A Longitudinal Investigation of Code-Related Emergent Literacy Skills in Children Who Are Deaf and Hard of Hearing Across the Preschool Years

Abstract

Purpose:

The purpose of this longitudinal investigation was to compare the developmental trajectories of code-related emergent literacy skills of children who are deaf and hard of hearing (DHH) who use amplification and spoken language across the preschool years.

Method:

Thirty children who are DHH and 31 children with typical hearing completed a language and emergent literacy assessment at 6-month intervals from age 4 through 6 years. Growth curve analysis was used to compare developmental trajectories between groups of the code-related skills of phonological awareness, phonological memory, phonological recoding, alphabet knowledge, and conceptual print knowledge.

Results:

Growth across the preschool years was observed on all code-related emergent literacy skills across groups. Children who are DHH scored consistently lower than children with typical hearing on phonological awareness, phonological memory, and conceptual print knowledge; no group differences were observed for phonological recoding or alphabet knowledge. No interactions of time and group were significant.

Conclusions:

Children who are DHH exhibit consistent deficits in phonological awareness, phonological memory, and conceptual print knowledge across the preschool years and begin formal literacy instruction with a weaker foundation in emergent literacy skills. Future work should focus on optimizing emergent literacy interventions for children who are DHH during the preschool years.

Supplemental Material:

https://doi.org/10.23641/asha.21998153

One long-standing area of concern for children who are deaf and hard of hearing (DHH) who use spoken language ¹ is the low literacy outcomes experienced by the population. Although many advances have been made both in amplification technology and in the science of reading, children who are DHH continue to experience difficulties in literacy acquisition. Not only are reading deficits present for children who are DHH (Geers & Hayes, 2011), but there is also a deceleration of reading growth over time (Wei et al., 2011). Recent work suggests that this deceleration may begin in early elementary school (Antia et al., 2020). Therefore, it is critical to take steps to optimize literacy outcomes for children who are DHH.

The focus in this article is on foundational code-related emergent literacy skills. Even before the onset of formal literacy instruction, code-related skills such as phonological processing and print knowledge set the stage for later success for children with typical hearing (CTH; NELP, 2008). Children who are DHH have known deficits during preschool in both of these code-related skill areas (Werfel, 2017). Specifically, previous work has reported that children who are DHH perform lower than their peers on measures of phonological awareness, phonological memory, and conceptual print knowledge (Werfel, 2017). Longitudinal work in this area, however, has been rare. The purpose of this study was to compare growth in code-related emergent literacy skills of children who are DHH and CTH, as an initial step toward identifying areas in which early interventions have the potential to be maximally impactful.

Theoretical Framework

This study draws on prevailing theoretical perspectives of literacy acquisition in CTH. The Simple View of Reading (Gough & Tunmer, 1986; Juel et al., 1986), is a formula that defines successful reading comprehension, successfully deriving meaning from text, as the product of fluent, accurate decoding, and language comprehension (Protopapas et al., 2012). Therefore, an individual must be able to both decode or identify written words and apply their language knowledge to understand the written message. Successful reading is unlikely to occur when deficits occur in either decoding or language comprehension.

The Rope Model of Reading (Scarborough, 2001), as well as subsequent work from authors of the Simple View of Reading, converge to support this formula and provide a more detailed picture of how early code-related skills combine with language comprehension to support skilled reading. Scarborough's (2001) Rope Model illustrates how strands of a rope can be used to represent the ways language comprehension and word recognition skills interweave to support skilled reading and text comprehension. Conceptual print knowledge is one of several strands of the language comprehension section of the skilled reading rope. The word recognition section of the rope includes strands that represent the code-related skills of phonological awareness and alphabet knowledge (e.g., Bradley & Bryant, 1983; Carson et al., 2019; Compton, 2000; Stahl & Murray, 1994). When the strands of language comprehension and word recognition are woven tightly, early reading skills become increasingly strategic and increasingly automatic, leading to fluent word recognition and text comprehension.

Frameworks developed by authors of the Simple View of Reading further describe the language comprehension and word recognition skills needed to achieve reading comprehension. In both the Reading Acquisition Framework (Hoover & Gough, 2000) and the Cognitive Foundations Framework (Tunmer & Hoover, 2019), early word recognition skills including concepts about print, letter knowledge, alphabet knowledge, and phonemic awareness underpin alphabetic coding. As in the Rope Model, code-related components support increasingly automatic word recognition. Word recognition skills then combine with language comprehension to allow for decoding and comprehension of written text. A large body of work with CTH has supported the role of phonological processing and print knowledge across decoding and comprehension (Layes et al., 2022; Lervåg et al., 2018; for a review, see NELP, 2008).

Early Language and Literacy Acquisition in Children Who Are DHH

Children who are DHH have deficits in early code-related skills that are already present during the preschool years (e.g., Easterbrooks et al., 2008; Lund et al., 2015; Moeller et al., 2007; Werfel et al., 2015). Previous research suggests that children who are DHH exhibit deficits in each category of code-related skills: phonological processing and print knowledge (Werfel, 2017). The performance of children who are DHH across each area of early code-related skills is detailed below.

Phonological Processing

Phonological processing consists of phonological awareness, phonological memory, and phonological recoding, or rapid naming (Whitehurst & Lonigan, 1998). As with oral language, deficits in some areas of phonological processing are present in children who are DHH (Ambrose et al., 2012; Easterbrooks et al., 2008; Werfel, 2017). Specifically, children who are DHH have deficits in phonological awareness, and mixed results have been reported for phonological memory; phonological recoding skills, however, appear to be intact.

