An Examination of the Relations Between Oral Vocabulary and Phonological Awareness in Early Childhood

Abstract

This article reports a post hoc analysis conducted as part of a larger study in which 61 typically developing, 3-, 4-, and 5-year-olds were assessed in phonological awareness (PA), vocabulary (i.e., receptive, expressive, and definitional), and grammatical skill at baseline and 3, 6, and 9 months later. The larger study’s purpose was to shed light on the theory of lexical reorganization claim that increases in receptive vocabulary provide the basis for PA development. Its results showed that expressive vocabulary accounted for additional PA growth in 4- and 5-year-olds but not in 3-year-olds. The post hoc study involved a closer analysis of PA growth in relation to expressive vocabulary. Its results suggest that increases in receptive vocabulary might be necessary as a foundation for initial PA development but that expressive-level vocabulary might be needed to hold words in memory to perform complex manipulations required in higher level PA tasks.

Keywords

vocabulary phonological awareness Theory of Lexical Reorganization

Background and Intent

The post hoc study presented here was part of a larger study that used individual growth modeling (IGM) to examine the relationship between vocabulary (i.e., receptive, expressive, definitional, and grammatical skill) and phonological awareness (PA) in 3-, 4- and 5-year-olds, over a year. The main, larger study (Cassano, 2013) had three findings: (a) receptive and expressive vocabulary and grammatical skill, though closely related, made independent contributions to PA growth for 3-, 4-, and 5-year-olds; (b) receptive vocabulary had an additional effect on PA growth for 3-year-olds but not for 4- or 5-year-olds; and (c) expressive vocabulary had an additional effect on PA growth for 4- and 5-year-olds but not for 3-year-olds.

It was unclear from the larger study, however, the specific level of PA skill at which receptive vocabulary (i.e., listening vocabulary) no longer made an additional contribution or why the additional contribution of expressive vocabulary (i.e., productive vocabulary) began to emerge. This lack of clarity existed because the PA tasks in the main study were collapsed into a composite variable (i.e., summed score) for the growth models. To recapture the needed information, it was necessary to arrange the individual PA tasks into a continuum of difficulty. This post hoc study was conducted to reveal a clearer picture of PA skill level attained in relation to variations in expressive vocabulary skill.

Supporting Research

Extant research has identified both PA and oral vocabulary as important precursors to reading skill (National Institute of Child Health and Human Development, Early Child Care Research Network [NICHD ECCRN], 2005; Sénéchal, Ouellette, & Rodney, 2006; Storch & Whitehurst, 2002). PA, measured in preschool, strongly predicts word recognition, beginning in first grade, when decoding is the major reading challenge (NICHD-ECCRN, 2005; Storch & Whitehurst, 2002). Kindergarten levels of oral vocabulary add to the prediction of reading comprehension, beginning in the third grade, when high levels of background knowledge and oral language skill are required to understand grade-level texts (Biemiller, 2006; Juel, 2006).

Children with low levels of PA and oral vocabulary often experience reading difficulties throughout elementary school. For example, Spira, Bracken, and Fischel (2005) found that 75% of children who scored below the 30th percentile in reading achievement at the end of first grade remained there through the fourth grade, or fell even lower. This rate of continued failure was consistent with, though slightly lower than, the rate of 88% identified in Juel’s (1988) study, which also followed first graders with reading difficulties through the fourth grade.

At present, PA instruction is emphasized heavily in many preschool and kindergarten classrooms (Neuman, 2006; Paris, 2005). This emphasis is likely due to the firmly established effect of PA on word-recognition skill (Storch & Whitehurst, 2002) and to the ease with which PA can be assessed (Paris, 2005). PA instruction is often limited to isolated tasks, such as clapping syllables in words or isolating a word’s phonemes, verbally (Lonigan, Allan, & Lerner, 2011). Devoting significant portions of instructional time to PA, alone, may increase PA and subsequent word reading skill, but at a high cost, especially if oral vocabulary is not given adequate attention (Dickinson, Golinkoff, & Hirsh-Pasek, 2010). Early vocabulary and content knowledge anchor the development of new knowledge (Juel, 2006). Thus, failing to provide young children with ample word and world learning opportunities can lead to “knowledge gaps” (Neuman, 2006). Providing a balance in early literacy programs, between the foundations of word recognition and comprehension, is an important goal. Support for this goal might increase if the relationship between vocabulary development and PA development could be understood in greater detail.

Theoretical Framework

The frequency and complexity of child-directed speech (Hart & Risley, 1995; Huttenlocher, Haight, Bryk, Seltzer, & Lyons, 1991; Rowe, 2012), as well as extended discourse opportunities (Dickinson & Tabors, 2001; Wells, 1985), are associated with vocabulary development. Less is known, however, about the developmental origins of PA, although the theory of lexical reorganization (LRT) claims that growth in PA occurs as a function of receptive vocabulary knowledge (Metsala, 2011; Metsala & Walley, 1998). According to LRT, early in development, when a child’s lexicon is small, words are stored holistically (i.e., little segmental detail is captured) and recognized by overall acoustic shape. This type of storage works because few sound similarities exist among words in a small lexicon (e.g., milk, banana, and kitty). When the lexicon increases, however, words begin to overlap in their phonology (e.g., cat, cut, and kite). The LRT asserts that these phonologically similar words are reorganized and restored, segmentally, to allow differentiation from one another. According to the LRT, this reorganization (i.e., from storage of larger to smaller linguistic units) facilitates greater attention to phonemes, which serves as a developmental substrate for skill development in PA.

Support for the LRT comes from studies that either identified relationships between PA and vocabulary in early childhood (Dickinson, McCabe, Anastasopoulos, Peisner-Feinberg, & Poe, 2003; Lonigan, 2007; Sénéchal et al., 2006; Schwarz, Bowey, & Burnham, 2006) or identified vocabulary and PA abilities as commensurate skills (Smith, McGregor, & DeMille, 2006; A. L. Williams & Elbert, 2003). Other research identified poor expressive vocabulary as a “limiting factor” in PA interventions. That is, children with lower levels of expressive vocabulary were less likely to benefit from intensive PA instruction than their more language-advantaged peers (Whiteley, Smith, & Connor, 2007).

