipulation in the current research is aimed atncreasing the frequency with which adult caregivers engage inarticular kinds of talk that previous work suggests should sup-ort children’s STEM learning. Also, because little is known aboutow museum staff and other educators can build on families’ fundsf knowledge (González, Moll, & Amanti, 2013) and facilitate familyearning conversations (see Pattison & Dierking, 2012, for discus-ion), this study is aimed at taking steps in addressing these gapss well.

The research took place in the context of a building construc-ion exhibit at a children’s museum. The emphasis on buildingngineering in this exhibit is important for several reasons. First,ngineering is an excellent domain in which to study STEM learn-ng in informal contexts. Children and adults are interested in andnjoy building projects of the sort used in this study (Benjamin,aden, & Wilkerson, 2010). Some have suggested early exposure

o fun and creative science and engineering projects as a way toncrease the quality and diversity of students pursuing engineer-ng and technology education paths (Carlson & Sullivan, 2004;unningham, 2009; Uttal et al., 2013). Second, young childrenave much to learn about properties of materials, stability, andracing that are essential to successful engineering (Davis, Ginns,

McRobbie, 2002). Third, engineering emphasizes STEM-relatedroblem-solving skills, including defining the problem, consideringifferent solutions, testing hypotheses, and so forth, and inte-rates science and mathematics in ways that make these topicsccessible to young children (NRC, 2009; Uttal et al., 2013). More-ver, the process of problem solving and feedback (physical andocial) is observable to both participants and researchers, as fam-lies work together to solve engineering problems. Finally, prior

ork (Benjamin, Haden, & Wilkerson, 2010) shows that there areeliable differences in how adult caregivers and children togetherolve simple engineering challenges, and these differences provideodels for what is likely more or less effective in inspiring and

acilitating STEM learning.It has been suggested that question-asking and answering is fun-

amental to supporting learning in informal environments (Borunt al., 1997; Falk & Dierking, 2002; Haden, 2010; NRC, 2009).iven this, one focus in this study was on increasing the numberf open-ended Who, What, Where, Why, and How type-questionsdults asked during interactions with their children in the build-ng exhibit. These so-called Wh-questions have been highlightedn prior research as important for enhancing children’s under-tanding and the encoding of information across multiple settingsFivush, Haden, & Reese, 2006; Haden, 2010; Jant, Haden, Uttal,

Babcock, 2014). Caregivers’ Wh-questions can call attention toalient aspects of an experience as it unfolds, and help them deter-ine what children may or may not know. Moreover, by requesting

ames, descriptions, actions, explanations, and so forth, caregiversan help children make sense of an experience in ways that may

ake it more accessible for use when future opportunities for

emembering and learning arise (Ellenbogen, 2002; Haden et al.,001; Jant et al., 2014). Caregivers’ open-ended questions can alsoe essential in motivating sustained engagement in science-related

rch Quarterly 29 (2014) 333–344

activities in ways that may be crucial for early science learning(Haden, 2010; Humphrey & Gutwill, 2005; Schauble, 1996).

Several studies have shown that in contrast to caregivers whoask few Wh-questions during interactions with their children,parents who ask many Wh-questions have children who showgreater understanding, retention, and subsequent recall of personalexperiences (Boland, Haden, & Ornstein, 2003; Hedrick, Haden,& Ornstein, 2009), including experiences in museums (Benjaminet al., 2010; Jant et al., 2014; Tessler & Nelson, 1994). Given theseresults, we wanted to increase the frequency with which adultsin this study asked Wh-questions as they interacted with chil-dren in the STEM-related exhibit. If adults could comply with ourinstructions to ask Wh-questions while building with children,then we expected that the children would demonstrate betterunderstanding and learning from the experience. We also exam-ined children’s responding to caregivers’ Wh-questions, becausechildren’s responsiveness during events has also been linked totheir later memory reports (Hedrick, Haden, et al., 2009; Ornstein,Haden, & Hedrick, 2004). Indeed, it has been suggested that itmight not just be the sheer frequency of Wh-questions asked byadult caregivers, but children’s responsiveness to these questions,that most strongly predicts learning and retention of information.Additionally, in cases where knowledge is lacking, and questionsare not met with child responses, Wh-questions may in turn leadto parents’ explanations that have been highlighted in museumresearch as also contributing substantially to children’s STEM learn-ing (Callanan & Jipson, 2001; Crowley & Callanan, 1998; Crowley,Callanan, Tenenbaum, et al., 2001; Crowley & Jacobs, 2002; NRC,2009; Tare, French, Frazier, Diamond, & Evans, 2011; Tenenbaum& Callanan, 2008; Tenenbaum, Callanan, Alba-Speyer, & Sandoval,2002).

Finally, during the building activity, we also indexed adults’talk with children about the scientific method (e.g., planning,testing ideas), technology (e.g., building materials, techniques),engineering (e.g., building strength, bracing), and math (e.g., quan-tity, height). In this way, the current study also connects withresearch that explores how the frequency of specific kinds of lan-guage inputs, such as spatial and relational language (Loewenstein& Gentner, 2005; Pruden, Levine, & Huttenlocher, 2011), num-ber words (Gunderson & Levine, 2011; Levine, Gunderson, &Huttenlocher, 2011; Levine, Suriyakham, Huttenlocher, Rowe, &Gunderson, 2011), emotion and mental state talk (Adams, Kuebli,Boyle, & Fivush, 1995; Laible, 2004; Lohmann & Tomasello, 2003;Rudek & Haden, 2005; Taumoepeau & Ruffman, 2008; Welch-Ross,Fasig, & Farrar, 1999), and so on, predicts children’s skills in relateddomains. In this case, we suggest that if our intervention – withoutexplicitly prompting for it – could increase adults’ STEM talk withchildren in the building exhibit, children’s STEM learning similarlymight be increased.

We were encouraged that our instructions could indeed gener-alize in this way because of Benjamin et al.’s (2010) findings from astudy involving a primarily European-American sample of 6-year-olds and their adult caregivers who were offered building and/orconversation instructions prior to entering a building exhibit. InBenjamin et al., families who received the information about a keyengineering principle – bracing – talked more about engineering(e.g., “Why do you think this is wobbling?” “This doesn’t seem verysturdy.”), when compared with those who did not receive build-ing instructions. Caregivers asked more Wh-questions when theyreceived conversation instructions than if not. Most important, thecombination of building and conversation instructions was linkedto the children’s increased talk about engineering (e.g., “We added a

triangle so it wouldn’t wobble.” “We needed to brace it.”) immediatelyafter building, and children’s remembering of engineering-relatedinformation 1-day and 2-weeks following the museum visit. In thecurrent study, we developed a coding system to capture a full range

Resea

ompt

we&iWeaBbiemtiwo

wiatftgSptrvttsWtoestcBesus(tornti

hosgMfwts

C.A. Haden et al. / Early Childhood

f STEM talk in the exhibit, including discussion of the scientificethod, technology, engineering and math, consistent with the

erspective that these all come together in the problem-solvingask families are presented with in the building exhibition.

This study also expands upon the growing pool of experimentalork linking conversation to children’s learning from personally

xperienced events (Boland et al., 2003; Jant et al., 2014; Tessler Nelson, 1994). The method of instruction was unique in that it

nvolved a specially created, facilitated STEM education program.e modeled the program after the kinds of on-the-exhibit-floor

ducational activities staff conduct daily at museums to person-lize and enhance visitors’ learning. This approach contrasts withenjamin et al.’s (2010) study that was conducted in a differentuilding construction exhibit than was used in this study, and

nvolved instructions from an examiner in a room away from thexhibit space. Here we aligned the aims of this research with theuseum’s goals of maximizing opportunities for STEM learning in

heir new building exhibit (c.f. Callanan, 2012). We developed andmplemented the instructional program in the exhibit in a way that

ould be sustainable for the museum, and felt like an integral partf the exhibit experience to visitors.

More specifically, prior to building, some families interactedith Inspector Sturdy – a researcher playing the role of a build-

ng inspector. Inspector Sturdy provided families with a tip about key engineering principle: the use of triangular cross-bracingo build strong buildings (Inspector Sturdy Build-Only). To otheramilies, Inspector Sturdy provided both the engineering tip and aip to ask Wh-questions (Inspector Sturdy Build + Talk). For a thirdroup (Build + Talk Signs), two signs that had been used by Inspectorturdy to summarize the engineering and conversation tips werelaced in the exhibit, although families were not explicitly directedo read these signs. Finally, there was a group of families whoeceived no instruction and saw no added signage, thereby pro-iding a semi-naturalistic comparison group. We hypothesized thathe provision of the building tip, either alone or in combination withhe conversation instructions, would result in families buildingturdy buildings featuring triangular bracing. Those who receivedh-question instructions were expected to ask more of these ques-

ions than those not receiving the conversation tip. Children’s ratef responding to these questions might or might not be differ-nt across groups, although if caregivers asked more questions inome conditions, then their children would have a greater oppor-unity to respond. Research has suggested that signage in exhibitsan also impact caregiver–child talk and behaviors (Allen, 2004;orun, 2002; Gutwill, 2006; Hohenstein & Tran, 2007; Leinhardtt al., 2002; Rowe, 2005), but we expected that Inspector Sturdy’signs alone would be less effective for these outcomes than whensed as part of the interactive program. Moreover, we hypothe-ized that when compared with those who received no instructionsControls), adults who received both the building and conversa-ion instructions from Inspector Sturdy would reference STEM mostften during their interactions with children in the exhibit. Thisesult would be important in demonstrating that both the engi-eering tip and prompts to ask Wh-questions did more than affecthe behaviors directly instructed and worked more generally toncrease STEM-relevant talk in the exhibit.

