Bridging the Gap:
Analyzing English Reading Proficiency Disparities Among Korean EFL Learners

doi:10.21203/rs.3.rs-5303313/v1

This study investigates the heterogeneity in English reading abilities among sixth-grade EFL learners in South Korea, focusing on their L2 reading profiles and the socio-educational factors predicting them. 598 students from mid- and low-income neighborhoods were assessed on five L2 literacy components: decoding, oral fluency, vocabulary knowledge, syntactic knowledge, and reading comprehension. Latent profile analysis identified distinct subgroups of learners, and multinomial logistic regression was used to analyze the key predictor variables influencing English reading proficiency. The results highlight a significant proficiency gap between students who began learning English in the third grade and those exposed earlier through private tutoring. Struggling readers showed deficits in decoding, while proficient readers excelled in fluency. The impact of early English education was limited compared to learning experiences in elementary school. The study suggests the need for targeted interventions to address reading disparities and provides pedagogical implications for improving English language instruction in Korean classrooms.

Educational Psychology

Linguistics

L2 Reading ability

L2 learning background

heterogenous groups

L2 decoding ability

latent profile analysis

Reading is a critical source of comprehensible input, significantly contributing to overall language skills development (Krashen, 1993). Researchers emphasize the importance of practical reading ability in foreign language learning, as it is the most effective way to achieve language learning goals (Carrell, 1988; Mikulecky & Jeffries, 1986). Developing reading skills is especially crucial for early learners in L1 and L2 contexts (Chall et al., 1990; Oladel, 2016).

Early engagement in reading is essential because difficulties in L2 reading tend to accumulate over time, necessitating appropriate educational interventions during early learning stages. Reading proficiency encompasses multiple sub-skills, and the absence of any one sub-skill can hinder overall reading achievement (Grabe, 2009; Jeon, 2011; Jeon & Yamashita, 2014; Koda, 2005). Critical components of reading comprehension include word decoding, reading fluency, vocabulary, and syntactic knowledge, which interact to compensate for deficiencies in specific sub-skills (Stanovich, 1980).

Unlike simple reading ability analysis, the sociocultural perspective also suggests variations in students' language proficiency stem from diverse educational backgrounds influenced by socioeconomic factors, private tutoring, family literacy environment, and affective domains (Lee, 2018). These factors significantly impact young learners' English reading abilities, forming the foundation for this study.

In South Korea, where parents are highly invested in their children's English education, there is significant concern regarding the gap in English language proficiency among young learners. This gap is influenced by early exposure to English education through private tutoring and other methods, leading to disparities in starting points and subsequent learning experiences (Park, 2012). Despite the importance of English education, it receives less attention in elementary schools than other subjects (Lee, 2018).

Building on this background, this study explores the diversity of English reading abilities among sixth-grade students in Korean elementary schools. Previous studies have primarily focused on variables associated with English proficiency, but this research focuses on individual students, uncovering "hidden groups" within potential hierarchies through profiling studies (Landers, 2008). While many existing studies have analyzed reading abilities or sociocultural backgrounds, there is a lack of research that examines the potential subgroups of L2 learners and their reading abilities through such an analysis.

Therefore, this study aims to investigate the heterogeneity of sixth-grade students in a Korean elementary school within an EFL environment. The research questions are:

1. How many distinct sub-groups exist among Korean EFL sixth graders regarding their L2 reading ability, and what characteristics does each group demonstrate concerning their L2 reading skills?

2. What is the predictive relationship between these subgroups and their English learning backgrounds?

Different Perspectives on Reading

The reading activity delves into two perspectives on reading: as a set of composite sub-skills and as a socially constructed activity. Traditional views focus on improving reading accuracy and speed, grounded in behaviorism and cognitive psychology, while recent perspectives emphasize the social context of reading.

Reading as a Set of Composite Sub-Skills

This perspective defines reading ability as an aggregate of several components: word decoding, reading fluency, vocabulary, syntactic knowledge, and reading comprehension (Grabe, 2014; Jeon, 2011; Jeon & Yamashita, 2014; Koda, 2005; Perfetti et al., 2005). These components interact to compensate for deficiencies in specific sub-skills, supporting overall reading comprehension. Word Decoding is essential for early readers, but its importance diminishes as other skills develop (Adlof et al., 2006; Cain & Oakhill, 2006; Gottardo et al., 1996; Perfetti et al., 2008; Siegel & Ryan, 1988).