Research Questions and Hypotheses

It is clear that vocabulary knowledge and PA are correlated. The nature of this correlation, however, as captured in the LRT, has not been well established. To explore the validity of the LRT, longitudinal research is needed to record the development of both PA and vocabulary (i.e., receptive and expressive) during critical periods of the preschool and kindergarten years. The larger study (Cassano, 2013), of which this post hoc study was a part, was designed to obtain these data. Although the 2013 study’s results indicated that receptive and expressive vocabulary and grammatical skill were related to growth in PA in 3-, 4-, and 5-year-olds, questions remained about an observed age-related change in the relations between receptive and expressive vocabulary, and PA. Specifically, the larger study did not explain the additional main effect found for expressive vocabulary, in 4- and 5-year-olds, only.

To explore this finding further, this post hoc study focused on three questions: (a) PA tasks involving smaller linguistic units are performance differences between older (i.e., 4- and 5-year-olds) and younger (i.e., 3-year-olds) children larger than when tasks involve larger linguistic units; (b) PA tasks involving smaller linguistic units, do children with higher levels of expressive vocabulary outperform children with lower levels of expressive vocabulary; and (c) do performance differences related to expressive vocabulary increase as task difficulty increases, especially as manipulations (e.g., blending, segmenting, and deleting), but not linguistic units, vary across tasks?

These three questions were related to three research hypotheses for this post hoc study about the likelihood that lexical reorganization is the only catalyst for the attainment of higher levels of PA. Specifically, it was hypothesized that (a) larger performance differences would be found between older participants (i.e., with larger vocabularies, on average) and younger participants (i.e., with smaller vocabularies, on average) when PA tasks involved smaller linguistic units; (b) children with higher levels of expressive vocabulary would outperform participants with lower levels on tasks involving smaller units; and (c) performance differences related to expressive vocabulary would not increase much, even as manipulations varied across PA tasks.

Method/Techniques

Sixty-one 3-, 4-, and 5-year-olds, from private-pay preschool/kindergartens, participated in the larger, yearlong study. Preschool directors identified 11% as low socio-economic status based on income-based tuition assistance criteria. They also identified 93% as Caucasian, 5% as Hispanic, and 2% as African American. The participants were all “typically developing,” spoke English as their primary language, and had no history of language or hearing delays, per parent report. Two participants received speech services for minor articulation errors (e.g., /wain/ for /rain/) of the kind researchers have found insignificant for PA development and task performance (Rvachew, Ohberg, Grawburg, & Heyding, 2003).

Table 1 outlines the comprehensive battery of measures used to assess PA, vocabulary (i.e., receptive, expressive, and definitional), and grammatical skill, at baseline, and at 3, 6, and 9 months later. All assessments were administered at each testing point by a trained examiner, at each child’s preschool.

Table 1.

Vocabulary and Phonological Awareness Measures.

Instrument	Domain Measured	Sample Prompts
Expressive Vocabulary Test-2 (EVT-2; K. T. Williams, 2007)	Expressive vocabulary	What do you see … [tree]? Tell me another name for … [money]?
Peabody Picture Vocabulary Test-4 (PPVT-4; Dunn & Dunn, 2007)	Receptive vocabulary	Point to … Show me …
Test of Preschool Early Literacy (TOPEL) definitional vocabulary subtest (Lonigan, Wagner, Torgesen, & Rashotte, 2007)	Definitional vocabulary	What do you call this [pocket]? What does it do?
Morphological Generation Task (Muter, Hulme, Snowling, & Stevenson, 2004)	Morphological closure (i.e., grammatical skill)	Here is one mouse. Here are two … [mice]
TOPEL phonological awareness subtest (Lonigan et al., 2007)	Phonological awareness (multilevel)	What is hot—dog together? What is f--ox? What is f—i—sh? What is playground without ground? What is heat without /t /?
Rhyme Detection (Fernandez-Fein & Baker, 1997)	Phonological awareness (rhyming)	Does cat rhyme with hat or does cat rhyme with bell?
Rhyme Production (Fernandez-Fein & Baker, 1997)	Phonological awareness (rhyming)	What rhymes with cat?
Syllable Segmentation (Mann & Liberman, 1984)	Phonological awareness (syllable)	Clap (or count) the syllables butterfly
Yopp-Singer test of Phoneme Segmentation (Yopp, 1995)	Phonological awareness (phoneme)	What are the sounds in dog?

Continuum of Task Difficulty

For the post hoc study, the 10 individual PA tasks used in the larger study were analyzed for difficulty level and then ranked, from easiest to hardest. This ranking was necessary for determining the specific level of PA skill development achieved by the 4- and 5-year-olds for whom an added effect of expressive vocabulary was found. Difficulty judgments (i.e., the placement of each task) were based on specific task features including (a) linguistic unit size (e.g., syllable, onset–rime, or phoneme; Liberman, Shankweiler, Fischer, & Carter, 1974; Treiman & Zukowski, 1991), (b) the task operation required to perform with the linguistic units (e.g., detecting, producing, blending, segmenting, or elision; Schatschneider, Francis, Foorman, Fletcher, & Mehta, 1999; Stahl & Murray, 1994; Yopp, 1988), (c) the response mode (e.g., verbal or nonverbal), and (d) the support/demand provided. Support/demand was operationalized as (a) any task presentation format (e.g., picture prompt or forced choice), (b) requirement (e.g., memory or motor demand), (c) prompt type (e.g., words or pseudo-words), or (d) answer format (e.g., 2- vs. 4-item, forced choice) that might likely increase or decrease the task difficulty.

Linguistic unit of analysis

Tasks were first organized according to linguistic unit size (i.e., from largest to smallest). Tasks involving smaller units were considered generally more difficult than tasks involving larger ones (Liberman et al., 1974; Treiman & Zukowski, 1991), although categorization challenges were encountered. For example, four blending items in the Test of Preschool Early Literacy (TOPEL; Lonigan, Wagner, Torgesen, & Rashotte, 2007) were segmented oddly (i.e., not in typical syllable, onset–rime, or phoneme units) when presented by the examiner (e.g., ca—t or sew—k). Children could easily use a gating procedure, similar to any outlined in speech recognition studies, to produce the correct word (Garlock, Walley, & Metsala, 2001). That is, because most of a word is presented in the first segment, “blending” the final phoneme segment provided by the tester would likely differ from both true phoneme blending tasks in which all of the individual phonemes in a three-phoneme word must be blended (e.g., /f/ /i/ /sh/) and onset–rime blending tasks (e.g., /k/ /at/). Thus, these items were categorized as “atypical” and assigned a lower difficulty rating than both onset–rime and phoneme blending items (i.e., Level 2 vs. Level 3 and Level 4, respectively).

