Examining the Acquisition of

Vocabulary Knowledge Depth

Among Preschool Students

Elizabeth B. Hadley ABSTR ACT

Well-­developed lexical representations are important for reading comprehen-
David K. Dickinson sion, but there have been no prior attempts to track growth in the depth of
Vanderbilt University, Nashville, knowledge of particular words. This article examines increases in depth of
Tennessee, USA vocabulary knowledge in 4–5-­year-­old preschool students (n = 240) who par-
ticipated in a vocabulary intervention that taught words through book read-
ing and book-­related play. At pretest and posttest, students defined words
Kathy Hirsh-Pasek verbally and by using gesture. Responses were coded for type of semantic
Temple University, Philadelphia, information given. There were significant increases in depth of knowledge
Pennsylvania, USA for all word types. Concrete nouns were learned significantly better than
all other word types, and verbs were learned significantly better than ab-
stract nouns and adjectives. Analysis of semantic content provided nuanced
Roberta Michnick Golinkoff information about word learning across word types. Synonyms and contextual
University of Delaware, Newark, USA information were learned well for all word types, whereas functional infor-
mation was learned best for concrete nouns. These results suggest that ease
of word learning may not be influenced solely by perceptual accessibility of
Kimberly T. Nesbitt words but also by the kind of instructional information that can be provided
University of New Hampshire, Durham, for different word types.
USA

A
number of studies have established the important connection
between reading comprehension and vocabulary knowledge
(Cain, Oakhill, & Bryant, 2004; Quinn, Wagner, Petscher, &
Lopez, 2015; Vellutino, Tunmer, Jaccard, & Chen, 2007). However,
there is a growing awareness that vocabulary knowledge is a complex
construct that cannot be understood solely in terms of breadth, or
number of words known (Christ, 2011; Schoonen & Verhallen, 2008).
Vocabulary breadth is a descriptor of the overall number of entries in
a learner’s lexicon, each of which may be known to a greater or lesser
extent. Vocabulary depth, a related but distinct aspect of word knowl-
edge, is a descriptor of how well the individual entries in one’s lexicon
are known (Anderson & Freebody, 1981). In other words, depth can be
defined as a learner’s richness of knowledge about individual words
and has also been shown to contribute to students’ ability to under-
stand what they read (Ouellette, 2006). However, depth has been less
frequently explored than breadth in the literature, with many vocabu-
lary intervention studies focusing on number of words learned, with-
out asking how much and what kind of knowledge students have
gained about individual words or whether this knowledge is of suffi-
ciently high quality to impact reading comprehension.
Reading Research Quarterly, 51(2)
pp. 181–198 | doi:10.1002/rrq.130
Although the concept of depth itself has several different dimen-
© 2015 International Literacy Association. sions, this article focuses on two key aspects: richness of semantic

181
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
representation of words and knowledge of use in typical then during a guided play activity, may also facilitate
contexts (Nation & Snowling, 1998; Ordóñez, Carlo, deeper word knowledge than learning a word in a single
Snow, & McLaughlin, 2002). We respond to the need for context (Bolger, Balass, Landen, & Perfetti, 2008).
more detailed information about what kinds of instruc- Henriksen (1999) described the process of gaining
tion can help foster depth. We present a nuanced deep word knowledge as network building: discovering
­portrait of preschoolers’ acquisition of deep word learn- links between the word in question and other related
ing, examining which kinds of semantic information terms. Henriksen gave an example of network building
children learned from instruction for words ranging in for the word hot: A child might learn antonyms for this
perceptual accessibility. Our data are drawn from a word (e.g., cold), synonyms (e.g., warm), words that vary
­vocabulary intervention designed to evaluate children’s in gradation (e.g., scalding, tepid), and multiple typical
word learning from shared book reading paired with contexts for the use of the word.
play sessions with varying levels of adult support
(Dickinson et al., 2013). This multiphase intervention
has been shown to have strong effects on children’s Deep Word Knowledge
depth and breadth of word knowledge (Dickinson et al.,
2013). Analyses reported in the present study focus on and Reading Comprehension
children’s depth of word learning, looking at results by Networks of knowledge associated with words have
both word type and semantic information. been identified as a key factor in the relationship be-
tween vocabulary knowledge and reading comprehen-
sion. Anderson and Freebody (1981) posited that
vocabulary and reading comprehension are related be-
Acquiring Deep Word Knowledge cause vocabulary serves as an indication of conceptual
The initial process of learning a new word has often knowledge. According to this theory, a person can un-
been described as fast mapping (Carey, 1978): the quick derstand what is read not only because he or she knows
learning of a few aspects of a word after only a few inci- individual word meanings but also because he or she
dental exposures. Fast-­ mapped information includes has built extensive networks of conceptual knowledge
the association between an object and a word label, lim- from which to draw on, of which vocabulary is the tip of
ited information about the context in which the new the iceberg. To improve reading comprehension, there-
word is encountered, and the ability to produce some of fore, vocabulary instruction must build deep, conceptu-
the phonemes in the word label (Dollaghan, 1985). ally rich knowledge. Neuman and Celano (2006)
Children with a fast-­mapped, limited understanding of suggested that once children begin to acquire this con-
a word may identify a picture of the word on a receptive ceptually rich knowledge, they become able to acquire
vocabulary measure but lack a deeper understanding of more knowledge at faster rates and that, conversely,
the word’s nuances and uses in multiple contexts. children lacking in conceptually rich knowledge fall
Therefore, although children may, in a superficial sense, further behind their peers.
know the word, their semantic knowledge may not be Perfetti’s (2007) lexical quality hypothesis, a theo-
extensive enough for them to use the word in real-­world retical model describing the reading process, similarly
settings or draw on it when comprehending text. highlights the centrality of networks of word knowl-
A deep understanding of word meanings is acquired edge in reading comprehension. The hypothesis views
gradually over time (Bion, Borovsky, & Fernald, 2013; comprehension as dependent on the ability to retrieve
Yu & Smith, 2012). Bloom (2000) described the rate of word identities, which in turn relies on the lexical qual-
children’s word learning not, as is often cited, as learn- ity of a word, or how much knowledge a reader has
ing 10 new words a day but as “learning one-­hundredth about the form and meaning of a particular word, as
of each of a thousand different words” (p. 25). He well as how closely these elements are connected to one
pointed to research showing that common verbs such as another. As a learner has more experience with a word
pour and fill are not fully understood even by children in a variety of contexts (Bolger et al., 2008), its phono-
as old as 7 (Gropen, Pinker, Hollander, & Goldberg, logical representation becomes more stable, more gram-
1991). Research by Medina, Snedeker, Trueswell, and matical classes and inflections of the word are learned,
Gleitman (2011), however, suggested that if the context and the meaning becomes incrementally more precise
in which a word is learned is highly informative, such as and less bound to context. High-­quality representa-
in a book-­reading session in which the story and pic- tions, or semantic networks in which elements of form
tures illustrate the word’s meaning, children may be and meaning are tightly connected to one another, can
able to gain a great deal of knowledge by encountering be quickly retrieved, whereas low-­quality representa-
the word just once. Encountering a word in multiple, tions threaten a reader’s retrieval speed and ability to
varied contexts, such as in a book-­reading session and comprehend a passage.

182 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
The lexical quality hypothesis provides a theoretical elementary students. Given the fact that early language
framework for understanding the relationship between ability is highly predictive of later language competence
high-­quality word knowledge and reading comprehen- (Dickinson & McCabe, 2001; Storch & Whitehurst,
sion, which has been demonstrated in several studies. A 2002), the lexical representations that children build
study that measured depth by evaluating fourth-­grade early on are likely to be of key importance once they
students’ ability to verbally define words showed that begin reading. Efforts to foster young children’s vocab-
depth predicted reading comprehension beyond the ulary learning, therefore, should focus not only on add-
­association explained by measures of breadth (Ouellette, ing new words to their lexicons but also on building
2006). Proctor, Silverman, Harring, and Montecillo rich, high-­quality representations of words. These two
(2012) corroborated these findings: Depth was a signifi- instructional goals are likely complementary: As learn-
cant predictor of reading comprehension for students in ers add new words to their lexicons, their networks of
grades 2–4, even after controlling for decoding and word knowledge become more refined and precise for
vocabulary breadth. In a longitudinal study, Roth,
­ distinguishing new entries from old ones (Carey, 1978),
Speece, and Cooper (2002) reported a strong correlation thereby increasing depth, and when learners gain rich
(r = .70) between kindergartners’ abilities to give oral knowledge about a number of aspects of a word, they
definitions (a measure of depth) and their second-­grade likely learn other, related words along the way, thereby
reading comprehension. Finally, the National Early increasing breadth.
Literacy Panel (2008) found that definitional vocabulary
(a depth of knowledge measure) was more predictive of
later decoding and reading comprehension than breadth. Semantic Networks
Studies with struggling readers at the elementary
level have further corroborated the importance of deep by Word Type
word knowledge for reading comprehension. Nation and Depth of word knowledge can be conceptualized as a
Snowling (1998) found that elementary school–age stu- rich network of semantic associations around a word
dents who had difficulties with reading comprehension that support semantic processing and reading compre-
scored the same as their peers on measures assessing hension. The content of these networks, however, is
phonological skills but did less well on tests that mea- thought to vary by form class (Miller & Fellbaum, 1991),
sured semantic abilities. These results indicate that stu- which may also have consequences for how well differ-
dents who struggled with reading comprehension did so ent words are learned. A number of studies have
not because of weak phonological skills but because they ­reported that verbs are more difficult to learn than
had limited semantic knowledge, which led to slower nouns (see Gentner, 1982; Imai et al., 2008), even for
­semantic processing and poor comprehension. adults (Gillette, Gleitman, Gleitman, & Lederer, 1999).
Landi and Perfetti (2007) suggested that this Maguire, Hirsh-­Pasek, and Golinkoff (2006) suggested
weakness in semantic knowledge is due to a lack of that all words lie on a continuum of concepts and that
relevant experience with words, such as repeated words are easier or more difficult to learn based not on
exposures to words’ phonological and semantic features. their form class but on how perceptually accessible they
Once a pattern of inadequate exposure to words is are, as determined by the factors of shape, individua-
established, it can have long-­reaching reciprocal effects: tion, concreteness, and imageability.
A paucity of high-­quality experiences with words leads Shape can be understood as the extent to which an
to weak semantic representations, which then leads to object or action has a reliable outline or contour (e.g., a
poor reading comprehension (Landi & Perfetti, 2007). cup has a more consistent shape than a person dancing,
As early as first grade, students who struggle with which has a more consistent shape than someone
reading have lower self-­concepts about reading and are thinking; Maguire et al., 2006). Individuation refers to
rated by teachers as more likely to be off task during the ease with which an item can be distinguished from
class reading activities and less likely to read others in a scene (Gentner & Boroditsky, 2001).
independently on their own (Morgan, Fuchs, Compton, Concreteness refers to whether something is a tangible
Cordray, & Fuchs, 2008). This lack of sufficient exposure object (Paivio, Yuille, & Madigan, 1968), and
to print then limits students’ ability to build high-­quality ­imageability to how readily one can produce a mental
representations of words, which continues the cycle image for that word (Maguire et al., 2006). These four
(Landi & Perfetti, 2007). Helping young students build ­elements, considered together, characterize a word’s
high-­quality representations of words through rich ­perceptual accessibility.
instruction and repeated exposures is therefore essential Because verbs tend to lie on the less perceptually
in providing a strong foundation for reading success. ­accessible end of this continuum, they are generally
These studies have shown the importance of deep more difficult to learn, although verbs such as walking
word knowledge in reading comprehension for that easily produce a mental image and have a reliable,

