View in EDS HTML Full Text PDF Full Text

Improving the Measures of Phonological Ability in the Russian Language: IRT and CART Modeling Application

Saved in:
Bibliographic Details
Title: Improving the Measures of Phonological Ability in the Russian Language: IRT and CART Modeling Application
Language: English
Authors: Ilia V. Markov, Ksenia S. Kharitonova, Elena L. Grigorenko
Source: Reading Research Quarterly. 2025 60(1).
Availability: Wiley. Available from: John Wiley & Sons, Inc. 111 River Street, Hoboken, NJ 07030. Tel: 800-835-6770; e-mail: cs-journals@wiley.com; Web site: https://www.wiley.com/en-us
Peer Reviewed: Y
Page Count: 18
Publication Date: 2025
Sponsoring Agency: National Institutes of Health (NIH) (DHHS)
Contract Number: R01DC007665
Document Type: Journal Articles
Reports - Research
Descriptors: Foreign Countries, Phonological Awareness, Phonology, Language Acquisition, Literacy, Short Term Memory, Item Response Theory, Predictor Variables, Test Construction, Drills (Practice), Auditory Tests
Geographic Terms: Russia
DOI: 10.1002/rrq.604
ISSN: 0034-0553
1936-2722
Abstract: Phonological awareness and phonological working memory are essential for successful language acquisition and development of literacy. Although this essence is language-universal, its degree varies for different languages, depending, in part, on language transparency. The current study analyzes the adapted versions of the pseudoword repetition test (assessing phonological working memory) and Rosner's Auditory Segmentation test (assessing phonological awareness) in a typically developing Russian native sample of children (n = 502). As a preparatory step to item analysis, we investigated the effects of grade and gender on performance using a mixed effects model. The initial item analysis was carried out using model comparison within the Item Response Theory model framework and threshold/slope analysis. The majority of the items in both assessments did not differentiate between students with different levels of phonological ability. Further item selection using regression tree models led to the formation of predictive and non-predictive item subsets for each assessment. After comparing the item subsets on various linguistic metrics, the differences were found in number of syllables and subsyllabic complexity for the pseudoword repetition test and elision segment position for the auditory segmentation test. The findings inform test development strategies in the cases of extremely low difficulty/discrimination of the items and outline a blueprint of pseudoword repetition and auditory segmentation test's adaptation for potentially detecting higher levels of phonological ability in transparent languages such as Russian.
Abstractor: As Provided
Entry Date: 2025
Accession Number: EJ1458551
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwE4aTLybPu44hB-R7SZFTNXAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDFXlERAeU0wUfB3jqgIBEICBm_75H_k-eVzlEOs-gvLx9y7A6aZZ11nTtiMrtcwpArAO2mreJdjzaoX4cUKRhZ95ZJMGl6kxtucZPfgIcapFcBz5SIzzvquYIhOmV1bKSdKh0-s-XwxClRVh4TifZtSoexvNzOUfRybicKGdF42T_6TVKZpRR-qpRjINaxwN4FY_Pzuif6pleVt0mfBLyaoG1SccGxPsS5iKsEI0
Text:
  Availability: 1
  Value: <anid>AN0183756703;[nrnu]01jan.25;2025Apr04.08:23;v2.2.500</anid> <title id="AN0183756703-1">Improving the Measures of Phonological Ability in the Russian Language: IRT and CART Modeling Application </title> <p>Phonological awareness and phonological working memory are essential for successful language acquisition and development of literacy. Although this essence is language‐universal, its degree varies for different languages, depending, in part, on language transparency. The current study analyzes the adapted versions of the pseudoword repetition test (assessing phonological working memory) and Rosner's Auditory Segmentation test (assessing phonological awareness) in a typically developing Russian native sample of children (n = 502). As a preparatory step to item analysis, we investigated the effects of grade and gender on performance using a mixed effects model. The initial item analysis was carried out using model comparison within the Item Response Theory model framework and threshold/slope analysis. The majority of the items in both assessments did not differentiate between students with different levels of phonological ability. Further item selection using regression tree models led to the formation of predictive and non‐predictive item subsets for each assessment. After comparing the item subsets on various linguistic metrics, the differences were found in number of syllables and subsyllabic complexity for the pseudoword repetition test and elision segment position for the auditory segmentation test. The findings inform test development strategies in the cases of extremely low difficulty/discrimination of the items and outline a blueprint of pseudoword repetition and auditory segmentation test's adaptation for potentially detecting higher levels of phonological ability in transparent languages such as Russian.</p> <p>We examined the effectiveness of phonological awareness (Rosner's Auditory Segmentation) and working memory (pseudoword repetition) tests in the Russian language for students with different levels of phonological ability. The findings suggest that the overall item difficulty is too low to fully capture the range of phonological ability in native Russian‐speaking children. Based on item analysis, we argue that increasing syllable count and subsyllabic complexity in pseudoword items, as well as using auditory segmentation items with middle or final elision will be most effective for detecting higher levels of ability.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-toc-0001.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-toc-0001.jpg" title="." /> </p> <p></p> <hd id="AN0183756703-3">Introduction</hd> <p>Phonology is a crucial area of study within the Russian language curriculum for students from the first to the eleventh grade. Various phonetic complexities of the Russian language, such as vowel reduction and the distinction between voiced and voiceless syllables, often contribute to confusion among learners. Notably, the Unified State Exam, a standardized exam undertaken upon high‐school completion, includes questions that target the knowledge of phonetic concepts and understanding of peculiarities of Russian phonetics. The language curriculum in Russian schools places an emphasis on the development of phonological awareness (PA), which is recognized as fundamental to the successful development of fluent reading and writing skills (Blachman, [<reflink idref="bib7" id="ref1">7</reflink>]; Rakhlin et al., [<reflink idref="bib64" id="ref2">64</reflink>]). Alongside PA, a domain‐specific temporary storage system known as the phonological working memory (PWM) represents a necessary resource for the successful development of reading ability. Research has shown that PWM not only correlates with PA but also predicts unique variance in reading skills (De Jong & Van Der Leij, [<reflink idref="bib20" id="ref3">20</reflink>]; Hansen & Bowey, [<reflink idref="bib37" id="ref4">37</reflink>]; Mann & Liberman, [<reflink idref="bib52" id="ref5">52</reflink>]).</p> <p>This study aims to conduct an item‐level analysis of the adapted Russian versions of Rosner's Test of Auditory Segmentation (RAS) and pseudoword repetition (PSW) assessments as measures of PA and PWM, respectively. It will also highlight the most difficult items and investigate linguistic elements that contribute to their level of difficulty, thus formulating a potential direction for the development of assessments for various aspects of phonological processing in the Russian language.</p> <p>Russian literacy instruction, which is based on the analytic‐synthetic approach, includes major components such as parsing spoken words into syllables and sounds by carrying out phonetic segmentation of words and developing PA, learning letter‐sound correspondences, and learning how to blend letter sounds into syllables and words (Kornev et al., [<reflink idref="bib45" id="ref6">45</reflink>]). These elements are vital for not only mastering the mechanics of reading but also for enhancing overall reading fluency and comprehension.</p> <p>While most Russian students learn to read during the first 2 years of formal schooling, this skill does not diminish the importance of continuing to focus on PA and PWM. In fact, reading fluency remains one of the main sources of individual differences through the entire formal schooling stage (Grigorenko & Elliott, [<reflink idref="bib36" id="ref7">36</reflink>]). Dorofeeva et al. ([<reflink idref="bib25" id="ref8">25</reflink>]) explored the role of phonological processing in reading among Russian‐speaking children aged 7–11, using seven phonological tests of varying complexity (phoneme discrimination, lexical decision, phoneme detection, pseudoword repetition, phoneme isolation, number of sounds, and changing sound) to assess their relationship with reading fluency and comprehension. Results indicated that more complex phonological tasks had lower accuracy rates and that the difficulty of these tasks was linked to decreased reading fluency, emphasizing that more complex phonological tests have a greater predictive value for reading fluency.</p> <hd id="AN0183756703-4">Features of Russian Phonetics</hd> <p>Among the 33 letters of the Russian alphabet, there are 10 vowels, 21 consonants, and 2 auxiliary letters. Four (е, ё, ю, я) of the ten (а, е, ё, и, о, у, ы, э, ю, я) vowels in Russian are formed by jotation (я = j + a, е = j + e, ё = j + о, ю = j + u); if one of these vowels follows a consonant that has a palatalized form, the consonant will be palatalized (with a few exceptions in borrowed words: <фонетика> ("fonetika"—phonetics)—the sound n would need to be palatalized because it is followed by a jotated vowel). Unlike English, Russian does not differentiate between short and long vowel sounds. Linguistic stress (or accent) plays a critical role in the pronunciation of vowels because unstressed vowels undergo a process called "vowel reduction." For example, two first vowels in the word <молоко> ("moloko"—milk) are unstressed, which leads to them being pronounced more like <a > rather than <o>. This feature of Russian vowels presents a significant challenge for young writers as they need to learn a set of rules related to unstressed vowels and strategies to spell them correctly, for example, finding a cognate in which the same vowel is in a stressed position (<молóчный> ("molóchniy"—of milk, adj.), but in other cases only memorizing such words works).</p> <p>All Russian consonants have the property of being soft or hard, with most of the consonants possessing both forms. However, some of the consonants (ж, ш, ц) are always hard, and some (й, ч, щ) are always soft. Confusion arises because there are two auxiliary letters in the Russian alphabet that only denote hardness (ъ) or softness (ь) of the preceding consonant, but in most cases, they do not need to be used as the same function is fulfilled by iotated vowels. Additionally, soft consonants are not always paired with iotated vowels, and hard consonants do not always go with hard‐series vowels.</p> <p>Consonants are also divided into voiced and voiceless based on the vocal cord vibration. This feature is heavily context‐dependent because sounds in consonant clusters can become voiced or voiceless because of the influence of adjacent consonants. Consonant clusters represent a separate set of issues associated with their spelling and reading: besides changing a form, some consonants can become completely silent in a cluster (л in the consonant cluster <лнц> of <солнце> ("solntse"—Sun) is silent). Importantly, changes in the sound quality due to palatalization, positional (de)voicing, and stress fluctuations are not reflected in orthography. Because many words in Russian either have complex codas (e.g., <торт> ("tort"—cake) or complex onsets (e.g., <труд> ("trud"—labor))), the mapping of graphemes to phonemes is easy, and the mapping of phonemes to graphemes is difficult (Zhukova & Grigorenko, [<reflink idref="bib83" id="ref9">83</reflink>]).</p> <p>Most orthographies can be placed on a depth continuum based on how transparent (shallow) or opaque (deep) they are. English orthography belongs to the side of the spectrum that hosts deep orthographies—in most cases, predicting the pronunciation of an English word based on its spelling is impossible. Although Russian orthography is complicated by the issues described above, it is sufficiently less deep than that of English.