Direct Vocabulary Instruction in Kindergarten: Teaching ...

University of Rhode IslandDigitalCommons@URI

Psychology Faculty Publications Psychology

2009

Direct Vocabulary Instruction in Kindergarten:Teaching for Breadth versus DepthMichael D. Coyne

D. Betsy McCoach

See next page for additional authors

Follow this and additional works at: https://digitalcommons.uri.edu/psy_facpubs

Terms of UseAll rights reserved under copyright.

This Article is brought to you for free and open access by the Psychology at DigitalCommons@URI. It has been accepted for inclusion in PsychologyFaculty Publications by an authorized administrator of DigitalCommons@URI. For more information, please contact [email protected].

Citation/Publisher AttributionCoyne, M. D., McCoach, D. B., Loftus, S., Zipoli, R., Jr. & Kapp, S. (2009). Direct Vocabulary Instruction in Kindergarten: Teachingfor Breadth, Versus Depth. Elementary School Journal, 110(1),1-48. doi: 10.1086/598840Available at: http://dx.doi.org/10.1086/598840

http://ww2.uri.edu/?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

http://ww2.uri.edu/?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.uri.edu?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.uri.edu/psy_facpubs?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.uri.edu/psy?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

https://digitalcommons.uri.edu/psy_facpubs?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

http://dx.doi.org/10.1086/598840

mailto:[email protected]

AuthorsMichael D. Coyne, D. Betsy McCoach, Susan Loftus, Richard Zipoli Jr., and Sharon Kapp

This article is available at DigitalCommons@URI: https://digitalcommons.uri.edu/psy_facpubs/1

https://digitalcommons.uri.edu/psy_facpubs/1?utm_source=digitalcommons.uri.edu%2Fpsy_facpubs%2F1&utm_medium=PDF&utm_campaign=PDFCoverPages

Direct Vocabulary Instruction in Kindergarten: Teaching for Breadth versus DepthAuthor(s): Michael D. Coyne, D. Betsy McCoach, Susan Loftus, Richard Zipoli Jr., andSharon KappSource: The Elementary School Journal, Vol. 110, No. 1 (September 2009), pp. 1-18Published by: The University of Chicago PressStable URL: http://www.jstor.org/stable/10.1086/598840 .

Accessed: 04/04/2013 14:18

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .http://www.jstor.org/page/info/about/policies/terms.jsp

.JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range ofcontent in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new formsof scholarship. For more information about JSTOR, please contact [email protected].

.

The University of Chicago Press is collaborating with JSTOR to digitize, preserve and extend access to TheElementary School Journal.

http://www.jstor.org

This content downloaded from 131.128.70.27 on Thu, 4 Apr 2013 14:18:49 PMAll use subject to JSTOR Terms and Conditions

http://www.jstor.org/action/showPublisher?publisherCode=ucpress

http://www.jstor.org/stable/10.1086/598840?origin=JSTOR-pdf

http://www.jstor.org/page/info/about/policies/terms.jsp


Direct VocabularyInstruction inKindergarten:Teaching for Breadthversus Depth

Michael D. CoyneD. Betsy McCoachUniversity of Connecticut

Susan LoftusUniversity of Rhode Island

Richard Zipoli Jr.West Hartford Public Schools, CT

Sharon KappMarlborough Public Schools, MA

Abstract

The purpose of this study was to compare 2methods for directly teaching word meanings tokindergarten students within storybook read-alouds that varied in instructional time anddepth of instruction along with a control condi-tion that provided students with incidental ex-posure to target words. Embedded instructionintroduces target word meanings during story-book readings in a time-efficient manner. Ex-tended instruction is more time intensive butprovides multiple opportunities to interact withtarget words outside the context of the story.Participants included 42 kindergarten studentswho were taught 9 target words, 3 with eachmethod. Target words were counterbalanced ina within-subjects design. Findings indicated thatextended instruction resulted in more full andrefined word knowledge, while embedded in-struction resulted in partial knowledge of targetvocabulary. Implications are discussed in rela-tion to the strengths and limitations of differentapproaches to direct vocabulary instruction inkindergarten and the trade-offs between instruc-tion that focuses on teaching for breadth versusdepth.

Children enter school with significant dif-ferences in vocabulary knowledge (Hart &Risley, 1995) and these differences growlarger in the early grades (Biemiller &Slonim, 2001). Converging evidence hassuggested that instruction in code-basedskills is insufficient to meet the needs ofstudents who are at risk for experiencingreading problems because of language andvocabulary difficulties (Catts, Hogan, &Adolf, 2005; Storch & Whitehurst, 2002;Vellutino, Tunmer, Jaccard, & Chen, 2007).Therefore, there is growing recognition ofthe importance of accelerating vocabularydevelopment in young school-age children

The Elementary School JournalVolume 110, Number 1© 2009 by The University of Chicago. All rights reserved.0013-5984/2009/11001-0001$10.00



through targeted and teacher-supported in-struction and other intervention efforts (Bi-emiller, 2001; Catts et al., 2005).

Researchers and practitioners interestedin accelerating academic achievement facea difficult challenge concerning how best toleverage scarce instructional time. In vo-cabulary intervention research, discussionsabout leveraging instructional time oftenrevolve around the trade-offs betweenteaching for breadth or depth. The purposeof this study was to evaluate two ap-proaches for supporting vocabulary learn-ing with kindergarten students in the con-text of storybook read-alouds. Specifically,we compared two methods for directlyteaching word meanings that varied in in-structional time and depth of instruction.We also compared these methods with acontrol condition that provided studentswith incidental exposure to target words.

Teaching for Breadth: EmbeddedVocabulary InstructionAt the beginning of kindergarten, typicallydeveloping students know thousands moreword meanings than their peers at risk oflanguage and learning difficulties (Hart &Risley, 1995). This vocabulary gap onlygrows larger in the primary grades. Forexample, Biemiller and Slonim (2001) esti-mated that by second grade, children withlarge vocabularies know approximately4,000 more root word meanings than chil-dren with delays in vocabulary develop-ment. Although it is impossible to teachdirectly all the words necessary to close thisgap (Anderson & Nagy, 1992), one goal ofdirect vocabulary instruction is to intro-duce students to as many new words aspossible.

Instructional approaches that allow theintroduction of many word meanings mostoften provide students with brief defini-tions of target words within the context oforal language experiences such as story read-alouds (e.g., Elley, 1989; Penno, Wilkinson, &Moore, 2002). This approach characterizes

the research of Biemiller and his colleagues.Biemiller and Boote (2006) conducted twostudies in which vocabulary learning amongprimary students improved when brief ex-planations of word meanings were providedduring repeated storybook readings. Knowl-edge of word meanings was tested by ratingverbal explanations of words presented incontext sentences. Children demonstrated a22% gain in instructed words compared to a12% gain for noninstructed words, indicatingthat explicit explanations resulted in a statis-tically significant increase in word learning.In a second study, two reviews of each wordmeaning were provided, including an oppor-tunity to review word meanings in new con-text sentences. Children in this study showedan average gain of 41%. Biemiller and Booteconcluded that teaching many word mean-ings without in-depth discussion appeared tobe an effective approach to direct vocabularyinstruction.

Embedded vocabulary instruction has anumber of benefits. First, it is time efficient.Providing brief definitions of target wordswithin the context of a story read-aloudtakes very little time, perhaps 1 minute perword (Coyne, McCoach, & Kapp, 2007).This allows the introduction of many wordmeanings during instructional time. Em-bedded vocabulary instruction can also beincorporated into story readings with min-imal disruption to the story flow. Finally,embedded instruction provides studentswith definitions within a meaningful andsupportive context (i.e., a story), two criti-cal features of effective vocabulary instruc-tion (National Reading Panel [NRP], 2000;Stahl & Fairbanks, 1986).