Another important aspect of our project is an examination ofow variations in family background affect interaction patterns andutcomes. Research suggests differences in parent–child conver-ations among families from different ethnic and socioeconomicroups (Fivush & Haden, 2003; Heath, 1982, 1983; Hoff, 2003;iller, Fung, Chen, & Boldt, 2012; NRC, 2009). For example, parents

rom higher socioeconomic backgrounds use a greater number ofords, more diverse vocabulary, and more complex syntax when

alking with and around their children than do parents from lowerocioeconomic backgrounds (Hart & Risley, 1992; Hoff, 2003, 2006;

rch Quarterly 29 (2014) 333–344 335

Huttenlocher, Vasilyeva, Waterfall, Vevea, & Hedges, 2007). Otherwork suggests ethnic variations in parents’ encouragement of theirchildren’s questions and participation in conversations (Goody,1978; Miller, 1982, 1994).

Parents from different backgrounds also vary in their teach-ing behaviors with their children, including the concepts theyteach, the level of their teaching, and the amount and valenceof feedback provided to their children (see Tamis-LeMonda, Sze,Ng, Kahana-Kalman, & Yoshikawa, 2013, for review). Some stud-ies find differences by ethnicity and schooling in the frequency ofexplanatory conversations about science (NRC, 2009; Siegel et al.,2007; Tenenbaum & Callanan, 2008; Tenenbaum et al., 2002).Importantly, the task or activity being performed can minimizeor increase these differences in interaction styles. For example, inGaskin’s work in a children’s museum (2008a, 2008b) providing abuilding goal (“See how close to the clouds you can get.”) increasedjoint collaborative engagement for European-American families,whereas Hispanic-American families demonstrated high levels ofcollaborative activity with or without a building goal. Mindful thatan important piece of the STEM education problem is increasingthe diversity of the pipeline of individuals entering science fields,this study featured an examination of the effectiveness of the fullinstructional program – involving Inspector Sturdy’s tips aboutengineering and conversation – for European-American, African-American, and Hispanic-American families relative to families fromthese three groups who were in the control condition.

In addition to scoring the frequency of types of adults’ talkand children’s responding during the construction projects, andgauging of the sturdiness of the resulting structures, we assessedlearning by observing families’ engagement with a novel featureof the exhibit. During the families’ building activity, a camera con-nected to a computer kiosk snapped timed photographs of them.After their buildings were built, families were invited throughcomputer prompts to select pictures from the set taken duringbuilding, and then to respond to questions about what they learned.From these photo-narratives we measured adults’ and children’stalk about STEM, hypothesizing that those in the Inspector Sturdyconditions might include more different types of talk about STEMin their photo-narratives than those in the control group. Takentogether, our analyses were aimed at determining the effectivenessof our Inspector Sturdy program in enhancing family conversationsand children’s learning about STEM.

1. Method

1.1. Participants

The sample consisted of 130 families (71 European-American,Non-Hispanic; 33 African-American; 26 Hispanic-American) whowere recruited as they entered the Skyline building exhibit at theChicago Children’s Museum. Families were invited to participate ifthey had one (target) child that was between four and eight yearsold. The average age of the target children (69 male; 61 female)was 6.42 years. Table 1 lists the number of families, average agesof the target children, and the number of adults and children inthe family group, by experimental and ethnic group. Each familyhad at least one parent in the group, and most often, if there wasa second adult, it was another parent. Additionally, gender of thetarget child, and the gender distribution of the adults in each fam-ily group is shown in Table 1. Parents reported via questionnaireboth family race/ethnicity (checking a box) and the primary lan-

guage spoken at home (filling in a blank). The primary languagespoken at home was English for 90% of the sample and Spanish for8.5%. During informed consent, families elected to participate inthe study in English (n = 120) or Spanish (n = 10). Of the 128 who

336 C.A. Haden et al. / Early Childhood Research Quarterly 29 (2014) 333–344

Table 1Participant demographics for full sample.

Demographic variable Build + Talk Build-only Signs Control

European-American

AfricanAmerican

HispanicAmerican

European-American

AfricanAmerican

HispanicAmerican

No. of families 23 15 11 16 17 15 18 15Child age 6.56 (1.16) 6.40 (1.18) 6.90 (1.37) 6.44 (1.03) 6.35 (1.00) 6.07 (1.22) 6.39 (1.42) 6.33 (1.05)No. of adults 1.60 (.58) 1.40 (.51) 1.36 (.50) 1.44 (.51) 1.59 (.51) 1.67 (.62) 1.44 (.51) 1.87 (.52)No. of children 2.13 (1.14) 2.07 (.70) 2.36 (1.12) 1.87 (.62) 1.88 (.78) 2.00 (1.00) 1.88 (.83) 2.27 (1.16)Gender of target child

Female 12 9 5 7 8 7 7 6Male 11 6 6 9 9 8 11 9

Gender of adults in visitor groupsFemale + Male 12 6 4 7 9 9 4 6Female-only 10 8 5 5 6 3 14 8

4

N

rAd8e(eA(car

1

Mat(wkwiWcaetr

1

1

AIBCiym

BdatS

Male-only 1 1 2

ote: SDs are in parentheses.

eported education information, 87% (65 European-American, 28frican American, 18 Hispanic American) reported holding a collegeegree or higher; 13% (5 European-American, 4 African American,

Hispanic American) reported high school as the highest level ofducation completed. Of the 92 who reported family incomes, 40%23 European-American, 9 African American, 5 Hispanic American)arned $100,000 or more; 27% (10 European-American, 8 Africanmerican, 7 Hispanic American) earned $50,000–100,000; and 33%

3 European-American, 15 African American, 12 Hispanic Ameri-an) earned less than $50,000/year. Participating families received

family pass for a free return visit to the museum, and the childreneceived stickers.

.2. Museum setting

The Skyscraper Challenge area of the Chicago Children’suseum’s Skyline exhibit provided the setting for the research. At

workstation in the Skyscraper Challenge, families built with plas-ic pieces, including triangle-shaped braces, struts, nuts, and boltssee Fig. 1a). The building challenge to build a tall, stable skyscraperas presented via an audio and video presentation on a computer

iosk at one end of the workstation (see Fig. 1b). Families selectedhether to hear the computer presentation in English or Span-

sh. As they built, a camera took digital pictures of the families.hen they finished building, families reviewed the pictures on the

omputer monitor, and selected six of the photos to use in telling photo-narrative about what they learned during their buildingxperience. The workstation utilized in the research (one of three inhe exhibit) is equipped with built-in equipment to video and audioecord families during the building activity and photo-narratives.

.3. Procedure

.3.1. Experimental manipulationAll families were randomly assigned to groups. European-

merican families participated in one of four conditions: (1)nspector Sturdy Build + Talk Instructions, (2) Inspector Sturdyuild-Only Instructions, (3) Build + Talk Signs, or (4) No Instructionontrol. African-American and Hispanic-American families partic-

pated in either the Inspector Sturdy Build + Talk or Control groups,ielding a 2 (Group: Build + Talk, Control) × 3 (Ethnicity) experi-ental design.All families assigned to the Inspector Sturdy Build + Talk and

uild-Only conditions interacted with a research assistant who was

ressed in a white lab coat, had a triangle-shaped magnifying glass,nd played the role of a building inspector named Inspector Sturdyo carry out the instructional program. Before they began thekyscraper Challenge, Inspector Sturdy presented these families

2 3 0 1

with a model skyscraper built with the exhibit’s plastic buildingmaterials (see Fig. 2a), and asked the target children to test if thestructure was sturdy or wobbly. Inspector Sturdy directed thefamilies’ attention to the yellow triangle-shaped brace pieces, aswell as blue and speckled struts that had been positioned in themodel to make triangular braces, and explained that this structurewas sturdy because it had been built with triangles.

Next, Inspector Sturdy presented a second model building(Fig. 2b) and asked the target children again to test out if theskyscraper was sturdy or wobbly. This model was wobbly, andInspector Sturdy pointed out to the families that even though it hadtriangles (mostly on top, for decoration), they were not in placeswhere they could make the building sturdy. Inspector Sturdy sum-marized the building instructions by showing the families a signwith photos of the two model structures and read aloud the buildingtip written on the sign: ‘Triangles make strong buildings.’

To families in the Inspector Sturdy Build + Talk condition,Inspector Sturdy offered an additional tip, encouraging the adults toask What?, Why?, Where?, and How? type-questions. She explainedthat families she had observed building sturdy buildings askedeach other for information. Inspector Sturdy showed families in theBuild + Talk condition a second sign with examples What (“Whatcould we do to make this stronger?”), Why (“Why do you think thisis wobbling?”), Where (“Where could we put a triangle piece?”), andHow (“How can we make this stand up?) questions, and encouragedthe adults to ask their children for information with questions likethese as they were building. Inspector Sturdy left the tip sign(s)adjacent to the building area in plain view while the families built.