</p> <p>Another point of comparison between languages is syllable complexity, which is partly connected to the orthographic difficulty of a language. Germanic languages tend to have more closed syllables and consonant clusters. For example, the most common types of monosyllables in English are CVC (cat; 43%), CVCC (cold; 21%), and CCVC (drum; 15%); CV type (do) is only 3.5% (Goswami, [<reflink idref="bib34" id="ref10">34</reflink>]). The syllabic boundaries in Russian are inherently unstable. Many high‐frequency words in Russian are multisyllabic with a variety of syllable types. However, open syllables constitute 78% of all syllable types in Russian, with over half of them being of a CV (consonant‐vowel) type (Bondarko, [<reflink idref="bib10" id="ref11">10</reflink>] and Bogomazov, [<reflink idref="bib8" id="ref12">8</reflink>], as cited in Kerek & Niemi, [<reflink idref="bib41" id="ref13">41</reflink>]). The syllabic pattern of Russian permits closed syllables and consonant clusters in both the onset and coda positions. The internal subsyllabic structure of the CVC syllable is body‐coda (Kogan & Saiegh‐Haddad, [<reflink idref="bib44" id="ref14">44</reflink>]). In Russian, a consonant cluster can comprise a maximum of five consonants, as seen in words like <бодрствование> ("bodrstvovanie,"—wakefulness); each written letter in these clusters typically represents an individual sound. The degree of syllable complexity, along with the orthography depth, may play a significant role in language acquisition and especially in learning to read (Grigorenko & Elliott, [<reflink idref="bib36" id="ref15">36</reflink>]), with the degree of orthographic transparency influencing the rate of learning to read in both normal and dyslexic readers (Caravolas, [<reflink idref="bib16" id="ref16">16</reflink>]). However, the findings on the role of PA in explaining individual and group differences in literacy development in languages with relatively transparent orthographies are mixed (Caravolas, [<reflink idref="bib16" id="ref17">16</reflink>]).</p> <hd id="AN0183756703-5">Phonological Working Memory and Phonological Awareness</hd> <p>Phonological working memory (PWM) is an essential component of phonological skills. It involves a set of cognitive processes that control the short‐term maintenance of language sounds. PWM is associated with various linguistic behaviors, including spoken language development (Adams & Gathercole, [<reflink idref="bib1" id="ref18">1</reflink>]), speech accuracy (Waring et al., [<reflink idref="bib80" id="ref19">80</reflink>]), vocabulary, and syntax development (Van der Schuit et al., [<reflink idref="bib77" id="ref20">77</reflink>]) in preschool children. It is comprised of the central executive and the phonological loop that is responsible for maintaining a phonological representation of a novel word (Baddeley, [<reflink idref="bib6" id="ref21">6</reflink>]; Gathercole & Baddeley, [<reflink idref="bib30" id="ref22">30</reflink>]). Phonological memory skill in pre‐reading children was demonstrated to be significantly linked with scores on a reading test at age 8 (Gathercole & Baddeley, [<reflink idref="bib31" id="ref23">31</reflink>]). The most common measures of PWM are pseudoword (or non‐word) repetition; sometimes, digit span or digit span‐running are used as well. In English, the pseudoword repetition (PWR) task is considered to be a powerful tool for describing language performance in both typical and atypical populations due to its correlation with standardized vocabulary assessments and its ability to identify children with or at risk for different language disorders (Coady & Evans, [<reflink idref="bib18" id="ref24">18</reflink>]).</p> <p>Phonological awareness (PA) is a metalinguistic skill that can be understood as an awareness of various phonological segments in speech. It plays a critical role in literacy acquisition in both L1 and L2 (Bowey, [<reflink idref="bib11" id="ref25">11</reflink>]; Engel de Abreu & Gathercole, [<reflink idref="bib28" id="ref26">28</reflink>]; Zifcak, [<reflink idref="bib84" id="ref27">84</reflink>]). Phonemic and phonological awareness are commonly distinguished in the literature based on the unit of analysis. Phonemic awareness is a subtype of phonological awareness, and it refers to the ability to distinguish sounds in a word and implicitly analyze the phonetic form of a word (Elkonin, [<reflink idref="bib27" id="ref28">27</reflink>]). Given that PA refers to an ability to separate the sound of the word from its spelling and to respond to the explicit sound structure of language units, it is reflected in a variety of assessments, each appropriate for different ages and ranging in levels of difficulty: asking the child to tap out the number of component sounds or syllables in a word ("phoneme tapping" or "tapping game" in Liberman et al., [<reflink idref="bib49" id="ref29">49</reflink>]), adding or deleting sounds in words (de Graaff et al., [<reflink idref="bib19" id="ref30">19</reflink>]), or judging which words in a list do not share a common sound or a sequence of sounds (Bradley & Bryant, [<reflink idref="bib12" id="ref31">12</reflink>]).</p> <p>Importantly, the debate outlined in Gathercole and Baddeley ([<reflink idref="bib31" id="ref32">31</reflink>]) over whether PWM and PA have the same or dissociable influences on reading development is not yet fully resolved. The results of a longitudinal study by Mann and Liberman ([<reflink idref="bib52" id="ref33">52</reflink>]) showed that the adequacy of PA and PWM in kindergarten may presage future reading ability in the first grade and that the performance on the corresponding assessments is correlated. Hansen and Bowey ([<reflink idref="bib37" id="ref34">37</reflink>]) used hierarchical multiple regression analyses to examine the unique contribution of phonological analysis (measured with a phonological oddity task) and verbal working memory assessments (as captured by Nonword Repetition, Sentence Imitation, and Rehearsal Rate) to the prediction of reading ability. It was observed that both predictors accounted for unique variations in each of the three reading assessments. Additionally, phonological analysis, but not verbal working memory, assessments particularly strongly predicted pseudoword reading skills. The results of the study suggest that although phonological analysis and verbal working memory skills share a substantial amount of common variance, their various assessments utilize somewhat different reading‐related skills. De Jong and Van der Leij ([<reflink idref="bib20" id="ref35">20</reflink>]) analyzed PA and verbal working memory as predictors of reading acquisition and showed that PA and verbal working memory have a substantial amount of variance in common. However, the study also revealed that the importance of phonological abilities was restricted to the first year of reading instruction. What is more, PA had an independent effect on reading achievement, whereas the effects of verbal working memory could be accounted for by PA.</p> <hd id="AN0183756703-6">Universal and Language‐specific Constraints</hd> <p>The positive relationship between PA and reading ability holds for different types of orthographies, from highly transparent to purely logographic (Goswami, [<reflink idref="bib33" id="ref36">33</reflink>]). The role of PA in different language‐related skills has been widely studied in many languages (e.g., German, as in Schmidt et al., [<reflink idref="bib69" id="ref37">69</reflink>]; Chinese, as in Newman et al., [<reflink idref="bib58" id="ref38">58</reflink>]; English, as in Lundberg et al., [<reflink idref="bib50" id="ref39">50</reflink>] and Treiman, [<reflink idref="bib75" id="ref40">75</reflink>]; Finnish, as in Silvén et al., [<reflink idref="bib71" id="ref41">71</reflink>]) with evidence converging on PA being one of the central components in reading and spelling development. However, the contribution of PA seems to vary in its strength and proposed mechanisms in different languages. It is argued that cross‐linguistic differences in the properties of different languages contribute to the observed heterogeneity of findings. Specifically, the type of orthography, structure and length or types of syllables, and other language‐specific phonological and orthographic properties may affect PA patterns across different languages (Koda, [<reflink idref="bib43" id="ref42">43</reflink>]).</p> <p>For example, Saiegh‐Haddad et al. ([<reflink idref="bib67" id="ref43">67</reflink>]) tested PA in the two languages of Russian and Hebrew sequential bilingual children (<emph>N</emph> = 20) with the aim to identify universal and language‐specific constraints on PA for these languages; they demonstrated that word length and stress were universal, while phoneme position and linguistic context were language‐specific predictors. The effect of word length on phoneme deletion was observed—deleting phonemes from bi‐syllabic words was more challenging than that from monosyllabic words, which is consistent with the well‐established cross‐linguistic hypothesis that longer words are harder for children to manipulate due to phonological and memory processing demands. Additionally, phonemes were easier to delete when they were embedded within stressed than within unstressed syllables. In both Russian and Hebrew, initial phonemes were found to be harder to delete than final phonemes, and the task of deleting initial phonemes was significantly more difficult for 4‐year‐olds than for 5‐year‐olds in both languages. However, the findings did not align with the evidence observed for other languages, suggesting a language‐specific nature of the feature (Saiegh‐Haddad et al., [<reflink idref="bib67" id="ref44">67</reflink>]).</p> <p>Interestingly, in a study by Petchko ([<reflink idref="bib62" id="ref45">62</reflink>]), typically developing Russian children in the first and second grades scored very high on a PA assessment. The test appeared too easy for this group, as their average ability level was nearly one standard deviation above the average difficulty of the test items. In contrast, the Rosner's Auditory Analysis Test, which measures PA in English, showed different results: first graders had a mean score of 17.6 (SD = 8.4) out of 40 items, while sixth graders had a mean score of 29.9 (SD = 6.9). It is not clear whether different languages present significantly different requirements for a speaker's phonological memory capacity. However, there is evidence in other language domains that might suggest so: between languages classified syntactically as left‐branching (LB, modifiers usually precede the head, initial information needs to be retained in working memory) or right‐branching (RB, the phrase begins with the subject and details are expanded on after it), LB speakers were better than RB speakers at recalling initial stimuli in a set of words, numbers, or pictures, but worse at recalling final stimuli (Amici et al., [<reflink idref="bib3" id="ref46">3</reflink>]).</p> <p>Considering all the aspects of Russian phonetics outlined above, we endeavor to conduct an item‐level analysis of the RAS and PSW assessments in a typically developing Russian native sample to find out what phonological features of the items increase the difficulty and discrimination of the assessments in children. Achieving adequate model fit for the selected statistical procedures outlined in the Methods section will lead to successful item performance evaluation and feature reduction necessary to demonstrate how the tests can be improved.</p> <hd id="AN0183756703-7">Methods</hd> <p></p> <hd id="AN0183756703-8">Tasks</hd> <p></p> <hd id="AN0183756703-9">Rosner's Test of Auditory Segmentation (RAS)</hd> <p>One assessment of PA and one of PWM adapted for the Russian phonetic system were used for the current study. Rosner's Test of Auditory Segmentation (RAS; Rosner & Simon, [<reflink idref="bib66" id="ref47">66</reflink>]) was chosen for adaptation as the assessment of PA. RAS quantifies the child's ability to identify the separate sounds in spoken words and the temporal sequence of those sounds by asking the child to delete sounds and voice what is left after the deletion. Children first practiced on two demonstration items, in which they needed to elide a segment from a word (e.g., "say <коса> without <к>"). The elision segments ranged from a single phoneme in the beginning, the middle, or the end of a word to a syllable in compound words—e.g., "say <водолей> without <водо>." The words were selected on the basis that the elimination of a phoneme, phoneme cluster, or syllable would result in another Russian word when pronounced. The items varied in length from one to four syllables. The outcome variable included the response time and accuracy (maximum 40 correct items). The scoring criteria for items were coded as 0–1‐2 (error—self‐correction—pass). For ease of further analysis, the self‐correction category was merged with the error to make each item binary. Table 5 gives an overview of reliability metrics for the RAS on the current study sample.