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 183
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
consistent shape are easier to learn than more abstract i­dentify superordinates and subordinates (Verhallen &
verbs such as thinking. Nouns also fall at various points Schoonen, 1993). Another type of hierarchical relation-
along the continuum. For example, the noun justice ship among concrete nouns is that of part–whole
may be quite difficult to learn, as it is highly abstract relations (Henriksen, 1999; Verhallen & Schoonen,
­
and difficult to form a mental image for, whereas a per- 1993). For a word such as fish, various characteristic
ceptually accessible, easily individuated object such as component parts may be included in a child’s semantic
sword may be quite simple to learn. Although verbs vary network, such as scales, fins, gills, and tail.
in perceptual accessibility, in the present study, eight of
the nine verbs tested fall on the more perceptually Perceptual Qualities
­accessible end of the continuum, so we analyzed them For concrete nouns, the object’s perceptual qualities
together (see Appendix A for more specific information constitute additional nodes in the semantic network.
about each word). The nouns tested vary in their per- Although gleaning perceptual information about an
ceptual accessibility, so we divided them into “abstract” object is often thought of as only a first step toward
and “concrete” categories (the abstract nouns are also gaining a deeper knowledge of word meaning
less perceptually accessible in terms of shape, individu- (Hollich, Hirsh-­Pasek, & Golinkoff, 2000), deciding
ation, and imageability, but we use the terms concrete which perceptual qualities are characteristic of par-
nouns and abstract nouns to reflect the common usage ticular nouns is a skill that reveals deeper conceptual
of these terms in the literature). We use the term word knowledge (Booth & Waxman, 2002). Perceptual in-
type rather than form class to reflect this division in formation about objects (e.g., cats have fur, armor is
nouns. made out of metal, gold is yellow) provides important
The idea that words fall along a continuum from fodder for the sophisticated process of categorizing
less to more perceptually accessible has important con- what type of object or material something is and how
sequences for theories of vocabulary depth. It indicates it can be differentiated from other similar objects or
that the types of semantic information available for materials.
words along the continuum will be qualitatively differ-
ent, so a learner’s semantic network for a concrete noun
will look different from his or her semantic network for Semantic Networks for Concrete
an abstract verb. This also suggests that the instruc- and Abstract Nouns, Verbs,
tional information that can be provided for different and Adjectives
word types will also be different. In our study, we
Synonyms
examine children’s learning of words that fall at
­
different points on the continuum, enabling us to
­ Another key aspect of a semantic network for concrete
­determine the impact of word type on learning. and abstract nouns, verbs, and adjectives is synonyms,
or the core meaning(s) of a word. A synonym can be
either a single word or a short, decontextualized
Semantic Networks definition when a single-­word synonym does not exist
for Concrete Nouns (Miller & Fellbaum, 1991). A verb’s core meaning is
Functional Information often expressed by using a synonym with a manner
qualification, such as “To (verb 1) is to (verb 2) in some
For concrete nouns, functional information, or infor-
manner” (Miller & Fellbaum, 1991). For example, the
mation about what an object does or is used for, has
meaning of the word gallop might be expressed in this
been found to be highly salient for preschool word
way: “To gallop is to run fast.”
learners (Booth, 2009; Greif, Nelson, Keil, & Gutierrez,
Knowledge of synonyms is often the deciding factor
2006). Preschoolers were found to be more likely to
in whether a child knows a word, demonstrating a
learn words that were described in terms of their func-
­decontextualized knowledge of word meaning. A stu-
tion (e.g., a shovel is used to dig) than in nonfunctional
dent’s ability to select synonyms for a word is a
terms (e.g., a shovel has a part that is made out of metal;
commonly used standardized measure of receptive
­
Booth, 2009; Nelson, O’Neil, & Asher, 2008).
­vocabulary knowledge (e.g., Test of Word Knowledge;
Wiig & Secord, 1992). In teaching vocabulary, giving
Hierarchical Information synonyms or short, decontextualized word meaning
Word knowledge also includes the understanding of explanations has been shown to help primary-­grade
­hierarchical relationships among concrete nouns. This students learn new words (Biemiller & Boote, 2006).
dimension involves the ability not simply to add nodes However, Beck, McKeown, and Kucan (2002) cautioned
to the semantic network but also to categorize the that providing decontextualized word meaning infor-
relationships among nodes, such as the ability to
­ mation is helpful only when couched in child-friendly

184 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
language and also paired with more contextualized ex- the word in a single context, to eventually learning to
amples of how a word is used. use the word flexibly in a range of contexts (Clark,
2010).
Gestures
Although knowledge about words is often thought of as Summary
only verbal, there is a growing awareness that embod- Deep word knowledge for concrete nouns includes
ied experiences of words may help support comprehen- functional, perceptual, and hierarchical information,
sion and that gestures serve as another way of synonyms, nonverbal information, and knowledge
representing the meanings of words (Glenberg, about how words are used. The concepts underlying ab-
Gutierrez, Levin, Japuntich, & Kaschak, 2004). Just as stract nouns and verbs are complex and relational and,
words serve as a label for underlying semantic infor- therefore, more easily described with synonyms or ges-
mation, gestures can serve as a similar kind of label, tures or through meaningful context. Adjectives lend
although they can also provide supplementary infor- themselves to networks of knowledge composed of syn-
mation about meaning in ways a word label does not onyms, gestures, and use in context.
(Göksun, Hirsh-­ Pasek, & Golinkoff, 2010). Gesture
plays an important supporting role in communication
because of its ability to clarify or supplement spoken
language, especially when that language is complex Measuring Depth
(McNeil, Alibali, & Evans, 2000). Pairing language of Word Knowledge
with gesture, rather than using language alone, was Depth of word knowledge is an important construct
shown to improve comprehension for preschoolers that is rarely measured in vocabulary interventions.
(McNeil et al., 2000) and for first and second graders Instead, many assess only whether a child can recog-
(Glenberg et al., 2004). McNeil and colleagues posited nize a word, using tasks such as asking children to se-
that gestures can act as a scaffold for verbal informa- lect a picture from a group that matches a word spoken
tion, helping to guide children’s attention to the by the tester (Dunn & Dunn, 1997; Ouellette, 2006).
­semantic content of complex language. Gesture may These kinds of measures do not account for the fact that
also function as an alternative way for children to ex- words can be known to a lesser or greater extent and
press knowledge before they can explain it verbally that differences in depth of knowledge have conse-
(Capone, 2007). Examining gestural representations of quences for reading comprehension (Perfetti, 2007).
knowledge, therefore, may show increases in student The vocabulary measure used in the present study
learning that would otherwise be overlooked (Verhallen is a definition task in which we ask students to tell us
& Schoonen, 1993). what they know about words. Wasik and Hindman
Gestures may be a particularly powerful way of (2014) suggested that for young emergent readers who
teaching the meanings of verbs. Although concrete learn words through oral language, not print, a
nouns have stable perceptual features, verbs are slightly more limited conception of depth is appropri-
­dynamic and require children to abstract the verbal ate: a facility with basic phonological, semantic, syn-
­essence of an action (Golinkoff et al., 2002). Gestures tactic, and contextual, but not orthographic, aspects
may be a useful way of representing verbs because they of a word. Definition tasks capture the quality of the
filter out the noise and distill an action into a limited, semantic and contextual information a child knows
more meaningful essence. about a word and may also indirectly tap into the
strength of his or her phonological and syntactic
Contextual Information representation. Consistent with the lexical quality
­
Finally, rich word knowledge must include not only ­hypothesis, definition tasks test the main criterion for
semantic information but also an ability to use a word a high-­quality representation: the ability to retrieve
in different contexts (Nagy & Townsend, 2012; Stahl word identities.
& Fairbanks, 1986). Nagy and Scott (2000) argued The following three questions are addressed:
that knowing a word means being able to do things
1. Did children’s depth of vocabulary knowledge for
with it, and the ability to use a word correctly in con-
target words increase, as compared with their
text shows a deep, applied knowledge of a word and its
knowledge of exposure and control words?
use. Like other elements of depth, the ability to appro-
priately use a word develops over time, progressing 2. Did increases in depth of knowledge for target
from a basic association with the word and its typical words vary by word type?
context of use (e.g., knowing that the word has some- 3. How did the kind of semantic information
thing to do with __; Dale, 1965), to being able to use learned vary by word type?