Families in the Signs condition met the researcher (not posing asInspector Sturdy) who conducted the consent procedures. The twosigns used to summarize the tips to families in the Inspector SturdyBuild + Talk condition were placed in the same location immedi-ately adjacent to the building area as they were for families in theInspector Sturdy conditions, but the researcher did not point outthe signs or provide any building or conversational instructions.The families in the Control condition provided informed consentbut otherwise engaged in the exhibit as visitors normally would,without building or conversation instructions, or Inspector Sturdy’ssigns.

1.3.2. Skyscraper ChallengeAll families participated in the Skyscraper Challenge building

activity. At the start of the building activity, the same goal wasprovide via a computer presentation to all families: “Everyone,

here’s your first job. Build a skyscraper together. Find a way to braceit so it won’t fall down. See how close you can get to the clouds.” Atimer appeared on the computer screen and families were told theyhad 12 min to finish their building. At the end, families could choose

C.A. Haden et al. / Early Childhood Research Quarterly 29 (2014) 333–344 337

build

tWtr(tg

1

ceFyu

Fig. 1. Skyscraper Challenge: (a)

o build for an additional 3 min, for a total of 15 min building time.hen the building activity was completed, Inspector Sturdy tested

he buildings of the families in the Inspector Sturdy conditions andegardless of the sturdiness, provided general, positive feedbacke.g., “Wow, you did a great job building!”), before they moved tohe photo-narrative task. Those in the other two conditions wereuided to begin the photo-narrative task by the researcher.

.3.3. Photo-narrativeAll families also participated in the photo-narrative task. The

omputer prompted the families to tell a narrative about their

xperience using the photos taken during their building activity.amilies were asked to select six pictures: one “picture that showsour team getting started”; a second, to be chosen by a grown-p; and four more pictures, the selection of which was to be “a

Fig. 2. Model structures built with the exhibit materials in the Skysc

ing area and (b) computer kiosk.

job for everyone.” Then, the families were asked to select a titlefor their story (Architects at Work; Teamwork Makes it Happen;From the Ground Up; or Chicago Construction). Then the computerprompted visitors to answer a question with each picture: (1) I’mcurious about this first picture. How did you figure out how to startbuilding? (2) What was each of you thinking as you built? (3) Whatproblems did your team have as you built? (4) Grown ups, this questionis for you: How did your team try to solve these problems? (5) Everyonetalk about your building. What did each of you learn from building?(6) Great job building together and telling your story. What do youthink you’ll remember from doing this?. The design of the Skyscraper

Challenge and photo-narrative task was such that the time engag-ing in the exhibit was approximately the same across families. Afterthe photo-narrative, one parent from each family completed a briefwritten survey of demographic information.

raper Challenge: (a) sturdy structure and (b) wobbly structure.

3 Resea

1

aPvitvgdrc2ac

1

ncamtttwatstwinbeture–p8

1

cstaao

1

aScTbaatSot

38 C.A. Haden et al. / Early Childhood

.4. Coding

Videotape records of the conversations during the buildingctivity (masked for condition) were scored using Noldus Observerro software. The photo-narratives (also masked) were coded fromerbatim transcripts of the audio records. Procedures for establish-ng reliability were the same for all coding systems. Rememberinghat there were sometimes multiple adults and children in theisitor group, we coded talk by any adult or child in the visitorroup according to the adult and child categories, respectively, asescribed below. Two researchers independently coded 20% of theecords. For each coding system, two trained researchers blind toondition (different coders for each system) independently coded0% of the records. Once reliability was established, no single reli-bility estimate was below 80% agreement (kappas > .75); one coderoded the remainder of the data with checks by the second coder.

.4.1. Talk during the building activityAdults’ open-ended Wh-questions were defined as requests for

ew (not previously provided) information (e.g., “What is thisalled?” “What are these used for?”). The children’s responses todults’ Wh-questions were coded if they provided on-topic infor-ation that had not been previously mentioned by members of

he group (e.g., an adult asks “What should we do next?” andhe child responds “Brace it!”). All Wh-questions, yes-no ques-ions, and statements made by adults during the building activityere summed to index the frequency of adults’ STEM talk, using

system developed as a part of this project. Specifically, STEMalk included: (1) Scientific Method – mentions of plans, steps,equences, experimentation, hypothesis testing, figuring some-hing out, trying something out, or redoing based on something notorking (e.g., “First we need a plan, then we can start testing our

deas”); (2) Technology – mentions of building materials, includinguts, bolts, struts, etc. (e.g., “How can we use the nuts to hold theolts in place?”); (3) Engineering-Triangles Make it Sturdy – refer-nce to triangles as making a structure sturdy (e.g., “We need to addriangles to make our building more sturdy.”; “The triangles helpeds big time in holding the building up.”); (4) Engineering-Other –eference to screwing, tightening, having trouble fitting the bolts,tc. (e.g., “Remember, righty tighty, lefty loosey”); and (5) Math

mentions of number, amount, measurement (e.g., “We need 4ieces to start.”). Inter-rater agreement for these codes averaged5% (kappas ≥ .75).

.4.2. Building sturdinessTwo coders viewed the video records and independently

ounted the (a) total number of building materials included in thetructure and the (b) total number of braced pieces, defined as theotal number of triangles and struts that were connected to form

triangle or X-shaped brace. Agreement in scoring the buildingsveraged 85.4% (kappas ≥ .77). A ratio of the total number of bracesver the total number of pieces was computed.

.4.3. Photo-narrativeThe photo-narrative coding system involved scoring of the

dults’ and children’s contributions to the story that involvedTEM content. Preliminary analyses revealed in this task that STEMontent was relatively infrequent, particularly for the children.herefore, the coding system for the adults’ STEM talk duringuilding (measured in terms of frequency) was modified to char-cterize the presence/absence of talk during the photo-narrativectivity in each of the five categories defined above: (1) Scien-

ific Method, (2) Technology, (3) Engineering-Triangles Make itturdy, (4) Engineering-Other, and (5) Math. Each speaker receivedne point for mentioning information that fit a category, withhe maximum score being 5 if information pertaining to all five

rch Quarterly 29 (2014) 333–344

categories was provided. For example, a child who during thephoto-narrative said “We found out triangles make it strong.”(Engineering-Triangles Make it Sturdy), “Sometimes we couldn’tget the nuts turning right.” (Engineering-Other), and “We neededto measure it.” (Math), received three points. Multiple commentsfitting one category received only one point for that category (e.g.,a parent who said “We kept trying different pieces.” and “We triedlots and lots of different things to make it work.” received one pointfor the presence of talk about the Scientific Method). Agreement inscoring the photonarratives averaged 87% (kappas ≥ .79).

2. Results

All hypotheses were tested using analyses of variance (ANOVAs).For each dependent measure, one, one-way ANOVA tested differ-ences among the Build + Talk, Build-only, Signs, and Control groupsfor the European-American sample only (n = 71). A second 2 (Group:Build + Talk, Control) × 3 (Ethnicity: European-American, African-American, Hispanic-American) ANOVA tested if the effects weredifferent for families of different ethnic backgrounds (n = 97). In allmain analyses, all significant effects identified in the four-groupANOVAs were replicated by significant group main effects in the2 × 3 ANOVAs. Therefore, for brevity, only main effects of Ethnic-ity and Group × Ethnicity differences are reported for the 2 × 3ANOVAs.

Each analysis was run first with child age as a covariate, thenwith the number of adults in the visitor group as a covariate,and third with the number of children in the visitor group as acovariate. Covariance analyses are reported only when a covari-ate significantly predicted the dependent measure. Main effectswere followed by pairwise tests with a Bonferroni adjustment formultiple comparisons (all ps < .05, unless otherwise noted).

2.1. Preliminary analyses

ANOVAs confirmed that random assignment resulted in equalgroups with respect to child age, the number of adults per visitorgroup, and the number of children per visitor group (see meansin Table 1; all Fs < 1.98, ps > .16). For family income and parenteducation data (see Section 1.1), Chi-square analyses revealed nofamily income or parent education differences among the fourexperimental groups, �2(6) = 3.25, and 5.17, respectively, ps ≥ .52.Hispanic-Americans were significantly more likely to report highschool as the highest level of education completed (standard resid-ual = 2.4), �2(2) = 9.21, p = .01, Cramér’s V = .27, but having obtaineda college degree or higher was not related to ethnicity. FurtherChi-square analyses established a significant association betweenethnicity and income such that that European-Americans were sig-nificantly less likely to report family incomes of <$50,000/year(standard residual = 2.6), and significantly more likely to reportincomes of >$100,000/year (standard residual = 2.2), �2(4) = 19.25,p < .001, Cramér’s V = .32. At the middle reported income level($50,000–100,000; n = 25), there were no differences among theethnic groups.