</p> <hd id="AN0183756703-10">Pseudoword Repetition Test (PSW)</hd> <p>The pseudoword repetition test was used as a measure of PWM (Gathercole & Baddeley, [<reflink idref="bib30" id="ref48">30</reflink>]). The test was adapted from the Children's Test of Non‐word Repetition (Gathercole & Baddeley, [<reflink idref="bib32" id="ref49">32</reflink>]). Each participant first practiced on two demonstration items and was then presented with items ranging from 2 to 5 syllables in length with an equal number of items of each length. To systematically vary the complexity of the syllable structure, half of the items contained consonant clusters in the initial syllable onset, and the remaining half contained no clusters, ranging from CV and VC syllables to CVC and to CVCC or CCVC. The test was administered by a native Russian speaker of the same regional dialect using live voice, which was used to ensure that the tested child was not distracted. The test was scored 'online' with each item judged as correctly or incorrectly repeated. The outcome variable included the response time and accuracy (maximum 40). The PSW test has the same scoring criteria that were transformed into binary for the analyses, and the reliability values are given in Table 5.</p> <hd id="AN0183756703-11">Sample Characteristics</hd> <p>The sample included children attending second, third, and fourth grades in seven public schools in a mid‐size city in Russia. Schools were selected from school clusters to represent typical schools in the city. The flyers at the participating schools were used to advertise the study, and families were monetarily compensated for participation. The final dataset includes 502 children (male = 261) studying in 2–4 grades and aged from 7 to 11 years old. The descriptive statistics for the age and the relevant test scores in the sample are given in Table 1, and Table 2 includes descriptive statistics for both tests in the 10th percentile of the sample, which is estimated to be the approximate upper threshold of developmental language disorder (DLD) prevalence (Wagner et al., [<reflink idref="bib79" id="ref50">79</reflink>]). This subset more closely adheres to the intended function of PSW and RAS tests, which is to determine possible signs of DLD. While the size of the subsample is insufficient to conduct a separate IRT analysis, raw endorsement rates can be reviewed to give context to the main analysis.</p> <p>1 TABLE Descriptive Statistics for the Study Sample</p> <p> <ephtml> <table><thead><tr><th align="left">Variable</th><th align="center"><italic>N</italic></th><th align="center">Min</th><th align="center">Max</th><th align="center">Median</th><th align="center">IQR</th><th align="center">Mean</th><th align="center">SD</th></tr></thead><tbody><tr><td align="left">Age</td><td align="left">495</td><td align="left">7</td><td align="left">11</td><td align="left">9</td><td align="left">1</td><td align="char" char=".">9.11</td><td align="char" char=".">0.93</td></tr><tr><td align="left">PSW</td><td align="left">495</td><td align="left">22</td><td align="left">40</td><td align="left">38</td><td align="left">3</td><td align="char" char=".">37.14</td><td align="char" char=".">2.69</td></tr><tr><td align="left">RAS</td><td align="left">495</td><td align="left">8</td><td align="left">40</td><td align="left">33</td><td align="left">6</td><td align="char" char=".">32.73</td><td align="char" char=".">5.08</td></tr></tbody></table> </ephtml> </p> <p>2 TABLE Descriptive Statistics for PSW and RAS Tests in the 10th Percentile</p> <p> <ephtml> <table><thead><tr><th align="left">Measure</th><th align="center"><italic>N</italic></th><th align="center">Min</th><th align="center">Max</th><th align="center">Median</th><th align="center">IQR</th><th align="center">Mean</th><th align="center">SD</th></tr></thead><tbody><tr><td align="left">PSW in 10th percentile</td><td align="left">45</td><td align="char" char=".">22.00</td><td align="char" char=".">33.00</td><td align="char" char=".">32.00</td><td align="char" char=".">3.00</td><td align="char" char=".">30.98</td><td align="char" char=".">2.69</td></tr><tr><td align="left">RAS in 10th percentile</td><td align="left">44</td><td align="char" char=".">8.00</td><td align="char" char=".">26.00</td><td align="char" char=".">24.00</td><td align="char" char=".">7.50</td><td align="char" char=".">21.50</td><td align="char" char=".">5.31</td></tr></tbody></table> </ephtml> </p> <hd id="AN0183756703-12">Statistical Procedures</hd> <p>The grade and gender variability in the sample was investigated using a random intercepts mixed model approach through the R package "lmerTest" (Kuznetsova et al., [<reflink idref="bib47" id="ref51">47</reflink>]), as the data potentially violate the assumptions required to use standard statistical methods such as ordinary least‐squares regression or ANOVA due to children being clustered within schools. Preliminary overviews of the error rates in each assessment were computed as the proportion of wrong answers in the total sample and the 10th percentile.</p> <p>The internal reliability of the assessments was investigated using several different metrics. Cronbach's coefficient alpha, a measure of squared correlation between observed scores and true scores, is appropriate for many types of scales. For the present study, the index is given in both standardized and unstandardized forms; the former is based on item covariance and is used when the scales are comparable (Yu, [<reflink idref="bib82" id="ref52">82</reflink>]). The MS statistic (Molenaar & Sijtsma, [<reflink idref="bib55" id="ref53">55</reflink>], [<reflink idref="bib56" id="ref54">56</reflink>]), unlike previous metrics, gives a direct estimate of the reliability of a test score instead of a lower bound (van der Ark, [<reflink idref="bib76" id="ref55">76</reflink>]). Kuder–Richardson 20 statistic (Kuder & Richardson, [<reflink idref="bib46" id="ref56">46</reflink>]) was also utilized due to its usefulness for dichotomous items.</p> <p>To investigate the differences between items, we utilized models within the Item Response Theory (IRT) framework. IRT is essential for item analysis as it supplies parameter estimates that provide information on the discriminating power separate from its difficulty (Steinberg & Thissen, [<reflink idref="bib72" id="ref57">72</reflink>]). Within the framework of IRT, we fit both 1 and 2 parameter logistic models (1PL and 2PL) using IRTRO (Cai et al., [<reflink idref="bib15" id="ref58">15</reflink>]) to each of the measures separately, comparing the models through the likelihood ratio test of the significance of variation among the slope parameters, which is computed as the difference between—2loglikelihood fit indices for the 1PL and 2PL models (Steinberg & Thissen, [<reflink idref="bib72" id="ref59">72</reflink>]). After model selection, thresholds and slopes of individual binary items were analyzed for the 2PL model to determine what items to select for possible scale improvement. Further attempts to subset the item pool were made through training a classification and regression tree model (CART) for the total score of each measure and a separate model for the proportion of pseudoword test total score and time of completion. The models were trained using 10‐fold cross‐validation with default parameters for the R package "rpart" (Therneau, [<reflink idref="bib74" id="ref60">74</reflink>]) using the "anova" method. Additional IRT model estimation to produce relevant graphs was done using the "ltm" package (Rizopoulos, [<reflink idref="bib65" id="ref61">65</reflink>]).</p> <hd id="AN0183756703-13">Linguistic Metrics of Item Difficulty</hd> <p>To derive conclusions from the results of the IRT and CART model analyses, we needed a measure of phonological complexity that is specific to the Russian language. Previous studies on quantitative measures of phonological complexity used metrics such as the total number of phonemes (Braginsky et al., [<reflink idref="bib13" id="ref62">13</reflink>], which included Russian), as well as more comprehensive measures created for the English language like the Index of Phonetic Complexity (Jakielski, [<reflink idref="bib39" id="ref63">39</reflink>]) or the Word Complexity Measure (Stoel‐Gammon, [<reflink idref="bib73" id="ref64">73</reflink>]). The methods of analysis differ sufficiently between tests, as many of the traditional linguistic metrics would not be applicable to non‐words in the PSW test that have no established role, meaning, or frequency in the language. Therefore, items in the PSW test were analyzed based on the number of syllables and subsyllabic structures, with syllable boundaries determined using Avanesov's ([<reflink idref="bib5" id="ref65">5</reflink>]) sonority sequencing principle. PSW items with consonant clusters were analyzed in terms of Net Auditory Distance (NAD) and cluster preferability for the Russian language by using the NAD Phonotactic Calculator (Dziubalska‐Kołaczyk & Pietrala, [<reflink idref="bib26" id="ref66">26</reflink>]). RAS items, on the other hand, possess a transformation aspect and use natural words as initial and final for the item, for which the usage frequency statistics were estimated in instances‐per‐million (IPM) using Russian National Corpus (RNC, Savchuk et al., [<reflink idref="bib68" id="ref67">68</reflink>]). Items are further differentiated in the elision location (initial, middle, and final) and the occurrence of a stress shift after transformation.</p> <p>To the best of our knowledge, there is no current comprehensive measure of phonological complexity in the Russian language that would consider all of the aforementioned factors. Therefore, we formulated a subsyllabic complexity score to be derived alongside the number of syllables, which includes a sum of all subsyllabic structure elements (simple and complex onsets and codas) weighed according to the number of consonants in the element.</p> <hd id="AN0183756703-14">Results</hd> <p>A random intercept mixed effects model was fitted to the data with grade level and gender as categorical predictors and the school as a grouping variable to estimate the degree of influence of the fixed effects on the test scores. For the PSW test model, no significant differences by gender or grade levels were found, with only the difference between grades 2 and 3 approaching significance (<emph>t</emph> = −1.882, <emph>p</emph> = .06, <emph>Satterthwaite df</emph> = 481.76), Table 3. For the RAS scores, a significant main effect of gender was found (<emph>t</emph> = −3.380, <emph>p</emph> < .01, <emph>Satterthwaite df</emph> = 493.92) with an interaction on grade level 4 approaching significance (<emph>t</emph> = 1.94, <emph>p</emph> = .054, <emph>Satterthwaite df</emph> = 493.950), Table 4. The distribution of values is demonstrated in Figure 1. These sample differences need to be considered while reviewing the effectiveness of the items.</p> <p>3 TABLE Mixed Effects Model for Main Effects of Gender and Grade Level on PSW Test Score and their Interaction</p> <p> <ephtml> <table><thead><tr><th align="left">Predictors</th><th align="center">Estimate</th><th align="center">Std. error</th><th align="center">Df</th><th align="center"><italic>t</italic></th><th align="center"><italic>p</italic> value</th></tr></thead><tbody><tr><td align="left">Intercept</td><td align="char" char=".">37.408</td><td align="char" char=".">0.424</td><td align="char" char=".">7.518</td><td align="char" char=".">88.133</td><td align="char" char="."><.001</td></tr><tr><td align="left">Grade 3</td><td align="char" char=".">−0.811</td><td align="char" char=".">0.431</td><td align="char" char=".">481.756</td><td align="char" char=".">−1.882</td><td align="char" char=".">.060</td></tr><tr><td align="left">Grade 4</td><td align="char" char=".">0.065</td><td align="char" char=".">0.434</td><td align="char" char=".">469.677</td><td align="char" char=".">0.149</td><td align="char" char=".">.882</td></tr><tr><td align="left">Male</td><td align="char" char=".">0.153</td><td align="char" char=".">0.403</td><td align="char" char=".">494.464</td><td align="char" char=".">0.379</td><td align="char" char=".">.705</td></tr><tr><td align="left">Grade 3: Male</td><td align="char" char=".">0.608</td><td align="char" char=".">0.585</td><td align="char" char=".">494.183</td><td align="char" char=".">1.040</td><td align="char" char=".">.299</td></tr><tr><td align="left">Grade 4: Male</td><td align="char" char=".">−0.423</td><td align="char" char=".">0.566</td><td align="char" char=".">494.515</td><td align="char" char=".">−0.746</td><td align="char" char=".">.456</td></tr></tbody></table> </ephtml> </p> <p>1 <emph>Note</emph>: Grade 2 and male used as baseline values. Bold values indicate the significance at the <emph>p</emph> < 0.05 level.</p> <p>4 TABLE Mixed Effects Model for Main Effects of Gender and Grade Level on RAS Test Score and their Interaction</p> <p> <ephtml> <table><thead><tr><th align="left">Predictors</th><th align="center">Estimate</th><th align="center">Std. error</th><th align="center"><italic>Df</italic></th><th align="center"><italic>t</italic></th><th align="center"><italic>p</italic> value</th></tr></thead><tbody><tr><td align="left">Intercept</td><td align="char" char=".">32.356</td><td align="char" char=".">1.009</td><td align="char" char=".">4.889</td><td align="char" char=".">32.068</td><td align="char" char="."