Reading fluency as the bridge between decoding and reading comprehension in Chinese children

Purpose Reading fluency has been considered an essential component of reading comprehension, but it is yet to be examined in a reading model in a non-alphabetic writing system. This study investigated whether reading fluency could be identified as a separate construct from decoding and examined the unique role of reading fluency in the Simple View of Reading (SVR). Method A total of 342 Cantonese-speaking Chinese children in grades 3–5 were recruited to participate in the study. They were assessed on word reading accuracy and fluency, morphological awareness, vocabulary knowledge, and reading comprehension. Results The confirmatory factor analysis results confirmed that reading fluency is a separate factor from decoding, linguistic comprehension, and reading comprehension. Furthermore, the structural equation modeling results revealed that reading fluency is a significant predictor of reading comprehension and a mediator between decoding and reading comprehension in the extended SVR model. Conclusion The findings extended previous research in alphabetic languages and supported reading fluency as the bridge between decoding and reading comprehension. The present study highlighted the importance of reading fluency in Chinese reading acquisition in a theoretical framework.


Introduction
e Simple View of Reading (SVR) states that reading comprehension is the product of two key factors: decoding (D), i.e., the ability to recognize and pronounce words, and linguistic comprehension or language comprehension (LC), i.e., understanding the meaning of words and sentences, (D x LC = RC; Gough and Tunmer, 1986;Hoover and Gough, 1990).While the SVR has been validated by over 150 studies across different languages with different orthographic depths (see Florit and Cain, 2011 for a meta-analysis), whether this model fully captures the components needed for successful reading comprehension remains to be determined (Sparks, 2021).Reading uency is conceptualized as the oral translation of the text with speed and accuracy (Fuchs et al., 2001), which is oen operationalized as 1-minute uency tasks requiring participants to read aloud a word list or a passage composed of high-frequency words.Based on the original SVR, studies in alphabetic languages have proposed the extended SVR, suggesting that in addition to Hsu et al. 10.3389/fpsyg.2023.1221396decoding and linguistic comprehension, reading uency may be a unique contributor to children's reading comprehension (Joshi and Aaron, 2000;Torgesen and Hudson, 2006;Hudson et al., 2008;Rasinski et al., 2011;Silverman et al., 2013;Torppa et al., 2016;Cadime et al., 2017;Kuhn and Schwanen ugel, 2019).However, the unique contributing role of reading uency in non-alphabetic languages, for example, Chinese, is yet to be examined.e present study examined the role of reading uency in reading comprehension in Chinese under the extended SVR framework.