As also shown in Table 1 (see also Section 1.1), the majorityof visitor groups included female adults only, or female and maleadults; very few groups included male adults only. Across the entiresample, controlling for number of adults in the visitor group, visitorgroups with female and male adults (M = 9.22, SD = 4.95) providedmore STEM information during the building activity than thosewith female adults only (M = 6.54, SD = 4.64), F(1, 115) = 9.14, p < .01,

�2 = .07. No other main or interactive effects of gender of adultswere found, and there were no child gender differences on any ofthe dependent measures, Fs < .73, ps > .39; therefore, gender is notconsidered further.

Resea

2

aiSr

2

tpqwdFmASltFqwc

i0WrtbrprAiBotTtFdgo

SittiapBiwnaA(op

iW

C.A. Haden et al. / Early Childhood

.2. Building activities

Presentation of the main analyses begins with the buildingctivities. First, we address whether talk during the building activ-ties varied as a function of experimental condition and ethnicity.econd, we consider how group membership and ethnicity wereelated to the sturdiness of the structures that families built.

.2.1. Talk during building activitiesThe top portion of Table 2 displays the means for parents’

otal Wh-questions during the building activity. As hypothesized,arents in the Build + Talk condition asked significantly more Wh-uestions than parents in the Build-only, Signs, and Control groups,ho were no different from one another, F(3, 67) = 7.64, p < .001,

= 1.18. As indicated by the significant main effect of ethnicity,(2, 91) = 17.58, p < .001, d = 1.25, European-American adults askedore Wh-questions (M = 5.60, SD = 4.52) than either the African-mericans (M = 1.27, SD = 1.57) or Hispanic-Americans (M = 1.73,D = 2.03), who were not different from each other. Neverthe-ess, the effects of the instructions were equivalent across groups;he Group × Ethnicity interaction was not statistically significant,(2, 91) = 1.22, p = .30. For each ethnic group, the number of Wh-uestions asked by adults in the Inspector Sturdy Built + Talk groupas more than two times the number asked by those in the control

ondition.Children’s responding to adults’ Wh-questions during the build-

ng activity was low across conditions and groups (M = 1.24, range–8). In fact, 44% of children did not provide any responses to adults’h-questions. We nonetheless examined both the frequency and

ate of children’s responding. Rate of responding was calculated ashe number of child responses to parents’ Wh-questions dividedy the total number of parents’ Wh-questions; therefore rate ofesponding controlled for variation in the number of Wh-questionsarents asked. European-American children were more frequentlyesponding to adults’ Wh-questions (M = 1.94) than were African-merican (M = .75) and Hispanic-American children (M = .95), as

ndicated by a main effect of ethnicity, F(2, 91) = 6.69, p < .01, d = .74.ut, for children’s frequency of responding there was no main effectf group, F(3, 67) = 1.94, p = .13, and no Group × Ethnicity interac-ion, F(2, 91) = .79, p = .45. Also, as shown in the middle portion ofable 2, there were no differences in the rate of children’s responseso adults’ Wh-questions by group or ethnicity, and no interactions,s < .96, ps ≥ .42. Thus, although children in the Build + Talk con-ition were being asked the most questions, and thus had thereatest opportunity to respond, the children’s frequency and ratef responding was not different across conditions.

The top portion of Table 2 also displays the means for adults’TEM talk during the building activity. We hypothesized that adultsn the Build + Talk group would talk the most about STEM duringhe building activity. In these analyses, the number of adults inhe family group was significantly related to the provision of STEMnformation, F(1, 66) = 5.70, p < .05, d = .59, and with this covari-te, the main effect of group was also significant, F(3, 66) = 4.27,

< .01, d = .91. As shown in the table, adults in the Inspector Sturdyuild + Talk group demonstrated significantly more STEM talk dur-

ng the building activity than any of the other three groups, whoere not different from one another. There was a main effect of eth-icity, F(2, 90) = 6.30, p < .01, d = .77, such that European-Americandults (M = 17.67, SD = 9.68) talked more about STEM than African-mericans (M = 11.19, SD = 7.89; p = .06) and Hispanic-Americans

M = 9.44, SD = 5.87, p < . 05), who were not different from eachther. There was no Group × Ethnicity interaction, F(2, 90) = .02,

= .98.In summary, in comparison to adults who did not participate

n the Inspector Sturdy Build + Talk program, those who did askedh-questions with higher frequency, thus offering their children

rch Quarterly 29 (2014) 333–344 339

more opportunities to provide responses. European-Americanchildren most frequently responded to their parents’ questions,however, children’s rate of responding (controlling for the numberof questions they were asked) was very low and not different acrossconditions or ethnic groups. Although not specifically instructedto do so, adults in the Inspector Sturdy Build + Talk condition morefrequently included STEM content in their talk during buildingthan did parents in the Control group. Notably, the buildinginstructions on their own did not lead to significant increases inadults’ STEM talk while building. Moreover, the lack of interactioneffects across these analyses suggests that the Inspector SturdyBuild + Talk program was equally effective for European-American,African-American, and Hispanic-American families. Overall,African-American and Hispanic-American families engaged in lessof the types of talk targeted by the intervention.

2.2.2. BuildingsThe total number of pieces and the ratio of braces to total

pieces in the families’ resulting structures are shown in the bottomportion of Table 2. The Signs group included the greatest num-ber of pieces overall, as indicated by a significant main effect ofgroup for total number of pieces in the structures, F(3, 65) = 3.83,p = .01, d = .84. Also, European-American (M = 24.97, SD = 10.65) andHispanic-American (M = 22.49, SD = 7.11) families included moretotal pieces than African-American families (M = 17.24, SD = 5.41),as indicated by the main effect of ethnicity, F(2, 89) = 7.44, p < .001,d = .81. There was no Group × Ethnicity interaction for total numberof pieces included in the structures F(2, 89) = .87, p = .42.

We hypothesized that families engaged in the Inspector Sturdyprogram, who were armed with building instructions about theengineering principle bracing, would incorporate the most cross-braces in their buildings. Confirming this hypothesis, the ratioof braces-to-total-pieces was higher for the Inspector SturdyBuild + Talk (M = .21, SD = .07) and Build-only (M = .22, SD = .10)groups compared with the Signs (M = .11, SD = .12) and Control(M = .10, SD = .10) conditions, as indicated by a significant maineffect of group, F(3, 65) = 6.13, p = .001, d = 1.06. The ratio of bracesto total pieces was also higher for European-American (M = .16,SD = .10) and Hispanic-American (M = .13, SD = .11) families thanfor African-American families (M = .07, SD = .08), as indicated bya significant main effect of ethnicity, F(2, 89) = 11.29, p < .001,d = 1.00. The interaction of Group × Ethnicity was not significant,F(2, 89) = 1.10, p = .34. Therefore, the building instructions providedby Inspector Sturdy helped European-American, African-American,and Hispanic-American families equally to brace their buildings tomake them stronger.

2.2.3. Photo-narrativesRecall that the adults and the children’s contributions to the

photo-narratives were scored for mentions of the five categories ofSTEM talk. Table 3 displays the mean number of categories of STEMtalk mentioned by adults and children in the photo-narratives.As shown in Table 3, scores were generally low on this measure.Nevertheless, there were scores across the whole of 0–5 range forboth children and adults. We hypothesized that the families in theInspector Sturdy Build + Talk condition would mention more dif-ferent types of STEM-related content in the photo-narrative taskthan the other groups. Indeed, the main effect of group for adults’talk about STEM was significant, F(3, 65) = 4.81, p = .05, d = .70, andfollow-up tests revealed that the adults in the Inspector SturdyBuild + Talk group tended to mention more categories of STEM talkthan those in the Control group (p = .08). In addition, as also illus-

trated in Table 3, children in the Inspector Sturdy Build + Talk andBuild-only groups mentioned more different categories of STEMthan the Control group, as indicated by the significant main effectof group, F(3, 65) = 5.77, p = .001, d = 1.03. No significant main effects

340 C.A. Haden et al. / Early Childhood Research Quarterly 29 (2014) 333–344

Table 2Conversation during the building activities and sturdiness of the resulting structures.

Build + Talk Build-only Signs Control

European-American

African-American

Hispanic-American

European-American

African-American

Hispanic-American

Adults’ verbal behaviorsWh-questions 6.91 (5.18) 1.87 (1.68) 3.00 (2.14) 2.56 (2.42) 2.17 (2.43) 3.60 (2.16) .78 (1.30) .80 (1.37)STEM talk 20.41 (9.71) 15.28 (7.66) 13.20 (4.47) 12.00 (6.68) 13.00 (7.65) 13.36 (8.19) 8.00 (6.63) 6.93 (5.43)

Children’s response frequency 2.22 (2.02) 1.00 (1.07) 1.64 (1.69) .94 (1.34) 1.35 (1.66) 1.67 (1.54) .50 (.98) .27 (.59)Children’s response ratio .41 (.28) .47 (.33) .45 (.21) .32 (.35) .55 (.39) .45 (.42) .47 (.51) .37 (.50)Buildings

Total pieces 27.81 (11.89) 18.40 (5.54) 26.09 (8.58) 26.63 (8.94) 32.29 (8.18) 21.00 (7.25) 16.28 (5.26) 19.80 (4.44)Brace ratio .21 (.69) .13 (.08) .22 (.10) .22 (.13) .11 (.12) .10 (.10) .01 (.04) .06 (.06)

Note: SDs are in parentheses.