><0.001</td></tr><tr><td align="left">Grade 3</td><td align="char" char=".">1.027</td><td align="char" char=".">0.788</td><td align="char" char=".">494.175</td><td align="char" char=".">1.303</td><td align="char" char=".">0.193</td></tr><tr><td align="left">Grade 4</td><td align="char" char=".">0.745</td><td align="char" char=".">0.795</td><td align="char" char=".">491.596</td><td align="char" char=".">0.938</td><td align="char" char=".">0.349</td></tr><tr><td align="left">Male</td><td align="char" char=".">−2.477</td><td align="char" char=".">0.733</td><td align="char" char=".">493.918</td><td align="char" char=".">−3.380</td><td align="char" char=".">0.001</td></tr><tr><td align="left">Grade 3: Male</td><td align="char" char=".">1.464</td><td align="char" char=".">1.065</td><td align="char" char=".">493.656</td><td align="char" char=".">1.374</td><td align="char" char=".">0.170</td></tr><tr><td align="left">Grade 4: Male</td><td align="char" char=".">1.995</td><td align="char" char=".">1.031</td><td align="char" char=".">493.950</td><td align="char" char=".">1.935</td><td align="char" char=".">0.054</td></tr></tbody></table> </ephtml> </p> <p>2 <emph>Note</emph>: Grade 2 and male used as baseline values. Bold values indicate the significance at the <emph>p</emph> < 0.05 level.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0001.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0001.jpg" title="1 Plots of the PSW and RAS Test—Linear Mixed Effects Model by Grade and Gender" /> </p> <p></p> <p>A brief preliminary overview of raw error rates between the total sample and the 10th percentile can be seen in Figures 2 and 3. A general pattern of difficulty in the total sample is followed by the 10th percentile subset rates in the PSW test and in the RAS test. Item‐level differences in the subset are elaborated upon in the Linguistic Analysis section.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0002.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0002.jpg" title="2 Comparison of the Error Rates between the Total Sample and the 10th Percentile in PSW Assessment" /> </p> <p></p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0003.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0003.jpg" title="3 Comparison of the Error Rates between the Total Sample and the 10th Percentile in RAS Assessment" /> </p> <p></p> <p>The reliability estimates for both tests are given in Table 5. While the RAS test is performing above generally acceptable levels, the PSW test shows slightly less desirable reliability scores. Although the reliability score is always expected to increase with an increase in the number of items, this metric represents the quality of the items first and foremost (Wells & Wollack, [<reflink idref="bib81" id="ref68">81</reflink>]).</p> <p>5 TABLE Test Reliability Estimates</p> <p> <ephtml> <table><thead><tr><th align="left">Scale</th><th align="center">Number of items</th><th align="center">Standardized Cronbach's alpha</th><th align="center">CI (2.5%–97.5%)</th><th align="center">Unstandardized Cronbach's alpha</th><th align="center">Molenaar Sijtsma statistic</th><th align="center">Kuder‐ Richardson 20</th></tr></thead><tbody><tr><td align="left">PSW</td><td align="left">40</td><td align="char" char=".">0.666</td><td align="char" char="–">0.666–0.697</td><td align="char" char=".">0.692</td><td align="char" char=".">0.71</td><td align="char" char=".">0.679</td></tr><tr><td align="left">RAS</td><td align="left">40</td><td align="char" char=".">0.813</td><td align="char" char="–">0.803–0.820</td><td align="char" char=".">0.835</td><td align="char" char=".">0.84</td><td align="char" char=".">0.812</td></tr></tbody></table> </ephtml> </p> <p>Two unidimensional IRT models were fitted for each of the measurements: a one‐parameter logistic model (1PL), which constrained all of the item slopes to the same value, and a two‐parameter (2PL) model, in which the item slopes were allowed to vary. The two models were compared with each other using a likelihood ratio goodness‐of‐fit difference in ‐2loglikelihood scores (the technique described in Steinberg & Thissen, [<reflink idref="bib72" id="ref69">72</reflink>]). The significance of variation among the slope parameters for the pseudoword test model is computed as 8681.53 (1PL)—8617.67 (2PL) = 63.86, which is distributed as <emph>χ</emph><sups>2</sups> on 39 <emph>df</emph>, <emph>p</emph> = .007. The significance of the difference in fit for the RAS test model is similarly computed as 16600.46 (1PL)—16517.51 (2PL) = 82.95, which is distributed as <emph>χ</emph><sups>2</sups> on 39 <emph>df</emph>, <emph>p</emph> < .001. Full information on model fit estimates is given in Table 6. The significant difference between models accounted for by freely estimated slopes suggests that there is a significant difference in variation between the items. While the model fit estimates given by the AIC and BIC statistics appear contradictory to the results of the likelihood ratio test, the latter is more appropriate for the comparison of nested models (Lewis et al., [<reflink idref="bib48" id="ref70">48</reflink>]). We also suggest that the inclusion of the slope parameter is informative in the case of the present investigation, as it provides us with more information about the ability of the items to differentiate between the levels of ability. The item information curves for both scales are given in Appendix A, and the item‐level diagnostic statistics (Orlando & Thissen, [<reflink idref="bib59" id="ref71">59</reflink>], [<reflink idref="bib60" id="ref72">60</reflink>]) for 1PL and 2PL models are given in Appendix B. Several items for both scales demonstrate significant <emph>p</emph> values indicating a lack of fit for the assessment (e.g., items "Rikulikiti" and "Hatiburiti" for the PSW test and items "(b)rod," "tsvet(ok)," "(p)loshka," and "ma(ku)shka" for RAS).</p> <p>6 TABLE Goodness‐of‐Fit Statistics for IRT Models</p> <p> <ephtml> <table><thead><tr><th align="left">Scale</th><th align="center">Fit metric</th><th align="center">1PL</th><th>2PL</th></tr></thead><tbody><tr><td align="left">PSW</td><td align="left">‐2loglikelihood</td><td align="char" char=".">8681.53</td><td align="char" char=".">8617.67</td></tr><tr><td align="left">AIC</td><td align="char" char=".">8763.53</td><td align="char" char=".">8777.67</td></tr><tr><td align="left">BIC</td><td align="char" char=".">8936.49</td><td align="char" char=".">9115.15</td></tr><tr><td align="left">RAS</td><td align="left">‐2loglikelihood</td><td align="char" char=".">16600.46</td><td align="char" char=".">16517.51</td></tr><tr><td align="left">AIC</td><td align="char" char=".">16682.46</td><td align="char" char=".">16677.51</td></tr><tr><td align="left">BIC</td><td align="char" char=".">16855.43</td><td align="char" char=".">17,015</td></tr></tbody></table> </ephtml> </p> <p>3 <emph>Note</emph>: AIC, akaike information criterion; BIC, bayesian information criterion.</p> <p>The results of the IRT analyses demonstrate a significant difference between some items in both assessments, suggesting that the scales can be restructured. However, it is difficult to decide which items need to be left out based on their IRT representation, as the model parameters for the majority of items are uninformative: the information curves demonstrate that most items do not tell us anything about the level of the latent construct of each child, and a minority of items allow us to discriminate between low and below average levels of ability.</p> <p>To facilitate the potential restructuring of the PSW and RAS assessments to reduce the response burden on children and more efficiently discriminate between higher levels of ability, we utilized CART modeling to determine the most useful items in predicting a high score for both assessments. This method has recently been proven useful in item analysis contexts (see Michel et al., [<reflink idref="bib54" id="ref73">54</reflink>]; Peute et al., [<reflink idref="bib63" id="ref74">63</reflink>]). Given that the time of completion is recorded for PSW but not for the RAS test, we fitted three separate models: total scores in PSW and RAS and a proportion of PSW total score to time taken. All models were trained through 10‐fold cross‐validation to avoid overfitting. As a result, the models were able to achieve <emph>R</emph><sups>2</sups> = 0.502 (relative—0.661) for the PSW total score model and <emph>R</emph><sups>2</sups> = 0.468 (relative—0.674) for the RAS total score model. Sufficiently less variance was explained by the PSW model that used a proportion of total score to completion time (<emph>R</emph><sups>2</sups> = 0.127), which is why this model is not utilized further, and training details are omitted. Figures 4 and 5 demonstrate that the improvement of the models was mostly linear, with the best performance achieved at CP = 0.01. Figures 6 and 7 show the resulting regression tree structures. Variable importance scores are demonstrated in Figures 8 and 9.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0004.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0004.jpg" title="4 PSW Total Score CART Model R2 by Iteration" /> </p> <p></p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0005.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0005.jpg" title="5 RAS Total Score CART Model R2 by Iteration" /> </p> <p></p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0006.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0006.jpg" title="6 PSW Total Score CART Model Tree Structure" /> </p> <p></p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0007.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0007.jpg" title="7 RAS Total Score CART Model Tree Structure" /> </p> <p></p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0008.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0008.jpg" title="8 Variable Importance Scores for the PSW Assessment" /> </p> <p></p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0009.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0009.jpg" title="9 Variable Importance Scores for the RAS Assessment" /> </p> <p></p> <hd id="AN0183756703-24">Linguistic Item Analysis</hd> <p>In order to reflect on the results obtained from the IRT and CART models, we compare the items in each assessment in terms of relevant phonological metrics. Figure 10 gives an overview of syllable counts and subsyllabic scores in the PSW assessment, where we can see a trend towards higher subsyllabic score values in items that were selected by the CART model. The observed trend is also supported by high NAD values in the consonant clusters for the items in the CART model subset—with the item "Rambatadzhika" having the largest sum of cluster auditory distances out of all items in the scale and being the second most important variable in the model, and "Isvontenie" demonstrating the second largest sum of cluster NADs and rank 9 on feature importance. We also note that most clusters in both subsets are judged as preferable by the principle devised in (Dziubalska‐Kołaczyk & Pietrala, [<reflink idref="bib26" id="ref75">26</reflink>])—the preferability decision is, however, given not on the basis of production difficulty but on the ability to be easily learned and transmitted. All distance values for item consonant clusters, as well as their preferability, are given in Table 7.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jan25/rrq604-fig-0010.