A second challenge in determining difficulty level occurred with the Yopp-Singer phoneme segmentation task (Yopp, 1995). Specifically, 7 of the 22 items are words comprised of a single-consonant onset and a vowel-only rime (e.g., lay, no, and zoo). Thus, for these two-phoneme items, the task requires onset–rime segmentation, not phoneme segmentation, despite the fact that the child must produce two distinct phonemes. The scoring directions do not make any distinction between these 7 items and the remaining 15, which do require phoneme segmentation. Thus, for this study, the onset–rime and phoneme-level segmentation items were categorized separately, with onset–rime segmentation items placed lower on the continuum of difficulty (Level 3 vs. Level 4).

A third challenge, which came to light during the administration of the phoneme segmentation task, involved scoring. It was not uncommon for a child to incorrectly segment a three-phoneme word (e.g., grew) into phonemes but correctly into onset and rime units (i.e., /gr/ /ew/). For the purpose of this study, any word in the phoneme segmentation task list that was segmented correctly into onset and rime was given a point in the onset–rime segmentation category. Although this decision resulted in wide variation from child to child in the number of items included in the onset–rime category, scoring the Yopp-Singer items in this manner provided more distinct levels in the PA tasks for the different difficulty categories, and more items, overall, for judging a child’s onset–rime segmentation skill.

Task operation

Consistent with previous research, the tasks for the post hoc study were first sorted into two task operation (i.e., linguistic unit manipulation) categories: synthesis and analysis (Stanovich, Cunningham, & Cramer, 1984). Synthesis tasks require the blending of syllables, onset–rime units, or phonemes, presented by the examiner, whereas analysis tasks require child judgment (e.g., Rhyme Detection), manipulation of segments (e.g., Phoneme Segmentation), or production of examples that match the target word provided (e.g., Rhyme Production). Further difficulty judgments were then made for the tasks within the analysis category. Production tasks were ranked as more difficult than detection tasks (e.g., Rhyme Production vs. Rhyme Detection), and elision tasks were ranked as more difficult than segmentation tasks (Syllable Elision vs. Syllable Segmentation; Stahl & Murray, 1994). Finally, difficulty ratings were assigned to the tasks within the synthesis (i.e., Level 1–4) and analysis (i.e., Level 1–6) categories. Although the six levels first obtained for the analysis tasks accounted most accurately for the full range of task features, it was necessary to reduce the number of levels to four to obtain a more even distribution of items across the different levels.

Response mode and task support/demand

Before finalizing difficulty ratings, response mode and task support/demand were examined for each PA measurement item. Response modes included verbal or nonverbal (i.e., pointing to a picture), and motor (i.e., clapping). Supportive response modes included picture prompts and forced- or multiple-choice formats (i.e., 2 or 4 items presented, respectively). Task demands (i.e., response modes that likely increase difficulty) included memory requirements (i.e., number of steps involved while remembering the target word) and motor responses (e.g., clapping syllables).

Although linguistic unit size and complexity of a task’s operation were the primary criteria used to determine task difficulty, consideration of the response mode and/or amount of support/demand led the researchers to give some tasks either a lower or a higher rating than it would have otherwise received. For example, Phoneme Elision is frequently considered a high-level PA task because it requires a complex, two-step operation on a small linguistic unit. In this study, however, half of the Phoneme Elision items required children to select the appropriate picture by pointing to one picture (i.e., nonverbal response) from among four choices (i.e., forced choice). Moreover, for the items requiring a verbal response, the correct response obtained, by deleting the targeted phoneme, was an actual word (e.g., “What is lamp without /p/”: lamb; Lonigan et al., 2007). Thus, for such items, there was both picture and semantic support for the correct response. This kind of support is not available with a task word such as lunch (i.e., “What is lunch without /ch/?”). Thus, the Phoneme Elision task was assigned a difficulty Level of 3, not 4.

A final challenge in establishing a continuum of PA task difficulty resulted from the varying number of items for each type of task. For example, of the 27 items on the TOPEL (Lonigan et al., 2007), only 2 required Phoneme Blending. In contrast, there were 42 Syllable Segmentation items on the assessment selected for the main study (Mann & Liberman, 1984). This unevenness created a sampling problem for the PA assessment. To achieve a better balance in the number of items, only half of the Syllable Segmentation items were used. Those selected included the same proportion of single syllable, multisyllabic, and compound word items as the full battery.

Table 2 displays the 11 PA tasks in the order of their estimated difficulty by operation category. Based on an analysis of response mode and task support/demand, higher mean scores were expected for the lower level skills within each operation category.

Table 2.

Estimated Levels of Difficulty by Operation, Linguistic Units, and Task Support/Demand.

Linguistic Unit/Operation	Item Total	Estimated difficulty	Response Mode	Support/Demand
Synthesis
Syllable blending	6	Level 1	Nonverbal Verbal	Half of forced choice (four pictures) semantic support
Atypical blending	4	Level 2	Nonverbal Verbal	Half of forced choice (four pictures)
Onset–rime blending	3	Level 3	Nonverbal Verbal	Two third of forced choice (four pictures)
Phoneme blending	2	Level 4	Verbal
Analysis
Syllable segment	21	Level 1	Vocal-motor	Motor requirement
Rhyme detection	10	Level 2	Verbal	Forced choice (two word pairs)
Syllable elision	6	Level 2	Nonverbal Verbal	Half of forced choice (four pictures) semantic support
Phoneme elision	6	Level 3	Nonverbal Verbal	Half of forced choice (four pictures) semantic support
Onset–rime segment	19	Level 3	Verbal
Rhyme production	8	Level 4	Verbal
Phoneme segment	15	Level 4	Verbal

Results

Raw scores were generated for each participant for each task, at each testing point. Because the number of items varied across tasks (i.e., from 2 to 21), mean percentage correct scores were calculated to allow for comparison. Descriptive statistics by age-group at baseline (i.e., Time 0) are presented in Table 3.

Table 3.

Obtained Item Difficulty Including Mean Percentage Correct (Standard Deviations) for Phonological Awareness Tasks at Baseline.