Reading fluency automates basic skills and is crucial for freeing cognitive resources for comprehension (Cunningham & Stanovich, 1998; Nasaji, 2003). Vocabulary and syntactic knowledge become more significant predictors of reading comprehension in later stages of reading development (Cho et al., 2019; Hutchinson et al., 2003; Muter et al., 2004; Roth et al., 2002; Verhoeven & van Leeuwen, 2008). Studies have shown that the predictive power of these component skills varies by grade (Garcia & Cain, 2014; Lonigan & Burgess, 2017; Lonigan & Schatschneider, 2018; Tilstra et al., 2009). For instance, word decoding is the most predictive in fourth grade but less so by sixth grade, where vocabulary and syntactic knowledge play more prominent roles. Reading comprehension is the ultimate goal, requiring the integration of all these skills to understand and interpret the text.

Reading as a Socially Constructed Activity

This perspective views reading as influenced by social and cultural factors, recognizing that reading practices and meanings are shaped by the social context, including cultural norms and values (Burgoyne et al., 2011; García, 2000; Jiménez et al., 1996). Higher socioeconomic status (SES) often correlates with improved access to reading materials and educational support, positively impacting L2 reading ability (Cheung & Anderson, 2003). Supplementary education, such as reading clubs and tutoring, can enhance L2 reading skills (Choi, 2007; Lee, 2002; Park, 2012). A supportive home environment, characterized by the availability of diverse reading materials, fosters the development of L2 reading abilities (Grabe, 2009). The presence of books and other reading resources at home plays a crucial role in encouraging reading habits and proficiency. Motivation and self-efficacy are critical factors influencing learners' reading ability. Learners with higher motivation and confidence in their reading skills are likelier to engage with texts and improve their L2 reading proficiency.

Understanding reading from both perspectives, composite sub-skills, and socially constructed activity—provides a comprehensive view of L2 reading development. While the traditional view focuses on the technical components of reading, the socio-cultural perspective highlights the broader social context, acknowledging reading development's dynamic and complex nature. By considering these perspectives, educators can better support the diverse needs of EFL learners, particularly in South Korea, where educational, social, and economic factors influence variations in English reading abilities.

Participants

The study involved 678 sixth-grade students from seven elementary schools located in mid-income areas across five provinces in Korea: Seoul, Incheon, Suwon, Changwon, and Jeonju. These students had been learning English for over three years as part of the regular curriculum. Schools were chosen to represent a cross-section of regions and socio-economic diversity. Students were required to have no diagnosed learning disabilities that could interfere with reading assessments. After excluding 80 students due to absences, missing data, or failure to submit questionnaires, 598 participants remained. All students attended public schools that adhered to the Korean National English Curriculum, ensuring a shared achievement goal across the sample. Parental consent was obtained from all participants, and ethical guidelines for research with minors were followed. Table 1 provides detailed information on the participating schools.

Table 1

*Information on Participation Schools*
	School A	School B	School C	School D	School E	School F	School G
¹ Total number (N)	570 (87)	537 (94)	470 (93)	438 (82)	932 (133)	626 (115)	528 (114)
² Region of schools	*Seoul	*Seoul	*Incheoun	*Incheoun	**Suwon	**Chun Ju	**Chang won
³Cooperative class	Impleme-ntation	Impleme-ntation			Impleme-ntation
⁴.Average number	20.4	18.5	20.5	19.9	24	25	24
Note					⁵.Innovation school

Total number of students in the school (Grades 1–6)

* Metropolitan city ** Medium-sized city

Schools where native English-speaking teachers are in schools and classes are conducted with Korean teachers.

Average number of students in a class

A type of school that devises and applies a new curriculum, unlike the uniform public education curriculum.

Procedure

Data collection took place at the start of the second semester, following the summer vacation in 2021, reflecting Korean sixth-grade students' typical English proficiency and learning experiences. Assessments were conducted over three English class sessions, with English teachers administering the tests based on standardized protocols. The researcher held two preparatory online meetings with the teachers to ensure consistent assessment administration across schools.

The study included five assessment tasks: one face-to-face test and four online or paper-based tests. The face-to-face test measured decoding and oral reading fluency, while the online tests assessed word knowledge, syntactic awareness, and reading comprehension. Each task took approximately 40 minutes to complete. To account for socioeconomic factors, a 10-minute questionnaire gathered data on participants' backgrounds and learning experiences, including their access to private English tutoring and parental education levels. The total assessment time was approximately 2 hours and 30 minutes, spread over three weeks. Oral reading tests were administered to volunteers outside of class, while online group tests were conducted during class. Data from students who did not participate were discarded. The assessments were conducted in a controlled classroom environment, ensuring minimal distractions.