Reading fluency as an additional component to the extended Simple View of Reading
e Simple View of Reading (SVR), i.e., D x LC = RC, suggests that decoding is a "bottom-up" skill-based component and linguistic comprehension is a "top-down" cognitive-based component in reading comprehension (Gough and Tunmer, 1986;Hoover and Gough, 1990).Linguistic comprehension involves understanding the meaning of words in contextualized sentences, which relies on the reader's prior semantic knowledge.Linguistic comprehension is usually measured using morphological awareness, oral vocabulary knowledge, and listening comprehension (Adlof et al., 2006;Tilstra et al., 2009), which is generally undisputed.Decoding emphasizes the accuracy and pattern recognition of letters and words.Studies, however, have employed different operationalizations of decoding in SVR.Some measured decoding using reading accuracy tasks in which participants are asked to read aloud a word list in the ascending order of difficulty levels without a time limit (Joshi and Aaron, 2000;Georgiou et al., 2009), while others have combined reading accuracy and uency measures to a single variable (Adlof et al., 2006;Kirby and Savage, 2008;Kershaw and Schatschneider, 2012).
Reading uency is an automatic-activation process, encompassing perceptual skills and a print-to-sound translation process, making it more complex than a pure decoding measure (Fuchs et al., 2001;Wolf and Katzir-Cohen, 2001).Researchers have suggested that uency might play a more prominent role in later grades in readers of opaque orthographies due to the demand for sound-print mapping as compared to readers of transparent orthographies (Wimmer et al., 1998;Aaron et al., 1999;Florit and Cain, 2011).According to the verbal efficiency theory (Perfetti, 1985), decoding must become automatic to allocate cognitive resources for reading comprehension.Reading uency has been found to signi cantly predict reading comprehension (Cutting and Scarborough, 2006;Silverman et al., 2013;Torppa et al., 2016).Silverman et al. (2013) examined the role of reading uency in a group of fourth-grade students in the United States.eir results showed that reading uency was a unique predictor of reading comprehension, which mediated the prediction of decoding and linguistic comprehension on reading comprehension.Torppa et al. (2016) andCadime et al. (2017) found similar results in a sample of Finnish kindergarteners to grade 3 and Portuguese grade 2 students, respectively.Studies conducted on Korean kindergarteners and grade 1 students have shown that reading uency is signi cantly associated with reading comprehension above and beyond decoding and linguistic comprehension in the extended SVR framework (Kim et al., 2014;Kim, 2015).In line with Florit and Cain's meta-analysis (2011), reading uency is a signi cant predictor of reading comprehension across transparent (Finnish: Torppa et al., 2016), intermediate (Portugal;Cadime et al., 2017), and opaque (English: Silverman et al., 2013) orthographies as well as languages in logographic shape (Korean: Kim et al., 2014).e Chinese language, however, lies in the further ends of the opaque orthographic consistency spectrum.Previous studies have not examined whether reading uency plays a unique role in reading comprehension under the extended SVR.e present study aimed to ll this literature gap by examining the role of reading uency in reading comprehension under the extended SVR in a sample of Chinese primary-grade students.
Furthermore, previous studies have identi ed the mediating role of reading uency in the extended SVR (Silverman et al., 2013;Torppa et al., 2016;Cadime et al., 2017).Based on the verbal efficiency theory (Perfetti, 1985), researchers examined whether reading uency mediates the relationship between decoding and reading comprehension.In the longitudinal study by Cadime et al. (2017) in a sample of Portuguese grade 2 students, the results showed a signi cant mediating role of reading uency between decoding and reading comprehension in grades 2 and 4.Moreover, Silverman et al. (2013) found that reading uency is a signi cant mediator between decoding, linguistic comprehension, and reading comprehension.e results supported the meta-analysis by Florit and Cain (2011) that orthographic transparency might in uence the role of reading uency in the extended SVR.In light of these diverse ndings, studies in Chinese have included reading uency in the SVR (Ho et al., 2017;Yan et al., 2021;Pan and Lin, 2022) in primarygrade children.None of them, however, have considered the unique mediating role of reading uency in the extended SVR.Despite growing evidence that reading uency might be a separate measure, reading uency has been consistently considered a measure of decoding.erefore, this study examined the factor structures of decoding and reading uency using con rmatory factor analysis and the unique predictive role of reading uency in the extended SVR using structural equation modeling.