Linguistic Unit/Operation	Estimation Level	Sample (N = 53)	3-Year-Olds (n = 17)	4-Year-Olds (n = 23)	5-Year-Olds (n = 13)
Syllable blending	1	.89 (.17)	.80 (.22)	.92 (.13)	.95 (.11)
Atypical blending	2	.75 (24)	.66 (.23)	.75 (.25)	.87 (.17)
Onset–rime blending	3	.61 (.32)	.45 (.33)	.63 (.29)	.79 (.26)
Phoneme blending	4	.46 (.45)	.24 (.36)	.48 (.49)	.73 (.33)
Syllable segmentation	1	.26 (.30)	.10 (.16)	.30 (.36)	.38 (.24)
Rhyme detection	2	.61 (.30)	.38 (.21)	.63 (.30)	.84 (.19)
Syllable elision	2	.55 (.37)	.25 (.25)	.56 (.35)	.91 (.16)
Phoneme elision	3	.31 (.29)	.12 (.17)	.29 (.28)	.61 (.18)
Onset–rime segmentation	3	.05 (.13)	0 (0)	.03 (.08)	.18 (.22)
Rhyme production	4	.40 (43)	0 (0)	.42 (.44)	.84 (.19)
Phoneme segmentation	4	.04 (.12)	0 (0)	.04 (.13)	.11 (.17)

Note. N = 53.

Performance Differences by Age-Group

The first question explored in this post hoc study focused on performance differences between older (i.e., 4- and 5-year-olds) and younger (i.e., 3-year-olds) children. An examination of baseline performance revealed three patterns, the first related to age. Specifically, 5-year-olds outperformed 4-year-olds on all tasks, and 4-year-olds outperformed 3-year-olds. Performance across tasks for each age-group, however, was consistent with Anthony and colleagues’ (2002). Anthony, Lonigan, Driscoll, Phillips, and Burgess’s (2003) description of PA development as occurring in overlapping stages, not in a lock-step sequence. The second pattern observed related to linguistic unit of analysis, particularly in the synthesis tasks. Higher means were observed for Syllable Blending than for Onset–Rime Blending, and for Onset–Rime Blending compared to Phoneme Blending (Treiman & Zukowski 1991). Third, a pattern related to the required task operation was also evident. As found in previous research, performance of each age-group was consistently better on synthesis tasks compared to analysis tasks (Stahl & Murray, 1994; Stanovich et al., 1984). Within the analysis tasks, each age-group performed markedly better on Rhyming Detection compared to Rhyming Production. In fact, none of the 3-year-olds produced a single rhyme accurately at the start of the study. The fourth and last pattern observed related to task support/demand (Stahl & Murray, 1994; Yopp, 1988). That is, when additional support was provided (e.g., forced choice) and/or the demands were low (e.g., nonverbal response—pointing), higher mean scores were often observed. Determining the role of task support/demand in performance, however, was not straightforward, in part because it was often confounded with response mode. The issues related to task support/demand are discussed further below.

A comparison of the results reported in Table 3 with the estimated levels of difficulty in Table 2 indicated that the children generally performed as predicted. That is, lower mean scores, represented as the mean percentage correct, were observed for tasks rated as more difficult. Syllable Segmentation tasks, however, were an exception. Specifically, 3- and 4-year-olds unexpectedly performed equally well on the Syllable Segmentation and Phoneme Elision tasks, while 5-year-olds performed better on the latter.

A close inspection of these tasks suggested that results might have been due to the response mode and task support/demand for the Phoneme Elision task, which made it simpler than expected, if considering only linguistic unit. Specifically, half of the Phoneme Elision tasks include picture prompts and forced choice. Although pictures are not provided on the remaining three tasks, the correct response is always a real word (e.g., “What is heat without the /t/?”; Lonigan et al., 2007). Perhaps some children completed this task using global sound similarities (Gombert, 1992; Muter, Hulme, Snowling, & Taylor, 1997), not the more complex two-step manipulation (i.e., segmenting the last phoneme from the word and then deleting it). For example, on the TOPEL (Lonigan et al., 2007), a child might have pointed to the picture of tea steaming in a cup because its name “sounds the most like” the target word in the prompt: “Point to tease without /z/.”

It is also possible that the use of meaningful segments in Phoneme Elision tasks reduces cognitive inhibition. That is, when a nonmeaningful segment remains as a result of correctly performing phoneme elision, very young children often have difficulty disregarding a well-established or preferred response (i.e., saying a complete word as children always do when speaking; Harnishfeger, 1995). Inhibition, though ordinarily difficult for 3- and 4-year-olds, is likely diminished in the PA tasks used in this study, because a meaningful word results. This situation might allow even 3-year-olds to perform better than expected.

In comparison, Syllable Segmentation, though involving larger and more easily discernable linguistic units, differed from Phoneme Elision items in task support/demand and response mode. Specifically, an accurate response requires the child to clap (i.e., motor demand) at each syllable juncture. Yet, to do this, a child must hold the word in memory as it is segmented (i.e., memory demand). In other words, the response mode (i.e., clapping) increases the task difficulty because it is a representation or demonstration of what has already taken place cognitively. Thus, the response mode for the Syllable Segmentation task likely increases its difficulty beyond what would be expected when considering size of linguistic unit and complexity of operation, alone.

Also notable during the administration of this task, all but one of the participants responded by clapping while simultaneously repeating the target word. Clapping while saying the word offers an additional motor demand, because clapping must be appropriately timed to occur in sync with the child’s verbal segmenting. Thus, although the response mode was originally recorded as “motor (i.e., clapping),” the term “vocal-motor matching” is a more accurate description of how the majority of the children responded. Perhaps they were simply mimicking what was modeled when the task was introduced (e.g., “Listen to the word, but—ter—fly [examiner claps at each syllable juncture]. Butterfly has three syllables or claps.”). Or, perhaps, and more likely, they repeated or subvocalized the target words to cope with memory demands (Baddeley, 1986). It seems unlikely to these researchers that clapping would aid segmentation. It seems more likely that it represents segmentation that has already occurred.

Interrelations Between Vocabulary and PA Performance

The second and third questions in the post hoc analyses addressed interrelations between PA task and expressive vocabulary performances. These questions were examined through correlations and age- and vocabulary-group comparisons. Bivariate correlations for each PA task, and for receptive and expressive vocabulary, were computed. The Atypical task was removed from all analyses because it is not a common PA measure. Onset–rime segmentation was also removed because the number of items varied for each participant.