Table 3Presence of STEM content in photo-narratives.

Content Build + Talk Build-only Signs Control

European-American

AfricanAmerican

HispanicAmerican

European-American

AfricanAmerican

HispanicAmerican

Adults’ STEM 1.95 (1.59) 1.60 (1.40) 1.00 (.67) 1.75 (1.43) 1.18 (1.18) .78 (.89) 1.17 (.78) .78 (1.30).12 (1

N

opp

iictotgad

3

itbnciStqcagmdhtottwirwt

Children’s STEM 1.45 (1.30) 1.07 (1.39) 1.20 (1.23) 2

ote: Maximum score equals 5. SDs are in parentheses.

f ethnicity, or Group × Ethnicity interactions were observed in thehoto-narratives for parent or child mentions of STEM, Fs ≤ 1.21,s ≥ .30.

In sum, the photo-narrative results suggest that the buildingnstructions with or without conversation instructions led childrenn families who received them to discuss a range of content thatould be characterized as STEM-related when asked to discuss whathey had learned during building. For these results, there were noverall differences among the ethnic groups. Moreover, once again,hese results did not vary by ethnicity, suggesting that across familyroups when prompted to discuss what they had learned, adultsnd children who received building instructions mentioned moreifferent aspects of STEM than those who received no instruction.

. Discussion

In this study, we used an experimental methodology to exam-ne family interactions in a STEM-related museum exhibit. Takenogether, our findings indicate, first, that with a fairly modest andrief facilitated educational program, it was possible to double theumber of Wh-questions adults asked children during a buildingonstruction activity in the exhibit. Compared with adults in fam-lies who did not receive conversation instructions from Inspectorturdy, those who did asked more Who, What, Where, Why and Howype-questions during the building activity. Second, children’s fre-uency and rate of responding was not different across groups;lose to half of the children in the sample did not respond tony adult Wh-questions. Third, in comparison to all three otherroups, adults in the Inspector Sturdy Build + Talk group producedore STEM-related talk during building. The building tip alone

id not appear to lead to increases in adults’ STEM talk. Fourth,owever, the building tip did facilitate families building activi-ies. Families who received the building tip had the highest ratiof braces-to-total-pieces in their completed buildings, showingheir use of the key engineering principle emphasized by Inspec-or Sturdy. Fifth, relative to those in the Control group, and with orithout the conversation tip, children who received the building

nstructions from Inspector Sturdy mentioned more types of STEM-elated content when in the photo-narrative task they reported onhat they had learned. Finally, we considered the effectiveness of

he Inspector Sturdy Build + Talk program for European-American,

.36) 1.24 (1.20) .36 (.50) .50 (.78) .78 (.97)

African-American, and Hispanic-American families. Ethnic groupdifferences were found for the frequency of adults’ Wh-questions,children’s frequency of responding to questions, adults’ STEM talkwhile building, and the extent to which families braced theirbuildings. Nevertheless, our results also show that the educationalprogram was equally effective across our diverse sample on con-versational and learning measures. We turn now to further discussthese findings and their implications for children’s informal STEMlearning.

3.1. Conversations about STEM

This study was designed to enhance family interactions in aSTEM-related children’s museum exhibit. This work connects withresearch on children’s memory of personally experienced eventsthat emphasizes the important role of parent–child conversationsduring and after events for children’s understanding and remem-bering of their experiences (Fivush et al., 2006; Haden, 2013;Ornstein et al., 2004). This study also links with research focusedon family conversations in museums that recommends particulartypes of talk that may provide mechanisms by which talk sup-ports science learning (Callanan & Jipson, 2001; Crowley, Callanan,Jipson, et al., 2001; Crowley, Callanan, Tenenbaum, et al., 2001;Falk & Dierking, 2002; Haden, 2010; Leinhardt et al., 2002; NRC,2009). Moreover, the study is based in sociocultural theory (Berk,2001; Gauvain, 2000; Rogoff, 1990; Vygotsky, 1978) that empha-sizes collaborative conversational exchanges between children andadults as a source of potentially powerful mediators of learning anddevelopment. Research in the sociocultural tradition has demon-strated important differences in conversational and teaching stylesamong families from different ethnic, socioeconomic, and school-ing backgrounds (Fivush & Haden, 2003; Heath, 1982, 1983; Milleret al., 2012; NRC, 2009; Rogoff, 2003; Tamis-LeMonda et al., 2013;Tenenbaum & Callanan, 2008; Tenenbaum et al., 2002). In thisstudy, we were interested in how family background might ormight not interact with our efforts to intervene to increase specifictypes of talk associated with STEM learning.

Concerning the forms of conversation we hoped to encouragein this study, we chose to focus on adults’ Wh-questions becausethey have been identified as a key component of an elaborativeconversational style that can facilitate understanding (Fivush et al.,

Resea

2rrf2mwtc2htSwqs(cicWeR

dttttnsgytia

t(Ntlfahawece

ldSrAmpiagJphie

C.A. Haden et al. / Early Childhood

006; Ornstein et al., 2004). Importantly, parents’ use of an elabo-ative conversational style has been tied to children’s learning andemembering in the U.S. and cross-culturally across a range of dif-erent kinds of personally experienced events (see Fivush & Haden,003; Fivush et al., 2006, for reviews). Observations of families inuseums in particular have shown that whereas some caregiversill, in a quiz-like fashion, ask narrowly focused questions that

hey know the answer to, others ask open-ended Wh-questions thatould receive a range of responses (Diamond, 1986; Falk & Dierking,002; Leinhardt & Knutson, 2004). Moreover, work in museumsas suggested that caregivers’ open-ended questions can facili-ate sustained engagement in exhibits (Humphrey & Gutwill, 2005;chauble, 1996). The results of this study add to a growing body ofork indicating that caregivers who ask many open-ended Wh-

uestions promote children’s understanding and learning aboutcience more than parents who ask few of these sorts of questionsBenjamin et al., 2010; Jant et al., 2014). The current study alsoontributes to a small but growing literature demonstrating thatt is possible – even in this case with minimal prompts – to suc-essfully encourage adult caregivers to increase their elaborativeh-question-asking of children during and after events (Benjamin

t al., 2010; Boland et al., 2003; Peterson, Jesso, & McCabe, 1999;eese & Newcombe, 2007).

Although increasing adults’ Wh-questions is important, chil-ren’s verbal responding to these questions may also be critical toheir representing experiences and information in ways that enablehem to be recalled and used in different or new contexts overime (Jant et al., 2014). One reason Wh-questions may be key ishat they are more cognitively challenging than closed-ended, yes-o type questions. Asking a child “What should we do to make thistronger?” or “How should we attach this piece?” invites the child toenerate descriptions, explanations, inferences, and so on in wayses-no questions do not. Nevertheless, if children do not respondo these Wh-questions, their impact may not be fully realized. Thessue thus becomes how we can create conditions in which parentssk questions that in turn are responded to by their children.

Several studies have identified natural variability in the extento which families engage in this form of joint verbal interactionHaden et al., 2001; Hedrick, San Souci, et al., 2009; Tessler &elson, 1994). Moreover, results from these studies demonstrate

hat parent–child joint talk is highly positively related to children’searning and remembering. For example, Tessler and Nelson (1994)ound that young preschoolers who were observed as they visited

museum only later recalled objects that the child and motherad talked about together during the experience. Jant et al. (2014)nd Benjamin et al. (2010) also recently reported that parents whoere encouraged to ask Wh-questions in natural history and sci-

nce exhibits, respectively, engaged in more joint talk with theirhildren, and had children who recalled more about their museumxperiences days and weeks later.

Against this backdrop of previous work linking joint talk toearning and subsequent remembering, that responding by chil-ren in the current study was low is of concern. The Inspectorturdy program did not lead to increases in the rate of children’sesponding to caregivers’ questions during the building activity.lthough we do not know precisely why this was the case, weight have better affected children’s responding by explicitly

rompting children to try to respond to their parents’ questions. Ast was, the children seemed more interested in building “now” andnswering questions later, and this may have been fueled by theoal to build a tall building with a time limit for completion. Also,ant et al. (2014) found that children who had the opportunity to

lay with key exhibit objects prior to visiting an exhibit in a naturalistory museum engaged in more joint talk with their parents

n the exhibit, compared with children who did not engage withxhibit objects beforehand. The building materials would have

rch Quarterly 29 (2014) 333–344 341

been relatively unfamiliar, and perhaps if the children had theopportunity to play with them before building, they might haveengaged more in conversations while building. Generally, what isneeded is what Sigel (1993) called distancing and what Goldstoneand Sakamoto (2003) called concreteness fading – conditions thatencourage children to focus less on the specific objects and moreon the abstract knowledge or concepts that can facilitate learning.Joint talk may provide a critical mechanism for distancing, learningand memory (Jant et al., 2014); therefore, it is important that futurework be directed at understanding and promoting conditions thatat once increase parents’ question-asking and children’s answeringduring informal learning experiences.