Phonological awareness, the ability to analyze and/or manipulate the sound structure of words, is one area of deficit within phonological processing for children who are DHH. Many research groups have reported that children who are DHH have deficits in their ability to perform sound analysis tasks compared to peers with typical hearing (Ambrose et al., 2012; Briscoe et al., 2001; Lund et al., 2015; Nittrouer et al., 2012; Spencer & Tomblin, 2009). Unsurprisingly, phonological awareness appears to develop more slowly for children who are DHH than for CTH (Easterbrooks et al., 2008; Kyle & Harris, 2011; Most et al., 2006; Sterne & Goswami, 2000; Werfel, 2017). Despite reports of slower development, it appears that children who are DHH follow the same general developmental pattern observed in CTH, such that the ability to analyze larger units of words (e.g., syllables) precedes smaller segments of words (e.g., phonemes; James et al., 2005).

Phonological memory is measured using tasks such as nonword repetition and digit span. Findings of phonological memory performance of children who are DHH have varied across task type. As with phonological awareness, on tasks of nonword repetition, many research groups have reported poorer performance by children who are DHH compared to their peers with typical hearing (Briscoe et al., 2001; Burkholder & Pisoni, 2003; Park & Lombardino, 2012). Briscoe et al. reported that deficits in nonword repetition were present in children who are DHH with and without speech sound production impairments. This lower performance on nonword repetition holds true even when children who are DHH have access to visual information of the speakers' face to lessen the burden on auditory access (Al-Salim et al., 2020). When phonological memory has been measured with digit span tasks, however, group differences between children who are DHH and CTH have not always been identified. For example, Briscoe et al. (2001) and Park and Lombardino (2012) both reported differences between groups on nonword repetition but not on digit span. Both studies included only school-age children with mild to moderate degrees of hearing loss. In studies that include children with more severe degrees of hearing loss, deficits in digit span appear across the age span (e.g., Geers & Hayes, 2011; Werfel, 2017).

Conversely, phonological recoding skills, most commonly measured using tasks of rapid naming, have widely been reported to be intact for children who are DHH. Phonological recoding is the ability to rapidly pull labels from long-term memory. Spencer and Tomblin (2009) reported that children with cochlear implants who ranged in age from 7 to 17 years did not differ from CTH on measures of rapid letter naming or rapid digit naming. Park and Lombardino (2012) reported similar findings for a group of children with mild to moderate hearing loss aged 7–12 years. Finally, Werfel (2017) reported that preschool children with moderate to profound hearing loss who used hearing aids or cochlear implants did not differ from peers with typical hearing on phonological recoding. Thus, across a range of ages and degrees of hearing loss, phonological recoding does not appear to be a deficit for this population.

Print Knowledge

Print knowledge consists of alphabet knowledge and conceptual print knowledge. Alphabet knowledge consists of knowledge of letter names and letter sounds, whereas conceptual print knowledge consists of knowledge of how print works and features of written words. To measure alphabet knowledge, one might, for example, ask a child to name a letter or provide a letter sound. To measure conceptual print knowledge, one might, for example, ask a child to indicate where on the page they would start to read or to explain why there is a space between words.

Print knowledge is generally considered an area of strength for children who are DHH (e.g., Ambrose et al., 2012; Easterbrooks et al., 2008); however, studies that have concluded that print knowledge is a strength have primarily measured alphabet knowledge. In some cases (e.g., Easterbrooks et al., 2008), only alphabet knowledge was assessed. In other cases (e.g., Ambrose et al., 2012; Tomblin et al., 2020), an omnibus measure of print knowledge, including both alphabet knowledge and conceptual print knowledge, was utilized. A recent item analysis of this standardized test, the Test of Preschool Early Literacy (Lonigan et al., 2007), however, provided evidence that the majority of items assessed alphabet knowledge and not conceptual print knowledge (Lund et al., 2020). Scarborough's (2001) Rope Model of Reading separates alphabet knowledge and conceptual print knowledge and posits that they contribute to different later reading skills. There is a need, therefore, to consider alphabet knowledge and conceptual print knowledge separately.

Children who are DHH consistently have alphabet knowledge skills that are similar to or higher than their peers with typical hearing (Ambrose et al., 2012; Cupples et al., 2014; Easterbrooks et al., 2008; Tomblin et al., 2020; Werfel et al., 2015). These findings span both letter name knowledge and letter sound knowledge. In the area of alphabet knowledge, it is clear that the traditional view of strength is supported.

For conceptual print knowledge, however, initial evidence points to deficits for children who are DHH. In separate samples of preschool children who are DHH, Werfel et al. (2015) and Werfel (2017) have reported that children who are DHH score lower on measures of conceptual print knowledge than CTH. There is a need for more research that evaluates print knowledge, including separately evaluating alphabet knowledge and conceptual print knowledge, in children who are DHH. This distinction is supported by the Rope Model of Reading and frameworks related to the Simple View of Reading, which identify alphabet knowledge and conceptual print knowledge as separate strands or components

This Study

Research has established that children who are DHH exhibit deficits in several early literacy skills; however, to date, investigations of preschool children who are DHH have not included comprehensive longitudinal investigations of the acquisition of the early code-related skills that are implicated in later literacy achievement for CTH. Understanding the development of these early skills and how they relate to later literacy achievement will provide foundational knowledge that can guide early identification of the children who are DHH and developing spoken language who are most at risk for later deficits. Therefore, the purpose of this study was to compare the developmental trajectories of early code-related skills of children who are DHH and CTH across the preschool years. Our hypothesis was that children who are DHH would have lower performance and slower growth than CTH across most code-related skills, with exceptions of alphabet knowledge and phonological recoding, which previous work suggested are areas of strength for children who are DHH.

Method

The University of South Carolina Institutional Review Board approved all study procedures.

Participants

Participants were the same 30 preschool children who are DHH (18 boys) and 31 preschool CTH (10 boys) as in Werfel et al. (2022). Recruitment for children who are DHH occurred via social media, preschool programs for children who are DHH, and speech-language pathology and audiology clinics. Recruitment for CTH occurred via social media, community preschools, and pediatrician offices. Recruitment occurred across the United States.

