View in EDS HTML Full Text PDF Full Text

Getting KnERDI with Language: Examining Teachers' Knowledge for Enhancing Reading Development in Code-Based and Meaning-Based Domains

Saved in:
Bibliographic Details
Title: Getting KnERDI with Language: Examining Teachers' Knowledge for Enhancing Reading Development in Code-Based and Meaning-Based Domains
Language: English
Authors: Davis, Dennis S., Samuelson, Courtney, Grifenhagen, Jill, DeIaco, Robyn, Relyea, Jackie (ORCID 0000-0002-7560-7136)
Source: Reading Research Quarterly. Jul-Sep 2022 57(3):781-804.
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: 24
Publication Date: 2022
Document Type: Journal Articles
Reports - Research
Education Level: Elementary Education
Descriptors: Reading Instruction, Elementary Education, Test Reliability, Test Validity, Teacher Effectiveness, Teacher Evaluation, Teacher Characteristics, Elementary School Teachers, Knowledge Level
DOI: 10.1002/rrq.445
ISSN: 0034-0553
Abstract: Much of the early research on teachers' knowledge for reading instruction used instruments that primarily emphasized code-based aspects of reading, unintentionally signaling that meaning-focused knowledge is not essential for teaching foundational reading to elementary-age children. In this study, we developed the Knowledge for Enhancing Reading Development Inventory (KnERDI; pronounced 'nerdy'), an instrument that measures teachers' knowledge in two domains: alphabetic code/word reading and meaning/connected text processes. Using this instrument, we sought to determine if teachers' knowledge of the two proposed domains was highly associated and if knowledge was related to level of education and amount of teaching experience. We report on the reliability and validity of the KnERDI and describe patterns in educators' performance on this instrument. We found that the KnERDI measured both domains with acceptable reliability in multiple samples of educators enrolled in an online professional development course. Educators scored higher on the meaning/connected text process subscale than on the alphabetic code/word reading subscale. The two subscales were strongly correlated, indicating that teachers' knowledge as measured on the KnERDI is a unidimensional construct. Advanced degree completion was not consistently related to educators' scores on the KnERDI, but there was a positive association between educators' performance on the instrument and their years of experience teaching early reading, particularly with respect to the code-based domain. We discuss potential uses of the KnERDI in future research and practice, positive trends in educators' performance on the instrument, and challenges in measuring educators' knowledge for supporting higher-level language and comprehension processes.
Abstractor: As Provided
Entry Date: 2022
Accession Number: EJ1340023
Database: ERIC
Full text is not displayed to guests.
FullText Links:
  – Type: pdflink
    Url: https://content.ebscohost.com/cds/retrieve?content=AQICAHj0k_4E0hTGH8RJwT4gCJyBsGNe_WN95AvKlDbXJGqwxwG32vqC_O4wPoA_cFoml7-aAAAA4zCB4AYJKoZIhvcNAQcGoIHSMIHPAgEAMIHJBgkqhkiG9w0BBwEwHgYJYIZIAWUDBAEuMBEEDPMVvIw_AYI8ftRSIwIBEICBm38AXlOp13pPvwF5ChQUmHlW-P8f3VVXOWBQWj3Mn_a03S41v7XNSKkyaVouyp9BLPqDYNPs4Pjb7i9LctqKqUBFWjTZtwAUCtKeljw6Dd7GPSNWexQhbuoSidxu01oLY6QHYUdkeQHtykb0aL2LrMx3R_OT1EbCiybY1IEkhuVmM5IEXpsl9EFJpLzD-OJCixYZzrXy3koM27cJ
Text:
  Availability: 1
  Value: <anid>AN0157693737;[nrnu]01jul.22;2025Apr04.08:23;v2.2.500</anid> <title id="AN0157693737-1">Getting KnERDI with Language: Examining Teachers' Knowledge for Enhancing Reading Development in Code‐Based and Meaning‐Based Domains </title> <p>Much of the early research on teachers' knowledge for reading instruction used instruments that primarily emphasized code‐based aspects of reading, unintentionally signaling that meaning‐focused knowledge is not essential for teaching foundational reading to elementary‐age children. In this study, we developed the Knowledge for Enhancing Reading Development Inventory (KnERDI; pronounced 'nerdy'), an instrument that measures teachers' knowledge in two domains: alphabetic code/word reading and meaning/connected text processes. Using this instrument, we sought to determine if teachers' knowledge of the two proposed domains was highly associated and if knowledge was related to level of education and amount of teaching experience. We report on the reliability and validity of the KnERDI and describe patterns in educators' performance on this instrument. We found that the KnERDI measured both domains with acceptable reliability in multiple samples of educators enrolled in an online professional development course. Educators scored higher on the meaning/connected text process subscale than on the alphabetic code/word reading subscale. The two subscales were strongly correlated, indicating that teachers' knowledge as measured on the KnERDI is a unidimensional construct. Advanced degree completion was not consistently related to educators' scores on the KnERDI, but there was a positive association between educators' performance on the instrument and their years of experience teaching early reading, particularly with respect to the code‐based domain. We discuss potential uses of the KnERDI in future research and practice, positive trends in educators' performance on the instrument, and challenges in measuring educators' knowledge for supporting higher‐level language and comprehension processes.</p> <p>Keywords: Comprehension; Decoding; Phonics; phonemic awareness; phonological awareness; Teacher education; professional development; Theoretical perspectives; 2‐Childhood</p> <p>Specialized knowledge of language concepts and literacy skills is essential for elementary teachers who are instructing young children in early reading. There is evidence that educators' knowledge of early reading development is associated with instructional practices (Jordan & Bratsch‐Hines, 2020; Piasta, Soto Ramirez, et al., 2020) as well as student reading outcomes (Carlisle et al., 2011; McCutchen et al., 2002; Piasta, Park, et al., 2020). Despite the importance of teachers' knowledge, developing tools that adequately measure this knowledge has posed a significant challenge, and there are unanswered questions about the structure of this knowledge and the components that make it up. While it is unlikely that any single measure would be able to thoroughly capture what teachers could know about the complex and multidimensional construct of reading (Compton‐Lilly et al., 2020), high‐quality measures can be useful tools for researchers and practitioners. Such tools can be helpful to inform practice in teacher preparation programs, to design and measure the effectiveness of professional development opportunities, and to explore the relations between teacher knowledge, instructional practice, and student achievement.</p> <p>In this study, we take inspiration from researchers who have proposed models of teacher knowledge that include but extend beyond code‐based aspects of reading (e.g., Carlisle et al., 2011; Phelps & Schilling, 2004). Building from the assumption that young children who are learning word‐level reading skills can and should learn how to read for meaning at the same time (Duke et al., 2021; Pearson et al., 2020), we propose that educators' specialized knowledge for teaching early reading includes (but is not limited to) two broad domains, conceptualized in this study as <emph>Alphabetic Code and Word Reading</emph> and <emph>Meaning and Connected Text Processes</emph>. Both domains include knowledge of language concepts important for teaching young children to read in English, including phonemes and letter–sound correspondence in the first domain, and morphology, semantics, syntax, and organizational structure of written text in the second domain. We evaluated the adequacy of this model by developing and testing an instrument that instantiates the proposed knowledge structure, called the Knowledge for Enhancing Reading Development Inventory (KnERDI; pronounced <emph>nerdy</emph>). In multiple samples of educators, we utilized the instrument to examine patterns in knowledge in both domains across different levels of teaching experience and education.</p> <hd id="AN0157693737-2">Measuring Teachers' Knowledge for Early Reading Instruction</hd> <p>The conceptualization of teachers' knowledge used in this study draws on a long tradition of research, conducted primarily in the field of mathematics education, which has established that educators need a strong understanding of the disciplines they teach (Hill et al., 2008). Scholars have proposed several categories of knowledge essential for teaching practice. For example, teachers need specialized content knowledge (SCK), a special type of subject matter knowledge that has been "unpacked" or "decompressed" in ways that enable teachers to organize effective learning opportunities for children (Ball et al., 2008, p. 400). SCK is distinct from common content knowledge that one might find outside of the teaching community, that is, among typical adults who know how to solve a math problem but do not have or need the specialized vocabulary for explaining why a particular strategy works for finding the solution (Hill et al., 2005; Schoen et al., 2017). SCK is also distinct from Shulman's (1986) construct of pedagogical content knowledge, which is often defined as knowledge of students and teaching practices essential for the content area being taught (Ball et al., 2008; Campbell et al., 2014). In this study, we refer to this range of categories of knowledge collectively as specialized knowledge for teaching, inclusive of SCK, and pedagogical content knowledge. When studying educators' specialized knowledge in any domain of teaching, it is important to consider their content expertise and their knowledge of instructional practices for making the content understandable to others.</p> <p>The specialized nature of teacher knowledge is particularly evident in early reading instruction. The knowledge needed to teach elementary children to read is distinct from the knowledge literate adults employ to read texts on their own (Phelps, 2009). For example, phonemic awareness instruction requires teachers to demonstrate for students how to count, blend, segment, and manipulate the individual speech sounds in words. Teachers must help students understand that letters represent sounds in words and that often the number of letters and phonemes in words may not match. To teach students how to properly decode, teachers must guide students to use their developing knowledge of orthography—the way words are spelled—and map speech sounds to spelling patterns to read words (Ehri, 2017).</p> <p>Simply teaching students how to read individual words is not enough. Teachers must also help students develop fluency, vocabulary, and comprehension skills as they guide students in reading for meaning and to build new knowledge as they read (Duke & Block, 2012). Just as teachers base their phonics and word reading instruction on a working model of how the language system is structured to encode sounds, they base their comprehension instruction on their understanding of the components of language and processes involved in meaning making (McElhone et al., 2017).</p> <p>The complexity of knowledge required to teach foundational reading skills poses a challenge for those wishing to design instruments to measure such knowledge. Questions remain regarding what exactly should be measured and how. In a discussion of complications that arise when developing teacher knowledge assessments, Reutzel and colleagues (2011) identified numerous possible domains of knowledge among elementary teachers, such as linguistic knowledge, knowledge of comprehension instruction, familiarity with children's literature, knowledge of how to support readers when they struggle, and teaching English learners. Other scholars have argued that research and discussions of early educators' knowledge should include a focus on knowledge of response to intervention (RTI) procedures (Al Otaiba et al., 2019; Spear‐Swerling & Cheesman, 2012), children's literature (Cunningham et al., 2004), dyslexia and reading difficulties (Washburn et al., 2011a; 2011b), and socially just teaching practices (Mosley Wetzel et al., 2020).</p> <p>Another challenge is determining how to effectively measure educators' specialized knowledge in practical and efficient ways. Some intensive measurement strategies, like interviews (e.g., McElhone et al., 2017) and video reflection tasks (e.g., Kucan et al., 2009) provide a more detailed portrait of educators' knowledge, especially their knowledge of complex comprehension practices and discussion‐based pedagogies. However, these methods are difficult and time‐intensive to administer in many professional development contexts. We have chosen to focus our efforts in this study on an easy‐to‐administer knowledge inventory measure. Many researchers have used knowledge inventories with multiple‐choice or short‐answer formats (Brady et al., 2009; Carlisle et al., 2011; Folsom et al., 2017). Valid and reliable measures that use this format should not be a substitute for more holistic, in‐depth measures of teachers' knowledge, but they can provide an important data point to add to the full picture of what teachers know about reading development.</p> <hd id="AN0157693737-3">Findings from Inventories of Teachers' Knowledge for Early Reading Instruction</hd> <p>Many of the knowledge inventories available in the reading research literature owe some part of their lineage to the Informal Survey of Linguistic Knowledge developed by Moats (1994). In that early study, the researcher measured the knowledge of 89 teachers on concepts such as letter–sound correspondences, phoneme and morpheme awareness, and phonics. The participants included classroom teachers, reading specialists, special education teachers, and teacher assistants with an average of 5 years of experience. Low scores on the survey, especially in the areas of phonics and morpheme awareness, led Moats to conclude that teachers' understanding of spoken and written language structure was insufficient for effective classroom instruction for beginning readers.</p> <p>Many subsequent knowledge instruments have primarily focused on teachers' knowledge of constrained skills related to letters, sounds, and word reading. Binks‐Cantrell and colleagues (2012) developed a survey of basic language constructs to measure teacher educators' and pre‐service teachers' knowledge and skills in the areas of phonological awareness, phonemic awareness, alphabetic principle/phonics, and morphology. Washburn, Joshi, and Binks‐Cantrell (2011a; 2011b) measured pre‐service and in‐service teachers' knowledge with this instrument and found that while most participants demonstrated implicit knowledge (such as high scores in syllable counting tasks), they lacked knowledge of principles and terminology related to systematic phonics instruction. More recently, Arrow and colleagues (2019) used the same instrument with lower elementary teachers in New Zealand and found similar patterns in teachers' performance.</p> <p>Bos and colleagues (2001) also developed and used a knowledge instrument focused on code‐based skills such as phonology and phonics to measure pre‐service and in‐service elementary and special education teachers' knowledge. They found that pre‐service teachers on average were able to answer 53% of questions correctly, with in‐service teachers averaging 60% correct. A study by Cheesman and colleagues (2009) measured 223 first‐year elementary teachers' knowledge of phonological awareness and phonics and concluded that a large proportion of teachers demonstrated limited understanding of phonics instruction and how to reliably count phonemes in written words with complex spellings.</p> <p>The trends reported in previous studies have been used to frame a bleak portrayal of literacy teachers, who are often unfairly depicted as lacking accurate knowledge of reading. A common conclusion offered by many of these studies is that teachers are licensed without learning enough about the language systems involved in teaching early reading. In our view, this is a decidedly deficit‐oriented conclusion, reflecting a tendency in the teacher knowledge literature to position teachers as struggling to understand their content. There are many grounds on which to challenge this deficit positioning of teachers. As it relates to the present study, a major problem with this conclusion is that many studies that make this argument only focus on one particular aspect of beginning reading instruction: code‐based word reading skills.</p> <p>There is scientific consensus that multiple domains of language‐related competencies are required for children's reading development, among them: effortless word reading (made possible by mastery of the alphabetic code) and language comprehension (Castles et al., 2018; Foorman et al., 2018; Gough & Tunmer, 1986). Teachers need robust knowledge of these two domains to support their students' learning. These are certainly not the only competencies required for reading development and teaching, but they are the two at the center of the present study. Some inventories that measure teacher knowledge for early reading instruction have included items related to code‐based and meaning‐based concepts. One strategy in this literature has been to include both domains in the measurement tool without assessing or reporting separate subscales, in essence, treating teachers' knowledge as a unidimensional construct. This was the case in a study by Folsom and colleagues (2017), who analyzed knowledge growth during a statewide professional development program in early literacy instruction. Their instrument consisted of items related to comprehension, fluency, writing, vocabulary, spelling, phonological and phonemic awareness, and phonics. Using a single combined measure, they found that teacher knowledge of early literacy increased as well as quality of instruction and student engagement.</p> <p>Other researchers have measured and reported on the domains as separate dimensions of knowledge. This research suggests that teachers' knowledge of word‐level and meaning‐based reading content is likely associated, but the strength of the association varies across studies. For example, Brady and colleagues (2009) examined the efficacy of a year‐long professional development (PD) for building first‐grade teachers' literacy knowledge. They used an instrument that included separate measurements of phonemic awareness, code concepts, fluency, and oral language (including reading comprehension). They found improvements in teachers' knowledge after the PD experience. They reported correlations among the knowledge measures, showing weak associations of knowledge of code concepts with fluency (<emph>r</emph> = .20) and oral language concepts (<emph>r</emph> = .05) prior to PD, and only slightly higher after PD (<emph>r</emph>s = .32 and.31, respectively).</p> <p>In another study in which researchers measured multiple dimensions of knowledge, Garet and colleagues (2008) assessed educators' knowledge in the five areas of reading outlined by the National Reading Panel (NICHD, 2000): phonemic awareness, phonics, and fluency (what they classified as word‐level content); and vocabulary and comprehension (meaning‐level content). They found that teachers who participated in the professional development institute outperformed a control group on measures of word‐level knowledge, and the impact of the professional development on educators' meaning‐level knowledge was positive although not statistically significant. Teachers' scores on the word‐level and meaning subscores were strongly associated (correlation = .83) in a model that adjusted for scale reliability, but only weakly correlated according to the unadjusted Pearson's <emph>r</emph> of.38.