There are also limitations to embeddedvocabulary instruction, primarily related tothe minimal instructional time allocated forteaching each word. Research has indicatedthat multiple and repeated exposure toword meanings is a critical feature of effec-tive vocabulary instruction (NRP, 2000;Stahl & Fairbanks, 1986). Limited time re-stricts the number of exposures studentshave to target words in embedded vocab-

2 THE ELEMENTARY SCHOOL JOURNAL

SEPTEMBER 2009



ulary instruction. These exposures are alsoconstrained by the story context. In otherwords, because students are introduced tonew words within a fixed narrative, theydo not receive experiences with or expo-sure to words in varied contexts. Finally,with embedded instruction, students donot actively engage in learning tasks thatrequire them to discriminate, manipulate,and interact with word meanings.

Teaching for Depth: ExtendedVocabulary InstructionAlthough breadth (i.e., the number of wordmeanings in a student’s lexicon) is one im-portant dimension of vocabulary knowl-edge, depth (i.e., how well the studentknows those word meanings) is anothersignificant dimension. Knowledge of eachword meaning exists on a continuum fromno knowledge to varying levels of partialknowledge to more complete and fullknowledge (Nagy & Scott, 2000; Schwanen-flugel, Stahl, & McFalls, 1997). Depth ofword knowledge has important implica-tions for listening or reading comprehen-sion. How well, or deeply, a word is knowndetermines whether or not it can be dis-criminated from other words and understoodin novel contexts or in different morphosyntac-tic forms. Therefore, another critical goal of di-rect vocabulary instruction is to help studentsdevelop sufficient depth of word knowledge tosupport comprehension.

Instructional approaches that focus ondeveloping depth of vocabulary knowl-edge most often provide students with ex-tended opportunities to discuss and inter-act with words outside story readings. Thisapproach characterizes the work of Beck,McKeown, and their colleagues (Beck,McKeown, & Kucan, 2002). In the first oftwo studies, Beck and McKeown (2007) re-ported that oral vocabularies of kindergar-ten and first-grade students were enhancedby providing rich instruction following sto-rybook readings. Rich instruction includedexplanations of word meanings and the

presentation of selected words in the con-text of the original story and new contexts.Students were also asked to discriminateamong potential exemplars of word mean-ings and to offer their own examples ofword meanings. Instructed children were com-pared to children who had not received vocab-ulary instruction on a picture-recognition taskthat involved making interpretations in novelcontexts. Children who received rich instruc-tion demonstrated significantly higher gains intarget words. In a second study, instructed stu-dents learned twice as many words when pro-vided with multiple interactions with wordsacross several days.

A number of benefits are associated withextended vocabulary instruction. Because ex-tended instruction allocates more instruc-tional time per word, students receive moreencounters with and exposure to target vo-cabulary. Extended vocabulary instructionalso provides students with opportunities tointeract with words outside the narrativeconstraints of the story. This allows teachersto give students examples of how targetwords can be used in multiple and novelcontexts. Activities associated with extendedinstruction also focus on enabling students toengage in rich dialogic interactions aroundwords and word meanings. The deep andrefined knowledge of word meanings that isthe goal of extended instruction may bettersupport comprehension across varied con-texts (NRP, 2000; Stahl & Fairbanks, 1986).Finally, there is an implicit hypothesis thatextended instruction may lead to greaterword consciousness or metalinguistic aware-ness (Nagy, 2007). Because this instructionencourages deep processing of word mean-ings and challenges students to move beyondmemorizing simple dictionary definitions tounderstanding words at a richer, more com-plex level, students may become more at-tuned to novel words they encounter andbetter able to infer word meanings inciden-tally (Nagy, 2007; Nagy & Scott, 2000).

The primary limitation of extended vo-cabulary instruction is the amount of timerequired to teach each target word. Ideally,

VOCABULARY INSTRUCTION 3



teachers would have unlimited instructionaltime to spend teaching vocabulary. How-ever, in reality, teachers have limited time.Therefore, an extended approach to vocab-ulary instruction compels teachers to focuson teaching fewer words. Finally, becauseevidence to guide teachers in choosingwhich words to target for instruction is lim-ited, extended instruction forces teachers tomake difficult decisions about what wordsto teach (Coyne, Simmons, Kame’enui, &Stoolmiller, 2004).

Evaluating the Effects of DifferentApproaches to VocabularyInstructionStudies have evaluated the effects of eitherembedded or extended vocabulary instruc-tional approaches separately on youngschool-age children’s vocabulary learning(e.g., Beck & McKeown, 2007; Biemiller &Boote, 2006; Coyne et al., 2004). Few stud-ies, however, have directly compared theseapproaches with each other. We recentlyconducted a study in which we evaluatedthe effects of both types of instruction onthe word learning of 32 kindergarten stu-dents from an urban school in the North-east (Coyne et al., 2007). Children listenedto three readings of a storybook in smallgroups of four for a week. All studentsreceived embedded instruction on threewords from the storybook and extendedinstruction on three additional words in awithin-subjects experimental design. Em-bedded instruction included providingsimple definitions of target words withinthe context of the story. Extended instruc-tion included providing simple definitionsas well as extended activities after the storyreading. Target vocabulary words receiv-ing embedded and extended instructionwere counterbalanced across groups tocontrol for word effects. Measures admin-istered at posttest and a 6-week delayedposttest included experimenter-designedassessments of expressive and receptiveknowledge of target word definitions and a

receptive measure of target words used innovel contexts. Results indicated that ex-tended instruction resulted in greater wordlearning than embedded instruction in fiveof six comparisons and that these differ-ences were maintained on delayed post-tests.

Although these findings support the effi-cacy of extended instruction, a number ofimportant considerations must be addressedwhen directly comparing the relative effectsof embedded and extended vocabulary in-struction. In intervention studies, researchersmost often use inferential statistics to deter-mine if instruction produces statistically sig-nificant differences between treatment andcomparison groups. In this case, one wouldbe interested in whether the additional in-vestment in time and intensity associatedwith extended instruction would producegreater effects on vocabulary outcome mea-sures (Coyne et al., 2007). However, becauseextended and embedded instruction havedifferent goals (i.e., depth vs. breadth of vo-cabulary knowledge), statistical tests tell onlypart of the story. To get a full picture of therelative benefits of the two types of instruc-tion, researchers need to describe the amountand quality of word learning that occurwithin each approach.

One indicator of word learning is thenumber of target words learned. For exam-ple, if 10 target words were introduced,how many of those words did studentslearn with each approach? Although thisappears to be a straightforward question,the answer is highly dependent on howresearchers define what it means to know aword. For example, one might say that astudent knows a word when she can rec-ognize its definition. Alternately, one mightsay that a student knows a word only whenshe can produce its definition. Therefore, asecond indicator of word learning is qualityor depth of word learning.