A priori inclusionary criteria ensured that no participants in either group had additional diagnoses known to affect language or cognition, and English was the language spoken in the home at least 70% of the time. No children had nonverbal intelligence below the average range, as measured by the Primary Test of Nonverbal Intelligence (Ehrler & McGhee, 2008). Chi-square analyses revealed that the DHH and CTH groups did not differ by distribution of ethnicity, race, or speech perception skill, measured by the Early Speech Perception Test (ESP; Moog et al., 2012). Most children in each group identified as White (> 80% in each group; p > .05) and not Hispanic or Latino (> 80% in each group; p > .05). Only one child in each group had speech perception skills lower than Consistent Word Identification, the highest score possible on the ESP. The groups differed on omnibus spoken language, measured by the Test of Early Language Development (Hresko et al., 1999, 2018) and speech sound production, measured by the Arizona Articulation Proficiency Scale (Fudala, 2000; Fudala & Stegall, 2017). Participant demographic information by group is displayed in Table 1.

Table 1.

Demographic information for participants by group.

Variable	DHH		CTH		t	p	d
	M (SD)	Range	M (SD)	Range
Age at study entry (months)	52.27 (4.41)	45–62	49.87 (4.34)	45–61	−2.138	.037	0.55
English use in home (percent)	94.44 (12.20)	70–100	97.43 (7.72)	70–100	1.117	.269	−0.29
Maternal education (years)	15.78 (2.72)	12–22	17.23 (2.20)	12–22	2.274	.027	−0.59
Nonverbal intelligence*	106.73 (16.68)	76–137	114.77 (11.42)	89–145	2.190	.033	−0.56
Omnibus spoken language*	89.60 (22.69)	41–125	115.42 (10.93)	88–133	5.632	< .001	−1.45
Speech production*	86.30 (11.33)	55–108	94.35 (9.21)	80–114	3.052	.003	−0.77

Note. DHH = children with who are deaf and hard of hearing; CTH = children with typical hearing.

^*Standard score.

All children who are DHH had been diagnosed with permanent bilateral hearing loss by a certified audiologist, were developing language primarily in spoken English, and used amplification. Sixteen children used at least one cochlear implant (12 bilateral, four bimodal), 12 children used bilateral hearing aids, and two children used bone-anchored hearing aids. Children's degree of hearing loss ranged from moderate to profound. Per parent report, all children who are DHH were receiving speech-language services. Additional audiologic information for the DHH group is displayed in Table 2. CTH passed a bilateral hearing screening at 20 dB HL.

Table 2.

Audiologic information of children who are deaf and hard of hearing.

Variable	M (SD)	Range	Mdn
Age at identification	7.67 (11.62)	0–36	1.5
Age at first hearing aid	11.97 (11.70)	1.5–36	5.00
Age at first implant (n = 16)	21.50 (11.91)	9–44	16.50

Note. All ages are reported in months.

Procedure

Beginning at approximately 4 years of age, participants completed language and early literacy assessment batteries at 6-month intervals through 6 years of age; testing occurred at approximately ages 4;0 [years;months], 4;6, 5;0, 5;6, and 6;0. Depending on behavior and attention, children completed the assessment battery in one or two assessment sessions at each time point. Total testing time for each time point was approximately 1.5–2 hr. Assessment sessions took place individually in a quiet room at the child's school, home, or a local library. Parents were permitted to be present if desired but were instructed to refrain from participating in the testing. Order of test administration was prerandomized for each participant.

Examiners included certified speech-language pathologists and speech-language pathology graduate students. The following training procedures were followed for each examiner: (a) read the administration chapter of the test manual for standardized measures, (b) receive 100% accuracy on a knowledge test on administration for each measure, (c) observe an assessment session by a trained lab member in person or via recording, (d) practice administering the assessments to a lab member, and (e) administer each measure to a certified speech-language pathologist with zero errors.

Measure

Of interest in this study were measures of phonological processing and print knowledge. Copies of experimental tasks that we developed are available in Supplemental Materials S1–S5.

Phonological Processing

Phonological awareness, phonological memory, and phonological recording were measured within the construct of phonological processing.

Phonological awareness. Phonological awareness was assessed using measures based on the Phonological Awareness Literacy Screening–Kindergarten (Invernizzi et al., 2004) Initial Sounds and Rhyme subtests. Alternate forms were created for use at each assessment time so that each child was not administered the same form more than once.

On the Initial Sounds subtest, the examiner showed children a picture of a target word along with pictures of three potential responses. The examiner said the target word and the three response choices out loud. The children were asked to select the picture of the word that began with the same sound as the target word (e.g., target: bee; nose, boat, dad). Children indicated their response by pointing to one of the three pictures. This subtest contained 10 test items, completed after three trial items.

On the Rhyme subtest, children were shown a picture of a target word along with pictures of three potential responses. The examiner said the target word and three response choices out loud. The children were asked to select the picture of the word that rhymed with the target word (e.g., target: pear; bear, pan, sun). Children indicated their response by pointing to one of the three pictures. This subtest contained 10 items, completed after three trial items.

In the analysis, the variables used from these measures consisted of number of correct items on each subtest (maximum of 10 each). Test–retest reliability of the published measure we based this task on is .94.

Phonological memory. To measure phonological memory, digit span and nonword repetition tasks were developed based on Gathercole and Adams (1993). Alternate forms were created for use at each assessment time.

On the digit span task, children were asked to repeat random strings of digits presented orally by the examiner. The child was instructed to repeat strings of digits of increasing length, beginning with combinations of two digits. Each digit span length had three items, and children advanced to the next digit span length when they correctly responded to two items at the previous length. When a child failed to accurately repeat two strings of digits of a certain length out of three attempts, testing was discontinued. The digit span score was the highest number of strings of digits the child was able to successfully repeat in two out of three attempts, following scoring procedures from Gathercole and Adams (1993).

