Contents lists available at ScienceDirect

Cognitive Development

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

Preschoolers’ word learning and story comprehension duringshared book reading

Sofia R. Jimenez⁎, Megan M. SaylorVanderbilt University, United States

A R T I C L E I N F O

Keywords:Word learningStory comprehensionJoint-book readingPreschoolersCognitive load

A B S T R A C T

Providing instruction about picture-book vocabulary is an effective way to support preschoolers’word learning. Often the instruction occurs while the book is being read, however this mayincrease cognitive load compared to providing vocabulary instruction before or after the book isread. If instruction provided during the story increases cognitive load, children might be morereliant on ancillary cognitive abilities to successfully learn the words and understand thestoryline. Three-to-five-year-olds (N = 83) were read a storybook that included instruction on sixnew words. Instruction was provided either during (inside) or before/after (outside) the story.Children were then tested on their ability to identify the named items, generalize to unfamiliarexemplars, and comprehend the story. Word learning and story comprehension were abovechance and similar across the inside and outside conditions. Memory skills were stronger pre-dictors of word learning in the inside condition than the outside condition. These results suggestthat interactive vocabulary instruction during the story may place more demands on pre-schoolers’ cognitive resources and may not be equally appropriate for all children.

1. Introduction

Before children can read on their own, joint book reading facilitates the development of language and later reading skills (e.g.,Biemiller, 2003; Landry et al., 2012). The effects of joint reading on children’s emerging literacy behaviors are long-lasting (Bus, VanIJzendoorn, & Pellegrini, 1995) and may be especially important for preschoolers from low-income families (Dickinson & Smith,1994). As a result, researchers have investigated which reading techniques best support vocabulary acquisition in the context of jointreading episodes. One popular technique is dialogic reading, which encourages adults to ask open-ended questions and elicit adialogue while reading to a child (Whitehurst et al., 1988). This technique involves adding extra-textual talk to provide instructionwhile reading the story. However, it is uncertain whether such interjections are necessary for vocabulary learning, or if breaking upthe story with instruction might be ineffective for children with less robust ancillary cognitive skills. In what follows, we investigatewhether children’s non-linguistic cognitive skills, such as memory, interact with the placement of instruction.

Dialogic reading was initially developed to help preschoolers be participatory during joint-reading sessions by encouraging extra-textual talk. One aspect of dialogic reading is asking children “wh-” questions such as “what do you think will happen” or, “where isshe?” to elicit extended dialogue during story time which might encourage the growth of expressive vocabulary. When using dialogicreading techniques, adults respond to what children say and scaffold questions to fit children’s abilities and familiarity with the bookthat is being read. Researchers have explored the efficacy of different aspects of dialogic reading since its development, includinglevel of interactivity (Justice, 2002; Walsh & Blewitt, 2006), and question types (Strasser, Larrain, & Lissi, 2013; Zucker, Justice,

http://dx.doi.org/10.1016/j.cogdev.2017.08.011Received 20 August 2015; Received in revised form 17 May 2017; Accepted 25 August 2017

⁎ Corresponding author.E-mail address: sofia.r.jimenez@vanderbilt.edu (S.R. Jimenez).

Cognitive Development 44 (2017) 57–68

Available online 07 September 20170885-2014/ © 2017 Elsevier Inc. All rights reserved.

MARK

Piasta, & Kaderavek, 2010). Previous intervention studies reveal that regularly engaging in dialogic reading can lead to gains in 3–6-year old children’s expressive vocabulary (e.g., Lever & Sénéchal, 2011; Whitehurst et al., 1994).

Dialogic reading has inspired studies that emphasize different elements of shared-book reading and extra-textual talk. For ex-ample, some studies with 3–5-year-old children encouraged supportive dialogue centered on the storyline and social themes (Aram,Fine, & Ziv, 2013). Other studies have focused on improving the quality of shared-book reading among educators (Milburn,Girolametto, Weitzman, & Greenberg, 2013; Wilkinson &Houston-Price, 2012) and parents (Hindman, Skibbe, & Foster, 2014; Landryet al., 2012). Although these reading techniques have normally been studied within typically developing monolingual populations,researchers have also investigated whether the technique supports language skills in and 3–5-year-old children with Autism SpectrumDisorder (Fleury, Miramontez, Hudson, & Schwartz, 2014), 5-year-old children with communication impairments (Colmar, 2014;Kaderavek, Pentimonti, & Justice, 2013), and bilingual 7–8-year-olds (Jiménez, Filippini, & Gerber, 2006).

Additional research has borrowed elements of the dialogic technique for a different purpose: to ask whether preschoolers learnspecific vocabulary items that are included in the story (Ewers & Brownson, 1999; Sénéchal & Cornell, 1993; Walsh & Blewitt, 2006).One common element across study procedures is the tendency to break-up the reading to ask preschoolers questions (e.g., Justice,2002; Sénéchal & Cornell, 1993; Walsh & Blewitt, 2006). The assumption is that questioning and instruction should occur proximal tothe vocabulary item in the text because it may increase children’s attention to vocabulary items and associated narrative content.However, such interjections might increase children’s cognitive load because they have to juggle multiple tasks during dialogicreading episodes.

When children are encouraged to be participatory, they have to manage many tasks at once: following the narrative presented inthe story, contributing to mini-conversations as the story is stopped and probes are introduced, and learning the new words presentedin the story. Consistent with cognitive load theory, balancing these different demands may tax preschoolers’ limited cognitive re-sources and lead to less optimal learning (Paas & Ayres, 2014; Paas & Sweller, 2012). One possibility is that such interjections willreduce story comprehension and word learning. This could occur because responding to questions during a story requires children toswitch between several tasks that necessitate tracking multiple pieces of information simultaneously (e.g., listening to the story,learning words, participating in a dialogue). This requires children to participate in a dialogue while they are also tasked with holdingaspects of the story in memory while also attempting to keep novel words and potential meanings in mind. In addition to loading theirmemory stores, children may use cognitive control, like inhibition, to switch between these tasks.