Teachers received thorough training through two preparatory online meetings, where the researcher outlined the specific protocols for each test. This training ensured consistency across all schools and minimized potential discrepancies in the administration process.

Instruments

The study assessed L2 (second language) reading-related abilities by evaluating six critical components of L2 literacy: L2 Nonword Reading, L2 Word Reading, L2 Passage Reading, L2 Vocabulary Knowledge, L2 Syntactic Knowledge, and L2 Reading Comprehension. Guided by NICHD research, the study emphasized the importance of decoding ability, particularly in relation to phonological skills. It used nonword reading and word reading tests to measure whether students relied more on phoneme-grapheme correspondence or whole-word recognition.

Decoding ability was further analyzed through the phonological route (grapheme-to-phoneme correspondence) and the lexical route (whole-word recognition), capturing different aspects of reading skills. In addition to decoding, the study focused on language comprehension, examining word knowledge and syntactic knowledge within the EFL (English as a Foreign Language) context. The six indices used to assess the students' L2 reading abilities were:

L2 Decoding Ability (Nonword and Word Reading)

L2 Oral Reading Fluency,

L2 Vocabulary Knowledge,

L2 Syntactic Knowledge, and

L2 Reading Comprehension.

These indices were used to classify students into distinct reading proficiency groups. All instruments were adapted from previously validated measures, ensuring reliability and validity. Before full-scale data collection, the tools were piloted with a smaller group of students to confirm their effectiveness. Table 2 summarizes the instruments and procedures used in the study.

Table 2

*A Summary of Instruments and Procedures*
Constructs	Measures	Source	Time (sec/min)	Session
L2 decoding	1. Non-word reading	¹ TOWRE	45 sec	1
L2 decoding	2. Word reading	¹ TOWRE	45 sec	1
L2 reading fluency	3. Oral passage reading	¹ DIBELS Step 1	1 min	1
L2 vocabulary knowledge	4. Vocabulary size test	² Researcher-made Nation	30min	2
L2 grammar knowledge L2 reading comprehension	5. Grammar knowledge test 6. Reading comprehension test	² Researcher-made ² Researcher-made	40min 40min	2 3
Note. ¹ Tasks to count the number of words read quickly in a limited time
² Tasks that can solve all problems within the specified time

Measuring Learners’ L2 Learning Backgrounds

The study's survey aimed to measure Korean EFL sixth graders' L2 learning backgrounds. It included questions about personal information, household environment, parental information, and English learning experiences before and during elementary school. The survey consisted of 24 questions, administered through Google Forms during online classes. A question about parents' educational background was removed due to low response rates. The survey used eight categorical and sixteen interval scales, detailed in Table 3.

Table 3

*Survey Information*
Domains	Types of Scale	Number of items	Reliability
Gender	Categorical	1	-
Location		1	-
Father Education		1	-
Mother Education		1	-
Household English book holdings	Interval	1	-
English support from parents		5	.82
English experience before elementary school		3	.83
Types of English learning activities taught in elementary school	Categorical	4	-
Elementary school English learning experiences	Interval	4	.89
Intrinsic motivation for current English	Interval	3	.75
Total		24

Statistical Analysis Plans

Latent profile analysis (LPA) was conducted to identify student subgroups based on their performance across six L2 literacy indices. Exploratory factor analysis (EFA) was used to ensure the factors' validity. Multinomial logistic regression examined the influence of socio-educational factors such as parental education, private tutoring, and regional differences on students' reading abilities, controlling for socio-economic background. Appropriate statistical techniques were applied to handle missing data from the 598 participants, ensuring reliable results. Two statistical packages were used: SPSS and Mplus. The SPSS program (version 22) was used for descriptive statistics, EFA, and multinomial logistic regression, while Mplus software (version 8.7) was used for LPA.

Descriptive Statistics of the Students’ L2 Reading Ability Measurements

The descriptive analysis of all L2 reading-related measures, including mean, standard deviation, skewness, kurtosis, and Cronbach alpha, has been reported. To satisfy the assumptions for preliminary analysis, skewness and kurtosis values were checked to confirm the data's normality. Table 4 shows descriptive statistics of each test.