On the nonword repetition task, children were asked to repeat 15 nonwords of various syllable lengths. Five words were one syllable, five were two syllables, and five were three syllables. The one- and two-syllable nonwords were selected from the Irvine Phonotactic Online Dictionary (IPhOD; Vaden et al., 2009), based on unstressed phonological neighborhood density (10–25), unstressed average biphoneme probability (.0001–.001), and word average positional probability (.02–.08). All syllables were either consonant–vowel–consonant or consonant–vowel. To create three syllable words, we added a common syllable onset or offset to a two-syllable nonword from the IPhOD nonword database. Our intention was that no nonwords contained fricatives or real words within the nonword (e.g., blowter includes blow and would not be included). ² Finally, nonwords were selected such that no more than two items within a syllable group on each list started with the same onset or contained the same rime. All 15 nonwords were administered to each child, regardless of performance. Words were presented in an auditory–visual format. The nonword repetition score was the total number of nonwords correctly repeated in full. Partial credit was not awarded.

Phonological recoding. Two rapid naming tasks based on Catts (1993) were used to assess phonological recoding. In the first measure, rapid animal naming, children were presented with a page containing color drawings of animals in six rows. Five animals were in each row: cow, dog, pig, bear, and mouse; each animal appeared once in each row in random order (30 animals total). Children were asked to name the animals on the page as fast as they could, going in order from left to right and top to bottom. The examiner used her finger to guide the child through the rows if the child had difficulty following the rows. Alternate forms were created for use at each assessment time. The second measure was rapid object naming. This measure was identical to the rapid animal naming measure except the pictures of the objects were top, bed, sun, cheese, and key. The rapid animal naming and rapid object naming scores were the number of seconds a child took to name all 30 animals or objects, respectively. Prior to completing this task, children were asked to name each of the five animals and objects. If they did not know one of the words, testing for that subtest was not completed. Additionally, an a priori decision was made that any attempt that included more than two errors was excluded from the analysis. The majority of missing data on these subtests were the result of more than two errors rather than a child not being able to name the animals in an untimed format.

Print Knowledge

Four print knowledge measures were administered: two measures of alphabet knowledge (letter name knowledge and letter sound knowledge), and two measures of conceptual print knowledge (concepts of print and concepts of written words).

Alphabet knowledge. Measures based on the Uppercase Letter Names and Lowercase Letter Names subtests of the Phonological Awareness Literacy Screening for Kindergarten (PALS-K; Invernizzi et al., 2004) were used to assess letter name knowledge. For each subtest, children were asked to provide the name of letters on a page with 26 uppercase or lowercase letters, respectively. Letters were presented in a predetermined, nonalphabetic order. Alternate forms were created for use at each assessment time. Test–retest reliability of the PALS-K is .92. The letter name score was the number of correctly named letters from each subtest in total (maximum of 52).

A measure based on the Letter Sounds subtest of the PALS-K was used to assess letter sound knowledge. After completing one practice item, children were presented with a page of 23 uppercase letters and three uppercase digraphs (CH, TH, and SH). The examiner instructed children to provide the sound made by each letter or letter pair. Alternate forms were created for use at each assessment time. For scoring, consistent with the PALS-K test manual, only short vowel sounds were considered correct for vowel letters; if children provided an appropriate long vowel sound for a vowel (e.g., /e/ for a), the examiner asked, “What other sound can this letter make?” Test–retest reliability of the PALS-K is .88. The letter sound score was the number of correct letter sounds provided (maximum of 26).

Conceptual print knowledge. To assess conceptual print knowledge, the Print Concepts subtest of the Preschool Print and Word Awareness Test (PWPA; Justice & Ezell, 2001) was used. The examiner and child participated in a shared book-reading activity with embedded questions about the print. For example, the examiner asked the child to show her where to start reading, asked about directionality of reading, and asked to explain where a character's dialogue was located within an illustration, among other questions. Items to assess knowledge of print concepts were spread throughout the book-reading activity. Interrater reliability is .99. The Print Concepts score was the number of points the child earned by correctly responding to the prompts during the book-reading activity, calculated according to published instructions (maximum of 18).

The Words in Print subtest of the Preschool Print and Word Awareness Test was used to assess conceptual knowledge of words in print. The examiner and child participated in a shared book-reading activity with embedded questions about written words. For example, the examiner asked the child to show her the big words versus little words, to identify the longest word on a given page, and to point out the space between two words, among other questions. Items to assess knowledge of written word concepts were spread throughout the book-reading activity. Interrater reliability is .99. The Words in Print score was the number of points the child earned by correctly responding to prompts during the book-reading activity, calculated according to published instructions (maximum of 12).

Reliability

Tests were double-scored by two different trained lab members. Disagreements, although rare, were discussed and attained 100% agreement. Study data were managed by REDCap electronic data capture tools (Harris et al., 2009) hosted at the University of South Carolina Arnold School of Public Health and Health Sciences South Carolina.

Data Preparation

To maximize the interpretability of analytic results, participants' scores on each measure were z-scored by full sample mean and standard deviation, equating 0 with the average score children in the study sample received for that measure. This also allowed a value of +1 to equal a score 1 SD above the sample average and −1 to equal a score 1 SD below the sample average.

Consistent with the broader theoretical constructs of interest, composite scores were constructed from the individual tasks administered to the children. Within the phonological processing domain, composite scores for phonological awareness, phonological memory, and phonological recoding were computed. For phonological awareness, children's z scores from the Rhyme and Initial Sounds subtests were averaged. For phonological memory, z scores from the digit span and nonword repetition tasks were averaged. For phonological recoding, z scores from the rapid animal naming and the rapid object naming were averaged. Within the print knowledge domain, composite scores for alphabet knowledge and conceptual print knowledge were constructed. For alphabet knowledge, z scores from the Letter Name Knowledge and Letter Sound Knowledge subtests were averaged. For conceptual print knowledge, z scores from the Print Concepts and Words in Print subtests were averaged.

In computing each composite, anytime a child had a score for one component but not the other, the composite score included only the measure that the child had completed. This approach was utilized to minimize the impact of any missing data on the results. Across the 249 Time × Child data points and the 10 subtests of interest (2,490 total data points), only 40 instances of this missing data pattern occurred. The majority of missing data (25 instances) resulted from invalid scores on the rapid naming subtests.