One useful model to inform this hypothesis was offered by Lee and Kinzie (2011) in an analysis of the effects of teachers’ questionsduring a science activity with 4–5-year-olds. Their proposal was that although teacher questions scaffold children’s learning, chil-dren’s ability to answer the questions remains constrained by their available cognitive resources. This occurs because answeringquestions requires multiple components (e.g., listening to the question, understanding it, creating and producing an appropriateresponse). To answer questions, children must have the capacity to retain many different pieces of information at once. Therefore,questions that are meant to aid children could inadvertently overload their memory and decrease their ability to learn from theactivity. Applying this to shared book reading: when questions are interjected into the text, additional cognitive skills may benecessary to enable optimal word learning and story comprehension. While there is evidence from 7 to 10-year-old children that bothmemory and executive function support reading comprehension (Cain, Oakhill, & Bryant, 2004; Sesma, Mahone, Levine,Eason, & Cutting, 2009), few (if any) studies have investigated how non-linguistic cognitive skills contribute to pre-literate children’sword learning and story comprehension.

One possibility is that the extent to which interjections lead to increases in cognitive load may be affected by when the ques-tioning occurs. In particular, children may have greater difficulty if questions are interjected during the story reading versus beingplaced before or after the reading. This might occur because when instruction occurs outside of the story context, children do notneed to balance the needs of attending to story content while they are answering questions. Previous work has not addressed thisquestion directly; although there are suggestions that the placement of interjections has little to no effect on children’s overall levelsof vocabulary learning (Blewitt, Rump, Shealy, & Cook, 2009) or story comprehension (Reese & Cox, 1999). The lack of overalldifferences in learning may indicate that load is not increased when stories are interrupted versus not. However, because previousstudies have not included separate measures of cognitive skills that may be affected by increases in cognitive load it is not clearwhether there are greater costs – in terms of the amount of memory necessary to be successful at learning – when stories areinterrupted with questions versus having questions outside of the story context. We test this possibility in the current study.

The central goal of the current study was to determine what effect cognitive abilities have on word learning and story com-prehension when questions are posed during the story versus before or after the reading. In one condition, interactive vocabularyinstruction was placed within the story (the inside condition). In the other conditions, instruction was placed either before or after thestory. Both vocabulary acquisition and story comprehension were used as measures of learning. Based on the cognitive load hy-pothesis outlined above, the prediction was that these abilities would be more likely to influence word learning and story com-prehension in the inside condition. To test this possibility, we included two measures of memory (forward digit span, self-orderedpointing) and a measure of inhibition (day-night stroop, Carlson, 2005; Cragg &Nation, 2007; Montgomery & Koeltzow, 2010;Simpson & Riggs, 2005). We also included a measure of receptive vocabulary (the Peabody Picture Vocabulary Test-IV; Dunn &Dunn,2007). We tested children between the ages of 3 and 5, because there is a great deal of variability in children’s vocabulary size andcognitive skills across this age period (e.g., Carlson, Moses, & Claxton, 2004; Deak, 2000; Justice 2002; Nicolopoulou, Cortina, Ilgaz,Cates, & de Sá, 2015; Zucker et al., 2010).

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

58

2. Method

2.1. Participants

Eighty-three preschool-aged children (range 36–72 months, M = 54 months, 38 Females, 45 Males) were recruited from childcarecenters and state birth records in the southeastern United States. Six additional children were recruited and not included in theanalysis for non-compliance (n = 1), biased responding (n = 4) and non-proficiency in English (n = 1). Participants were typicallydeveloping, had intact hearing, and heard English in their household at least 70% of the time. Participants were randomly assigned toone of three between-subjects conditions: Inside (n = 41), Before (n = 22), or After (n = 20) with the constraints that age and genderwere equal across conditions.

Demographic surveys revealed that 48% of mothers had a post graduate degree or some graduate school, 45% had a collegedegree or some college, and 2% had a technical/AA degree. Four participants’ parents did not respond to this question. 45% ofparticipants’ families reported an income of $150,000 or more per year, 17% reported an income between $100,000 to $150,000 peryear, 13% reported an income between $50,000 and $100,000 per year, and 14% of families reported an income of less than $50,000per year. Ten participants’ parents did not respond to this question.

2.2. Design

The procedure began with a pretest that evaluated children’s knowledge of the target novel words that appeared in the storybook.After the pretest, children in both conditions were read a picture book and underwent label training. The two conditions only differedwith regard to the placement of the label training. In the inside condition, the label training occurred when the words appeared in thestory. In the two outside conditions, children were taught the words either before or after the storybook was read in its entirety.Following this, participants were tested on the target words in a receptive vocabulary test, a generalization test, an expressive vocabularytest, and a story comprehension test. The story comprehension test either came before the three vocabulary tests or after the threevocabulary tests. After the vocabulary tests, participants were tested on general cognitive and linguistic measures with the digit span,day-night stroop, self-ordered pointing task, and PPVT-IV.

2.3. Label training

2.3.1. Inside conditionIn the inside condition, the researcher provided participants with information about what the word meant when she came to a

novel target word in the text of the book. The instructional dialogue was interactive and followed the following pattern:

Experimenter: “Do you know what an ocarina is?”

Child responds.

Experimenter: “An ocarina is a musical instrument you can play with your mouth. Do you see that?”

Child responds.

Experimenter: Can you say ocarina?

Child repeats word.

If children did not point to the item when asked, the experimenter pointed to the target item in the illustration. Participants werenot corrected if they mispronounced target words or if they failed to say the target word when prompted. Additionally, to control forthe amount of time spent on each novel word, the same script was used for each target word and extraneous pausing and extra-textualtalk was avoided. Children heard the novel target word four times: three times during instruction and once when the researcher readthe page with the novel word on it. Please see the Appendix A for the text of the story and the definitions that were provided tochildren.

2.3.2. Before conditionAn identical label training procedure and instructional dialogue was used in the before condition, except that the instructional

dialogue used in the inside condition was provided after the pretest, but before the story was read. Children were taught the six targetnovel words, one at a time, in the same order that they appeared in the story. Children were also shown the picture that accompaniedthe text to equate the presentation of the label training across conditions. Then the book was read without any interjections.

2.3.3. After conditionThe participants in the after condition were read the storybook straight through after the pretest. After this they received label

training using the same instructional dialogue used in both the inside and before conditions.