We also studied if the conversation and building tips might leadfamilies to talk more about the scientific method, technology, engi-neering, and math. Benjamin et al. (2010) found that receivinginformation about bracing structures and prompts to ask Wh-questions was linked to measures of engineering talk, but they didnot code for STEM talk overall. We discovered in the current studythat overall discussion of STEM while building was higher whenfamilies received both conversation and building instructions. Thisoffering of STEM information by parents may be especially criti-cal in the context of the novel building activity. As we observed,the children’s responding to adults’ Wh- questions was rather low.The adults’ provision of information about STEM during buildingmay therefore have been key to extending the discussion of sci-ence, technology, engineering, and math, even when the childrenwere not able to provide requested information (c.f., Reese, Haden,& Fivush, 1993). Extrapolating from recent work that has examinedif the production of specific types of talk (e.g., number terms, men-tal state words, spatial terms, etc.) predicts children’s later skills inthe related domain (Gunderson & Levine, 2011; Levine, Gunderson,et al., 2011; Levine, Suriyakham, et al., 2011; Pruden et al., 2011;Taumoepeau & Ruffman, 2008), increasing the amount of care-givers’ talk that features STEM-related content might improvechildren’s STEM understanding and learning.

3.2. Buildings and photo-narratives

One way we measured what families and children learned aboutbuilding from the intervention was to look at the sturdiness ofthe structures they built. As expected, in contrast to families whodid not receive building instructions from Inspector Sturdy, fami-lies who did built sturdier structures. Benjamin et al. (2010) alsoshowed that building instructions alone were sufficient to enhancefamilies building outcomes in a (different) building constructionexhibit. Basically, in the present study, as in Benjamin et al. (2010),families were provided with information about cross-bracing –albeit in a different demonstration format across studies – and theywere able to use this information in constructing their own struc-tures. However, it is interesting to note that families in the Signscondition in the current study, who saw pictures of these modelstructures in the absence of verbal instructions, were not able touse the signage in the same way. Benjamin et al. found that view-ing physical models of structures with cross-bracing throughoutalso did not prompt families to brace their own buildings. Effortsto draw attention to the signs, or changing them to involve chal-lenge questions (e.g., What if?) might have been more effective (seeGutwill, 2006; Hohenstein & Tran, 2007).

The photo-narratives provided a unique window onto fami-lies’ reflections on their exhibit interactions and learning. Thesenarratives evidence what families understood about building andengineering concepts. Moreover, the opportunity to review and

reflect on their experiences immediately following the buildingactivity provides the chance for families to further elaborate andeven modify their initial understandings. Indeed, constructing nar-ratives about personally experienced events can be a powerful

3 Resea

mwHhh

hemuacotoanfravwefdaWl

3

atmsp

EdHmwdoofphntie

eseTbqiotctk

42 C.A. Haden et al. / Early Childhood

eans for adding layers of comprehensibility to events beyondhat is available from direct experience (Bruner, 1991; Fivush &aden, 1997; Nelson & Fivush, 2004). Reflective narratives mayelp families to make their experiences more meaningful, and per-aps more memorable (Fivush et al., 2006).

The photonarrative results show that compared with those whoad not received any building instruction, children – and to a lesserxtent, adults – in the building instruction conditions includedore different categories of STEM talk in their reports. Somewhat

nexpectedly, there was no difference between children who hadnd had not experienced the building tip in combination with theonversation tip. The rate of occurrence of STEM talk types was lowverall. Nevertheless, children who received the building instruc-ions from Inspector Sturdy seemed to form a better understandingf the STEM-related aspects of the building activity that they couldccess and use to tell a story about their experiences. The photo-arrative was a highly structured task, and may not have provided a

ull reading of what the children knew and learned from their expe-ience. Greater differences in STEM talk might have be found if ourssessments had included natural, parent–child reminiscing con-ersations days and weeks following the museum visit, a result thatould be consistent with prior work (Benjamin et al., 2010; Boland

t al., 2003; Ellenbogen, 2002; Jant et al., 2014). Additionally, in theuture, it would also be interesting to know if the opportunity foriscussion of the experience immediately after building provides

positive post-encoding experience (Bauer, King, Larkina, Varga, &hite, 2012) that might strongly predict later retention and use of

earned information in different contexts over time.

.3. Patterns of ethnic differences

In all of the results reported here, when there was a differencecross experimental conditions, that difference was found consis-ently across the three ethnic groups. However, the fact that in some

easures there are ethnic differences while in others there are notuggests that there is more to be learned by comparing the differingatterns seen in STEM behavior across the three ethnicities.

In terms of amount of talk during the building activity,uropean-American adults asked more Wh-questions and pro-uced more STEM talk overall than either African-American orispanic-American adults. European-American children providedore frequent responses, but when the number of questions theyere asked was controlled, there were no differences in chil-ren’s rate of question-answering by ethnic group. In measuresf building behavior, Hispanic-Americans used a similar numberf pieces as European-Americans, while African-Americans usedewer. Hispanic-Americans also had a similar ratio of braces-to-ieces as the European-Americans, while, again, African-Americansad a lower ratio. Finally, during the photo-narratives, there wereo ethnic group differences in the number of STEM categories men-ioned by adults or children. Thus, despite the differences in theirnteraction and their building, at the end of the activity, the differentthnic groups were similar in their talk about STEM.

Other verbal and physical behaviors may indicate other differ-nces in interactions patterns by family background than thosetudied here, such as parents’ teaching behaviors (Tamis-LeMondat al., 2013), and explanations (NRC, 2009; Siegel et al., 2007;enenbaum & Callanan, 2008). Moreover, to tease apart family-ackground-specific characteristics, one would need to do a moreualitative and exploratory analysis of the families’ behaviors dur-

ng building and the narrative component that is beyond the scopef this report. But the interesting patterns reported here suggest

hat such an analysis would be worth doing. This analysis mightontribute to our understanding of how different families may usehe resources of a museum exhibit in ways that reflect funds ofnowledge, and cultural understandings about how children learn

rch Quarterly 29 (2014) 333–344

as well as their everyday practices of how to interact together dur-ing a shared activity.

3.4. Conclusions and implications

Overall, the results provide evidence that adult caregivers canand do engage children in informal STEM-related learning activi-ties in ways that can support STEM understanding. Moreover, theprogram we designed was effective across families from diversebackgrounds in fostering interactions that could enhance chil-dren’s STEM learning. The sample sizes for the African-Americanand Hispanic-American samples especially limited our ability tosystematically study schooling, socioeconomic status, prior expe-rience visiting museums, and a host of other background factorsthat may combine to influence how parents talk about STEM withtheir children. Also, our observations were limited to the museum,and cannot speak to the ways these families may have interactedaround science activities in a variety of contexts, including thosethat may be more familiar (e.g., at home). Nevertheless, our resultssuggest that providing adult caregivers with information in muse-ums that can support their efforts to help children learn STEM maybe important. Increasingly, there is discussion in the informal learn-ing literature (NRC, 2009) that caregivers may not always see theirrole in the learning process. In some cases, they may view schools asthe primary source of children’s STEM learning. In others, caregiversmay lack understanding themselves – in this case of engineeringprinciples and general information related to STEM – that couldsupport authentic problem solving with their children. Our resultssuggest, however, that it is possible to scaffold adult scaffolders ofchildren’s learning to promote children’s understanding and inter-est in STEM.

Finally, despite growing interest, there is a dearth of researchthat would inform the work of educators and other on-the-groundprofessionals in museums in their efforts to mediate STEM learningin museums (Pattison & Dierking, 2012). The research that exists(Mony & Heimlich, 2008; Rosenthal & Blankman-Hetrick, 2002),together with the results of this study, suggest that empiricallybased, staff-facilitated interactive programs can encourage conver-sation and collaboration to enhance children’s STEM learning inmuseums. Thus, together with efforts to better understand howto encourage families’ use of the physical contexts and signage inexhibits, it is important to study characteristics of the social contextthat can make for successful staff-facilitated interactive programs.Indeed, one of the most robust findings in the informal learningliterature is the important role that families play in children’s learn-ing. This study contributes to understanding the mechanisms bywhich this is so. The work is also important in suggesting ways thatmuseums can leverage what families bring to the museum to in asuccessful and seamless manner facilitate children’s STEM learning.

Acknowledgements

This research was supported in part by the National ScienceFoundation (NSF) under grants #0452550 and #1123411. J. Geddes’work on this project was also supported by a fellowship awardedby the Loyola Undergraduate Research Opportunities Program. Wewould like to thank Jennifer Farrington, Tsivia Cohen, and Rick Gar-mon of the Chicago Children’s Museum; student research assistantsValerie Flores, Sandra Vanegas, Erin Brown, and Karina Lima; andthe families who participated in the study.

References

Adams, S., Kuebli, J., Boyle, P. A., & Fivush, R. (1995). Gender differences inparent–child conversations about past emotions: A longitudinal investigation.Sex Roles, 33(5–6), 309–323.