Reading acquisition in Chinese
As a more opaque language, Chinese has a morpho-syllabic writing system (DeFrancis, 1984).Unlike alphabetic languages, Chinese characters (the smallest unit of Chinese words) are more visually complex.Each character represents a morpheme and syllable (DeFrancis, 1984;Hoosain, 1991).Over 80% of Chinese characters consist of semantic and phonetic radicals.Semantic radicals carry the meaning of the character, while the phonetic radical provides the phonetic cue of the character although the phonetic cue is far from reliable.e abundance of homophones and homographs in Chinese might further complicate the reading process.Chinese readers must attend to contextual information in sentences or passages to disambiguate the meaning of words.Moreover, the lack of boundaries between characters might put an excessive cognitive load on Chinese reading comprehension (Liu, 2014).Given the less reliance on phonetic cues in Chinese  -Chang et al., 2003;Ho et al., 2004;Yeung et al., 2013).Moreover, the learning and instruction of Chinese vary across different Chinese-speaking regions.Chinese readers from mainland China, whose official spoken language is Putonghua, learn to read Chinese through the pinyin phonetic coding system.However, no phonetic coding system is used by Chinese readers from Hong Kong, where Cantonese is the most common spoken language.erefore, Hong Kong Chinese readers rely more on rote memory and lexical retrieval efficiency to learn to read Chinese.Furthermore, traditional Chinese characters are used in Taiwan and Hong Kong, which are more visually complex than simpli ed Chinese characters used in China.e differences in Chinese languages between regions might in uence the relationship between cognitive-linguistic and literacy skills (McBride-Chang et al., 2005;McBride, 2015).e present study focused on Chinese children from Hong Kong who spoke Cantonese and wrote traditional Chinese characters as their mother tongue to examine the extended SVR in Chinese.

The Simple View of Reading in Chinese
e SVR with reading uency as a measure of decoding has been examined widely in Chinese, particularly in Cantonese-speaking children (Ho et al., 2012(Ho et al., , 2017;;Yeung et al., 2013;Pan and Lin, 2022).Yeung et al. (2013) examined the SVR in Chinese in a sample of grade 4 Cantonese-speaking children in Hong Kong.e results showed that reading comprehension was signi cantly predicted by decoding (measured by word reading accuracy) and syntactic skills (measured by word and sentence orders).Linguistic (measured by morphological awareness) and rapid naming did not signi cantly predict reading comprehension.Although rapid naming is a speeded measure, previous studies have shown that rapid naming is a signi cant predictor of reading uency, which measures the lexical retrival speed (Georgiou et al., 2016).Yeung et al. (2016) further examined the SVR in their longitudinal study of grade 1 Cantonese-speaking children in Hong Kong.Text-reading uency and word-reading accuracy were measured as the decoding variable.Linguistic comprehension was measured by oral vocabulary, narrative, and syntactic skills.e results showed that in line with SVR, decoding and linguistic comprehension were signi cant predictors of reading comprehension.However, the in uence of reading uency was not examined independently.Furthermore, Ho et al. (2017) examined the extended SVR by including rapid naming as a third predictor in a sample of grades 1 to 3 Cantonese-speaking children in Hong Kong.e con rmatory factor analysis showed a four-factor structure: rapid naming, decoding, linguistic comprehension, and reading comprehension, and that word-reading uency regressed on both rapid naming and decoding.e structural equation model results showed that decoding and linguistic comprehension but not rapid naming were signi cant predictors of reading comprehension.
Consistent with Yeung et al. 's (2013) ndings, rapid naming might not be the proximal predictor of reading comprehension; instead, rapid naming is possibly a distal predictor via decoding (Yeung et al., 2013;Ho et al., 2017).erefore, rapid naming was not included in this study.Since the role of reading uency in the extended SVR in Chinese is still underexplored, the present study aimed to examine the proximal variables of reading comprehension based on the extended SVR: decoding, linguistic comprehension, and reading uency.