Analytic Plan

Children's scores on the five composite measures were examined descriptively and plotted by time and group to visualize change over time. Using the lme4 package (Bates et al., 2015) in the R environment (R Core Team, 2020), growth models were estimated for each composite score using a mixed-effect regression framework (McNeish & Matta, 2018). Across all models, time was centered at the first assessment time point, which is corresponding with approximately age 4 years for the participants. Each one-unit increase in time was equivalent to a single 6-month interval, which aligned with the testing windows. Group membership was included as a fixed effect, with children who are DHH being the reference group. All models were evaluated for evidence of nonlinear growth and interactions between group and time. Model fit and appropriateness was evaluated considering (a) evidence of misfit in the model variance parameters (e.g., negative variance); (b) model assumptions including homogeneity of variance, residual normality, and linearity; (c) model estimates, confidence intervals (CIs), R ² variance explained, and model comparisons where more parsimonious models were preferred; and (d) model alignment with the theoretical frameworks of this study.

Results

Descriptive information for the individual measures is provided by group and time point in Table 3 for phonological processing variables and Table 4 for print knowledge variables. When examined across all five time points, participants' composite scores for phonological awareness were slightly skewed with a somewhat platykurtic distribution (i.e., reduced central peak with wider tails). No evidence of outliers or difference in variance by group (i.e., heteroscedasticity) was observed. Similar trends were noted for the print knowledge measures of alphabet knowledge and conceptual print knowledge. For participants' composite phonological memory scores, a normal distribution with no evidence of skew, outliers, or heteroscedasticity by group was observed. For phonological recoding, the distribution exhibited a slight negative skew with some values outside 2 SDs of the sample mean, suggesting potential outliers. Re-examination of the values revealed that they likely reflected the true abilities of children in the population. As such, the values were retained.

Table 3.

Descriptive statistics of phonological processing variables by group across assessment points.

Time point	DHH mean (SD)	CTH mean (SD)	t	p	d
Phonological awareness: rhyme
Time 1	3.13 (1.69)	4.95 (3.17)	2.226	.033	−0.72
Time 2	6.20 (2.78)	6.81 (2.56)	0.830	.410	−0.23
Time 3	7.07 (2.88)	8.20 (2.34)	1.674	.099	−0.43
Time 4	7.37 (2.94)	9.00 (1.94)	2.392	.021	−0.65
Time 5	7.62 (3.19)	9.17 (1.64)	2.007	.054	−0.61
Phonological awareness: initial sounds
Time 1	3.07 (2.05)	4.19 (2.06)	1.614	.116	−0.55
Time 2	4.88 (2.28)	5.89 (2.62)	1.476	.146	−0.41
Time 3	5.30 (2.67)	6.70 (2.90)	1.945	.057	−0.50
Time 4	6.48 (2.78)	7.58 (2.53)	1.498	.140	−0.41
Time 5	7.33 (3.14)	9.17 (1.64)	2.406	.023	−0.73
Phonological memory: digit span
Time 1	2.60 (0.91)	4.95 (6.48)	1.391	.173	−0.51
Time 2	3.38 (1.17)	4.00 (0.98)	2.101	.041	−0.57
Time 3	3.60 (0.81)	4.06 (1.18)	1.783	.080	−0.45
Time 4	4.11 (1.17)	4.83 (1.17)	2.238	.030	−0.61
Time 5	4.43 (1.21)	4.87 (0.97)	1.342	.187	−0.40
Phonological memory: nonword repetition
Time 1	3.27 (2.15)	7.30 (2.27)	5.311	< .001	−1.82
Time 2	6.23 (3.17)	9.18 (2.91)	3.567	.001	−0.97
Time 3	6.40 (3.44)	9.60 (2.76)	3.973	< .001	−1.03
Time 4	7.07 (3.27)	10.42 (2.28)	4.339	< .001	−1.19
Time 5	6.86 (3.25)	12.36 (2.50)	6.253	< .001	−1.90
Phonological recoding: rapid animal naming
Time 1	53.64 (20.74)	66.63 (25.81)	1.572	.125	0.55
Time 2	52.40 (18.19)	51.48 (12.86)	−0.208	.836	0.06
Time 3	54.11 (21.68)	46.57 (14.73)	−1.521	.135	−0.41
Time 4	43.63 (17.11)	43.32 (16.79)	−0.066	.948	−0.02
Time 5	41.67 (16.18)	40.91 (14.75)	−0.162	.872	−0.05
Phonological recoding: rapid object naming
Time 1	54.67 (28.61)	54.58 (14.95)	−0.011	.991	0.00
Time 2	48.54 (16.39)	49.62 (15.05)	0.249	.804	0.07
Time 3	51.39 (18.42)	45.29 (13.69)	−1.408	.165	−0.38
Time 4	43.50 (16.07)	38.76 (10.06)	−1.268	.210	−0.35
Time 5	36.10 (9.09)	34.26 (8.93)	−0.675	.503	−0.20

Note. DHH = children who are deaf and hard of hearing; CTH = children with typical hearing.

Table 4.

Descriptive statistics of print knowledge variables by group across assessment points.