Despite the small sample size (N = 53), receptive vocabulary was significantly correlated with the majority of the PA tasks (Figure 1). The exceptions were for Syllable Segmentation (r = .19) and Syllable Blending (r = .22). The strongest correlations between receptive vocabulary and PA tasks were observed for Rhyme Production (r = .63), Syllable Elision (r = .61), and Rhyme Detection (r = .61). Expressive vocabulary was moderately to strongly related to all of the PA tasks, with the strongest correlations observed for Syllable Elision (r = .72), Rhyme Production (r = .63), and Rhyme Detection (r = .61). The PA tasks with the weakest correlations with expressive vocabulary were Syllable Blending (r = .28) and Phoneme Segmentation (r = .39).

Figure 1.

Correlations of between phonological awareness tasks and receptive and expressive vocabulary (N = 53).

For ease of comparison, the correlation coefficients for receptive and expressive vocabulary and each PA task are displayed in a bar graph (Figure 2), with the PA tasks arranged by obtained item difficulty (Table 3). Visual inspection of Figure 2 suggests that the relationship between PA and each form of vocabulary varied with PA task difficulty.

Figure 2.

A comparison of the correlations between receptive and expressive vocabulary on phonological awareness tasks at the first testing point.

The differences between the correlation coefficients for vocabulary predictors and each PA task were explored by subtracting the larger correlation coefficient (i.e., Pearson r) from the smaller (i.e., difference scores). With only two exceptions, Rhyme Production and Phoneme Elision, the absolute differences scores were larger for expressive vocabulary (range = 0–.23). Although none of these differences reached statistical significance, in part due to sample size, some findings are interesting to think about. Specifically, difference scores related to expressive vocabulary were the largest for Syllable Segmentation (.29) and Phoneme Segmentation (.09; i.e., the most difficult tasks) as well as Syllable Elision (.09; i.e., the task with the strongest correlation with expressive vocabulary). The results of these analyses suggest that although both receptive and expressive vocabulary are correlated significantly with performance on most PA tasks, the correlations with expressive vocabulary were larger, though not significantly, for some of the more complex tasks.

Next, the sample was divided at the mean into High and Low performance on expressive vocabulary (i.e., one group above the mean; the other below, at first and final testing points, see Figure 3). Although the groups were categorized as High and Low performance groups, all participants were typically developing, with age-appropriate expressive vocabulary scores. In other words, these data do not compare children with average vocabulary abilities to children with below average abilities.

Figure 3.

Comparison of performance on phonological awareness tasks for High and Low performance groups in expressive vocabulary.

The mean age for the High group was higher (in months) than for the Low group (i.e., 43 compared to 51 months at Time 0 and 53 compared to 63 months at Time 3). The mean percentage correct scores, as well as the difference between the scores, were then calculated for each group for each PA task.

With regard to the second research question, children with higher levels of expressive vocabulary frequently outperformed children with lower levels on PA tasks involving smaller linguistic units. As Table 4 shows, larger performance differences were found between the High/Low vocabulary groups when Phoneme Blending (24% and 9% difference at Time 0 and Time 3, respectively) was compared to Syllable Blending (9% and 2%), but not when Phoneme Elision (28% and 19% difference at Time 0 and Time 3, respectively) was compared to Syllable Elision (37% and 14%).

Table 4.

Mean Percentage Correct Scores and Percentage Differences in Performance by Vocabulary Group (High vs. Low) for the First and Final Testing Point.

		Time 0			Time 3
Linguistic Unit/Operation	Max Score	Low (n = 28)	High (n = 28)	Percentage Difference	Low (n = 25)	High (n = 27)	Percentage Difference
Synthesis
Syllable blending	6	.83	.92	.09	.97	.99	.02
Onset–rime blending	3	.49	.72	.23	.59	.83	.24
Phoneme blending	2	.35	.59	.24	.54	.63	.09
Analysis
Syllable elision	6	.35	.72	.37	.76	.90	.14
Phoneme elision	6	.17	.45	.28	.51	.70	.19
Rhyme detection	10	.43	.70	.30	.76	.88	.12
Syllable segment	21	.15	.33	.18	.50	.63	.13
Rhyme production	8	.17	.60	.43	.53	.78	.25
Phoneme segment	16	0	.08	.08	.09	.21	.11

The third research question explored whether performance differences related to expressive vocabulary increased as PA task difficulty increased. At baseline (i.e., Time 0), the greatest performance differences were observed on the Syllable Elision and Rhyme Production tasks (37% and 43% difference, respectively)—the same two tasks for which the strongest correlations with expressive vocabulary were observed (r = .72 and r = .63, respectively, Figure 1). The tasks with the smallest performance differences at baseline were Syllable Blending and Phoneme Segmentation (9% and 8%, respectively)—the same two tasks that showed the smallest correlations with expressive vocabulary (r = .28 and r = .39). Performance differences decreased from Time 0 to Time 3, as mean scores increased for both groups on all PA tasks. One way to view this result is to consider that, as expressive vocabulary increased in the low group, smaller difference scores were observed.

Summary of the Post Hoc Study Results

Exploratory post hoc analyses were conducted to examine more closely the age-related differences found in the relationship between PA and receptive and expressive vocabulary found in the growth models identified in the main study (Cassano, 2013). The results suggest that higher levels of expressive vocabulary (i.e., higher scores on the measures) are likely required to complete PA tasks involving the most difficult operations and higher task demands, even when the level of linguistic unit in a task is large (i.e., presumably easy to detect and manipulate). This is potentially an important finding for clarifying when the LRT exerts its effect, and what else is required beyond this effect, for more difficult PA tasks. Results from the post hoc study suggest that the “what else” might include expressive-level vocabulary.