The present study
Findings from previous studies in Chinese reading have suggested that decoding and linguistic comprehension play important roles in SVR.Despite the plethora of studies on Chinese reading, very few examined reading uency as an independent measure in a Chinese reading comprehension model.Furthermore, of those that included reading uency, the measure was only examined at either word (Ho et al., 2017) or text level (Yeung et al., 2016) and was submerged into decoding.erefore, the present study aimed to examine the unique role of reading uency, measured by word and text-level measures to understand Chinese reading comprehension better based on the extended SVR.
e rst aim of the present study was to investigate whether reading uency, measured by word-and text-reading uency, could be a separate construct from decoding using con rmatory factor analysis (CFA).Some studies have considered reading uency a separate variable from decoding (Torgesen and Hudson, 2006;Silverman et al., 2013).On the other hand, some research studies have considered reading uency to measure decoding skills (Yeung et al., 2016;Ho et al., 2017).Based on the verbal efficiency theory (Perfetti, 1985), we expected that reading uency would be a separate factor from decoding, such that a four-factor model, reading uency, decoding, linguistic comprehension, and reading comprehension, would t the data better than a three-factor model, decoding + reading uency, linguistic comprehension, and reading comprehension.
e second aim was to elucidate the unique role of reading uency in the extended SVR model in Chinese reading through structural equation modeling (SEM) analysis.Guided by previous studies (Cutting and Scarborough, 2006;Silverman et al., 2013;Torppa et al., 2016), we expected that reading uency would be a unique predictor of reading comprehension.Moreover, reading uency would signi cantly mediate the relationship between decoding and reading comprehension.By achieving these objectives, we expect to understand better the unique role of reading uency in the extended SVR in Chinese reading.

Procedure and materials
Written informed consent was obtained from the parents of the children before proceeding with the study.All measures were administered to individual children in a 45-min session by trained research assistants in a quiet room at the schools.

Word-reading fluency
Modeled aer Ho et al. 's (2000) study, this task assessed children's reading speed of context-free words.Participants were asked to read aloud a wordlist of 110 high-frequency Chinese twocharacter words arranged in ascending order of difficulty as fast and as accurately as possible in 1 min.ese words were selected from the Leixcal Lists for Chinese Learning provided the Hong Kong's Education Bureau for children in grades 1 to 3. One mark was awarded for each correctly read word.e reliability was calculated based on Gulliksen's formula for a time-limited task, r lower =0.76 (Chan and Yeung, 2020).

Text-reading fluency
Modeled aer Ho et al. 's (2000) study, this task assessed children's reading speed of connected text.Participants were asked to read aloud a passage of 310 Chinese characters as fast and accurately as possible in 1 min.ese words were selected from the Lexical Lists for Chinese Learning for children in grades 1 to 3. One mark was awarded for each correctly read character.e reliability was calculated based on Gulliksen's formula for a time-limited task, r lower = 0.90 (Chan and Yeung, 2020).

Character reading
e character reading task assessed children's decoding skills of single Chinese characters.Children were required to read aloud a wordlist of 60 Chinese characters arranged in ascending order of difficulty.e characters were selected from the Lexical Lists for Chinese Learning for children in grades 2 to 6.One mark was awarded to each correctly read character.e maximum possible score was 60, and Cronbach's alpha was 0.94.

Word reading
Modeled aer Ho et al. 's (2000) study, this task assessed children's decoding skills of Chinese words.Participants were required to read aloud a wordlist of 150 Chinese two-character words arranged in ascending order of difficulty.ese words were selected from the Lexical Lists for Chinese Learning for children in grades 2 to 6 and were non-existent in the character reading task.One mark was awarded to each correctly read a word.e maximum possible score was 150, and Cronbach's alpha was 0.96.

Vocabulary knowledge
is task was modi ed aer McBride-Chang et al. 's (2008) study and assessed children's vocabulary depth and breath.For each trial, participants were required to provide an oral explanation of a Chinese word, for example, "公平" (fair), an explanation would be "平等" (equal) or "公正" (just).is task comprised 16 Chinese words selected from the Lexical Lists for Chinese Learning.Two points were used to score participants' answers based on the clarity of explanations.e maximum possible score was 32, and Cronbach's alpha was 0.70.

Morphological awareness
Modeled aer Liu and McBride-Chang's (2010) study, this task assessed children's ability to morpheme a construction.For each trial, participants were required to construct a new Chinese word based on a two-sentence scenario, for example, "用手開嘅槍叫手槍, 用腳開嘅槍我哋會點叫佢呀?"(A gun that is triggered by hand is called a handgun.If there is a gun that is triggered by foot, what do we call it?),and the correct answer was "腳槍" (foot gun).is task consisted of 38 items.One mark was awarded for each correctly produced word.e maximum possible score was 38, and Cronbach's alpha was 0.71.

