You are on page 1of 17

Early Childhood Research Quarterly 59 (2022) 186–202

Contents lists available at ScienceDirect

Early Childhood Research Quarterly


journal homepage: www.elsevier.com/locate/ecresq

Review

A meta-analysis of teacher language practices in early childhood


classrooms
Elizabeth Burke Hadley a,∗, Erica M. Barnes b, Brenton M. Wiernik c, Mukhunth Raghavan c
a
Department of Language, Literacy, Ed.D., Exceptional Education, and Physical Education, University of South Florida, 4202 East Fowler Ave, EDU 105,
Tampa, FL 33620
b
Department of Literacy Teaching and Learning, University at Albany, SUNY; Albany, NY, United States
c
Department of Psychology, University of South Florida, Tampa, FL, United States

a r t i c l e i n f o a b s t r a c t

Article history: The purpose of this meta-analysis was to determine the relationships between teacher language practices
Received 23 July 2021 in early childhood classrooms, using a systemic functional linguistic theoretical framework. We included
Revised 1 December 2021
studies with typically developing preschool or kindergarten-aged children that conducted 1 or more ob-
Accepted 9 December 2021
servations of teachers in which teachers’ talk was systematically coded for strategies that supported oral
Available online 6 January 2022
language development. Studies also had to present at least one correlation between types of teacher lan-
Keywords: guage practice variables (e.g., open-ended questions and use of rare vocabulary). In a meta-analysis of 30
early childhood studies and 539 correlations, we identified correlations between 11 key types of teacher language prac-
oral language development tices. We then conducted a factor analysis using the overall meta-analytic mean correlations, finding that
vocabulary teacher language practices loaded onto 2 factors: an Emergent Academic Language dimension character-
meta-analysis ized by abstract talk with a greater variety of word types, and a Bridge Language dimension characterized
by scaffolding practices. Last, we explored how teacher language practices varied by activity setting and
group size, finding that shared book-reading enhanced the use of more Emergent Academic Language
practices, and that small group time included fewer beneficial language practices as compared to other
settings.
© 2021 Elsevier Inc. All rights reserved.

1. Introduction child talk, providing models for complex sentence structure and
high-level vocabulary use, and familiarizing children with an emer-
1.1. A meta-analysis of teacher language practices in early childhood gent academic language register (e.g., Hadley, Barnes, & Hwang,
classrooms 2021; Bowne Yoshikawa, & Snow, 2017; Cabell, Justice, McGinty,
DeCoster, & Forston, 2015; Deshmukh et al., 2019; Dickinson, Hofer,
Early childhood classrooms have a unique role to play in sup- Barnes, & Grifenhagen, 2014; Farrow, Wasik, & Hindman, 2020).
porting children’s language development. Positioned in between Foundational research has confirmed the importance of certain
the more formal schooling environment of elementary school and teacher language practices in early childhood for children’s aca-
children’s home language environments, early childhood teachers demic language and literacy development (Cazden, 1988; Dickinson
can mediate children’s introduction into the kinds of language val- & Smith, 1994; Dickinson & Tabors, 2001), finding that there
ued in school, while also providing opportunities for the nurturing, are long-term associations between a supportive preschool class-
playful interactions that are more characteristic of home learning room language environment and children’s language and reading
environments (e.g., Authors, 2020; Justice, McGinty, Zucker, Cabell, skills in kindergarten and fourth-grade (Dickinson & Porche, 2011).
& Piasta, 2013; Tompkins, Zucker, Justice, & Binici, 2013; Toub et al., Dozens of studies have grown from this early work, examining the
2018). Teachers’ talk in early childhood classrooms (ECCs) supports types and frequencies of teacher language practices in early child-
children’s language development in multiple ways: building con- hood classrooms, as well as their relationships to children’s oral
ceptual knowledge about the world, encouraging children to grow language growth (e.g., Bratsch-Hines, Burchinal, Peisner-Feinberg,
in their expressive language skills, responding to and affirming & Franco, 2019; Cabell, DeCoster, LoCasale-Crouch, Hamre, & Pi-
anta, 2013; Gámez, 2015; Hindman, Connor, Jewkes, & Morrison,
2008; Justice, Mashburn, Hamre, & Pianta, 2008; Silverman & Cran-

Corresponding author. dell, 2010). Considered together, these studies confirm the impor-
E-mail address: hadleye@usf.edu (E. Burke Hadley). tance of teacher language practices for children’s oral language

https://doi.org/10.1016/j.ecresq.2021.12.002
0885-2006/© 2021 Elsevier Inc. All rights reserved.
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

development, which in turn serves as the foundation for later read- The present review focuses on uncovering key types and pat-
ing skills. terns of teacher language practices, and variation of those practices
Although there is a large body of work on how EC teacher across instructional settings, rather than examining how those
language practices support oral language development, there have practices relate to child outcomes. We chose to focus on the re-
been few systematic reviews or meta-analyses on this line of lationships between teacher language practices only in this study
work. There are several potential reasons for this lack of synthesis. for conceptual as well as practical reasons: many of the stud-
First, studies examining teacher language practices in early child- ies we synthesize here also report observational data on teachers
hood classrooms (ECCs) are typically correlational and descrip- only, without linking those practices to children’s learning. We also
tive in nature, which can be more difficult to synthesize statisti- found that, in the studies that did correlate teacher language prac-
cally than interventional work. Second, the kinds of teacher lan- tices to children’s outcomes, there were very few studies that re-
guage practice variables typically coded for in these studies are, ported the information necessary for calculating effect sizes for the
on their face, fairly heterogenous, with a wide range of teacher relationships between teacher language practices and child out-
language practice types measured and inconsistent terminology comes.
used across studies. Finally, the reporting of statistics needed for We define early childhood classrooms in this review as those
meta-analysis is inconsistent, with few studies reporting the full that serve preschool and kindergarten-aged children (roughly 3 to
correlation matrices needed to calculate effect sizes. In previous 6 years old). Our primary interest here is in investigating teacher
work (Hadley, Barnes, & Hwang, 2021), we have begun synthe- language practices in pre-school, which we define as instruction
sizing this body of research by conducting a systematic review that takes place before the more formalized environment of el-
of teacher language practices and child outcomes, using the find- ementary school. Although kindergarten has become increasingly
ings to develop a theory of early childhood classroom language academic in the last decade in the United States, and there-
development. The present study builds on this previous work fore is arguably more similar to first grade than preschool in its
by using the theoretical framework developed in that paper, but structure and content (Bassok, Latham, & Rorem, 2016), kinder-
here we focus exclusively on teacher language practices, using garten remains the first formalized schooling experience for many
a meta-analytic approach to better understand how these prac- young children in the United States and elsewhere (Friedman-
tices cluster together and vary across typical classroom settings. Krauss et al., 2020). Therefore, we cast a relatively wide net in our
The correlational meta-analysis presented here has several pur- search, looking for research conducted in classrooms described as
poses: to further develop a theoretical framework for research on preschool, pre-k, pre-kindergarten, kindergarten, Head Start, child-
early childhood teachers’ oral language practices, to identify pat- care/daycare or even simply “early childhood.”
terns across teacher language practices, and to inform future re-
search that measures the quality of classroom language environ- 2. Theoretical framework: early childhood classrooms as hybrid
ments. language spaces
Our approach here, which uses meta-analysis with correlational
data, is relatively uncommon in educational research, where meta- Our theoretical framework here is rooted in systemic functional
analyses are typically conducted to synthesize effect sizes from ex- linguistics (SFL), which views language as inextricably tied to and
perimental interventions. However, we use meta-analysis, rather shaped by the purposes for which it is used, the places in which it
than more qualitative methods of synthesis, to summarize this area is used, and the participants with whom it is used (Hadley, Barnes,
of research for several reasons. First, the data we synthesize here & Hwang, 2021). Central to SFL is the notion of language regis-
are well-suited to meta-analysis as this literature includes mea- ters, or constellations of co-occurring lexical and grammatical fea-
sures of a variety of teacher language practices which we can clas- tures that serve different social purposes of language and vary
sify according to a taxonomy developed for the study. We can then across contexts (Biber, 1995; Halliday, 1978; Schleppegrell, 2004).
synthesize quantitative correlations among measures of different Young children possess a toolkit of culturally-shaped home lan-
language practice categories and apply meta-analytic factor anal- guage resources that have allowed them to effectively communi-
ysis to identify empirically-grounded, higher-order dimensions of cate with their caregivers, family, and community members. Upon
teacher language practices. Importantly, most studies in this lit- school entry, children become members of a new language com-
erature do not measure all of the language practice categories in munity that has its own set of expectations, norms, and valued
our taxonomy; meta-analysis allows us to integrate quantitative ways of speaking and listening, while also sharing their own indi-
findings from many studies to permit broader multivariate anal- vidual ways with words in the classroom space. The resulting “hy-
yses than would be possible from any single study (Viswesvaran & bridity,” or blend of home and school language practices present
Ones, 1995). Meta-analysis of this literature also has an advantage in the preschool classroom (Authors, 2020; Gutiérrez, Baquedano-
in that it can avoid one of the major limitations of meta-analysis López, & Tejeda, 1999), can create opportunities for further lan-
as a research synthesis method—publication bias. In many litera- guage development. In the following section, we discuss some of
tures, studies with small or no treatment effects can be judged the key characteristics of home and school language practices, with
as “unsuccessful” and therefore remain unpublished or unreported the understanding that these 2 registers are not dichotomous, but
and unavailable for inclusion in meta-analyses. This leads to in- rather exist on a continuum from more colloquial to more aca-
flated effect sizes from meta-analytic models. However, the corre- demic in nature (Snow & Uccelli, 2009).
lations we use here are typically not the focal topic of the study Home language practices can be defined as the kinds of lan-
and are instead reported as descriptive information. This type of guage practices families use together at home to accomplish the
incidentally-reported statistic is not subjected to the same selec- activities of everyday life, maintain relationships, and communi-
tion pressure, and so meta-analyses of these data generally do not cate information (van Kleeck, 2006). Such practices are shaped
suffer from publication bias (Mathur & VanderWeele, 2021). Thus, by the cultural, social, and socioeconomic backgrounds of fami-
quantitative synthesis can allow us to examine patterns of rela- lies and their beliefs about children’s development, learning, and
tionships among teacher language practices in a precise and trans- role in the family (Johnston & Wong, 2002; van Kleeck, 2007). De-
parent way. In contrast, qualitative syntheses of these relationships velopmental research on home language practices has often fo-
can be inaccurate due to their inability to capture the strength of cused on variation in the linguistic features of caregiver speech,
correlational relationships and/or account for differences in sample as measured by variables such as the number of words, the
sizes (e.g., Bushman & Wells, 2001). number of different words, and the mean length of caregiver