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

59

2.4. Materials

Participants were read an illustrated storybook created for the purposes of this experiment. It contained six low-frequency targetwords (i.e. rebar, gingham, Nicaragua, rickshaw, ocarina and pitaya) that were not easily nameable using a higher-frequency sy-nonym. The words were chosen to avoid conflict from a familiar label (i.e., that might activate the mutual exclusivity bias, Jaswal,2010). Pilot testing determined that these words were unfamiliar for same aged children who did not participate in the study. Eachword appeared in the story only one time.

The storyline included a setting and an episode, adhering to story grammar. The story followed a brother and sister who traveledto Nicaragua with their mother. The children became separated from their mom during their trip, but the three reunited at the end.The story was written to maintain preschoolers’ interest and to create an ecologically valid joint reading situation. It contained 252words with 13 being polysyllabic (3 syllables or more).

For the pretest, receptive vocabulary posttest, expressive vocabulary posttest, and generalization test, cards were created that depictedtarget and distractor objects as they appeared in the storybook. The cards for the generalization test depicted the same items in aphotograph (rather than using the illustrations used in the book). See Fig. 1.

2.5. Vocabulary tests

2.5.1. PretestTo ensure the target words were unfamiliar to the participants, they were pretested in a three-option forced-choice task for all six

target novel words. In particular, they were asked to select a card (“Show me the ocarina.”) when presented with three options: thetarget (ocarina) and two distractors (a different target item (Nicaragua) and an object that was illustrated in the storybook, but notnamed by the researcher). If children chose more than one card, the final item they selected was coded as their response. The order ofwords presented was counterbalanced across participants, as were the locations of the three cards. Children received a score of 1 ifthey selected the item the target item in response to the researcher’s request. Children could therefore score between 0 and 6 on thismeasure.

2.5.2. Receptive vocabulary posttestAfter the story was read and the vocabulary instruction was completed, participants were retested on the target words in a three-

item forced choice task that was identical to the pretest, apart from the items being presented in a different counterbalanced order.Scoring was conducted as in the pretest.

2.5.3. Generalization testThe generalization test was also a three-item forced choice measure. The procedure, counterbalancing, and scoring were all

identical to the pretest and receptive vocabulary posttest. The only difference was that the test items were depictions of real pho-tographs of the target items and distractors. This test was included to determine if children recognized that the label used in the textreferred to an example of the named item that was not depicted in the book. Correct selection of the target in this case might indicatethat children learned a category label rather than simply associating the name with the illustration in the text.

Fig. 1. Examples of vocabulary cards for the target-word “ocarina” and the two distractors. The pictures on the top row were used for pretest and receptive vocabularyposttest and were taken from the storybook. The pictures on bottom row were used for generalization.

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

60

2.5.4. Expressive vocabulary testFor the expressive vocabulary test, participants were asked to name an item shown on a card (“What is this called?”). The

depictions of the target items matched their presentation in the storybook. If children said they did not know or did not answer afterfive seconds, the experimenter provided a hint. The hint provided was the first syllable of the target word (e.g., “It starts with oc-”). Ifchildren still did not produce the word after the hint and five additional seconds, the experimenter moved onto the next word. Theorder of the presentation of the words was counterbalanced across participants. Participants received a score of 1 for correctly namingthe target item without a hint. If they named the target after the hint they received half a point. Children’s scores could thereforerange between 0 and 6.

2.5.5. Story comprehensionDuring the story comprehension test, children were asked six questions about events in the story. They were provided with three

answer choices (see Appendix A for full list of comprehension questions). These questions targeted events that occurred from thebeginning to the end of the story and did not include any of the target words. For example, one question was “Who told Ben and Amythat they were going on a trip?” Children were offered three options for the answer to the question: One of the choices was correct(their mom), one was a near distractor that included an element of the story (their teacher), and one choice was a far distractor thatdid not appear in the story (their dog). The answer options were counterbalanced. The story comprehension questions were presentedin the same order for each child.

Twenty percent of the participants’ pretest, receptive vocabulary posttest, generalization and story comprehension tasks (n = 16)were recoded by a research assistant. There was 97% agreement between the initial coding and the reliability coders’ judgments ofwhether children selected the correct item.

2.6. Cognitive and linguistic measures

After children completed the label training and vocabulary tests, the cognitive and linguistic measures were administered. Allparticipants were tested in the following order: digit span, day-night stroop, self-ordered pointing task, and then the PPVT-IV.Children did all tasks in one session on one day, unless they were tested in childcare centers. If they were tested in childcare centers,the PPVT-IV was administered on a different day.

2.6.1. Digit span (memory)Children were told they were going to play a game where the experimenter said some numbers and then they were asked to repeat

the numbers right after. Participants were trained with a number sequence (8, 2) and did not begin the task until they correctlyrepeated back the two-digit sequence (by repeating “eight, two”). After this practice round, they were given two numbers to say, theirresponses were recorded, and they were given two more numbers to say. Then they were given two trials with three digits, then twotrials with four digits, etc. until they made a mistake on both digit sequences in the same level. Children’s score was the number oftrials they correctly repeated back before making a mistake on a pair of items of the same length (e.g., a child received a score of 6 ifthey correctly repeated back both of the two digit items, both of the three digit items, both of the four digit items, but made an erroron both of the five digit items). The number of sequences a child correctly repeated was recorded as their score. Cain et al. (2004)report an α = 0.82 for this task.

2.6.2. Day-night stroop (inhibition)Twenty day-night cards were created for a day-night stroop task to measure inhibition. The day card depicted a yellow sun with a

white background while the night card had a white crescent moon with a black background. In this task children were to say “day”when presented with the night card and “night” when presented with day card. Participants practiced the task until they demonstratedunderstanding of the rules. Participants were given one point for correct responding. Their scores could therefore range between 0and 20. Opposite responding and self-corrects were recorded as errors. There were twenty trials. The Kruder–Richardson Reliabilitywas reported as 0.93 by Chasiotis, Kiessling, Hofer, and Campos (2006).