Resea

A

B

B

B

B

B

B

BC

C

C

C

C

C

C

C

C

D

D

D

E

F

F

F

F

F

G

G

G

G

G

G

G

C.A. Haden et al. / Early Childhood

llen, S. (2004). Designs for learning: Studying science museum exhibits that domore than entertain. Science Education, 88(1), S17–S33.

auer, P. J., King, J. E., Larkina, M., Varga, N. L., & White, E. A. (2012). Characters andclues: Factors affecting children’s extension of knowledge through integrationof separate episodes. Journal of Experimental Child Psychology, 111(4), 681–694.

erk, L. E. (2001). Awakening children’s minds: How parents and teachers can make adifference. New York, NY: Oxford University Press.

enjamin, N., Haden, C. A., & Wilkerson, E. (2010). Enhancing building, conversa-tion, and learning through caregiver–child interactions in a children’s museum.Developmental Psychology, 46(2), 502–515.

oland, A. M., Haden, C. A., & Ornstein, P. A. (2003). Boosting children’s memory bytraining mothers in the use of an elaborative conversational style as an eventunfolds. Journal of Cognition and Development, 4(1), 39–65.

orun, M. (2002). Object-based learning and family groups. In S. G. Paris (Ed.), Per-spectives on object-centered learning in museums (pp. 245–260). Mahwah, NJ:Erlbaum.

orun, M., Chambers, M. B., Dristas, J., & Johnson, J. I. (1997). Enhancing familylearning through exhibits. Curator, 40(4), 279–295.

runer, J. S. (1991). The narrative construction of reality. Critical Inquiry, 18(1), 1–21.allanan, M. A. (2012). Conducting cognitive developmental research in museums:

Theoretical issues and practical considerations. Journal of Cognition and Devel-opment, 13(2), 137–151.

allanan, M. A., & Jipson, J. (2001). Explanatory conversations and young children’sdeveloping scientific literacy. In K. Crowley, C. D. Schunn, & T. Okada (Eds.),Designing for science: Implications from everyday, classroom, and professional sci-ence (pp. 21–49). Mahwah, NJ: Erlbaum.

allanan, M. A., & Oakes, L. M. (1992). Preschoolers’ questions and parents’ expla-nations: Causal thinking in everyday activity. Cognitive Development, 7(2),213–233.

arlson, L. E., & Sullivan, J. F. (2004). Exploiting design to inspire interest in engineer-ing across the K-16 engineering curriculum. International Journal of EngineeringEducation, 20(3), 372–378.

rowley, K., & Callanan, M. A. (1998). Identifying and supporting collaborative sci-entific thinking in parent–child interactions. Journal of Museum Education, 23(1),12–17.

rowley, K., Callanan, M. A., Jipson, J. L., Galco, J., Topping, K., & Shrager, J. (2001).Shared scientific thinking in everyday parent–child activity. Science Education,85(6), 712–732.

rowley, K., Callanan, M. A., Tenenbaum, H. R., & Allen, E. (2001). Parents explainmore often to boys than to girls during shared scientific thinking. PsychologicalScience, 12(3), 258–261.

rowley, K., & Jacobs, M. (2002). Islands of expertise and the development of fam-ily scientific literacy. In G. Leinhardt, K. Crowley, & K. Knutson (Eds.), Learningconversations in museums (pp. 333–356). Mahwah, NJ: Erlbaum.

unningham, C. M. (2009). Engineering is elementary. Bridge, 30(3), 11–17. Retrievedfrom http://www.mos.org/eie/pdf/newsitems/bridge fall2009.pdf

avis, R. S., Ginns, I. S., & McRobbie, C. J. (2002). Elementary school students’understanding of technology concepts. Journal of Technology Education, 14(1.).Retrieved from http://scholar.lib.vt.edu/ejournals/JTE/v14n1/davis.html

iamond, J. (1986). The behavior of family groups in science museums. Curator: TheMuseum Journal, 29(2), 139–154.

uncan, G. J., Dowsett, C. J., Claessens, A., Magnuson, K., Huston, A. C., Klebanov, P.,et al. (2007). School readiness and later achievement. Developmental Psychology,43(6), 1428. http://dx.doi.org/10.1037/0012-1649.43.6.1428

llenbogen, K. M. (2002). Museums in family life: An ethnographic case study. In G.Leinhardt, K. Crowley, & K. Knutson (Eds.), Learning conversations: Explanationand identity in museums (pp. 81–101). Mahwah, NJ: Erlbaum.

alk, J. H., & Dierking, L. D. (1992). The museum experience. Ann Arbor, MI: EdwardsBrothers.

alk, J., & Dierking, L. (2002). Lessons without limit: How free-choice learning is trans-forming education. Walnut Creek, CA: AltaMira Press.

ivush, R., & Haden, C. A. (1997). Narrating and representing experience:Preschooler’s developing autobiographical recounts. In P. van den Broek, P. J.Bauer, & T. Bourg (Eds.), Developmental spans in event comprehension and rep-resentation: Bridging fictional and actual events (pp. 169–198). Mahwah, NJ:Erlbaum.

ivush, R., & Haden, C. A. (Eds.). (2003). Autobiographical memory and the constructionof a narrative self: Developmental and cultural perspectives. Mahwah, NJ: Erlbaum.

ivush, R., Haden, C. A., & Reese, E. (2006). Elaborating on elaborations: The role ofmaternal reminiscing style in cognitive and socioemotional development. ChildDevelopment, 77(6), 1568–1588.

askins, S. (2008a). Designing exhibitions to support families’ cultural understand-ings. Exhibitionist, 27(1), 11–19.

askins, S. (2008b). The cultural meaning of play and learning in children’s muse-ums. Hand to Hand, 22(4), 1–2.

auvain, M. (2000). The social context of cognitive development. New York, NY: Guil-ford Press.

elman, R., & Brenneman, K. (2004). Science learning pathways for young children.Early Childhood Research Quarterly, 19(1), 150–158.

leason, M. E., & Schauble, L. (1999). Parents’ assistance of their children’s scientificreasoning. Cognition and Instruction, 17(4), 343–378.

oldstone, R. L., & Sakamoto, Y. (2003). The transfer of abstract principlesgoverning complex adaptive systems. Cognitive Psychology, 46, 414–466.http://dx.doi.org/10.1016/S0010-0285(02)00519-4

onzález, N., Moll, L. C., & Amanti, C. (Eds.). (2013). Funds of knowledge: Theorizingpractices in households, communities, and classrooms. New York, NY: Routledge.

rch Quarterly 29 (2014) 333–344 343

Goody, E. (1978). Questions and politeness: Strategies in social interaction. New York,NY: Cambridge University Press.

Gunderson, E. A., & Levine, S. C. (2011). Some types of parent number talk countsmore than others: Relations between parents’ input and children’s cardinal-number knowledge. Developmental Science, 14(5), 1021–1032.

Gutwill, J. (2006). Labels for open-ended exhibits: Using questions and suggestionsto motivate physical activity. Visitor Studies Today, 9(1), 1–9.

Haden, C. A. (2010). Talking about science in museums. Child Development Perspec-tives, 4(1), 62–67.

Haden, C. A. (2013). Interactions of knowledge and memory in the development ofskilled remembering. In P. Bauer, & R. Fivush (Eds.), Wiley-Blackwell handbookon children’s memory (pp. 809–835). New York, NY: Wiley-Blackwell.

Haden, C. A., Ornstein, P. A., Eckerman, C. O., & Didow, S. M. (2001). Mother–child con-versational interactions as events unfold: Linkages to subsequent remembering.Child Development, 72(4), 1016–1031.

Hart, B., & Risley, T. R. (1992). American parenting of language-learning children:Persisting differences in family–child interactions observed in natural homeenvironments. Developmental Psychology, 28(6), 1096–1105.

Heath, S. B. (1982). What no bedtime story means: Narrative skills at home andschool. Language in Society, 11(2), 49–76.

Heath, S. B. (1983). Ways with words: Language, life, and work in communities andclassrooms. New York, NY: Cambridge University Press.

Hedrick, A. M., Haden, C. A., & Ornstein, P. A. (2009). Elaborative talk during andafter an event: Conversational style influences children’s remembering. Journalof Cognition and Development, 10(3), 188–209.

Hedrick, A. M., San Souci, P., Haden, C. A., & Ornstein, P. A. (2009). Mother–child jointconversational exchanges during events: Linkages to children’s event memoryover time. Journal of Cognition and Development, 10(3), 143–161.

Hoff, E. (2003). The specificity of environmental influence: Socioeconomic statusaffects early vocabulary development via maternal speech. Child Development,74(5), 1368–2137.

Hoff, E. (2006). How social contexts support and shape language development. Devel-opmental Review, 26(1), 55–88.

Hohenstein, J., & Tran, L. U. (2007). Use of questions in exhibit labels to generateexplanatory conversation among science museum visitors. International Journalof Science Education, 29(12), 1557–1580.

Humphrey, T., & Gutwill, J. P. (Eds.). (2005). Fostering active prolonged engagement:The art of creating APE exhibits. San Francisco, CA: The Exploratorium.

Huttenlocher, J., Vasilyeva, M., Waterfall, H., Vevea, J., & Hedges, L. (2007). Thevarieties of speech to young children. Developmental Psychology, 43(5), 1062–1083.

Jant, E. A., Haden, C. A., Uttal, D. H., & Babcock, E. (2014). Conversation and objectmanipulation influence children’s learning in a museum. Child Development (inpress).

Laible, D. J. (2004). Mother–child discourse surrounding a child’s past behavior at 30months: Links to emotional understanding and early conscience developmentat 36 months. Merrill-Palmer Quarterly, 50(2), 159–180.