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq604-fig-0010.jpg" title="10 Syllable Lengths and Subsyllabic Scores for PSW Items Included (Top) and Excluded (Bottom) from the CART Model" /> </p> <p></p> <p>7 TABLE Net Auditory Distances for Consonant Clusters and their Preferability in the PSW Assessment</p> <p> <ephtml> <table><thead><tr><th align="left" /><th align="center">Item</th><th align="center">Consonant cluster</th><th align="center">Sum of cluster distances</th><th>Product of syllable number and total distance</th><th align="center">Preferability of cluster</th></tr></thead><tbody><tr><td align="left">Model</td><td align="left">Eletmokavaj</td><td align="left">etmɒ</td><td align="left">43</td><td align="char" char=".">215</td><td align="left">Yes</td></tr><tr><td align="left">Rambatadzhika</td><td align="left">amba, adʒɨ</td><td align="left">76</td><td align="char" char=".">380</td><td align="left">Yes, Yes</td></tr><tr><td align="left">Tatruarnat</td><td align="left">atɹu, aɹna</td><td align="left">63</td><td align="char" char=".">252</td><td align="left">Yes, No</td></tr><tr><td align="left">Eburanzik</td><td align="left">anzi</td><td align="left">26</td><td align="char" char=".">104</td><td align="left">Yes</td></tr><tr><td align="left">Shchivomotka</td><td align="left">ɒtka</td><td align="left">40</td><td align="char" char=".">160</td><td align="left">Yes</td></tr><tr><td align="left">Isvontenie</td><td align="left">isvɔ, ɔnte</td><td align="left">73</td><td align="char" char=".">365</td><td align="left">Yes, Yes</td></tr><tr><td align="left">Despakenie</td><td align="left">espa</td><td align="left">43</td><td align="char" char=".">215</td><td align="left">Yes</td></tr><tr><td align="left">Zarnok</td><td align="left">ʌɹnɔ</td><td align="left">21</td><td align="char" char=".">42</td><td align="left">No</td></tr><tr><td align="left">Odrivarec</td><td align="left">ədɹi</td><td align="left">27</td><td align="char" char=".">108</td><td align="left">No</td></tr><tr><td align="left">Celitra</td><td align="left">itɹa</td><td align="left">36</td><td align="char" char=".">108</td><td align="left">No</td></tr><tr><td align="left">Krosmynya</td><td align="left">ʌsmɨ</td><td align="left">33</td><td align="char" char=".">99</td><td align="left">Yes</td></tr><tr><td align="left">Non‐Model</td><td align="left">Ungalizi</td><td align="left">unɡa</td><td align="left">33</td><td align="char" char=".">132</td><td align="left">Yes</td></tr><tr><td align="left">Posvorg</td><td align="left">ɔswɔ, ɔɹɡ</td><td align="left">51</td><td align="char" char=".">102</td><td align="left">No</td></tr><tr><td align="left">Morvashki</td><td align="left">əɹwʌ</td><td align="left">10</td><td align="char" char=".">30</td><td align="left">Yes</td></tr><tr><td align="left">Isturonka</td><td align="left">istu, ɔnkʌ</td><td align="left">72</td><td align="char" char=".">288</td><td align="left">Yes, Yes</td></tr><tr><td align="left">Trambor</td><td align="left">ʌmbɔ</td><td align="left">35</td><td align="char" char=".">70</td><td align="left">Yes, Yes</td></tr><tr><td align="left">Astrumizm</td><td align="left">astɹu, izm</td><td align="left">56</td><td align="char" char=".">168</td><td align="left">No, No</td></tr><tr><td align="left">Estratist</td><td align="left">ɛstɹæ, ist</td><td align="left">52</td><td align="char" char=".">156</td><td align="left">No, No</td></tr><tr><td align="left">Tanmat</td><td align="left">ʌnmʌ</td><td align="left">24</td><td align="char" char=".">48</td><td align="left">Yes</td></tr><tr><td align="left">Ktovo</td><td align="left">ktɔ</td><td align="left">22</td><td align="char" char=".">44</td><td align="left">No</td></tr></tbody></table> </ephtml> </p> <p>The points of comparison between the items in the RAS test, on the other hand, are not as clear—the elision component precludes us from analyzing the difficulty of production of constituent words only. In that aspect, a clear difference is found in the elision type—the model subset items predominantly feature middle elision, while items not selected by the CART model overwhelmingly lean towards initial elision. No significant differences in magnitude between items were observed in the categories of subsyllabic complexity, number of syllables, and IPM frequency of the initial and final words of the item. However, the pattern of differences in phonological complexity characteristics of the RAS test items between the ones selected by the CART model and other items still mostly correlates with the pattern for the PSW test. The distribution of IPM frequencies in the items is less informative on its own: the patterns of distribution follow Zipf's law (Zipf, [<reflink idref="bib85" id="ref76">85</reflink>]), with shorter words, on average, being more frequent. All characteristics of RAS item sets are given in Table 8.</p> <p>8 TABLE Proposed Linguistic Item Difficulty Metrics for the RAS Assessment</p> <p> <ephtml> <table><thead><tr><th align="left" /><th align="center">Mean number of syllables</th><th align="center">Mean subsyllabic complexity score</th><th align="center">Mean frequency of initial word (IPM)</th><th align="center">Mean frequency of final word (IPM)</th><th align="center">Elision type (mode)</th><th align="center">Stress shift (%)</th></tr></thead><tbody><tr><td align="left">Model Set</td><td align="char" char=".">2.17</td><td align="char" char=".">3.17</td><td align="char" char=".">59.43</td><td align="char" char=".">324.46</td><td align="left">Middle</td><td align="char" char=".">45.83</td></tr><tr><td align="left">Non‐Model Set</td><td align="char" char=".">1.96</td><td align="char" char=".">3.10</td><td align="char" char=".">92.90</td><td align="char" char=".">61.31</td><td align="left">Initial</td><td align="char" char=".">37.5</td></tr></tbody></table> </ephtml> </p> <p>A few small differences in item endorsement can be seen in the 10<sups>th</sups> percentile by total raw score as compared to the total sample. The highest error rates in the subsample for the PSW test are demonstrated by the "Tatruarnat" and "Krosmynya," with the overall pattern of highest error items fully aligned with variable importance scores for the CART model (Figure 2). For the RAS test, the item with the highest rate of error is "makar(on)," which was unexpected due to its low variable importance score in the overall CART model (Figure 3). However, it possesses all the identifying characteristics for items with higher difficulty outlined above, featuring both a rarely seen final elision and a relatively high number of syllables. The same can be seen for the item "ma(ku)shka," which, despite showing the highest error score in the total sample and one of the highest in the 10th percentile, is not featured in the most important variable list for the CART model. This item also features a middle elision segment.</p> <hd id="AN0183756703-26">Discussion</hd> <p>Neither PSW nor RAS performance demonstrated a significant effect of age on the test score as a result of the preliminary mixed model analysis. That finding is expected due to the age range of the sample and aligns with the existing literature: a study of complex phonological tasks in Russian (pseudoword repetition task being one of them) in 7–11 years old (<emph>N</emph> = 105) did not observe an effect of age on performance (Dorofeeva et al., [<reflink idref="bib25" id="ref77">25</reflink>]). However, substantial variability in RAS performance is observed in Russian children of preschool age (mean age of 5.2 years, SD = 0.9) when PA is emerging developmentally (Grigorenko, [<reflink idref="bib35" id="ref78">35</reflink>]).</p> <p>We observed a significant effect of gender for RAS performance but not for PSW performance. The evidence about gender differences in phonological processing is inconsistent: some studies found no differences (Burt et al., [<reflink idref="bib14" id="ref79">14</reflink>]; Hecht & Greenfield, [<reflink idref="bib38" id="ref80">38</reflink>]), whereas others demonstrate that girls have an advantage in PA (Dorofeeva et al., [<reflink idref="bib25" id="ref81">25</reflink>]; Lundberg et al., [<reflink idref="bib51" id="ref82">51</reflink>]). The results of the analysis, therefore, do not seem significant enough to invalidate any further conclusions to be made on the item's performance.</p> <p>IRT model comparison between 1PL and 2PL models has proven that a significant variation in discrimination between items exists, with the most likely source of it being the high slope of some of the more complex items (e.g., "Rambatadzhika" for the PSW test or "u(shche)l'ya" for the RAS) and an extremely low slope for the overwhelming majority of the items in both scales that would bias the common slope for the 1PL model. For the 2PL model, the information curves for the PSW have many shallow slopes (e.g., items "Notso," "Peneta," "Shezera," and "Upu"), with thresholds for most items being sufficiently below the mean on the latent trait axis. The discrimination rates for the RAS are marginally better, with some item thresholds (items "Ma(ku)shka" and "u(shche)l'ya") seen slightly below or above the mean on the latent construct. The overall conclusion that can be made from IRT analysis of the PSW and RAS assessments in this sample would be its marked low difficulty and discrimination—items in these measures provide us almost no information on any of the respondents with scores at or above the mean on the latent construct.</p> <p>We attempted to restructure the scales through the formulation of predictive and non‐predictive item subsets using CART modeling. We analyzed the linguistic features of items in these subsets to arrive at a common characteristic predictive of high item difficulty. We encourage the usage of this method for similar item selection, as this model type not only highlights the most difficult items, it also points out easier ones appropriate for discriminating between lower levels of ability, arriving at a balanced item set, the structure of which can be preserved and expanded with similar entries. For the PSW and RAS tests, linguistic analysis results have determined what features would be important for the new items added to the scale—maximizing those metrics will likely sufficiently increase the difficulty and discrimination of the items.</p> <p>For the PSW test, we observed a trend towards a higher number of syllables and subsyllabic structure scores in the items included in the CART model subset. Other similar investigations also report the high importance of these metrics. A significant difference in the performance on non‐word repetition test items, including a higher number of consonant clusters, was found between a specific language impairment (SLI) group (typically referred to in modern studies as DLD) and both age‐matched and language‐matched typically developing groups (Archibald & Gathercole, [<reflink idref="bib4" id="ref83">4</reflink>]). An inquiry into the effect of subsyllabic structure was also made by Kavitskaya et al. ([<reflink idref="bib40" id="ref84">40</reflink>]), demonstrating that the pseudoword repetition performance was affected by syllable structure complexity in children with SLI. A similar decrease in performance was found while comparing children with SLI, who typically develop controls on consonant cluster production in word‐medial compared to word‐initial position and in unstressed syllables compared to stressed syllables. All groups, however, differed in new cluster production in the initial rather than middle position (Marshall & van der Lely, [<reflink idref="bib53" id="ref85">53</reflink>]). There is also linguistic evidence showing that producing a correct phoneme in an unstressed syllable in Russian requires finding a cognate for the word, which would be challenging both in cases of typical and atypical development (Kornev et al., [<reflink idref="bib45" id="ref86">45</reflink>]). The highest NAD values for consonant clusters are mostly found in highly predictive items, with the preferability of the cluster seemingly not playing an important role. One possible extension of the current study's analysis would be the phonotactic probability of varying phoneme combinations, including consonant‐vowel combinations.</p> <p>For the RAS test, a clear difference is found in the elision type—the model subset items predominantly feature middle elision, while items not selected by the CART model overwhelmingly lean towards initial elision. This conclusion is supported by the literature, showing that initial phonemes are universally easier for children to isolate and delete, even in a sufficiently transparent language (Aidinis & Nunes, [<reflink idref="bib2" id="ref87">2</reflink>]; Papadopoulos et al., [<reflink idref="bib61" id="ref88">61</reflink>]), with an investigation into phoneme‐related neural congruency components using EEG signals suggesting that the final elision seems to require an additional and distinct set of neural processes beyond the ones required to process the initial elision (Christoforou et al., [<reflink idref="bib17" id="ref89">17</reflink>]). We also suppose that the presence of a stress shift in the final word of the item after the elision is completed would make the item more difficult, but research on that aspect of the RAS is scarce. A closer look at the CART model structure might partially point to this, with the item with the highest variable importance rank—"di(re)ktor," featuring a middle elision and a stress shift, as well as the item "nos(ok)" used to classify the upper 66% of observations, also including a stress shift and a final elision.</p> <p>The overall pattern of differences in high and low‐performing items for both tests was also observed in the 10<sups>th</sups> percentile of the sample. Certain items that show a high error rate in the 10<sups>th</sups> percentile despite having low predictive potential for the overall sample demonstrate the same profile of characteristics—such as the item "solo(mi)nka" in the RAS assessment featuring a high number of syllables and a middle elision. While the main focus of the article is on item performance in the total sample as it relates to the features of pseudowords or elision exercises most children would find difficult, we include these observations to better contextualize these findings.