187
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

utterances (e.g., Vernon-Feagans et al., 2020). Ethnographic re- and audience do not share common knowledge (Michaels, 1981;
search has found other discourse-level, culturally-shaped differ- Schleppegrell, 2004; Snow & Uccelli, 2009).
ences in caregiver speech, such as the extent to which parents ECCs are positioned “in between” home and the more formal
specifically direct their speech to children (Sperry et al., 2019), schooling environment of the upper elementary grades, and blend
the frequency with which they tell personal stories (Miller, Cho, many of the features of each setting. For example, the focus on de-
& Bracey, 2005), and the kinds of questions they ask children velopmentally appropriate practice in ECCs means that warm, re-
(Scollon & Scollon, 1981). Certain socioculturally-shaped home sponsive relationships between teachers and children are valued:
language practices, such as parents’ commenting practices dur- indeed, the first standard for high-quality EC programs set by the
ing shared book-reading, closely mirror school language practices National Association for the Education of Young Children (NAEYC)
(Heath, 1982; Hindman et al., 2008; Hindman, Skibbe, & Foster, focuses on relationships, described as “teaching staff engag[ing]
2014), and the children whose home language practices align more in warm, friendly conversations with the children. . .encourag[ing]
closely with the language practices valued in school may therefore and recogniz[ing] children’s work and accomplishments.” Academic
have an advantage in communicating and learning through school learning is also valued in ECCs, with the use of activities “designed
language registers (Michaels, 1981; van Kleeck, 2014). to help children get better at reasoning, solving problems, get-
Academic language, as identified in the upper elementary and ting along with others, using language, and developing other skills”
middle school grades, involves the use of language that differs (NAEYC Standard 2).
from more conversational talk in terms of its lexical diversity and Given the unique affordances of the ECC environment, we ar-
structural complexity. Academic language typically includes the ar- gue here that ECCs can function as a hybrid space in which mul-
ticulation of precise meanings, use of rare vocabulary words, and tiple kinds of home and school language practices are employed
discussion topics beyond the here-and-now (Snow & Uccelli, 2009; to apprentice children into the new language community of school
Uccelli, Galloway, Barr, Meneses, & Dobbs, 2015). While these (Hadley, Barnes, & Hwang, 2021; Gutiérrez et al., 1999), and that,
features of academic language have been observed in texts in- rather than a problem to be solved, this hybridity can function
tended for older children, researchers have identified emergent as a scaffold or springboard for young children into a new lan-
academic language registers in preschool classrooms, including the guage community. These practices might include the use of emer-
use of complex syntax, rare words, and talk about abstract con- gent academic language by teachers (e.g., Dickinson et al., 2014),
cepts (Dickinson et al., 2014). Recent longitudinal research has dis- children’s use of their own home language practices, especially in
covered another important precursor of later academic language, settings such as play and mealtime that connect to home activities
finding that caregiver-child conversations about decontextualized for many children (e.g., Barnes, Grifenhagen, & Dickinson, 2020;
topics, or those that focus on abstract ideas that are not physi- Leyva & Skorb, 2017) and the use of interactive language practices
cally present in children’s environment, are a significant predictor by teachers that bridge or connect home and school language prac-
of academic language proficiency in seventh grade (Uccelli, Demir- tices (Gort & Sembiante, 2015).
Lira, Rowe, Levine, & Goldin-Meadow, 2019). However, not all chil-
dren have equal opportunities at home to practice emergent aca-
3. Dimensions of hybrid language spaces in early childhood
demic language registers (Schleppegrell, 2004). Given that aca-
classrooms
demic language can be either a gateway or gatekeeper to ac-
cessing the text children are expected to learn from in school
In the present meta-analysis, we investigated and summarized
(Uccelli, Phillips Galloway, Aguilar, & Allen, 2020), it is of critical
studies that examined the early childhood teacher language prac-
importance that educators understand the linguistic demands of
tices that were hypothesized to support children’s oral language
the academic environment and are able to help all children gain
development. In our initial review, we found that a wide range of
facility with school-valued language practices, beginning at chil-
terms were used to describe teacher language practices. Through
dren’s school entry.
an iterative coding process (see details in the Method section be-
The extent to which school and home language practices over-
low), drawing on findings from our previous related systematic re-
lap or diverge within different communities is still an open
view (Hadley, Barnes, & Hwang, 2021), as well as prior work in this
question, but available research indicates several key areas in
area (Rowe & Snow, 2020), we identified several major categories
which school language practices are distinct from many chil-
of teacher language practices that we hypothesize support chil-
dren’s home language experiences. First, the content of conversa-
dren’s acquisition of school-valued oral language skills. We briefly
tions between children and caregivers at home typically focuses
discuss each category below.
on more concrete, literal, and “common-sense” topics (Halliday,
1993), in contrast to academic language, which is characterized
by abstract and generalized observations about the world com- 3.1. Linguistic features
municated in specialized and technical language (Vygotsky, 1978;
Schleppegrell, 2004). The nature of participation structures at Perhaps the most well-researched dimension of teacher lan-
home vs school also differ. Children often engage in more dyadic guage in ECCs are the linguistic features of teacher language, also
language interactions with caregivers at home, therefore afford- described as the “data-providing” features of talk (Justice, Jiang, &
ing more opportunities for child talk (Hindman et al., 2008). In Strasser, 2018) or structural features of language (Dickinson et al.,
contrast, classroom settings involve more time engaged in didactic 2014). While the study of linguistic features could theoretically in-
whole-group settings in which a teacher instructs many children volve the study of any number of word-, sentence-, or discourse-
at once, and individual children may have less time to engage in level features, the research to date has typically focused on linguis-
contingent conversations with an adult (Early et al., 2010). Finally, tic features that are characteristic of emergent academic language,
running alongside these differences in participation structures is a such as information about the complexity of teachers’ syntax, the
fundamental difference in the tenor (Halliday & Hasan, 1989), or number of words used (tokens), number of different word types,
relationship between the speaker and audience, in home vs school the number of rare words or academic vocabulary, and lexical di-
language registers, with home language interactions characterized versity. Hearing the linguistic features associated with an emer-
by warmth, references to shared knowledge and experiences, and gent academic language register, such as sentences with multiple
the growth of personal relationships, whereas academic language clauses, helps young children to learn the forms of language that
relies on an impersonal stance and an assumption that the speaker are valued in a school setting.

188
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

3.2. Conceptual talk tings. For example, during small group time when a child sees a
picture of a volcano and says, “It’s going to explode!,” a teacher
A second facet of an ECC hybrid language space concerns the might affirm and scaffold their statement by replying, “Yes, that
content or topics of conversation. While talk at home typically dis- volcano’s going to erupt!” Michaels (1981) identified sharing time
cusses topics that are immediate, including perceptually present as a distinct discourse setting in which children were expected to
objects and personal or shared experiences, academic language display their knowledge in a certain valued style. When teachers
discusses abstract, general ideas that are further removed from were skilled at supporting children’s talk, and could help children
the present time and space (Snow, Lawrence, & White, 2009; to use the expected style of “sharing time talk” by asking ques-
van Kleeck, 2014). This includes inferential talk, or using informa- tions, sharing time messages were co-constructed by teachers and
tion that is not immediately present or directly stated to predict, children, giving children scaffolded practice in making their ideas
summarize, interpret, and explain (van Kleeck, 2014). For exam- explicit and using complex syntax (Michaels, 1981), much as is ex-
ple, a teacher might explain during a shared book-reading session, pected in an academic language register.
“So now we know that fruits have seeds and they can grow on Interactive talk includes 2 kinds of talk: modeling strategies,
trees.” We call this kind of language “conceptual talk” here be- in which a teacher builds on child talk to provide a model
cause it is typically about concepts and ideas rather than physi- for emergent academic language (e.g., extensions and expan-
cally present objects; it is often referred to as “decontextualized” sions), and strategies that encourage children to speak and to en-
talk in the literature because of its removal from children’s imme- gage in extended conversations (Justice, Jiang, & Strasser, 2018),
diate, physical context (Rowe, 2012; Uccelli et al., 2019), although sometimes described as elicitations (Cabell et al., 2015). When
some have argued that this term is imprecise because no lan- preschool teachers used more elicitation strategies such as ask-
guage is truly context-free (Schleppegrell, 2004). Other researchers ing open-ended questions (e.g., “Can you tell me about your pic-
have referred to similar kinds of abstract talk as high-level ture?”) or waiting expectantly for children to speak, children talked
(Barnes & Dickinson, 2017a), conceptual (Rowe & Snow, 2020), an- more, used longer sentences, and used a wider variety of words
alytic (Dickinson, 2011), or inferential (Mascareño, Snow, Deunk, & (Girolametto & Weitzman, 2002). These child-driven conversations
Bosker, 2016). One specific type of abstract talk involves the in- help teachers to learn about children’s own individual language
struction and discussion of vocabulary words. This explicit focus practices and open up a space for socializing children into aca-
on teaching and learning word meanings is distinct from teachers’ demic language practices through authentic engagement in talk
largely unconscious use of sophisticated vocabulary as described about topics that children are interested in.
above in the section on linguistic features of talk. Rather, this vo-
cabulary talk involves the explicit teaching of new word mean-
4. Variation in teacher language based on activity setting and
ings, including definitions and examples of use in context (e.g.,
group size
Silverman & Crandell, 2010). Words become the focus of discus-
sion and an object of thought, which may promote word con-
Academic language is setting specific, such that different
sciousness or metalinguistic awareness, both of which have been
academic disciplines each have a distinct academic register
associated with children’s language development (Nagy, 2007;
(Schleppegrell, 2004). For younger children, for whom content area
Neugebauer, Coyne, McCoach, & Ware, 2017).
divisions are not as clear-cut, language practices have been found
to vary across different instructional settings, with the nature of
3.3. Literal talk
talk influenced by participants, materials, and topics of study, as
well as the size of the group (Dickinson et al., 2014). Studies de-
Literal talk, sometimes described as low-level, concrete, or con-
scribing adults’ use of emergent academic language practices re-
textualized talk, exists on the opposite end of the spectrum from
veal differences among instructional settings involving different
conceptual talk, as it focuses on topics in the “here-and-now.” Lit-
disciplines at the word and/or sentence level (Barnes & Dick-
eral talk provides a number of physical supports for interpret-
inson, 2017b; Dickinson et al., 2014; Weizman & Snow, 2001),
ing and understanding the messages that are communicated, such
and discourse level (e.g., Nagy & Townsend, 2012). Language-
as pointing, gestures, and people and objects that are present in
focused studies in pre-kindergarten and kindergarten classrooms
the immediate environment. Literal talk is also more familiar to
have largely focused on shared book reading settings, and have
many children because of its use in most home language environ-
found that this setting promotes the use of linguistic features
ments (e.g., Hindman et al., 2008). While literal talk has some-
typical of emergent academic language, as well as many concep-
times been dismissed as unhelpful for language development in
tual and interactive talk practices (Barnes & Dickinson, 2017b;
preschool as many children no longer need such explicit support
Dickinson & Porche, 2011; Mol, Bus, & de Jong, 2009). A more lim-
for understanding language, such talk can provide a means of ac-
ited set of studies have evaluated teacher language across the day
cessing more conceptual, abstract language, particularly for chil-
in multiple activity settings (e.g., Cabell et al., 2013), including play
dren who are still in an emergent stage of language acquisition
activities, science lessons, and circle or morning meeting time, al-
(Hindman, Wasik, & Erhart, 2012).
though consistent patterns of talk across those settings have not
yet been identified.
3.4. Interactive talk

The final dimension of a hybrid language space we discuss here 4.1. Previous meta-analyses and syntheses
is the use of interactive language practices, or talk strategies that
both encourage child talk and extend or respond to children’s con- To our knowledge, our recent systematic review on teacher lan-
versational offerings. Interactive talk practices are distinct from the guage and child outcomes (Hadley, Barnes, & Hwang, 2021) was
other language practices discussed here in that they are responsive the first to systematically summarize this line of research to date.
in nature and can help to “bridge” or scaffold children’s language In our previous systematic review, we identified the key strands
to align with school-valued language practices. In the context of of teacher language practices that we describe above and analyzed
supportive interactions, teachers can encourage children to try out findings from previous studies to develop the theory of ECCs as
new sentence structures, use complex words, and implicitly or ex- hybrid language spaces. We build on that work in the present pa-
plicitly teach them the kinds of talk that are valued in certain set- per by using meta-analytic methods to examine the relationships