Sentence comprehension
is task assessed children's sentence comprehension skills.Participants were required to read a sentence silently and answer one multiple-choice question.is task consisted of 20 sentences and multiple-choice questions.One point was awarded for each correct answer.e maximum possible score was 20, and Cronbach's alpha was 0.76.

Passage comprehension
e passage comprehension task was used to assess children's reading comprehension skills.Participants were required to read two narrative passages silently and answer six multiple-choice questions in each passage.ese questions were designed to focus on (1) retrieval of information, (2) making direct inferences, (3) interpreting and integrating presented information; and (4) evaluating content, language, and contextual components, based on the Progress in International Reading Literacy Study (Mullis et al., 2007).One point was awarded for each correct answer.e maximum possible score was 12, and Cronbach's alpha was 0.67.

Raven's standard progressive matrices
Raven's standard progressive matrices were used as an intelligence assessment (Raven, 2003).For each item, children were presented with a visual stimulus with a missing part and were required to select the best matching piece out of six or eight options.is task consisted of 60 items.One mark was awarded for each correct answer.

Data analysis
e full information maximum likelihood (FIML) was used to handle missing data (1.5% in this study).In addition, descriptive statistics and correlation aer controlling for the grade level and IQ were performed for all measures.To examine whether reading uency is a separate construct from decoding, a CFA using maximum likelihood was calculated to examine the three-factor (Model 1A) and four-factor (Model 1B) structure of reading uency, decoding, linguistic comprehension, and reading comprehension.e original SVR model was tested to examine the validity of the SVR model in Chinese reading (Model 2).Moreover, three SEMs using maximum likelihood were conducted to examine the role of reading uency in the SVR model in Chinese reading.Model 3A examined the unique role of reading uency, Model 3B examined the mediating role of reading uency between decoding and reading comprehension, and Model 3C examined the indirect effect of reading uency on reading comprehension via decoding.Furthermore, moderation analyses were conducted to examine whether the relationship between reading uency and reading comprehension was moderated by the grade level and gender of participants.
Model ts were determined using multiple indices: χ 2 test, root mean square of approximation (RMSEA), standard root mean residual (SRMR), comparative t index (CFI), and Akaike information criterion (AIC).A non-signi cant χ 2 value, RMSEA < 0.08, SRMR < 0.06, and CFI > 0.95, indicated a good t to the data (Hu and Bentler, 1999).Model ts were compared using χ 2 tests and AIC values, where a 2-unit less in AIC values indicated better model ts.e indirect and moderating effects were examined using bias-corrected bootstrap with 2,000 resamples.e signi cance of indirect effects was determined using pvalues and 95% con dence intervals (CIs), where the 95% CI not passing zero indicates a signi cant indirect effect (Preacher and Hayes, 2008;Memon et al., 2018).

Descriptive statistics and correlations
e means, standard deviations, and correlations aer controlling for grade levels and IQ for all measures are shown in Table 1.Reading comprehension measures were signi cantly associated with all measures, rs ≥ 0.22, and signi cantly intercorrelated, r = 0.38.Reading uency measures were signi cantly associated with decoding measures and linguistic comprehension measures, rs ≥ 0.24, and signi cantly intercorrelated, r = 0.82.Decoding measures were signi cantly associated with linguistic comprehension measures, rs ≥ 0.37, and signi cantly intercorrelated, r = 0.80.Linguistic comprehension measures were signi cantly intercorrelated, r = 0.38.