Nagy and Scott (2000) suggested thatvocabulary learning is incremental. Ac-cording to this theoretical model, a stu-dent’s knowledge of a word’s meaning de-


SEPTEMBER 2009



velops incrementally from no knowledge,through varying levels of partial knowl-edge, to more full and complete knowl-edge. In other words, “knowing a word isnot an all-or-nothing proposition” (Beck etal., 2002, p. 9). Beck, McKeown, and Oman-son (1987) developed the following contin-uum of levels of word knowledge: (a) noknowledge; (b) general sense; (c) narrow,context-bound knowledge; (d) generalizedreceptive knowledge; and (e) rich, decon-textualized knowledge of a word’s mean-ing and its relation to other words. Similarlevels of word knowledge have been pro-posed by Calfee and Drum (1986) and Dale(1965). It is important to note, however, thatthese are only descriptive examples of whatdiffering levels of word knowledge mightlook like. In reality, connectionist models ofword learning suggest that the develop-ment of word knowledge is more likely tobe truly incremental, with knowledge of aword’s meaning becoming gradually morerefined with every new exposure to thatword (Landauer & Dumais, 1997).

In the current study, we were interestedin replicating and extending our previousfindings by attempting to better characterizethe amount and quality of word learning thatresulted from embedded and extended vo-cabulary instruction. We developed mea-surement tasks that would capture differentlevels of partial word knowledge. In thisway, we believed that we would be betterable to describe the breadth and depth ofvocabulary learning.

Responsiveness to VocabularyInterventionResearch on shared storybook reading hassuggested that students who are at risk forlanguage and literacy difficulties and havesmaller initial vocabularies are less likelythan their peers with larger vocabularies tolearn words incidentally while listening tostories (Coyne et al., 2004; Nicholson &Whyte, 1992; Robbins & Ehri, 1994; Senechal,Thomas, & Monker, 1995). Research on direct

vocabulary instruction with young studentshas produced inconsistent findings on the re-lation between individual differences and re-sponsiveness to vocabulary interventions.Although a few studies have showed that allstudents responded similarly to instruction(Coyne et al., 2004; Elley, 1989), other re-searchers have found that at-risk students areless responsive to vocabulary instruction(Coyne et al., 2007; Penno et al., 2002).

We were interested in whether studentsmost at risk for language and reading disabil-ities responded differentially to embeddedand extended vocabulary instruction. Overallreceptive vocabulary knowledge is a strongpredictor of language and comprehensionoutcomes (Scarborough, 2005; Storch &Whitehurst, 2002). Therefore, we used stu-dents’ pretest Peabody Picture VocabularyTest (PPVT; Dunn & Dunn, 1997) scores as anindicator of risk.

Purpose of the StudyThe purpose of this study was to compare theeffects of two interventions that differed intheir approach to direct vocabulary instruc-tion for young students. The goal of embed-ded instruction is to promote breadth of vo-cabulary knowledge by introducing targetword meanings during storybook readingsin a time-efficient manner. The goal of ex-tended instruction is to promote depth of vo-cabulary learning by allocating more instruc-tional time per target word so that studentscan have multiple opportunities to interactwith words outside the context of the story.We compared these approaches both statisti-cally and descriptively to better characterizethe amount and quality of word learning thatresulted from each. We also compared bothapproaches with a control condition consist-ing of incidental exposure to target wordswithin the story reading. Our goal was toprovide data so that researchers and practi-tioners can make more informed decisionsabout the benefits associated with differentapproaches to vocabulary instruction, as wellas the trade-offs in instructional time and in-




tensity that are required of each. Addition-ally, we were also interested in whether stu-dents maintained vocabulary knowledgegained through instruction at an 8-week de-layed posttest. Finally, we examined whetherstudents’ responsiveness to vocabulary in-struction was moderated by general recep-tive vocabulary knowledge measured at pre-test.

MethodParticipantsParticipants in this study included 42

kindergarten students enrolled in a PreK–8elementary school in a large city in theNortheast. Approximately 69% of the stu-dents in the school were Hispanic, 24%were black, and 6% were white. Approxi-mately 65% were eligible for free orreduced-price lunch. On the 2005 state mas-tery test, 8% of fourth graders met the stategoal in reading compared to 53.8% of stu-dents statewide.

Consent forms were sent home to par-ents of all 58 students in the three kinder-garten classrooms. Fifty consent formswere signed and returned to the teachers.Because of space constraints, we randomlyselected 45 students from this pool to takepart in the study. Three students were ab-sent for two of three intervention sessionsand were therefore not included in theanalyses. The 42 participants included 23females and 19 males; 12 students wereHispanic, 8 were African American, 3 wereCaucasian, and 1 student was Asian. Theaverage age of participants was 5 years,4 months and ranged from 4 years, 10months to 6 years, 1 month.

DesignWe used an experimental design with

two within-subjects factors: type of instruc-tion (story words taught with extended in-struction, story words taught with embed-ded instruction, and story words receivingincidental exposure) and time (posttest and

delayed posttest). In a within-subjects de-sign, each student receives each instruc-tional condition and serves as her own con-trol. There are a number of benefits of usingthis method to investigate vocabulary in-struction. First, because each student re-ceived all three instructional procedures,we were able control for between-subjectvariability, thus increasing our power con-siderably. Second, in vocabulary instruc-tion, directly teaching the meaning of oneword does not immediately affect learningthe meanings of other words, especiallyduring short, focused intervention studies.Therefore, carryover effects that often makewithin-subjects designs unfeasible in edu-cational research are eliminated. Third,within-subjects designs control for almostall threats to internal validity, making thismethod particularly rigorous and trustwor-thy. A number of research teams have usedwithin-subjects designs successfully in re-search on vocabulary instruction and inter-vention (e.g., Beck & McKeown, 2007; Bi-emiller & Boote, 2006), and we have usedthis method in our previous research(Coyne et al., 2007).

Instructional ConditionsWe developed the intervention around

the storybook Goldilocks by James Marshall(1998). We selected nine target words usingthe following procedures. First, we identi-fied words in the story that appeared onlyonce and were located near pictures thatillustrated their meanings. We then re-placed these words with low-frequencysynonyms that kindergartners would beunlikely to know. Our goal in selecting tar-get words was twofold. First, we wantedtarget words that were unfamiliar to stu-dents but whose meanings they would beable to understand. These are the featuresof words that Beck et al. (2002) character-ized as Tier II words—words they recom-mend for teaching directly to primary-grade students. Second, to assess thestrength of the intervention accurately, we


SEPTEMBER 2009



wanted to ensure that students had no ex-perience with any of the target words.Therefore, we selected rare words of verylow frequency. The nine target words in-cluded six nouns (weald, duvet, domicile, par-lor, lass, and shards) and three adjectives(torrid, fatigued, and dismayed).

It is important to note that, although ourwords were selected to approximate thekinds of Tier II words that would be tar-geted for instruction in kindergarten class-rooms, we do not suggest the target wordswe used be taught to kindergarten stu-dents, because of their low frequency. In-stead, we recommend that teachers selecthigher-frequency Tier II words that stu-dents would be likely to encounter often inacademic discourse (Beck et al., 2002).

We developed three instructional con-ditions, two intervention conditions (i.e.,words taught with embedded instructionor extended instruction) and a comparisoncondition (i.e., words receiving only inci-dental exposure). We created three versionsof the intervention in which target wordswere counterbalanced across conditions.Each version included three words thatwere taught using extended instruction,three words taught using embedded in-struction, and three words receiving inci-dental exposure. Target words appeared inthe story in a random order to control fororder effects. Each group of three wordsincluded two nouns and one adjective.Children were assigned randomly to ver-sion A, version B, or version C of the inter-vention, and instructional procedures wereidentical across the three versions.