</p> <p>Additionally, Spear‐Swerling and Cheesman (2012) developed an instrument that measured performance on three subscales: phonemic awareness and phonics; fluency, vocabulary, and comprehension; and RTI and assessment. The questionnaire was administered to 142 practicing elementary and special education teachers, and results showed significant differences in mean scores among the three scales. Participants scored highest on the fluency/vocabulary/comprehension (flu/voc/comp) subscale and lower on the phonemic awareness/phonics (PA/ph) and RTI/assessment subscales. Regarding associations among the subscales, the researchers found a moderate correlation between flu/voc/comp and PA/ph of.58.</p> <p>Researchers who have directly evaluated the dimensionality of teachers' knowledge with respect to code‐based and meaning‐based domains have reached conflicting conclusions about the appropriateness of measuring these domains separately. For example, Phelps and Schilling (2004) found that teachers' knowledge for teaching word analysis and text comprehension operated as distinct dimensions. Their instrument included items in two content domains: comprehension (morphology, vocabulary, comprehension strategies and questions, genre, and fluency) and word analysis (phonemic awareness, letter–sound relationships, word frequency, and decoding). Similarly, despite the high correlation of teachers' word‐level and meaning‐level knowledge described above, Garet et al. (2008) found support for a two‐factor model over a one‐factor model using confirmatory factor analysis (CFA).</p> <p>In contrast, findings from other studies have suggested that teachers' knowledge of reading across these domains should be treated as a single construct. Carlisle and colleagues (2011) concluded that their Teachers' Knowledge of Reading and Reading Practices survey, which measured linguistic knowledge of word reading and reading comprehension and also included items related to academic content and pedagogy, formed a unidimensional scale. Likewise, Folsom and colleagues (2017) combined their items, covering content and pedagogy in the multiple areas of reading, in a single‐factor model.</p> <p>Thus, while there is robust evidence that reading itself is a multidimensional process (Hindman et al. 2020; Kendeou et al., 2016), there is not consensus about the dimensions of knowledge that make up teachers' understanding of early reading instruction and how strongly associated teachers' knowledge of word‐level code‐based skills might be with their knowledge of higher‐level language and comprehension concepts. The variation in magnitude of knowledge association reported in the literature raises questions about the linkages between these knowledge domains among teachers. Given the negative portrayal of teacher knowledge common in the existing literature, which, as reviewed above, has come largely from studies that emphasize teachers' knowledge of phonemes and phonics generalizations, it is important to know if teachers perform similarly when assessed in other domains related to meaning making. If their knowledge in both areas is highly correlated, it may be the case that teachers' specialized knowledge for early reading development reflects one underlying construct. However, if the relation between the two is weak, or if it is common for teachers to know a lot more about one area than the other, it might be more appropriate to conceptualize this knowledge along multiple dimensions when developing strategies and materials to support teachers' knowledge growth.</p> <p>We identify two challenges in the existing literature that might contribute to inconsistent findings regarding the nature and structure of teachers' knowledge. First, researchers have differed in the approaches used to examine associations between code‐based and meaning‐based knowledge domains. For instance, Spear‐Swerling and Cheesman (2012) reported Pearson correlation coefficients between raw subscores on their knowledge measure, while Garet et al. 2008) reported both raw and reliability‐corrected correlations. It may be the case that low subscale reliability has led to attenuated estimates of knowledge association in some studies.</p> <p>A second challenge is that researchers who have constructed meaning‐focused subscales have differed in the competencies they have chosen to include. For instance, Garet et al. (2008) defined their scales according to the National Reading Panel categories and included items related to vocabulary and reading comprehension, while Carlisle et al. (2011) included a more granular set of competencies that included morphology, text analysis, and fluency. These differences in scale construction reflect the lack of consensus in the field about how to best define comprehension. Many of the existing knowledge inventories treat reading comprehension as a self‐evident concept, listing it as a facet of what teachers should know without clarifying a definition or what teachers should know about it.</p> <hd id="AN0157693737-4">Associations of Education Level and Years of Experience with Knowledge</hd> <p>Teacher knowledge instruments have been used to measure the association between knowledge and practice, to evaluate and improve teacher preparation programs and professional development efforts, and to improve student outcomes. Underpinning these research purposes is the hypothesis that teacher preparation, professional development or additional training, and years of experience matter for teacher knowledge. Encouragingly, in a study of teachers in early childhood settings, Piasta, Soto Ramirez, and colleagues (2020) recently found educators' knowledge of early reading was linearly associated with quality of literacy instruction and did not reach a plateau. Thus, even educators with deep literacy knowledge may still be able to grow and expand this knowledge to improve their classroom practices. While these authors' findings apply only to teachers in prekindergarten classrooms, there are significant overlaps in the knowledge teachers use in early childhood settings and elementary classrooms.</p> <p>A few studies have found that amount of teaching experience has a significant relation to educator knowledge. In Bos and colleagues' 2001 study, educators with more than 11 years of experience outperformed in‐service teachers with 1 to 5 years of experience on an assessment that primarily measured code‐based aspects of language. Spear‐Swerling and Cheesman (2012) also found that teachers with more years of experience significantly outperformed less experienced teachers in the domains of phonics/phonemic awareness knowledge and RTI and assessment practices, but not in the domains of fluency, vocabulary, and comprehension. In a study of kindergarten and first‐grade teachers working in rural schools, Jordan, Bratsch‐Hines, and Vernon‐Feagans (2018) found that teaching experience was correlated with content and pedagogical knowledge.</p> <p>Studies of the association between education level and performance on knowledge inventories have produced varied findings. In an analysis of the teacher knowledge measure used in the Carlisle et al. (2011) study, Kelcey (2011) reported that teachers who scored higher on the instrument tended to be those with a master's degree in literacy or certification in reading. These characteristics were associated with standardized mean difference effect sizes of 0.51 and 0.80, respectively.</p> <p>Other studies have found that level of education and experience may not have a significant relation to educator knowledge. In recent work, McMahan and colleagues (2019) examined variations in educators' knowledge in five domains: phonemic awareness, phonological sensitivity, decoding, encoding, and morphology. The authors found no statistically significant differences in performance between teachers with a bachelor's degree and those with a master's degree, and years of experience did not predict higher performance on their knowledge instrument. Pittman et al. (2020) found that teaching experience was associated with better performance on only one subscore in their measure (syllable counting), but not on other items related to morphology or phonology.</p> <p>Overall, the existing studies offer a mixed picture of the association between knowledge and level of teacher preparation and years of classroom experience. One limitation in the literature is that many researchers have drawn their samples from a single school district or locale, reducing the generalizability of their findings. Another limitation is that with only a few exceptions, much of the research that reports differences in knowledge across education and experience levels has primarily focused on teachers' knowledge of phonology and code‐based concepts. Future research addressing these limitations can help clarify the field's understanding of how educators' specialized knowledge might be developed over time or through specialized training.</p> <hd id="AN0157693737-5">Research Questions</hd> <p>Across the robust existing research, three important trends inform the present study. First, there is more emphasis in the literature on educators' knowledge of phonology and the alphabetic code for word reading compared to higher‐level language structures and processes involved in reading for meaning. As a result of this trend, there has been little examination of the association between educator's knowledge of these two broad domains of early literacy, and there are conflicting findings in the literature regarding the appropriateness of measuring teacher knowledge in these domains as a unidimensional or multidimensional construct. Finally, there have been mixed findings about the association of advanced preparation and years of experience with teacher knowledge, creating a need for further examination using larger, geographically diverse samples of educators.</p> <p>In this study, we examined educators' knowledge as evidenced on the KnERDI, which includes two subscales, one related to alphabetic code and word reading, and the other related to higher‐level language structures and processes for understanding connected texts. We were guided by three main research questions:</p> <p></p> <ulist> <item> Can educators' knowledge for teaching early reading be reliably measured with an instrument that contains two subscales, corresponding to two major competency domains involved in children's reading development?</item> <p></p> <item> To what extent is educators' knowledge of the two proposed subscales associated or disassociated?</item> <p></p> <item> What is the relation between educators' knowledge for teaching early reading and their level of education and years of teaching experience?</item> </ulist> <hd id="AN0157693737-6">Proposed Model of Language‐Related Components for Teaching Early Reading</hd> <p>We designed the KnERDI to include items in eight important facets of reading development that make up two broad domains (see Table 1): (a) decoding/phonics and phonological/phonemic awareness (domain of alphabetic code and word reading); and (b) reading fluency, morphological awareness, vocabulary/semantics, syntactic awareness, text structure awareness, and mental model building (domain of meaning and connected text processes). These facets were selected in alignment with recent scholarship that has clarified some of the essential code‐ and meaning‐based competencies involved in reading (e.g. Kendeou et al., 2016; Language and Reading Research Consortium et al., 2019). We find it important to note that, by design, the KnERDI is only intended to measure a specific portion of the knowledge teachers might use to inform their early reading instruction. This instrument emphasizes linguistic structures and processes of comprehension, but these are not the only facets of teacher knowledge that are important for teaching children to read, as indicated in the literature review above. The KnERDI should not in any way be construed as a comprehensive model of reading teacher knowledge.</p> <p>1 TABLEExample Items for Each Facet in the Two Subscales</p> <p> <ephtml> <table><thead><tr><th align="left">Alphabetic Code and Word Reading</th><th align="left">Meaning and Connected Text Processes</th></tr></thead><tbody><tr><td align="left">Decoding/phonics (36 items tested, 22 remained)Which of the following prompts should a teacher use to help a first‐grade reader decode an unfamiliar word?<list list-type="Bullet"><list-item><p>Use the syntactic cueing system to predict the word based on sentence structure (6.8%)</p></list-item><list-item><p>Tell the child to skip the word and read the rest of the sentence (3.1%)</p></list-item><list-item><p>Say and blend the sounds of spelling patterns in the word (88.0%)*</p></list-item><list-item><p>Give a clue about the meaning of the word (1.6%)</p></list-item></list>Phonological/phonemic awareness (9 items tested, 6 remained)Students need to develop strong phonemic awareness because it will help them be prepared to:<list list-type="Bullet"><list-item><p>Break spoken words into syllables (17.2%)</p></list-item><list-item><p>Learn to recognize letters in the alphabet (1.1%)</p></list-item><list-item><p>Read irregularly spelled words (5.4%)</p></list-item><list-item><p>Learn letter–sound relationships (76.1%)*</p></list-item></list></td><td align="left">Fluency (8 items tested, 5 remained)Which aspects of fluency are assessed with a WCPM (words correct per minute) measure?<list list-type="Bullet"><list-item><p>rate and accuracy (89.4%)*</p></list-item><list-item><p>rate and expression (2.1%)</p></list-item><list-item><p>accuracy and phrasing (3.1%)</p></list-item><list-item><p>speed and understanding (5.0%)</p></list-item></list>Morphology (11 items tested, 9 remained)Which of these words has a suffix that changes its part of speech?<list list-type="Bullet"><list-item><p>Cautioned (24.8%)</p></list-item><list-item><p>Expects (7.5%)</p></list-item><list-item><p>Writer (57.1%)*</p></list-item><list-item><p>Smarter (9.8%)</p></list-item></list>Syntax (4 items tested, 4 remained)A teacher wants to help a student who is having trouble understanding complex sentences in the book he is reading. Which is the best strategy for the teacher to use?<list list-type="Bullet"><list-item><p>Model how to pay attention to connective words like prepositions and conjunctions (60.1%)*</p></list-item><list-item><p>Teach new academic content words that the child might encounter in a complex sentence (15.1%)</p></list-item><list-item><p>Practice echo‐reading examples of complex sentences aloud (18.7%)</p></list-item><list-item><p>Provide extra practice with independent reading materials (4.9%)</p></list-item></list>Text structure (5 items tested, 5 remained)Why is it helpful to explicitly teach expository text structure to children?<list list-type="Bullet"><list-item><p>To help them identify important text features (10.9%)</p></list-item><list-item><p>To help them use the organization of a text to improve their understanding (75.3%)*</p></list-item><list-item><p>To help them correctly identify which structures are found in a text (8.6%)</p></list-item><list-item><p>To help them find the most important section of a text (4.3%)</p></list-item></list>Vocabulary/semantics (8 items tested, 5 remained)Which of these statements is true about explicit teaching of new vocabulary words?<list list-type="Bullet"><list-item><p>It is ineffective; children need to develop knowledge of too many words for explicit instruction to make a difference in their performance (3.4%)</p></list-item><list-item><p>It is ineffective; children are not able to learn deep knowledge about new words from explicit instruction in most situations (6.1%)</p></list-item><list-item><p>It is necessary but not sufficient; children need explicit teaching of specific words along with opportunities for incidental word learning (54.8%)*</p></list-item><list-item><p>It is necessary and sufficient; if implemented effectively, it can help children who are delayed in vocabulary learn enough words to catch up with their peers (34.9%)</p></list-item></list>Mental modeling/comprehension (12 items tested, 10 remained)Which of these competencies is usually most important for a reader to successfully comprehend a text?<list list-type="Bullet"><list-item><p>Knowledge of vocabulary and language used in the text (74.9%)*</p></list-item><list-item><p>Strategic awareness (11.1%)</p></list-item><list-item><p>Personal connection to the text (11.9%)</p></list-item><list-item><p>The difficulty of the questions asked by the teacher (1.9%)</p></list-item></list></td></tr></tbody></table> </ephtml> </p> <p>1 Note</p> <p>2 The instrument included a combination of original items and items sourced or modified from published studies, including Binks‐Cantrell et al., 2012; Brady et al., 2009; Carlisle et al., 2011; Folsom et al., 2017; Moats, 1994; Moats & Foorman, 2003; Piasta et al., 2009; Spear‐Swerling & Zibulsky, 2014.</p> <p>In the first domain, decoding/phonics refers to knowledge of the correspondence of letters and sounds in the English alphabetic writing system and the use of this knowledge to read words, both in context and out of context (Ehri, 2017). This facet also includes knowledge of word learning processes and the importance of developing automaticity, or effortless word reading, through repeated opportunities to encounter words in print and to analyze their letter–sound correspondences (Davis et al., 2021). Phonological/phonemic awareness, an important prerequisite skill for efficient word reading, refers to a reader's ability to identify and manipulate the sounds in language.</p> <p>In the second domain, reading fluency is the ability to read a text with appropriate speed, accuracy, and expression. It relies on the development of effortless word reading and language knowledge, and enables reading for understanding (Foorman et al., 2016). Morphology refers to knowledge of how words are made up of units of meaning, called morphemes, including root words, prefixes, and suffixes. Awareness of morphemic structure helps children decode complex words and to infer the meaning of unfamiliar academic words (Goodwin & Ahn, 2013).</p> <p>Syntactic knowledge refers to knowledge of how sentences can be structured in a language. As people read, they use their expectations for sentence structure, largely derived from their spoken language development, to help them assemble words into meaningful ideas (Castles et al., 2018; Kendeou et al., 2016). This knowledge includes awareness of how connective words signal relations among ideas within and across sentences to support text‐based inferences (Mesmer and Rose‐McCully, 2018). Text structure is included in our model of teacher knowledge because readers can also use their sensitivity to the organizational patterns in written text to support their understanding of an author's message (Roehling et al., 2017). Text structure awareness is particularly important for reading expository texts, which often include multiple organizational structures to connect ideas, including compare–contrast, cause and effect relationships, chronology, and description (Williams et al., 2014). Along with syntax and text structure, readers draw on their semantic knowledge of words to understand the meaning of a text. A large body of research has established positive associations between vocabulary and reading comprehension (NICHD, 2000; Quinn et al., 2015).