Table 4

*Descriptive Statistics (N = 598)*
Mea sures		Mean	SD	Min	Max	Skewness	Kurto sis	Cronbach alpha
		Decoding Ability
NW	66	29.24	15.195	0	64	.032	− .610	-
WR	108	49.80	20.507	0	90	− .514	− .287	-
		Oral Reading Fluency
PR	170	85.99	45.150	0	170	.060	− .740	-
		Written Linguistic Knowledge
WK	35	25.09	8.638	3	35	− .829	− .467	.95
SK	25	14.48	5.174	3	25	.004	− .785	.74
		Reading Comprehension
RC	25	17.88	5.265	0	25	-1.026	.437	.76
Note. NW = L2 non-word reading; WR = L2 word reading; PR = L2 passage reading; WK = L2 word knowledge; SK = L2 syntactic knowledge; RC = L2 reading comprehension. The reliability of NW, WR, PR cannot be obtained from the current study, but the publisher-reported reliability by Kuder-Richardson Formula is .90 to .92.

The present study examined the correlation coefficients of oral reading indicators, including non-word reading, word reading, and passage reading. While these indicators represent specific aspects of L2 reading abilities, it is acknowledged that they may not be completely independent of each other. However, since the main objective is to identify essential elements for reading comprehension, the study still holds significant research value even with relatively high correlations. For example, meta-analyses like the one conducted by Jeon and Yamashita (2024) review various studies on L2 reading comprehension and confirm significant correlations between different reading skills, such as decoding and oral fluency, and overall reading comprehension. Despite these correlations, they emphasize the importance of each skill in predicting comprehension outcomes, underscoring their individual contributions within a broader framework of reading proficiency. Additionally, the work of Hoover & Gough (1990) on the "Simple View of Reading" posits that reading comprehension is the product of decoding and linguistic comprehension, implying a natural interdependence among oral reading skills, even if they are highly correlated.

In addition, in latent profile analysis (LPA), ensuring the independence of indicators is crucial for accurate model specification and interpretation. A commonly accepted practice is to treat correlations below 0.8 as an indication that the variables are sufficiently independent. When correlations exceed this threshold, it may indicate multicollinearity, which can reduce the distinctiveness of latent profiles (Choueiry, 2020). This rule of thumb is consistent with broader guidelines in statistical modeling, such as those applied in regression analysis, where variance inflation factors (VIF) or correlation coefficients above 0.8 signal potential issues with variable redundancy (James et al., 2013). The threshold also aligns with Hill and Robertson’s (1968) work, which explores the relationship between linkage disequilibrium and correlation, demonstrating that higher correlations can confound the interpretation of independent variables. The correlation values for each indicator are presented in Table 5.

Table 5

*Coefficients among L2 Reading Ability Variables (N = 598)*
	1	2	3	4	5	6
1 NW	1
2 WR	.799^***	1
3 PR	.798^***	.789^***	1
4 WK	.585^***	.705^***	.692^***	1
5 SK	.624^***	.715^***	.729^***	.677^***	1
6 RC	.580^***	.705^***	.704^***	.714^***	.726^***	1
Note. Correlation coefficients are all significant at the .001 level. NW = L2 non-word reading; WR = L2 word reading; PR = L2 passage reading; WK = L2 word knowledge; SK = L2 syntactic knowledge; RC = L2 reading comprehension.

Research Question 1: L2 Reading Ability Profiles of the Korean 6th Graders

The first research question involved a two-step process to identify distinct sub-groups based on their profiles of English reading ability. Firstly, a latent profile analysis (LPA) was conducted using six indicators, including L2 non-word reading, L2 word reading, L2 oral passage reading, L2 word knowledge, L2 syntactic knowledge, and L2 reading comprehension. The most appropriate model was selected based on the latent profile analysis results. Next, each profile group was named to represent the specific characteristics of the group.

1Six indicators (L2 nonword reading, L2 word reading, L2 oral passage reading, L2 word knowledge, L2 syntactic knowledge, and L2 reading comprehension skills) were used in the EFL learners' English reading ability profile analysis model in this study. For LPA analysis, model suitability was judged by sequentially reviewing the suitability indices from Profile 1 to Profile 8 in Table 6.

Model Selection and Interpretation

Regarding the LPA, a model must be selected and interpreted first. A series of LPAs was conducted, following McLachlan and Peel’s (2000) suggestions of appropriate statistical tests and indices to determine an optimal number of profiles to retain. The present study adopted AIC, BIC, and SABIC indices for model fit type, entropy index for classification quality, LMR, and BLRT for model comparison validation. AIC, BIC, and SABIC are all information criteria that assess the goodness of fit of a model by balancing the model's complexity against its ability to explain the data. A lower value indicates a better fit for these indices. The entropy index measures the precision with which individuals are classified into their respective latent groups. A higher value indicates a better classification quality. LMR and BLRT are both likelihood ratio tests that compare the fit of a model with k-1 latent groups to that of a model with k-1 latent groups. By combining these indices, the study can determine the optimal number of potential groups for subcomponent analysis of reading skills.