Time point	DHH mean (SD)	CTH mean (SD)	t	p	d
Alphabet knowledge: letter names
Time 1	22.73 (19.22)	27.52 (20.35)	0.712	.481	−0.24
Time 2	32.81 (17.03)	37.04 (15.70)	0.949	.347	−0.26
Time 3	41.03 (12.73)	40.90 (12.38)	−0.041	.967	0.01
Time 4	46.04 (7.55)	48.73 (3.85)	1.668	.103	−0.45
Time 5	48.95 (4.49)	50.09 (2.07)	1.093	.281	−0.33
Alphabet knowledge: letter sounds
Time 1	5.60 (8.24)	7.14 (7.64)	0.578	.567	−0.19
Time 2	8.85 (7.97)	11.54 (8.39)	1.205	.234	−0.33
Time 3	13.83 (7.67)	14.37 (8.03)	0.263	.793	−0.07
Time 4	17.36 (6.31)	19.54 (5.87)	1.312	.195	−0.36
Time 5	22.00 (4.18)	23.09 (3.70)	0.914	.366	−0.28
Conceptual print knowledge: print concepts
Time 1	4.20 (3.10)	5.95 (2.40)	1.914	.064	−0.63
Time 2	6.81 (3.37)	11.67 (3.26)	5.336	< .001	−1.47
Time 3	8.47 (4.27)	12.32 (3.39)	3.911	< .001	−1.00
Time 4	11.36 (3.82)	14.58 (2.76)	3.525	.001	−0.97
Time 5	12.33 (4.61)	15.70 (1.94)	3.102	.005	−0.95
Conceptual print knowledge: words in print
Time 1	1.47 (1.96)	3.81 (2.18)	3.311	.002	−1.13
Time 2	3.35 (2.42)	5.70 (3.10)	3.081	.003	−0.85
Time 3	4.20 (2.94)	6.39 (3.16)	2.797	.007	−0.72
Time 4	6.68 (3.68)	8.04 (2.86)	1.507	.138	−0.41
Time 5	7.43 (3.50)	10.30 (2.06)	3.283	.003	−1.00

Note. DHH = children who are deaf and hard of hearing; CTH = children with typical hearing.

Change Over Time

The growth model for phonological memory was computed as planned in the lme4 package. For the four composite scores that did not meet baseline criteria of a normal distribution (i.e., phonological awareness, phonological recoding, alphabet knowledge, conceptual print knowledge), models were estimated twice. First, analyses were conducted using lme4 (Bates et al., 2015). Models were then re-estimated using the robustlmm R package (Koller, 2016) for robust estimation. Model parameters and assumptions were examined and compared for evidence of substantial misfit and robustness of the obtained estimates. The robustlmm package was designed as a direct analogue for the lme4 package, and therefore, results are reported similarly.

Results from the phonological processing growth models are provided in Table 5 and depicted in Figure 1. Overall, regardless of hearing status, children tended to exhibit similar rates and forms of growth in phonological awareness, phonological memory, and phonological recoding, as evidenced by the lack of interaction effects. However, there was a significant gap in the scores achieved by children who are DHH compared to the CTH both for phonological awareness and for phonological memory. On average, children who are DHH scored 0.46 SD lower (95% CI [0.11, 0.82], p = .011) than CTH on phonological awareness and 0.77 SD lower (95% CI [0.47–1.08], p < .001) on phonological memory. These gaps did not change across the duration of the study, although, generally, growth in phonological awareness slowed slightly across the 2 years of the study (−0.06, 95% CI [−0.09, −0.02], p = .001). Conversely, for phonological recoding, there were no significant differences in scores or growth between children who are DHH and those with typical hearing (0.09, 95% CI [−0.46, 0.28], p = .635).

Table 5.

Phonological processing growth models.

	Phonological awareness (robust)			Phonological memory			Phonological recoding (robust)
Predictors	Est.	CI (95%)	p value	Est.	CI (95%)	p value	Est.	CI (95%)	p value
Intercept	−1.11	[−1.40, −0.83]	< .001	−0.99	[−1.22, −0.75]	< .001	0.61	[0.33, 0.89]	< .001
Time (centered at 1)	0.60	[0.45, 0.74]	< .001	0.29	[0.25, 0.33]	< .001	−0.31	[−0.35, −0.26]	< .001
Time: Quadratic	−0.06	[−0.09, −0.02]	.001
Group (CTH)	0.46	[0.11, 0.82]	.011	0.77	[0.47, 1.08]	< .001	−0.09	[−0.46, 0.28]	.635
Random effects
σ2	0.17			0.17			0.17
τ00	0.43 ChildID			0.33 ChildID			0.48 ChildID
ICC	0.72			0.65			0.74
N	61 ChildID			61 ChildID			61 ChildID
Observations	246			248			243
Marginal R 2/conditional R 2	0.323/0.808			0.367/0.780			0.200/0.790

Note. The intercept may be interpreted as the number of standard deviations below the sample mean a child with hearing loss would be predicted to score on phonological awareness at Time 1 (approximately 4 years of age). Bolded values indicate statistical significance. Est. = estimate; CI = confidence interval; CTH = children with typical hearing; ICC = intraclass correlation coefficient.

Figure 1.

Comparison of growth in phonological processing skills across groups. CHL = children who are deaf and hard of hearing; CNH = children with typical hearing.

Results from the print knowledge growth models are provided in Table 6 and depicted in Figure 2. Again, regardless of hearing status, children exhibited similar overall rates and forms of growth both in alphabet knowledge and conceptual print knowledge. No interaction effects were identified. For alphabet knowledge, there was not a significant difference between the scores of children who are DHH compared to CTH (0.18, 95% CI [−0.19, 0.56], p = .338), but overall growth in alphabet knowledge slowed over the 2 years of the study (−0.04, 95% CI [−0.0, −0.01], p = .021). For conceptual print knowledge, there was a significant difference between the groups' scores, with CTH scoring approximately 0.62 SD (95% CI [0.31, 0.93], p < .001) higher than children who are DHH at the first time point. This gap did not appear to change over the course of the study.

Table 6.

Print knowledge growth models.

Predictors	Alphabet knowledge (robust)			Conceptual print knowledge (robust)
	Est.	CI (95%)	p value	Est.	CI (95%)	p value
Intercept	−0.99	[−1.29, −0.69]	< .001	−1.23	[−1.47, −1.00]	< .001
Time (centered at Time 1)	0.57	[0.42, −0.73]	< .001	0.45	[0.40, 0.49]	< .001
Group (CTH)	0.18	[−0.19, 0.56]	.338	0.62	[0.31, 0.93]	< .001
Time: quadratic	−0.04	[−0.08, −0.01]	.021
Random effects
σ2	0.20			0.18
τ00	0.49 ChildID			0.31 ChildID
ICC	0.71			0.64
N	61 ChildID			61 ChildID
Observations	249			248
Marginal R 2/conditional R 2	0.297/0.798			0.469/0.807

Note. The intercept may be interpreted as the number of standard deviations below the sample mean a child with hearing loss would be predicted to score on phonological awareness at Time 1 (approximately 4 years of age). Est. = estimate; CI = confidence interval; CTH = children with typical hearing; ICC = intraclass correlation coefficient.