To review, if lexical reorganization is the only catalyst for attainment of PA skill, then larger performance differences would be expected on tasks involving smaller linguistic units (e.g., phonemes) compared to larger linguistic units (e.g., syllables and onset–rimes), and, therefore, between older (i.e., 4- and 5-year-olds) and younger (i.e., 3-year-olds) age-groups, because the LRT claims that words overlapping in phonological structure are restored from larger units to smaller ones. This hypothesis was only partially supported by the post hoc analyses. Higher levels of performance were typically observed for (a) older children, (b) larger linguistic units, and (c) less challenging operations (e.g., blending vs. segmentation), although variations in task support/demand frequently “disrupted” these patterns (e.g., Syllable Segmentation vs. Phoneme Deletion). That is, as the operation/task demands increased (e.g., involved additional memory and/or motor demands), performance usually decreased, particularly for younger children. When all of the task features (i.e., linguistic unit, operation, response mode, and task support/demand) were considered, 3-year-olds typically performed well on tasks involving larger linguistic units; but, unexpectedly in terms of the LRT, they also performed well on tasks involving smaller linguistic units when task support was high (e.g., picture prompts, forced choice, and segments that were actual words), and/or task demand was low (e.g., nonverbal—pointing, detection) not high (i.e., verbal or motor response, production or segmentation) regardless of linguistic unit size. This good performance on smaller linguistic units is not consistent with the LRT explanation of PA skill development.

With regard to the second research hypothesis and the LRT, it was expected that when the PA task requirements included operations with smaller linguistic units (e.g., phonemes), the performance of children with higher levels of expressive vocabulary (e.g., above the mean) should exceed the performance of children with lower levels (e.g., below the mean). This hypothesis was supported only for the less difficult tasks, not for the more complex tasks. This result is inconsistent with the claims of the LRT. Specifically, as LRT asserts, with vocabulary growth, phonologically similar words must be reorganized and restored segmentally to allow differentiation from one another. This reorganization prompts greater attention to and learning about phonemes. Thus, differences between High/Low vocabulary groups should be evident on all, not just some, phoneme-level tasks. Given the small sample size, however, replication with a larger sample is required before any firm conclusions can be drawn.

With regard to the third research hypothesis and the LRT, it was expected that PA performance differences related to expressive vocabulary should increase little, if at all, as PA task difficulty increases, even as the required operation varies across tasks because the LRT bases reorganization of words in the lexicon on receptive- not expressive-level vocabulary (Metsala, 1999). This hypothesis was not supported by this analysis. Variations in both the size and the strength of the correlations between the vocabulary predictors and the PA tasks, as well as the results of the High/Low vocabulary-group comparisons, suggest that some PA tasks, more than others, require children to draw from their expressive vocabulary stores. For example, one of the larger vocabulary-group performance differences was observed for Rhyme Production. Thus, although performance on Rhyme Production improved for all children over the course of the yearlong study (Cassano, 2013) and was strongly related to both receptive and expressive vocabulary, considerable performance differences were evident when expressive vocabulary was considered. This finding was not unexpected when considering the cognitive challenge of the task. Specifically, Rhyme Production requires children to draw from their expressive vocabulary stores to produce an appropriate word. In addition, the target word must be kept in memory, while the child searches vocabulary stores. If not known well (i.e., at least the expressive level), a child is likely to forget the target.

Implications

The findings of this post hoc and the larger original study raise important questions about critical assumptions on which LRT is based. Specifically, according to the theory, increases in the size of a child’s receptive vocabulary prompt segmental reorganization of phonologically similar words (i.e., words that differ by a single phoneme) in the lexicon. The LRT, however, does not indicate that expressive vocabulary might become increasingly more important for PA task performance, as children move from 3 to 4 and 5 years of age. In other words, at present, the LRT does not specify any limits to which lexical reorganization might support PA skill development.

The increase in the power of expressive vocabulary, in this study, to predict PA growth in 4- and 5-year-olds suggests that PA development might require more than continued rounds of lexical reorganization. Perhaps the ability to use a word expressively, acquired from multiple experiences, helps to solidify its phonological representation (i.e., sound structure), which, in turn, makes its segmental structure more accessible for conscious manipulation as is required in higher level PA tasks. The results of this post hoc study suggest that larger expressive vocabulary stores are likely required to complete PA tasks involving both the most difficult operations and the higher task demands, even if the linguistic unit is large (e.g., Syllable Elision). Thus, the LRT, with its central focus on lexical reorganization, might not provide the full foundation needed to support progress in PA to the higher levels.

The results from both the main study and the post hoc analyses have implications mostly for future research not for early childhood practices. Nevertheless, a few cautious words will be offered about implications for classroom practices, before the implications for future research are discussed.

Implications for Classroom Practices

This study, like many previous studies, found strong correlations between oral vocabulary and PA. Although this study cannot support the conclusion that this relationship is causal (i.e., increasing vocabulary size leads to greater PA development), its findings do suggest that early childhood teachers and program developers might consider carefully their use of curricula that assume PA development comes primarily or only from explicit instruction using PA tasks, especially if the amount of instructional time devoted to these tasks leaves little time for opportunities to support oral vocabulary development. In the event that the relationship between oral vocabulary and PA is causal, an instructional program that is more balanced between PA skill development and vocabulary development might be prudent.

Because this study also found an increasingly stronger relationship between expressive vocabulary and PA in older (i.e., 4- and 5-year-olds) compared to younger (i.e., 3-year-olds) children, a classroom focus on expressive-level knowledge of words, when supporting children’s vocabulary development, rather than a focus primarily or only on receptive-level knowledge, would seem wise. Given that expressive-level vocabulary knowledge is also known to provide better support for reading comprehension than what receptive knowledge provides (National Early Literacy Panel, 2008), a move in instruction in the early years toward supporting a higher level of vocabulary knowledge is already warranted.

Implications for Future Research

The implications for future research include the need for two distinct, yet related, lines of investigation. First, future research is needed to delineate the actual levels of difficulty in PA tasks used with young children, and second, the role of receptive and expressive vocabulary on PA performance needs more specification.

Identify a continuum of PA task difficulty

Various measurement tools, if used together, as in the current study, provide material from which a continuum of PA tasks can be created. As already discussed, however, it is difficult to create an ideal, fine-grained continuum of tasks using currently available items. Moreover, the presumed difficulty level of each task, determined at first in this study by analyzing each task in relation to multiple criteria, must be tested. As was found in these post hoc analyses, some tasks were found to be either easier or more difficult than expected when judgments were based on criteria used initially to place each task onto a continuum.

Other researchers have documented issues related to PA assessments. For example, Cassady, Smith, and Putnam (2008) assert that, “disparate findings in the research on phonological awareness development have their roots in limitations imposed by instrumentation or measurement” (p. 530). Although these researchers’ main concern was for adequate assessment tools to inform PA instruction and the identification of children with potential reading difficulties, it is also important to have fine-grained PA measures for conducting research on vocabulary and PA relations. If the task difficulty levels are not correctly determined to establish an accurate continuum, conclusions drawn about the relationship between receptive and expressive vocabulary, and PA development, are compromised.