2.6.3. Self-ordered pointing task (memory)This task measured visual working memory and was modeled after a task described by Cragg and Nation (2007). Children were

told that they would see pages with pictures of things on them and that they would be asked to point to a new thing every time theexperimenter flipped the page. Children were shown one page at a time and asked to point to a different item on each page. Itemsappeared in different locations on each page, to ensure participants were tracking the items and not the locations. The task wasdivided into three increasingly difficult levels, each with its own stimuli. The first level had four pages with four items randomlyarranged on each page (i.e., lamp, glasses, hammer, and umbrella). In the second level eight different fruit items appeared in differentlocations on eight pages and in the third level ten animals appeared in different locations on ten pages. Children completed eachdifficulty level only once. Participants were given one point for each new item they pointed to. Children could receive a score rangingbetween 0 and 22 for the self-ordered pointing task. Cragg and Nation (2007) report that α= 0.88 for this task.

2.6.4. Peabody picture vocabulary test-IV (PPVT-IV)This measure was used to measure children’s receptive vocabulary size (Dunn &Dunn, 2007). The PPVT-IV requires that children

point to the correct picture (given four options) when presented with a word. Administration of the measures and scoring followed

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

61

the guidelines in the PPVT-IV manual. The split-half reliability of the PPVT-IV for 3–5-year-old children ranges from 0.94 to 0.96.

2.6.5. QuestionnairesParents were also given questionnaires about the home literacy environment of the child, including how many hours per week the

child was read to and how many children’s books were in the home.

3. Results

We investigated whether eliciting an instructional dialogue inside of a story is important for word learning and story compre-hension, and if there are cognitive costs associated with interrupting the storyline. In particular, we tested whether placing theinstruction inside the story incurred a larger cognitive load than having the instruction before or after the story.

3.1. Preliminary analysis

3.1.1. Cognitive measuresOur cognitive measures were correlated with age such that performance increased with age and mostly did not differ across

gender. We found that age was positively correlated with digit span (r(80) = 0.49, p < 0.001), the self-ordered pointing task (r(79)= 0.27, p < 0.25), inhibition (r(78) = 0.50, p < 0.001) and raw PPVT-IV scores (r(80) = 0.60, p < 0.001). Therefore, children’sperformance on our cognitive measured showed the expected age related increase. There were no gender differences in children’sresponses to the digit span (t(78) = 0.05, p = 0.96), inhibition (t(76) = 1.16, p = 0.25), or raw PPVT-IV scores (t(78) = 1.18,p = 0.24). Unexpectedly, we found that females (M = 19.92, sd = 1.73) in our sample outperformed males (M = 18.51, sd = 2.34)on the self-ordered pointing task (t(77) = 2.98, p = 0.004).

3.1.2. Demographic measures by conditionPreliminary measures revealed that there were no differences between the before or after conditions with respect to both word

learning, independent samples t(40) = 0.83, p = 0.41, d = 0.26, and story comprehension: t(40) = 0.25, p = 0.81, d = 0.08. As aresult, we collapsed across these two presentation orders into the outside condition in the analyses below. Independent sample t-testsdetermined that the participants in the inside and outside conditions were matched on age, sex, maternal education and income,number of children’s books in the home, number of hours read to per week as well as PPVT-IV, digit-span, inhibition, and self-orderedpointing task scores. Refer to Table 1 for means and standard deviations for each of these variables

3.2. Condition comparisons

To investigate whether children learned words based on the instruction we provided, we first conducted a 2 (Condition: inside,outside) X 2 (Phase: receptive vocabulary pretest, receptive vocabulary posttest) repeated-measures mixed ANOVA. Condition wasbetween participants and Phase was within. The analysis revealed a significant effect of Phase, which was the result of an increase inselection of the target referent across the pretest (M = 2.31, SD = 1.00) and posttest (M = 3.52, SD = 1.36), F(1,81) = 35.85,p < 0.001, ƞp2 = 0.31. There was no effect of Condition, F(1,81) = 1.08, p = 0.30, ƞp2 = 0.01 and there was no interaction be-tween Condition and Phase, F(1,81) = 0.74, p = 0.39, ƞp2 = 0.01. These results clarify that children learned words based on theinstruction we offered and that children learned words equally well across the inside and outside conditions. In line with this result,there were no differences in performance across conditions on the generalization items, t(81) = 0.22, p = 0.83, d = 0.05, or for theexpressive vocabulary test, t(80) = 0.54, p = 0.58, d = 0.12. See Table 2 for summary statistics.

Table 1Means and standard deviations of participant demographics and cognitive measures by condition.

Inside condition (N = 41) Outside condition (N = 42)

M SD Range M SD Range

Age (in months) 54.87 8.68 36–72 53.57 8.61 38–72Sex (proportion female) 0.46 0.45Maternal Education* 5.73 1.34 3–7 5.56 1.43 3–7Income* 5.74 1.84 1–7 5.56 1.93 1–7Number of Children’s books 132 67 15–300 117 93 3–500Hours read per week 3.73 2.65 1–16 4.20 2.68 2–16PPVT-IV raw score 99.05 23.74 46–135 95.56 23.58 27–138Digit-span 6.35 1.48 3–9 6.10 1.52 3–10Inhibition 14.76 4.09 2–20 14.32 4.64 5–20Self-Ordered Pointing Task 19.07 1.69 15–22 19.24 2.65 12–22

Note: *Maternal education and income were given a value from 1 to 7. For maternal education, options were: some high school (1), high school (2), some college (3),technical or AA degree (4), college degree (5), some graduate school (6), and post graduate degree (7). Income levels were: less than $15,000 (1), $15,000–$30,000(2), $30,000–$50,000 (3), $50,000–$75,000 (4), $75,000–$100,000 (5), $100,000–$150,000 (6), $150,000 or more (7).

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

62

To evaluate the reliability of children’s performance during the receptive vocabulary posttest and receptive vocabulary pretesttests against chance were conducted. Children selected the target at above chance levels in the posttest, when they picked the novelword from the illustrations they had seen before, in the outside (t(41) = 7.89, p < 0.001, d = 2.47) and inside (t(40) = 6.53,p < 0.001, d = 2.06) conditions. Additionally, children identified the target during generalization trials at above chance levels inboth the outside (t(41) = 4.42, p < 0.001, d = 1.32) and inside (t(40) = 4.79, p < 0.001, d = 1.51) conditions. Unexpectedly,children’s selection of the target was above chance levels during the pretest in the outside (t(41) = 2.44, p = 0.02, d = 0.76), but notthe inside (t(40) = 1.77, p = 0.09, d = 0.56) condition.