</p> <p>The final facet in the second domain foregrounds the crucial role of mental model building in comprehension. Text comprehension is a broad concept that can mean many things (Davis et al., 2015; Kendeou et al., 2016), but at a minimum, it means a reader has built a coherent mental representation of the text by generating inferences that help establish causal relations among ideas (Kintsch, 2019; McMaster et al., 2012). These inferences link ideas together across different parts of the text and with the reader's prior knowledge, resulting in the construction of a mental model that accurately represents what the text means, not just what the text says. Construction of a mental model can be supported by the application of strategies, including comprehension monitoring and activation of background knowledge (Shanahan et al., 2010).</p> <p>The KnERDI includes items on content knowledge and application of knowledge, as both are important aspects of specialized knowledge for teaching according to the theory guiding this study. Content knowledge items (~33% of items) asked about a concept or skill important for early reading (e.g., which is the best definition of <emph>sight words</emph>?). Application items asked about teaching practices or interpretations of young readers' responses during an instructional activity (e.g., which is the most appropriate approach for helping third graders improve their identification of multisyllabic words?). Previous research has acknowledged a need to measure teachers' knowledge using content items and practice‐based items. For instance, Folsom et al. (2017) measured both types of knowledge but ultimately constructed single‐factor models that blend these dimensions together. We did not make any distinctions between these two item types in our analyses because we did not conceptualize these as dimensions of knowledge that could be measured separately in the knowledge inventory format.</p> <hd id="AN0157693737-7">Method</hd> <p>We tested our proposed model and addressed our research questions using three different forms of the KnERDI administered to three independent samples of educators. We consulted existing knowledge inventories (e.g., Binks‐Cantrell et al., 2012; Brady et al., 2009; Carlisle et al., 2011; Folsom et al., 2017; Moats, 1994; Moats & Foorman, 2003; Piasta et al., 2009; Spear‐Swerling & Zibulsky, 2014) to develop a pool of 93 items measuring knowledge in the eight facets that make up the proposed model. We hypothesized two subscales for our instrument based on the domains of early reading development that are important for literacy instruction: (a) alphabetic code and word reading; and (b) meaning and connected text processes.</p> <p>All items were multiple choice with four answer options and one correct answer. They were scored dichotomously as correct or incorrect. Experts with experience in foundational reading instruction reviewed the item pool to revise for clarity and accuracy. Sample items are found in Table 1. In order to test all 93 items without placing too much of a burden on respondents, we constructed two parallel forms of the inventory for this study. We paired each item with its closest similar item and then randomly placed one of each pair on each form. Some items were included on both forms. This resulted in Forms A and B, each consisting of 54 items. Items on each form were randomly sequenced and presented without headings or labels that would alert respondents to their corresponding facets.</p> <p>Respondents accessed the inventory as an untimed online survey. The sample was drawn from participants enrolled in a free online professional development course for educators on the topic of teaching foundational reading skills in Kindergarten through third grade. Our university's Institutional Review Board granted approval for conducting research using de‐identified data from the online course.</p> <p>Forms A and B were administered to educators enrolled in the spring 2020 course. They accessed the inventory during the introductory unit prior to engaging with the course content. Each participant was randomly assigned to complete Form A or Form B. At the time of data retrieval, there were 6,288 people enrolled in the course, of which 3,853 had actively participated in the introductory unit. Of these, 3,107 people from 47 US states and more than 40 countries completed the knowledge inventory, yielding a total response rate of 80.1% of active enrollees and 49% of all enrollees. The sample of respondents was made up primarily of classroom teachers with elementary experience. They had an average of 13 years of teaching experience (<emph>SD</emph> = 8.7), and nearly half (47%) had 6 or more years of experience teaching foundational reading skills to children. Additional demographic information for the analytic sample is found in Table 2.</p> <p>2 TABLESample Characteristics</p> <p> <ephtml> <table><thead><tr><th align="left" /><th><p>Forms A and B</p><p>(<italic>N</italic> = 3,107)</p></th><th><p>Form C</p><p>(<italic>N</italic> = 141)</p></th></tr><tr><th align="left">Characteristic</th><th><italic>n</italic></th><th><italic>%</italic></th><th><italic>n</italic></th><th><italic>%</italic></th></tr></thead><tbody><tr><td align="left">Country</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">United States</td><td align="char" char=".">2,846</td><td align="char" char=".">91.6</td><td align="char" char=".">129</td><td align="char" char=".">91.5</td></tr><tr><td align="left">Canada</td><td align="char" char=".">106</td><td align="char" char=".">3.4</td><td align="char" char=".">‐</td><td align="char" char=".">‐</td></tr><tr><td align="left">Australia</td><td align="char" char=".">58</td><td align="char" char=".">1.9</td><td align="char" char=".">‐</td><td align="char" char=".">‐</td></tr><tr><td align="left">Other<xref ref-type="fn" rid="tfn3" /></td><td align="char" char=".">97</td><td align="char" char=".">3.1</td><td align="char" char=".">12</td><td align="char" char=".">8.5</td></tr><tr><td align="left">State<xref ref-type="fn" rid="tfn4" /> (in United States; n = 2,846)</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">North Carolina</td><td align="char" char=".">2,130</td><td align="char" char=".">68.6</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Arizona</td><td align="char" char=".">66</td><td align="char" char=".">2.1</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">New York</td><td align="char" char=".">56</td><td align="char" char=".">1.8</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Pennsylvania</td><td align="char" char=".">53</td><td align="char" char=".">1.7</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Wisconsin</td><td align="char" char=".">46</td><td align="char" char=".">1.5</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Texas</td><td align="char" char=".">43</td><td align="char" char=".">1.4</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">California</td><td align="char" char=".">41</td><td align="char" char=".">1.3</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Tennessee</td><td align="char" char=".">33</td><td align="char" char=".">1.1</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Ohio</td><td align="char" char=".">31</td><td align="char" char=".">1.0</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Massachusetts</td><td align="char" char=".">31</td><td align="char" char=".">1.0</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Colorado</td><td align="char" char=".">30</td><td align="char" char=".">1.0</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Other US state<xref ref-type="fn" rid="tfn5" /></td><td align="char" char=".">286</td><td align="char" char=".">9.2</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Highest education level completed</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">2 years of college or less</td><td align="char" char=".">402</td><td align="char" char=".">12.9</td><td align="char" char=".">4</td><td align="char" char=".">2.8</td></tr><tr><td align="left">4‐year degree</td><td align="char" char=".">1,404</td><td align="char" char=".">45.2</td><td align="char" char=".">53</td><td align="char" char=".">37.5</td></tr><tr><td align="left">Master's degree</td><td align="char" char=".">1,233</td><td align="char" char=".">39.7</td><td align="char" char=".">79</td><td align="char" char=".">56.0</td></tr><tr><td align="left">Doctorate</td><td align="char" char=".">68</td><td align="char" char=".">2.2</td><td align="char" char=".">5</td><td align="char" char=".">3.5</td></tr><tr><td align="left">Primary area</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Classroom teaching</td><td align="char" char=".">1,868</td><td align="char" char=".">60.1</td><td align="char" char=".">66</td><td align="char" char=".">46.8</td></tr><tr><td align="left">Special education</td><td align="char" char=".">351</td><td align="char" char=".">11.3</td><td align="char" char=".">20</td><td align="char" char=".">14.2</td></tr><tr><td align="left">Student/preservice teacher</td><td align="char" char=".">59</td><td align="char" char=".">1.9</td><td align="char" char=".">7</td><td align="char" char=".">4.9</td></tr><tr><td align="left">School or district administration</td><td align="char" char=".">76</td><td align="char" char=".">2.4</td><td align="char" char=".">5</td><td align="char" char=".">3.5</td></tr><tr><td align="left">Other school personnel (counselor, instructional technology, librarian)</td><td align="char" char=".">90</td><td align="char" char=".">2.9</td><td align="char" char=".">6</td><td align="char" char=".">4.3</td></tr><tr><td align="left">Professional development/curriculum and instruction support</td><td align="char" char=".">248</td><td align="char" char=".">7.9</td><td align="char" char=".">15</td><td align="char" char=".">10.6</td></tr><tr><td align="left">Research/teacher education</td><td align="char" char=".">27</td><td align="char" char=".">0.9</td><td align="char" char=".">5</td><td align="char" char=".">3.5</td></tr><tr><td align="left">Other</td><td align="char" char=".">388</td><td align="char" char=".">12.5</td><td align="char" char=".">17</td><td align="char" char=".">12.1</td></tr><tr><td align="left">Years of experience teaching foundational reading skills</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">None</td><td align="char" char=".">309</td><td align="char" char=".">9.9</td><td align="char" char=".">15</td><td align="char" char=".">10.6</td></tr><tr><td align="left">Less than 2 years</td><td align="char" char=".">560</td><td align="char" char=".">18.0</td><td align="char" char=".">17</td><td align="char" char=".">12.1</td></tr><tr><td align="left">2–5 years</td><td align="char" char=".">776</td><td align="char" char=".">25.0</td><td align="char" char=".">44</td><td align="char" char=".">31.2</td></tr><tr><td align="left">6–15 years</td><td align="char" char=".">917</td><td align="char" char=".">29.5</td><td align="char" char=".">42</td><td align="char" char=".">29.8</td></tr><tr><td align="left">More than 15 years</td><td align="char" char=".">545</td><td align="char" char=".">17.5</td><td align="char" char=".">23</td><td align="char" char=".">16.3</td></tr><tr><td align="left">Grade level specialization</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Elementary only</td><td align="char" char=".">2,518</td><td align="char" char=".">81.0</td><td align="char" char=".">87</td><td align="char" char=".">61.7</td></tr><tr><td align="left">Elementary and middle or high school</td><td align="char" char=".">127</td><td align="char" char=".">4.1</td><td align="char" char=".">16</td><td align="char" char=".">11.3</td></tr><tr><td align="left">Middle or high school only</td><td align="char" char=".">134</td><td align="char" char=".">4.3</td><td align="char" char=".">17</td><td align="char" char=".">12.1</td></tr><tr><td align="left">Post‐secondary only</td><td align="char" char=".">7</td><td align="char" char=".">0.2</td><td align="char" char=".">4</td><td align="char" char=".">2.8</td></ tr><tr><td align="left">Pre‐Kindergarten (PreK) only</td><td align="char" char=".">66</td><td align="char" char=".">2.1</td><td align="char" char=".">3</td><td align="char" char=".">2.1</td></tr><tr><td align="left">PreK and elementary</td><td align="char" char=".">136</td><td align="char" char=".">4.4</td><td align="char" char=".">9</td><td align="char" char=".">6.4</td></tr><tr><td align="left">PreK and middle or high school</td><td align="char" char=".">55</td><td align="char" char=".">1.8</td><td align="char" char=".">2</td><td align="char" char=".">1.4</td></tr><tr><td align="left">Missing</td><td align="char" char=".">64</td><td align="char" char=".">2.1</td><td align="char" char=".">3</td><td align="char" char=".">2.1</td></tr></tbody></table> </ephtml> </p> <ulist> <item>3 a Includes more than 40 additional countries with less than 1% each.</item> <item>4 b Information about state was not available for the Form C sample, but based on historical trends of the online course, we can infer that the majority were from NC.</item> <item>5 c Includes 36 additional states with less than 1% each.</item> </ulist> <p>After completing the analyses using Forms A and B of the KnERDI, we applied our model and instrument to a new sample of respondents. First, we constructed a revised version of the KnERDI instrument (Form C), containing 43 items selected from the items that survived the item analysis process. This version included 19 items in the alphabetic code/word reading subscale and 24 items in the meaning/connected text processes subscale. This version was made available in the introductory unit of a subsequent session of the online professional development course for educators.</p> <p>Form C of the instrument was completed by 141 educators. These participants were from a subsample of 414 course enrollees (34% response rate) who had been given access to this version of the instrument at the time of data retrieval (other enrollees were given access to a different version which was not used in the present study). Like the larger sample used in the initial analyses, these educators were primarily from the United States (91.5%). They had an average of 14.2 years of teaching experience overall (<emph>SD</emph> = 8.6), slightly higher than the original samples for Forms A and B. See Table 2.</p> <hd id="AN0157693737-8">Data Analysis and Results</hd> <p>In this section, we describe the analytic strategies we used and the results as they relate to our three research questions. We begin with Forms A and B to evaluate the adequacy of our theory and instrument for measuring teachers' specialized knowledge for early reading in two dimensions (RQ1). We also report what our findings from Forms A and B show about the level of association of educators' knowledge across the dimensions (RQ2) and the relation between knowledge as measured on the KnERDI and education level and years of experience (RQ3). Finally, we describe the extent to which these findings were replicated in a third sample with Form C of the instrument.</p> <hd id="AN0157693737-9">Item Analysis and Subscale Development</hd> <p>We computed corrected item‐total correlations for the items corresponding to each of the two theorized subscales (alphabetic code/word reading and meaning/connected text processes). The goal of this analysis was to identify items that had sufficient internal consistency with their proposed subscales. We removed items if they did not have a positive correlation of.25 or higher with the remaining items in the subscale. We conducted these analyses first with Form A to determine if internally consistent subscales corresponding to the proposed domains could be developed. Then we repeated the process using Form B to confirm the same structure could be developed with a different sample. The item analysis process resulted in the removal of 13 items from the original 54 on each form, leaving 41 total items on each form. In addition, we examined the percentage of respondents who answered each item correctly as an indicator of item difficulty. The goal of this analysis was to identify items for additional scrutiny that were too easy (answered correctly by more than 90%) or too difficult (less than 25% correct). No items were removed from the subscales based on difficulty. One item on Form A (How many phonemes are in the word <emph>fixes</emph>?) was answered correctly by fewer than 25% of the sample, but it was retained in the scale because it includes an important concept in teachers' knowledge of English phonemes (the letter x usually corresponds to two spoken sounds) and met the inter‐item consistency criterion discussed above.</p> <p>Using the surviving items, we computed subscale scores for each domain by summing the number of items answered correctly. We also computed the total score by summing across the items on both subscales. We did not separate content and pedagogy into distinct categories in our model because this distinction was not theoretically aligned with the purpose of this study.</p> <hd id="AN0157693737-10">Examining Reliability and Patterns of Performance on the Proposed Subscales</hd> <p>Evidence of internal consistency reliability was examined using the Kuder–Richardson reliability index (KR‐20), analogous to Cronbach's alpha but used when items are scored dichotomously (correct or incorrect), for the subscales and for the whole instrument. Also, we obtained reliability information by administering both initial forms (A and B) of the instrument to independent samples of educators. Before conducting the analyses, we established the following criteria for acceptable reliability: (a) reliability coefficients approaching.80 or higher; and (b) patterns of performance on Form A of the instrument that are replicated in the Form B sample.</p> <p>Forms A and B of the KnERDI were found to measure the two proposed domains with acceptable reliability in a large sample of educators. On Form A, we constructed a subscale for alphabetic code/word reading made up of 14 items, with internal consistency reliability of.76. The second subscale, meaning/connected text processes, included 27 items with a reliability coefficient of.81. The reliability for the entire 41‐item instrument was.87. This pattern of results was replicated with an independent sample and a partially different set of items in Form B.</p> <p>Descriptive statistics for Forms A and B of the KnERDI are shown in Table 3. The scores did not differ significantly across Forms A and B on either subscale or overall. For alphabetic code/word reading, <emph>t</emph>(2911) = 1.29, <emph>p </emph>= .20; for meaning/connected text processes, <emph>t</emph>(2787) = 0.75, <emph>p </emph>= .45; and for overall score, <emph>t</emph>(2661) = .65, <emph>p </emph>= .52. The subscale scores were positively correlated (<emph>r </emph>= .70 and.68 on Forms A and B, respectively). Educators correctly answered 69% of the items on average. On both forms, they answered a significantly higher percentage of items correctly on the meaning/connected text processes scale than on the alphabetic code/word reading scale, Form A, <emph>t</emph>(1392) = 12.97, <emph>p </emph>< .001; Form B, <emph>t</emph>(1269) = 10.72, <emph>p </emph>< .001. In practical terms, these were modest differences of about 5%, equivalent to a paired‐groups standardized mean difference of around 0.28.