All students listened to three readingsof the story and were exposed to all ninetarget words during each reading. Studentsin version A were taught weald, torrid, andduvet through extended instruction; domi-cile, parlor, and fatigued through embeddedinstruction; and lass, shards, and dismayedthrough incidental exposure. Students inversion B were taught domicile, parlor, andfatigued through extended instruction; lass,shards, and dismayed through embedded in-

struction; and weald, torrid, and duvetthrough incidental exposure. Students inversion C were taught lass, shards, and dis-mayed through extended instruction; weald,torrid, and duvet through embedded in-struction; and domicile, parlor, and fatiguedthrough incidental exposure.

The intervention was delivered in smallgroups of three to four students by fourgraduate students who were members ofthe research team. Each graduate studenttaught one group receiving version A of theintervention, one group receiving versionB, and one group receiving version C. Theintervention was delivered in three ses-sions over the course of 1 week. The story-book was read during each session. Eachsession lasted approximately 30 minutes,with 15 minutes dedicated to reading thestorybook and introducing target wordsand 15 minutes dedicated to postreadingvocabulary activities.

Prior to the intervention, graduate stu-dents were trained to implement it. Trainingfocused on fidelity of implementation andthe standardization of administration acrossinterventionists. Procedures for maintainingstudent attention and eliciting student re-sponses were explained. Trainers modeledimplementation of the intervention and pro-vided graduate students opportunities topractice administration with corrective feed-back.

Embedded instruction. Graduate stu-dents introduced words receiving embed-ded instruction prior to the storybook read-ing and instructed students to repeat eachtarget word aloud. Students were told tolisten for the target words and raise theirhands when they heard the words in thestory. When a word was encountered in thestorybook, the interventionist asked stu-dents to identify the target word and thenreread the sentence containing the word (e.g.,“Oh, good. Some of you raised your hands!What word did you hear? Yes, lass. ‘Oncethere was a lass named Goldilocks.’ ”). Stu-dents were then provided with a simple def-inition of the word (e.g., “A lass is a little




girl.”). Next, the graduate student reread thesentence and replaced the target word withits definition (e.g., “Now I’ll say the sentenceagain with the words that mean lass. ‘Oncethere was a little girl named Goldilocks.’ ”).The interventionist then showed the studentsthe supportive picture that accompanied thetarget word (e.g., “In the picture we can seethat Goldilocks is a lass, or a little girl.”).Finally, students were prompted to pro-nounce the target word together once again(e.g., “Everyone say lass.”). These procedureswere repeated for each reading of the story-book, so that each target word receiving em-bedded instruction was introduced and de-fined a total of three times. Teaching wordswith embedded instruction took approxi-mately 30 seconds per word over the threereadings. Procedures for words taught withembedded instruction were similar to thosein other studies that provided this type ofinstruction (Biemiller & Boote, 2006; Justice,Meier, & Walpole, 2005; Penno et al., 2002).

Extended instruction. Words receivingextended instruction were introduced priorto the storybook reading and defined dur-ing the storybook reading using the sameprocedures as those used for words in theembedded instruction condition. In ex-tended instruction, however, graduate stu-dents led students in activities followingthe storybook reading that allowed them toexperience the target words in differentcontexts. Each activity began by reintro-ducing the target word using the context ofthe story and the supportive picture. Oncethe word was reviewed, additional exam-ples of the word’s usage were provided(e.g., “Other things could also be torrid. If apan on the stove got very hot, the panwould be torrid. A fire is also torrid, or veryhot.”).

During the first two intervention sessions,students participated in three activities, onefor each word receiving extended instruction.Activities required both group and individ-ual student responses. In one activity, stu-dents were shown various pictures and wereasked to determine whether or not the pic-

ture was a positive example of the targetword. Other activities required students torespond to yes or no and open-ended ques-tions regarding the target words. Examplesof yes or no questions include, “Would woodburning in a fire be torrid?” and, “Is the in-side of a refrigerator torrid?” Examples ofopen-ended questions include, “After walk-ing and walking, she finally reached the topof the tall building. Why was she fatigued?”and, “We built a snowman in the parlor.What is silly about that sentence?” Graduatestudents followed up children’s responses inall activities by either confirming the stu-dent’s response (e.g., “Yes! A pizza would betorrid, or hot, if it just came out of the oven.”)or asking the student to justify his or herresponse (e.g., “Why is that picture an exam-ple of a duvet?”).

During the third intervention session,students participated in activities that com-bined the three target words receiving ex-tended instruction. For example, studentswere asked open-ended questions contain-ing two target words, such as, “Would youneed a duvet if it was torrid in yourhouse?” Student responses were followedup with confirmation or correction from theinterventionist. Students were also givenprompts to extend or expand their re-sponses when correct in order to encouragethem to demonstrate their understandingof both target words. Providing extendedinstruction took approximately 15 minutesper word over the three readings (5 min-utes per word per reading).

Incidental exposure. Words in the inci-dental exposure condition appeared in thestory but were not taught directly. Studentsheard these words in the context of the storyonce per storybook reading. Interventionistsdid not directly discuss these words at anytime.

Fidelity of ImplementationCritical components of the interventions

were identified and an observation check-list was developed to document and eval-


SEPTEMBER 2009



uate fidelity of implementation (Gersten,Baker, & Lloyd, 2000; Gresham, MacMillan,Beebe-Frankenberger, & Bocian, 2000). Ex-amples of critical components incorporatedinto the fidelity checklist included whetheror not graduate students delivered each in-structional element, modeled proceduresappropriately, maximized opportunities torespond, provided error correction, andread storybooks with enthusiasm. Theproject director observed each interven-tionist during one of the three instructionalsessions. Fidelity of implementation aver-aged above 90% for each graduate student.

MeasuresThe National Reading Panel (2000) con-

cluded that specific vocabulary growth is bestassessed through researcher-developed mea-sures because they are more sensitive to gainsachieved through instruction than are stan-dardized tests. For this study, we developedmeasures that assessed children’s knowledgeof the nine words targeted in the interventionconditions. Our goal was to develop assess-ment tasks that would capture different levelsof word knowledge. We used a standardizedinstrument, the Peabody Picture VocabularyTest (PPVT), to assess general vocabularyknowledge. All instruments were administeredindividually.

Expressive target word definitions(expressive definitions). The expressive-definitions measure assessed expressiveknowledge of the nine target word defi-nitions. Students were asked to provide adefinition for each target word. For exam-ple, for the word domicile, students wereasked, “What does the word domicilemean?” Students’ responses were re-corded verbatim. If a student failed torespond after 5 seconds, or replied “Idon’t know,” the administrator asked afollow-up question, “Can you tell meanything about the word domicile?” Chil-dren were given 2 points for a completeresponse (e.g., “A domicile is a house”), 1point for a partial or related response

(e.g., “A domicile is inside”), and 0 pointsfor an unrelated response, or no response.The maximum score a student could re-ceive for each condition was 6 (i.e., 2points possible for each of three words).

Target words used in novel contexts/fullknowledge (context/full knowledge). Thecontext/full knowledge measure assessed re-ceptive knowledge of the target words pro-vided in neutral contexts. This test was de-signed to assess high levels of target wordknowledge by requiring children to makefiner discriminations about word meanings.Students were asked questions that requireda yes or no answer (Beck & McKeown, 2007).Each target word was represented by fourquestions—two positive examples of theword’s usage and two negative examples.For example, for dismayed, the positive exam-ples were, “If you lost your toy, would yoube dismayed?” and “Would you be dis-mayed if you got in trouble for somethingthat you did not do?” The two negative ex-amples were, “If your mother were dis-mayed, would she be smiling?” and “If yourmom made your favorite dinner, would yoube dismayed?” The four questions for eachtarget word were separated and distributedacross the measure. Students received 1 pointfor each correct answer and 0 points for eachincorrect answer. Each student received aseparate total score for target words intro-duced within each instructional condition(i.e., embedded, extended, incidental expo-sure). The maximum score a student couldreceive for each condition was 12 (i.e., 4points possible for each of three words).