Figure 2.

Comparison of growth in print knowledge skills across groups. CHL = children who are deaf and hard of hearing; CNH = children with typical hearing.

Sensitivity Analyses

Some group differences were observed between DHH and CTH on key background variables that are commonly associated with literacy development. In particular, maternal education was, on average, lower among DHH compared to CTH: t(59) = 2.27, p = .027, d = −0.59. Children's scores for nonverbal intelligence, as measured by the Primary Test of Nonverbal Intelligence (Ehrler & McGhee, 2008), was also, on average, lower among DHH compared to CTH: t(59) = 2.19, p = .033, d = −0.56. Although these values should be interpreted with caution, particularly given that there is limited evidence that nonverbal intelligence assessments function reliably for DHH, we elected to conduct post hoc sensitivity analyses to assess the robustness of the findings. Detailed results are reported in the Supplemental Materials S6 and S7. No substantial differences were observed in overall model results with one exception. The group difference in phonological awareness performance between DHH and CTH was not statistically significant at p < .05 after accounting for nonverbal IQ and maternal education. However, given the low statistical power to detect child-level effects in the small-to-moderate range, no conclusive inferences can be drawn regarding this analysis. The CI [−0.07, 0.65] for the group estimate is not consistent with a null difference between the groups. Replication with a larger sample is needed.

As a further test of robustness, we conducted one additional post hoc sensitivity analyses. In this second sensitivity analysis, we examined age of amplification as a covariate. Notably, age of amplification for CTH was entered as “0” to reflect typical hearing from birth, whereas age of amplification for each DHH varied depending on the timeline on which they received amplification. Results are provided in Supplemental Materials S8 and S9. Age of amplification emerged as a potentially more accurate predictor of differences in phonological awareness and conceptual print knowledge compared to the overall group comparison, though this was not the case for phonological memory.

Discussion

This study compared the developmental trajectories of code-related skills for children who are DHH and CTH across the preschool years. Our primary motivation was to identify preschool code-related skills known to impact later literacy acquisition for CTH for which children who are DHH had lower performance and/or slower growth. Our a priori hypothesis was that children who are DHH would have lower performance and slower growth across most code-related skills, except for alphabet knowledge and phonological recoding.

We present several findings of interest. First, differences between groups in level of skill were observed for phonological awareness skills, composed of a measure of rhyme and a measure of initial sounds, as well as for phonological memory skills, composed of a measure of digit span and a measure of nonword repetition; however, contrary to our hypothesis, rate of growth for these skills over time did not differ between groups. Second, group differences were also observed for level of skill in conceptual print knowledge; again, the rate of growth did not differ between groups. Third, as expected, there was no difference in performance between groups in level of skill or rate of growth for phonological recoding or alphabet knowledge. Therefore, our hypothesis of lower performance was confirmed, whereas our hypothesis of slower growth was not.

Phonological Processing Trajectories Across the Preschool Years

Within phonological processing, we report different patterns of performance across component skills. Phonological awareness and phonological memory scores were lower for children who are DHH across all time points during the preschool years; however, the rates of growth did not differ from those of the CTH. Therefore, children who are DHH begin the preschool period already behind their peers with typical hearing in phonological awareness and phonological memory, and although they make gains across ages 4–6 years, these gains are not sufficient to close the gap in performance by the time of school entry. Conversely, and similar to previous findings (Park & Lombardino, 2012; Spencer & Tomblin, 2009; Werfel, 2017), there was no difference in performance for phonological recoding across groups, suggesting that this skill is an area of strength for children who are DHH.

Phonological awareness. Our phonological awareness composite consisted of measures of rhyme and initial sound matching. Children who are DHH performed lower than CTH on these measures throughout the preschool years. This finding is consistent with previous work that has reported deficits in phonological awareness skills for children who are DHH, even those who use amplification and spoken language (Ambrose et al., 2012; Briscoe et al., 2001; Lund et al., 2015; Nittrouer et al., 2012; Spencer & Tomblin, 2009; Werfel, 2017). The growth of these skills and gap in performance between the two groups was almost half a standard deviation at each time point and remained consistent throughout the study.

A novel finding of this study is that the magnitude of difference between groups on measures of phonological awareness is consistent across the preschool years, when measured at 6-month intervals. This consistent gap supports the conclusion of James et al. (2005) that children who are DHH develop phonological awareness with the same trajectory as peers with typical development but not on the same time scale. Despite not falling further behind across the preschool years in phonological awareness, the children who are DHH in this study did not make gains to close the gap with their peers with typical hearing. Evidence of this persistent delay supports arguments by Wang et al. (2008) and Paul et al. (2009) that the foundation of phonological awareness skills in children who are DHH may be insufficient to support successful proficient literacy development. Because phonological awareness is a critical foundational skill to decoding (e.g., Adams, 1990), it is vital that future research explore ways to accelerate phonological awareness acquisition during the preschool years for children who are DHH.

Phonological memory. Consistent with much previous research, phonological memory was an area of weakness for children who are DHH who use amplification and spoken language, demonstrating impaired performance compared to peers with typical hearing (Briscoe et al., 2001; Burkholder & Pisoni, 2003; Park & Lombardino, 2012). Similar to phonological awareness skill development, children who are DHH did not catch up to the typical hearing group. The gap in performance remained present across all time points. This is consistent with previous findings of impaired performance on digit span (Geers & Hayes, 2011; Werfel, 2017) and nonword repetition (Briscoe et al., 2001; Burkholder & Pisoni, 2003; Park & Lombardino, 2012) tasks for children who are DHH.