In this study, the values of the three information indexes (AIC, BIC, and SABIC) continued to decrease across the range of models considered but only marginally between the six-class solution and the eight-class one. For model comparison validation, the adjusted likelihood ratio test (LMR) became no significant with the seven-class model, meaning that the six-class model was optimal. On the other hand, there was no discriminating difference in the BLRT index in all profile models. In classification quality, the entropy values were higher than the suggested 0.80 value and almost identical among the two-class and eight-class solutions.

It is necessary to examine further the meaning of the fitness of each index concerning the heterogeneity between profiles. Looking at the six fitness indices (AIC, BIC, SABIC, Entropy, LRT, BLRT) in Table 6 displays the fit of profile models. Comprehensively, looking at all profiles, the AIC, BIC, and SABIC values were sequentially decreased from the Profile 2 model to the Profile 8 model, and the BLRT results of all profiles were significant (p ≺ .001). In addition, the average posterior probability range of all profiles presented in Table 6 was .79 to .91, indicating that the average posterior probability value was close to 1.0, confirming that the classification error was small (Nagin, 1999). Finally, the Profile 7 model showed the lowest AIC, BIC, and SABIC values; LMR and BLRT were also significant. Therefore, the best-fit model indices from Table 6 can be seen in Profile 7, although the LMR index is not statistically significant.

However, more than simply analyzing the model fit indices is required to determine the number of profiles. Nylund et al. (2007) proposed specific considerations for determining the number of profiles, and this study applied those considerations to determine the number of profiles. Firstly, special attention was given to Profile 6 and Profile 7 models among all the profiles, as they exhibited detailed characteristics of the student group that slightly deviated from the average of all indicators and showed statistical significance. Secondly, the profile graph pattern in Fig. 1 was analyzed, and the Profile 6 model was found to have more diverse group information than other profiles, making it a potential candidate for analysis. Thirdly, the study ensured that the minimum number of students in each profile exceeded 30. In the case of the Profile 6 model, all profiles had a hierarchical frequency exceeding 30 (minimum frequency 7.2% = 43 students), thereby avoiding any classification ratio falling below the minimum frequency. However, the group with the lowest score in the Profile 7 model did not meet the minimum frequency for latent profile classification (n = 30) at 4.2% (n = 25) (Berlin et al., 2014; Ferguson et al., 2020; Nylund, 2007; Pearson et al., 2015; Peugh & Fan, 2013; Tein et al., 2013; Williams & Kibowski, 2016).

The Profile 6 model was ultimately chosen based on the following criteria. The six-profile model demonstrates high practicality as it exhibits a distribution that approximates normality, with a minimum classification rate of 7.2%, surpassing the 5% threshold, and a minimum frequency exceeding 30. Table 6 indicates that the Profile 6 model exhibits relatively even and diverse information, with classification rates ranging from 7.2–28.8%. Furthermore, the Profile 6 model achieved a classification Entropy of .90 after careful analysis and selection. Table 6 shows the model fit of latent profile analysis.

An Elbow Plot can validate the model fit based on these information indices. Nylund-Gibson & Choi (2018) suggested plotting the information index values such as AIC, BIC, and SAIC and visually displaying the values to provide for straightforward interpretation. In general, they continue to decrease for each additional class added. The plot elbow can be particularly useful for inspecting for an “elbow” at the point of “diminishing returns” in model fit. In this study, Fig. 1 shows a sharp decrease up to the six profile and a gentle decline from the Profile 6 model.