</p> <p>The data on consonant acquisition by Russian monolingual children shows a great variability in the time and order of acquisition by different children with the age range from 1 year and 3 months to 6 years at the latest (Kistanova, [<reflink idref="bib42" id="ref90">42</reflink>]). Thus, low difficulty and discrimination for both tests are most likely to be explained by phonological development, which mostly occurs at an earlier age than the range included in our sample. Additionally, as demonstrated in the study of PA in illiterate adults, strong awareness of speech as a sequence of phonemes does not arise spontaneously as an epiphenomenon of general cognitive growth; it requires direct instructional support (Morais et al., [<reflink idref="bib57" id="ref91">57</reflink>]) and strongly depends on the acquisition of the alphabet (de Santos Loureiro et al., [<reflink idref="bib21" id="ref92">21</reflink>]). Relatively high performance on both RAS and PSW is expected in a sample of typically developing children involved in formal schooling. If a sufficiently high score threshold is set for both tests, they can be viewed as effective in detecting developmental language disability. However, we want to use the results of the present study to demonstrate how difficult it might be to develop new items and analyze their performance on such skewed assessments. If the assessments were to be expanded with more difficult items, the observed trends in linguistic characteristics show that certain success can be achieved through a linear increase in such features as syllable length, subsyllabic complexity, number and complexity of consonant clusters, and introducing more unusual elision segments.</p> <hd id="AN0183756703-27">Limitations and Future Directions</hd> <p>A possible limitation of the study is the significant gender difference in the RAS assessment. Future attempts at item‐level analysis of the Russian language adaptations of PSW and RAS should include separate models for each gender to provide a more in‐depth assessment and potentially locate differences in the acquisition of certain phonemic contrasts, provided that sufficient power can be achieved. While the limitation concerning the age range of the sample was partially overcome through CART modeling, a certain adjustment to it might result in a more clearly defined latent construct continuum. Multiple other measures of PA and PWM that have shown to be effective in different languages (such as rapid automatized naming) can be added to the assessment pool to capture more variance in the target constructs.</p> <p>Another promising future direction of research can be the development of new PA and PWM assessments, which would introduce higher difficulty items created through linear increases in the observed linguistic complexity metrics. One use for such assessments may be their potential to detect high levels of ability—"phonological talent" (Selinker, [<reflink idref="bib70" id="ref93">70</reflink>]; as cited in Biedroń & Pawlak, [<reflink idref="bib9" id="ref94">9</reflink>]) or giftedness. The study by Filippova ([<reflink idref="bib29" id="ref95">29</reflink>]) analyzed whether PA in L1 is related to reading abilities in L2. Phonological abilities (as measured by phoneme isolation task, number of sounds task, and changing a sound in a pseudoword task) in Russian children (<emph>n</emph> = 37, 9–11 years old) could predict reading and spelling skills in English as a second language. High PWM is known as one characteristic feature of hyperpolyglots, which assists in overcoming the decline in phonetic abilities after the sensitive period and continuing to acquire new sounds and words successfully (Biedroń & Pawlak, [<reflink idref="bib9" id="ref96">9</reflink>]). Several neuroimaging studies also confirm that successful learning of native phonetic contrasts is predictive of foreign phonetic contrast learning, which results in the recruitment of the same areas that are involved during the processing of native contrasts (Díaz et al., [<reflink idref="bib23" id="ref97">23</reflink>]; Díaz et al., [<reflink idref="bib24" id="ref98">24</reflink>]). In addition to that, a recent consensus is currently forming on the role of phonological skills developing independently from general intelligence and not compensated by it (van Viersen et al., [<reflink idref="bib78" id="ref99">78</reflink>]; as cited in Desvaux et al., [<reflink idref="bib22" id="ref100">22</reflink>]). Therefore, we suppose that it is possible to develop an assessment that can discriminate between above‐average and high levels of phonological ability and may help to detect a distinct linguistic talent useful for foreign language acquisition and acquiring multiple languages. Such a version of the PSW test would include 4–5 syllable pseudowords with complex onsets and/or codas, containing consonant clusters with a large sum of phonological distances, while the RAS test should include 3–4 syllable words with predominantly middle or final elision and a stress shift in the final word that remain after elision. One of the advantages of such an assessment would be its potential for early application (before the start of formal foreign language learning), as it uses only the native language's phonological system. The modified versions of these assessments can be used by educators and school psychologists to examine the levels of phonological ability across the whole spectrum and administer appropriate interventions if necessary.</p> <hd id="AN0183756703-28">Conclusion</hd> <p>Pseudoword repetition and Rosner's Auditory Segmentation assessments are effective in discriminating between lower levels of phonological awareness and phonological working memory constructs in typically developing Russian children, as demonstrated by the IRT analysis results. In order to tap into higher levels of the construct, both scales need to be restructured and have more difficult items added. The difficulty of pseudoword tasks can be increased by including a large number of syllables (<reflink idref="bib4" id="ref101">4</reflink>, 5) and simultaneously increasing the subsyllabic complexity (feature multiple complex onsets or codas), while the auditory segmentation items would increase in difficulty by featuring middle or final elision and possibly a shift in vowel stress, as suggested by the structural features of the items deemed highly important by the CART models. The findings and methods of the present study can be used to develop and update phonological assessments to differentiate between more distinct levels of reading ability in typically developing and SLI children.</p> <hd id="AN0183756703-29">Acknowledgments</hd> <p>The authors express their gratitude to the two anonymous peer reviewers and the special issue editors for giving their valuable feedback that has immensely enhanced the quality of this article. They are also grateful to Lauren Elderton for her editorial assistance.</p> <hd id="AN0183756703-30">Funding Information</hd> <p>This research was supported by National Institutes of Health Grant R01 DC007665 (E.L.G., Principal Investigator) and the Russian Science Foundation (grant no. 18‐18‐00451; E.L.G., Principal Investigator). Grantees undertaking such projects are encouraged to freely express their professional judgment. Therefore, this article does not necessarily reflect the position or policies of the abovementioned agencies, and no official endorsement should be inferred.</p> <hd id="AN0183756703-31">Conflict of Interest Statement</hd> <p>The authors report no known conflicts of interest.</p> <hd id="AN0183756703-32">Ethics Statement</hd> <p>Approval STUDY00000093 from the University of Houston.</p> <hd id="AN0183756703-33">Data Availability Statement</hd> <p>The data are not publicly available but can be shared with individuals who enter into a data‐sharing contract with the University of Houston.</p> <p>GRAPH: Data S1.</p> <ref id="AN0183756703-34"> <title> References </title> <blist> <bibl id="bib1" idref="ref18" type="bt">1</bibl> <bibtext> Adams, A. M., & Gathercole, S. E. (1995). Phonological working memory and speech production in preschool children. Journal of Speech, Language, and Hearing Research, 38 (2), 403 – 414.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref87" type="bt">2</bibl> <bibtext> Aidinis, A., & Nunes, T. (2001). The role of different levels of phonological awareness in the development of reading and spelling in Greek. Reading and Writing, 14, 145 – 177.</bibtext> </blist> <blist> <bibl id="bib3" idref="ref46" type="bt">3</bibl> <bibtext> Amici, F., Sánchez‐Amaro, A., Sebastián‐Enesco, C., Cacchione, T., Allritz, M., Salazar‐Bonet, J., & Rossano, F. (2019). The word order of languages predicts native speakers' working memory. Scientific Reports, 9 (1), 1124. https://doi.org/10.1038/s41598‐018‐37654‐9</bibtext> </blist> <blist> <bibl id="bib4" idref="ref83" type="bt">4</bibl> <bibtext> Archibald, L. M. D., & Gathercole, S. E. (2006). Nonword repetition: A comparison of tests. Journal of Speech, Language, and Hearing Research, 49 (5), 970 – 983. https://doi.org/10.1044/1092‐4388(2006/070)</bibtext> </blist> <blist> <bibl id="bib5" idref="ref65" type="bt">5</bibl> <bibtext> Avanesov, R. I. (1954). О слогоразделе и строении слога в русском языке [About the syllabic division and syllable structure in the Russian language]. Voprosy yazykoznaniya, 6, 88 – 101.</bibtext> </blist> <blist> <bibl id="bib6" idref="ref21" type="bt">6</bibl> <bibtext> Baddeley, A. (1992). Working memory. Science, 255 (5044), 556 – 559.</bibtext> </blist> <blist> <bibl id="bib7" idref="ref1" type="bt">7</bibl> <bibtext> Blachman, B. A. (2000). Phonological awareness. In M. L. Kamil, P. B. Mosenthal, P. D. Pearson, & R. Barr (Eds.), Handbook of reading research (Vol. 3, pp. 483 – 502). Routledge.</bibtext> </blist> <blist> <bibl id="bib8" idref="ref12" type="bt">8</bibl> <bibtext> Bogomazov, G. M. (2001). Sovremennyj RusskijJazyk. Fonetika. [Modern Russian Language. Phonetics]. Мoscow:Vlados.</bibtext> </blist> <blist> <bibl id="bib9" idref="ref94" type="bt">9</bibl> <bibtext> Biedroń, A., & Pawlak, M. (2016). New conceptualizations of linguistic giftedness. Language Teaching, 49 (2), 151 – 185.</bibtext> </blist> <blist> <bibtext> Bondarko, L. V. (1998). Fonetika sovremennogo russkogo jazyka. In [phonetics of modern Russian]. St. St. Petersburg University.</bibtext> </blist> <blist> <bibtext> Bowey, J. A. (2000). Recent developments in language acquisition and reading research: The phonological basis of children's reading difficulties. The Educational and Developmental Psychologist, 17 (1), 5 – 31.</bibtext> </blist> <blist> <bibtext> Bradley, L., & Bryant, P. E. (1978). Difficulties in auditory organisation as a possible cause of reading backwardness. Nature, 271 (5647), 746 – 747.</bibtext> </blist> <blist> <bibtext> Braginsky, M., Yurovsky, D., Marchman, V. A., & Frank, M. C. (2019). Consistency and variability in children's word learning across languages. Open Mind, 3, 52 – 67.</bibtext> </blist> <blist> <bibtext> Burt, L., Holm, A., & Dodd, B. (1999). 1999: Phonological awareness skills of 4‐year‐old British children: An assessment and developmental data. International Journal of Language and Communication Disorders, 34, 311 – 335.</bibtext> </blist> <blist> <bibtext> Cai, L., Du Toit, S. H. C., & Thissen, D. (2011). IRTPRO: Flexible, multidimensional, multiple categorical IRT modeling. Scientific Software International.</bibtext> </blist> <blist> <bibtext> Caravolas, M. (2005). The nature and causes of dyslexia in different languages. In M. J. Snowling & C. Hulme (Eds.), The science of reading: A handbook (pp. 336 – 355). Blackwell Publishing. https://doi.org/10.1002/9780470757642.ch18</bibtext> </blist> <blist> <bibtext> Christoforou, C., Theodorou, M., Fella, A., & Papadopoulos, T. C. (2023). Phonological ability and neural congruency: Phonological loop or more? Clinical Neurophysiology, 156, 228 – 241.</bibtext> </blist> <blist> <bibtext> Coady, J. A., & Evans, J. L. (2008). Uses and interpretations of non‐word repetition tasks in children with and without specific language impairments (SLI). International Journal of Language & Communication Disorders, 43 (1), 1 – 40.</bibtext> </blist> <blist> <bibtext> de Graaff, S., Hasselman, F., Verhoeven, L., & Bosman, A. M. (2011). Phonemic awareness in Dutch kindergartners: Effects of task, phoneme position, and phoneme class. Learning and Instruction, 21 (1), 163 – 173.</bibtext> </blist> <blist> <bibtext> De Jong, P. F., & Van der Leij, A. (1999). Specific contributions of phonological abilities to early reading acquisition: Results from a Dutch latent variable longitudinal study. Journal of Educational Psychology, 91 (3), 450 – 476.</bibtext> </blist> <blist> <bibtext> de Santos Loureiro, C., Braga, L. W., do Nascimento Souza, L., Nunes Filho, G., Queiroz, E., & Dellatolas, G. (2004). Degree of illiteracy and phonological and metaphonological skills in unschooled adults. Brain and Language, 89 (3), 499 – 502.</bibtext> </blist> <blist> <bibtext> Desvaux, T., Danna, J., Velay, J. L., & Frey, A. (2024). From gifted to high potential and twice exceptional: A state‐of‐the‐art meta‐review. Applied Neuropsychology: Child, 13 (2), 165 – 179.