Target words used in novel contexts/partial knowledge (context/partial knowl-edge). The partial-knowledge measure as-sessed receptive knowledge of the targetwords provided in supportive contexts.This test was designed to detect low targetword knowledge. Items were developed sothat students could correctly answer ques-tions without full knowledge of the wordmeanings. Children were asked to respondto questions that required a yes or no an-swer (Beck & McKeown, 2007). Each target




word was represented by four questions—two that required a positive response andtwo that required a negative response. Forexample, for the word parlor, questions thatrequired a positive response were, “Haveyou ever been in a parlor?” and “Couldsomeone have a parlor in their house?”Questions that required a negative re-sponse were, “Could you put a parlor in abag?” and “Could you find a parlor in yourkitchen?” The four questions for each of thenine target words were separated and dis-tributed across the measure. Students re-ceived 1 point for each correct answer and0 points for each incorrect answer. Themaximum score a student could receive foreach condition was 12 (i.e., 4 points possi-ble for each of three words).

Receptive target word definitions (recep-tive definitions). The receptive-definitionsmeasure assessed receptive knowledge of thenine target word definitions. Students wereasked questions that required a yes or noanswer (Beck & McKeown, 2007). Each targetword was represented by two questions—one that corresponded to the correctdefinition and one that corresponded to anincorrect definition. For example, the twoquestions for the target word duvet were, “Isa duvet a warm blanket?” and “Is a duvet afast car?” Children received 1 point for eachcorrect answer and 0 points for each incorrectanswer. The two questions for each of thenine target words were separated and dis-tributed across the measure. Each student re-ceived a separate total score for target wordsintroduced within each instructional condi-tion (i.e., embedded, extended, incidental ex-posure). The maximum score a student couldreceive for each condition was 6 (i.e., 2 pointspossible for each of three words).

Peabody Picture Vocabulary Test—III(PPVT). The PPVT is a norm-referencedmeasure of receptive vocabulary. We used itto characterize children’s overall receptivevocabulary knowledge prior to the interven-tion. Students are presented with four pic-tures and are asked to point to the picturethat best represents the word given by the

examiner. Standardized scores (mean � 100;SD � 15) are computed based on number ofitems correct and the student’s chronologicalage. Reliability of the PPVT is satisfactory,with alternate-forms reliability coefficientsranging from .88 to .96 and test-retest reliabil-ity coefficients ranging from .91 to .94. Over-all, research had suggested high reliabilityand validity for the PPVT (Salvia & Ys-seldyke, 1998).

Data Collection and ScoringData collection took place at pretest, post-

test, and delayed posttest. Members of theresearch team administered all assessmentsindividually to students in a quiet location.We collected pretest data approximately 1week prior to the start of the intervention.Measures administered at pretest includedthe PPVT and the context/full knowledgeand context/partial knowledge measures.We collected posttest data between 1 and 5days after the third reading of the storybook;these data included the expressive defini-tions, context/full knowledge, context/par-tial knowledge, and receptive definitionsmeasures. The delayed posttest was admin-istered 8 weeks after posttest and includedthe expressive and receptive definitions mea-sures. We administered measures in the or-der they were described. We collected post-test assessments in one 30-minute session.Pretest and delayed posttest assessmentswere each collected in one 20-minute session.

Data collectors were required to dem-onstrate at least 90% reliability for admin-istration. All measures were scored by onemember of the research team. The projectdirector randomly selected and indepen-dently scored 20% of the assessment proto-cols to check for scoring reliability. Agree-ment was 100%.

Results

We administered both the context/fullknowledge and context/partial knowl-edge assessments at pretest. Students’ to-


SEPTEMBER 2009



tal scores on the preassessments indicatedthat they were not performing above thechance level on either the context/fullknowledge assessment (mean � 17.97 outof 36; t � .057, p � .96) or the context/partial knowledge subtest (mean � 18.46out of 36; t � 1.51, p � .14). Scores at thechance level would indicate that studentsperformed no better than guessing on theyes/no questions. We also compared stu-dents’ scores on the full and partialknowledge subtests by condition usingrepeated-measures analysis of variance(ANOVA). These two analyses indicatedthat the pretest scores on the full knowl-edge and partial knowledge assessmentswere similar across conditions ( p � .16for full knowledge; p � .32 for partialknowledge). Scores on the pretest mea-sures indicated that students had no mea-surable knowledge of target word mean-ings prior to receiving the intervention andthat there were no differences between stu-dents’ knowledge of words across instruc-tional conditions. Participants’ mean PPVTscore measured at pretest was 95.24 (SD �9.73).

Comparison of InstructionalConditionsTo determine whether there were statis-

tically significant differences among the

means of words taught in the three condi-tions and whether these differences per-sisted over time, we conducted a series ofrepeated-measures ANOVAs on the mea-sures we collected at posttest and delayedposttest (receptive and expressive defini-tions). The two within-subjects variableswere time (posttest and delayed posttest)and instructional condition (words taughtwith extended instruction, words taughtwith embedded instruction, and words re-ceiving incidental exposure). Descriptivestatistics for these analyses are reported inTable 1. Cohen’s d effect sizes for the com-parisons between the instructional condi-tions are reported in Table 2.

For the expressive-definitions measure,the main effects of word learning condition(Wilks’s � � .284; F � 45.42, p � .001), andtime (Wilks’s � � .467; F � 42.25, p � .001),and the interaction between time and con-dition (Wilks’s � � .518; F � 16.76, p � .001)were significant. The linear trend was sig-nificant, indicating that the mean score ofwords taught with extended instructionwas higher than that for embedded instruc-tion, which in turn was higher than that forincidental exposure at both posttest and de-layed posttest. However, the interaction re-vealed that students also lost the greatestamount of knowledge between posttestand delayed posttest for words taught with

TABLE 1. Means and Standard Deviations at Posttest and Delayed Posttest, by Condition

Measure

ExtendedInstruction

EmbeddedInstruction

IncidentalExposure

M SD M SD M SD

Expressive definitions:a

Posttest 3.79 2.06 1.47 1.74 .24 .63Delayed posttest 1.82 1.78 1.11 1.48 .24 .88

Receptive definitions:a

Posttest 3.97* 1.38 3.58* 1.00 3.03 1.06Delayed posttest 4.06* 1.37 3.61* 1.08 3.08 1.08

Context/full knowledgeb 7.54* 1.79 6.34 1.62 5.98 1.44Contest/partial knowledgeb 7.85* 1.75 6.90* 1.62 6.15 1.42

aRange � 0 to 6.bRange � 0 to 12.*Scores above chance level (p � .0033).




extended instruction when compared to theother two conditions.

For the receptive-definition measure,the effect of word learning condition wassignificant (Wilks’s � � .695; F � 7.478, p �.002). The effects of time (Wilks’s � � .994;F � .198, p � .66) and the interaction be-tween time and condition (Wilks’s � � .999,F � .022, p � .98) were not significant.Therefore, there were differences in vocab-ulary scores among the word learning con-ditions, and these differences were main-tained over time. The linear trend wassignificant, indicating that extended in-struction produced higher scores than em-bedded instruction, which led to better per-formance than incidental exposure.