Table 6

Model Fit Indices of Latent Profile Analysis
Number of Profile	¹AIC	²BIC	³SABIC	Entropy	⁴LMR	⁵BLRT	Latent Classification Rate (%)
							1	2	3	4	5	6	7	8
2	8459.17	9019.50	8959.17	0.90	0.0008	0.00	36	64
3	7481.87	7963.92	7881.37	0.91	0.0006	0.00	16.0	49.6	34.4
4	7144.28	7449.06	7344.28	0.91	0.0009	0.00	10.7	30	43.6	15.7
5	7001.13	7311.12	7184.123	0.89	0.0208	0.00	9.0	15.5	25.8	31.6	18.1
6	6934.42	7163.94	7014.72	0.90	0.0455	0.00	10	16.8	7.2	28.8	22.0	15.2
7	6252.34	6923.91	6752.48	0.89	0.0513	0.00	8.9	13.5	4.2	15.6	25.7	19.7	12.4
8	6123.72	6423.31	6352.32	0.79	0.0678	0.023	6.9	9.5	3.2	14.4	21.3	17.4	17.4	9.9
Note. ¹ AIC = Akaike information criterion ^{; 2}BIC =Bayesian Information Criterion; ³SABIC=Sample adjusted BIC; ⁴ LMR = adjusted Lo-Mendell-Rubin likelihood ratio ; ⁵ BLRT = Bootstrap Likelihood Ratio Test (p ≺ .001), Correlation coefficients are all significant at .001 level. NW = L2 non-word reading; WR = L2 word reading; PR = L2 passage reading; WK = L2 word knowledge; SK = L2 syntactic knowledge; RC = L2 reading comprehension

Naming Based on the Characteristics of the Six Profiles

After selecting the six profiles as the best-fit model in LPA, Fig. 2 analyses each profile based on the Z-scores of each reading subskill. The overall pattern of the six profiles demonstrates a parallel trend, but Profile 3 and Profile 4 exhibit a crossing pattern.

In Table 7, the first profile (n = 60, 10%) is the group that has difficulty with all the L2 decoding words, syntactic, and reading comprehension. Therefore, this profile group has been named ‘Overall severe deficit L2 readers’. Despite the four-year English education at school, this group does not seem to have basic L2 literacy skills across the six L2 indicators.

In the second profile group (n = 100, 16.8%), all L2 literacy sub-skills were below average. However, compared to ‘Overall severe deficit L2 readers’, they seemed to have the decoding ability even though they were not high enough compared to other profiles. The peculiarity of this profile is that L2 oral reading skills, such as L2 word decoding ability and L2 oral reading fluency, are at a relatively low level. In contrast, vocabulary, L2 grammar, and L2 reading comprehension were higher than the former three skills. Therefore, this profile group was named ‘Severe deficit in L2 oral decoding and reading skills readers’. One intriguing feature of this group is that although basic L2 decoding skills are not well learned, this group seems to have some basic vocabulary and syntactic knowledge compared to Profile 3 and Profile 1.

The third profile group (n = 43, 7.2%) is slightly below average in decoding and reading fluency ability but has significantly lower L2 linguistic knowledge, such as vocabulary and syntax knowledge. Therefore, this profile group was named ‘Severe deficit in L 2 linguistic knowledge readers’.

The remaining 4, 5, and 6 profiles were groups with average or above-average reading ability. The fourth profile (n = 173, 28.8%) group was named ‘Average L2 readers’ because the six L2 reading-related indicators were in the middle of the six groups.

The fifth profile group (n = 132, 22%) was named ‘Above average L2 readers’ because all the L2 indicators were slightly above average.

Finally, the sixth profile group (n = 90, 15.4%) was named ‘Proficient L2 readers’ because it was the highest in all six L2 indicators. This group was 1 or 1.5 SD higher than the average L2 readers (Profile 4 and 5 in the six L2-related reading indicators).

The overall profile characteristics range from Profile 1, which shows a lack of decoding ability and limited development in other specific reading abilities, to Profile 6, which demonstrates the highest proficiency across all skills. These profiles exhibit a leveled pattern in terms of proficiency levels. However, Profile 2 and Profile 3 exhibit a crossover pattern in the oral passage reading measure, indicating contrasting weaknesses in specific reading abilities. Conversely, these two profiles demonstrate contrasting strengths in less vulnerable (closer to average) particular skills. Table 7 presents the characteristics of each reading profile, while Fig. 3 illustrates the features of the six distinct profiles.