</p> <p>3 TABLEDescriptive Statistics and Reliability for the KnERDI Subscales and Total Score</p> <p> <ephtml> <table><thead><tr><th align="left" /><th align="left">Form A</th><th align="left">Form B</th><th align="left">Form C</th></tr><tr><th align="left">Scale</th><th align="left"><italic>N</italic></th><th align="left">Reliability</th><th align="left"><italic>Mean</italic> items correct (<italic>SD</italic>)</th><th align="left"><italic>Mean</italic> % of items correct (<italic>SD</italic>)</th><th align="left"><italic>N</italic></th><th align="left">Reliability</th><th align="left"><italic>Mean</italic> items correct (<italic>SD</italic>)</th><th align="left"><italic>Mean</italic> % of items correct (<italic>SD</italic>)</th><th align="left"><italic>N</italic></th><th align="left">Reliability</th><th align="left"><italic>Mean</italic> items correct (<italic>SD</italic>)</th><th align="left"><italic>Mean</italic> % of items correct (<italic>SD</italic>)</th></tr></thead><tbody><tr><td align="left">Alphabetic code and word reading<xref ref-type="fn" rid="tfn11" /></td><td align="char" char=",">1,542</td><td align="char" char=".">.76</td><td align="left">9.09 (3.05)</td><td align="left">64.92 (21.78)</td><td align="char" char=".">1371</td><td align="char" char=".">.75</td><td align="left">10.63 (3.24)</td><td align="left">66.45 (20.28)</td><td align="char" char=".">141</td><td align="char" char=".">.78</td><td align="left">12.29 (3.72)</td><td align="left">64.69 (19.57)</td></tr><tr><td align="left">Meaning and connected text processes<xref ref-type="fn" rid="tfn12" /></td><td align="char" char=",">1,440</td><td align="char" char=".">.81</td><td align="left">19.06 (4.86)</td><td align="left">70.59 (18.01)</td><td align="char" char=".">1349</td><td align="char" char=".">.80</td><td align="left">17.77 (4.46)</td><td align="left">71.10 (17.84)</td><td align="char" char=".">141</td><td align="char" char=".">.75</td><td align="left">17.08 (3.89)</td><td align="left">71.16 (16.22)</td></tr><tr><td align="left">All items<xref ref-type="fn" rid="tfn13" /></td><td align="char" char=",">1,393</td><td align="char" char=".">.87</td><td align="left">28.31 (7.22)</td><td align="left">69.06 (17.61)</td><td align="char" char=".">1270</td><td align="char" char=".">.86</td><td align="left">28.49 (7.03)</td><td align="left">69.48 (17.15)</td><td align="char" char=".">141</td><td align="char" char=".">.86</td><td align="left">29.37 (6.98)</td><td align="left">68.30 (16.23)</td></tr></tbody></table> </ephtml> </p> <ulist> <item>11 a Number of items = 14 on Form A, 16 on Form B, and 19 on Form C.</item> <item>12 b Number of items = 27 on Form A, 25 on Form B, and 24 on Form C.</item> <item>13 c Number of items = 41 on Forms A and B, 43 on Form C.</item> </ulist> <p>As a sensitivity analysis, we checked to see if the patterns we observed in the full sample were evident if we only used the subset of participants from the United States and from the state of North Carolina (the most represented country and state in our dataset). The same patterns we have reported continue to hold. Specifically, in the US sample, we computed reliability coefficients for the two subscales similar to the overall findings (.76 for alphabetic code/word reading;.82 for meaning/connected text processes in Form A;.74 and.80 in form B). For the NC‐only sample, reliabilities were.73 and.80 in Form A and.71 and.79 in Form B. In the US‐only sample, the difference in performance favoring the meaning/connected text subscale was found to be about the same as in the whole sample (71% vs. 65% in Form A; 71% vs. 66% in Form B). In the NC sample, the differences are similar (69% vs. 63% in Form A; 69% vs. 64% in form B). Finally, the correlations between the two subscales in these subsamples are around the same magnitude as reported in the full study (Form A: <emph>r </emph>= .71 in the US sample and.68 in the NC sample; Form B: <emph>r </emph>= .68 in US and.65 in NC).</p> <hd id="AN0157693737-11">Examining Validity</hd> <p>To examine the validity of our hypothesized model, we conducted a confirmatory factor analysis (CFA) in Lisrel 9.30. CFA was chosen as the preferred method for these analyses because we were primarily interested in examining the dimensional structure (one‐factor vs. two‐factor) of the latent construct measured on the KnERDI. We specified a model in which scores on decoding and phonological/phonemic awareness loaded on the latent variable of alphabetic code/word reading; and the remaining facets loaded on the latent variable of meaning/connected text processes. We combined the data from Forms A and B of the KnERDI to analyze a single model. For each form, we computed raw scores (sums) for the surviving items in each of the eight facets of knowledge included in the KnERDI. Then we computed standardized scores (z‐scores) for each facet. These served as the observed indicators used to generate the sample variance–covariance matrix for the CFA.[<reflink idref="bib1" id="ref1">1</reflink>]</p> <p>Bivariate correlations among the indicator scores are shown in Table 4. The distributions for all the indicator means showed evidence of skewness, and all but three (mental modeling, decoding, and text structure) showed evidence of kurtosis. Weighted least squares estimation was used in the CFA, recommended when observed indicators show evidence of nonnormality (Brown, 2006). Using the CFA model, we examined validity in three ways (goodness of fit, meaningfulness of factor loadings, and discriminant validity) with pre‐planned criteria. Goodness of fit for the CFA model was evaluated using the standardized root mean square residual (SRMR), root mean square error of approximation (RMSEA), comparative fit index (CFI), and adjusted goodness‐of‐fit index (AGFI). Guided by Brown (2006) and Hu and Bentler (1999), criteria for good model fit were set as: SRMR below.08, RMSEA less than.05, and both CFI and AGFI greater than.95.</p> <p>4 TABLECorrelation Matrix for Indicators Used in the Confirmatory Factory Analysis</p> <p> <ephtml> <table><thead><tr><th align="left" /><th align="left">Mental modeling</th><th align="left">Decoding</th><th align="left">Fluency</th><th align="left">Morph</th><th align="left">Syntax</th><th align="left">Text Structure</th><th align="left">Vocab</th></tr></thead><tbody><tr><td align="left">Decoding</td><td align="char" char=".">.56</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Fluency</td><td align="char" char=".">.47</td><td align="char" char=".">.49</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Morph</td><td align="char" char=".">.49</td><td align="char" char=".">.51</td><td align="char" char=".">.40</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Syntax</td><td align="char" char=".">.39</td><td align="char" char=".">.38</td><td align="char" char=".">.29</td><td align="char" char=".">.37</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Text Structure</td><td align="char" char=".">.45</td><td align="char" char=".">.43</td><td align="char" char=".">.39</td><td align="char" char=".">.39</td><td align="char" char=".">.30</td><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Vocab</td><td align="char" char=".">.44</td><td align="char" char=".">.44</td><td align="char" char=".">.38</td><td align="char" char=".">.42</td><td align="char" char=".">.29</td><td align="char" char=".">.36</td><td align="char" char="." /></tr><tr><td align="left">PA</td><td align="char" char=".">.44</td><td align="char" char=".">.55</td><td align="char" char=".">.38</td><td align="char" char=".">.42</td><td align="char" char=".">.28</td><td align="char" char=".">.29</td><td align="char" char=".">.33</td></tr></tbody></table> </ephtml> </p> <ulist> <item>6 Note</item> <item>14 <emph>N</emph>s range from 2,872 to 3,028. Morph = morphology. Vocab = vocabulary/semantics. PA = phonological/phonemic awareness. Correlations calculated using standardized (z) scores for all variables with <emph>M</emph> = 0 and <emph>SD</emph> = 1.</item> <item>15 All correlations are significant, <emph>p</emph> < .001.</item> </ulist> <p>The fit indices for the CFA model were within the criteria we established for good model fit (SRMR = .016; RMSEA = .022; 90% confidence interval for RMSEA [0.0136 ‐ 0.0308]; CFI = .979; AGFI = .992). An exception is the chi‐square goodness‐of‐fit index (<emph>x<sups>2</sups></emph> = 43.87, <emph>df</emph> = 19, <emph>p</emph> = .001). A significant chi‐square value is often interpreted as an indication of poor model fit. However, there are known limitations and biases of this indicator, particularly for large sample sizes, which makes the other indices reported above more appropriate and robust for examining model fit (Brown, 2006). Inspection of the standardized residuals did not reveal any serious issues with the fit of the modeled solution. The largest residuals were ‐2.47, 2.43, and 2.26, well within the +/‐2.58 boundary used to identify areas of localized poor fit. All other residuals had an absolute value less than 2.0 (<emph>M</emph> = 0.38). Modification indices were not used to revise the model because the suggested modifications did not substantially change the fit of the model and were not theoretically justifiable.[<reflink idref="bib2" id="ref2">2</reflink>]</p> <p>We examined the magnitude of the factor loadings for each path in the model to understand the extent to which the theory‐driven latent constructs explain meaningful portions of the variance in the indicators. As a criterion for validity, we established that each factor should explain a significant and meaningful portion of variance in all of its proposed indicators. We chose to define standardized factor loadings of.40 as the minimum threshold for meaningfulness.</p> <p>The parameter estimates are presented in Figure 1 and Table 5. We have represented the latent constructs as ovals and the observed indicator scores as rectangles. Each indicator is modeled as a function of the correlation with its corresponding latent factor plus error variance. Because the indicator scores used in the analysis were standardized, the parameter estimates shown on each path in the figure can be interpreted as standardized values similar to correlation coefficients. Their corresponding <emph>R<sups>2</sups></emph> values (in Table 5) estimate the proportion of variance in the indicator explained by the latent construct.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jul22/rrq445-fig-0001.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq445-fig-0001.jpg" title="1 Path Diagram of Two‐Factor Model" /> </p> <p></p> <p>5 TABLESummary Statistics and Parameter Estimates from the Two‐factor Model</p> <p> <ephtml> <table><thead><tr><th align="left">Factor</th><th align="left">Factor loading</th><th align="left"><italic>SE</italic></th><th align="left"><italic>Z</italic> value</th><th align="left"><italic>R<sup>2</sup></italic></th><th align="left"><italic>AVE</italic></th></tr></thead><tbody><tr><td align="left">Alphabetic code and word reading by:</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char=".">.56</td></tr><tr><td align="left">Decoding</td><td align="char" char=".">.823</td><td align="char" char=".">0.018</td><td align="char" char=".">45.37</td><td align="char" char=".">.70</td><td align="char" char="." /></tr><tr><td align="left">PA</td><td align="char" char=".">.636</td><td align="char" char=".">0.016</td><td align="char" char=".">38.70</td><td align="char" char=".">.41</td><td align="char" char="." /></tr><tr><td align="left">Meaning and connected text processes by:</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char=".">.38</td></tr><tr><td align="left">Fluency</td><td align="char" char=".">.589</td><td align="char" char=".">0.021</td><td align="char" char=".">28.42</td><td align="char" char=".">.38</td><td align="char" char="." /></tr><tr><td align="left">Morphology</td><td align="char" char=".">.667</td><td align="char" char=".">0.017</td><td align="char" char=".">39.52</td><td align="char" char=".">.46</td><td align="char" char="." /></tr><tr><td align="left">Syntax</td><td align="char" char=".">.488</td><td align="char" char=".">0.019</td><td align="char" char=".">25.59</td><td align="char" char=".">.24</td><td align="char" char="." /></tr><tr><td align="left">Text structure</td><td align="char" char=".">.566</td><td align="char" char=".">0.020</td><td align="char" char=".">25.59</td><td align="char" char=".">.33</td><td align="char" char="." /></tr><tr><td align="left">Vocabulary</td><td align="char" char=".">.559</td><td align="char" char=".">0.020</td><td align="char" char=".">28.77</td><td align="char" char=".">.33</td><td align="char" char="." /></tr><tr><td align="left">Mental modeling</td><td align="char" char=".">.724</td><td align="char" char=".">0.017</td><td align="char" char=".">41.66</td><td align="char" char=".">.55</td><td align="char" char="." /></tr></tbody></table> </ephtml> </p> <ulist> <item>7 Note</item> <item>16 <emph>N</emph> = 2,663 after listwise deletion of cases with missing data.</item> <item>17 AVE = average variance extracted. PA = phonologica/phonemic awareness.</item> <item>18 All factor loadings are significant, <emph>p</emph>s < .001.</item> </ulist> <p>All estimates were statistically significant (<emph>p</emph>s < .001), positive, and of meaningful size, consistent with expectations from theory. Decoding had the strongest association with its proposed latent factor (<emph>R<sups>2</sups></emph> = .70), followed by mental modeling (<emph>R<sups>2</sups></emph> = .55). The remaining indicators had small to moderately sized associations with their corresponding latent variable (<emph>R<sups>2</sups></emph>s range from.24 to.46). These patterns are consistent with the assertion that these facets of teacher knowledge are suitable indicators of the two constructs, albeit with varying levels of strength.</p> <p>Our final check on the model was to examine the discriminant validity of the two proposed latent factors in the CFA model. If teachers' knowledge corresponds to separate dimensions as defined in our theory, then two conditions of discriminant validity should be met. First, the separate underlying abilities in the multi‐dimensional model should have small to moderate, but not high, correlations with each other. Second, the average variance extracted (<emph>AVE</emph>) for each factor should be greater than the amount of variance shared between the latent factors. The <emph>AVE</emph> quantifies the amount of shared variance among the indicators for each factor; it is calculated as the amount of variation that a construct explains in the observed variables (sum of squared standardized factor loadings) in relation to the total variance including error (Carter, 2016; Fornell & Larcker, 1981). Neither of these conditions of discriminant validity was met.</p> <p>The correlation of the two latent constructs was high (<emph>r</emph> = .89). A correlation of this size is indicative of poor discriminant validity and signals the indicators may correspond to a unidimensional construct. As shown in Table 5, the <emph>AVE</emph> for the alphabetic code/word reading factor indicates that an estimated 56% of the variance in the indicators is explainable by the construct. For the meaning/connected text processes factor, the <emph>AVE</emph> was much lower, with only 38% of the variance explained by the proposed construct. Both values are lower than the amount of variance shared between the constructs, estimated to be 79% (0.89^2). Thus, it appears from the two‐factor model there is so much overlap between the constructs that they can best be thought of as one higher‐order factor.</p> <p>We ran a simplified CFA model in which all eight indicators were loaded onto a single latent construct (Figure 2 and Table 6). This model provided a reasonably good fit to the data (SRMR = .024; RMSEA = .039; 90% confidence interval for RMSEA [.0317 ‐.0468]; CFI = .932; AGFI = .982), with the exception of the chi‐squared test (<emph>x<sups>2</sups></emph> = 101.26, <emph>df</emph> = 20, <emph>p </emph>< .001), as indicated above. Compared to the single‐factor model, the original two‐factor model provided a better overall fit to the data (<emph>x<sups>2</sups> difference</emph> = 57.39, <emph>df</emph> = 1, <emph>p</emph> < .001). However, due to the high correlation of the subscales, we chose to retain the parsimonious one‐factor model as the best way to depict the structure of the knowledge measured on the KnERDI.</p> <p> <img src="https://imageserver.ebscohost.com/img/embimages/rdk/NRNU/01Jul22/rrq445-fig-0002.jpg?ephost1=dGJyMNXb4kSepq84yOvqOLCmsE6epq5Srqa4SK6WxWXS" alt="rrq445-fig-0002.jpg" title="2 Path Diagram of Single‐Factor Model" /> </p> <p></p> <p>6 TABLEParameter Estimates from the Single‐factor Model</p> <p> <ephtml> <table><thead><tr><th align="left">Knowledge facet</th><th align="left">Factor loading</th><th align="left"><italic>SE</italic></th><th align="left"><italic>Z</italic> value</th><th align="left"><italic>R<sup>2</sup></italic></th></tr></thead><tbody><tr><td align="left">Reading knowledge by:</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /></tr><tr><td align="left">Decoding</td><td align="char" char=".">0.773</td><td align="char" char=".">0.017</td><td align="char" char=".">46.59</td><td align="char" char=".">.62</td></tr><tr><td align="left">PA</td><td align="char" char=".">0.623</td><td align="char" char=".">0.016</td><td align="char" char=".">38.70</td><td align="char" char=".">.40</td></tr><tr><td align="left">Fluency</td><td align="char" char=".">0.579</td><td align="char" char=".">0.020</td><td align="char" char=".">28.43</td><td align="char" char=".">.37</td></tr><tr><td align="left">Morphology</td><td align="char" char=".">0.660</td><td align="char" char=".">0.017</td><td align="char" char=".">39.64</td><td align="char" char=".">.45</td></tr><tr><td align="left">Syntax</td><td align="char" char=".">0.486</td><td align="char" char=".">0.019</td><td align="char" char=".">25.87</td><td align="char" char=".">.24</td></tr><tr><td align="left">Text structure</td><td align="char" char=".">0.558</td><td align="char" char=".">0.019</td><td align="char" char=".">28.78</td><td align="char" char=".">.33</td></tr><tr><td align="left">Vocabulary</td><td align="char" char=".">0.554</td><td align="char" char=".">0.019</td><td align="char" char=".">28.57</td><td align="char" char=".">.32</td></tr><tr><td align="left">Mental modeling</td><td align="char" char=".">0.720</td><td align="char" char=".">0.017</td><td align="char" char=".">41.88</td><td align="char" char=".">.54</td></tr></tbody></table> </ephtml> </p> <ulist> <item>8 Note</item> <item>19 <emph>N</emph> = 2,663 after listwise deletion of cases with missing data.</item> <item>20 All factors loadings are significant, <emph>p</emph>s < .001.</item> </ulist> <p>The single‐factor model shows a similar pattern as the two‐factor model, with decoding and mental modeling explained best by the latent factor (62% and 54% of variance explained, respectively) relative to the other indicators. We contend that there are conceptual and practical reasons to examine teachers' knowledge of alphabetic code/word reading and meaning/connected text processes separately, but they are too highly related to be considered psychometrically distinct factors as measured on our instrument.