3.3. Responding to story comprehension items

There were no differences between conditions for children’s responses to story comprehension items in the outside and insideconditions (See Table 2 for means and standard deviations), independent samples t(81) = 1.60, p = 0.11, d = 0.36. Story com-prehension was above chance in both the outside (t(41) = 8.02, p < 0.001, d = 2.51) and the inside (t(40) = 9.63, p < 0.001,d = 3.05) conditions. There was a modest positive correlation between story comprehension and children’s responses to the receptivevocabulary posttest, r(83) = 0.28, p = 0.01, and generalization items, r(83) = 0.34, p = 0.002. Children who learned the wordseffectively also understood and remembered the storyline better.

3.4. What skills support children’s learning?

To investigate whether learning the words and understanding the story required more resources in the inside condition comparedto the outside condition, we conducted two hierarchical multiple regressions. The set of variables entered into the regressions waschosen based on an examination of bivariate correlations between predictor and outcome variables. Predictor variables that werecorrelated with the outcome variable at p < 0.10 were included (see Table 3 for correlations). Digit span and the self-orderedpointing task were combined into a single memory measure to reduce the number of predictors in our regression. Interaction termsfor the measures of cognitive ability that were entered into the regressions, as we had a priori predictions about differential effects ofthese variables on children’s word learning story comprehension in the inside versus outside conditions.

3.5. Predicting receptive word learning

The outcome variable for word learning was a combination of receptive vocabulary posttest and generalization scores since thesetwo measures both measured word learning and were correlated (r(83) = 0.45, p < 0.001). To predict word learning, age in monthswas entered in the first step and PPVT-IV, inhibition, memory, condition and two interaction terms: memory by condition andinhibition by condition were entered in the second step.

Table 2Means and standard deviation of learning measures by condition.

Inside condition (N = 41) Outside condition (N = 42)

M SD M SD

Pretest 2.33 1.16 2.32 0.85Receptive vocabulary posttest 3.33 1.33 3.69 1.39Generalization 3.18 1.58 3.10 1.61Story comprehension 3.88 1.25 3.45 1.17Elicited production 1.01 0.95 1.13 1.08

Note: Means are out of 6 and chance selection is 2, except for Elicited production.

Table 3Correlations between measures for entire sample.

Variable 1 2 3 4 5 6 7 8

1. Pretest – −0.08 −0.02 −0.00 0.13 0.08 0.12 0.012. Word learning – 0.37** 0.29** 0.33** 0.30** 0.42** 0.093. Story Comprehension – 0.43** 0.54** 0.22 0.55* 0.31**

4. Age (months) – 0.49** 0.50** 0.60** 0.205. Memory (digit span and SOPT) – 0.22 0.56** 0.32**

6. Inhibition – 0.40** 0.107. PPVT-IV – 0.47**

8. Maternal Education –

Note: PPVT-IV raw scores are reported.** Correlation is significant at the 0.01 level (two-tailed).* Correlation is significant at the 0.05 level (two-tailed).

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

63

The hierarchical multiple regression predicting word learning was significant (F(6,70) = 3.99, p < 0.001). At step 1 of theregression age in months accounted for a significant amount of variability R2 = 0.11, F (1,76) = 9.26, p < 0.001. In step two, theadditional predictors, PPVT-IV, inhibition, memory, condition, memory by condition interaction and inhibition by condition inter-action significantly predicted more of the variance in word learning than age alone ΔR2 = 0.18, F(6,70) = 2.88, p < 0.05. Thememory by condition interaction emerged as the only significant independent predictor (β = 0.29, p < 0.05; See Table 4). Weinterpreted the significant interaction between memory and condition by plotting it (Cohen, Cohen, West, & Aiken, 2003) and found a

Table 4Hierarchical regression predicting Word learning and Story Comprehension.

Predicting Word Learning B SE(B) β T Sig. R R2 ΔR2

Step 1 0.33 0.11 0.11Age (months) 0.10 0.03 0.33 3.04* 0.003

Step 2 0.53 0.29 0.18Age (months) 0.01 0.04 0.02 0.18 0.861Condition −0.62 0.51 −0.12 −1.22 0.225PPVT-IV 0.03 0.02 0.23 1.65 0.103Memory −0.09 0.48 −0.03 −0.20 0.844Inhibition 0.05 0.09 0.08 0.56 0.578Memory by Condition 1.40 0.69 0.29 2.04* 0.046Inhibition by Condition 0.08 0.12 0.09 0.63 0.528

Predicting Story Comprehension B SE(B) β t Sig. R R2 ΔR2

Step 1 0.48 0.23 0.23Age (months) 0.05 0.02 0.37 3.56* 0.001Maternal Education 0.22 0.09 0.25 2.40* 0.019

Step 2 0.66 0.43 0.20Age (months) 0.01 0.02 0.03 0.29 0.775Maternal Education 0.04 0.09 0.05 0.45 0.657Condition 0.41 0.22 0.17 1.85 0.068PPVT-IV 0.02 0.01 0.29 2.18* 0.033Memory 0.40 0.20 0.26 1.98 0.052Memory by Condition 0.32 0.29 0.14 1.14 0.259

* Predictor significant at 0.05 level (two-tailed).

Fig. 2. Plot of the relationship between word learning and memory in the inside condition r2 linear = 0.38, and in the outside condition r2 linear = 0.01.

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

64

crossover interaction (See Fig. 2). In line with our predictions, word learning was correlated with memory the inside condition (r(41)= 0.61, p < 0.001), but not in the outside condition (r(38) = 0.09, p = 0.57).

3.6. Predicting story comprehension

We conducted a hierarchical multiple regression on story comprehension that included age in months, maternal education in thefirst step, and memory, PPVT-IV, condition and the interaction between memory and condition in the second step.

The hierarchical multiple regression predicting story comprehension was significant (F(6,70) = 8.95, p < 0.001). At step 1 ofthe regression age in months and maternal education accounted for a significant amount of variability R2 = 0.23, F (2,74) = 11.31,p < 0.001. In step two, the additional variables, PPVT-IV, condition, memory, and memory by condition interaction accounted forsignificantly more variance in story comprehension ΔR2 = 0.20, F(4,70) = 6.18, p < 0.001. The raw PPVT-IV score emerged as theonly significant predictor (β = 0.29, p < 0.05; See Table 4). Contrary to our predictions the interaction between condition andmemory did not significantly predict story comprehension.