Phonological recoding. Differences did not emerge between children who are DHH and CTH on phonological recoding tasks. Children who are DHH demonstrated similar growth pattern on this skill to their peers with typical hearing. This finding corroborates previous findings that children who are DHH do not show impairments in rapid automatic naming performance or development (Park & Lombardino, 2012; Spencer & Tomblin, 2009; Werfel, 2017).

Print Knowledge Trajectories Across the Preschool Years

Alphabet knowledge. Alphabet knowledge did not emerge as area of weakness for children who are DHH. Children who are DHH developed alphabet knowledge skills at a rate consistent with their peers with typical hearing. This finding is consistent with previous research (Ambrose et al., 2012; Cupples et al., 2014; Easterbrooks et al., 2008; Tomblin et al., 2020; Werfel et al., 2015). We conclude, as many others have, that alphabet knowledge is an area of strength for preschool children with hearing loss.

Conceptual print knowledge. In this study, we were particularly interested in evaluating conceptual print knowledge separate from alphabet knowledge. Deficits in performance on conceptual print knowledge were found for children who are DHH compared with CTH, although not rate of growth. This finding contradicts some previous research findings (Ambrose et al., 2012; Easterbrooks et al., 2008; Tomblin et al., 2020), but supports preliminary evidence of weaknesses in this skill area of children who are DHH reported in the works of Werfel et al. (2015) and Werfel (2017).

The finding of deficits for children who are DHH in conceptual print knowledge but not alphabet knowledge highlights the importance of the conclusions of the item analysis conducted by Lund et al. (2020). Lund et al. (2020) reported that many common measures of preschool print knowledge do not sufficiently separate items that measure conceptual print knowledge from those that measure alphabet knowledge to obtain a valid assessment of a child's conceptual print knowledge. Additionally, there appears to be an interaction of vocabulary knowledge and performance on particular types of conceptual print knowledge tasks, such that children without sufficient understanding of conceptual words like first, before, or longest are unlikely to be successful at attaining conceptual print knowledge skills commensurate with their peers (Lund et al., 2020). Werfel et al. (2022) reported that children who are DHH have lower vocabulary knowledge across the preschool years. It is vital, therefore, that professionals who work with children who are DHH select measures carefully to ensure that separate scores of alphabet knowledge and conceptual print knowledge can be obtained, and that intervention approaches consider targeting conceptual vocabulary as well as conceptual print knowledge skills.

This finding also provides valuable information regarding assessment for clinical practice. Historically, children who are DHH have performed similar to their peers on measures of print knowledge when that skill was assessed using a standardized measure (Ambrose et al., 2012; Easterbrooks et al., 2008). In this study, a nonstandardized measure, the PWPA, was used to assess conceptual print knowledge. This measure scores each item depending on the child's response while engaged in a book-reading activity. Our finding that children who are DHH exhibited a deficit in conceptual print knowledge compared to their peers underscores the importance of utilizing evaluative methods that assess the skill in question in ecologically valid ways. Omnibus measures, while timesaving, may insufficiently distinguish strengths from weaknesses, and test items may not adequately reflect how children use knowledge in their day-to-day lives. Our finding also highlights the potential significance of utilizing naturalistic assessment methods. Identifying print concepts during a book-reading session with a physical book more closely approximates the skills required to have sufficiently developed print concept knowledge than answering multiple-choice questions in a standardized assessment.

Clinical Implications

The low literacy achievement of children who are DHH has long been identified as an area of need. The findings of this study indicate areas of early literacy skills in children who are DHH that could be targeted to bridge the gap in reading attainment at an early age, including phonological awareness and conceptual print knowledge. Importantly, this study identified the area of conceptual print knowledge as an area of weakness for children who are DHH. Code-related emergent literacy skills are related to reading outcomes in elementary school for CTH (NELP, 2008; Senechal & LeFevre, 2002). Therefore, children who are DHH likely would benefit from therapy targeting phonological awareness and conceptual print knowledge to improve future reading performance.

Limitations

We present the following limitations. First, the groups of children who are DHH and CTH were not perfectly matched. The groups in this study differed on demographic variables of nonverbal intelligence, maternal education, and age at enrollment. Therefore, caution is needed in interpreting the specific magnitudes of differences observed between the participant groups. Demographic influences on study findings among children who are DHH should continue to be explored. Additionally, the representativeness of race and ethnicity in this sample may not match all communities, which should be kept in mind when interpreting findings. Although all children who are DHH were enrolled in speech and/or language services at the time of study enrollment, we did not collect information on the specific nature of these services, or on the training and expertise of the service providers.

Another consideration is that this article included a relatively small sample of participants and focused specifically on examining differences between two groups of children over 2 years. Although the findings provide an important foundation for future investigations detailing how code-related literacy skills develop among children with hearing loss, the current work had limited statistical power. This specifically limited our ability to assess child-level predictors that could contribute to clearer understanding of skill development among children who are DHH. At the child level, we were powered at 0.80 to detect moderate-to-large effect sizes given the current sample size. This limited the capacity to evaluate potential child-level covariates and interactions between group and growth. We recommend that future work consider how these and other additional potentially important factors, such as socioeconomic status and age of amplification, contribute to the development of children's code-related literacy skills.

Conclusions

Several key findings emerged from this study. First, deficits are apparent in preschool for phonological awareness, phonological memory, and conceptual print knowledge for children who are DHH who use amplification and spoken language. Second, phonological recoding and alphabet knowledge did not differ for this population and were comparable to skills of CTH. Many research groups have previously reported print knowledge as a whole as an area of strength for children who are DHH. Our findings do not support that conclusion. Instead, our study shows that when considered separately, distinct strengths in alphabet knowledge and weaknesses in conceptual print knowledge are present. This weakness would not emerge as an area of need if it was only assessed in tandem with alphabet knowledge, an area of relative strength. Early intervention that focuses on phonological and conceptual print knowledge skills is needed for children who are DHH.

Data Availability Statement

The data sets generated during and/or analyzed during this study are available from the corresponding author on reasonable request.

Funding Statement

This work was supported in whole by the National Institutes of Health (R03DC014535 to K.L.W.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.