</p> <hd id="AN0157693737-14">Examining Knowledge Association or Disassociation</hd> <p>To further examine the degree of association between educators' knowledge of the two domains measured on the KnERDI, we identified how common it was in our sample for educators to perform much higher in one domain compared to the other. This was a pre‐planned analysis that enhanced our interpretation of the subscale correlations and low discriminant validity observed in the CFA model. Given the extensive attention that has recently been placed on teachers' knowledge of code‐based aspects of early reading, and minimal attention to higher‐level language and text processes, we were interested in quantifying the incidence of teachers having high knowledge in one area and low knowledge in the other.</p> <p>We split participants into high‐ and low‐knowledge groups, relative to the others in the sample. The high/low grouping variable was created by splitting the sample at the median score of each subscale (High = above median, Low = below median). The median was chosen as a meaningful cut‐point for this analysis because our goal was to create relatively equal groups of high and low performance in a sample in which the subscale scores were slightly negatively skewed. We recognize that the chosen cut‐point is somewhat arbitrary, but in the absence of other ways of benchmarking high and low performance on this measure of teacher knowledge, the median was a logical way to define these categories in our sample. A common problem with median splits is that scores adjacent to the median on either side get treated as categorically different from each other despite their proximity. To avoid this problem when assigning high and low designations in our sample, we assigned all cases with subscale scores at or adjacent (+/‐ 1 raw point) to the medians as a separate category, defined as average performance.</p> <p>We defined cases as having associated levels of knowledge if they were categorized similarly in both KnERDI subscales. Disassociated knowledge was defined as scoring high on one subscale and low on the other. Results are shown in Table 7. Around half of participants (50.5%) scored at or adjacent to the median of one of the subscales on Forms A and B. Individuals in this category were considered to have associated knowledge in the two areas (i.e., average performance in one area and high or low in the other, but no extreme high‐low difference). An additional 43% of the Form A and B sample had associated knowledge of the two areas measured on the KnERDI (i.e., high–high or low–low). Thus, nearly all educators who scored high in one subscale scored high in the other, and vice versa. Across Forms A and B, only 6% of the total sample did not fit this pattern of associated knowledge. We interpret this small percentage as evidence that it is extremely rare for teachers to develop knowledge in one area independently of the other area.</p> <p>7 TABLEFrequency of Associated and Disassociated Scores on Two Dimensions of Knowledge</p> <p> <ephtml> <table><thead><tr><th align="left" /><th align="left">Number of Educators</th></tr><tr><th align="left"><p>Forms A and B</p><p>(<italic>N</italic> = 2,663)</p></th><th align="left"><p>Form C</p><p>(<italic>N</italic> = 141)</p></th></tr></thead><tbody><tr><td align="left">Associated</td><td align="left" /><td align="left" /></tr><tr><td align="left">Near median on either subscale</td><td align="left">1,345 (50.5%)</td><td align="left">70 (49.6%)</td></tr><tr><td align="left">Low on both subscales</td><td align="left">673 (25.2%)</td><td align="left">32 (22.7%)</td></tr><tr><td align="left">High on both subscales</td><td align="left">479 (18%)</td><td align="left">28 (19.9%)</td></tr><tr><td align="left">Total</td><td align="left">2,497 (93.8%)</td><td align="left">130 (92.2%)</td></tr><tr><td align="left">Disassociated</td><td align="left" /><td align="left" /></tr><tr><td align="left">Low alphcode, high txproc</td><td align="left">88 (3.3%)</td><td align="left">7 (5.0%)</td></tr><tr><td align="left">High alphcode, low txproc</td><td align="left">78 (2.9%)</td><td align="left">4 (2.8%)</td></tr><tr><td align="left">Total</td><td align="left">166 (6.2%)</td><td align="left">11 (7.8%)</td></tr></tbody></table> </ephtml> </p> <ulist> <item>9 Note</item> <item>21 Alphcode = alphabetics and word reading subscale; txproc = meaning and connected text processes subscale; high = above subscale median; low = below subscale median</item> </ulist> <hd id="AN0157693737-15">Examining Relations between Knowledge and Education and Experience Levels</hd> <p>When enrolling in the online professional development course, educators completed a survey in which they indicated their highest degree obtained by selecting one of these four options: 2 years of college or less, 4‐year degree, master's degree, or doctorate. They also indicated their years of experience in two ways. First, they provided their total years of experience in education. These responses were treated as a continuous variable in our analyses. Second, they indicated their years of experience teaching foundational reading in grades K‐3 by selecting one of five categorical options: none, less than 2 years, between 2 and 5 years, between 6 and 15 years, or more than 15 years.</p> <p>We used one‐way analysis of variance (ANOVA) to compare educators' KnERDI scores across the two categorical variables (education level and experience teaching K‐3 reading). <emph>P</emph>‐values were evaluated using a Bonferroni‐adjusted test‐level <emph>alpha</emph> of.008 to adjust for multiple comparisons (family‐wise error rate of.05 across six separate ANOVAs for each sample). We combined the master's and doctoral categories into one group for the analyses because there were few educators with doctorates in the sample. We examined the association between KnERDI scores and total years of experience in education using bivariate correlations. Despite the mixed results in previous research, we hypothesized that educators with advanced degrees would outperform those with 4‐year degrees or less on both subscales and overall. Furthermore, we expected educators with more teaching experience overall and more experience teaching foundational reading would outperform those with less experience.</p> <p>The subgroup comparisons are shown in Table 8. On Forms A and B, teachers with advanced degrees performed higher on both subscales and overall compared to teachers with 4‐year degrees only, who in turn outperformed respondents who had yet to earn a 4‐year degree. The <emph>F</emph>‐values were significant for all pairwise contrasts, as reported in Table 8. To interpret the strength of association between education level and total score, we evaluated partial <emph>eta</emph>‐squared as an estimate of the proportion of variance explained. These values were.19 on Form A and.17 on Form B, which we interpret as small associations. Expressed as an effect size, the average difference between teachers with a graduate degree and those with a 4‐year degree was 0.38 and 0.33 standard deviations on Forms A and B, respectively.</p> <p>8 TABLESubgroup Comparisons on Alphabetics/Word Reading, Meaning/Connected Text Processes, and Total Scores</p> <p> <ephtml> <table><thead><tr><th align="left" /><th align="left"><p><italic>n</italic></p></th><th align="left"><p>Alphabetics/word reading,</p><p><italic>M</italic> (<italic>SD</italic>)</p></th><th align="left"><p>Meaning/connected text processes,</p><p><italic>M</italic> (<italic>SD</italic>)</p></th><th align="left"><p>Total score,</p><p><italic>M</italic> (<italic>SD</italic>)</p></th></tr><tr><th align="left" /><th align="left">Form A</th><th align="left">Form B</th><th align="left">Form C</th><th align="left">Form A (14 items)</th><th align="left">Form B (16 items)</th><th align="left"><p>Form C (19</p><p>items)</p></th><th align="left">Form A (27 items)</th><th align="left">Form B (25 items)</th><th align="left"><p>Form C</p><p>(24 items)</p></th><th align="left">Form A (41 items)</th><th align="left">Form B (41 items)</th><th align="left">Form C (43 items)</th></tr></thead><tbody><tr><td align="left">Education level</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /></tr><tr><td align="left">< 4‐year degree</td><td align="char" char=".">162</td><td align="char" char=".">158</td><td align="char" char=".">‐</td><td align="left">6.50<xref ref-type="fn" rid="tfn23" /><sub>1</sub>(2.98)</td><td align="left">8.05<xref ref-type="fn" rid="tfn24" /><sub>1</sub>(3.17)</td><td align="left">‐</td><td align="left">13.94<xref ref-type="fn" rid="tfn26" /><sub>1</sub>(4.53)</td><td align="left">13.34<xref ref-type="fn" rid="tfn27" /><sub>1</sub>(4.24)</td><td align="left">‐</td><td align="left">20.44<xref ref-type="fn" rid="tfn29" /><sub>1</sub>(6.89)</td><td align="left">21.39<xref ref-type="fn" rid="tfn30" /><sub>1</sub>(6.68)</td><td align="left">‐</td></tr><tr><td align="left">4‐year degree</td><td align="char" char=".">637</td><td align="char" char=".">573</td><td align="char" char=".">53</td><td align="left">9.08<sub>2</sub>(2.83)</td><td align="left">10.63<sub>2</sub>(3.06)</td><td align="left">12.32<xref ref-type="fn" rid="tfn25" /><sub>1</sub>(3.29)</td><td align="left">18.95<sub>2</sub> (4.50)</td><td align="left">17.74<sub>2</sub> (4.06)</td><td align="left">17.30<xref ref-type="fn" rid="tfn28" /><sub>1</sub>(3.51)</td><td align="left">28.03<sub>2</sub> (6.71)</td><td align="left">28.36<sub>2</sub> (6.41)</td><td align="left">29.62<xref ref-type="fn" rid="tfn31" /><sub>1</sub>(6.00)</td></tr><tr><td align="left">Master's or higher</td><td align="char" char=".">594</td><td align="char" char=".">539</td><td align="char" char=".">84</td><td align="left">10.0<sub>3</sub>(2.74)</td><td align="left">11.47<sub>3</sub>(3.02)</td><td align="left">12.45<sub>1</sub>(3.86)</td><td align="left">20.76<sub>3</sub> (4.06)</td><td align="left">19.21<sub>3</sub> (3.96)</td><td align="left">17.18<sub>1</sub>(4.05)</td><td align="left">30.76<sub>3</sub> (6.18)</td><td align="left">30.68<sub>3</sub> (6.36)</td><td align="left">29.63<sub>1</sub>(7.34)</td></tr><tr><td align="left">Experience teaching foundational reading to children</td><td align="char" char="." /><td align="char" char="." /><td align="char" char="." /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /><td align="left" /></tr><tr><td align="left">None</td><td align="char" char=".">146</td><td align="char" char=".">118</td><td align="char" char=".">15</td><td align="left">7.17<xref ref-type="fn" rid="tfn32" /><sub>1</sub>(2.87)</td><td align="left">8.42<xref ref-type="fn" rid="tfn33" /><sub>1</sub>(3.07)</td><td align="left">10.47<sub>1,2</sub><xref ref-type="fn" rid="tfn34" />(4.50)</td><td align="left">17.36<xref ref-type="fn" rid="tfn35" /><sub>1</sub>(5.25)</td><td align="left">15.68<xref ref-type="fn" rid="tfn36" /><sub>1</sub>(4.83)</td><td align="left">15.80<xref ref-type="fn" rid="tfn37" /><sub>1</sub>(5.98)</td><td align="left">24.53<xref ref-type="fn" rid="tfn38" /><sub>1</sub>(7.54)</td><td align="left">24.11<xref ref-type="fn" rid="tfn39" /><sub>1</sub>(7.31)</td><td align="left">26.27<xref ref-type="fn" rid="tfn40" /><sub>1</sub>(10.03)</td></tr><tr><td align="left">< 2 years</td><td align="char" char=".">236</td><td align="char" char=".">247</td><td align="char" char=".">17</td><td align="left">8.15<sub>2</sub>(2.97)</td><td align="left">10.0<sub>2</sub>(3.14)</td><td align="left">10.18<sub>1</sub>(3.57)</td><td align="left">17.92<sub>1,2</sub>(5.26)</td><td align="left">17.34<sub>2</sub>(4.34)</td><td align="left">16.35<sub>1</sub>(3.86)</td><td align="left">26.07<sub>1,2</sub>(7.71)</td><td align="left">27.34<sub>2</sub>(6.71)</td><td align="left">26.53<sub>1</sub>(6.56)</td></tr><tr><td align="left">2–5 years</td><td align="char" char=".">352</td><td align="char" char=".">314</td><td align="char" char=".">44</td><td align="left">8.76<sub>2</sub>(2.98)</td><td align="left">10.31<sub>2</sub>(3.11)</td><td align="left">12.27<sub>1,2</sub>(3.11)</td><td align="left">18.77<sub>2</sub>(4.76)</td><td align="left">17.43<sub>2</sub>(4.45)</td><td align="left">16.55<sub>1</sub>(3.19)</td><td align="left">27.53<sub>2</sub>(7.17)</td><td align="left">27.75<sub>2</sub>(6.91)</td><td align="left">28.82<sub>1</sub>(5.71)</td></tr><tr><td align="left">6–15 years</td><td align="char" char=".">438</td><td align="char" char=".">352</td><td align="char" char=".">42</td><td align="left">10.12<sub>3</sub>(2.64)</td><td align="left">11.55<sub>3</sub>(3.04)</td><td align="left">13.05<sub>2</sub>(3.86)</td><td align="left">20.03<sub>3</sub>(4.18)</td><td align="left">18.57<sub>3</sub>(4.17)</td><td align="left">17.86<sub>1</sub>(3.68)</td><td align="left">30.15<sub>3</sub>(6.23)</td><td align="left">30.12<sub>3</sub>(6.65)</td><td align="left">30.90<sub>1</sub>(6.83)</td></tr><tr><td align="left">More than 15 years</td><td align="char" char=".">221</td><td align="char" char=".">239</td><td align="char" char=".">23</td><td align="left">10.36<sub>3</sub>(2.64)</td><td align="left">11.62<sub>3</sub>(3.06)</td><td align="left">13.70<sub>2</sub>(3.20)</td><td align="left">20.47<sub>3</sub>(4.37)</td><td align="left">18.73<sub>3</sub>(4.21)</td><td align="left">18.04<sub>1</sub>(3.67)</td><td align="left">30.82<sub>3</sub>(6.38)</td><td align="left">30.35<sub>3</sub>(6.68)</td><td align="left">31.74<sub>1</sub>(6.34)</td></tr></tbody></table> </ephtml> </p> <ulist> <item>10 Note</item> <item>22 <emph>P</emph>‐values from one‐way ANOVAs were evaluated using a Bonferroni adjusted test‐level alpha of.008 to adjust for multiple comparisons (family‐wise error rate of.05 / 6 comparisons). Means with subscripts (numerals) that are the same within the column for a demographic characteristic (education or experience) are not significantly different from one another, but those with different subscripts are.</item> <item>23 a <emph>F</emph>(<reflink idref="bib2" id="ref3">2</reflink>, 1539) = 125.622; <emph>p</emph> =.000; all pairwise contrasts (Tukey Honestly Significant Difference [HSD]) are significant, <emph>p</emph>s <.001.</item> <item>24 b <emph>F</emph>(<reflink idref="bib2" id="ref4">2</reflink>, 1368) = 78.811; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.001.</item> <item>25 c <emph>F</emph>(<reflink idref="bib1" id="ref5">1</reflink>, 135) = 0.042; <emph>p =</emph>.838.</item> <item>26 d <emph>F</emph>(<reflink idref="bib2" id="ref6">2</reflink>, 1437) = 166.531; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.001.</item> <item>27 e <emph>F</emph>(<reflink idref="bib2" id="ref7">2</reflink>, 1345) = 141.930; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.001.</item> <item>28 f <emph>F</emph>(<reflink idref="bib1" id="ref8">1</reflink>, 135) = 0.033; <emph>p</emph> =.855.</item> <item>29 g <emph>F</emph>(<reflink idref="bib2" id="ref9">2</reflink>, 1390) = 160.764; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.001.</item> <item>30 h <emph>F</emph>(<reflink idref="bib2" id="ref10">2</reflink>, 1267) = 127.927; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.001.</item> <item>31 i <emph>F</emph>(<reflink idref="bib1" id="ref11">1</reflink>, 135) =.000; df=1,135; <emph>p</emph> =.994.</item> <item>32 j <emph>F</emph>(<reflink idref="bib4" id="ref12">4</reflink>, 1537) = 57.517; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.007, except for: Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years.</item> <item>33 k <emph>F</emph>(<reflink idref="bib4" id="ref13">4</reflink>, 1366) = 32.901; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.01, except for: Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years.</item> <item>34 m <emph>F</emph>(<reflink idref="bib4" id="ref14">4</reflink>, 136) = 3.818; <emph>p</emph> =.006; the following pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.05: Less than 2 years vs. 6‐15 years; Less than 2 years vs. more than 15 years.</item> <item>35 n <emph>F</emph>(<reflink idref="bib4" id="ref15">4</reflink>, 1435) = 20.103; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.007, except for: None = Less than 2 years; Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years.</item> <item>36 o <emph>F</emph>(<reflink idref="bib4" id="ref16">4</reflink>, 1343) = 14.864; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.01, except for: Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years.</item> <item>37 p <emph>F</emph>(<reflink idref="bib4" id="ref17">4</reflink>, 136) = 1.557; <emph>p</emph> =.189.</item> <item>38 q <emph>F</emph>(<reflink idref="bib4" id="ref18">4</reflink>, 1388) = 33.336; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.007, except for: None = Less than 2 years; Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years.</item> <item>39 r <emph>F</emph>(<reflink idref="bib4" id="ref19">4</reflink>, 1265) = 24.522; <emph>p</emph> =.000; all pairwise contrasts (Tukey HSD) are significant, <emph>p</emph>s <.01, except for: Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years.</item> <item>40 s F(<reflink idref="bib4" id="ref20">4</reflink>, 136) = 2.825; <emph>p</emph> =.027.</item> </ulist> <p>Also, we found a pattern of higher performance among teachers with more experience. First, teachers' self‐reported number of total years in education was positively but weakly correlated with performance in alphabetic code/word reading (<emph>r</emph> =.21 , <emph>p</emph> <.001 on Form A; <emph>r</emph> =.11, <emph>p</emph> <.001 on Form B), meaning/connected text processes (<emph>r</emph> =.19, <emph>p</emph> <.001 on Form A; <emph>r</emph> =.13, <emph>p</emph> <.001 on Form B), and overall (<emph>r</emph> =.22, <emph>p</emph> <.001 on Form A; <emph>r</emph> =.14, <emph>p</emph> <.001 on Form B). Using the categorical variable of the amount of experience teaching reading to K‐3 students as the predictor, we obtained significant <emph>F</emph>‐values from the ANOVAs for both KnERDI subscales on Forms A and B, as shown in Table 8. Descriptively, higher scores in alphabetic code/word reading and meaning/connected text processes were associated with higher levels of experience.</p> <p>The total score on Form A was associated with higher levels of experience teaching early reading, <emph>F</emph>(<reflink idref="bib4" id="ref21">4</reflink>, 1388) = 33.336, <emph>p</emph> <.001. The pairwise contrasts for the total score (Tukey honestly significant difference test) were significant, with three exceptions. Teachers with no experience were not different from those with less than 2 years; teachers with less than 2 years of experience were not different from those with 2‐5 years; and teachers at the top two experience levels (6‐15 years and more than 15 years) were not different from each other. On Form B, a similar pattern was evident (<emph>F</emph>[<reflink idref="bib4" id="ref22">4</reflink>, 1265] = 24.522; <emph>p</emph> <.001; all pairwise contrasts were significant, <emph>p</emph>s <.01, except for: Less than 2 years = 2‐5 years; 6‐15 years = more than 15 years). The partial <emph>eta</emph>‐squared values were.09 for Form A and.07 for Form B, indicating a relatively small association. As a point of reference, the difference between teachers with less than 2 years of experience teaching foundational reading in grades K‐3 and those with 6‐15 years (the modal category in our sample), is estimated to be around 0.57 standard deviations on Form A and 0.40 standard deviations on Form B.