189
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

between these strands of teacher language practices and examine conversation∗ or “book reading” or instruction or speech (terms for
how they vary across settings. teacher’s language practices) AND classroom or teach∗ (terms about
Other recent studies, while not meta-analyses or systematic re- setting). These searches yielded a total of 3,821 articles. Next, we
views, have provided conceptual overviews of the field that we hand-searched the references of relevant articles from the initial
draw on here. For example, Rowe and Snow (2020) theorized that database search published in the last 2 years and searched for ad-
3 strands of high-quality caregiver input were related to children’s ditional articles by authors who had more than one article in our
language development, including linguistic, conceptual, and inter- sample after full-text screening, yielding an additional 26 articles.
active talk, then used a developmental framework to trace the In total, we had 2699 articles after duplicates were removed (see
use of these talk features with children from birth to age 5. The Fig. 1).
present study builds on Rowe and Snow’s (2020) work by extend-
ing it to a classroom context, employing systematic review meth-
5.3. Study selection
ods to more comprehensively survey the existing research, and us-
ing meta-analysis to determine the correlational relationships be-
All titles and abstracts were reviewed by the first and second
tween and among these strands.
author using the above inclusion and exclusion criteria. We had
high interrater reliability (92% agreement) as to whether studies
4.2. Research questions
should be included. Disagreements were resolved through discus-
sion or by adding ambiguous studies to the full-text screening
The present systematic review and meta-analysis asks the fol-
pool.
lowing questions:
After the initial screening of titles and abstracts, 231 articles re-
What are the relationships between the different types of early
mained. The full-text of these articles was again screened for el-
childhood teacher language practices across studies?
igibility using the inclusion and exclusion criteria above. An ad-
Which underlying factors, or clusters, of teacher language prac-
ditional 201 articles were deemed ineligible during this second
tices can be identified across studies?
round of full-text screening (see Fig. 1). We had high interrater
Which patterns of teacher language practices are evident across
reliability for exclusion of articles (86% agreement), and we again
different activity settings and with different participants?
resolved disagreements through discussion.
After full-text screening, we were left with a pool of 60 stud-
5. Method
ies that met inclusion and exclusion criteria. Twenty-five of these
studies met other inclusion and exclusion criteria but did not
5.1. Inclusion and exclusion criteria
report full descriptive or correlational information necessary for
computing effect sizes. We emailed corresponding authors to re-
We used the following inclusion criteria to select studies for
quest data and received responses with data for 2 studies, meaning
our sample: (1) child participants were typically developing chil-
that 23 studies in total were excluded based on missing or incom-
dren in preschool, kindergarten, Head Start, daycare/childcare, or
plete data.
early childhood classrooms; (2) the study primarily took place in
In total, 30 studies met our inclusion and exclusion criteria and
a classroom, rather than home, setting; (3) the study included one
were coded using the criteria described in the following section.
or more observations of teachers in which teacher’s talk was sys-
tematically counted and/or coded for strategies that supported oral
language development; (4) the study presented at least one corre- 5.4. Coding procedure
lation between teacher language variables, and 5) the study was
written in English. We did not exclude studies based on the ages We developed a detailed codebook to extract relevant informa-
of children, but rather excluded only studies that took place in a tion from studies, including citation and study information. First,
classroom setting outside of preschool, kindergarten, or early child- we coded for citation information such as publication type, year of
hood (e.g., studies in first grade classrooms). We excluded articles publication, and authors. We then coded key features of the study
that examined teacher language practices by using a global ob- design, sample characteristics, teacher language variables, potential
servation system such as the Classroom Assessment Scoring Sys- moderator variables such as activity setting and group size, and
tem (CLASS, Pianta, La Paro, & Hamre, 2008) or the Early Literacy effect sizes. Table 1 summarizes the key characteristics of the arti-
and Language Classroom Observation scale (ELLCO, Smith, Dickin- cles; a table with a full summary of all characteristics of individual
son, Sangeorge, & Anastasopoulos, 2002), as results from these ob- studies are available upon request from the authors.
servational tools typically reported an aggregated score that char-
acterized the overall quality of the classroom language environ-
5.5. Study design
ment, rather than reporting the frequency of the more fine-grained
teacher language practices that are the focus of this review. We in-
For the study design, we noted if each study was descrip-
cluded both published and unpublished studies in our sample to
tive/observational or whether data were drawn from larger inter-
limit potential publication bias.
ventions intended to support children’s oral language development
and/or teacher’s instructional practices. Studies were coded as ob-
5.2. Information sources
servational if data were drawn from a larger intervention study
but were collapsed across treatment and control conditions (e.g.,
Articles for this meta-analysis were identified in 2 ways. First, a
Barnes, Dickinson, & Grifenhagen, 2017; Bowne et al., 2017).
literature search was conducted by searching the databases ERIC,
Education Source, and PsycInfo. We used the earliest possible
start date through June 2021. Titles, abstracts, and keywords were 5.6. Sample characteristics
searched using the following terms: syntax or grammar or vocab-
ulary or comprehension or “oral language” or “phon∗ awareness” or We coded the participant characteristics for both teachers and
discourse (oral language terms) AND preschool or pre-k or prekinder- children. Few child or teacher characteristics were reported consis-
garten or kindergarten or “early childhood” or “Head Start” (chil- tently across studies, but we coded for child age, child grade, gen-
dren’s age terms) AND language or talk or input or linguistic or der, race/ethnicity information, the number of low-socioeconomic

190
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Fig. 1. PRISMA flow diagram of study screening and selection.

Table 1
Characteristics of samples included in meta-analyses.

Characteristic N Mean SD Min Max

Included studies 30
Year 30 2015.93 3.48 2008.00 2021.00
Child sample size 28 304.89 154.49 70.00 592.00
Teacher sample size 30 41.33 20.72 13.00 122.00
Mean child age (months) 27 55.15 8.15 40.80 73.40
Mean teacher experience (years) 24 12.52 3.70 4.90 19.86
Country N %
USA 28 93%
Singapore 1 3%
Switzerland 1 3%
Study design
Observational 25 83%
Intervention 5 17%
Child grade
Pre-K 21 70%
Kindergarten 7 23%
Pre-K and kindergarten 2 7%
Low SES students
Mixed 10 45%
All 12 55%
Dual-language learner students
0%–1% 6 33%
9%–45% 9 50%
100% 3 17%

Note. N = for continuous variables, number of studies reporting information on sample


characteristic; for categorical variables, number of studies at each variable level.

status (SES) children in the sample, and the number of dual lan- coded for their race/ethnicity, gender, highest level of education at-
guage learners (DLLs) in the sample. To be coded as low-SES, chil- tained, and their years of experience.
dren had to be described in the sample as meeting specific cri-
teria for low-SES status, such as a family income below 150% of 5.7. Activity setting and group size
the poverty line, receiving free or reduced-price lunch, or attending
Head Start or other publicly-funded preschools designated for fam- We coded for the activity setting by examining the instructional
ilies living in poverty. Children were coded as DLLs when the study context for the teacher language practices that were observed. For
noted that children in their sample spoke a different language at example, some studies recorded and coded teacher talk within
home, in addition to the language used at school. For teachers, we shared book-reading, while others recorded talk during mealtime,

191
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Table 2
Teacher language practice codes, descriptions, and examples.

Macro Code Description Examples of variables included

Literal talk Questions or comments that focus on topics in the here-and-now Literal questions,
Concrete talk
Management talk Talk designed to manage children’s behavior or describe rules and Directions,
procedures Procedural talk,
Behavior-disapproving talk
Conceptual talk Questions or comments that focus on topics beyond the Abstract talk,
here-and-now and/or seeks to build knowledge of concepts. Topic-focused talk,
Includes talk about vocabulary words. Vocabulary talk
Vocabulary talk Talk about vocabulary words, their meanings, and examples of usage Definitions of words,
Examples of word use,
Word consciousness talk
Questions and Elicitations All questions or prompts in which a child response is expected or Open or closed-ended questions, Literal questions,
desired Inferential questions, Communication-facilitating talk
High-level questions Questions that ask for abstract thinking or extended responses Inferential questions,
Open-ended questions,
Vocabulary-related questions
Linguistic Features Lexical and grammatical features of teacher speech Types and tokens.
Syntactic complexity,
Use of rare vocabulary
Word Types and Tokens Count of the number of words and unique word types in teacher Total number of words,
speech Number of different word types, Type/token ratio
Syntax Measure of the complexity of teacher speech Mean length of utterance (MLU),
Number of clauses per T-unit
Modeling Talk that provides a model of school-valued language Modeling the use of new vocabulary,
Orally narrating child’s actions,
Modeling use of new materials
Responsive talk Talk that responds to a child’s utterance or overture Responding to child talk,
Repeating or confirming,
Active listening

and still others recorded talk throughout the school day without 5.9. Effect sizes and sample sizes
specifying the settings. We developed a list of common activity
settings from the studies, then coded each teacher language vari- For each study, we coded correlations among the teacher lan-
able with the setting in which it was observed. Activity setting guage variables measured, as well as the teacher sample sizes asso-
codes included (1) shared book-reading, (2) language/literacy in- ciated with these correlations. All eligible studies reported teacher
struction, which comprised explicit instruction on literacy skills sample sizes; however, 30 studies did not report correlations be-
such as letter names and sounds, (3) play activities such as socio- tween teacher language variables or statistics that could be trans-
dramatic play or play with manipulatives such as playdough, (4) formed into correlations and were therefore excluded from our fi-
mealtimes, (5) science, or (6) not reported. nal sample.
Group size codes reflected how many children were present at
the time the teacher talk was recorded and included (1) whole
5.10. Reliability of coding
class, (2) small group, (3) mixed (observations that included both
whole class and smaller group time), and (4) not reported.
Both the first and second author coded all studies individually
and met weekly throughout the coding process to resolve disagree-
5.8. Teacher language variables
ments and minimize coding drift. Across all variables, the mean in-
terrater agreement was 89%, with values of 85% for participant and
Teacher language variables were described in various ways
demographic variables and 90% for variable characteristics. Dis-
across studies, with little consistency in terminology even for sim-
agreements were resolved through discussion and consulting the
ilar constructs. In an iterative, inductive process, we developed a
original article.
coding scheme to group together similar variables using macro
codes. First, we copied all teacher language variable names and
verbatim descriptions into a spreadsheet. After reviewing variables, 5.11. Data analysis
we identified several common types of teacher language practices
across studies, then independently coded all variables applying 5.11.1. Meta-analysis
those codes. After discussion, we added additional codes for vari- We estimated separate meta-analysis models for each pair of
ables that were not well-represented by existing codes and com- teacher language practice codes. For each pair of teacher lan-
bined other codes that were overly specific. We then coded all guage practice codes, we fit multilevel random effects meta-
teacher language variables again with the finalized set of codes. analysis models to account for effect size dependency within
Table 2 gives a list of final teacher language codes and descrip- studies, with random intercepts for each study. We computed
tions. Note that these codes are not mutually exclusive and that estimated sampling error variances for each study using the
more than one code was often applied to a variable (e.g., open- sample-size weighted mean correlation for the construct pair
ended questions were coded using both the abstract talk and the (Bakbergenuly et al., 2019; Schmidt & Hunter, 2015). We es-
question codes). timated the amount of heterogeneity (i.e., τ ) using the re-
In addition to the type of teaching language, we also coded stricted maximum-likelihood estimator (Viechtbauer, 2005). We
other information about the teacher language measures, including computed confidence intervals for meta-regression coefficients and
the reliability of coding for that talk, number and duration of ob- mean treatment effects using the Knapp and Hartung (2003) t-
servations, and means and standard deviations. distribution method and for random effects variance components