Although there is considerable research on the developmental progression of PA related to the linguistic unit of analysis in tasks (Liberman et al., 1974; Treiman & Zukowski, 1991), and additional research on how PA should be operationalized and defined (Anthony, Lonigan, Driscoll, Phillips, & Burgess, 2003; Stanovich et al., 1984), there is presently an insufficient research base to allow the disentangling of the effects of linguistic unit, required operation, response mode, and task support/demand on PA task difficulty. This situation clouds the interpretation of performance scores. When conducting research that seeks to understand the relationships between vocabulary and PA, and especially the relationship between expressive-level vocabulary and development of higher level PA, it is important to know exactly what each PA task measures, and where it actually falls on a continuum of PA task difficulty. A clearly vetted battery of PA tasks would be immensely useful for future studies of vocabulary and PA relationships. Such a battery does not now exist.

Specify the role of vocabulary in PA performance

Future investigations must also examine how receptive- and expressive-level knowledge of specific words impacts performance on PA tasks across a range of difficulty, as well as why poor expressive vocabulary is a limiting factor in PA performance (Whiteley et al., 2007). For example, it would be useful to know the extent to which children’s familiarity with the specific words used on an assessment (i.e., receptive or expressive knowledge), not just a child’s overall vocabulary size, impacts performance on PA tasks across a range of difficulty.

The relationship between familiarity (i.e., high vs. low age of acquisition [AOA]) and performance on PA tasks has already been explored in research (e.g., Metsala, 1999; Sosa & Stoel-Gammon, 2012), and evidence suggests that performance on PA tasks is higher for more familiar (i.e., lower AOA) words. An important caveat, however, is that, in AOA studies, the familiarity ratings are assigned to the words based on studies of general vocabulary development in young children. Any participant’s actual familiarity with a word on PA tasks has not been examined. As a consequence, we know little about the word-level vocabulary demands of PA tasks frequently used in research (or in classroom practice).

These suggested investigations are needed to obtain information useful for revising PA assessments and deepening our understanding of the relationships between PA and vocabulary. The knowledge gained through future research along the lines described here could provide important information about the adequacy of the theory of lexical reorganization in explaining the developmental of PA. More information about this theory could then help to inform instructional practices used in the early years.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Funding was provided by the Woman’s Guild Grant (Boston University, 2010), Judith A. Schickedanz Scholarship (Boston University, 2010), Helen M. Robinson Grant (International Reading Association, 2009).

References

Anthony

J. L.

Lonigan

C. J.

Burgess

S. R.

Driscoll

Phillips

B. M.

Cantor

B. G.

(2002). Structure of preschool phonological sensitivity: Overlapping sensitivity to rhyme, words, syllables, and phonemes. Journal of Experimental Child Psychology, 82, 65–92.

Anthony

J. L.

Lonigan

C. J.

Driscoll

Phillips

B. M.

Burgess

S. R.

(2003). Phonological sensitivity: A quasi-parallel progression of word structure units and cognitive operations. Reading Research Quarterly, 38, 470–487.

Baddeley

A. D.

(1986). Working memory. Oxford, England: Clarendon Press.

Biemiller

(2006). Vocabulary development and instruction: A prerequisite for school learning. In Neuman

Dickinson

(Eds.), Handbook of early literacy research (Vol. II, pp. 41–51). New York, NY: Guilford.

Cassady

J. C.

Smith

L. L.

Putman

S. M.

(2008). Phonological awareness development as a discrete process: Evidence for an integrative model. Reading Psychology, 29, 508–533.

Cassano

C. M.

(2013). An examination of growth in vocabulary and phonological awareness in early childhood: An individual growth model approach (Doctoral dissertation). Available from ProQuest. (UMI No. 3575346)

Dickinson

D. K.

Golinkoff

R. M.

Hirsh-Pasek

(2010). Speaking out for language: Why language is central to reading development. Educational Researcher, 39, 305–310.

Dickinson

D. K.

McCabe

Anastasopoulos

Peisner-Feinberg

E. S.

Poe

M. D.

(2003). The comprehensive language approach to early literacy: The interrelationships among vocabulary, phonological sensitivity, and print knowledge among preschool-aged children. Journal of Educational Psychology, 95, 465.

Dickinson

D. K.

Tabors

P. O.

(2001). Beginning literacy with language: Young children learning at home and school. Baltimore, MD: Brookes.

10.

Dunn

L. M.

Dunn

L. M.

(2007). Peabody picture vocabulary test-4. Circle Pines, MN: Pearson Assessments.

11.

Fernandez-Fein

Baker

(1997). Rhyme and alliteration sensitivity and relevant experiences among preschoolers from diverse backgrounds. Journal of Literacy Research, 29, 433–459.

12.

Gombert

J. E.

(1992). Metalinguistic development. Chicago, IL: University of Chicago Press.

13.

Garlock

V. M.

Walley

A. C.

Metsala

J. L.

(2001). Age-of-acquisition, word frequency, and neighborhood density effects on spoken word recognition by children and adults. Journal of Memory and Language, 45, 468–492.

14.

Harnishfeger

K. K.

(1995). The development of cognitive inhibition. In Dempster

F. N.

Brainerd

C. J.

(Eds.), Interference and inhibition in cognition (pp. 175–204). Amsterdam, the Netherlands: Elsevier.

15.

Hart

Risley

(1995). Meaningful differences in the everyday experiences of young American children. Baltimore, MD: Brookes.

16.

Huttenlocher

Haight

Bryk

Seltzer

Lyons

(1991). Early vocabulary growth: Relation to language input and gender. Cognitive Psychology, 45, 337–375.

17.

Juel

(1988). Learning to read and write: A longitudinal study of fifty-four children from first through fourth grade. Journal of Educational Psychology, 80, 437–447.

18.

Juel

(2006). The impact of early school experiences on initial reading. In Dickinson

D. K.

Neuman

S. B.

(Eds.), Handbook of early literacy research (Vol. II, pp. 410–426). New York, NY: Guilford.

19.

Liberman

I. Y.

Shankweiler

Fischer

F. W.

Carter

(1974). Explicit syllable and phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18, 201–212.