</p> <hd id="AN0157693737-16">Replicating the Findings in Form C</hd> <p>An important feature of the design of this study is that we examined our proposed model and research questions using three samples (i.e., Forms A, B, and C). Having verified the adequacy of the model of teacher knowledge instantiated on the KnERDI instrument and the patterns evident in the model in Forms A and B, we turn our attention to the Form C sample. As shown in Table 3, this form of the instrument had similar reliability as the other forms. Internal consistency estimates were.86 overall and.78 and.75 for the two subscales.</p> <p>The subscale scores were positively correlated (<emph>r</emph> =.68, <emph>p</emph> <.001). Educators correctly answered 68% of the items on average. As with the other two forms, they answered a significantly higher percentage of items correctly on the meaning/connected text processes scale than on the alphabetic code/word reading scale, <emph>t</emph>(<reflink idref="bib140" id="ref23">140</reflink>) =5.265; <emph>p</emph> <.001. Similar to Forms A and B, this difference was about 6%, a paired‐group standardized mean difference of 0.33. Also, educators who completed Form C showed a high level of knowledge association across the two subscales (Table 7). Only 11 respondents (7.8%) had high knowledge in one domain and low in the other.</p> <p>Compared to the large samples in Forms A and B, the smaller size of the Form C sample reduced the statistical power of the subgroup comparisons. Descriptively, there was a pattern of higher performance in the Form C sample for more experienced teachers of K‐3 reading, shown in Table 8. However, these descriptive differences were not all statistically significant (using a test‐level Type I error rate of.008 to account for multiple comparisons). The ANOVA showed a significant relation between experience level and the alphabetic code/word reading scale, <emph>F</emph>(<reflink idref="bib4" id="ref24">4</reflink>, 136) = 3.818; <emph>p</emph> =.006, but not for the meaning/connected text scale, <emph>F</emph>(<reflink idref="bib4" id="ref25">4</reflink>, 136) = 1.557, <emph>p</emph> =.189 or overall, <emph>F</emph>(<reflink idref="bib4" id="ref26">4</reflink>, 136) = 2.825; <emph>p</emph> =.027.</p> <p>In contrast to Forms A and B, the correlations between total number of years in education and KnERDI performance were weak and nonsignificant (<emph>r</emph> =.10, <emph>p</emph> =.249 for alphabetic code/word reading; <emph>r</emph> =.00, <emph>p</emph> =.958 for meaning/connected text; <emph>r</emph> =.06; <emph>p</emph> =.52 overall). Also in contrast to the results seen in Forms A and B, there were no differences found in the Form C sample in relation to education level; <emph>F</emph>(<reflink idref="bib1" id="ref27">1</reflink>, 135) = 0.042; <emph>p</emph> =.838 for alphabetic code/word reading; <emph>F</emph>(<reflink idref="bib1" id="ref28">1</reflink>, 135) = 0.033; <emph>p</emph> =.855 for meaning/connected text; <emph>F</emph>(<reflink idref="bib1" id="ref29">1</reflink>, 135) =.000; <emph>p</emph> =.994 overall. Teachers with 4‐year degrees and teachers with graduate degrees scored virtually identically in this sample.</p> <hd id="AN0157693737-17">Discussion</hd> <p>In this study, we proposed a model of educators' specialized knowledge of early reading instruction that includes both word reading and connected text processes. We developed and tested the KnERDI to examine the adequacy of this model and to fulfill the need for the development of a reliable instrument to measure educators' knowledge corresponding to these two major domains involved in children's reading development. We also explored the high level of association of educators' knowledge in the two domains and analyzed the relation between educators' knowledge and their years of experience and level of education. We now discuss our findings in relation to previous research and consider the implications of these findings for educators and researchers.</p> <hd id="AN0157693737-18">Using the KnERDI to Measure Teacher Knowledge</hd> <p>The KnERDI successfully met both of the criteria for reliability established prior to the study. It reliably measured educators' knowledge in the two domains, and these results were replicated across multiple forms of the instrument with large samples of educators with a wide range of experience and degree levels. Furthermore, the results from the confirmatory factor analysis revealed good model fit. Thus, this study offers an additional measurement tool to researchers and practitioners who are interested in studying teachers' knowledge for enhancing reading development.</p> <p>Although the KnERDI provided reliable subscores for the two domains we theorized, there was not sufficient discriminant validity in our model to claim with confidence that the structure of teachers' knowledge in these domains is multidimensional. Based on our findings, we conclude that teachers' knowledge of alphabetic code/word reading and meaning/connected text processes is best thought of as a single construct on the KnERDI. The pattern of highly associated domain knowledge evident in our model aligns with previous research in which multiple facets of reading and language were combined in a single teacher knowledge score (e.g., Carlisle et al., 2011).</p> <p>In order to address some of the limitations found in previous research that have led to inconsistent findings about the dimensionality of teachers' knowledge, we carefully examined the association between the proposed subscales using two different strategies. First, by examining the association in a two‐factor CFA model, we were able to provide an error‐adjusted estimate of the correlation of the two domains. This estimate is higher (.89 or 79% shared variance) than the bivariate correlation of the raw subscale scores (.70 and.68 on the two forms, or around just under 50% shared variance). Had we only interpreted the latter, we might have underestimated the degree of shared variance between teachers' knowledge of code‐ and meaning‐based concepts, thereby understating the inseparability of the two domains of knowledge. Also, by quantifying the small percentage of our sample with a disassociated pattern of knowledge across the two domains, we have offered a practical estimate of the likelihood that teachers will know a lot more about one area than the other (only 6% in our large sample). Taken together, these findings suggest that a teacher's knowledge in one area is highly predictive of their knowledge in the other area.</p> <p>Why might teachers' knowledge be so tightly associated across code‐ and meaning‐based concepts related to early reading development? One possibility is that the KnERDI subscales truly measure one underlying construct—that is, teachers' knowledge of language structures and meaning‐making processes spanning sublexical, word, and connected text levels—and that there is no meaningful distinction between the two proposed domains. Another possibility is that the two domains are conceptually distinct, but their high degree of association is due to teachers' professional development‐seeking practices or learning opportunities. It may be the case that teachers who have benefited from professional learning opportunities to gain knowledge in one area have had opportunities in the other area as well (and conversely, teachers who have not had access to PD in one area have also not had it in the other area). Future research could tease apart these possibilities by providing teachers with PD that targets one domain but not the other and examining if their performance changes differentially on the two subscales.</p> <p>Future research should also examine additional dimensions of teacher knowledge in addition to the content domains we have proposed, including a distinction between content knowledge and pedagogical knowledge. This is an open question that has garnered recent attention in the teacher knowledge literature in reading. Jordan et al. (2018) found that educators' content and pedagogical content knowledge could be measured independently as two distinct domains in their sample of kindergarten and first‐grade teachers. Phillips and colleagues (2020) administered a knowledge survey to 248 teachers and assistant teachers in classrooms of 3 to 5‐year‐old children. They interpreted their CFA results to suggest that models in which four scales (content and pedagogical content knowledge of vocabulary and language) were treated separately provided a better fit to the data than models forcing a single factor. However, they also found significant and moderate to large correlations between the factor scores for content and pedagogical content knowledge in both areas they measured (.85 for vocabulary and.68 for language).</p> <hd id="AN0157693737-19">Participants' Performance across the Two Subscales</hd> <p>Despite the high correlation of the two subscales on the KnERDI, there are still practical reasons for using such instruments to measure educators' knowledge of alphabetic code/word reading and meaning/connected text processes on separate subscales. Research on reading development has shown decoding skills and language comprehension make independent contributions to children's ability to read proficiently (Foorman et al., 2018; Petscher et al., 2020). It is insufficient for researchers and teacher educators to think of reading teachers as teachers of code‐based skills only. Teachers may tap into one part of their knowledge to teach concepts such as how to segment and blend sounds in order to decode words, and they may utilize different expertise to educate students on practices such as using text structure to build a mental model of a text. Using the KnERDI to measure knowledge in the two different domains may be useful for targeting professional development efforts to increase knowledge in a particular domain in which educators need more support.</p> <p>Our results showed teachers demonstrated superior performance on the items in the meaning/connected text processes scale compared to alphabetic code/word reading. This finding aligns with the findings of Spear‐Swerling and Cheesman (2012) in one of the few existing studies that reported direct comparisons between teachers' knowledge in code‐based and meaning‐based aspects of reading. Unfortunately, neither our findings nor the existing literature provide an empirical basis for interpreting the meaningfulness of the size of the differences we observed in educators' knowledge across these domains. The differences we observed were small, between five and six percentage points on our measure, an effect size of around.28 <emph>SDs</emph>. An important contribution of the present study is that we have demonstrated that these differences, although modest in size, were consistent across three different administrations of the KnERDI. Additional research will be needed to better understand if these differences are found on other measures of teachers' knowledge and if they are large enough to matter in practice.</p> <hd id="AN0157693737-20">Relations between Teacher Knowledge and Education and Experience Levels</hd> <p>Our third research question examined whether teachers' education level and years of experience were associated with performance on the KnERDI. In interpreting these associations, we prioritized results that were consistent across all three independent samples used in this study (Forms A, B, and C). We also prioritized the total KnERDI scores over the subscales in our interpretations because of the high correlation between the two subscales. We draw two main conclusions from these analyses.</p> <p>First, advanced degree completion does not appear to be consistently related to educators' knowledge as measured on the KnERDI. We found that educators with advanced degrees significantly outperformed educators who only had 4‐year degrees in the two larger samples, but not in the third replication sample. This difference does not appear to be due solely to the reduced statistical power of the smaller sample used for Form C, as teachers in this sample with 4‐year and advanced degrees scored nearly identically. The mixed results we obtained could be due to demographic differences between the samples. Teachers in the Form C sample had about one extra year of experience compared to those who took Forms A and B (14 compared to 13 years). They also had a higher proportion of teachers specializing outside of the elementary grades and a higher proportion of teachers with master's degrees compared to Forms A and B. In any case, our multi‐sample design produced patterns that echo a larger trend across the literature—the association between graduate degrees and knowledge varies across samples, possibly due to other sample characteristics and to the content area of the degrees.</p> <p>Second, we found a consistent association between educators' performance on the KnERDI and their years of experience teaching K‐3 reading, but not their overall teaching experience (irrespective of grade level and content). The association was most consistent on the alphabetic code/word reading subscale. Across all three samples, increased experience in K‐3 reading was associated with higher knowledge in this area. The strength of this relation was modest. For the meaning/connected text process subscale and overall KnERDI score, descriptive differences were observed in all three samples, such that more experienced teachers outperformed less experienced teachers. However, these contrasts were not large enough to be statistically significant in the Form C sample due to the smaller sample size.</p> <p>Our correlational design does not permit us to make claims about the causes of differences in knowledge in our sample. Also, the question remains of whether or not the differences across subgroups observed on the KnERDI matter for teaching practice or student learning. Although it seems logical that teachers who have high levels of knowledge of foundational language and literacy concepts will be more effective in teaching early reading, the association between knowledge, teaching practice, and student reading outcomes has been difficult to empirically establish. Several of the studies reviewed in earlier sections of this article exemplify the mixed results found in the literature. McCutchen et al. (2002) found significant correlations between teachers' phonological knowledge and kindergarten students' word‐level skills on end‐of‐year assessments, but there were no significant correlations for grades 1 and 2. In another study, McCutchen et al. (2009) found that teachers' linguistic knowledge did relate to student performance for readers receiving intervention in grades 3–5. In a study of first‐grade teachers, Piasta and colleagues (2009) found that teacher knowledge was not directly predictive of student reading gains. Instead, they found that the interaction between knowledge and practice was associated with changes in students' performance in word reading. Students benefited when they received more decoding instruction from a teacher with high knowledge. Carlisle et al. (2011) examined the relation between teacher knowledge and students' reading growth and found only partial support for a meaningful association between the two variables. They found that for children in first grade, teacher knowledge had a small but significant association with comprehension (standardized regression coefficient of 0.08) but not word analysis; and in second and third grades, there were no statistically significant relations.</p> <p>Piasta, Park, et al. (2020) proposed that the logic underlying many such studies is that improvements in teacher knowledge will produce benefits for students' reading outcomes through its associations with practice. In other words, teacher knowledge is related to teacher practice, which is related to student learning. They found some support for this hypothesis in a mediation analysis with early childhood educators. Teaching practice mediated the association between teacher knowledge and student outcomes in print concepts, but not other measures of emergent literacy skills (letter naming, phonological awareness, and oral language). Overall, teachers' content knowledge explained only 4% to 6% of the variance in students' learning gains in their study.</p> <p>Our interpretation of the existing literature is that teachers' knowledge is one factor among many factors that contribute to students' reading achievement. Knowledge informs and enables practice but does not guarantee student learning. We caution against using the KnERDI (or any other knowledge measure) in ways that might lead to exaggerated claims about the practical importance of the knowledge it represents.</p> <hd id="AN0157693737-21">Challenges to Measuring Knowledge of Meaning and Connected Text Processes</hd> <p>Although teachers performed higher on the subscale that included meaning‐based items compared to the other subscale, the factor analysis revealed our instrument did a better job measuring knowledge of the alphabetic code and word reading, in terms of the amount of variance explained in the indicators for the facets associated with the two dimensions. The items related to syntax, fluency, text structure, and vocabulary were associated with relatively low parameter estimates in the CFA model. In some ways, this is not surprising. Compared to the code‐based domain, we had less research to draw on when theorizing the facets and developing items for meaning and connected text processes. The observed fragilities in explanatory power on the KnERDI are concerning, and they signal a need for more scholarship to address the challenges associated with measuring educators' knowledge in the meaning‐based domain.</p> <p>One possible explanation that might account for differences in measurement quality between the concepts assessed in the two dimensions is the degree of constraint of the dimensions in children's development. Skills related to the alphabetic code and word reading are constrained, meaning they can be broken down into smaller subskills with clear developmental endpoints most learners can reach within a reasonable amount of time (Paris, 2005). This level of clarity makes it relatively straightforward to codify what teachers need to understand about these skills to inform instruction.</p> <p>In contrast, skills related to meaning‐making processes are less constrained. They are not as easy to discretely list along a clear learning trajectory. They develop over many years and never reach a developmental endpoint (e.g., readers continue developing new vocabulary and world knowledge to support their mental models of texts across their whole lives). Furthermore, there is substantial disagreement in the field about the best way to define and measure reading comprehension (Clemens & Fuchs, 2021). One prominent way of defining reading comprehension in common usage—constructing meaning from text—is made up of three concepts (construction, meaning, and text) that have also been hard to clearly define (see Gavelek & Whittingham, 2017). Skills related to comprehension are broader and messier than code‐based skills, which has made it harder to construct viable theories for examining what teachers should know about them.