Table 7

*Characteristics of Each Profile*
Profiles	Names	Characteristics
1	Overall severe deficit L2 readers	A group with difficulties in all subcomponents of L2 reading skills, including basic decoding skills
2	Severe deficit in L2 oral decoding and reading skills of readers	A group with slightly below-average oral reading skills and below-average syntactic knowledge and reading comprehension. A crossover pattern is observed with Profile 3 at the point of the oral passage reading indicator.
3	Severe deficit in L 2 linguistic knowledge readers	A group with below-average oral reading skills and slightly below-average syntactic knowledge and reading comprehension. A crossover pattern is observed with Profile 2 at the point of the oral passage reading indicator.
4	Average L2 readers	A group with average reading comprehension and subcomponent skills
5	Above average L2 readers	A group with slightly above-average reading comprehension and sub-component skills
6	Proficient L2 readers	A group with better-than-average reading comprehension and subcomponent skills

Research Question 2:

Learning Background Affecting English Reading Ability of EFL Learners

The second research question aimed to investigate the relationship between the L2 reading profiles of sixth-grade students and their English learning backgrounds. To achieve this, the participants completed a questionnaire, and the responses were analyzed using exploratory factor analysis (EPA) to identify common factors. The validity of the factors was examined, and multinomial logistic regression analysis was used to determine which learning background predictors affected the participants' English reading ability.

To validate the survey results and identify predictors of English reading ability among Korean EFL sixth graders, an exploratory factor analysis (EFA) was conducted on 16 Likert scale items. Following Hatcher's (2013) procedure, commonality values were calculated, and shared factors grouped survey questions. Items with loading values below 0.4 were excluded, leading to the extraction of three sub-factors. Item 16 was excluded for its low explanatory power.

The factor matrix showed loading values for the three retained factors. Factors with eigenvalues greater than 1.00 were considered significant, with Fig. 4 indicating that Factors 1 to 4 met this criterion. However, the scree test suggested retaining three factors due to a noticeable eigenvalue drop after the third factor. This approach was deemed appropriate for the analysis, and the three factors were recommended for further rotation and interpretation, as they explained a significant portion of the variance.

This study identified three key factors influencing the English reading ability of EFL learners: 'English learning environment at home,' 'English experiences before elementary school,' and 'English experiences during elementary school.' The reliability (Cronbach's alpha) for each factor was .784, .859, and .601, respectively. Figure 4 shows a scree plot of factor analysis, and Table 8 shows the factor analysis.

Table 8

*Factor Matrix Questionnaire and Items by Extracted Factors*
Factors	Commonality	Loading Value			Questionnaire
Factors	Commonality	1	2	3	Questionnaire
English environments at home	.525	.744			Are your parents interested in English?
	.485	.730			Have your parents ever read a book in English?
	.408	.634			Do your parents usually emphasize English a lot?
	.320	.596			Do your parents help you learn English a lot?
	.343	.571			Have you ever talked to your parents in English?
	.301	.509			How many books in English (excluding textbooks and reference books) can you read at home?
English learning experiences before elementary school	.849		.939		Were you good at English when you were in kindergarten?
	.797		.923		Did you like English when you were learning English in kindergarten?
	.467		.564		How much English did you learn in kindergarten?
English learning experiences during elementary school	.577			.715	Did you study English a lot in elementary school?
English learning experiences during elementary school	.330			.625	Do you feel the need to study English?
	.387			.569	How have you been studying English lately?
	.246			.540	How many years did you attend an English academy in elementary school?
	.178			.458	Did you learn much English in elementary school in places other than school?
	.388			.441	How many books did you read in English in elementary school? (Excluding English textbooks and workbooks)
Note. Loading value > .40; Extraction Method = Maximum likelihood; Rotation Method: Direct Oblimin

A multinomial logistic regression analysis was conducted to investigate the predictive relationship between the English learning experiences of sixth-grade L2 learners and their L2 reading profiles. The study included three factors (i.e., English environments at home, English learning experiences before elementary school, and English learning experiences during elementary school) as category variables and gender as an individual variable. For the analysis of predictive variables related to English reading ability, the three factors were entered as categorical variables, with a 4-point scale ranging from the best (4) to the most vulnerable (1). However, to improve readability, the three category variables were converted into variables with three categories, still ranging from 3 to 1. Table 9 shows the predictor variables.

Table 9

*Predictor Variables*
Variables	M	Sd	Min	Max	Coding	Reliability
Gender	.55	.498	0	1	Male = 1, Female = 0	-
¹ English environment at home	14.4	.645	1.98	24	17–24: Much supportive
					13–16: Moderately supportive	.784
					0–12: Rarely supportive
² English learning experiences before elementary school	6.84	.880	.00	12	9–12: Much experienced
					5–8: Moderately experienced	.859
					0–4: Rarely experienced
³ English learning experiences during elementary school	16.38	1.116	.83	24	17–24: Much experienced
					13–16: Moderately experienced	.701
					0–12: Rarely experienced
Note.¹ English environment at home (the physical and emotional English language environments at home were included). ² English learning experiences before elementary school (all kindergarten private tutoring received before public schooling). ³ English learning experiences during elementary school (additional English experience other than school English education after the first grade when public education began).