4. General discussion

Despite the popular convention of teaching words to children as they appear in storybooks, the current findings suggest that thisplacement might come at the cost of increasing cognitive load during word learning tasks. Children relied on memory skills to learnwords in the inside condition, where vocabulary instruction occurred during the story, more so than in the outside condition, in whichvocabulary instruction took place before or after the story was read. The effect of interjections about word meaning was specific toword learning; story comprehension was not similarly affected.

Overall, interactive vocabulary instruction was effective in facilitating children’s word learning, and children learned wordsequally well when instruction occurred during, before, or after the story. In addition, the comprehension of the storyline wasequivalent across our presentation conditions and was positively correlated with measures of word learning. This indicates thatparticipants who understood and remembered the story events also retained the information about the target words, which suggeststhat story comprehension and word learning are complementary skills during joint book reading (Reese & Cox, 1999; Strasser et al.,2013).

We found support for our hypothesis that placement of instruction differentially affected cognitive load during word learning.Children in our sample came from a broad range of ages—spanning the preschool period of three to five years. Although age wascorrelated with word learning, it was not a significant independent predictor of word learning when other cognitive and linguisticskills were taken into account. Instead, the interaction between participants’ memory and condition was a better predictor of wordlearning. If participants were in the inside condition, their memory was a correlated with word learning, but if they were in the outsidecondition their memory did not correlate with word learning. Memory abilities may thus play a larger role in vocabulary acquisitionwhen words are taught during the story, rather than when they are taught before or after the story.

Cognitive load theory may explain this result, because in the inside condition children had to keep more items in mind at once.These results suggest that more traditional dialogic instruction may put a larger demand on preschoolers’ cognitive load than whenthe story is separate from vocabulary instruction. Although we did not specifically target a population of children with low memoryabilities, the variation in memory abilities among typically-developing children was sufficient to show this effect. Future studiesshould clarify whether populations of children with impaired memory abilities, such as children with low birth weight (Baron, Kerns,Müller, Ahronovich, & Litman, 2012), have increased difficulty learning words when instruction is interjected in a storybook.

The influence of cognitive abilities and instructional placement on story comprehension was less clear, however. While theinteraction between memory and condition was not a significant predictor of story comprehension, we did find that vocabulary sizepredicted story comprehension across both conditions, even when controlling for age and maternal education. This is a sensiblefinding: the ability to correctly respond to the story comprehension questions relies on children’s receptive language skills. Onereason cognitive skills did not affect story comprehension in the same way as word learning is that there might be specificity in theeffects of interjections on increasing cognitive load. In other words, because the interjections were about vocabulary (not the nar-rative) children were not faced with the same costs associated with balancing multiple tasks. A future study that includes interjectionsthat relate to the narrative could test this possibility.

Children were exposed to the pictures of the target words for the same amount of time in the two conditions. In the insidecondition, they saw the picture of the target word one time (since story instruction happened during story reading), in the outsidecondition they saw the pictures twice (once for the word instruction and once for the story reading). Importantly, in the outsidecondition, the duration of exposure to the pictures was the same as in the inside condition – it was just divided into two shortersegments of presentation. Children in both conditions heard the same information about the target items with the same number ofmentions of the target word. In designing the study, we prioritized the duration of exposure to the target word pictures over thenumber of times children were exposed. Given these considerations, we believe our design carefully isolated the effects of placementof instruction on word learning and provide a strong test of the role of cognitive abilities supporting learning during instruction.

There are several remaining questions about the role of instructional placement on children’s learning during storybook reading.Although the current findings suggest that the proximity of information about word meaning does not differentially affect learning,there are some circumstances where having the information close might matter more. One such case is non-fiction books withexpository text (e.g., a book that introduces new kinds of flowers, Justice, 2002). Such presentations of novel words may make thetarget words’ meaning difficult to infer without proximal extra-textual talk. It will also be important to determine if instructional

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

65

placement influences preschoolers’ learning in a more naturalistic parent or teacher led reading context. In subsequent studies it willbe important to investigate whether the role of instructional placement changes based on the story context and reading environment.

Future studies could also probe the effects of mass versus distributed practice of target items on vocabulary learning fromstorybooks by parametrically manipulating exposure time and presentation frequency. Previous research on the efficacy of the twopractice styles is mixed. In one study, two-year-olds learn novel words more efficiently if they were repeated across successivesentences (e.g. “Look at the fep. The fep is neat.”), than if they were distributed throughout different sentences (e.g. “Look at the fep,The coro is neat.”) with other novel words (Schwab & Lew-Williams, 2016). This finding is highlights the utility of mass practice. Incontrast, other studies have found that distributed practice facilitates word learning among 3-year-old children (e.g.,Vlach & Sandhofer, 2012; Vlach, Sandhofer, & Kornell, 2008). Differences in the methods used may account for the different findings(e.g., the first study used a looking time procedure while the second body of work used the train-test procedure that is typical of wordlearning studies with preschoolers). One possibility for future study is to use our storybook procedure to carefully control thepresentation across massed and distributed conditions to test whether mass versus distributed practice is a more effective tool forteaching vocabulary during shared book reading

All together the current study revealed that children’s memory abilities predicted performance on word learning tasks when thewords were taught inside the narrative, but not when they are taught separately from the narrative. This indicates that the cognitiveload might have been increased by the interjections of vocabulary instruction into the story. We also found that interjecting defi-nitions into stories may not be necessary to engender word learning, as word learning was equal across conditions. One possibility isthat interjections could prove difficult for children with low memory capacity, such as children who are born prematurely and thoseborn with a low-birth weight (Baron et al., 2012). This work clarifies that there is value to investigating children’s label learning inthe context of ancillary cognitive abilities. Such approaches yield a more nuanced understanding of the factors supporting early orallanguage skills.

Acknowledgements

This research was supported by the Vanderbilt Psychology and Human Development departmental funds as well a NationalScience Foundation—Graduate Research Fellowship awarded to the first author (1445197). We would like to sincerely thank themembers of the Language Development Lab for their help with recruitment, data input and organization as well as the parents andpreschools that allowed their children in to participate in this study.