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 185
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Methods TABLE 1
Demographic Information for the Sample
Study Design and Characteristic N Percentage
Research Participants Gender
The present study was conducted as part of Read, Play,
Female 130 54.2
Learn, a project aimed at increasing the vocabulary
knowledge of preschoolers from low-­ income back- Male 110 45.8
grounds through a combined method of book reading Race/ethnicity
and play (Dickinson et al., 2013; Weisberg et al., 2015).
The goal of the larger intervention was to test the effi- African American 131 54.6
cacy of play combined with book reading as a method Hispanic 55 22.9
of vocabulary instruction; however, the present study
Caucasian 33 13.8
focuses specifically on increases in depth of children’s
word knowledge by word type (concrete and abstract Other 16 6.6
nouns, verbs, and adjectives) during the course of this No response 5 2.1
intervention, without examining the efficacy of the
book-­reading and play methods in detail. English learner
This study uses a within-­subjects research design in No 204 85.0
which children served as their own controls. To deter-
Yes 36 15.0
mine whether children learned a significant amount
about words during the intervention, they were assessed Note. English learner = children who speak a language other than English
at home.
on three kinds of words: target words, which were part
of the book text, used in the play sessions, and explicitly
defined; exposure words, which were not explicitly
Farmer Duck by Martin Waddell and Pumpkin Soup by
taught or defined but were in the book read-­a louds and
Helen Cooper. All four books were comparable in terms
used in the play sessions; and control words, which were
of the pictorial representations of most target words,
not used in the intervention at all. This design allowed
text complexity, and length.
us to test whether the effects of the intervention were
Ten target words per book—abstract and concrete
due to merely hearing words used repeatedly or if addi-
nouns, verbs, and adjectives—were selected using the
tional teaching of the words made a significant differ-
following procedures. As an initial step, we identified
ence in students’ learning.
words in the story that were considered Tier 2, or so-
Data come from 240 four-­and five-­year-­old stu-
phisticated words of high utility (Beck et al., 2002), and
dents (mean age = 4 years 11.3 months, standard devia-
would therefore need additional explanation for chil-
tion = 4.8 months). Eighty-­five of these students were
dren to understand them fully. Additional target words
enrolled in seven Head Start classrooms in a mid-­
were inserted in the texts because all four books lacked
Atlantic U.S. state, and 155 were enrolled in 11 pre-
10 total Tier 2 words. Because some of the books had
school classrooms from a state-­ funded program for
minimal text, these adaptations typically involved
low-­income families in a Southeastern U.S. city. The
­adding sentences with Tier 2 words that described the
sample included only students who did not have an
action depicted in the book’s illustrations. For example,
Individualized Education Plan and who understood
the book Farmer Duck includes several illustrations of
enough English to be able to follow directions, as
the duck doing work around the farm without any text
­reported by their teacher. Table 1 summarizes the de-
describing his actions. We added sentences such as
mographics for the sample. The intervention was deliv-
“[The duck] took his shovel and dug the weeds out,”
ered by nine female intervention specialists, all of whom
thereby providing a fuller description of the book’s ac-
possessed a bachelor’s or master’s degree plus experi-
tion without significantly altering the story line.
ence in early childhood settings.
Next, we considered whether the words could be
easily explained in child-­ friendly terms and were
Materials: Book and Word Selection semantically and phonologically distinct from one
­
The book-­reading and play intervention was developed ­another. We also cross-­referenced our selection with
around two themes (dragon and farm), which were cho- Biemiller’s (2010) list of words, which are rated in terms
sen for their appeal to young children and opportuni- of appropriateness for instruction by grade level. Nine
ties for play. Two books per theme were read aloud to target words did not appear on Biemiller’s list. Of the 31
students: The Knight and the Dragon by Tomie dePaola target words that were on the list, 61% were character-
and Dragon for Breakfast by Eunice McMullen, or ized as at least level T2—high-­priority words that are

186 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
typically known by more advanced students by the end including perceptual, functional (e.g., “The specta-
of second grade and not known by at-­risk students. cles help the king see better”), and hierarchical infor-
Approximately half of the target words also appeared mation whenever possible
on the Dale–Chall (Chall & Dale, 1995) list of words • The use of gesture, whenever possible, to kines-
known by 80% of fourth-­grade students. thetically reinforce meaning (e.g., “Can you pre-
Finally, to ensure that the target words were suffi- tend you are wearing spectacles like this?” while
ciently difficult, words that over 30% of students from the teacher makes spectacles with rounded
the previous iteration of the experiment identified cor- fingers)
rectly at pretest were replaced. We used the same meth-
• An example of a word in a context other than the
ods to select 17 exposure words and 16 control words
one used in the story (e.g., “Look, your teacher
that were comparable in difficulty to the taught words
wears spectacles, too!”)
and contained similar proportions of the four word
types (concrete and abstract nouns, verbs, and adjec-
During the third and fourth readings, students’ verbal
tives). Because of the significant cognitive demand of the
participation was encouraged to reinforce each word’s
vocabulary measure used in this study, in which stu-
phonological and meaning representations (e.g., “What
dents were asked to define words, we randomly selected
was the king wearing to help him see?”).
a subset of 21 target words, 10 exposure words, and eight
control words for testing on this measure. See
Appendix A for a list of all words assessed, along with Play Conditions
their word type, difficulty, and descriptive information. A 10-­minute play session immediately followed each
book reading. Three play conditions were developed to
test the effects of adult-­supported play on students’
Procedures word learning. Because the effects of the different play
The intervention was conducted over a two-­ month conditions are not the focus of the present study, we in-
­period, from April to May 2012. All students were indi- cluded play condition as a covariate in our analyses,
vidually pretested and posttested by members of the re- controlling for its effects.
search team for knowledge of the target vocabulary
within one week prior to and following the interven-
tion, respectively. Students were randomly assigned to Instrument and Scoring
one of three play conditions within classrooms, and New Word Definition Test—Modified
classrooms were randomly assigned to one of the two To measure students’ depth of knowledge of target
themes. Books within each theme were counterbal- words, an experimenter-­designed measure was devel-
anced. Intervention specialists read aloud to mixed-­ oped and administered at pretest and posttest. This
gender groups of three students in a quiet location measure was adapted from Blewitt, Rump, Shealy, and
outside the classroom for four consecutive days during Cook’s (2009) New Word Definition Test, which we
the week. Immediately following each book reading, ­renamed as the New Word Definition Test—Modified
play sessions were conducted. The current article does (NWDT–M) to reflect our adaptations to the coding
not focus on the main effects of different intervention scheme, namely, additional categories for gestures and
methods. contextual information. This definition task employs
an informal rather than a formal definition task (Snow,
Book Reading Cancino, De Temple, & Schley, 1991). Our focus here is
Intervention specialists read two books aloud to not the structure or form of children’s definitions,
­students four times over the course of the intervention. which may reveal more about their metalinguistic skills
Each target word was explained as part of every book than their knowledge of words. Therefore, the
reading, once during each reading as the words NWDT–M does not track the extent to which children
­occurred in the text and once after each reading was give adult-­like definitions of words, a skill that requires
completed as part of a vocabulary and plot review. The practice with the form of definitions (Read, 2004), but
explanation consisted of the following: instead codes for the amount of accurate semantic and
contextual information that students provided for each
• Drawing students’ attention to a word by pointing word. The results of the NWDT–M allow for an under-
to the picture, which also helps illustrate meaning standing of the kinds of information that preschool stu-
(e.g., “Look, the king is wearing spectacles” while dents learn about words ranging in perceptual
pointing to the glasses in the picture) accessibility.
• Definitional information delivered in concise, child- Students were asked to define concrete and abstract
friendly language (e.g., “Spectacles are glasses”), nouns, verbs, and adjectives verbally or by using

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 187
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
gestures. Students were not tested on all target, expo- context, because the student used a typical example to
sure, and control words on this measure due to time explain how a shovel could be used, along with seman-
constraints and the cognitive demands on children. tic information. Basic context, worth only 0.5 point,
The NWDT–M test forms for the dragon and farm was a simple association between a target word and a
themes were similar in the number of items, number of typical context, without any use of semantic informa-
words per word type, and difficulty of words. In a previ- tion. For example, students frequently said, “Santa
ous phase of this study, we identified words that were Claus,” for chimney, a response that does not include
known by more than 30% of students at pretest, and any semantic information but still contains an associa-
eliminated those words for the present study to ensure tion with a typical context in which the target word is
that all words were of a similar difficulty level. used. Incorrect or irrelevant responses received a score
For each word, students were asked, “What is (a) of 0. See Appendix B for more examples of coded
—?” and a follow-­up question, “Can you show me or tell ­student responses.
me anything else about —?” If a student did not respond
to a question, the tester moved on to the next word. All
student responses were transcribed by testers and also
video-­or audiotaped. A coding scheme was developed
Results
(adapted from Blewitt et al., 2009) to categorize and We performed multilevel regression models to address
score student responses for the number of information each of our three research questions, testing for (1) over-
units given. Coding was conducted by research assis- all growth in depth of knowledge, comparing students’
tants, and 20% of all forms were randomly selected and learning of target words to that of exposure and control
checked for reliability against a master coder after every words; and (2) growth in depth of knowledge for target
four forms were completed. Overall percentage agree- words only by word type and (3) by word type and
ment averaged 93.2%, with a mean Cohen’s Kappa value ­semantic information category. Multilevel model proce-
of .82. dures (Raudenbush & Bryk, 2002) were used to account
for interdependency among study observations.