Next, we conducted a series of one-wayrepeated-measures ANOVAs on the mea-sures we collected at posttest only (con-text/full knowledge and context/partialknowledge). At posttest, there were signif-icant differences among the three condi-tions on the context/full knowledge mea-sure (Wilks’s � � .683, p � .001). The lineartrend was statistically significant (p � .001),indicating that students’ scores were higherfor words taught with extended instruc-tion, followed by words taught with em-bedded instruction, followed by words re-ceiving incidental exposure. There werealso significant differences among the threeconditions on the context/partial knowl-edge measure (Wilks’s � � .66, p � .001),and the linear trend was significant (p �

.001), showing the same trend as for thecontext/full knowledge measure.

Finally, we conducted a series of one-sample t-tests to determine whether stu-dents scored above the chance level on thedichotomous yes/no measures of wordlearning (receptive definitions, context/fullknowledge, and context/partial knowl-edge). Scores at the chance level would in-dicate that students performed no betterthan guessing on the yes/no questions. Weused a Bonferroni adjustment (.05/15) toaccount for the large number of statisticaltests being conducted. Therefore, probabil-ity levels below .0033 were considered sig-nificant. Scores above the chance level areindicated with an asterisk in Table 1.

For the receptive-definitions measureadministered at posttest, students scoredabove the chance level on words that weretaught using extended (t � 4.96, p � .001)and embedded (t � 3.81, p � .001) instruc-tion. However, students scored at thechance level on words that received inci-dental exposure (t � �.44, p � .66). Resultswere similar for the receptive measure atdelayed posttest. Again, students per-formed above the chance level on wordstaught using extended (t � 5.02, p � .001)and embedded (t � 3.6, p � .001) instruc-tion. However, they scored at the chancelevel on words that received incidental ex-posure (t � .453, p � .65).

For the context/full knowledge mea-sure, children scored above the chance

TABLE 2. Cohen’s d Effect Sizes for Comparisons between Instructional Conditions at Posttestand Delayed Posttest

MeasureExtended vs.Embedded

Extended vs.Incidental

Embedded vs.Incidental

Expressive definitions:Posttest 1.34** 2.57** .87**Delayed posttest .44* 1.18** .72**

Receptive definitions:Posttest .70** .97** .24*Delayed posttest .33 .84** .52

Context/full knowledge .38** .91** .63**Contest/partial knowledge .56** 1.07** .49*

*p � .05.**p � .01.


SEPTEMBER 2009



level on words taught using extended in-struction (t � 5.50, p � .001). However, theyperformed at the chance level on wordsthat received embedded instruction (t �1.35, p � .19) and incidental exposure (t ��.11, p � .91). For the context/partialknowledge measure, students scored abovethe chance level on words that were taughtusing extended (t � 6.77, p � .001) andembedded (t � 3.56, p � .001) instruction.However, they performed at the chancelevel on words that received incidental ex-posure (t � �.66, p � .51).

Characterizing Word LearningWe were also interested in describing

the effects of the instructional conditions onbreadth and depth of vocabulary learning.Breadth refers to how many words, on av-erage, students learned in each condition,and depth indicates how well they learnedthose words.

To characterize breadth of word knowl-edge, we examined the mean scores on eachmeasure of target word learning and esti-mated how many words were learned basedon the criteria for that assessment. For exam-ple, a mean score on the expressive-definitions measure of 6 points out of 6 forwords from one of the three approacheswould indicate that students had learned allthree words taught in that condition (i.e.,they produced complete 2-point definitionsfor each of the three target words). A meanscore of 4 would suggest that childrenlearned the equivalent of two of the threewords. Finally, a mean of less than 2 wouldindicate that students, on average, could notproduce a complete definition of any word.

For the full and partial knowledge con-text measures, there were four yes/noquestions for each word. Therefore, 12would be a perfect score and indicate thatstudents learned all three of the wordstaught in that condition according to thecriterion of these measures. A score of 8would mean that children learned theequivalent of two of the three words (i.e.,

four correct yes/no questions for each oftwo words). If analyses showed that themean score was statistically no differentfrom chance, this would suggest that stu-dents, on average, did not learn any ofthe words. On the receptive-definitionsmeasure, a mean score of 6 out of 6 wouldmean that students correctly answered twoyes/no questions about each of the threetarget words in that condition. In contrast,a score of 3 on these measures would indi-cate that students performed no better thanchance.

To characterize depth of word learning,we examined mean scores across measuresof target word learning that assessed levelsof partial word knowledge. Table 3 sum-marizes our characterization of amountand quality of word learning within thethree conditions across each measure. Tosimplify interpretation, we rounded offmean scores on each measure before esti-mating the number of words learned.

Responsiveness to InstructionThe PPVT scores predicted responsive-

ness to instruction on receptive and expres-sive target word measures for both theextended and embedded instructional condi-tions. Students with higher initial PPVTscores demonstrated greater word learningacross the two instructional conditions andthe two instructional measures. The relationbetween PPVT scores and students’ scores onthe expressive-definitions measure was mod-erately strong for both words taught withextended instruction (r � .49) and wordstaught with embedded instruction (r � .41).We tested the difference between the twodependent correlation coefficients for signifi-cance using the formula Glass and Hopkins(1996) recommended. This difference was notsignificant (t � .57, p � .10), suggesting thatthe relation between PPVT and expressive-definitions scores did not differ across thetwo instructional conditions. The relation be-tween PPVT and scores on the receptive-definition measure was also moderately




strong for both the extended (r � .45) andembedded (r � .49) conditions. This differ-ence was not significant (t � .35, p � .10),indicating that the relation between PPVTand receptive-definition scores also did notdiffer across the two instructional conditions.

DiscussionComparison of InstructionalConditionsOur results indicated that there were

statistically significant differences at post-test favoring words taught with extendedand embedded instruction over words re-ceiving only incidental exposure duringstory reading on all measures. These find-ings are consistent with a growing body ofresearch documenting the efficacy of di-rectly teaching word meanings to youngstudents within oral language activitiessuch as storybook readings (Elley, 1989;Justice et al., 2005; Penno et al., 2002; Walsh& Blewitt, 2006). Moderate to large effectsizes for these comparisons indicate thatdirect instruction of vocabulary results inreliably greater word learning in kinder-garten students than does incidental expo-sure by itself.

Even with the short duration of thestudy, we found significant differencesacross all measures at posttest between

words taught with extended and embed-ded instruction. Although few studies havecompared these types of instruction, ourresults converge with other findings sug-gesting that increasing instructional timeand providing more exposure to target vo-cabulary in varied contexts lead to en-hanced word learning (Beck & McKeown,2007; Biemiller & Boote, 2006; Coyne et al.,2007).

Breadth and Depth of Word LearningAlthough the results of statistical tests

are important, we were particularly inter-ested in describing the effects of embeddedand extended instruction. We wanted toexamine the relative benefits related to theamount and quality of word learning thatoccurred within each approach. We charac-terized breadth by examining mean scoreson each measure and estimating, on aver-age, the number of words students in eachof the three conditions learned. We de-scribed depth by looking at students’ meanscores across measures that required vary-ing levels of word knowledge to answerassessment items correctly.