Appendix A

Target words and Definitions

1. Rebar: metal sticks inside cement and sidewalk2. Gingham: pattern of colored squares3. Nicaragua: a different country south of the U.S4. Rickshaw: a bike where people can ride in front5. Ocarina: a musical instrument people play with their mouth6. Pitaya: a sour purple fruit

Storybook Text (target words are underlined. * indicates where vocabulary instruction took place in the inside condition)

1. Once there was a brother and sister named Ben and Amy that lived in a cold, grey city. Every day they walked through the city toget to school.

2. They passed a corner where rebar was sticking out of the sidewalk. * Ben liked to hop over it.3. One day as they got back to their house from school, Ben and Amy’s mom set a pitcher down on the gingham tablecloth. * She

told them they were going on a trip to a faraway place where she would teach English.4. Soon they were on a plane going Nicaragua. *5. They arrived in a village near by a huge volcano and a beautiful green forest.6. On their way to their mother’s new school where she would teach, Ben and Amy hopped onto a rickshaw. *7. While they rode through the streets a parade of ocarina players crossed their path. * They lost their mom!8. Ben and Amy didn’t know what to do alone in a place they had never been before.9. They tried calling for their mom and looked around the market, but nobody knew who their mom was or where she went. Nobody

could help them.10. They got very thirsty so they stopped for pitaya drinks from a nice lady. * She told them the way to the school.11. After a few minutes of walking they found the red schoolhouse right by the volcano!12. Finally, they found their mom and met all the local kids. Everyone danced and celebrated.13. The End.

Story Comprehension Questions (answers in bold)

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

66

1. When the story started, where did Ben and Amy live?a In a cityb In a treec In a village

2. Who told Ben and Amy that they were going on a trip?a Their teacherb Their momc Their dog

3. What was Ben and Amy’s mom going to do in the new place?a Danceb Cookc Teach

4. How did Ben and Amy lose their mom?a A paradeb They ran awayc On an airplane

5. What did the nice lady tell Ben and Amy when they got drinks from her?a What birds eatb How to get to the schoolc They were on a sidewalk

6. Where was the school?a On the beachb In the marketc By the volcano

References

Aram, D., Fine, Y., & Ziv, M. (2013). Enhancing parent-child shared book reading interactions: Promoting references to the book’s plot and socio-cognitive themes.Early Childhood Research Quarterly, 28(1), 111–122. http://dx.doi.org/10.1016/j.ecresq.2012.03.005.

Baron, I. S., Kerns, K. A., Müller, U., Ahronovich, M. D., & Litman, F. R. (2012). Executive functions in extremely low birth weight and late-preterm preschoolers:effects on working memory and response inhibition. Child Neuropsychology, 18(6), 586–599.

Biemiller, A. (2003). Vocabulary: Needed if more children are to read well. Reading Psychology, 24(3–4), 323–335. http://dx.doi.org/10.1080/02702710390227297.Blewitt, 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

Educational Psychology, 101(2), 294–304.Bus, A. G., Van IJzendoorn, M. H., & Pellegrini, A. D. (1995). Joint book reading makes for success in learning to read: A meta-analysis on intergenerational

transmission of literacy. Review of Educational Research, 65(1), 1–21.Cain, K., Oakhill, J., & Bryant, P. (2004). Children’s reading comprehension ability: Concurrent prediction by working memory, verbal ability, and component skills?

Journal of Educational Psychology, 96(1), 31–42.Carlson, S. M., Moses, L. J., & Claxton, L. J. (2004). Individual differences in executive functioning and theory of mind: An investigation of inhibitory control and

planning ability. Journal of Experimental Child Psychology, 87(4), 299–319. http://dx.doi.org/10.1016/j.jecp.2004.01.002.Carlson, S. M. (2005). Developmentally sensitive measures of executive function in preschool children. Developmental Neuropsychology, 28(2), 595–616. http://dx.doi.

org/10.1207/s15326942dn2802.Chasiotis, A., Kissling, F., Hofer, J., & Campos, D. (2006). Theory of mind and inhibitory control in three cultures: Conflict inhibition predicts false belief understanding

in Germany, Costa Rica and Cameroon. International Journal of Behavioral Development, 30(3), 249–260. http://dx.doi.org/10.1177/0165025406066759.Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation analysis for the behavioral sciences (3rd ed.). New York, NY: Routledge.Colmar, S. H. (2014). A parent-based book-reading intervention for disadvantaged children with language difficulties. Child Language Teaching and Therapy, 30, 79–90.

http://dx.doi.org/10.1177/0265659013507296.Cragg, L., & Nation, K. (2007). Self-ordered pointing as a test of working memory in typically developing children. Memory (Hove, England), 15(5), 526–535. http://dx.

doi.org/10.1080/09658210701390750.Deak, G. O. (2000). The growth of flexible problem solving: Preschool children use changing verbal cues to infer multiple word meanings. Journal of Cognition and

Development, 1(2), 157–191. http://dx.doi.org/10.1207/S15327647JCD010202.Dickinson, D. K., & Smith, M. W. (1994). Long-term effects of preschool teachers’ book readings on low-income children’s vocabulary and story comprehension.

Reading Research Quarterly, 29(2), 104–122.Dunn, L. M., & Dunn, D. M. (2007). PPVT-IV: Peabody picture vocabulary test – fourth edition. Circle Pines, MN: American Guidance Service.Ewers, C., & Brownson, S. (1999). Kindergarteners’ vocabulary acquisition as a function of active vs. passive storybook reading, prior vocabulary, and working

memory. Reading Psychology, 20(1), 11–20. http://dx.doi.org/10.1080/027027199278484.Fleury, V. P., Miramontez, S. H., Hudson, R. F., & Schwartz, I. S. (2014). Promoting active participation in book reading for preschoolers with Autism Spectrum

Disorder: A preliminary study. Child Language Teaching and Therapy, 30(3), 273–288. http://dx.doi.org/10.1177/0265659013514069.Hindman, A. H., Skibbe, L. E., & Foster, T. D. (2014). Exploring the variety of parental talk during shared book reading and its contributions to preschool language and

literacy: Evidence from the Early Childhood Longitudinal Study-Birth Cohort. Reading and Writing, 27(2), 287–313. http://dx.doi.org/10.1007/s11145-013-9445-4.