192
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

using profile likelihood (Viechtbauer, 2007). For all analyses, we re- concern for the current meta-analyses. Correlations among teacher
port results when the number of studies was k ≥ 2. language variables are rarely the central focus of the study. In-
stead, these correlations are reported incidentally as part of stud-
5.11.2. Moderator analyses ies focused on testing other hypotheses (e.g., the relationships of
We used the subgroup analysis approach to test moderator ef- teacher language variables with child outcomes), or are part of
fects, estimating a separate meta-analysis model for each level of larger observational studies aimed at capturing typical practice in
a moderator variable (Schmidt & Hunter, 2015). Our primary mod- early childhood classrooms. Accordingly, the correlations that are
erator analyses focused on group size (whole class, small group, the focus of the current study are unlikely to have been subject to
mixed) and activity setting (book reading, literacy instruction, play, selective reporting or similar biases.
circle time, meal time, group activity). As exploratory analyses, we
also examined moderating effects of the percentage of students in
5.11.5. Analysis software
the sample that were low SES or dual-language learners (DLL). Al-
We fit models using R (version 4.1.0; R Core Team, 2021). We
though these percentage variables are continuous, the range of val-
computed sampling error variances using the psychmeta package
ues in the included samples was limited, so we grouped studies
(version 2.6.0.90 0 0; Dahlke & Wiernik, 2017/2021, 2019) and es-
according to whether all students were low SES/DLL, no students
timated meta-analyses with the metafor package (version 3.1-3;
were low SES/DLL, or students had a mixture of SES/DLL statuses.
Viechtbauer, 2010, 2021). We imputed missing correlations for fac-
tor analyses with the filling package (version 0.2.2; You, 2020)
5.11.3. Meta-analytic factor analyses
and estimated factor analyses using the psych package (ver-
We used the overall meta-analytic mean correlation matrix as
sion 2.1.6; Revelle, 2021). We used the readxl (version 1.3.1;
input to exploratory factor analysis models to examine patterns of
Wickham & Bryan, 2019), readr (version 1.4.0; Wickham & Hes-
relationships across teacher language variable types. We used max-
ter, 2020), here (version 1.0.1; Müller, 2020), and dplyr (version
imum likelihood factor analysis with direct oblimin rotation. We
1.0.7; Wickham, François, Henry, & Müller, 2021) packages for data
used several methods to determine the number of factors to retain.
management, and the ggplot2 (version 3.3.5; Wickham, 2016), see
First, we examined a scree plot using parallel analysis (Horn, 1965,
(version 0.6.5; Lüdecke, Patil, et al., 2021), and ggtext (version 0.1.1;
using the harmonic mean sample size across the meta-analyses as
Wilke, 2020) packages for plotting.
the sample size for the simulated data (Viswesvaran & Ones, 1995).
In parallel analysis, eigenvalues for the observed correlation matrix
are compared to eigenvalues for randomly simulated uncorrelated 6. Results
data; the number of factors with eigenvalues larger than those for
random data is taken as an upper bound on the number of fac- 6.1. Correlations among teacher language practice types
tors to retain. Parallel analysis is similar to the “eigenvalue > 1”
Kaiser criterion, but is more stable at smaller sample sizes and in Overall meta-analysis results among teacher variables are illus-
the presence of minor factors. trated in Fig. 2, which reports the overall correlations for each pair
In addition to parallel analysis, we also compared Bayesian of teacher language practices. Full meta-analytic results tables, in-
information criterion (BIC), Revelle and Tonkin’s Very Simple cluding the confidence intervals and number of studies for each
Structure (VSS) criterion, and Velicer’s minimum average par- pair of correlations, are available upon request from the authors.
tials (MAP) values for solutions with varying numbers of factors For overall analyses, the number of studies reporting between each
(Revelle, 2009). BIC estimates the cross-validated log-likelihood pair of teacher language variables ranged from k = 2 to 23, with
factor models, penalizing models for increasing complexity (lower total teacher sample sizes ranging from N = 56 to 911. In general,
values are better, ࢞BIC > 6 is generally taken to indicate noticeably correlations among different teacher language practice types were
better fit). VSS and MAP aim to identify the most interpretable fac- small to moderate in magnitude (|r̄|ranged 0.03 to 0.47). Correla-
tor solutions. VSS identifies the number of factors where each in- tions with other language variables were strongest for Vocabulary
dicator most clearly shows a strong loading on a single factor and Talk (r̅ = 0.30 across language variables) and Modeling (r̅ = 0.27)
very weak loadings on other factors. MAP finds the number of fac- and weakest for Types and Tokens (r̅ = 0.14) and High-Level Ques-
tors where the average squared residual correlation among indica- tions (r̅ = 0.15). These patterns of correlations indicate that al-
tors after extracting factors is minimized. Each of these methods though many aspects of teacher language are related, each of the
has advantages and disadvantages, so consensus among methods variables we consider captures unique aspects of teacher language.
and interpretability of the final factor solution are used to choose When examining convergent correlations among alternative
among possible solutions (Revelle, 2009). measures of a single teacher language variable, we see that some
Eight of the 55 remaining correlations were not reported in any aspects of teacher language are more homogenous than others.
included study. Plausible values for these correlations were im- Complex Syntax (r̅ = 0.63 [95% CI 0.44, 0.83]), Linguistic Features
puted using nuclear norm optimization (Candès & Recht, 2009). (r̅ = 0.48 [95% CI 0.29, 0.66]), and Types and Tokens (r̅ = 0.44 [95%
This approach relies on the same assumption as factor analysis— CI 0.15, 0.73]) were the most homogeneous categories, showing the
that the correlation matrix has an underlying low-dimensional highest convergent correlations within-category. Responsive Talk
structure—so it is an appropriate imputation method for factor (r̅ = −0.06 [95% CI −0.22, 0.11]) and Conceptual Talk (r̅ = 0.13 [95%
analysis. The imputed complete correlation matrix is available CI −0.01, 0.28]) were the most heterogeneous, with measures in
upon request from the authors. For the factor analyses, we ex- these groups converging only weakly. Low within-category conver-
tracted factors using maximum likelihood and rotated them using gent correlations may reflect poor measure reliability, as well as
direct oblimin. variation in measurement modalities and variable conceptualiza-
tion across measures.
5.11.4. Publication bias and selective reporting Correlations among teacher language variables were generally
Typical methods for assessing potential for publication bias in quite heterogeneous across studies (mean τ = 0.24 across meta-
meta-analysis (e.g., the PET-PEESE test for funnel plot asymme- analyses), suggesting the presence of moderators. Some pairs of
try, cumulative meta-analysis; cf. Carter, Schönbrodt, Gervais, & teacher language variables were not measured together in any sin-
Hilgard, 2019) have not yet been extended for use in multilevel gle study (e.g., Complex Syntax and Management Talk), and other
meta-analysis. However, publication bias is unlikely to be a major pairs were measured together in relatively few studies, leading to

193
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Fig. 2. Overall mean correlations among teacher language practices. Note. Darkness of shading indicates the magnitude of the correlation. Positive correlations in blue.
Negative correlations in red.

Table 3 Table 4
Criterion values for determining the number of factors to retain in meta-analytic Meta-analytic factor loadings for teacher language practices for 2-factor solution
factor analysis.
Macro code Emergent academic Bridge h2
Number of Eigenvalue Median Random VSS MAP BIC language language
factors Eigenvalue
Literal talk .12 0.38 0.17
1 2.35 0.77 0.50 0.037 −13 Management talk −0.15 0.89 0.77
2 0.95 0.30 0.54 0.036 −77 Conceptual talk 0.30 0.19 0.15
3 0.36 0.23 0.52 0.053 −67 Vocabulary talk 0.36 0.63 0.59
4 0.12 0.15 0.56 0.080 −56 Questions and Elicitations 0.13 0.51 0.29
5 0.07 0.08 0.50 0.102 −38 High-level questions −0.09 0.43 0.18
6 0.02 0.02 0.51 0.132 −19 Linguistic Features 0.62 −0.04 0.38
Word Types and Tokens 0.63 −0.13 0.38
Note. Values computed using harmonic mean n = 192 as the sample size. Paral- Syntax 0.51 −0.02 0.26
lel analysis indicates to retain only factors whose eigenvalue is larger than the Modeling 0.45 0.25 0.30
corresponding median eigenvalue for simulated random data. VSS indicates to Responsive talk 0.26 0.30 0.18
retain the first solution where the criterion value is larger than the criterion Proportion of variance accounted for 0.15 0.18
value for the next larger solution. MAP indicates to retain the solution with the Factor intercorrelation 0.15
smallest criterion value. BIC indicates to retain the solution with the smallest
criterion value (࢞BIC > 6 is generally taken as indicating a substantial differ- Note. Factors extracted using maximum likelihood and rotated using direct oblimin.
ence in fit). Indicated numbers of factors for each criterion highlighted in bold. Factor loadings ≥ .30 highlighted in bold. h2 = communality.

wide confidence intervals (e.g., Questions/Elicitations and Manage-


factor represents teachers’ use of school-valued language practices
ment Talk, k = 3, r̅ = 0.47 [95% CI −0.12, 1.00]). These patterns
that reflect the discussion of abstract ideas and concepts and the
indicate a need for further research exploring patterns of covari-
use of certain linguistic features such as longer sentences and a
ation across these aspects of teacher language, as well as factors
greater variety of word types. Therefore, we labeled this factor the
that moderate these relationships.
Emergent Academic Language register. This second factor was de-
fined by Questions and Elicitations, High-Level Questions, Respon-
6.2. Factor analysis of teacher language variable intercorrelations sive Talk, Literal Talk, and Management Talk. This factor represents
interactive language practices that encourage child talk, support
To aid in interpreting patterns of correlations across teacher and extend child talk, and use more concrete language practices
language variables, we conducted a factor analysis using the overall to scaffold children’s emergent language. We labeled this factor
meta-analytic mean correlations. Criterion statistics used to deter- the Bridge Language register because of its role in helping chil-
mine the number of factors to retain are shown in Table 3. The dren make connections between their own individual language
parallel analysis scree plot is shown in Fig. 3. Parallel analysis sug- practices and those valued in a school setting. Vocabulary Talk
gested 2 or 3 factors. VSS, MAP, and BIC each suggested 2 factors. showed strong loadings onto both factors, with a stronger load-
Examination of the loading’s matrix showed that a third factor was ing onto Bridge Language. The 2 broad teacher language factors
defined by a single variable (Literal Talk), so the 2-factor solution were only weakly correlated (r = 0.15), indicating that teachers’
was retained. Factor loadings are shown in Table 4. frequencies of Emergent Academic Language and Bridge Language
The first factor was defined by Conceptual Talk, Types and To- reflect largely independent aspects of their language use. Impor-
kens, Complex Syntax, Linguistic Features, and Modeling Talk. This tantly, these 2 factors account for only 33% of the total variance

194
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Fig. 3. Parallel analysis screen plot.

in the teacher language variables, and teacher language variables −0.37, +0.06] in small group settings). This pattern may reflect that
generally retained substantial amounts of unique variance separate teacher’s tend to use fewer beneficial language practices in small
from that shared with the broad factors. This means that, although groups than in other, larger group settings, but may also mean
the 2 broad registers of Emergent Academic Language and Bridge that teachers use more individualized language practices that are
Language are useful for describing certain patterns of teachers’ lan- tailored to children’s needs when working directly with smaller
guage use, there is also value in considering more specific variables groups, but more general speech patterns when addressing the
and their unique antecedents and impacts. class as a whole. As with the activity setting analysis, many cells in
this moderator analysis were empty, and more research is needed
6.3. Moderation of teacher language variable correlations across to understand the process through which teachers modify their
contexts language practices to address different configurations of children.

We examined setting and group size as potential moderators of 6.3.2. Student sample composition
teacher language variable intercorrelations. These results are illus- In addition to the above moderator variables, as exploratory
trated in Figs. 4A and 4B. analyses, we also considered whether teacher language variable in-
tercorrelations were moderated by the student sample composition
6.3.1. Activity settings and group size in terms of low SES students or dual-language learner students.
Direct comparison of teacher language variable correlations was The number of studies contributing to these moderator analyses is
possible between book reading and literacy instruction settings for small, so results should be viewed as suggestive. Figure S1 (in the
several pairs of language variables. In general, correlations were online supplemental materials) depicts the correlations between
similar across these settings, with correlations among most vari- pairs of teacher language practices for 100% low-SES vs mixed-SES
ables being somewhat to moderately higher in literacy instruction samples (in panel A) and for 100% DLL vs mixed-DLL samples (in
settings than in book-reading. However, correlations were generally panel B). For SES, correlations were overall similar across 100% low
higher in book-reading than literacy instruction for 2 key elements SES vs mixed SES samples. However, most variable pairs were only
of Emergent Academic Language: Conceptual Talk and Vocabulary studied in 100% low SES samples, so the impact of SES as a mod-
Talk. Importantly, many pairs of variables were only assessed in erator variable could not be studied for most relationships.
book reading contexts or in play contexts. Thus, more research is For dual-language learner composition, correlations were gen-
needed to understand the impact of setting, particularly play vs erally similar across sample composition. Most correlations were
more structured settings, on patterns of teacher language usage. somewhat higher in mixed dual-language learner samples com-
The empty cells in Fig. 4A provide a visual indication of areas most pared to 100% DLL samples. This pattern may reflect a similar pro-
in need of future research. cess as posited for group size above—when working with 100% DLL
Direct comparison of teacher language variable correlations groups, teachers may use more individualized speech tailored to
across group sizes was possible for most variables. For most lan- specific student needs than they use when working with commin-
guage variable pairs, correlations were smaller or even negative gled student samples. One exception to this pattern was the cor-
in small group settings compared to whole class or mixed set- relation between Conceptual Talk with Questions and Elicitations,
tings (e.g., for Conceptual Talk and High-Level Questions, r̅ = 0.60 which were more highly correlated in 100% DLL samples (r̅ = 0.41
[95% CI 0.05, 1.00] in whole class settings, but r̅ = −0.15 [95% CI [95% CI 0.28, 0.54]) vs mixed DLL samples (r̅ = −0.05 [95% CI

195
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Fig. 4. Comparison across book reading literacy instruction and play setting. Note. Darkness of shading indicates the magnitude of the correlation. Positive correlations in
blue. Negative correlations in red.