Results suggested that, with embeddedinstruction, students demonstrated mea-surable word learning on approximatelytwo-thirds of the words introduced. How-

TABLE 3. Breadth of Word Learning across Measures, by Condition

Condition/Time per WordExpressiveDefinitions

Context/FullKnowledge

Context/PartialKnowledge

ReceptiveDefinitions

Incidental (30 seconds):Words learneda 0 0 0 0Mean score .24 5.98* 6.15* 3.03*Perfect score 6.00 12.00 12.00 6.00

Embedded (3 minutes):Words learneda 0 0 2 2Mean score 1.47 6.34* 6.90 3.58Perfect score 6.00 12.00 12.00 6.00

Extended (15 minutes):Words learneda 2 2 2 2Mean score 3.79 7.54 7.84 3.97Perfect score 6.00 12.00 12.00 6.00

aStudents could potentially learn three words for each measure. Number of words learned was estimated.*Scores below chance level.


SEPTEMBER 2009



ever, word learning was evident on onlytwo of our four measures, suggesting thatstudents possessed relatively little knowl-edge of those words. For example, studentsrecognized correct and incorrect definitionsof target words and answered yes/no ques-tions that required low levels of wordknowledge. However, they were unable toproduce definitions of target words or re-spond to yes/no questions that requiredadditional word knowledge. For wordstaught with extended instruction, studentsalso learned approximately two out ofthree words. In contrast to embedded in-struction, however, this finding was consis-tent across all measures, suggesting thatstudents developed more complete knowl-edge of these words.

These descriptive findings help to illu-minate both the strengths and limitations ofdifferent approaches to direct vocabularyinstruction. Embedded instruction (i.e.,providing brief definitions of words withinthe context of a story) is time efficient (i.e.,30–60 seconds per word per story reading)and allows the introduction of many wordmeanings within limited instructional time.However, word learning appears to be lim-ited to partial knowledge of word mean-ings. Extended instruction (i.e., providingopportunities to discuss and interact withwords outside of story reading) requiresmore instructional time (i.e., 5 minutes perword per story reading) but provides mul-tiple exposures to target words in variedcontexts that result in more full and refinedknowledge of these words.

Durability of Word LearningWe administered a delayed posttest 8

weeks after the conclusion of the interven-tion to examine the durability of wordlearning without planned review or rein-forcement. We selected low-frequency vo-cabulary words that kindergarten studentswould not know and that were unlikely tobe encountered incidentally during school.We also confirmed with classroom teachers

that target words were not discussed inclass between the posttest and the delayedposttest.

Results indicated that there was no effectof time on the receptive-definitions mea-sures, suggesting that students’ ability to rec-ognize the meanings of target words did notdecrease between posttest and delayed post-test. On the expressive-definitions measures,however, there was a significant interactionbetween time and instructional condition in-dicating that students’ ability to produce def-initions of target words taught with extendedinstruction decreased differentially com-pared to words taught with embedded in-struction and incidental exposure. Scores onthe expressive-definitions measure for wordstaught using embedded instruction did notdecrease between posttest and delayed post-test. These findings are consistent with ourprevious research and suggest that wordlearning is robust over time but that higherand more complete levels of word knowl-edge may be more susceptible to deteriora-tion, at least initially (Coyne et al., 2007). Inour study, students did not receive formalreview or even encounter target words inci-dentally between posttest and delayed post-test. Word knowledge may not become per-manent or fully established unless studentsreceive continued encounters with targetwords over time or are provided withthoughtful and systematic review.

Responsiveness to VocabularyInstructionIn our previous research, we found that

individual differences in initial receptive vo-cabulary knowledge strongly predicted re-sponsiveness to extended vocabulary instruc-tion (Coyne et al., 2007). We replicated thisfinding in the current study where students’pretest PPVT scores were highly correlatedwith target vocabulary outcome measures.We found a similar relation between initialvocabulary knowledge and response to em-bedded instruction. These findings suggestthat students most at risk of language and




literacy difficulties are less responsive to di-rect vocabulary instruction than their typi-cally developing peers (Penno et al., 2002).

Implications for PracticeOur findings underscore the trade-offs

between different approaches to direct vo-cabulary instruction and reinforce the im-portance of carefully considering learninggoals when making decisions about vocab-ulary instruction for young students. Forexample, embedded instruction may helpstudents establish initial lexical representa-tions of a large number of new vocabularywords. However, our results suggest thatthe limited exposure to target words asso-ciated with this approach is unlikely to re-sult in deeper word knowledge. Therefore,if the goal is to support students in devel-oping fuller and more complete wordknowledge, extended instruction would bemore effective, although it requires moreinstructional time and limits the number ofwords that can be introduced.

We believe there is a place for both em-bedded and extended vocabulary instructionin primary classrooms. We have describedthis as a tri-level approach to vocabulary in-struction (Coyne et al., 2007). It includes (a)reading storybooks to children that containvaried and complex vocabulary, (b) provid-ing embedded instruction on a subset of tar-geted words contained in the storybook, and(c) providing extended instruction on a sec-ond set of words from the story.

Embedded instruction should targetwords that students will encounter often inacademic discourse but are not immedi-ately critical for comprehending the story.Embedded instruction may help studentsestablish an initial lexical representation ofwords and provide a foundation that willassist them in refining and consolidatingword knowledge when they encounterwords subsequently, either incidentally orthrough additional instruction. Words cho-sen for extended instruction, in contrast,should be immediately essential for under-

standing important ideas and concepts inthe story. Extended instruction would helpstudents develop more complete wordknowledge that would support improvedlistening comprehension of the story. A tri-level approach such as this would leverageinstructional time efficiently to target theequally important goals of helping studentsdevelop both breadth and depth of vocab-ulary knowledge.

We found that higher and more completelevels of word knowledge were more dif-ficult to maintain when students did notreceive continued encounters with targetwords. This suggests that teachers who pro-vide direct vocabulary instruction shouldconsider ways to review target words overtime and ensure that students continue to beexposed to words after they have been intro-duced. This may be especially important formore sophisticated or low-frequency targetwords that students are less likely to encoun-ter incidentally.

Finally, our findings revealed that stu-dents with less receptive vocabulary knowl-edge did not make learning gains as large astheir peers with more developed vocabular-ies. This finding indicates that even directlyteaching target words is insufficient to closethe vocabulary gap among students (Bi-emiller, 2001). The growing evidence sup-porting the strong relation between initial vo-cabulary knowledge and responsiveness toinstruction highlights the need to intensifyinstruction for students most at risk for lan-guage and literacy difficulties. It is likely thatthese students will require additional inter-vention above and beyond general classroominstruction to make gains similar to theirpeers who are not at risk (Loftus, Coyne, Mc-Coach, & Zipoli, 2008).

Limitations and Future ResearchThis study was of very short duration,

and students were taught the meanings ofonly three words in each condition. Addi-tionally, although we selected words thatwould approximate the kinds of Tier II


SEPTEMBER 2009



words that would typically be taught inkindergarten classrooms, our target vocab-ulary consisted of low-frequency wordsthat may not have been representative ofhigher-frequency Tier II words.

Although many studies have used story-book readings as a context for teaching vo-cabulary, there are some possible limitationsto this approach as well. Even though pro-viding brief definitions of vocabulary duringa story reading takes very little time, theseinterruptions may disrupt listening compre-hension, especially if they occur frequently.The possibility that vocabulary instructionmay interfere with comprehension may alsolimit the number of words that can be intro-duced during a given story reading.

Finally, we examined only proximalmeasures of target word learning. It will beimportant for future research to replicatethese findings as well as to investigate theeffects of long-term implementation of bothembedded and extended instruction onboth proximal measures of word learningand more distal measures of language andliteracy such as generalized vocabularyknowledge and listening comprehension.

References

Anderson, R. C., & Nagy, W. E. (1992). Thevocabulary conundrum. American Educator,16(4), 14–18, 44–47.