Coding Scheme Descriptive Statistics

We used seven information unit categories to score stu-
The descriptive statistics in Table 2 show that there was
dent responses for semantic content and contextual in-
variance in students’ vocabulary knowledge. The distri-
formation: perceptual qualities, functional information,
bution was skewed, so log transformations were per-
part/whole, synonyms, gestures, meaningful context,
formed to improve the normality of the distribution.
and basic context. Each information unit was worth 1
Because students were randomly assigned to a
point except for basic context, which was worth 0.5
theme (dragon or farm) and the words tested for each
point. The first three categories were used for concrete
theme were different, it was necessary to determine
nouns only. Perceptual qualities included properties
whether the farm and dragon NWDT–M test forms
such as how something looks, smells, tastes, feels, or
were comparable. An independent samples t-­test on
sounds. Functional information included any process,
NWDT–M pretest and posttest scores for the dragon
purpose, or use for concrete nouns and answers the
and farm themes indicated that there was no significant
question, “What do you do with it?” Part/whole de-
difference in mean pretest NWDT–M scores in the two
scribed a distinct part of a target word or the whole that
themes, t(240) = 0.72, p = .473, 95% confidence interval
the target word was a part of. The remaining categories
[−0.55, 1.42], but that there was a significant difference
were used for all word types. Synonyms included any
at posttest, t(240) = −4.00, p < .001, 95% confidence in-
word or short phrase that was equivalent to the word
terval [−5.68, −1.93]. Because of this difference, we in-
being explained, and provided decontextualized mean-
cluded theme as a covariate in all subsequent analyses.
ing information. Gestures included gestures, actions, or
facial expressions (e.g., the teacher uses a scary face to
illustrate the word fierce) that showed knowledge of the Psychometric Properties of Measure
word’s meaning. Both of the test forms demonstrate acceptable internal
We also coded for two types of use in context. consistency at pretest (farm theme: Cronbach’s α = .73;
Meaningful context included responses that showed dragon theme: Cronbach’s α = .76) and posttest (farm
knowledge of the target word in a typical, meaningful theme: Cronbach’s α = .76; dragon theme: Cronbach’s
context, along with semantic information. For example, α = .86). We also evaluated the validity of the NWDT–M
if a student said, “A shovel is used to dig up weeds in a by comparing the test scores on a concurrent measure
garden,” “used to dig” would be scored for function, and of a highly related construct (Cronbach, 1971; Messick,
“weeds in a garden” would be scored for meaningful 1989), in this case, the receptive vocabulary measure

188 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
TABLE 2
New Word Definition Test—Modifieda Descriptive Statistics: Average Number of Information Units per Word by
Level of Instruction (n = 240)
Percentage
of the sample
with responses
Standard above floor
Word type Test period Minimum Maximum Mean deviation level
All words Pretest 0 1.17 0.20 0.21 85.4

Posttest 0 2.02 0.47 0.37 96.7

Target words Pretest 0 1.19 0.13 0.20 60.4

Posttest 0 2.15 0.55 0.48 90.8

Exposure words Pretest 0 2.50 0.38 0.42 68.8

Posttest 0 2.50 0.49 0.45 80.8

Control words Pretest 0 1.50 0.19 0.23 59.6

Posttest 0 1.25 0.25 0.26 64.6

Note. Above floor-­level responses are those coded as receiving a score greater than 0.
a
Researcher-­modified version of Blewitt, Rump, Shealy, and Cook’sb New Word Definition Test. bBlewitt, P., Rump, K.M., Shealy, S.E., & Cook, S.A.
(2009). Shared book reading: When and how questions affect young children’s word learning. Journal of Education & Psychology, 101(2), 294–304.

used in this study. Although the NWDT–M measures The model accounted for three nesting levels in the
depth of word knowledge and the receptive measure data: Level of instructionijk (target, exposure, and con-
evaluates breadth of knowledge, studies have shown trol) was repeated within childrenjk (n = 240), and chil-
that these two constructs are distinct but related. The dren were nested within play groupsk (n = 85). For
correlations between the NWDT–M and receptive mea- parsimony, the classroom random effects were aggre-
sure were statistically significant (r = .41 at pretest, gated at the reading playgroup level. Level of instruction
r = .62 at posttest), demonstrating a moderately strong was dummy coded with control words as the reference
relationship between them but also indicating that they group, which were contrasted to taught (γ100) and expo-
do not measure exactly the same construct. sure (γ200) words. To look at residualized gains, students’
pretest vocabulary scores (γ300) were included as a co-
Overall Growth in variate, along with age at pretest (γ010), book theme (γ001),
Depth of Knowledge and play condition (γ002). Book theme was coded with
the dragon theme as the reference group, and play condi-
Although our primary interest in this study was to
tion was coded with free play as the reference group.
­examine children’s growth in depth of knowledge by
Theme was included as a covariate because the words
word type and semantic information category, it was
used in the two themes were different from one another.
necessary first to determine whether they learned a sig-
Analysis indicated that after accounting for the
nificant amount about taught words in general before
model covariates of pretest vocabulary knowledge,
we carried out more detailed analyses. To answer our
γ300 = 0.59, standard error (SE) = 0.04, p < .001; age at
first research question about overall growth in depth of
pretest, γ 010 = 0.003, SE = 0.001, p < .001; theme,
knowledge, we compared children’s learning of taught
γ 001 = −0.03, SE = 0.01, p < .001; and play condition,
words with their learning of exposure and control
γ 002 (free play vs. guided play = 0.03, SE = 0.01,
words. Using the following multilevel regression model,
p = .001, free play vs. directed play = 0.03, SE = 0.01,
we tested whether vocabulary gains varied by level of
p = .001), students knew more taught words at post-
instruction (target, exposure, and control words):
test than control words, γ100 = 0.10, SE = 0.01, p < .001,
Posttestijk = γ000 + (γ100 × Taughtijk ) and more exposure words than control words,
γ200 = 0.04, SE = 0.01, p < .001. Post hoc pairwise com-
+ (γ200 × Exposureijk )
parisons with least significant difference adjustments
+ (γ300 × Pretestijk ) + (γ010 × Agejk ) (1) also confirmed that students knew more taught
+ (γ001 × Themek ) + (γ002 × Conditionk ) words than exposure words (p < .001). On average,
after controlling for covariates, students gave 4.68
+ U00k + U0jk + rijk more information units at posttest for the target

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 189
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
words as a whole, or 0.42 more information units per type was dummy coded with concrete nouns as the
target word. The pretest–posttest effect sizes were ­reference group, which was contrasted to verbs (γ100),
1.22 for target words, 0.26 for exposure words, and abstract nouns (γ200), and adjectives (γ300).
0.22 for control words. This analysis indicated that students showed sig-
nificantly greater growth in their knowledge of con-
Growth in Depth of Knowledge crete nouns as compared with verbs, γ100 = −0.05,
SE = 0.01, p < .001; abstract nouns, γ200 = −0.12,
by Word Type SE = 0.01, p < .001; and adjectives, γ300 = −0.13, SE = 0.01,
The first analysis determined that students had indeed p < .001. Post hoc pairwise comparisons with least sig-
shown significantly greater growth in their knowledge nificant difference adjustments revealed that students
of taught words than exposure and control words. also learned s­ignificantly more about verbs than ab-
Further analyses were conducted only on target words stract nouns (p < .001) and adjectives (p < .001). There
and examined students’ growth in knowledge of target was no ­significant ­difference in the learning of adjec-
words in more detail. Our second research question pro- tives and abstract nouns. Students showed significant
posed to investigate how students’ learning varied by growth in ­k nowledge for each of the four word types
word type. Using the following multilevel regression from pretest to posttest (p < .001). Figure 1 shows the
model, we tested whether vocabulary gains varied by pretest–­posttest effect sizes for each word type and the
word type (concrete nouns, abstract nouns, verbs, and significant c­ ontrasts in learning by word type. Table 3
adjectives). lists d
­ escriptive i­nformation for target words by word
Posttestijk = γ000 + (γ100 × Verbsijk ) + (γ200 type.

× AbstractNounsijk ) + (γ300 × Adjectivesijk )

Growth in Depth of Knowledge
+ (γ400 × Pretestijk ) + (γ010 × Agejk ) (2) by Word Type and Semantic
+ (γ001 × Themek ) + (γ002 × Conditionk ) Information Category
+ U00k + U0jk + rijk Our third research question asked how the kind of
­semantic information learned varied by word type. To
The model accounted for three nesting levels in the address this question, we tested whether significant
data; word typeijk (concrete nouns, abstract nouns, gains were made from pretest to posttest for each se-
verbs, and adjectives) was repeated within childrenjk, mantic information category by word type, using the
and children were nested within play groupsk. Word following multilevel regression model:

FIGURE 1
Cohen’s d Pretest–Posttest Effect Sizes for Concrete Nouns, Verbs, Abstract Nouns, and Adjectives
1.4 (2, 3, 4)
1.24
1.2
Cohen's d effect size

(1, 3, 4)
1 0.89
0.8 (1, 2)
0.65 (1, 2)
0.56
0.6

0.4

0.2

0
Concrete Nouns Verbs Abstract Nouns Adjectives
(n = 9) (n = 6) (n = 3) (n = 3)
Word type
Note. Reference group is level of instruction at time 1 (pretest). (1) = significant difference (p < .001) from concrete nouns; (2) = significant difference
(p < .001) from verbs; (3) = significant difference (p < .001) from abstract nouns; and (4) = significant difference (p < .001) from adjectives.