20.

Lonigan

C. J.

(2007). Vocabulary development and the development of phonological awareness skills in preschool children. In Wagner

R. K.

Muse

A. E.

Tannenbaum

K. R.

(Eds.), Vocabulary acquisition: Implications for reading comprehension (pp. 15–31). New York, NY: Guilford.

21.

Lonigan

C. J.

Allan

N. P.

Lerner

M. D.

(2011). Assessment of preschool early literacy skills: Linking children’s educational needs with empirically supported instructional activities. Psychology in the Schools, 48, 488–501.

22.

Lonigan

C. J.

Wagner

R. K.

Torgesen

J. K.

Rashotte

C. A.

(2007). TOPEL: Test of preschool early literacy. Austin, TX: PRO-ED.

23.

Mann

Liberman

(1984). Phonological awareness and verbal short-term memory. Journal of Learning Disabilities, 17, 592–599.

24.

Metsala

J. L.

(1999). Young children’s phonological awareness and nonword repetition as a function of vocabulary development. Journal of Educational Psychology, 91, 3–19.

25.

Metsala

J. L.

(2011). Lexical reorganization and the emergence of phonological awareness. In Neuman

S. B.

Dickinson

D. K.

(Eds.), Handbook of early literacy research (pp. 66–82). New York, NY: Guilford.

26.

Metsala

J. L.

Walley

A. C.

(1998). Spoken vocabulary growth and the segmental restructuring of lexical representations: Precursors to phonemic awareness and early reading ability. In Metsala

J. L.

Ehri

L. C.

(Eds.), Word recognition in beginning literacy (pp. 89–120). Mahwah, NJ: Lawrence Erlbaum.

27.

Muter

Hulme

Snowling

M. J.

Stevenson

(2004). Phonemes, rimes, vocabulary, and grammatical skills as foundations of early reading development: Evidence from a longitudinal study. Developmental Psychology, 40, 665–681.

28.

Muter

Hulme

Snowling

Taylor

(1997). Segmentation, not rhyming, predicts early progress in learning to read. Journal of Experimental Child Psychology, 65, 370–396.

29.

National Early Literacy Panel. (2008). Developing early literacy: Report of the National Early Literacy Panel. Washington, DC: National Institute for Literacy.

30.

NICHD-ECCRN. (2005). Pathways to reading: The role of oral language in the transition to reading. Developmental Psychology, 41, 428–442.

31.

Neuman

S. B.

(2006). The knowledge gap: Implications for early education. In Dickinson

D. K.

Neuman

S. B.

(Eds.), Handbook of early literacy research (Vol. II, pp. 29–40). New York, NY: Guilford.

32.

Paris

S. G.

(2005). Reinterpreting the development of reading skills. Reading Research Quarterly, 40, 184–202.

33.

Rowe

M. L.

(2012). A longitudinal investigation of the role of quantity and quality of child-directed speech in vocabulary development. Child Development, 83, 1762–1774.

34.

Rvachew

Ohberg

Grawburg

Heyding

(2003). Phonological awareness and phonemic perception in 4-year-old children with delayed expressive phonology skills. American Journal of Speech-Language Pathology, 12, 463–471.

35.

Schatschneider

Francis

D. J.

Foorman

B. R.

Fletcher

J. M.

Mehta

(1999). The dimensionality of phonological awareness: An application of item response theory. Journal of Educational Psychology, 91, 439.

36.

Schwarz

I. C.

Bowey

J. A.

Burnham

(2006). Phoneme sensitivity predicts vocabulary size in 2 1/2- to 3-year-olds. Paper presented at the 11th Australasian International Speech Science and Technology Conference SST, Auckland, New Zealand.

37.

Sénéchal

Ouellette

Rodney

(2006). The misunderstood giant: On the predictive role of early vocabulary to future reading. In Dickinson

D. K.

Neuman

S. B.

(Eds.), Handbook of early literacy research (Vol. II, pp. 173–182). New York, NY: Guilford.

38.

Smith

B. L.

McGregor

K. K.

Demille

(2006). Phonological development in lexically precocious 2-year-olds. Applied Psycholinguistics, 27, 355–375.

39.

Sosa

A. V.

Stoel-Gammon

(2012). Lexical and phonological effects in early word production. Journal of Speech, Language, and Hearing Research, 55, 596–608.

40.

Spira

E. G.

Bracken

S. S.

Fischel

J. E.

(2005). Predicting improvement after first grade reading difficulties: The effects of oral language, emergent literacy, and behavior skills. Developmental Psychology, 41, 225–234.

41.

Stahl

S. A.

Murray

B. A.

(1994). Defining phonological awareness and its relationship to early reading. Journal of Educational Psychology, 86, 221.

42.

Stanovich

K. E.

Cunningham

A. E.

Cramer

B. B.

(1984). Assessing phonological awareness in kindergarten children: Issues of task comparability. Journal of Experimental Child Psychology, 38, 175–190.

43.

Storch

S. A.

Whitehurst

G. J.

(2002). Oral language and code-related precursors to reading: Evidence from a longitudinal structural model. Developmental Psychology, 38, 934–947.

44.

Treiman

Zukowski

(1991). Levels of phonological awareness. In Brady

S. A.

Shankweiler

D. P.

(Eds.), Phonological processes in literacy: A tribute to Isabel Y. Liberman (pp. 67–83). Hillsdale, NJ: Lawrence Erlbaum.

45.

Wells

(1985). The meaning makers: Children learning language and using language to learn. Portsmouth, NH: Heinemann.

46.

Whiteley

H. E.

Smith

C. D.

Connors

(2007). Young children at risk of literacy difficulties: Factors predicting recovery from risk following phonologically based intervention. Journal of Research in Reading, 30, 249–269.

47.

Williams

K. T.

(2007). Expressive vocabulary test (2nd ed.). Circle Pines, MN: AGS.

48.

Williams

A. L.

Elbert

(2003). A prospective longitudinal study of phonological development in late talkers. Language, Speech, and Hearing Services in Schools, 34, 138–153.

49.

Yopp

H. K.

(1988). The validity and reliability of phonemic awareness tests. Reading Research Quarterly, 23, 159–177.

50.

Yopp

H. K.

(1995). A test for assessing phonemic awareness in young children. The Reading Teacher, 49, 20–29.