Jaswal, V. K. (2010). Explaining the disambiguation effect: Don’t exclude mutual exclusivity. Journal of Child Language, 37(1), 95–113. http://dx.doi.org/10.1017/S0305000909009519.

Jiménez, T. C., Filippini, A. L., & Gerber, M. M. (2006). Shared reading within Latino families: An analysis of reading interactions and language use. Bilingual ResearchJournal, 30(2), 431–452. http://dx.doi.org/10.1080/15235882.2006.10162884.

Justice, L. M. (2002). Word exposure conditions and preschoolers’ novel word learning during shared storybook reading. Reading Psychology, 23(2), 87–106. http://dx.doi.org/10.1080/027027102760351016.

Kaderavek, J. N., Pentimonti, J. M., & Justice, L. M. (2013). Children with communication impairments: Caregivers’ and teachers’ shared book-reading quality andchildren’s level of engagement. Child Language Teaching and Therapy, 30(3), 289–302. http://dx.doi.org/10.1177/0265659013513812.

Landry, S. H., Smith, K. E., Swank, P. R., Zucker, T., Crawford, A. D., & Solari, E. F. (2012). The effects of a responsive parenting intervention on parent-childinteractions during shared book reading. Developmental Psychology, 48(4), 969–986. http://dx.doi.org/10.1037/a0026400.

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

67

Lee, Y., & Kinzie, M. B. (2011). Teacher question and student response with regard to cognition and language use. Instructional Science, 40(6), 857–874.Lever, R., & Sénéchal, M. (2011). Discussing stories: On how a dialogic reading intervention improves kindergartners’ oral narrative construction. Journal of

Experimental Child Psychology, 108(1), 1–24. http://dx.doi.org/10.1016/j.jecp.2010.07.002.Milburn, T. F., Girolametto, L., Weitzman, E., & Greenberg, J. (2013). Enhancing preschool educators’ ability to facilitate conversations during shared book reading.

Journal of Early Childhood Literacy, 14. http://dx.doi.org/10.1177/1468798413478261.Montgomery, D. E., & Koeltzow, T. E. (2010). A review of the day-night task: The Stroop paradigm and interference control in young children. Developmental Review,

30(3), 308–330. http://dx.doi.org/10.1016/j.dr.2010.07.001.Nicolopoulou, A., Cortina, K. S., Ilgaz, H., Cates, C. B., & de Sá, A. B. (2015). Using a narrative- and play-based activity to promote low-income preschoolers’ oral

language, emergent literacy, and social competence. Early Childhood Research Quarterly, 31, 147–162. http://dx.doi.org/10.1016/j.ecresq.2015.01.006.Paas, F., & Ayres, P. (2014). Cognitive load theory: A broader view of the role of memory in learning and education. Education Psychology Review, 24, 27–45.Paas, F., & Sweller, J. (2012). An evolutionary upgrade of cognitive load theory: Using the human motor system and collaboration to support the learning of complex

cognitive tasks. Education Psychology Review, 24, 27–45.Reese, E., & Cox, A. (1999). Quality of adult book reading affects children’s emergent literacy. Developmental Psychology, 35(1), 20–28. http://dx.doi.org/10.1037//

0012-1649.35.1.20.Sénéchal, M., & Cornell, E. H. (1993). Vocabulary acquisition through shared reading experiences. Reading Research Quarterly, 28(4), 360–374.Schwab, J. F., & Lew-Williams, C. (2016). Repetition across successive sentences facilitates young children’s word learning. Developmental Psychology, 52(6), 879–886.Sesma, H. W., Mahone, E. M., Levine, T., Eason, S. H., & Cutting, L. E. (2009). The contribution of executive skills to reading comprehension. Child Neuropsychology,

15(3), 232–246. http://dx.doi.org/10.1080/09297040802220029.Simpson, A., & Riggs, K. J. (2005). Inhibitory and working memory demands of the day-night task in children. British Journal of Developmental Psychology, 23(3),

471–486. http://dx.doi.org/10.1348/026151005X28712.Strasser, K., Larraín, A., & Lissi, M. R. (2013). Effects of storybook reading style on comprehension: The role of word elaboration and coherence questions. Early

Education & Development, 24(5), 616–639. http://dx.doi.org/10.1080/10409289.2012.715570.Vlach, H. A., & Sandhofer, C. M. (2012). Distributing learning over time: The spacing effect in children’s acquisition and generalization of science concepts. Child

Development, 83(4), 1137–1144.Vlach, H. A., Sandhofer, C. M., & Kornell, N. (2008). The spacing effect in children’s memory and category induction. Cognition, 109(1), 163–167 Elsevier B.V.

Retrieved from http://dx.doi.org/10.1016/j.cognition.2008.07.013.Walsh, B.a., & Blewitt, P. (2006). The effect of questioning style during storybook reading on novel vocabulary acquisition of preschoolers. Early Childhood Education

Journal, 33(4), 273–278. http://dx.doi.org/10.1007/s10643-005-0052-0.Whitehurst, G. J., Falco, F. L., Lonigan, C. J., Fischel, J. E., Debaryshe, B. D., & Caulfield, M. (1988). Accelerating language development through picture book reading.

Developmental Psychology, 24(4), 552–559.Whitehurst, G. J., Arnold, D. S., Epstein, J. N., Angell, A. L., Smith, M., & Fischel, J. E. (1994). A picture book reading intervention in day care and home for children

from low-Income families. Developmental Psychology, 30(5), 679–689.Wilkinson, K. S., & Houston-Price, C. (2012). Once upon a time, there was a pulchritudinous princess...: The role of word definitions and multiple story contexts in

children’s learning of difficult vocabulary. Applied Psycholinguistics, 34(03), 591–613. http://dx.doi.org/10.1017/s0142716411000889.Zucker, T. A., Justice, L. M., Piasta, S. B., & Kaderavek, J. N. (2010). Preschool teachers’ literal and inferential questions and children’s responses during whole-class

shared reading. Early Childhood Research Quarterly, 25(1), 65–83. http://dx.doi.org/10.1016/j.ecresq.2009.07.00.

S.R. Jimenez, M.M. Saylor Cognitive Development 44 (2017) 57–68

68