</bibtext> </blist> <blist> <bibtext> Díaz, B., Baus, C., Escera, C., Costa, A., & Sebastián‐Gallés, N. (2008). Brain potentials to native phoneme discrimination reveal the origin of individual differences in learning the sounds of a second language. Proceedings of the National Academy of Sciences, 105 (42), 16083 – 16088.</bibtext> </blist> <blist> <bibtext> Díaz, B., Mitterer, H., Broersma, M., & Sebastián‐Gallés, N. (2012). Individual differences in late bilinguals' L2 phonological processes: From acoustic‐phonetic analysis to lexical access. Learning and Individual Differences, 22 (6), 680 – 689.</bibtext> </blist> <blist> <bibtext> Dorofeeva, S. V., Laurinavichyute, A., Reshetnikova, V., Akhutina, T. V., Tops, W., & Dragoy, O. (2020). Complex phonological tasks predict reading in 7 to 11 years of age typically developing Russian children. Journal of Research in Reading, 43 (4), 516 – 535.</bibtext> </blist> <blist> <bibtext> Dziubalska‐Kołaczyk, K., & Pietrala, D. (2014). The NAD phonotactic calculator–an online tool to calculate cluster preference in English, polish and other languages. In H. Sauer & P. P. Chruszczewski (Eds.), Mostly medieval: In memory of Jack Fisiak. (Beyond Language 5) (pp. 445 – 458). Æ Academic Publishing.</bibtext> </blist> <blist> <bibtext> Elkonin, D. B. (1973). U.S.S.R. In J. Downing (Ed.), Comparative reading: Cross‐ national studies of behavior and processes in reading and writing (pp. 551 – 579). Macmillan.</bibtext> </blist> <blist> <bibtext> Engel de Abreu, P. M., & Gathercole, S. E. (2012). Executive and phonological processes in second‐language acquisition. Journal of Educational Psychology, 104 (4), 974 – 986.</bibtext> </blist> <blist> <bibtext> Filippova, N. (2022). Phonological awareness and rapid automatized naming as predictors of development of reading and spelling skills among Russian‐speaking learners of English (Doctoral dissertation, LCC tarptautinis universitetas.).</bibtext> </blist> <blist> <bibtext> Gathercole, S. E., & Baddeley, A. D. (1990). Phonological memory deficits in language disordered children: Is there a causal connection? Journal of Memory and Language, 29, 336 – 360.</bibtext> </blist> <blist> <bibtext> Gathercole, S. E., & Baddeley, A. D. (1993). Phonological working memory: A critical building block for reading development and vocabulary acquisition? European Journal of Psychology of Education, 8 (3), 259 – 272.</bibtext> </blist> <blist> <bibtext> Gathercole, S. E., & Baddeley, A. D. (1996). The Children's test of nonword repetition. Psychological Corporation.</bibtext> </blist> <blist> <bibtext> Goswami, U. (1999). The relationship between phonological awareness and orthographic representation in different orthographies. In M. Harris & G. Hatano (Eds.), Learning to read and write: A cross‐linguistic perspective (pp. 134 – 156). Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Goswami, U. (2010). A psycholinguistic grain size view of reading acquisition across languages. In Reading and dyslexia in different orthographies (pp. 41 – 60). Psychology Press.</bibtext> </blist> <blist> <bibtext> Grigorenko, E. L. (2011). Этап «дочтения», его характеристики и их роль в овладении чтением [Pre‐reading, its characteristics and their role in reading mastery]. Voprosy Psikhologii, 6, 122 – 130.</bibtext> </blist> <blist> <bibtext> Grigorenko, E. L., & Elliott, J. J. (2012). Чтение о чтении [Reading about reading]. Aist.</bibtext> </blist> <blist> <bibtext> Hansen, J., & Bowey, J. A. (1994). Phonological analysis skills, verbal working memory, and reading ability in second‐grade children. Child Development, 65 (3), 938 – 950.</bibtext> </blist> <blist> <bibtext> Hecht, S. A., & Greenfield, D. B. (2002). Explaining the predictive accuracy of teacher judgments of their students' reading achievement: The role of gender, classroom behavior, and emergent literacy skills in a longitudinal sample of children exposed to poverty. Reading and Writing, 15, 789 – 809.</bibtext> </blist> <blist> <bibtext> Jakielski, K. J. (2022). The index of phonetic complexity: At‐a‐glance scoring system, terminology, instructions, & data forms. Building Speech and Quantifying Complexity. https://digitalcommons.augustana.edu/csdbuildingspeech/10</bibtext> </blist> <blist> <bibtext> Kavitskaya, D., Babyonyshev, M., Walls, T., & Grigorenko, E. (2011). Investigating the effects of syllable complexity in Russian‐speaking children with SLI. Journal of Child Language, 38 (5), 979 – 998.</bibtext> </blist> <blist> <bibtext> Kerek, E., & Niemi, P. (2012). Grain‐size units of phonological awareness among Russian first graders. Written Language & Literacy, 15 (1), 80 – 113.</bibtext> </blist> <blist> <bibtext> Kistanova, E. V. (2021). The Acquisition of Russian Word‐Initial Consonant Clusters in a Russian‐American English bilingual child: An overview of one Child's exceptional production. Journal of Slavic Linguistics, 29 (3), 1 – 12.</bibtext> </blist> <blist> <bibtext> Koda, K. (2007). Reading and language learning: Crosslinguistic constraints on second language reading development. Language Learning, 57, 1 – 44.</bibtext> </blist> <blist> <bibtext> Kogan, N., & Saiegh‐Haddad, E. (2023). The internal structure of the syllable in Russian and in Hebrew: Evidence from monolingual kindergarteners. Applied PsychoLinguistics, 44 (1), 101 – 129.</bibtext> </blist> <blist> <bibtext> Kornev, A. N., Rakhlin, N., & Grigorenko, E. L. (2010). Dyslexia from a cross‐linguistic and cross‐cultural perspective: The case of Russian and Russia. Learning Disabilities: A Contemporary Journal, 8 (1), 41 – 69.</bibtext> </blist> <blist> <bibtext> Kuder, G. F., & Richardson, M. W. (1937). The theory of the estimation of test reliability. Psychometrika, 2 (3), 151 – 160.</bibtext> </blist> <blist> <bibtext> Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2017). lmerTest package: Tests in linear mixed effects models. Journal of Statistical Software, 82 (13), 1 – 26. https://doi.org/10.18637/jss.v082.i13</bibtext> </blist> <blist> <bibtext> Lewis, F., Butler, A., & Gilbert, L. (2011). A unified approach to model selection using the likelihood ratio test. Methods in Ecology and Evolution, 2 (2), 155 – 162.</bibtext> </blist> <blist> <bibtext> Liberman, I. Y., Shankweiler, D., Fischer, F. W., & Carter, B. (1974). Explicit syllable and phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18 (2), 201 – 212.</bibtext> </blist> <blist> <bibtext> Lundberg, I., Frost, J., & Petersen, O. P. (1988). Effects of an extensive program for stimulating phonological awareness in preschool children. Reading Research Quarterly, 23, 263 – 284.</bibtext> </blist> <blist> <bibtext> Lundberg, I., Larsman, P., & Strid, A. (2012). Development of phonological awareness during the preschool year: The influence of gender and socio‐economic status. Reading and Writing, 25, 305 – 320.</bibtext> </blist> <blist> <bibtext> Mann, V. A., & Liberman, I. Y. (1984). Phonological awareness and verbal short‐term memory. Journal of Learning Disabilities, 17 (10), 592 – 599.</bibtext> </blist> <blist> <bibtext> Marshall, C. R., & van der Lely, H. K. J. (2009). Effects of word position and stress on onset cluster production: Evidence from typical development, specific language impairment, and dyslexia. Language, 85 (1), 39 – 57.</bibtext> </blist> <blist> <bibtext> Michel, P., Baumstarck, K., Loundou, A., Ghattas, B., Auquier, P., & Boyer, L. (2018). Computerized adaptive testing with decision regression trees: An alternative to item response theory for quality of life measurement in multiple sclerosis. Patient Preference and Adherence, 12, 1043 – 1053.</bibtext> </blist> <blist> <bibtext> Molenaar, I. W., & Sijtsma, K. (1984). Internal consistency and reliability in Mokken's nonparametric item response model. Tijdschrift Voor Onderwijsresearch, 9 (5), 257 – 268.</bibtext> </blist> <blist> <bibtext> Molenaar, I. W., & Sijtsma, K. (1988). Mokken's approach to reliability estimation extended to multicategory items. Kwantitatieve Methoden: Nieuwsbrief Voor Toegepaste Statistiek En Operationele Research, 9 (28), 115 – 126.</bibtext> </blist> <blist> <bibtext> Morais, J., Cary, L., Alegria, J., & Bertelson, P. (1979). Does awareness of speech as a sequence of phones Arise spontaneously? Cognition, 7, 323 – 331. https://doi.org/10.1016/0010‐0277(79)90020‐9</bibtext> </blist> <blist> <bibtext> Newman, E. H., Tardif, T., Huang, J., & Shu, H. (2011). Phonemes matter: The role of phoneme‐level awareness in emergent Chinese readers. Journal of Experimental Child Psychology, 108 (2), 242 – 259.</bibtext> </blist> <blist> <bibtext> Orlando, M., & Thissen, D. (2000). Likelihood‐based item‐fit indices for dichotomous item response theory models. Applied Psychological Measurement, 24 (1), 50 – 64.</bibtext> </blist> <blist> <bibtext> Orlando, M., & Thissen, D. (2003). Further investigation of the performance of S‐X2: An item fit index for use with dichotomous item response theory models. Applied Psychological Measurement, 27 (4), 289 – 298.</bibtext> </blist> <blist> <bibtext> Papadopoulos, T. C., Kendeou, P., & Spanoudis, G. (2012). Investigating the factor structure and measurement invariance of phonological abilities in a sufficiently transparent language. Journal of Educational Psychology, 104 (2), 321 – 336.</bibtext> </blist> <blist> <bibtext> Petchko, K. (2009). Predicting reading achievement in a transparent orthography: Russian children learn to read. Doctoral Dissertation, Temple University, PA:(3344384).</bibtext> </blist> <blist> <bibtext> Peute, L., Scheeve, T., & Jaspers, M. (2020). Classification and regression tree and computer adaptive testing in cardiac rehabilitation: Instrument validation study. Journal of Medical Internet Research, 22 (1), e12509.</bibtext> </blist> <blist> <bibtext> Rakhlin, N., Cardoso‐Martins, C., & Grigorenko, E. L. (2014). Phonemic awareness is a more important predictor of orthographic processing than rapid serial naming: Evidence from Russian. Scientific Studies of Reading, 18 (6), 395 – 414.</bibtext> </blist> <blist> <bibtext> Rizopoulos, D. (2006). Ltm: An R package for latent variable modeling and item response theory analyses. Journal of Statistical Software, 17 (5), 1 – 25. https://doi.org/10.18637/jss.v017.i05</bibtext> </blist> <blist> <bibtext> Rosner, J., & Simon, D. P. (1971). The auditory analysis test: An initial report. Journal of Learning Disabilities, 4 (7), 384 – 392.</bibtext> </blist> <blist> <bibtext> Saiegh‐Haddad, E., Kogan, N., & Walters, J. (2010). Universal and language‐specific constraints on phonemic awareness: Evidence from Russian–Hebrew bilingual children. Reading and Writing, 23, 359 – 384.</bibtext> </blist> <blist> <bibtext> Savchuk, S. O., Arkhangelsky, T. A., Bonch‐Osmolovskaya, A. A., Donina, O. V., Kuznetsova, Y. N., Lyashevskaya, O. N., Orekhov, B. V., & Podryadchikova, M. V. (2024). Национальный корпус русского языка 2.0: новые возможности и перспективы развития. [Russian National Corpus 2.0: new opportunities and perspectives for development] Voprosy Yazikosnaniya, 2, 7–34.</bibtext> </blist> <blist> <bibtext> Schmidt, B. M., Breuer‐Küppers, P., Vahlhaus‐Aretz, D., Obergfell, A. L., & Schabmann, A. (2023). Prosodic sensitivity and phoneme awareness as predictors of reading fluency in German. Reading and Writing, 36 (1), 223 – 239.</bibtext> </blist> <blist> <bibtext> Selinker, L. (1972). Interlanguage. IRAL, 10, 209 – 231</bibtext> </blist> <blist> <bibtext> Silvén, M., Poskiparta, E., Niemi, P., & Voeten, M. (2007). Precursors of reading skill from infancy to first grade in Finnish: Continuity and change in a highly inflected language. Journal of Educational Psychology, 99 (3), 516 – 531.</bibtext> </blist> <blist> <bibtext> Steinberg, L., & Thissen, D. (2013). Item response theory. In J. S. Comer & P. C. Kendall (Eds.), The Oxford handbook of research strategies for clinical psychology (pp. 336 – 373). Oxford University Press.</bibtext> </blist> <blist> <bibtext> Stoel‐Gammon, C. (2010). The word complexity measure: Description and application to developmental phonology and disorders. Clinical Linguistics & Phonetics, 24 (4–5), 271 – 282.</bibtext> </blist> <blist> <bibtext> Therneau, T. (2022). Atkinson, B. _rpart: Recursive partitioning and regression Trees_. Rpackage Version, 4, 19.</bibtext> </blist> <blist> <bibtext> Treiman, R. (1991). Children's spelling errors on syllable‐initial consonant clusters. Journal of Educational Psychology, 83 (3), 346 – 360.</bibtext> </blist> <blist> <bibtext> van der Ark, L. (2010). Computation of the Molenaar Sijtsma statistic. In A. Fink, B. Lausen, W. Seidel, & A. Ultsch (Eds.), Advances in data analysis, data handling and business intelligence (pp. 775 – 784). Springer. https://doi.org/10.1007/978‐3‐642‐01044‐6_71</bibtext> </blist> <blist> <bibtext> Van der Schuit, M., Segers, E., Van Balkom, H., & Verhoeven, L. (2011). How cognitive factors affect language development in children with intellectual disabilities. Research in Developmental Disabilities, 32 (5), 1884 – 1894.