190 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
TABLE 3
New Word Definition Test—Modifieda Descriptives: Average Number of Information Units per Target Word by
Word Type (n = 240)
Percentage
of the sample
with responses
Standard above floor
Word type Test period Minimum Maximum Mean deviation level
Concrete noun Pretest 0 2.33 0.18 0.32 43.8

Posttest 0 3.00 0.78 0.70 85.0

Verb Pretest 0 2.50 0.22 0.38 45.4

Posttest 0 2.00 0.57 0.55 75.4

Adjective Pretest 0 2.00 0.06 0.23 7.9

Posttest 0 2.00 0.24 0.42 31.7

Abstract noun Pretest 0 2.00 0.05 0.22 9.6

Posttest 0 2.33 0.30 0.49 35.8

Note. Above floor-­level responses are those coded as receiving a score greater than 0.
a
Researcher-­modified version of Blewitt, Rump, Shealy, and Cook’sb New Word Definition Test. bBlewitt, P., Rump, K.M., Shealy, S.E., & Cook, S.A.
(2009). Shared book reading: When and how questions affect young children’s word learning. Journal of Education & Psychology, 101(2), 294–304.

Vocabijk = γ000 + (γ100 × Observationijk ) did not show significant growth in those categories.
+ (γ010 × Agejk ) + (γ001 × Themek ) (3) Table 4 shows the growth in depth of knowledge from
pretest to posttest for taught words in all word types by
+ (γ002 × Conditionk ) + U00k + U0jk + rijk
semantic information category, along with the percent-
The model accounted for three nesting levels in the ages of words in each category taught using that kind of
data; assessment observationijk was repeated within chil- semantic information and effect sizes for each category.
drenjk, who were nested within play groupsk. The
­independent variable of interest is Observation (γ100) and
represents the contrast of pretest and posttest scores for Discussion
each semantic information category by word type. Depth of knowledge is an important and distinct facet
Separate models were conducted for each semantic infor- of vocabulary knowledge that supports reading com-
mation category by word type. Because separate tests were prehension (Ouellette, 2006). Because children who al-
run for each semantic information type (19 tests), we used ready have rich vocabulary knowledge are better able to
a Bonferroni-­adjusted α level of .003 per test (.05/19) to de- acquire more rich vocabulary knowledge, and those
termine significance. The kind of semantic information who lack that knowledge fall further behind (Neuman
that students learned differed by word type (see Table 4). & Celano, 2006), there is a pressing need for efforts that
Students showed significant growth in their knowl- focus on building vocabulary depth in young children.
edge of all semantic information categories for concrete However, there is very little information about the kind
nouns, learning functional information best, followed of instruction that fosters this learning. We know that
by meaningful context, synonyms, part–whole rela- increased encounters with words build depth (Beck &
tions, gestures, perceptual qualities, and basic context. McKeown, 2007), but there has been little research ad-
Students also showed growth in all semantic informa- dressing the question of which specific kinds of infor-
tion categories for verbs, learning synonyms best, fol- mation about words are best learned by children,
lowed by meaningful context, gestures, and basic therefore adding to their depth of knowledge. The pres-
context. For abstract nouns, students showed signifi- ent study addresses this gap by showing that certain
cant growth only in their knowledge of synonyms and kinds of input are especially helpful in fostering depth
meaningful context. They showed no growth in knowl- and that these kinds of input vary by word type.
edge of the basic context and gesture categories for The words we taught in this study fell at different
abstract nouns. Finally, students showed growth in
­ points along the conceptual continuum, ranging in
knowledge of synonyms for adjectives. Although mean- their perceptual accessibility. Concrete nouns such as
ingful context was taught for all of the adjectives in the handkerchief were easy to observe and individuate, had
study, and gesture was taught for 67% of them, students a consistent shape, and were highly imageable. In

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 191
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
TABLE 4
Unstandardized Coefficients (standard errors) and Effect Sizes for Students’ Growth in Knowledge of the Target
Words by Word Type and Semantic Information Type Used to Teach the Words (n = 240)
Word type and semantic Percentage of the words taught using
information category the semantic information category Coefficient (b) d
Concrete noun

Function 89 0.21 (0.02) 0.96*

Meaningful context 100 0.15 (0.01) 0.84*

Part/whole 33 0.12 (0.01) 0.83*

Synonym 100 0.07 (0.01) 0.65*

Gesture 11 0.04 (0.01) 0.53*

Perceptual quality 56 0.05 (0.01) 0.40*

Basic context 0.03 (0.01) 0.25*

Verb

Synonym 100 0.12 (0.01) 0.79*

Meaningful context 100 0.09 (0.01) 0.53*

Gesture 67 0.05 (0.01) 0.42*

Basic context 0.04 (0.01) 0.29*

Abstract noun

Synonym 100 0.05 (0.01) 0.49*

Meaningful context 100 0.04 (0.01) 0.38*

Basic context <0.01 (0.01) 0.04

Gesture 0 <0.01 (0.02) 0.03

Adjective

Synonym 100 0.05 (0.01) 0.46*

Basic context 0.02 (0.01) 0.15

Meaningful context 100 0.01 (0.01) 0.13

Gesture 67 <0.01 (0.02) 0.11

Note. d = Cohen’s d standardized mean difference effect size. The reference group for effect sizes is level of knowledge at time 1 (pretest).
*p < .001.

contrast, verbs such as returning and abstract nouns functional information for them. This variability in in-
such as plan had no consistent shape, were difficult to struction and in word types provides an opportunity to
picture, and could not be physically manipulated or as examine children’s depth of word learning in a detailed
easily observed in the world. Over the course of the in- way, looking at their relative learning of words by both
tervention, all of these words were taught by providing a word type and the categories of semantic information
short verbal definition and contextual information, but that students were able to learn for each word type.
beyond these common features, we did not systemati-
cally control the kind of information supplied about
words, because different words lent themselves to dif- Results by Word Type
ferent kinds of supportive information. Indeed, it would Our results are consistent with Maguire and colleagues’
be nearly impossible to fully vary word type by infor- (2006) theory that perceptual accessibility, which in-
mation type. For example, although words such as re- cludes the factors of shape, imageability, concreteness,
turning and plan can be defined, it is difficult to use an and individuation, predicts the ease with which words
iconic gesture to represent these words or to supply are learned. The words that showed the most growth in

192 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
learning were those that were the most perceptually ac- significant growth for all word types except adjectives.
cessible. Students learned the most about concrete nouns Finally, the gesture category showed significant growth
(Cohen’s d = 1.24), followed by verbs (of which all but one for verbs (Cohen’s d = 0.42) but not for adjectives, in
were perceptually accessible; Cohen’s d = 0.89), abstract spite of the instruction given.
nouns (Cohen’s d = 0.65), and finally adjectives (Cohen’s These results demonstrate that not only were the
d = 0.56). This finding applies not only to form class dif- concrete nouns better learned because of their percep-
ferences but also to differences within form classes; in tual accessibility, but they also naturally lent themselves
this study, the highly abstract concepts labeled by certain to fuller, more varied kinds of instructional informa-
nouns were more difficult to learn than concrete verbs. tion. The less perceptually accessible words, in contrast,
were not only less imageable, less concrete, more diffi-
Semantic Information Categories cult to individuate, and without a consistent shape but
were also more difficult to define in terms of function
The conceptual continuum theory posits that percep-
or discuss as a part or whole. Therefore, our data sug-
tual accessibility is the primary determinant of ease of
gest that speed of learning of words may reflect the con-
learning. Our examination of the kinds of information
verging effects of both their perceptual accessibility and
that we could naturally supply about different words
the type of information that can be provided.
made evident that the types of information that can be
provided about word meanings may be a second factor
that affects word learning and may help explain differ- Educational Implications
ential learning of varying types of words. Our results can help inform vocabulary instruction by
We found that certain types of semantic informa- guiding the type of information that teachers use to ex-
tion were more often retained than others. The most plain new words to preschoolers. For the concrete nouns
perceptually accessible category of words, concrete in our study, students found functional information
nouns, was taught with a synonym, an explanation of highly salient. This finding is in line with Booth’s
what the word meant in context, and a reference to a (2009) findings that providing functional information
picture in the book (see Table 4). All but one of the con- for objects is a powerful way to help increase preschool-
crete nouns were also explained using functional infor- ers’ depth of word knowledge. To support word learn-
mation (e.g., “We use nostrils to breathe”), about half ing for concrete nouns, then, teachers should not only
were explained using perceptual qualities (e.g., “Nostrils exploit their perceptual accessibility by using pictures
look like little holes”), and a third of the words were ex- and pointing out important parts of the object but
plained by pointing out a part of the object or the whole should also explain an object’s function. Perceptual in-
of which the object is a part (e.g., “Scales are on a fish’s formation can also serve as a gateway to conceptual in-
body”; see Table 4). Students’ learning of concrete nouns formation (Booth, 2009): for example, describing what
reflects this instruction. They showed significant armor looks like (e.g., hard, made out of metal) natu-
growth in their knowledge of function, context, syn- rally leads into a description of what it is used for (to
onym, part–whole relationships, gesture, and percep- protect a person’s body in a fight).
tual qualities. Interestingly, function was the category These results also make it clear that verbs (even the
that showed the most growth (Cohen’s d = 0.96) for concrete ones used in this study), abstract nouns, and
concrete nouns, and although only a third of the words adjectives simply have fewer readily describable features
were explained using part–whole relationships, this as compared with concrete nouns. In spite of this limi-
type of semantic information also had a large effect size tation, the verbs taught in this study were well learned
(Cohen’s d = 0.83). (Cohen’s d = 0.89), most likely because they are rela-
The less perceptually accessible words in the tively accessible perceptually. Because of these features,
study—abstract nouns, verbs, and adjectives—were also we were able to teach students simple, easy-­to-­remember
taught with a synonym and an explanation of what the gestures for many of the verbs. This suggests that teach-
word meant in context. Pictures from the book were ing verbs through gesture or other forms of embodied
referenced for two thirds of the abstract nouns and learning can indeed serve as a helpful scaffold for the
verbs and about a third of the adjectives. Gesture was verbal information provided (McNeil et al., 2000).
another important element of instruction for these The growth in learning for abstract nouns (Cohen’s
words: Two thirds of the verbs and adjectives were la- d = 0.65) and adjectives (Cohen’s d = 0.56) was modest
beled with a gesture illustrating the word’s meaning. but still significant. Given that these types of words are
Consistent with the instruction given, synonym was the abstract, have few describable features, and are difficult
best learned category for verbs (Cohen’s d = 0.79), ab- to explain, how can we help children learn them? One
stract nouns (Cohen’s d = 0.49), and adjectives (Cohen’s important takeaway from these results is that both syn-
d = 0.46). The meaningful context category also showed onyms and meaningful context were well learned for