196
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

−0.37, +0.27]). As with the other moderators, many cells for these the more objective nature of the measures. Like syntactic complex-
analyses were empty, so more research is needed to understand ity, the amount and diversity of teacher language can also be auto-
how teachers tailor their speech patterns for students with differ- matically calculated using CLAN. These results suggest that the Lin-
ent language backgrounds. guistic Features variables are likely more comparable across studies
than some of the more subjective and conceptually broad measures
7. Discussion such as Responsive Talk (r̅ = 0.24) and Literal Talk (r̅ = 0.22).
However, even some of the teacher language categories that
The aim of the present systematic review and meta-analysis were more subjective (in that they required judgment and
was to categorize and identify relationships between the EC decision-making by coders) showed strong relationships with other
teacher language practices related to children’s oral language de- variables of the same type. For example, there were relatively high
velopment, then track the variation of these practices across con- correlations for variables within the Vocabulary (r̅ = 0.37) and
texts. In this review, we use an SFL perspective as an organizing Conceptual Talk (r̅ = 0.36) categories across studies. Vocabulary
framework to examine how a wide variety of individual teacher Talk was a tightly defined category that included definitions of new
language practices supportive of oral language development may words, examples of word use, and word consciousness talk. Simi-
cohere into patterns of use, or registers. This perspective helps us larly, Conceptual Talk included vocabulary talk, talk about the con-
to understand individual teacher language practices not as unre- tent areas such as science, and abstract talk that focused on topics
lated, discrete skills, but as "ways with words" (Heath, 1982) that beyond the here-and-now. Because these types of teacher language
cluster together in characteristic language registers and vary across practices were focused on discussing a particular, readily identifi-
contexts of use. This paper contributes to the field by synthesizing able topic area, they may have been easier to identify and code re-
the vast literature on EC teacher language, distilling a heterogenous liably than more diffuse language support strategies such as those
set of variables into several key strands of teacher language prac- included in the Responsive Talk category (e.g., active listening, re-
tices, and then examining whether those strands further intertwine sponding to child talk, repeating or confirming).
into registers of use. We discuss detailed findings below.
7.2. Registers of use in the EC classroom
7.1. Relationships between types of teacher language practices
Our second research question sought to identify potential clus-
Our first research question examined the relationships among ters of teacher language practices from the 11 macro codes, with
all the different types of EC teacher language practices. Using prior the goal of highlighting registers of use in the EC language environ-
research (Hadley, Barnes, & Hwang, 2021; Rowe & Snow, 2020; ment. The factor analysis yielded a 2-dimensional model of teacher
Justice, Jiang, & Strasser, 2018), and iterative coding of the teacher language in EC classrooms. The first dimension, which we call an
language variables, we developed 11 different macro codes for Emergent Academic Language register (Dickinson et al., 2014) be-
types of teacher language. The patterns of correlations revealed cause it contains many features characteristic of the language of
small to moderate associations between different teacher language schooling, included the Conceptual Talk, Types and Tokens, Com-
types, indicating that the macro codes captured related but unique plex Syntax, Linguistic Features, and Modeling Talk variables. The
aspects of teacher language. We discuss several of these correla- second dimension, which we call the Bridge Language register be-
tions here, focusing on the relationships replicated at least 5 times cause it includes language practices that help children connect
in our sample (k ≥ 5) and those that showed strong positive cor- their home and school language practices, includes Questions and
relations with one another (r̅ ≥ 0.40). One consistent pattern was Elicitations, High-Level Questions, Responsive Talk, Literal Talk, and
the strong relationships between different kinds of complex Lin- Management Talk. Vocabulary Talk evenly loaded on to both di-
guistic Features. For example, the Types/Tokens code and the Lin- mensions.
guistic Features code were highly correlated (k = 8, r̅ = 0.43), in- Each of these registers contains a constellation of lexical, gram-
dicating that teachers who used more linguistically complex talk matical, and discourse-level features that constitute a variety of
also tended to talk more and use a greater variety of word types. talk customized for a specific purpose and with specific partici-
In other words, several markers of more complex speech (use of pants. The Emergent Academic Language register included the dis-
rare vocabulary, more diversity in word types, longer sentences) cussion of abstract ideas (e.g., new vocabulary words, science talk,
appeared to cluster together. Next, teachers’ use of Questions and talk about character motivations during book-reading), and as part
Elicitations was highly related to Responsive Talk (k = 5, r̅ = 0.41), of this conceptual talk, certain linguistic features naturally arise
indicating that teachers who followed up on and responded to – longer sentences, more rare vocabulary, and greater variety in
child talk also tended to ask more questions, suggesting an over- word types. It is important to note here that the Linguistic Fea-
all characteristic pattern of interactive conversation with children. tures category here typically only included linguistic features that
We also examined the correlations between variables within are associated with more academic talk. In other words, linguis-
the same category (e.g., the relationships between different vari- tic features that might serve as markers of other kinds of registers
ables coded as “Complex Syntax” across studies). These relation- (e.g., the use of active voice, the use of contractions) were typ-
ships allow us to see which teacher language types were mea- ically not measured by these studies. Interestingly, the Emergent
sured more reliably across studies and also which were conceptu- Academic Language register also included Modeling Talk, such as
ally more homogeneous within a single category. Variables within teachers modeling the use of new vocabulary or emotion words,
the Complex Syntax category were highly related to one another orally narrating their own or children’s actions, and commenting
(r̅ = 0.63), likely because this variable was typically measured on child talk. Such talk may be an important feature of an emer-
the same way across studies and used objective measures, such gent academic language register in that teachers explicitly demon-
as calculating the mean length of utterances using the Comput- strate and encourage children in the use of the language practices
erized Language Analysis (CLAN) program (MacWhinney, 20 0 0). that are most suited to the setting and activity at hand.
The Complex Syntax category was also conceptually quite nar- The Bridge Language register contains elements that encour-
row, incorporating a very specific dimension of teacher language ages child talk (Questions, High-Level Questions), builds on that
(i.e., the average length of their sentences). The Linguistic Fea- talk (Responsive Talk), and builds gateways into more advanced
tures and Types/Tokens variables showed similarly high relation- talk (Literal and Management talk). We call this register “Bridge
ships between variables within each category, which also reflected Language” because we see one of the central functions of this

197
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

register as “bridging to make connections between the known [e.g., now. However, the vocabulary talk in our sample of studies typ-
children’s individual ways with words] and the new [e.g., school- ically involved explicit instruction of new vocabulary words (e.g.,
valued language practices]” (Rogoff et al., 1993, p. 8). Teacher Ques- Gerde & Powell, 2009; Silverman & Crandell, 2010), a practice that
tions and Elicitations are one key variable in this register, although is in keeping with bridging children into school-valued language
the use of teacher questions can vary widely in their nature, from practices in that such instruction equips children with the vocab-
quick, closed-ended prompts (e.g., “do you think Corduroy felt ulary necessary for accessing complex texts and precisely commu-
sad?”) that elicit only one-word answers from children, to more nicating ideas within content areas. Therefore, vocabulary talk is
demanding, abstract requests to make predictions or infer cause likely a practice that crosses over and connects both registers, in
and effect (e.g., “What do you think Corduroy will do after the that it pairs Emergent Academic Language practices (e.g., the use
toy store closes?”). Both kinds of questions may serve a purpose in of rare words, discussion of abstract ideas) with Bridge Language
the EC classroom, in that they open a space for child talk and for practices (e.g., explicit support for learning and mastering new lan-
teachers to learn about that talk, although closed-ended prompts guage forms).
may be most appropriate for children at an earlier stage of lan- The results of this factor analysis help to support our theo-
guage development. For example, Deshmukh et al. (2019) found retical framework for language learning in EC classrooms, which
that teachers who used closed-ended prompts during large group views the hybridity, or blend between home and school language
shared book-reading were more likely to receive a one-word re- practices in ECs, as beneficial for children’s language learning
sponse from children, but these responses were also nearly always (Hadley, Barnes, & Hwang, 2021a; Authors, 2020). In our previous
accurate. In contrast, teachers who asked more challenging wh- systematic review, we identified 3 primary strands of EC teacher
and why questions were more likely to receive a multi-word re- language practices that were positively related to children’s oral
sponse, but these responses were also more likely to be inaccurate. language development: (1) conceptual talk, (2) interactive talk,
Because children’s responses to high-level questions are longer, and (3) linguistic features. The present meta-analysis and factor
they give teachers a more sustained opportunity to learn about analysis further confirms and builds on this work by identifying
and fine-tune their speech to children’s individual language prac- 2 main registers of talk emerging from the broad literature on
tices (e.g., Justice et al., 2013), which is valuable even if children’s EC teacher language and oral language development, namely the
responses are unconventional. These inaccurate responses in turn Emergent Academic Language and Bridge Language registers. The
give teachers the opportunity to use responsive language practices Emergent Academic Language register is characterized by teacher-
such as scaffolding, recasting, and expanding on children’s utter- led, school-valued forms of talk that focus on abstract topics, use
ances in the moment, providing proximal feedback on the kind of longer sentences, more rare vocabulary. In the Bridge Language
language that is most useful, clear, and helpful in the setting at register, teachers use a variety of interactive, supportive language
hand. A recent study’s findings suggests that interactive talk prac- practices to encourage child talk (e.g., questions and prompts), use
tices, as compared to other kinds of teacher talk, may be espe- talk that meets children where they are (literal talk), and scaffolds
cially helpful for children’s language development. Justice, Jiang, & them into a new language community (responsive talk). Future re-
Strasser, 2018 examined several dimensions of the preschool class- search should identify the characteristic linguistic features associ-
room language environment, including teachers’ interactive talk, ated with Bridge Language to further aid in the identification and
the linguistic features of teacher talk, and global classroom qual- measurement of this register.
ity as measured by the CLASS, and found that only interactive talk Of course, as the present meta-analysis only examines teacher
was significantly associated with growth in children’s vocabulary. language practices, rather than tracing these practices’ relation-
This indicates the importance of teacher language practices that ships with child outcomes, it is possible that one or the other talk
not only provide a model for emergent academic language use, but register may not actually be facilitative of children’s oral language
that also actively engage children in conversations. skills, or that different strands may support different aspects of
Literal talk and management talk also loaded on to the Bridge oral language (e.g., syntax, discourse, vocabulary). Future research
Language dimension. Literal talk is a necessity in the classroom should investigate potential relationships with child outcomes in
in that teachers and children must talk about the here-and-now different settings and with children with varied language resources
of the classroom world to accomplish tasks and build comprehen- and family backgrounds.
sion (e.g., Kleeck et al., 2006), but too often, such talk dominates
in EC classrooms at the expense of more challenging conceptual 7.3. Variation in teacher language practices across places
talk (e.g., Justice et al., 2008). However, conceptualizing literal talk
as part of an overall orientation to bridging children’s home and Next, in keeping with our theoretical perspective on language
school language practices may help teachers to use literal talk in a as contextually situated and shaped by the participants, places,
more fruitful manner, such as using literal talk as a gateway into and purposes for talk, we examined how teacher language var-
school-valued language. For example, Hindman et al. (2012) found ied across classroom activity settings and group sizes. Our inten-
that, for children with lower standardized vocabulary scores, teach- tion was to also examine variation in language use with chil-
ers’ use of more literal, contextualized talk during book-reading dren from different demographic backgrounds, but because of the
was especially beneficial for their vocabulary development, as limited information reported about participants and the relatively
compared to their peers. These children also, however, benefited small number of studies, our findings on participants should be
from teachers’ use of decontextualized talk, indicating that pair- considered only as preliminary. We discuss our results for activity
ing both literal and abstract talk about books may help to facilitate settings and group sizes below.
children’s ability to discuss the more conceptual topics. The provi-
sion of multiple kinds of talk offers multiple entry points into the 7.4. Activity settings
conversation, thereby allowing greater access for children with a
wide variety of language repertoires. A small but growing number of studies have examined mean
The results for the vocabulary talk variable, which loaded on differences in teacher language practices across classroom settings,
to both factors, merits further unpacking. Vocabulary talk is well- such as book-reading, play, or literacy instruction time (Bratsch-
aligned with an Emergent Academic Language register in that it Hines et al., 2019; Cabell et al., 2013; Dickinson et al., 2014; Kook
is a subset of the conceptual talk category and involves the dis- & Greenfield, 2020). The moderator analyses in the present pa-
cussion of ideas (and words) that are removed from the here-and- per take a slightly different approach. Rather than examining mean