Structural equation modeling
Structural equation modeling (SEM) was used to examine the role of reading uency in the SVR.We rst examined the original   Because both Model 3A and 3B showed an excellent t to the data, a model t comparison was conducted for the selection of the preferred model in this study.Table 2 shows the model tting results of the above-examined models.A comparison of the overall t statistics and model paths between Model 3A and 3B did not show a statistically better model; ∆χ 2 = 2.50, ∆df = 1, p >0.05, and ∆AIC = 0.50.According to the verbal efficiency theory (Perfetti, 1985) and the hypothesis from Fuchs et al. (2001), reading uency must be achieved to free up cognitive resources for reading comprehension, and hence, the development of reading uency is preceded by decoding and mediating the relationship between decoding and reading comprehension.erefore, based on the theory, Model 3B was preferred in this study, supporting the mediating role of reading uency in the SVR.

Discussion
e present study examined the distinctiveness of reading uency from decoding and the role that reading uency plays in the extended SVR model in Chinese children.e results of CFA supported the hypothesis that reading uency is a separate factor from decoding.Moreover, SEM results revealed that reading uency is a unique predictor of reading comprehension, mediating between decoding and reading comprehension.

Reading fluency as a separated factor from decoding
Word reading and reading uency are de ned as reading accuracy and automaticity, respectively (Fuchs et al., 2001).However, previous studies in Chinese have consistently considered reading uency as a measure of decoding skills rather than a separate measure (Yeung et al., 2016;Ho et al., 2017).Only one study examined the factor structure using CFA (Ho et al., 2017).Moreover, previous studies in Chinese have employed either word-or textreading uency and submerged reading uency into decoding (Yeung et al., 2016;Ho et al., 2017).e present study is the rst to examine the extended SVR in Chinese, employing word-and text-reading uency measures simultaneously to represent the reading uency domain and enhance the generalizability of the variable.In line with the extended SVR (Silverman et al., 2013), the results from the CFA showed that the four-factor model tted the data better, suggesting that reading uency is a separate skill from decoding, linguistic comprehension, and reading comprehension in the Chinese language.Silverman et al. (2013) suggested that reading accuracy and uency develop simultaneously in early childhood and independently in fourth grade and beyond.Participants in this study were in grades 3 to 5, consistent with Silverman et al. 's (2013) ndings.Previous studies in Chinese reading have been conducted with children in grades 1 to 3 (Yeung et al., 2016;Ho et al., 2017); therefore, it is possible that the grade level of participants might in uence the factor structure of reading uency and decoding.e present study warranted that reading uency is separated from word reading in the Chinese writing system under the extended SVR framework.In addition, compared to the previous studies in Chinese SVR, our results suggested that the independent development between reading accuracy and uency is plausible aer junior elementary years.Future longitudinal studies are needed to further explore the development of reading accuracy and uency and their in uence on reading comprehension in the extended SVR.

Role of reading fluency in the Simple View of Reading
e present study is the rst to examine the mediating role of reading uency in Chinese reading.In line with the verbal efficiency theory (Perfetti, 1985) and previous ndings of the extended SVR in alphabetic languages (Tilstra et al., 2009;Silverman et al., 2013), the present study lled the gap in the literature that reading uency is an additional unique contributor to reading comprehension in Chinese and advancing the extended SVR in Chinese.Moreover, the results of the present study showed that reading uency signi cantly mediated the relationship between decoding and reading comprehension.
According to the verbal efficiency theory (Perfetti, 1985), one must achieve automatic word decoding to allocate more cognitive resources for reading comprehension.erefore, readers with faster word decoding, i.e., better reading uency, would be better in reading comprehension.is result is in line with previous studies in English, Finnish, Portuguese, and Korean (Joshi and Aaron, 2000;Silverman et al., 2013;Kim et al., 2014;Torppa et al., 2016;Cadime et al., 2017).In addition, the present nding of reading uency as the mediator between decoding and reading comprehension also echoes the study by Silverman et al. (2013) in English readers and supports Pikulski and Chard's (2005) proposition of reading uency as the bridge for children to proceed from decoding at the word level to comprehension at the text level.e mediating role that reading uency plays in the context of the Chinese language suggested a diminished effect of decoding in Chinese reading comprehension when reading uency is accounted for, which is a similar pattern observed in English readers (Silverman et al., 2013).Once children are capable of automating the process of word recognition, their decoding skill is insufficient to support their comprehension Finally, moderation analyses revealed that grade levels and gender did not signi cantly moderate the contribution of reading uency, decoding, and linguistic comprehension on reading comprehension.Previous studies have shown that girls consistently perform better than boys in reading comprehension and read more books than boys in primary schools (Logan andJohnston, 2009, 2010).However, studies about extended SVR have not yet examined the stability of SVR across gender.e results of the present study showed that the extended SVR is stable across gender, where the paths do not differ between boys and girls.While the present study was not a longitudinal study, our moderation analyses revealed that the grade level was not a signi cant moderator of the present extended SVR model.e model remains signi cantly stable across intermediate grade levels.