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 193
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
almost every word type (students did not show growth preschool students learned and expressed semantic in-
in meaningful context for adjectives). The large effect formation about Tier 2 words with only a verbal prompt.
sizes for synonyms suggests that children can learn and Furthermore, we did not test students at a later date
provide brief, simple definitions, further justifying the for maintenance of vocabulary knowledge. Further
use of clear word meaning explanations when new vo- studies exploring the instructional implications of dif-
cabulary words are introduced to children (Biemiller & ferent kinds of vocabulary instruction should explore
Boote, 2006). This result held for all form classes in the whether more extensive types of vocabulary instruction
study, showing that even when words are fairly abstract, lead to better retention of knowledge as opposed to
children are able to learn something about a word’s brief, less comprehensive instruction.
­essential meaning through instruction.
The meaningful context category also showed sig-
nificant growth for all word types except adjectives,
suggesting that children not only need clear semantic
Conclusions
information about words but also remember and use in- The present study adds to the research on children’s lan-
formation about the typical contexts in which words are guage acquisition by examining the factors that lead to
used. This finding supports instructional methods that depth of learning by word type. We respond to the need
emphasize the importance of both giving definitions in the field for reports of vocabulary interventions that
and teaching vocabulary in context (Beck et al., 2002; discuss not only how many words children have learned
Biemiller & Boote, 2006; Coyne, Simmons, Kame’enui, but also how much, what kind of information has been
& Stoolmiller, 2004). This may be especially important learned about different types of words, and how this in-
for words that are difficult to learn. Hearing a difficult, formation can be used to better tailor ­vocabulary in-
highly abstract word (e.g., plan in this study) used in struction. Furthermore, studies do not ­always report the
context multiple times (in this study, at least eight times) type of words taught. Given that depth of learning may
allows children to progressively refine their knowledge interact with word type, our study reinforces the impor-
of the nuances of its meaning. However, average growth tance of attending to word type when planning and re-
per word was somewhat limited, suggesting that eight porting results of vocabulary interventions. Our results
exposures may not be sufficient. Bolger et al. (2008) suggest not only that some words are learned more
found that adult learners had higher quality knowledge quickly and with greater depth because they are more
of words when encountering them in multiple varied perceptually accessible but also that these perceptually
contexts as opposed to a single context multiple times. accessible words also lend themselves to a greater variety
In this study, preschool students’ encounters with words of highly salient instructional information types. Highly
in related but different contexts, such as the book-­ abstract words, then, are not only more difficult for chil-
reading and play settings, may have had additional ben- dren to learn on their own but also more difficult to learn
efits in helping students refine their word knowledge. through instruction. Students must gain a deep knowl-
edge of highly abstract, conceptually complex words to
achieve academic success (Snow & Uccelli, 2009), and
Limitations our results demonstrate that clear information about
The number of words for each word type here is small, meaning and use of words in meaningful contexts can
particularly for the adjectives and abstract noun catego- help support learning. Further efforts must concentrate
ries, and the findings here may not be applicable to ad- on ways to foster depth of knowledge for the words that
jectives or abstract nouns that are significantly more students will need most as they progress through school.
concrete or more abstract than the ones used in this
study. We have provided the specific words used (see NOTE
Appendix A) to help guide interpretation. The research reported here was supported by the Institute of
Further studies should also look at the learning of Education Sciences, U.S. Department of Education, through Grant
abstract verbs to more fully explore the impact of cer- R305A110128 to Vanderbilt University, Temple University, the
tain types of instruction on words along the conceptual University of Delaware, and Lehigh University.
continuum.
It is also important to note that students’ increases REFERENCES
in word knowledge were relatively small (about 0.42 in- Anderson, R.C., & Freebody, P. (1981). Vocabulary knowledge. In
formation units per word), given that students could J.T. Guthrie (Ed.), Comprehension and teaching: Research reviews
(pp. 77–117). Newark, DE: International Reading Association.
theoretically score a nearly unlimited number of points Beck, I.L., & McKeown, M.G. (2007). Increasing young low-­income
for each word (although the highest score for an indi- children’s oral vocabulary repertoires through rich and focused
vidual word was 6 points). However, the demands of the instruction. The Elementary School Journal, 107(3), 251–271.
definition task are significant, and it is meaningful that doi:10.1086/511706