198
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

differences, we examine differences in correlations among types of ceptual Talk and High-Level Questions. These results suggest that
teacher language variables across settings. That is, we examined small group time as currently implemented may not positively en-
whether a particular setting influenced various kinds of teacher hance teacher language practices as much as other settings, and
langauge to be more strongly linked with each other than another that certain beneficial language practices, such as requesting lan-
setting. When a pair of teacher language variables have a higher guage from children or repeating their utterances, tend to happen
correlation in one setting vs another, we interpret the setting with more frequently in large group time than in small group (Bratsch-
the higher correlation as enhancing or promoting both kinds of Hines et al., 2019) or other settings (Cabell et al., 2013). There-
talk, as compared to the other setting. fore, the small group space appears to represent a missed op-
Several key patterns emerge from Fig. 4A, which shows a com- portunity for language development, as this participation struc-
parison of teacher language correlations across book reading, liter- ture is one of the few classroom spaces in which teachers and
acy instruction, and play settings (each setting is represented by a children can engage in the kinds of extended, scaffolded conver-
different color of circle). Examining the correlations between Vo- sations that help children to acquire the language of schooling.
cabulary Talk and other kinds of teacher language practices for the For example, previous studies have found that children and teach-
book-reading vs the literacy instruction setting reveals a pattern in ers appear to “mirror” the structure and conceptual complexity of
which 2 key elements of Emergent Academic Language, Vocabulary other’s use of language in small group play settings (Justice et al.,
Talk and Conceptual Talk, both appear to be enhanced more within 2013; Tompkins, Zucker, Justice, & Binici, 2013). However, small
a book-reading setting, as contrasted with the literacy instruction group instruction as it is often implemented typically involves
setting. In the studies in our sample, shared book-reading typi- more skills-based instruction with fewer opportunities for conver-
cally involved a teacher reading out loud, and discussing, a narra- sation (Dickinson et al., 2014).
tive text with children. In contrast, literacy instruction consisted of A possible alternative explanation for the smaller correlations
more skills-based activities such as identifying letters. These find- with small group sizes as opposed to whole group or mixed set-
ings accord with Dickinson et al. (2014), who found evidence of tings is that teachers tend to talk more, and in more character-
an emergent language register during shared book-reading in Head istic and consistent patterns, in larger groups, whereas their talk
Start classrooms, but found that skills-based literacy instruction in- within small group settings may be more variable, better individ-
volved fewer markers of academic speech. ualized to the children in the small groups, and involve more lis-
Fig. 4A also provides a look at gaps in the field in terms tening to child talk. More research is needed to understand the
of settings that have and have not been examined in EC lan- affordances of small vs. larger groups for shaping teacher language
guage research. In our sample, book-reading was far and away the practices, and how these affordances can be further leveraged to
most commonly investigated setting in terms of teacher language support child language development.
use. Indeed, for certain kinds of teacher language practices, book-
reading was the only setting within which these types of talk were 7.6. Limitations
examined. For example, Management Talk and High-Level Ques-
tions have exclusively been examined within a book-reading set- One major limitation of the present review is that the studies
ting, and Literal Talk and Questions and Elicitations have mostly we reviewed chiefly focused on teacher language, with few stud-
been examined only within book-reading. The field’s strong fo- ies reporting details about child language. Many studies were un-
cus on teacher language during the book-reading setting is war- able to report detailed information about child language given the
ranted in that book-reading has been repeatedly found to exert methodological limitations involved with understanding and tran-
a strong influence on children’s oral language development (e.g., scribing child language in the noisy classroom environment. How-
Flack, Field, & Horst, 2018; Hindman et al., 2008). However, book- ever, gaining a full understanding of how teachers can leverage in-
reading occupies only about 5% of the school day for preschool- dividual children’s home language practices is an important goal
ers (Early et al., 2010), and about 10 minutes per school day for for future research.
kindergarteners (Wright, 2011). Understanding the language affor- The present meta-analysis was also limited by the relatively
dances of other settings, such as free choice time, play or recess, small number of studies that reported the information necessary
mealtimes, and art, are important in gaining a more comprehen- to calculate effect sizes. We urge all authors to report complete
sive picture of the classroom language environment. For example, correlation matrices so that statistical summaries of this area of
previous research has indicated the promise of science activities research are possible.
and other content-area instruction for promoting language devel- Additionally, many studies reported only minimal details about
opment (Cabell et al., 2013), and more work in this area is needed child and teacher participants and did not clearly describe the ob-
to more fully explore whether, as some studies suggest, talking servational contexts or report findings by setting or group size. We
about content-area topics drives the use of meta-level talk and argue that these contextual features are essential to understand-
emergent academic language registers by teachers (Barnes & Dick- ing results, given the variation in talk reported across contexts and
inson, 2018). Newer studies have also suggested that settings that with different participants in previous research (Dickinson, Darrow,
are more characteristic of home activities, such as play or meal- & Tinubu, 2008; Hadley & Dickinson, 2019; Pelatti, Piasta, Justice,
times, may also afford unique opportunities for leveraging chil- & O’Connell, A., 2014).
dren’s home language practices (e.g., Barnes et al., 2020; Leyva & While we did not exclude international studies from our search,
Skorb, 2017). the majority of the studies in our final sample took place in the
United States. Therefore, our results are primarily generalizable
7.5. Group sizes only to early childhood classrooms in a U.S. context.

We examined correlations between teacher language types 7.7. Educational implications


across 3 group sizes: whole class, small group, and mixed (i.e.,
studies that aggregated findings across a variety of group sizes). Our findings here have several implications for practice. First,
Fig. 4B shows a big picture view of how the 3 settings compare our theoretical framework and results suggest the potential
to one another. Overall, our pattern of findings revealed smaller or promise of professional development efforts that have more holis-
negative correlations in small group as compared to whole class tic understandings of language use in the classroom. While pre-
or mixed settings for several beneficial practices, such as Con- vious approaches often have focused on using certain beneficial

199
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

practices in isolation (e.g., asking open-ended questions, defin- Barnes, E. M., Grifenhagen, J. F., & Dickinson, D. K. (2020). Mealtimes in Head
ing vocabulary words), a focus on registers asks teachers to think Start pre-k classrooms: Examining language-promoting opportunities in a hy-
brid space. Journal of Child Language, 47(2), 337–357. https://doi.org/10.1017/
about the purpose of their talk, the place or setting in which it is S03050 0 09190 0 0199.
used, and the participants at hand, when planning a language-rich Bassok, D., Latham, S., & Rorem, A. (2016). Is kindergarten the new first grade? AERA
activity. Teachers can then adopt clusters of language practices that Open, 2(1), 2332858415616358. https://doi.org/10.1177/2332858415616358.
Biber, D. (1995). Dimensions of register variation: A cross-linguistic comparison. Cam-
are suited for the task at hand. For example, a teacher might plan bridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511519871.
to communicate ideas about light and shadow through a hands-on Binici, S. (2014). Preschool teachers’ inferential questions during shared reading and
science activity, and therefore plan to use precise, technical lan- their relation to low-income children’s reading comprehension at kindergarten and
first grade [Doctoral dissertation. Columbus, Ohio: The Ohio State University].
guage to define key words, discuss abstract ideas, and ask ques-
Bowne, J. B., Yoshikawa, H., & Snow, C. E. (2017). Relationships of teachers’ language
tions that help children to build their knowledge of new concepts and explicit vocabulary instruction to students’ vocabulary growth in kinder-
(i.e., all practices within an academic language register). Another garten. Reading Research Quarterly, 52(1), 7–29. https://doi.org/10.1002/rrq.151.
Bratsch-Hines, M. E., Burchinal, M., Peisner-Feinberg, E., & Franco, X (2019). Fre-
important implication for practice that can be drawn from this re-
quency of instructional practices in rural prekindergarten classrooms and asso-
view is the idea of ECCs as a hybrid space, where a diverse range ciations with child language and literacy skills. Early Childhood Research Quar-
of home and school language practices can serve as an engine of terly, 47, 74–88. https://doi.org/10.1016/j.ecresq.2018.10.001.
development and a source of growth. In this view, responsive, nur- Bushman, B. J., & Wells, G. L. (2001). Narrative impressions of literature: The avail-
ability bias and the corrective properties of meta-analytic approaches. Person-
turing language practices, such as listening to children and learn- ality and Social Psychology Bulletin, 27(9), 1123–1130. https://doi.org/10.1177/
ing about their own ways with words, are valued and prioritized 0146167201279005.
alongside more traditional academic language learning. Cabell, S. Q., DeCoster, J., LoCasale-Crouch, J., Hamre, B. K., & Pianta, R. C. (2013).
Variation in the effectiveness of instructional interactions across preschool
classroom settings and learning activities. Early Childhood Research Quarterly,
8. Conclusion 28(4), 820–830. https://doi.org/10.1016/j.ecresq.2013.07.007.
Cabell, S. Q., Justice, L. M., McGinty, A. S., DeCoster, J., & Forston, L. D. (2015).
Teacher–child conversations in preschool classrooms: Contributions to chil-
The results from this meta-analysis are a first step in identify- dren’s vocabulary development. Early Childhood Research Quarterly, 30(Part A),
ing patterns of teacher language that may support children’s learn- 80–92. https://doi.org/ 10.1016/j.ecresq.2014.09.004
ing of school-valued language registers. From the settings and lan- Candès, E. J., & Recht, B. (2009). Exact matrix completion via convex optimiza-
tion. Foundations of Computational Mathematics, 9(6), 717–772. https://doi.org/
guage practices studied to date, we found evidence of 2 teacher 10.1007/s10208- 009- 9045- 5.
language registers in EC classrooms, namely Emergent Academic Carter, E. C., Schönbrodt, F. D., Gervais, W. M., & Hilgard, J. (2019). Cor-
Language and Bridge Language registers. We also found some recting for bias in psychology: A comparison of meta-analytic methods.
Advances in Methods and Practices in Psychological Science, 2(2), 115–144.
emerging indicators of teacher language variation across classroom
https://doi.org/10.1177/2515245919847196
activity settings, suggesting that particular settings may facilitate Cazden, C. B. (1988). Classroom discourse: The language of teaching and learning.
the use of certain kinds of talk. However, much more work is Heinemann. https://archive.org/details/classroomdiscour0 0 0 0cazd.
Dahlke, J. A., & Wiernik, B. M. (2019). psychmeta: An R package for psychomet-
needed in this complex line of research, particularly work that at-
ric meta-analysis. Applied Psychological Measurement, 43(5), 415–416. https://doi.
tends to the constellation of contextual factors that shape the con- org/10.1177/0146621618795933.
versations between teachers and children. Dahlke, J. A., & Wiernik, B. M. (2021). psychmeta: Psychometric meta-
analysis toolkit (2.6.0.90 0 0) [R package]. https://github.com/psychmeta/
psychmeta (Original work published 2017)
Author Note Deshmukh, R. S., Zucker, T. A., Tambyraja, S. R., Pentimonti, J. M., Bowles, R. P.,
& Justice, L. M. (2019). Teachers’ use of questions during shared book read-
Erica M. Barnes https://orcid.org/0 0 0 0-0 0 03-0818-9285 ing: Relations to child responses. Early Childhood Research Quarterly, 49, 59–68.
https://doi.org/10.1016/j.ecresq.2019.05.006.
Elizabeth Burke Hadley https://orcid.org/0 0 0 0-0 0 02-2880-6175 Dickinson, D. K. (2011). Teachers’ language practices and academic outcomes of
Brenton M. Wiernik https://orcid.org/0 0 0 0-0 0 01-9560-6336 preschool children. Science, 333(6045), 964–967. https://doi.org/10.1126/science.
Mukhunth Raghavan https://orcid.org/0 0 0 0-0 0 02-4785-9387 1204526.
Dickinson, D. K., Darrow, C. L., & Tinubu, T. A. (2008). Patterns of teacher-child con-
versations in Head Start classrooms: Implications for an empirically grounded
ncited References approach to professional dvelopment. Early Education and Development, 19(3),
396–429. https://doi.org/10.1080/10409280802065403.
Dickinson, D. K., Hofer, K. G., Barnes, E. M., & Grifenhagen, J. F. (2014). Examin-
Binici, 2014, Wasik and Hindman, 2014, Yu and Smith, 2012, ing teachers’ language in Head Start classrooms from a Systemic Linguistics
Zhang, 2014, Dahlke and Wiernik, 2019, Dahlke and Wiernik, 2021, Approach. Early Childhood Research Quarterly, 29(3), 231–244. https://doi.org/10.
Gámez et al., 2017, Hindman et al., 2019, Kassambara, 2019, 1016/j.ecresq.2014.02.006.
Dickinson, D. K., & Porche, M. V. (2011). Relation between language experiences
Lipsky, 2013, Neugebauer et al., 2017, Neuman et al., 2018,
in preschool classrooms and children’s kindergarten and fourth-grade language
Pentimonti et al., 2017, Phillips et al., 2017, La Paro and and reading abilities. Child Development, 82(3), 870–886. https://doi.org/10.1111/
Hamre, 2008, Piasta et al., 2012, Rojas, 2021, Saffran, 2003, j.1467-8624.2011.01576.x.
Dickinson, D. K., & Smith, M. W. (1994). Long-term effects of preschool teachers’
Sun et al., 2020, Studhalter et al., 2021, Turnbull et al., 2009,
book readings on low-income children’s vocabulary and story comprehension.
Uccelli et al., 2013, Reading Research Quarterly, 29(2), 105–122. https://doi.org/10.2307/747807.
Dickinson, D. K., & Tabors, P. O. (2001). Putting the pieces together: The impact
References of preschool on children’s language and literacy development in kindergarten.
In Beginning literacy with language: Young children learning at home and school
Bakbergenuly, I., Hoaglin, D. C., & Kulinskaya, E. (2019). Simulation study of esti- (pp. 257–287). Paul H. Brookes Publishing.
mating between-study variance and overall effect in meta-analysis of standardized Early, D. M., Iruka, I. U., Ritchie, S., Barbarin, O. A., Winn, D.-M. C., Crawford,
mean difference (1903.01362 [stat.ME]). arXiv. http://arxiv.org/abs/1903.01362 . G. M., Frome, P. M., Clifford, R. M., Burchinal, M., Howes, C., Bryant, D. M.,
Barnes, E. M., & Dickinson, D. K. (2017a). The impact of teachers’ commenting & Pianta, R. C. (2010). How do pre-kindergarteners spend their time? Gen-
strategies on children’s vocabulary growth. Exceptionality, 25(3), 186–206. https: der, ethnicity, and income as predictors of experiences in pre-kindergarten
//doi.org/10.1080/09362835.2016.1196447. classrooms. Early Childhood Research Quarterly, 25(2), 177–193. https://doi.org/
Barnes, E. M., & Dickinson, D. K. (2017b). The relationship of Head Start teachers’ 10.1016/j.ecresq.2009.10.003.
academic language use and children’s receptive vocabulary. Early Education and Farrow, J., Wasik, B. A., & Hindman, A. H. (2020). In Exploring the unique contribu-
Development, 28(7), 794–809. https://doi.org/10.1080/10409289.2017.1340069. tions of teachers’ syntax to preschoolers’ and kindergarteners’ vocabulary learning:
Barnes, E. M., & Dickinson, D. K. (2018). Relationships among teachers’ use of mental 51 (pp. 178–190). Early Childhood Research Quarterly.
state verbs and children’s vocabulary growth. Early Education and Development, Flack, Z. M., Field, A. P., & Horst, J. S. (2018). The effects of shared storybook reading
29(3), 307–323. https://doi.org/10.1080/10409289.2018.1440844. on word learning: A meta-analysis. Developmental Psychology, 54(7), 1334–1346.
Barnes, E. M., Dickinson, D. K., & Grifenhagen, J. F. (2017). The role of teachers’ https://doi.org/10.1037/dev0 0 0 0512.
comments during book reading in children’s vocabulary growth. The Journal Friedman-Krauss, A., Barnett, W. S., Garver, K. A., Hodges, K. S., Weisenfeld, G. G., &
of Educational Research, 110(5), 515–527. https://doi.org/10.1080/00220671.2015. Gardiner, B. A. (2020). The State of Preschool 2020. https://nieer.org/wp-content/
1134422. uploads/2021/04/YB2020_Full_Report.pdf.