Beck, I. L., & McKeown, M. G. (2007). Increasingyoung low-income children’s oral vocabu-lary repertoires through rich and focusedinstruction. Elementary School Journal, 107(3),251–271.

Beck, I. L., McKeown, M. G., & Kucan, L. (2002).Bringing words to life: Robust vocabulary in-struction. New York: Guilford.

Beck, I. L., McKeown, M. G., & Omanson, R. C.(1987). The effects and uses of diverse vo-cabulary instructional techniques. In M. G.McKeown & M. E. Curtis (Eds.), The nature ofvocabulary acquisition (pp. 147–163). Hills-dale, NJ: Erlbaum.

Biemiller, A. (2001). Teaching vocabulary: Early,direct, and sequential. American Educator,25(1), 24–29.

Biemiller, A., & Boote, C. (2006). An effectivemethod for building vocabulary in primary

grades. Journal of Educational Psychology,98(1), 44–62.

Biemiller, A., & Slonim, N. (2001). Estimatingroot word vocabulary growth in normativeand advantaged populations: Evidence for acommon sequence of vocabulary acquisi-tion. Journal of Educational Psychology, 93(3),498–520.

Calfee, R. C., & Drum, P. A. (1986). Research onteaching reading. In M. C. Wittrock (Ed.),Handbook of research on teaching (3d ed., pp.804–849). New York: Macmillan.

Catts, H. W., Hogan, T. P., & Adolf, S. M. (2005).Developmental changes in reading andreading disabilities. In H. W. Catts & A. G.Kahmi (Eds.), The connections between lan-guage and reading disabilities (pp. 25–40).Mahwah, NJ: Erlbaum.

Coyne, M. D., McCoach, D. B., & Kapp, S. (2007).Vocabulary intervention for kindergartenstudents: Comparing extended instructionto embedded instruction and incidental ex-posure. Learning Disabilities Quarterly, 30,74–88.

Coyne, M. D., Simmons, D. C., Kame’enui, E. J.,& Stoolmiller, M. (2004). Teaching vocabu-lary during shared storybook readings: Anexamination of differential effects. Exception-ality, 12(3), 145–162.

Dale, E. (1965). Vocabulary measurement: Tech-niques and major findings. Elementary En-glish, 42, 895–901.

Dunn, L., & Dunn, L. (1997). Peabody PictureVocabulary Test—Revised. Circle Pines, MN:American Guidance Service.

Elley, W. B. (1989). Vocabulary acquisition fromlistening to stories. Reading Research Quar-terly, 24(2), 174–187.

Gersten, R., Baker, S., & Lloyd, J. W. (2000).Designing high quality research in specialeducation: Group experimental design. Jour-nal of Special Education, 34(1), 2–18.

Glass, G. V., & Hopkins, K. D. (1996). Statisticalmethods in education and psychology. Boston:Allyn & Bacon.

Gresham, F. M., MacMillan, D. L., Beebe-Frankenberger, M. E., & Bocian, K. M.(2000). Treatment integrity in learning dis-abilities intervention research: Do we reallyknow how treatments are implemented?Learning Disabilities Research and Practice,15(4), 198–205.

Hart, B., & Risley, R. T. (1995). Meaningful differ-ences in the everyday experience of young Amer-ican children. Baltimore: Brookes.

Justice, L. M., Meier, J., & Walpole, S. (2005).Learning new words from storybooks: Anefficacy study with at-risk kindergartners.




Language, Speech, and Hearing Services inSchools, 36, 17–32.

Landauer, T. K., & Dumais, S. T. (1997). A solu-tion to Plato’s problem: The latent semanticanalysis theory of acquisition, induction andrepresentation of knowledge. PsychologicalReview, 104(2), 211–240.

Loftus, S., Coyne, M. D., McCoach, D. B., &Zipoli, R. (2008). Effects of a Tier 2 vocabularyintervention on the word knowledge of kinder-garten students at risk for language and literacydifficulties. Manuscript submitted for publi-cation.

Marshall, J. (1998). Goldilocks and the three bears.New York: Puffin.

Nagy, W. E. (2007). Metalinguistic awarenessand the vocabulary-comprehension connec-tion. In R. K. Wagner, A. E. Muse, & K. R.Tannenbaum (Eds.), Vocabulary acquisition:Implications for reading comprehension (pp. 52–77). New York: Guilford.

Nagy, W. E., & Scott, J. A. (2000). Vocabularyprocesses. In M. L. Kamil, P. B. Mosenthal,P. D. Pearson, & R. Barr (Eds.), Handbook ofreading research (Vol. 3, pp. 269–284). Mah-wah, NJ: Erlbaum.

National Reading Panel. (2000). Teaching childrento read: An evidence-based assessment of the sci-entific research literature on reading and its im-plications for reading instruction. Reports of thesubgroups. Bethesda, MD: National Instituteof Child Health and Human Development.

Nicholson, T., & Whyte, B. (1992). Matthew ef-fects in learning new words while listeningto stories. In C. K. Kinzer & D. J. Leu (Eds.),Literacy research, theory, and practice: Viewsfrom many perspectives. Forty-first yearbook ofthe National Reading Conference (pp. 499–503).Chicago: National Reading Conference.

Penno, J. F., Wilkinson, I. A. G., & Moore, D. W.(2002). Vocabulary acquisition from teacherexplanation and repeated listening to sto-

ries: Do they overcome the Matthew effect?Journal of Educational Psychology, 94(1), 23–33.

Robbins, C., & Ehri, L. C. (1994). Reading story-books to kindergartners helps them learnnew vocabulary words. Journal of EducationalPsychology, 86(1), 54–64.

Salvia, J., & Ysseldyke, J. E. (Eds.). (1998). Assess-ment (7th ed.). Boston: Houghton Mifflin.

Scarborough, H. (2005). Developmental relation-ships between language and reading: Recon-ciling a beautiful hypothesis with some uglyfacts. In H. W. Catts & A. G. Kamhi (Eds.),The connections between language and readingdisabilities (pp. 3–24). Mahwah, NJ: Erlbaum.

Schwanenflugel, P. J., Stahl, S. A., & McFalls,E. L. (1997). Partial word knowledge andvocabulary growth during reading compre-hension. Journal of Literacy Research, 29, 531–553.

Senechal, M., Thomas, E., & Monker, J. (1995).Individual differences in 4-year-old chil-dren’s acquisition of vocabulary during sto-rybook reading. Journal of Educational Psy-chology, 87, 218–229.

Stahl, S. A., & Fairbanks, M. M. (1986). The ef-fects of vocabulary instruction: A model-based meta-analysis. Review of EducationalResearch, 56(1), 72–110.

Storch, S. A., & Whitehurst, G. J. (2002). Orallanguage and code-related precursors toreading: Evidence from a longitudinal struc-tural model. Developmental Psychology, 38(6),934–947.

Vellutino, F. R., Tunmer, W. E., Jaccard, J. J., &Chen, R. (2007). Components of readingability: Multivariate evidence for a conver-gent skills model of reading development.Scientific Studies of Reading, 11(1), 3–32.

Walsh, B. A., & Blewitt, P. (2006). The effect ofquestioning style during storybook readingon novel vocabulary acquisition of pre-schoolers. Early Childhood Education Journal,33(4), 273–278.


SEPTEMBER 2009



Direct Vocabulary Instruction in Kindergarten: Teaching ...

Documents

Transcript of Direct Vocabulary Instruction in Kindergarten: Teaching ...