Limitations and future directions
ere are several limitations to the present study.One caveat of this study was the lack of data on younger and beginning readers, hindering the extent of the present nding.As suggested by Silverman et al. (2013), the role of reading uency becomes more salient aer grade fourth, and there is a lack of empirical evidence to support that grade fourth is the transition period of the development of reading uency.erefore, future studies should consider comparing the role of reading uency in the SVR between younger and older readers.Additionally, the construct of linguistic comprehension in this study did not include any listening comprehension measures, oen included in studies of SVR models, to account for children's higher mental processes at the sentence and discourse levels.Future research studies should include listening comprehension and discourse measures at the text level to further unravel the components of reading comprehension.Another limitation was that the present research only examined oral reading uency concerning rate and accuracy.Including measures of silentreading uency may also add additional variances to reading (Price et al., 2016).e developmental effect could not be examined in the present study without longitudinal follow-up to examine grade-level effects and the potential changes in the mediating relationship of the extended SVR model.

Conclusion and implications
Despite the limitations, this study contributes to the eld by extending the current understanding of SVR and reading uency in Chinese children's reading development.e present study was the rst to examine reading uency thoroughly in the SVR model in a language with opaque orthography.e results of the present study extended the SVR and demonstrated that reading uency, the ability to retrieve words quickly, could be a separate reading skill from decoding, which may play a unique role in reading comprehension and a bridging role between decoding and reading comprehension.In general, the ndings underscore the critical role of reading uency in children's reading development.Moreover, the results suggest the signi cance of incorporating reading uency measures at various linguistic levels to obtain a comprehensive assessment of children's uency performance.ese ndings hold important implications for education.It is recommended that classroom assessments and intervention programs consider reading uency as a core component of the curriculum.By focusing on reading uency, educators can better address the needs of children who demonstrate pro cient decoding skills but continue to struggle with text comprehension.Intervention programs prioritizing reading uency can facilitate children's reading development from decoding individual words to comprehending texts.In conclusion, this study provides valuable insights into the importance of reading uency in the reading development of Chinese children.

Participants
Participants were 342 Cantonese-speaking Chinese children in grades 3-5 (M age = 10.04 years; SD age = 0.98 years; girls = 163) from three primary schools in Hong Kong.Children were recruited by sending invitation letters to the parents of all children in grades Frontiers in Psychology 03 frontiersin.org3-5 from the participating schools.All the children were screened to identify typically developing Chinese children without learning disabilities or sensory impairments.

FIGURE 1
FIGURE 1Confirmatory factor analysis examining the factors in the Simple View of Reading.*** p < 0.001.

FIGURE 2
FIGURE 2Structural equation modeling examining the original Simple View of Reading in Chinese.** p < 0.01, *** p < 0.001.Covariances between independent variables were omitted for readability.

FIGURE 3
FIGURE 3 Structural equation modeling examining the role of reading fluency in the Simple View of Reading in Chinese.NS, Not significant, *p < 0.05, **p < 0.01, ***p < 0.001.Covariances between independent variables were omitted for readability.