200
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Gámez, P. B. (2015). Classroom-based English exposure and English Language Learn- els. [Manuscript submitted for publication] https://raw.githubusercontent.com/
ers’ expressive language skills. Early Childhood Research Quarterly, 31, 135–146. openjournals/joss-papers/joss.03393/joss.03393/10.21105.joss.03393.pdf.
https://doi.org/10.1016/j.ecresq.2015.01.007. MacWhinney, B. (20 0 0). The CHILDES Project: Tools for analyzing talk (3rd ed.).
Gámez, P. B., Neugebauer, S. R., Coyne, M. D., McCoach, D. B., & Ware, S. (2017). Lawrence Erlbaum Associates 10.4324/9781315805672.
Linguistic and social cues for vocabulary learning in Dual Language Learners and Mascareño, M., Snow, C. E., Deunk, M. I., & Bosker, R. J. (2016). Language complexity
their English-only peers. Early Childhood Research Quarterly, 40, 25–37. https: during read-alouds and kindergartners’ vocabulary and symbolic understanding.
//doi.org/10.1016/j.ecresq.2017.01.003. Journal of Applied Developmental Psychology, 44, 39–51. https://doi.org/10.1016/j.
Gerde, H. K., & Powell, D. R. (2009). Teacher education, book-reading practices, and appdev.2016.02.001.
children’s language growth across one year of head start. Early Education and Mathur, M. B., & VanderWeele, T. J. (2021). Estimating publication bias in meta-
Development, 20(2), 211–237. https://doi.org/10.1080/10409280802595417. analyses of peer-reviewed studies: A meta-meta-analysis across disciplines and
Girolametto, L., & Weitzman, E. (2002). Responsiveness of child care providers in journal tiers. Research Synthesis Methods, 12(2), 176–191. https://doi.org/10.1002/
interactions with toddlers and preschoolers [Article]. Language, Speech & Hear- jrsm.1464.
ing Services in Schools, 33(4), 268–290. https://doi.org/10.1044/0161-1461(2002/ Michaels, S. (1981). Sharing time”: Children’s narrative styles and differential ac-
022). cess to literacy. Language in Society, 10(3), 423–442. https://doi.org/10.1017/
Gort, M., & Sembiante, S. F. (2015). Navigating hybridized language learning spaces S0047404500008861.
through translanguaging pedagogy: Dual language preschool teachers’ languag- Miller, P. J., Cho, G. E., & Bracey, J. R. (2005). Working-class children’s experience
ing practices in support of emergent bilingual children’s performance of aca- through the prism of personal storytelling. Human Development, 48(3), 115–135.
demic discourse. International Multilingual Research Journal, 9(1), 7–25. https: Mol, S. E., Bus, A. G., & de Jong, M. T. (2009). Interactive book reading in
//doi.org/10.1080/19313152.2014.981775. early education: A tool to stimulate print knowledge as well as oral lan-
Gutiérrez, K. D., Baquedano-López, P., & Tejeda, C. (1999). Rethinking diversity: Hy- guage. Review of Educational Research, 79(2), 979–1007. https://doi.org/10.3102/
bridity and hybrid language practices in the third space. Mind, Culture, and Ac- 0034654309332561.
tivity, 6(4), 286–303. https://doi.org/10.1080/10749039909524733. Müller, K. (2020). here: A simpler way to find your files (1.0.1) [R package]. https:
Hadley, E. B., & Dickinson, D. K. (2019). Cues for word-learning during shared book- //CRAN.R-project.org/package=here.
reading and guided play in preschool. Journal of Child Language, 46(6), 1202– Nagy, W. (2007). Metalinguistic awareness and the vocabulary-comprehension
1227. https://doi.org/10.1017/S03050 0 09190 0 0552. connection. In Vocabulary acquisition: Implications for reading comprehension
Halliday, M. A. K. (1978). Language as social semiotic: The social interpretation of (pp. 52–77). Guilford Press.
language and meaning. Edward Arnold. https://1lib.us/book/3046653/a369a8. Nagy, W., & Townsend, D. (2012). Words as tools: Learning academic vocabulary as
Halliday, M. A. K. (1993). Towards a language-based theory of learning. Linguistics language acquisition. Reading Research Quarterly, 47(1), 91–108. https://doi.org/
and Education, 5(2), 93–116. https://doi.org/10.1016/0898-5898(93)90026-7. 10.1002/rrq.011.
Halliday, M. A. K., & Hasan, R. (1989). Language, context, and text: Aspects of language Neugebauer, S., Coyne, M., McCoach, B., & Ware, S. (2017). Teaching beyond the
in a social-semiotic perspective. Oxford: Oxford University Press. https://doi.org/ intervention: the contribution of teacher language extensions to vocabulary
10.1044/1092-4388(2002/074). learning in urban kindergarten classrooms. Reading and Writing, 30(3), 543–567.
Heath, S. B. (1982). What no bedtime story means: Narrative skills at https://doi.org/10.1007/s11145- 016- 9689- x.
home and school. Language in Society, 11(1), 49–76. https://doi.org/10.1017/ Neugebauer, S. R., Gámez, P. B., Coyne, M. D., McCoach, D. B., Cólon, I. T., &
S0047404500009039. Ware, S. (2017). Promoting word consciousness to close the vocabulary gap in
Hindman, A. H., Connor, C. M., Jewkes, A. M., & Morrison, F. J. (2008). Untangling young word learners. Elementary School Journal, 118(1), 28–54. https://doi.org/
the effects of shared book reading: Multiple factors and their associations with 10.1086/692986.
preschool literacy outcomes. Early Childhood Research Quarterly, 23(3), 330–350. Neuman, S. B., Kaefer, T., & Pinkham, A. M (2018). A double dose of disadvantage:
https://doi.org/10.1016/j.ecresq.2008.01.005. Language experiences for low-income children in home and school. Journal of
Hindman, A. H., Skibbe, L., & Foster, T. (2014). Exploring the variety of parental talk Educational Psychology, 110(1), 102–118. https://doi.org/10.1037/edu0 0 0 0201.
during shared book reading and its contributions to preschool language and Pelatti, C. Y., Piasta, S. B., Justice, L. M., & O’Connell, A. (2014). Language- and
literacy: evidence from the Early Childhood Longitudinal Study-Birth Cohort. literacy-learning opportunities in early childhood classrooms: Children’s typical
Reading & Writing, 27(2), 287-313. https://doi.org/ 10.1007/s11145-013-9445-4. experiences and within-classroom variability. Early Childhood Research Quarterly,
Hindman, A. H., Wasik, B. A., & Bradley, D. E. (2019). How classroom conversations 29(4), 445–456. https://doi.org/10.1016/j.ecresq.2014.05.004.
unfold: Exploring teacher–child exchanges during shared book reading. Early Pentimonti, J. M., Justice, L. M., Yeomans-Maldonado, G., McGinty, A. S., Slocum, L.,
Education and Development, 30(4), 478–495. https://doi.org/10.1080/10409289. & O’Connell, A (2017). Teachers’ use of high-and low-support scaffolding strate-
2018.1556009. gies to differentiate language instruction in high-risk/economically disadvan-
Hindman, A. H., Wasik, B. A., & Erhart, A. C. (2012). Shared book reading and Head taged settings. Journal of Early Intervention, 39(2), 125–146. https://doi.org/10.
Start preschoolers’ vocabulary learning: The role of book-related discussion and 1177/1053815117700865.
curricular connections. Early Education & Development, 23(4), 451–474. https:// Phillips, B. M., Zhao, Y., & Weekley, M. J. (2017). Teacher language in the preschool
doi.org/10.1080/10409289.2010.537250. classroom: Initial validation of a classroom environment observation tool.
Horn, J. L. (1965). A rationale and test for the numbers of factors in factor analysis. Early Education and Development, 1–19. https://doi.org/10.1080/10409289.2017.
Psychometrika, 30(2), 179–185. https://doi.org/10.1007/bf02289447. 1408371.
Johnston, J. R., & Wong, M.-Y. A. (2002). Cultural Differences in Beliefs and Practices Pianta, R., La Paro, K. M., & Hamre, B. (2008). Classroom assessment scoring system
Concerning Talk to Children. Journal of Speech, Language, and Hearing Research, (CLASS) manual. Pre-K. Paul H. Brookes Publishing.
45(5), 916–926. Piasta, S. B., Justice, L. M., Cabell, S. Q., Wiggins, A. K., Turnbull, K. P., & Curen-
Justice, L. M., Jiang, H., & Strasser, K. (2018). Linguistic environment of preschool ton, S. M. (2012). Impact of professional development on preschool teachers’
classrooms: What dimensions support children’s language growth? Early Child- conversational responsivity and children’s linguistic productivity and complex-
hood Research Quarterly, 42, 79–92. https://doi.org/10.1016/j.ecresq.2017.09.003. ity. Early Childhood Research Quarterly, 27(3), 387–400. https://doi.org/10.1016/j.
Justice, L. M., Mashburn, A. J., Hamre, B. K., & Pianta, R. C. (2008). Quality of lan- ecresq.2012.01.001.
guage and literacy instruction in preschool classrooms serving at-risk pupils. R Core Team. (2021). R: A language and environment for statistical computing
Early Childhood Research Quarterly, 23(1), 51–68. https://doi.org/10.1016/j.ecresq. (4.1.0) [Computer software]. R Foundation for Statistical Computing http://www.
20 07.09.0 04. r-project.org/.
Justice, L. M., McGinty, A. S., Zucker, T., Cabell, S. Q., & Piasta, S. B. (2013). Bi- Revelle, W. (2009). An introduction to psychometric theory with applications in R.
directional dynamics underlie the complexity of talk in teacher–child play- http://www.personality-project.org/r/book.
based conversations in classrooms serving at-risk pupils. Early Childhood Re- Revelle, W. (2021). psych: Procedures for psychological, psychometric, and personal-
search Quarterly, 28(3), 496–508. https://doi.org/10.1016/j.ecresq.2013.02.005. ity research (2.1.6) [R package]. Northwestern University https://CRAN.R-project.
Kassambara, A. (2019). ggcorrplot: Visualization of a correlation matrix using “ggplot2” org/package=psych.
(0.1.3) [R package]. https://CRAN.R-project.org/package=ggcorrplot. Rogoff, B., Mistry, J., Göncü, A., Mosier, C., Chavajay, P., & Heath, S. B. (2002). Guided
Knapp, G., & Hartung, J. (2003). Improved tests for a random effects meta-regression participation in cultural activity by toddlers and caregivers p. i-179. Monographs
with a single covariate. Statistics in Medicine, 22(17), 2693–2710. https://doi.org/ of the Society for Research in Child development.
10.1002/sim.1482. Rojas, N. M. (2021). Patterns of preschool teachers’ use of discourse strategies with
Kook, J. F., & Greenfield, D. B. (2020). Examining variation in the quality of in- individual Spanish-speaking dual language learners. Journal of Educational Psy-
structional interaction across teacher-directed activities in head start class- chology. https://doi.org/10.1037/edu0 0 0 0481.
rooms. Journal of Early Childhood Research, 19(2), 128–144. https://doi.org/10. Rowe, M. L. (2012). A longitudinal investigation of the role of quantity and quality
1177/2F1476718X20942956. of child-directed speech in vocabulary development. Child Development, 83(5),
Leyva, D., & Skorb, L. (2017). Food for thought: Family food routines and literacy in 1762–1774. https://doi.org/10.1111/j.1467-8624.2012.01805.x.
Latino kindergarteners. Journal of Applied Developmental Psychology, 52, 80–90. Rowe, M. L., & Snow, C. E. (2020). Analyzing input quality along three dimensions:
https://doi.org/10.1016/j.appdev.2017.07.001. Interactive, linguistic, and conceptual. Journal of Child Language, 47(1), 5–21.
Lipsky, M. G. (2013). Head Start teachers’ vocabulary instruction and language com- https://doi.org/10.1017/S03050 0 09190 0 0655.
plexity during storybook reading: Predicting vocabulary outcomes of students Saffran, J. R. (2003). Statistical language learning: Mechanisms and constraints. Cur-
in linguistically diverse classrooms. Early Education & Development, 24(5), 640– rent Directions in Psychological Science, 12(4), 110–114. https://doi.org/10.1111/
667. https://doi.org/10.1080/10409289.2012.721112. 1467-8721.01243.
Lüdecke, D., Patil, I., Ben-Shachar, M. S., Wiernik, B. M., Waggoner, P., & Schleppegrell, M. J. (2004). The language of schooling: A functional linguistics perspec-
Makowski, D. (2021). see: An R package for visualizing statistical mod- tive. Routledge. https://doi.org/10.4324/9781410610317.