</p> <p>The conceptual model guiding the development of the KnERDI is an important contribution to a literature in which few studies have carefully theorized what teachers might be expected to understand about early comprehension development. The six facets we have proposed in the meaning/connected text process subscale of KnERDI offer a slightly broader theory‐driven approach to measuring teacher knowledge compared to some of the previously developed knowledge inventories. In particular, our inclusion of items related to mental modeling is an important contribution. Of the facets we proposed in this domain, mental modeling was associated with the highest proportion of explained variance. To answer the KnERDI items related to this facet, teachers need to know about some of the important processes that enable readers to construct coherent understandings of text ideas, such as generating inferences to link ideas together within sentences and paragraphs, and monitoring ongoing understanding. Few previous inventories have directly addressed these concepts as important aspects of what teachers should know to enhance students' construction of meaning.</p> <p>We also recognize that the KnERDI and similar instruments measure only a small portion of the knowledge that educators need and use for text comprehension instruction. Finding appropriate ways to account for reader‐external factors is a common challenge in the measurement of comprehension in children (e.g., Pearson et al., 2014) that also extends to measurement of teacher knowledge. Although we endeavored to create items that captured a broad range of meaning‐based competencies, the items on the KnERDI prioritize reader‐internal factors and do not sufficiently address factors related to social practices and contexts in which texts are called upon. The measure does not assess what teachers know about how readers' cultural experiences shape the knowledge, assumptions, and purposes they bring to a text or how to ensure that reading instruction leverages the knowledge and resources children bring to the classroom from their families and communities. We encourage future research that examines how these essential aspects of teachers' knowledge relate to the language and process knowledge assessed on the KnERDI, and importantly, how such knowledge might be made more central in the teacher knowledge literature.</p> <hd id="AN0157693737-22">Highlighting the Positive</hd> <p>As outlined in our literature review, there is a tendency in the research on teacher knowledge to highlight the areas in which teachers do not know enough. We approached this study with a more optimistic viewpoint on what teachers know and their possibilities for continued professional growth. We are committed to critically examining the way that we as researchers interpret and discuss educators' performance on knowledge instruments without contributing to deficit viewpoints about teachers.</p> <p>Although performance on single items cannot be interpreted as robust evidence of knowledge, we were encouraged by our sample's performance on several items. For instance, when asked to select the best prompt for helping a child decode an unfamiliar word, 88% of teachers correctly selected <emph>say and blend the sounds of spelling patterns in the word</emph> rather than other prompts that included skipping or guessing the word based on context. On one item that asked respondents to determine what a series of words have in common (graph, stack, break, frost, and trial), 89% accurately answered that they begin with a consonant blend. We were also encouraged to find that 81% of teachers chose the correct description of morphological awareness: <emph>the ability to recognize and manipulate the smaller meaningful parts that build words</emph>.</p> <p>We also found positive trends on items related to text structure and meaning‐making processes. For example, 75% of the sample answered correctly that explicit teaching of expository text structures is important because it helps students learn to use the organization of the text to improve their understanding. When asked how often readers make inferences to support comprehension, 83% correctly selected <emph>almost constantly during reading, to help link ideas together to form a mental model of the text</emph> (other answer choices included: S<emph>ometimes before reading, to help activate prior knowledge needed for approaching the text</emph>; <emph>Occasionally, during reading, as a fix‐up strategy when the text becomes challenging for the reader</emph>; <emph>Only after reading, when answering a question or discussing the text</emph>). We are hopeful that teachers who understand the fundamental role of text structure and inference making can use this knowledge when modeling comprehension processes for children. We consider these positive findings in light of the emphasis in recent education reform efforts that target teacher knowledge as a way of improving instructional practices and student outcomes.</p> <hd id="AN0157693737-23">Implications</hd> <p>The findings of this study can provide useful insights for teacher education practice and research. Teacher educators may benefit from administering the KnERDI and using the results to inform course design or curriculum innovations. When used for these purposes, scores can be interpreted relative to the samples reported in this article. For instance, using the results reported in Table 8, teacher educators could compare the scores of a group of beginning teachers to the average score of teachers in our sample with similar levels of experience teaching early reading. A revised version of the KnERDI is available from the lead author by request. However, we caution that the KnERDI should only be used informally, for groups of teachers rather than individuals, and without high‐stakes consequences. Also, users should keep in mind that the KnERDI measures some, but certainly not all, aspects of reading that teachers should have opportunities to learn.</p> <p>As reviewed in previous sections, knowledge inventories have been used to measure how educators' knowledge of concepts such as phonological awareness and decoding relates to their instructional practice. The KnERDI can be used in future studies to illuminate how knowledge across a broader range of concepts might relate to various instructional decisions and student outcomes in reading. Also, it would be useful in future research to examine how educators fare on the subscales of the KnERDI relative to other similar measurement tools. This research strategy could establish evidence of concurrent validity of the instrument, but more importantly, it could offer additional insights about the structure of educators' knowledge. For instance, researchers could administer multiple instruments to teachers and examine the dimensions of teacher knowledge that might not be revealed on any individual instrument. This is especially needed because our findings suggest the facets we proposed in our meaning and connected text subscale provide only a partial picture of this domain. There is still a need for additional theory and research to better characterize knowledge related to text comprehension and other lesser constrained competencies in reading.</p> <hd id="AN0157693737-24">Limitations</hd> <p>This study has important limitations that should be considered. First, the educators who completed the KnERDI are not a representative sample of all educators. They were drawn from participants in an online course who were seeking out additional professional development, and may best represent teachers who chose to pursue learning opportunities in evidence‐based reading instruction. Also, the validation of our instrument was limited to content validity through expert review and construct validity with respect to dimensionality only. We do not yet know how our instrument aligns with other similar measures or how it might be predictive of effective classroom instruction. Our method does not support any direct comparisons between performance on our instrument compared to other instruments reported in the literature. Differences could be due to many factors, including the relative difficulty of items used on the instruments.</p> <p>The conclusions drawn about teacher knowledge in this study are limited to a relatively narrow range of knowledge related to literacy instruction that can be assessed in an inventory format. Any conclusions drawn from this study about how much teachers know or do not know should be made with caution. While it has become common in the literature to use knowledge inventories to provide evidence that teachers do not know enough, research has not determined a precise cutoff of required knowledge for effective teaching. It is not appropriate to say, for example, that less than 70% correct on a particular domain demonstrates limited or poor knowledge, when our field has not come to a firm understanding of how to benchmark knowledge for effective practice. Similar to the ways in which academic proficiency benchmarks for children are socially and often arbitrarily constructed, we acknowledge the challenge and limitations of creating such benchmarks for teachers, when so many other aspects of knowledge are required for effective teaching.</p> <hd id="AN0157693737-25">Conclusion</hd> <p>Kozak and Martin‐Chang, 2019 argue that "teachers' knowledge is the binding that connects the various domains of reading instruction into one compelling story" (p. 330). We speculate that increasing educators' knowledge will continue to play a significant role in the conversation around efforts to improve early reading instruction. If researchers, teacher educators, and other stakeholders are likely to continue to measure educator knowledge using inventories with multiple‐choice formats for convenience and ease of measurement, it is critical that they use instruments that reliably reflect the domains that our field has generally agreed upon are pivotal to early reading instruction. We believe that the KnERDI comes closer than many other knowledge instruments to capturing the complex understanding of higher‐level aspects of language structure required for early elementary teachers to help all students develop into confident and capable readers.</p> <ref id="AN0157693737-26"> <title> Footnotes </title> <blist> <bibl id="bib1" idref="ref1" type="bt">1</bibl> <bibtext> The internal consistency reliability indices for these facet scores were as follows in Form A: Decoding,.69; PA,.49; Fluency,.44; Morphology,.47; Syntax,.32; Text structure,.44; Vocabulary,.45; and Mental modeling,.53. In Form B, reliabilities were: Decoding,.68; PA,.50; Fluency,.44; Morphology,.44; Syntax,.30; Text structure,.30; Vocabulary,.38; and Mental modeling,.57. The instrument was not designed to measure these specific facets with reliability. On an exploratory basis, to evaluate the implications of the prespecification of the facet scores and their low reliabilities on the model, we also ran a CFA model in which the individual items loaded on the two latent factors (for Form A only). We have reported and depicted the CFA model with facet scores because it is easier to interpret, but the model with individual items produced similar results. Thirty‐seven of the 41 items had factor loadings above 0.40. The model showed reasonably good fit (SRMR =.049; RMSEA =.069; CFI =.977; AGFI =.982). The correlation between latent constructs was.90.</bibtext> </blist> <blist> <bibl id="bib2" idref="ref2" type="bt">2</bibl> <bibtext> Our purpose was to test a specific theory‐driven model, not to iteratively identify the model with maximum model fit. The two modification indices suggested for the model were: 1) add error covariances between PA and text structure; and 2) add an additional path from text structure to the alphabetic code dimension. Neither modification made meaningful changes in model fit or theoretical interpretation compared to the original model.</bibtext> </blist> </ref> <ref id="AN0157693737-27"> <title> References </title> <blist> <bibtext> Al Otaiba, S., Baker, K., Lan, P., Allor, J., Rivas, B., Yovanoff, P., & Kamata, A. (2019). Elementary teacher's knowledge of response to intervention implementation: A preliminary factor analysis. Annals of Dyslexia, 69 (1), 34 – 53.</bibtext> </blist> <blist> <bibtext> Arrow, A.W., Braid, C., & Chapman, J.W. (2019). Explicit linguistic knowledge is necessary, but not sufficient, for the provision of explicit early literacy instruction. Annals of Dyslexia, 69 (1), 99 – 113. https://doi.org/10.1007/s11881‐018‐00168‐0</bibtext> </blist> <blist> <bibl id="bib3" type="bt">3</bibl> <bibtext> Ball, D.L., Thames, M.H., & Phelps, G. (2008). Content knowledge for teaching: What makes it special? Journal of Teacher Education, 59, 389 – 407. https://doi.org/10.1177/0022487108324554</bibtext> </blist> <blist> <bibl id="bib4" idref="ref12" type="bt">4</bibl> <bibtext> Binks‐Cantrell, E., Joshi, R.M., & Washburn, E.K. (2012). Validation of an instrument for assessing teacher knowledge of basic language constructs of literacy. Annals of Dyslexia, 62 (3), 153 – 171. https://doi.org/10.1007/s11881‐012‐0070‐8</bibtext> </blist> <blist> <bibl id="bib5" type="bt">5</bibl> <bibtext> Bos, C., Mather, N., Dickson, S., Podhajski, B., & Chard, D. (2001). Perceptions and knowledge of preservice and inservice educators about early reading instruction. Annals of Dyslexia, 51 (1), 97 – 120. https://doi.org/10.1007/s11881‐001‐0007‐0</bibtext> </blist> <blist> <bibl id="bib6" type="bt">6</bibl> <bibtext> Brady, S., Gillis, M., Smith, T., Lavalette, M., Liss‐Bronstein, L., Lowe, E., North, W., Russo, E., & Wilder, T.D. (2009). First grade teachers' knowledge of phonological awareness and code concepts: Examining gains from an intensive form of professional development and corresponding teacher attitudes. Reading & Writing, 22 (4), 425 – 455. https://doi.org/10.1007/s11145‐009‐9166‐x</bibtext> </blist> <blist> <bibl id="bib7" type="bt">7</bibl> <bibtext> Brown, T.A. (2006). Confirmatory factor analysis for applied research. Guilford Press.</bibtext> </blist> <blist> <bibl id="bib8" type="bt">8</bibl> <bibtext> Campbell, P.F., Nishio, M., Smith, T.M., Clark, L.M., Conant, D.L., Rust, A.H., DePiper, J.N., Frank, T.J., Griffin, M.J., & Choi, Y. (2014). The relationship between teachers' mathematical content and pedagogical knowledge, teachers' perceptions, and student achievement. Journal for Research in Mathematics Education, 45 (4), 419 – 459.</bibtext> </blist> <blist> <bibl id="bib9" type="bt">9</bibl> <bibtext> Carlisle, J.F., Kelcey, B., Rowan, B., & Phelps, G. (2011). Teachers' knowledge about early reading: Effects on students' gains in reading achievement. Journal of Research on Educational Effectiveness, 4 (4), 289 – 321. https://doi.org/10.1080/19345747.2010.539297</bibtext> </blist> <blist> <bibtext> Carter, S.R. (2016). Using confirmatory factor analysis to manage discriminant validity issues in social pharmacy research. International Journal of Clinical Pharmacy, 38 (3), 731 – 737. https://doi.org/10.1007/s11096‐016‐0302‐9</bibtext> </blist> <blist> <bibtext> Castles, A., Rastle, K., & Nation, K. (2018). Ending the reading wars: Reading acquisition from novice to expert. Psychological Science in the Public Interest, 19 (1), 5 – 51. https://doi.org/10.1177/1529100618772271</bibtext> </blist> <blist> <bibtext> Cheesman, E.A., McGuire, J.M., Shankweiler, D., & Coyne, M. (2009). First‐year teacher knowledge of phonemic awareness and its instruction. Teacher Education and Special Education, 32 (3), 270 – 289.</bibtext> </blist> <blist> <bibtext> Clemens, N.H., & Fuchs, D. (2021). Commercially developed tests of reading comprehension: Gold standard or fool's good? Reading Research Quarterly. https://doi.org/10.1002/rrq.415</bibtext> </blist> <blist> <bibtext> Compton‐Lilly, C.F., Mitra, A., Guay, M., & Spence, L.K. (2020). A confluence of complexity: Intersections among reading theory, neuroscience, and observations of young readers. Reading Research Quarterly, 55 (S1), S185 – S195. https://doi.org/10.1002/rrq.348</bibtext> </blist> <blist> <bibtext> Cunningham, A.E., Perry, K.E., Stanovich, K.E., & Stanovich, P.J. (2004). Disciplinary knowledge of K‐3 teachers and their knowledge calibration in the domain of early literacy. Annals of Dyslexia, 54 (1), 139 – 167. https://doi.org/10.1007/s11881‐004‐0007‐y</bibtext> </blist> <blist> <bibtext> Davis, D.S., Jones, J.S., & Samuelson, C. (2021). Is it time for a hard conversation about cueing systems and word reading in teacher education? Reading and Writing Quarterly, 37 (4), 301 – 316. https://doi.org/10.1080/10573569.2020.1792813</bibtext> </blist> <blist> <bibtext> Davis, D.S., McElhone, D., & Tenore, F.B. (2015). A dialogic account of reader‐text interactions. English Teaching: Practice & Critique, 14 (3), 335 – 349. https://doi.org/10.1108/ETPC‐04‐2015‐0026</bibtext> </blist> <blist> <bibtext> Duke, N.K., & Block, M.K. (2012). Improving reading in the primary grades. The Future of Children, 22(2), 55 – 72.</bibtext> </blist> <blist> <bibtext> Duke, N.K., Ward, A.E., & Pearson, P.D. (2021). The science of reading comprehension instruction. The Reading Teacher, 74 (6), 663 – 672.</bibtext> </blist> <blist> <bibtext> Ehri, L. (2017). Orthographic mapping and literacy development revisited. In K. Cain, D.L. Compton, & R.K. Parrila (Eds.), Theories of reading development (pp. 127 – 146). John Benjamins.</bibtext> </blist> <blist> <bibtext> Folsom, J.S., Smith, K.G., Burk, K., & Oakley, N. (2017). Educator outcomes associated with implementation of Mississippi's K‐3 Early Literacy Professional Development Initiative. REL 2017‐270. Regional Educational Laboratory Southeast.</bibtext> </blist> <blist> <bibtext> Foorman, B., Beyler, N., Borradaile, K., Coyne, M., Denton, C.A., Dimino, J., Furgeson, J., Hayes, L., Henke, J., Justice, L., Keating, B., Lewis, W., Sattar, S., Streke, A., Wagner, R., & Wissel, S. (2016). Foundational skills to support reading for understanding in kindergarten through 3rd grade (NCEE 2016‐4008). National Center for Education Evaluation and Regional Assistance (NCEE), Institute of Education Sciences, U.S. Department of Education. https://ies.ed.gov/ncee/wwc/Docs/PracticeGuide/wwc_foundationalreading_040717.pdf</bibtext> </blist> <blist> <bibtext> Foorman, B., Petscher, Y., & Herrera, S. (2018). Unique and common effects of decoding and language factors in predicting reading comprehension in grades 1–10. Learning and Individual Differences, 63, 12 – 23. https://doi.org/10.1016/j.lindif.2018.02.011</bibtext> </blist> <blist> <bibtext> Fornell, C., & Larcker, D.F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18 (1), 39 – 50. https://doi.org/10.1177/002224378101800104</bibtext> </blist> <blist> <bibtext> Garet, M.S., Cronen, S., Eaton, M., Kurki, A., Ludwig, M., Jones, W., Uekawa, K., Falk, A., Bloom, H., Doolittle, F., Zhu, P., & Sztejnberg, L. (2008). The impact of two professional development interventions on early reading instruction and achievement (NCEE 2008‐4030). National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. https://ies.ed.gov/ncee/pdf/20084030.pdf</bibtext> </blist> <blist> <bibtext> Gavelek, J.R., & Whittingham, C.E. (2017). Meaning making in the 21st century: The sociogenesis of reading comprehension. In S.E. Israel (Ed.) Handbook of research on reading comprehension (2nd ed., pp. 166 – 190). Guilford.</bibtext> </blist> <blist> <bibtext> Goodwin, A.P., & Ahn, S. (2013). A meta‐analysis of morphological interventions in English: Effects on literacy outcomes for school‐age children. Scientific Studies of Reading, 17 (4), 257 – 285. https://doi.org/10.1080/10888438.2012.689791</bibtext> </blist> <blist> <bibtext> Gough, P.B., & Tunmer, W.E. (1986). Decoding, reading, and reading disability. Remedial and Special Education, 7 (1), 6 – 10.