Leinhardt, G., Crowley, K., & Knutson, K. (2002). Learning conversations in museums.Mahwah, NJ: Erlbaum.

Leinhardt, G., & Knutson, K. (2004). Listening in on museum conversations. WalnutCreek, CA: AltaMira Press.

Levine, S. C., Gunderson, E. A., & Huttenlocher, J. (2011). Number development incontext: Variations in home and school input during the preschool years. In N.L. Stein, & S. W. Raudenbush (Eds.), Developmental cognitive science goes to school(pp. 189–202). New York, NY: Taylor and Francis.

Levine, S. C., Suriyakham, L. W., Huttenlocher, J., Rowe, M. L., & Gunderson, E. A.(2011). What counts in the development of young children’s number knowl-edge? Developmental Psychology, 46(5), 1309–1319.

Loewenstein, J., & Gentner, D. (2005). Relational language and the development ofrelational mapping. Cognitive Psychology, 50(4), 315–353.

Lohmann, H., & Tomasello, M. (2003). The role of language in the develop-ment of false belief understanding: A training study. Child Development, 74(4),1130–1144.

McGuigan, F., & Salmon, K. (2004). The time to talk: The influence of the timingof adult–child talk on children’s event memory. Child Development, 75(3), 669–686.

McGuigan, F., & Salmon, K. (2006). The influence of talking on showing and telling:Adult–child discussion and children’s verbal and nonverbal recall. Applied Cog-nitive Psychology, 20(3), 365–381. http://dx.doi.org/10.1002/acp.1183

Miller, P. J. (1982). Amy, Wendy, and Beth: Learning language in South Baltimore.Austin, TX: University of Texas Press.

Miller, P. J. (1994). Narrative practices: Their role in socialization and self-construction. In U. Neisser, & R. Fivush (Eds.), The remembering self: Constructionand accuracy in the self-narrative (pp. 158–190). New York, NY: Cambridge Uni-versity Press.

Miller, P. J., Fung, H., Chen, E. C., & Boldt, B. R. (2012). How socialization happens onthe ground: Narrative practices as alternate socializing pathways in Taiwaneseand European-American families. Monographs of the Society for Research in ChildDevelopment, 77 (1, Serial No. 302).

Mony, P. R. S., & Heimlich, J. E. (2008). Talking to visitors about conservation: Explor-ing message communication through docent-visitor interactions at zoos. Visitor

Studies, 11(2), 151–162.

National Research Council. (2007). Taking science to school: Learning and teachingscience in grades K-8. In R. A. Duschl, H. A. Scheingruber, & A. W. Shouse (Eds.),Committee on science learning Kindergarten through eighth grade. Washington,DC: The National Academies Press.

3 Resea

N

N

O

P

P

P

P

R

R

R

R

R

R

R

R

S

S

44 C.A. Haden et al. / Early Childhood

ational Research Council. (2009). Learning science in informal environments: Peo-ple, places, and pursuits. In P. Bell, B. Lewnstein, A. W. Shouse, & M. A. Feder(Eds.), Committee on learning science in informal environments. Washington, DC:The National Academies Press.

elson, K., & Fivush, R. (2004). The emergence of autobiographical memory: A socialcultural developmental theory. Psychological Review, 111(2), 486–511.

rnstein, P. A., Haden, C. A., & Hedrick, A. M. (2004). Learning to remember: Social-communicative exchanges and the development of children’s memory skills.Developmental Review, 24(4), 374–395.

almquist, S. D., & Crowley, K. (2007). From teachers to testers: Parents’ role in childexpertise development in informal settings. Science Education, 91(5), 712–732.

attison, S. A., & Dierking, L. D. (2012). Exploring staff facilitation that supports familylearning. Journal of Museum Education, 37(3), 69–80.

eterson, C., Jesso, B., & McCabe, A. (1999). Encouraging narratives in preschoolers:An intervention study. Journal of Child Language, 26(1), 49–67.

ruden, S. M., Levine, S. C., & Huttenlocher, J. (2011). Children’s spatial thinking: Doestalk about the spatial world matter? Developmental Science, 14(6), 1417–1430.

eese, E., Haden, C. A., & Fivush, R. (1993). Mother–child conversations about thepast: Relationships of style and memory over time. Cognitive Development, 8(4),403–430.

eese, E., & Newcombe, R. (2007). Training mothers in elaborative reminiscingenhances children’s autobiographical memory and narrative. Child Development,78(4), 1153–1170.

igney, J., & Callanan, M. (2011). Patterns in parent–child conversations about ani-mals at a marine science center. Cognitive Development, 26(1), 155–171.

ogoff, B. (1990). Apprenticeship in thinking: Cognitive development in social context.New York, NY: Oxford University Press.

ogoff, B. (2003). The cultural nature of human development. New York, NY: OxfordUniversity Press.

osenthal, E., & Blankman-Hetrick, J. (2002). Conversations across time: Familylearning in a living history museum. In G. Leinhardt, K. Crowley, & K. Knutson(Eds.), Learning conversations in museums (pp. 305–330). Mahwah, NJ: Erlbaum.

owe, S. (2005). Using multiple situation definitions to create hybrid activity space.In S. Norris, & R. H. Jones (Eds.), Discourse in action: Introducing mediated discourseanalysis (pp. 123–134). New York, NY: Routledge.

udek, D., & Haden, C. A. (2005). Mothers’ and preschoolers’ mental state languageduring reminiscing over time. Merrill-Palmer Quarterly, 51(4), 557–583.

anders, M. (2009). STEM, STEM education, STEMmania. Technology Teacher,68, 20–26. Retrieved from http://www.iteaconnect.org/Publications/AAAS/TTT%20STEM%20Article 1.pdf

chauble, L. (1996). The development of scientific reasoning in knowledge-rich con-texts. Developmental Psychology, 32(1), 102–119.

rch Quarterly 29 (2014) 333–344

Siegel, D., Esterly, J., Callanan, M., Wright, R., & Navarro, R. (2007). Conversationsabout science across activities in Mexican-descent families. International Journalof Science Education, 29(12), 1447–1466.

Sigel, I. E. (1993). The centrality of a distancing model for the development of rep-resentational competence. In R. Cocking, & A. Renninger (Eds.), The developmentand meaning of psychological distance (pp. 141–158). Hillsdale, NJ: Erlbaum.

Tamis-LeMonda, C. S., Sze, I. N. L., Ng, F. F. Y., Kahana-Kalman, R., & Yoshikawa, H.(2013). Maternal teaching during play with four-year-olds: Variation by ethnic-ity and family resources. Merrill-Palmer Quarterly, 59(3), 361–398.

Tare, M., French, J., Frazier, B. N., Diamond, J., & Evans, E. M. (2011). Explanatoryparent–child conversation predominates at an evolution exhibit. Science Educa-tion, 95(4), 720–744.

Taumoepeau, M., & Ruffman, T. (2008). Stepping stones to others’ minds: Maternaltalk relates to child mental state language and emotion understanding at 15, 24,and 33 months. Child Development, 79(2), 284–302.

Tenenbaum, H. R., & Callanan, M. A. (2008). Parents’ science talk to theirchildren in Mexican-descent families residing in the United States. Inter-national Journal of Behavioral Development, 32(1), 1–12. http://dx.doi.org/10.1177/0165025407084046

Tenenbaum, H. R., Callanan, M., Alba-Speyer, C., & Sandoval, L. (2002). The role of edu-cational background, activity, and past experiences in Mexican-descent families’science conversations. Hispanic Journal of Behavioral Sciences, 24, 225–248.

Tenenbaum, H. R., Snow, C. E., Roach, K., & Kurland, B. (2005). Talking and readingscience: Longitudinal data on sex differences in mother–child conversations inlow-income families. Journal of Applied Developmental Psychology, 26(1), 1–19.

Tessler, M., & Nelson, K. (1994). Making memories: The influence of joint encodingon later recall by young children. Consciousness and Cognition: An InternationalJournal, 3(3–4), 307–326.

Uttal, D. H., Haden, C. A., Wilkerson, E., Hoffman, P., Marcus, M., & Senior, A. (2013).Building bridges: How parent–child conversation and hands-on activity impactschildren’s STEM learning and transfer. Paper presented at the American Educa-tional Research Association, San Francisco, CA, May.

Welch-Ross, M. K., Fasig, L., & Farrar, M. J. (1999). Predictors of preschoolers’self-knowledge: Reference to emotion and mental states in mother–child con-versation about past events. Cognitive Development, 14(4), 401–422.

Wood, D. J., Bruner, J. S., & Ross, G. (1976). The role of tutoring in problem solving.Journal of Child Psychiatry and Psychology, 17(2), 89–100.

Valle, A., & Callanan, M. A. (2006). Similarity comparisons and relational analogies inparent–child conversations about science topics. Merrill-Palmer Quarterly, 52(1),96–124. http://dx.doi.org/10.1353/mpq.2006.0009

Vygotsky, L. S. (1978). In M. Cole, V. John-Steiner, S. Scribner, & E. Souberman (Eds.),Mind in society. Cambridge, MA: Harvard University Press.