201
E. Burke Hadley, E.M. Barnes, B.M. Wiernik et al. Early Childhood Research Quarterly 59 (2022) 186–202

Schmidt, F. L., & Hunter, J. E. (2015). Methods of meta-analysis: Correcting error and van Kleeck, A. (2014). Distinguishing between casual talk and academic talk begin-
bias in research findings (3rd ed.). Sage 10.4135/9781483398105.. ning in the preschool years: An important consideration for speech-language
Scollon, R., & Scollon, S. (1981). Narrative, literacy, and face in interethnic communi- pathologists. American Journal of Speech-Language Pathology, 23(4), 724–741.
cation. Norwood, NJ: Ablex. https://doi.org/10.1044/2014_AJSLP-14-0032.
Silverman, R., & Crandell, J. D. (2010). Vocabulary practices in prekindergarten van Kleeck, A., Vander Woude, J., & Hammett, L. (2006). Fostering literal and infer-
and kindergarten classrooms. Reading Research Quarterly, 45(3), 318–340. ential skills in head start preschoolers with language impairment using scripted
https://doi.org/10.1598/RRQ.45.3.3. book-sharing discussions. American Journal of Speech-Language Pathology, 15, 85–
Smith, M., Dickinson, D., Sangeorge, A., & Anastasopoulos, L. (2002). Early literacy 95. https://doi.org/10.1044/1058-0360(20 06/0 09).
and language classroom observation scale (ELLCO). Paul H. Brookes Publishing. Vernon-Feagans, L., Bratsch-Hines, M., Reynolds, E., & Willoughby, M. (2019). How
Sperry, D. E., Sperry, L. L., & Miller, P. J. (2019). Reexamining the Verbal Environ- Early Maternal Language Input Varies by Race and Education and Predicts Later
ments of Children From Different Socioeconomic Backgrounds. Child Develop- Child Language. Child Development, 0(0). https://doi.org/10.1111/cdev.13281.
ment, 90(4), 1303–1318. https://doi.org/10.1111/cdev.13072. Viechtbauer, W. (2005). Bias and efficiency of meta-analytic variance estimators in
Snow, C. E., Lawrence, J. F., & White, C. (2009). Generating knowledge of academic the random-effects model. Journal of Educational and Behavioral Statistics, 30(3),
language among urban middle school students. Journal of Research on Educa- 261–293. https://doi.org/10.3102/10769986030 0 03261.
tional Effectiveness, 2(4), 325–344. https://doi.org/10.1080/19345740903167042. Viechtbauer, W. (2007). Confidence intervals for the amount of heterogeneity in
Snow, C. E., & Uccelli, P. (2009). The challenge of academic language. In D. R. Olson, meta-analysis. Statistics in Medicine, 26(1), 37–52. https://doi.org/10.1002/sim.
& N. Torrance (Eds.), The Cambridge handbook of literacy (pp. 112–133). Cam- 2514.
bridge: Cambridge University Press. https://doi.org/10.1017/CBO9780511609664. Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package.
008. Journal of Statistical Software, 36(3). https://doi.org/10.18637/jss.v036.i03.
Sun, H., Toh, W., & Steinkrauss, R (2020). Instructional strategies and lin- Viechtbauer, W. (2021). metafor: Meta-analysis package for R (3.1-3) [R package].
guistic features of kindergarten teachers’ shared book reading: The case https://github.com/wviechtb/metafor
of Singapore. Applied Psycholinguistics, 41(2), 427–456. https://doi.org/10.1017/ Viswesvaran, C., & Ones, D. S. (1995). Theory testing: Combining psychometric
S0142716420 0 0 0 053. meta-analysis and structural equations modeling. Personnel Psychology, 48(4),
Studhalter, U. T., Leuchter, M., Tettenborn, A., Elmer, A., Edelsbrunner, P. A., & Saal- 865–885. https://doi.org/10.1111/j.1744-6570.1995.tb01784.x.
bach, H. (2021). Early science learning: The effects of teacher talk. Learning and Vygotsky, L. S. (1978). Mind in Society Development of Higher Psychological Processes.
Instruction, 71. https://doi.org/10.1016/j.learninstruc.2020.101371. Harvard University Press. https://doi.org/10.2307/j.ctvjf9vz4.
Tompkins, V., Zucker, T. A., Justice, L. M., & Binici, S. (2013). Inferential talk during Wasik, B. A., & Hindman, A. H. (2014). Understanding the active ingredients in an
teacher–child interactions in small-group play. Early Childhood Research Quar- effective preschool vocabulary intervention: An exploratory study of teacher
terly, 28(2), 424–436. https://doi.org/10.1016/j.ecresq.2012.11.001. and child talk during book reading. Early Education and Development, 1–22.
Toub, T. S., Hassinger-Das, B., Nesbitt, K. T., Ilgaz, H., Weisberg, D. S., Hirsh-Pasek, K., https://doi.org/10.1080/10409289.2014.896064.
Golinkoff, R. M., Nicolopoulou, A., & Dickinson, D. K. (2018). The language of Weizman, Z. O., & Snow, C. E. (2001). Lexical input as related to children’s vocab-
play: Developing preschool vocabulary through play following shared book- ulary acquisition: Effects of sophisticated exposure and support for meaning.
reading. Early Childhood Research Quarterly, 45, 1–17. https://doi.org/10.1016/j. Developmental Psychology, 37(2), 265–279. https://doi.org/10.1037/0012-1649.37.
ecresq.2018.01.010. 2.265.
Turnbull, K. P., Anthony, A. B., Justice, L., & Bowles, R. (2009). Preschoolers’ exposure Wickham, H. (2016). ggplot2: Elegant graphics for data analysis (2nd ed.). Springer
to language stimulation in classrooms serving at-risk children: The contribution International Publishing. https://doi.org/10.1007/978- 3- 319- 24277- 4.
of group size and activity context. Early Education & Development, 20(1), 53–79. Wickham, H., & Bryan, J. (2019). readxl: Read Excel files (1.3.1) [R package]. https:
https://doi.org/10.1080/10409280802206601. //readxl.tidyverse.org.
Uccelli, P., Demir-Lira, Ö. E., Rowe, M. L., Levine, S., & Goldin-Meadow, S. (2019). Wickham, H., François, R., Henry, L., & Müller, K. (2021). dplyr: A grammar of data
Children’s early decontextualized talk predicts academic language proficiency manipulation (1.0.7) [R package]. https://dplyr.tidyverse.org/.
in midadolescence. Child Development, 90(5), 1650–1663. https://doi.org/10.1111/ Wickham, H., & Hester, J. (2020). readr: Read rectangular text data (1.4.0) [R package].
cdev.13034. https://readr.tidyverse.org/.
Uccelli, P., Dobbs, C. L., & Scott, J. (2013). Mastering academic language: Organization Wilke, C. O. (2020). ggtext: Improved text rendering support for “ggplot2” (0.1.1) [R
and stance in the persuasive writing of high school students. Written Communi- package]. https://CRAN.R-project.org/package=ggtext.
cation, 30(1), 36–62. https://doi.org/10.1177/0741088312469013. Wright, T. S. (2011). What classroom observations reveal about oral vocabulary in-
Uccelli, P., Galloway, E. P., Barr, C. D., Meneses, A., & Dobbs, C. L. (2015). Beyond struction in kindergarten [Doctoral dissertation]. University of Michigan http:
vocabulary: Exploring cross-disciplinary academic-language proficiency and its //hdl.handle.net/2027.42/84619.
association with reading comprehension. Reading Research Quarterly, 50(3), 337– You, K. (2020). filling: Matrix completion, imputation, and inpainting methods (0.2.2)
356. https://doi.org/10.1002/rrq.104. [R package]. https://CRAN.R-project.org/package=filling.
Hadley, E. B., Barnes, E. M., & Hwang, H. J. (2021). A systematic review of teacher Yu, C., & Smith, L. B. (2012). Modeling cross-situational word–referent learn-
language practices and child oral language outcomes in early childhood class- ing: Prior questions. Psychological Review, 119(1), 21–39. https://doi.org/10.1037/
rooms. Under review. a0026182.
Uccelli, P., Phillips Galloway, E., Aguilar, G., & Allen, M. (2020). Amplifying and af- Zhang, C. (2014). Understanding the relations between dimensions of literacy teaching
firming students’ voices through CALS-informed instruction. Theory Into Practice, and preschool children’s literacy skills. West Lafayette, Indiana: [Doctoral disser-
59(1), 75–88. https://doi.org/ 10.1080/00405841.2019.1665413 tation, Purdue University].
van Kleeck, A. (2007). Home talk and school talk: Helping teachers recognize cul-
tural mismatch. The ASHA Leader, 12(13), 23–24.

202

You might also like