</bibtext> </blist> <blist> <bibtext> Hill, H.C., Ball, D.L., & Schilling, S.G. (2008). Unpacking pedagogical content knowledge: Conceptualizing and measuring teachers' topic‐specific knowledge of students. Journal for Research in Mathematics Education, 39 (4), 372 – 400.</bibtext> </blist> <blist> <bibtext> Hill, H.C., Rowan, B., & Ball, D.L. (2005). Effects of teachers' mathematical knowledge for teaching on student achievement. American Educational Research Journal, 42 (2), 371 – 406.</bibtext> </blist> <blist> <bibtext> Hindman, A.H., Morrison, F.J., Connor, C.M., & Connor, J.A. (2020). Bringing the science of reading to preservice elementary teachers: Tools that bridge research and practice. Reading Research Quarterly, 55 (S1), S197 – S206.</bibtext> </blist> <blist> <bibtext> Hu, L., & Bentler, P.M. (1999). Cut‐off criteria for fit indexes in co‐variance structural analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6 (1), 1 – 55.</bibtext> </blist> <blist> <bibtext> Jordan, R.L.P., & Bratsch‐Hines, M. (2020). Associations of reading knowledge with kindergarten and first grade teachers' reported instructional strategies. Literacy Research and Instruction, 59(4), 277–297, https://doi.org/10.1080/19388071.2020.1774689</bibtext> </blist> <blist> <bibtext> Jordan, R.L.P., Bratsch‐Hines, M., & Vernon‐Feagans, L. (2018). Kindergarten and first grade teachers' content and pedagogical knowledge of reading and associations with teacher characteristics at rural low‐wealth schools. Teaching and Teacher Education, 74, 190 – 204.</bibtext> </blist> <blist> <bibtext> Kelcey, B. (2011). Assessing the effects of teachers' reading knowledge on students' achievement using multilevel propensity score stratification. Educational Evaluation and Policy Analysis, 33 (4), 458 – 482.</bibtext> </blist> <blist> <bibtext> Kendeou, P., McMaster, K.L., & Christ, T.J. (2016). Reading comprehension: Core components and processes. Policy Insights from the Behavioral and Brain Sciences, 3 (1), 62 – 69. https://doi.org/10.1177/2372732215624707</bibtext> </blist> <blist> <bibtext> Kintsch, W. (2019). Revisiting the construction‐integration model of text comprehension and its implications for instruction. In D.E. Alvermann, N.J. Unrau, M. Sailors, & R.B. Ruddell (Eds.), Theoretical models and processes of literacy (7th ed., pp. 178 – 203). Routledge.</bibtext> </blist> <blist> <bibtext> Kozak, S., & Martin‐Chang, S. (2019). Preservice teacher knowledge, print exposure, and planning for instruction. Reading Research Quarterly, 54 (3), 323 – 338.</bibtext> </blist> <blist> <bibtext> Kucan, L., Palincsar, A.S., Khasnabis, D., & Chang, C.I. (2009). The Video Viewing Task: A source of information for assessing and addressing teacher understanding of text‐based discussion. Teaching and Teacher Education, 25 (3), 415 – 423.</bibtext> </blist> <blist> <bibtext> Language and Reading Research Consortium (LARRC), Jiang, H., & Logan, J. (2019). Improving reading comprehension in the primary grades: Mediated effects of a language‐focused classroom intervention. Journal of Speech, Language, and Hearing Research, 62 (8), 2812 – 2828. https://doi.org/10.1044/2019_JSLHR‐L‐19‐001</bibtext> </blist> <blist> <bibtext> McCutchen, D., Green, L., Abbot, R.D., & Sanders, E.A. (2009). Further evidence for teacher knowledge: Supporting struggling readers in grades three through five. Reading & Writing, 22, 401 – 423.</bibtext> </blist> <blist> <bibtext> McCutchen, D., Harry, D.R., Cunningham, A.E., Cox, S., Sidman, S., & Covill, A.E. (2002). Reading teachers' knowledge of children's literature and English phonology. Annals of Dyslexia, 52 (1), 207 – 228. https://doi.org/10.1007/s11881‐002‐0013‐x</bibtext> </blist> <blist> <bibtext> McElhone, D., Tenore, F.B., & Davis, D.S. (2017). How teachers mobilize and transform their conceptualizations of reading comprehension into representations of instructional practice. Literacy Research: Theory, Method, and Practice, 66 (1), 232 – 247. https://doi.org/10.1177/2381336917718502</bibtext> </blist> <blist> <bibtext> McMahan, K.M., Oslund, E.L., & Odegard, T.N. (2019). Characterizing the knowledge of educators receiving training in systematic literacy instruction. Annals of Dyslexia, 69 (1), 21 – 33.</bibtext> </blist> <blist> <bibtext> McMaster, K.L., van den Broek, P., Espin, C. A., White, M.J., Rapp, D.N., Kendeou, P., Bohn‐Gettler, C.M., & Carlson, S. (2012). Making the right connections: Differential effects of reading intervention for subgroups of comprehenders. Learning and Individual Differences, 22 (1), 100 – 111. https://doi.org/10.1016/j.lindif.2011.11.017</bibtext> </blist> <blist> <bibtext> Mesmer, H.A., & Rose‐McCully, M.M. (2018). A closer look at close reading: Three under‐the‐radar skills needed to comprehend sentences. The Reading Teacher, 71 (4), 451 – 461.</bibtext> </blist> <blist> <bibtext> Moats, L.C. (1994). The missing foundation in teacher education: Knowledge of the structure of spoken and written language. Annals of Dyslexia, 44, 81 – 102. https://doi.org/10.1007/BF02648156</bibtext> </blist> <blist> <bibtext> Moats, L.C., & Foorman, B.R. (2003). Measuring teachers' content knowledge of language and reading. Annals of Dyslexia, 53 (1), 23 – 45. https://doi.org/10.1007/s11881‐003‐0003‐7</bibtext> </blist> <blist> <bibtext> Mosley Wetzel, M., Skerrett, A., Maloch, B., Flores, T.T., Infante‐Sheridan, M., Murdter‐Atkinson, J., Godfrey, V.C., & Duffy, A. (2020). Resisting positionings of struggle in "science of teaching reading" discourse: Counterstories of teachers and teacher educators in Texas. Reading Research Quarterly, 55 (S1), S319 – S330. https://doi.org/10.1002/rrq.358</bibtext> </blist> <blist> <bibtext> National Institute of Child Health and Human Development (NICHD) (2000). Report of the National Reading Panel: Teaching children to read: An evidence‐based assessment of the scientific research literature on reading and its implications for reading instruction. National Institutes of Health.</bibtext> </blist> <blist> <bibtext> Paris, S.G. (2005). Reinterpreting the development of reading skills. Reading Research Quarterly, 40 (2), 184 – 202. https://doi.org/10.1598/RRQ.40.2.3</bibtext> </blist> <blist> <bibtext> Pearson, P.D., Palincsar, A.S., Biancarosa, G., & Berman, A.I. (Eds.). (2020). Reaping the rewards of the reading for understanding initiative. National Academy of Education.</bibtext> </blist> <blist> <bibtext> Pearson, P.D., Valencia, S.W., & Wixson, K. (2014). Complicating the world of reading assessment: Toward better assessments for better teaching. Theory into Practice, 53 (3), 236 – 246.</bibtext> </blist> <blist> <bibtext> Petscher, Y., Cabell, S.Q., Catts, H.W., Compton, D.L., Foorman, B.R., Hart, S.A., Lonigan, C.J., Phillips, B.M., Schatschneider, C., Steacy, L.M., Terry, N.P., & Wagner, R.K. (2020). How the science of reading informs 21 st ‐Century education. Reading Research Quarterly, 55 (S1), S267 – S282. https://doi.org/10.1002/rrq.352</bibtext> </blist> <blist> <bibtext> Phelps, G. (2009). Just knowing how to read isn't enough! Assessing knowledge for teaching reading. Educational Assessment, Evaluation and Accountability, 21 (2), 137 – 154. https://doi.org/10.1007/s11092‐009‐9070‐6</bibtext> </blist> <blist> <bibtext> Phelps, G., & Schilling, S. (2004). Developing measures of content knowledge for teaching reading. The Elementary School Journal, 105 (1), 31 – 48. https://doi.org/10.1086/428764</bibtext> </blist> <blist> <bibtext> Phillips, B.M., Oliver, F., Tabulda, G., Wood, C., & Funari, C. (2020). Preschool teachers' language and vocabulary knowledge: Development and predictive associations for a new measure. Dyslexia, 26, 153 – 172.</bibtext> </blist> <blist> <bibtext> Piasta, S.B., Connor, C.M., Fishman, B.J., & Morrison, F.J. (2009). Teachers' knowledge of literacy concepts, classroom practices, and student reading growth. Scientific Studies of Reading, 13, 224 – 248. https://doi.org/10.1080/10888430902851364</bibtext> </blist> <blist> <bibtext> Piasta, S.B., Park, S., Farley, K.S., Justice, L.M., & O'Connell, A.A. (2020). Early childhood educators' knowledge about language and literacy: Associations with practice and children's learning. Dyslexia, 26 (2), 137 – 152. https://doi.org/10.1002/dys.1612</bibtext> </blist> <blist> <bibtext> Piasta, S.B., Soto Ramirez, P., Farley, K.S., Justice, L.M., & Park, S. (2020). Exploring the nature of associations between educators' knowledge and their emergent literacy classroom practices. Reading & Writing, 33 (6), 1399 – 1422. https://doi.org/10.1007/s11145‐019‐10013‐4</bibtext> </blist> <blist> <bibtext> Pittman, R.T., Zhang, S., Binks‐Cantrell, E., Hudson, A., & Joshi, R.M. (2020). Teachers' knowledge about language constructs related to literacy skills and student achievement in low socio‐economic status schools. Dyslexia, 26 (2), 200 – 219.</bibtext> </blist> <blist> <bibtext> Quinn, J.M., Wagner, R.K., Petscher, Y., & Lopez, D. (2015). Developmental relations between vocabulary knowledge and reading comprehension: A latent change score modeling study. Child Development, 86 (1), 159 – 175.</bibtext> </blist> <blist> <bibtext> Reutzel, D.R., Dole, J.A., Read, S., Fawson, P., Herman, K., Jones, C.D., Sudweeks, R., & Fargo, J. (2011). Conceptually and methodologically vexing issues in teacher knowledge assessment. Reading & Writing Quarterly, 27 (3), 183 – 211. https://doi.org/10.1080/10573569.2011.560098</bibtext> </blist> <blist> <bibtext> Roehling, J.V., Hebert, M., Nelson, J.R., & Bohaty, J.J. (2017). Text structure strategies for improving expository reading comprehension. The Reading Teacher, 71 (1), 71 – 82. https://doi.org/10.1002/trtr.1590</bibtext> </blist> <blist> <bibtext> Schoen, R.C., Bray, W., Wolfe, C., Tazaz, A., & Nielsen, L. (2017). Developing an assessment instrument to measure early elementary teachers' mathematical knowledge for teaching. The Elementary School Journal, 118 (1), 55 – 81.</bibtext> </blist> <blist> <bibtext> Shanahan, T., Callison, K., Carriere, C., Duke, N.K., Pearson, P.D., Schatschneider, C., & Torgesen, J. (2010). Improving reading comprehension in kindergarten through 3rd grade: A practice guide (NCEE 2010‐4038). National Center for Education Evaluation and Regional Assistance, Institute of Education Sciences, U.S. Department of Education. <ulink href="http://whatworks.ed.gov/publications/practiceguides">http://whatworks.ed.gov/publications/practiceguides</ulink></bibtext> </blist> <blist> <bibtext> Shulman, L.S. (1986). Those who understand: Knowledge growth in teaching. Educational Researcher, 15 (2), 4 – 14. https://doi.org/10.3102/0013189X015002004</bibtext> </blist> <blist> <bibtext> Spear‐Swerling, L., & Cheesman, E. (2012). Teachers' knowledge base for implementing response‐to‐intervention models in reading. Reading and Writing, 25 (7), 1691 – 1723. https://doi.org/10.1007/s11145‐011‐9338‐3</bibtext> </blist> <blist> <bibtext> Spear‐Swerling, L., & Zibulsky, J. (2014). Making time for literacy: teacher knowledge and time allocation in instructional planning. Reading and Writing, 27 (8), 1353 – 1378. https://doi.org/10.1007/s11145‐013‐9491‐y</bibtext> </blist> <blist> <bibtext> Washburn, E.K., Joshi, R.M., & Binks‐Cantrell, E. (2011a). Are preservice teachers prepared to teach struggling readers? Annals of dyslexia, 61 (1), 21 – 43.</bibtext> </blist> <blist> <bibtext> Washburn, E.K., Joshi, R.M., & Binks‐Cantrell, E.S. (2011b). Teacher knowledge of basic language concepts and dyslexia. Dyslexia, 17 (2), 165 – 183. https://doi.org/10.1002/dys.42</bibtext> </blist> <blist> <bibtext> Williams, J.P., Pollini, S., Nubla‐Kung, A.M., Snyder, A.E., Garcia, A., Ordynans, J.G., & Atkins, J.G. (2014). An intervention to improve comprehension of cause/effect through expository text structure instruction. Journal of Educational Psychology, 106 (1), 1 – 17. https://doi.org/10.1037/a0033215</bibtext> </blist> </ref> <aug> <p>By Dennis S. Davis; Courtney Samuelson; Jill Grifenhagen; Robyn DeIaco and Jackie Relyea</p> <p>Reported by Author; Author; Author; Author; Author</p> <p></p> <p>DENNIS S. DAVIS (corresponding author) is an associate professor in the Department of Teacher Education and Learning Sciences at North Carolina State University, Raleigh, NC, USA; email.</p> <p>COURTNEY SAMUELSON is a doctoral student in literacy and English language arts education (LELA) at North Carolina State University, Raleigh, NC, USA; email.</p> <p>JILL GRIFENHAGEN is an assistant professor in the Department of Teacher Education and Learning Sciences at North Carolina State University, Raleigh, NC, USA; email.</p> <p>ROBYN DEIACO is a doctoral candidate in literacy and English language arts education (LELA) at North Carolina State University, Raleigh, NC, USA; email.</p> <p>JACKIE RELYEA is an assistant professor in the Department of Teacher Education and Learning Sciences at North Carolina State University, Raleigh, NC, USA; email.</p> </aug> <nolink nlid="nl1" bibid="bib140" firstref="ref23"></nolink>
Header DbId: eric
DbLabel: ERIC
An: EJ1340023
AccessLevel: 3
PubType: Academic Journal
PubTypeId: academicJournal
PreciseRelevancyScore: 0
IllustrationInfo
Items – Name: Title
  Label: Title
  Group: Ti
  Data: Getting KnERDI with Language: Examining Teachers' Knowledge for Enhancing Reading Development in Code-Based and Meaning-Based Domains
– Name: Language
  Label: Language
  Group: Lang
  Data: English
– Name: Author
  Label: Authors
  Group: Au
  Data: <searchLink fieldCode="AR" term="%22Davis%2C+Dennis+S%2E%22">Davis, Dennis S.</searchLink><br /><searchLink fieldCode="AR" term="%22Samuelson%2C+Courtney%22">Samuelson, Courtney</searchLink><br /><searchLink fieldCode="AR" term="%22Grifenhagen%2C+Jill%22">Grifenhagen, Jill</searchLink><br /><searchLink fieldCode="AR" term="%22DeIaco%2C+Robyn%22">DeIaco, Robyn</searchLink><br /><searchLink fieldCode="AR" term="%22Relyea%2C+Jackie%22">Relyea, Jackie</searchLink> (ORCID <externalLink term="https://orcid.org/0000-0002-7560-7136">0000-0002-7560-7136</externalLink>)
– Name: TitleSource
  Label: Source
  Group: Src
  Data: <searchLink fieldCode="SO" term="%22Reading+Research+Quarterly%22"><i>Reading Research Quarterly</i></searchLink>. Jul-Sep 2022 57(3):781-804.
– 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: 24
– Name: DatePubCY
  Label: Publication Date
  Group: Date
  Data: 2022
– Name: TypeDocument
  Label: Document Type
  Group: TypDoc
  Data: Journal Articles<br />Reports - Research
– Name: Audience
  Label: Education Level
  Group: Audnce
  Data: <searchLink fieldCode="EL" term="%22Elementary+Education%22">Elementary Education</searchLink>
– Name: Subject
  Label: Descriptors
  Group: Su
  Data: <searchLink fieldCode="DE" term="%22Reading+Instruction%22">Reading Instruction</searchLink><br /><searchLink fieldCode="DE" term="%22Elementary+Education%22">Elementary Education</searchLink><br /><searchLink fieldCode="DE" term="%22Test+Reliability%22">Test Reliability</searchLink><br /><searchLink fieldCode="DE" term="%22Test+Validity%22">Test Validity</searchLink><br /><searchLink fieldCode="DE" term="%22Teacher+Effectiveness%22">Teacher Effectiveness</searchLink><br /><searchLink fieldCode="DE" term="%22Teacher+Evaluation%22">Teacher Evaluation</searchLink><br /><searchLink fieldCode="DE" term="%22Teacher+Characteristics%22">Teacher Characteristics</searchLink><br /><searchLink fieldCode="DE" term="%22Elementary+School+Teachers%22">Elementary School Teachers</searchLink><br /><searchLink fieldCode="DE" term="%22Knowledge+Level%22">Knowledge Level</searchLink>
– Name: DOI
  Label: DOI
  Group: ID
  Data: 10.1002/rrq.445
– Name: ISSN
  Label: ISSN
  Group: ISSN
  Data: 0034-0553
– Name: Abstract
  Label: Abstract
  Group: Ab
  Data: Much of the early research on teachers' knowledge for reading instruction used instruments that primarily emphasized code-based aspects of reading, unintentionally signaling that meaning-focused knowledge is not essential for teaching foundational reading to elementary-age children. In this study, we developed the Knowledge for Enhancing Reading Development Inventory (KnERDI; pronounced 'nerdy'), an instrument that measures teachers' knowledge in two domains: alphabetic code/word reading and meaning/connected text processes. Using this instrument, we sought to determine if teachers' knowledge of the two proposed domains was highly associated and if knowledge was related to level of education and amount of teaching experience. We report on the reliability and validity of the KnERDI and describe patterns in educators' performance on this instrument. We found that the KnERDI measured both domains with acceptable reliability in multiple samples of educators enrolled in an online professional development course. Educators scored higher on the meaning/connected text process subscale than on the alphabetic code/word reading subscale. The two subscales were strongly correlated, indicating that teachers' knowledge as measured on the KnERDI is a unidimensional construct. Advanced degree completion was not consistently related to educators' scores on the KnERDI, but there was a positive association between educators' performance on the instrument and their years of experience teaching early reading, particularly with respect to the code-based domain. We discuss potential uses of the KnERDI in future research and practice, positive trends in educators' performance on the instrument, and challenges in measuring educators' knowledge for supporting higher-level language and comprehension processes.
– Name: AbstractInfo
  Label: Abstractor
  Group: Ab
  Data: As Provided
– Name: DateEntry
  Label: Entry Date
  Group: Date
  Data: 2022
– Name: AN
  Label: Accession Number
  Group: ID
  Data: EJ1340023
PLink https://search.ebscohost.com/login.aspx?direct=true&site=eds-live&db=eric&AN=EJ1340023
RecordInfo BibRecord:
  BibEntity:
    Identifiers:
      – Type: doi
        Value: 10.1002/rrq.445
    Languages:
      – Text: English
    PhysicalDescription:
      Pagination:
        PageCount: 24
        StartPage: 781
    Subjects:
      – SubjectFull: Reading Instruction
        Type: general
      – SubjectFull: Elementary Education
        Type: general
      – SubjectFull: Test Reliability
        Type: general
      – SubjectFull: Test Validity
        Type: general
      – SubjectFull: Teacher Effectiveness
        Type: general
      – SubjectFull: Teacher Evaluation
        Type: general
      – SubjectFull: Teacher Characteristics
        Type: general
      – SubjectFull: Elementary School Teachers
        Type: general
      – SubjectFull: Knowledge Level
        Type: general
    Titles:
      – TitleFull: Getting KnERDI with Language: Examining Teachers' Knowledge for Enhancing Reading Development in Code-Based and Meaning-Based Domains
        Type: main
  BibRelationships:
    HasContributorRelationships:
      – PersonEntity:
          Name:
            NameFull: Davis, Dennis S.
      – PersonEntity:
          Name:
            NameFull: Samuelson, Courtney
      – PersonEntity:
          Name:
            NameFull: Grifenhagen, Jill
      – PersonEntity:
          Name:
            NameFull: DeIaco, Robyn
      – PersonEntity:
          Name:
            NameFull: Relyea, Jackie
    IsPartOfRelationships:
      – BibEntity:
          Dates:
            – D: 01
              M: 01
              Type: published
              Y: 2022
          Identifiers:
            – Type: issn-print
              Value: 0034-0553
          Numbering:
            – Type: volume
              Value: 57
            – Type: issue
              Value: 3
          Titles:
            – TitleFull: Reading Research Quarterly
              Type: main
ResultId 1