194 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Beck, I.L., McKeown, M.G., & Kucan, L. (2002). Bringing words to Dunn, L.M., & Dunn, L.M. (1997). Peabody Picture Vocabulary Test
life: Robust vocabulary instruction. New York, NY: Guilford. (3rd ed.). Circle Pines, MN: American Guidance Service.
Biemiller, A. (2010). Words worth teaching: Closing the vocabulary Gentner, D. (1982). Why nouns are learned before verbs: Linguistic
gap. Columbus, OH: SRA/McGraw-Hill. relativity versus natural partitioning. In S.A. Kuczaj (Ed.),
Biemiller, A., & Boote, C. (2006). An effective method for building Language development: Vol. 2. Language, thought, and culture
meaning vocabulary in primary grades. Journal of Educational (pp. 301–334). Hillsdale, NJ: Erlbaum.
Psychology, 98(1), 44–62. doi:10.1037/0022-­0663.98.1.44 Gentner, D., & Boroditsky, L. (2001). Individuation, relativity and
Bion, R.A.H., Borovsky, A., & Fernald, A. (2013).Fast mapping, slow early word learning. In M. Bowerman, & S.C. Levinson (Eds.),
learning: Disambiguation of novel word–object mappings in rela- Language, culture, and cognition: Vol. 3. Language acquisition
tion to vocabulary learning at 18, 24, and 30 months. Cognition, and conceptual development (pp. 215–256). New York, NY:
126(1), 39–53. doi:10.1016/j.cognition.2012.08.008 Cambridge University Press.
Blewitt, P., Rump, K.M., Shealy, S.E., & Cook, S.A. (2009). Shared Gillette, J., Gleitman, H., Gleitman, L., & Lederer, A. (1999). Human
book reading: When and how questions affect young children’s simulations of vocabulary learning. Cognition, 73(2), 135–176.
word learning. Journal of Education & Psychology, 101(2), 294– doi:10.1016/S0010-­0277(99)00036-­0
304. doi:10.1037/a0013844 Glenberg, A.M., Gutierrez, T., Levin, J.R., Japuntich, S., & Kaschak,
Bloom, P. (2000). How children learn the meanings of words. M.P. (2004). Activity and imagined activity can enhance young
Cambridge, MA: MIT Press. children’s reading comprehension. Journal of Educational
Bolger, D.J., Balass, M., Landen, E., & Perfetti, C.A. (2008). Context Psychology, 96(3), 424–436. doi:10.1037/0022-­0663.96.3.424
variation and definitions in learning the meanings of words: An Göksun, T., Hirsh-Pasek, K., & Golinkoff, R.M. (2010). How do pre-
instance-­ based learning approach. Discourse Processes, 45(2), schoolers express cause in gesture and speech? Cognitive
122–159. doi:10.1080/01638530701792826 Development, 25(1), 56–68. doi:10.1016/j.cogdev.2009.11.001
Booth, A.E. (2009). Causal supports for early word learning. Child Golinkoff, R.M., Chung, H.L., Hirsh-Pasek, K., Liu, J., Bertenthal,
Development, 80(4), 1243–1250. doi:10.1111/j.1467-­8624.2009.01328.x B.I., Brand, R., … Hennon, E. (2002). Young children can extend
Booth, A.E., & Waxman, S.R. (2002). Word learning is ‘smart’: motion verbs to point-­light displays. Developmental Psychology,
Evidence that conceptual information affects preschoolers’ 38(4), 604–614. doi:10.1037/0012-­1649.38.4.604
­extension of novel words. Cognition, 84(1), B11–B22. doi:10.1016/ Greif, M.L., Nelson, D.G.K., Keil, F.C., & Gutierrez, F. (2006). What
S0010-­0277(02)00015-­X do children want to know about animals and artifacts? Domain-­
Cain, K., Oakhill, J., & Bryant, P. (2004). Children’s reading com- specific requests for information. Psychological Science, 17(6),
prehension ability: Concurrent prediction by working memory, 455–459. doi:10.1111/j.1467-­9280.2006.01727.x
verbal ability, and component skills. Journal of Educational Gropen, J., Pinker, S., Hollander, M., & Goldberg, R. (1991). Syntax
Psychology, 96(1), 31–42. doi:10.1037/0022-­0663.96.1.31 and semantics in the acquisition of locative verbs. Journal of
Capone, N.C. (2007). Tapping toddlers’ evolving semantic represen- Child Language, 18(1), 115–151. doi:10.1017/S0305000900013325
tation via gesture. Journal of Speech, Language, and Hearing Henriksen, B. (1999). Three dimensions of vocabulary development.
Research, 50(3), 732–745. doi:10.1044/1092-­4388(2007/051) Studies in Second Language Acquisition, 21(2), 303–317. doi:10.1017/
Carey, S. (1978). The child as word learner. In M. Halle, J. Bresnan, S0272263199002089
& A. Miller (Eds.), Linguistic theory and psychological reality Hollich, G., Hirsh-Pasek, K., & Golinkoff, R.M. (2000). II. The emer-
(pp. 264–293). Cambridge, MA: MIT Press. gentist coalition model. Monographs of the Society for Research in
Chall, J.S., & Dale, E. (1995). Readability revisited: The new Dale- Child Development, 65(3), 17–29. doi:10.1111/1540-­5834.00092
Chall readability formula. Cambridge, MA: Brookline. Imai, M., Li, L., Haryu, E., Okada, H., Hirsh-Pasek, K., Golinkoff, R.M.,
Christ, T. (2011). Moving past “right” or “wrong” toward a contin- & Shigematsu, J. (2008). Novel noun and verb learning in Chinese-­,
uum of young children’s semantic knowledge. Journal of Literacy English-­, and Japanese-­speaking children. Child Development, 79(4),
Research, 43(2), 130–158. doi:10.1177/1086296X11403267 979–1000. doi:10.1111/j.1467-­8624.2008.01171.x
Clark, E.V. (2010). First language acquisition (2nd ed.). Cambridge, Landi, N., & Perfetti, C.A. (2007). An electrophysiological investigation
UK: Cambridge University Press. of semantic and phonological processing in skilled and less-­skilled
Coyne, M.D., Simmons, D.C., Kame’enui, E.J., & Stoolmiller, M. comprehenders. Brain and Language, 102(1), 30–45. doi:10.1016/
(2004). Teaching vocabulary during shared storybook readings: j.bandl.2006.11.001
An examination of differential effects. Exceptionality, 12(3), 145– Maguire, M.J., Hirsh-Pasek, K., & Golinkoff, R.M. (2006). A unified
162. doi:10.1207/s15327035ex1203_3 theory of word learning: Putting verb acquisition in context. In
Cronbach, L.J. (1971). Test validation. In R.L. Thorndike (Ed.), K. Hirsh-Pasek, & R.M. Golinkoff (Eds.), Action meets word: How
Educational measurement (2nd ed., pp. 443–507). Washington, children learn verbs (pp. 364–391). New York, NY: Oxford
DC: American Council on Education. University Press.
Dale, E. (1965). Vocabulary measurement: Techniques and major McNeil, N.M., Alibali, M.W., & Evans, J.L. (2000). The role of ges-
findings. Elementary English, 42(8), 895–901, 948. ture in children’s comprehension of spoken language: Now they
Dickinson, D.K., & McCabe, A. (2001). Bringing it all together: The need it, now they don’t. Journal of Nonverbal Behavior, 24(2),
multiple origins, skills and environmental supports of early 131–150. doi:10.1023/A:1006657929803
­literacy. Learning Disabilities Research & Practice, 16(4), 186–202. Medina, T.N., Snedeker, J., Trueswell, J.C., & Gleitman, L.R. (2011).
doi:10.1111/0938-­8982.00019 How words can and cannot be learned by observation.
Dickinson, D.K., Turner, K.A., Collins, M.F., Nicolopoulo, A., Proceedings of the National Academy of Sciences of the United
Golinkoff, R.M., Hirsh-Pasek, K., … Rivera, B.L. (2013, April). States of America, 108(22), 9014–9019. doi:10.1073/pnas
More word learning occurs when book reading is followed by teacher-­ .1105040108
supported play. Paper presented at the annual meeting of the Messick, S. (1989). Validity. In R.L. Linn (Ed.), Educational mea-
American Educational Research Association, San Francisco, CA. surement (3rd ed., pp. 13–103). New York, NY: American Council
Dollaghan, C. (1985). Child meets word: “Fast mapping” in pre- on Education/Macmillan.
school children. Journal of Speech and Hearing Research, 28(3), Miller, G.A., & Fellbaum, C. (1991). Semantic networks of English.
449–454. doi:10.1044/jshr.2803.454 Cognition, 41(1–3), 197–229. doi:10.1016/0010-­0277(91)90036-­4

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 195
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
Morgan, P.L., Fuchs, D., Compton, D.L., Cordray, D.S., & Fuchs, In E. Bialystock (Ed.), Language processing in bilingual chil-
L.S. (2008). Does early reading failure decrease children’s reading dren (pp. 90–112). Cambridge, UK: Cambridge University
motivation? Journal of Learning Disabilities, 41(5), 387–404. Press.
doi:10.1177/0022219408321112 Snow, C.E., & Uccelli, P. (2009). The challenge of academic lan-
Nagy, W.E., & Scott, J.A. (2000). Vocabulary processes. In M.L. guage. In D.R. Olson, & N. Torrance (Eds.), The Cambridge hand-
Kamil, P.B. Mosenthal, P.D. Pearson, & R. Barr (Eds.), book of literacy (pp. 112–133). New York, NY: Cambridge
Handbook of reading research (Vol. 3, pp. 269–284). Mahwah, University Press.
NJ: Erlbaum. Stahl, S.A., & Fairbanks, M.M. (1986). The effects of vocabulary in-
Nagy, W.E., & Townsend, D. (2012). Words as tools: Learning aca- struction: A model-­based meta-­a nalysis. Review of Educational
demic vocabulary as language acquisition. Reading Research Research, 56(1), 72–110. doi:10.3102/00346543056001072
Quarterly, 47(1), 91–108. doi:10.1002/RRQ.011 Storch, S.A., & Whitehurst, G.J. (2002). Oral language and code-­
Nation, K., & Snowling, M.J. (1998). Semantic processing and the related precursors to reading: Evidence from a longitudinal
development of word-­recognition skills: Evidence from children structural model. Developmental Psychology, 38(6), 934–947.
with reading comprehension difficulties. Journal of Memory and doi:10.1037/0012-­1649.38.6.934
Language, 39(1), 85–101. doi:10.1006/jmla.1998.2564 Vellutino, F.R., Tunmer, W.E., Jaccard, J.J., & Chen, R. (2007).
National Early Literacy Panel (2008). Developing early literacy: Components of reading ability: Multivariate evidence for a con-
Report of the National Early Literacy Panel. Washington, DC: vergent skills model of reading development. Scientific Studies of
National Institute for Literacy. Reading, 11(1), 3–32. doi:10.1080/10888430709336632
Nelson, D.G.K., O’Neil, K.A., & Asher, Y.M. (2008). A mutually Verhallen, M., & Schoonen, R. (1993). Lexical knowledge of mono-
­facilitative relationship between learning names and learning con- lingual and bilingual children. Applied Linguistics, 14(4), 344–
cepts in preschool children: The case of artifacts. Journal of Cognition 363. doi:10.1093/applin/14.4.344
and Development, 9(2), 171–193. doi:10.1080/15248370802022621 Wasik, B.A., & Hindman, A.H. (2014). Understanding the active in-
Neuman, S.B., & Celano, D. (2006). The knowledge gap: Implications gredients in an effective preschool vocabulary intervention: An
of leveling the playing field for low-­income and middle-­income exploratory study of teacher and child talk during book reading.
children. Reading Research Quarterly, 41(2), 176–201. doi:10.1598/ Early Education and Development, 25(7), 1035–1056. doi:10.1080/
RRQ.41.2.2 10409289.2014.896064
Ordóñez, C.L., Carlo, M.S., Snow, C.E., & McLaughlin, B. (2002). Weisberg, D.S., Ilgaz, H., Hirsh-Pasek, K., Golinkoff, R., Nicolopoulou,
Depth and breadth of vocabulary in two languages: Which vo- A., & Dickinson, D.K. (2015). Shovels and swords: How realistic and
cabulary skills transfer? Journal of Educational Psychology, 94(4), fantastical themes affect children’s word learning. Cognitive
719–728. doi:10.1037/0022-­0663.94.4.719 Development, 35, 1–14. doi:10.1016/j.cogdev.2014.11.001
Ouellette, G.P. (2006). What’s meaning got to do with it: The role of Wiig, E.H., & Secord, W. (1992). Test of Word Knowledge. San
vocabulary in word reading and reading comprehension. Journal Antonio, TX: Psychological.
of Educational Psychology, 98(3), 554–566. doi:10.1037/0022-­0663 Yu, C., & Smith, L.B. (2012). Modeling cross-­situational word–refer-
.98.3.554 ent learning: Prior questions. Psychological Review, 119(1), 21–39.
Paivio, A., Yuille, J.C., & Madigan, S.A. (1968). Concreteness, doi:10.1037/a0026182
­imagery, and meaningfulness values for 925 nouns. Journal of
Experimental Psychology, 76(1, Pt. 2), 1–25. doi:10.1037/h0025327 Submitted March 31, 2015
Perfetti, C. (2007). Reading ability: Lexical quality to comprehen- Final revision received September 13, 2015
sion. Scientific Studies of Reading, 11(4), 357–383. doi:10.1080/ Accepted September 14, 2015
10888430701530730
Proctor, C.P., Silverman, R.D., Harring, J.R., & Montecillo, C. ELIZABETH B. HADLEY (corresponding author) is a doctoral
(2012). The role of vocabulary depth in predicting reading com- candidate in the Department of Teaching, Learning, and
prehension among English monolingual and Spanish-­ English Diversity at Vanderbilt University, Nashville, Tennessee, USA;
bilingual children in elementary school. Reading and Writing, e-­mail elizabeth.b.hadley@vanderbilt.edu.
25(7), 1635–1664. doi:10.1007/s11145-­011-­9336-­5
Quinn, J.M., Wagner, R.K., Petscher, Y., & Lopez, D. (2015). DAVID K. DICKINSON is the Margaret Cowan Chair of and
Developmental relations between vocabulary knowledge and a professor in the Department of Teaching, Learning, and
reading comprehension: A latent change score modeling study. Diversity at Vanderbilt University, Nashville, Tennessee, USA;
Child Development, 86(1), 159–175. doi:10.1111/cdev.12292 e-­mail david.dickinson@vanderbilt.edu.
Raudenbush, S., & Bryk, A. (2002). Hierarchical linear models:
Applications and data analysis methods (2nd ed.). Newbury Park, KATHY HIRSH-PASEK is the Debra and Stanley Lefkowitz
CA: Sage. Faculty Fellow in the Department of Psychology at Temple
Read, J. (2004). Plumbing the depths: How should the construct of
University, Philadelphia, Pennsylvania, USA; e-­mail
vocabulary knowledge be defined? In P. Bogaards, & B. Laufer
(Eds.), Vocabulary in a second language: Selection, acquisition, khirshpa@temple.edu.
and testing (pp. 209–227). Philadelphia, PA: John Benjamins.
ROBERTA MICHNICK GOLINKOFF is the Unidel H. Rodney
Roth, F.P., Speece, D.L., & Cooper, D.H. (2002). A longitudinal anal-
ysis of the connection between oral language and early reading. Sharp Professor in the Departments of Psychology and
The Journal of Educational Research, 95(5), 259–272. doi:10.1080/ Linguistics and Cognitive Science at the University of
00220670209596600 Delaware, Newark, USA; e-­mail roberta@udel.edu.
Schoonen, R., & Verhallen, M. (2008). The assessment of deep word
knowledge in young first and second language learners. Language KIMBERLY T. NESBITT is an assistant professor in the
Testing, 25(2), 211–236. doi:10.1177/0265532207086782 Department of Human Development and Family Studies at the
Snow, C.E., Cancino, H., De Temple, J., & Schley, S. (1991). University of New Hampshire, Durham, USA; e-­mail
Giving formal definitions: A linguistic or metalinguistic skill? kimberly.nesbitt@unh.edu.