</bibtext> </blist> <blist> <bibtext> van Viersen, S., de Bree, E. H., Kroesbergen, E. H., Slot, E. M., & de Jong, P. F. (2015). Risk and protective factors in gifted children with dyslexia. Annals of Dyslexia, 65, 178 – 198.</bibtext> </blist> <blist> <bibtext> Wagner, R. K., Zirps, F. A., Edwards, A. A., Wood, S. G., Joyner, R. E., Becker, B. J., ... Beal, B. (2020). The prevalence of dyslexia: A new approach to its estimation. Journal of Learning Disabilities, 53 (5), 354 – 365.</bibtext> </blist> <blist> <bibtext> Waring, R., Liow, S. R., Eadie, P., & Barbara, D. O. D. D. (2019). Speech development in preschool children: Evaluating the contribution of phonological short‐term and phonological working memory. Journal of Child Language, 46 (4), 632 – 652.</bibtext> </blist> <blist> <bibtext> Wells, C. S., & Wollack, J. A. (2003). An instructor's guide to understanding test reliability. Testing & Evaluation Services University of Wisconsin.</bibtext> </blist> <blist> <bibtext> Yu, C. H. (2001). An introduction to computing and interpreting Cronbach coefficient alpha in SAS. In Proceedings of 26th SAS user group international conference, Cary, NC (pp. 22 – 25). SAS Institute Inc.</bibtext> </blist> <blist> <bibtext> Zhukova, M., & Grigorenko, E. (2019). Developmental dyslexia in Russian. In L. Verhoeven, C. Perfetti, & K. Pugh (Eds.), Developmental dyslexia across languages and writing systems (pp. 133 – 151). Cambridge University Press.</bibtext> </blist> <blist> <bibtext> Zifcak, M. (1981). Phonological awareness and reading acquisition. Contemporary Educational Psychology, 6 (2), 117 – 126.</bibtext> </blist> <blist> <bibtext> Zipf, G. K. (1949). Human behavior and the principle of least effort (pp. xi (p. 573). Addison‐Wesley Press.</bibtext> </blist> </ref> <aug> <p>By Ilia V. Markov; Ksenia S. Kharitonova and Elena L. Grigorenko</p> <p>Reported by Author; Author; Author</p> <p></p> <p>Ilia V. Markov is a Doctoral Candidate in the Department of Psychology, University of Houston, Houston, TX, USA. email:</p> <p>Ksenia S. Kharitonova is a Graduate Student in the Department of Psychology, University of Houston, Houston, TX, USA. email:</p> <p>Elena L. Grigorenko (corresponding author) is a Hugh Roy and Lillie Cranz Cullen Distinguished Professor of Psychology in the Department of Psychology, University of Houston, Houston, TX, USA. email:</p> </aug> <nolink nlid="nl1" bibid="bib64" firstref="ref2"></nolink> <nolink nlid="nl2" bibid="bib20" firstref="ref3"></nolink> <nolink nlid="nl3" bibid="bib37" firstref="ref4"></nolink> <nolink nlid="nl4" bibid="bib52" firstref="ref5"></nolink> <nolink nlid="nl5" bibid="bib45" firstref="ref6"></nolink> <nolink nlid="nl6" bibid="bib36" firstref="ref7"></nolink> <nolink nlid="nl7" bibid="bib25" firstref="ref8"></nolink> <nolink nlid="nl8" bibid="bib83" firstref="ref9"></nolink> <nolink nlid="nl9" bibid="bib34" firstref="ref10"></nolink> <nolink nlid="nl10" bibid="bib10" firstref="ref11"></nolink> <nolink nlid="nl11" bibid="bib41" firstref="ref13"></nolink> <nolink nlid="nl12" bibid="bib44" firstref="ref14"></nolink> <nolink nlid="nl13" bibid="bib16" firstref="ref16"></nolink> <nolink nlid="nl14" bibid="bib80" firstref="ref19"></nolink> <nolink nlid="nl15" bibid="bib77" firstref="ref20"></nolink> <nolink nlid="nl16" bibid="bib30" firstref="ref22"></nolink> <nolink nlid="nl17" bibid="bib31" firstref="ref23"></nolink> <nolink nlid="nl18" bibid="bib18" firstref="ref24"></nolink> <nolink nlid="nl19" bibid="bib11" firstref="ref25"></nolink> <nolink nlid="nl20" bibid="bib28" firstref="ref26"></nolink> <nolink nlid="nl21" bibid="bib84" firstref="ref27"></nolink> <nolink nlid="nl22" bibid="bib27" firstref="ref28"></nolink> <nolink nlid="nl23" bibid="bib49" firstref="ref29"></nolink> <nolink nlid="nl24" bibid="bib19" firstref="ref30"></nolink> <nolink nlid="nl25" bibid="bib12" firstref="ref31"></nolink> <nolink nlid="nl26" bibid="bib33" firstref="ref36"></nolink> <nolink nlid="nl27" bibid="bib69" firstref="ref37"></nolink> <nolink nlid="nl28" bibid="bib58" firstref="ref38"></nolink> <nolink nlid="nl29" bibid="bib50" firstref="ref39"></nolink> <nolink nlid="nl30" bibid="bib75" firstref="ref40"></nolink> <nolink nlid="nl31" bibid="bib71" firstref="ref41"></nolink> <nolink nlid="nl32" bibid="bib43" firstref="ref42"></nolink> <nolink nlid="nl33" bibid="bib67" firstref="ref43"></nolink> <nolink nlid="nl34" bibid="bib62" firstref="ref45"></nolink> <nolink nlid="nl35" bibid="bib66" firstref="ref47"></nolink> <nolink nlid="nl36" bibid="bib32" firstref="ref49"></nolink> <nolink nlid="nl37" bibid="bib79" firstref="ref50"></nolink> <nolink nlid="nl38" bibid="bib47" firstref="ref51"></nolink> <nolink nlid="nl39" bibid="bib82" firstref="ref52"></nolink> <nolink nlid="nl40" bibid="bib55" firstref="ref53"></nolink> <nolink nlid="nl41" bibid="bib56" firstref="ref54"></nolink> <nolink nlid="nl42" bibid="bib76" firstref="ref55"></nolink> <nolink nlid="nl43" bibid="bib46" firstref="ref56"></nolink> <nolink nlid="nl44" bibid="bib72" firstref="ref57"></nolink> <nolink nlid="nl45" bibid="bib15" firstref="ref58"></nolink> <nolink nlid="nl46" bibid="bib74" firstref="ref60"></nolink> <nolink nlid="nl47" bibid="bib65" firstref="ref61"></nolink> <nolink nlid="nl48" bibid="bib13" firstref="ref62"></nolink> <nolink nlid="nl49" bibid="bib39" firstref="ref63"></nolink> <nolink nlid="nl50" bibid="bib73" firstref="ref64"></nolink> <nolink nlid="nl51" bibid="bib26" firstref="ref66"></nolink> <nolink nlid="nl52" bibid="bib68" firstref="ref67"></nolink> <nolink nlid="nl53" bibid="bib81" firstref="ref68"></nolink> <nolink nlid="nl54" bibid="bib48" firstref="ref70"></nolink> <nolink nlid="nl55" bibid="bib59" firstref="ref71"></nolink> <nolink nlid="nl56" bibid="bib60" firstref="ref72"></nolink> <nolink nlid="nl57" bibid="bib54" firstref="ref73"></nolink> <nolink nlid="nl58" bibid="bib63" firstref="ref74"></nolink> <nolink nlid="nl59" bibid="bib85" firstref="ref76"></nolink> <nolink nlid="nl60" bibid="bib35" firstref="ref78"></nolink> <nolink nlid="nl61" bibid="bib14" firstref="ref79"></nolink> <nolink nlid="nl62" bibid="bib38" firstref="ref80"></nolink> <nolink nlid="nl63" bibid="bib51" firstref="ref82"></nolink> <nolink nlid="nl64" bibid="bib40" firstref="ref84"></nolink> <nolink nlid="nl65" bibid="bib53" firstref="ref85"></nolink> <nolink nlid="nl66" bibid="bib61" firstref="ref88"></nolink> <nolink nlid="nl67" bibid="bib17" firstref="ref89"></nolink> <nolink nlid="nl68" bibid="bib42" firstref="ref90"></nolink> <nolink nlid="nl69" bibid="bib57" firstref="ref91"></nolink> <nolink nlid="nl70" bibid="bib21" firstref="ref92"></nolink> <nolink nlid="nl71" bibid="bib70" firstref="ref93"></nolink> <nolink nlid="nl72" bibid="bib29" firstref="ref95"></nolink> <nolink nlid="nl73" bibid="bib23" firstref="ref97"></nolink> <nolink nlid="nl74" bibid="bib24" firstref="ref98"></nolink> <nolink nlid="nl75" bibid="bib78" firstref="ref99"></nolink> <nolink nlid="nl76" bibid="bib22" firstref="ref100"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1458551
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Improving the Measures of Phonological Ability in the Russian Language: IRT and CART Modeling Application
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Ilia+V%2E+Markov%22">Ilia V. Markov</searchLink><br /><searchLink fieldCode="AR" term="%22Ksenia+S%2E+Kharitonova%22">Ksenia S. Kharitonova</searchLink><br /><searchLink fieldCode="AR" term="%22Elena+L%2E+Grigorenko%22">Elena L. Grigorenko</searchLink>
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Reading+Research+Quarterly%22"><i>Reading Research Quarterly</i></searchLink>. 2025 60(1).
– Name: Avail
  Label: Availability
  Group: Avail
  Data: Wiley. Available from: John Wiley & Sons, Inc. 111 River Street, Hoboken, NJ 07030. Tel: 800-835-6770; e-mail: cs-journals@wiley.com; Web site: https://www.wiley.com/en-us
– Name: PeerReviewed
  Label: Peer Reviewed
  Group: SrcInfo
  Data: Y
– Name: Pages
  Label: Page Count
  Group: Src
  Data: 18
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2025
– Name: SourceSuprt
  Label: Sponsoring Agency
  Group: SrcSuprt
  Data: National Institutes of Health (NIH) (DHHS)
– Name: NumberContract
  Label: Contract Number
  Group: NumCntrct
  Data: R01DC007665
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Foreign+Countries%22">Foreign Countries</searchLink><br /><searchLink fieldCode="DE" term="%22Phonological+Awareness%22">Phonological Awareness</searchLink><br /><searchLink fieldCode="DE" term="%22Phonology%22">Phonology</searchLink><br /><searchLink fieldCode="DE" term="%22Language+Acquisition%22">Language Acquisition</searchLink><br /><searchLink fieldCode="DE" term="%22Literacy%22">Literacy</searchLink><br /><searchLink fieldCode="DE" term="%22Short+Term+Memory%22">Short Term Memory</searchLink><br /><searchLink fieldCode="DE" term="%22Item+Response+Theory%22">Item Response Theory</searchLink><br /><searchLink fieldCode="DE" term="%22Predictor+Variables%22">Predictor Variables</searchLink><br /><searchLink fieldCode="DE" term="%22Test+Construction%22">Test Construction</searchLink><br /><searchLink fieldCode="DE" term="%22Drills+%28Practice%29%22">Drills (Practice)</searchLink><br /><searchLink fieldCode="DE" term="%22Auditory+Tests%22">Auditory Tests</searchLink>
– Name: Subject
  Label: Geographic Terms
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Russia%22">Russia</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1002/rrq.604
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0034-0553<br />1936-2722
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Phonological awareness and phonological working memory are essential for successful language acquisition and development of literacy. Although this essence is language-universal, its degree varies for different languages, depending, in part, on language transparency. The current study analyzes the adapted versions of the pseudoword repetition test (assessing phonological working memory) and Rosner's Auditory Segmentation test (assessing phonological awareness) in a typically developing Russian native sample of children (n = 502). As a preparatory step to item analysis, we investigated the effects of grade and gender on performance using a mixed effects model. The initial item analysis was carried out using model comparison within the Item Response Theory model framework and threshold/slope analysis. The majority of the items in both assessments did not differentiate between students with different levels of phonological ability. Further item selection using regression tree models led to the formation of predictive and non-predictive item subsets for each assessment. After comparing the item subsets on various linguistic metrics, the differences were found in number of syllables and subsyllabic complexity for the pseudoword repetition test and elision segment position for the auditory segmentation test. The findings inform test development strategies in the cases of extremely low difficulty/discrimination of the items and outline a blueprint of pseudoword repetition and auditory segmentation test's adaptation for potentially detecting higher levels of phonological ability in transparent languages such as Russian.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2025
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1458551
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1458551
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1002/rrq.604
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 18
    Subjects:
      – SubjectFull: Foreign Countries
        Type: general
      – SubjectFull: Phonological Awareness
        Type: general
      – SubjectFull: Phonology
        Type: general
      – SubjectFull: Language Acquisition
        Type: general
      – SubjectFull: Literacy
        Type: general
      – SubjectFull: Short Term Memory
        Type: general
      – SubjectFull: Item Response Theory
        Type: general
      – SubjectFull: Predictor Variables
        Type: general
      – SubjectFull: Test Construction
        Type: general
      – SubjectFull: Drills (Practice)
        Type: general
      – SubjectFull: Auditory Tests
        Type: general
      – SubjectFull: Russia
        Type: general
    Titles:
      – TitleFull: Improving the Measures of Phonological Ability in the Russian Language: IRT and CART Modeling Application
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Ilia V. Markov
      – PersonEntity:
          Name:
            NameFull: Ksenia S. Kharitonova
      – PersonEntity:
          Name:
            NameFull: Elena L. Grigorenko
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2025
          Identifiers:
            – Type: issn-print
              Value: 0034-0553
            – Type: issn-electronic
              Value: 1936-2722
          Numbering:
            – Type: volume
              Value: 60
            – Type: issue
              Value: 1
          Titles:
            – TitleFull: Reading Research Quarterly
              Type: main
ResultId 1