196 | Reading Research Quarterly, 51(2)

 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
A PPENDIX A

Word Type, Difficulty, and Means for Target,

Exposure, and Control Words Tested on the
NWDT–M Measure
Living Word On the
Abstract Vocabulary Dale–Chall
or Level of difficulty common Tier Pretest M Posttest M
Word type Word concrete? instruction Theme ratinga word listb? 2c? (SD) (SD)
Adjective accidentally Abstract Exposure Dragon T2 Yes Yes 0.32 (0.51) 0.30 (0.43)
(or adverb)
intelligent Abstract Target Dragon T2 No Yes 0.02 (0.20) 0.21 (0.46)

peaceful Abstract Target Farm T2 Yes Yes 0.17 (0.47) 0.46 (0.66)

recent Abstract Control Farm T6 No Yes 0.00 (0.00) 0.07 (0.25)

wearily Abstract Target Farm T2 Yes Yes 0.02 (0.18) 0.08 (0.32)

Noun celebration Abstract Control Dragon E No Yes 0.64 (0.76) 0.81 (0.81)

chimney Concrete Target Farm E Yes Yes 0.33 (0.67) 0.87 (0.82)

curtain Concrete Control Farm Yes Yes 0.58 (0.71) 0.67 (0.80)

enemies Abstract Target Dragon E Yes Yes 0.14 (0.39) 0.44 (0.62)

field Concrete Exposure Farm E Yes Yes 0.46 (0.81) 0.58 (0.94)

foolishness Abstract Target Dragon No Yes 0.08 (0.28) 0.21 (0.45)

handkerchief Concrete Target Dragon Yes Yes 0.02 (0.20) 0.86 (1.07)

hedge Concrete Control Farm T6 No Yes 0.17 (0.42) 0.30 (0.64)

heel Concrete Control Dragon L2 Yes Yes 0.00 (0.00) 0.00 (0.00)

lane Concrete Target Farm E Yes Yes 0.07 (0.34) 0.48 (0.86)

nostrils Concrete Target Dragon No Yes 0.05 (0.31) 1.15 (1.21)

plan Abstract Exposure Farm E Yes Yes 0.23 (0.49) 0.25 (0.49)

pliers Concrete Control Dragon D No Yes 0.08 (0.33) 0.16 (0.49)

pond Concrete Exposure Dragon T2 Yes Yes 0.83 (1.03) 1.00 (1.02)

quarrel Abstract Target Farm T6 No Yes 0.00 (0.00) 0.27 (0.55)

quilt Concrete Exposure Dragon E Yes Yes 0.13 (0.41) 0.29 (0.74)

scales Concrete Target Dragon Yes Yes 0.05 (0.20) 0.52 (0.88)

servants Concrete Target Dragon Yes Yes 0.24 (0.58) 0.75 (1.07)

shield Concrete Exposure Dragon No Yes 0.45 (0.93) 0.60 (0.93)

stool Concrete Exposure Farm E Yes Yes 0.69 (0.98) 0.78 (0.90)

throne Concrete Target Dragon L2 Yes Yes 0.06 (0.37) 0.77 (1.10)

tip Concrete Exposure Farm T2 Yes Yes 0.09 (0.34) 0.13 (0.31)

valley Concrete Target Dragon E Yes Yes 0.02 (0.11) 0.57 (0.95)

weeds Concrete Target Farm L2 Yes Yes 0.22 (0.63) 0.76 (1.02)

wheelbarrow Concrete Exposure Farm No Yes 0.47 (0.88) 0.69 (0.98)

(continued)

Examining the Acquisition of Vocabulary Knowledge Depth Among Preschool Students | 197
 19362722, 2016, 2, Downloaded from https://ila.onlinelibrary.wiley.com/doi/10.1002/rrq.130 by <shibboleth>-member@monash.edu.au, Wiley Online Library on [20/08/2025]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License
(continued)

Living Word On the

Abstract Vocabulary Dale–Chall
or Level of difficulty common Tier Pretest M Posttest M
Word type Word concrete? instruction Theme ratinga word listb? 2c? (SD) (SD)
Verb charging Concrete Target Dragon E Yes Yes 0.04 (0.23) 0.43 (0.71)

chuckling Concrete Target Dragon E No Yes 0.05 (0.24) 0.62 (0.72)

emerging Concrete Target Dragon T2 No Yes 0.01 (0.10) 0.49 (0.81)

fetching Concrete Target Farm No Yes 0.39 (0.63) 0.67 (0.75)

galloping Concrete Target Dragon E Yes Yes 0.25 (0.56) 0.64 (0.77)

plummeting Concrete Control Farm No Yes 0.01 (0.04) 0.00 (0.00)

returning Abstract Target Farm E Yes Yes 0.34 (0.51) 0.52 (0.63)

scowling Concrete Control Dragon T6 No Yes 0.02 (0.13) 0.03 (0.16)

sobbing Concrete Exposure Farm T2 Yes Yes 0.01 (0.09) 0.11 (0.40)
Note. D = words known by fewer than 40% of students by the end of grade 6; E = words known by most students at the end of grade 2; L2 = low-­priority
words known by 40–80% of students by the end of grade 2; M = mean; SD = standard deviation; T2 = high-­priority words known by 40–80% of students
by the end of grade 2; T6 = words known by 40–80% of students by the end of grade 6.
a
Biemiller, A. (2010). Words worth teaching: Closing the vocabulary gap. Columbus, OH: SRA/McGraw-­Hill. bChall, J.S., & Dale, E. (1995). Readability
revisited: The new Dale–Chall readability formula. Cambridge, MA: Brookline. cBeck, I.L., & McKeown, M.G. (2007). Increasing young low-­income
children’s oral vocabulary repertoires through rich and focused instruction. The Elementary School Journal, 107(3), 251–271.

A PPENDIX B

Examples of Student Responses

and Codes Assigned
Word type Target word Student response Information unit(s) coded for
Concrete noun basket “You carry stuff with it.” Function

chimney “Made of bricks.” Part/whole

nostrils Points to nostrils and sniffs. Gesture

shield “[A] shield protects you when you get in a fight Function, meaningful context
with a dragon and he blows fire at you.”

throne “A throne is golden.” Perceptual quality

Verb chuckling “A quiet laugh.” Synonym

fetching “I throw the ball to my dog, and he fetches Meaningful context
it and gives it to me.”

returning “Run away and go back. In the story, the farmer Synonym, meaningful context
ran away, and he never returned.”

sobbing “You crying.” Synonym

Abstract noun foolishness “Foolishness means that you’re acting crazy.” Synonym

Adjective intelligent “Means that you could build a science fair project.” Basic context

198 | Reading Research Quarterly, 51(2)