The study used multinomial logistic regression analysis to determine which predictor variables most significantly influence English reading ability. This method, which accommodates more than two categories in the dependent variable, assessed model performance through model fitting information and pseudo-R-squared values, such as McFadden's and Cox and Snell's pseudo-R². Significant coefficients in the full model indicated that the additional variables improved the prediction of the dependent variable. The − 2-log likelihood (846.729) was smaller than that of the intercept-only model (886.582), suggesting a better fit. Pseudo R-squared values indicated that the model explained between 14.6% and 40.5% of the variance. The likelihood ratio test confirmed that models with more predictor variables had a better fit, as indicated by significant test statistics.

Table 9 outlines the outcomes of the multinomial logistic analysis, identifying factors influencing English reading ability among sixth-grade EFL students across six distinct profiles. The predictors considered include gender, L2 literacy environment, English learning experience before elementary school, and English learning experience during elementary school.

Profile 1 (Overall severe deficit L2 readers)

This profile served as the reference group with odds ratios for all variables set to 1 for comparison.

Profile 2 (Severe deficit in L2 oral decoding and reading skills)

No significant predictor variables were identified compared to Profile 1.

Profile 3 (Severe deficit in L2 linguistic knowledge readers)

A very supportive home environment significantly increases the odds ratio by 4.55 times compared to the least supportive environment (p < .01; OR = 4.550). Female students are significantly less likely to be in Profile 3 than male students, with the odds ratio decreasing by 74% (p < .01; OR = 0.263).

Profile 4 (Average readers)

Students with extensive English experience during elementary school have a 2.588 times higher odds ratio than those with rare experience (p < .05; OR = 2.588). Moderately experienced students also have a higher odds ratio of 2.445 than rarely experienced students (p < .05; OR = 2.445).

Profile 5 (Above average readers)

Extensive experience during elementary school results in a 2.784 times higher odds ratio compared to rare experience (p < .05; OR = 2.784). A very supportive home environment increases the odds ratio by 2.601 times compared to the least supportive environment (p < .05; OR = 2.601).

Profile 6 (Proficient readers)

A highly supportive home environment increases the odds ratio by 4.554 times (p < .01; OR = 4.554). Extensive experience during elementary school increases the odds ratio by 6.628 times (p < .01; OR = 6.628). Moderately experienced students have an odds ratio of 3.298 compared to rarely experienced students (p < .05; OR = 3.29).

Table 9

Multinomial Logistic Regression Results
^{Reading Profile & Variables}	Profile 2 vs Profile 1			Profile 3 vs Profile 1			Profile 4 vs Profile 1			Profile 5 vs Profile 1			Profile 6 vs Profile 1
	Exp(B)	Exp(B) 95% Confidence Interval		Exp(B)	Exp(B) 95% Confidence Interval		Exp(B)	Exp(B) 95% Confidence Interval		Exp(B)	Exp(B) 95% Confidence Interval		Exp(B)	Exp(B) 95% Confidence Interval
		Lower	Upper		Lower	Upper		Lower	Upper		Lower	Upper		Lower	Upper
Gender (reference: Male)
Female	1.10	.574	2.11	.26 *	.105	.658	.710	.387	1.302	1.013	.540	1.901	.859	.436	1.69
Male	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
L2 home literacy environment (reference: Least supportive)
Very Supportive	1.557	.514	4.714	4.550 **	1.201	17.245	1.672	.600	4.659	2.601 *	.910	7.440	4.554 *	.816	7.999
Moderately Supportive	1.340	.581	3.092	.895	.271	2.962	1.172	.539	2.550	1.090	.477	2.493	1.033*	.406	2.629
Not Supportive	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
English experience before elementary school (reference: Rarely experienced)
Much experienced	1.049	.452	2.435	1.487	.530	4.172	1.065	.482	2.353	1.312	.558	3.081	1.461	.575	3.715
Moderately experienced	.885	.380	2.059	.498	.147	1.683	1.162	.531	2.542	1.658	.715	3.842	1.430	.560	3.655
Rarely experienced	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
English experiences during elementary school (reference: Rarely experienced)
Much experienced	.910	.350	2.364	.899	.268	3.015	2.588 *	1.010	6.631	2.784 *	1.027	7.545	6.628 **	1.952	22.506
Moderately experienced	.874	.377	2.024	1.060	.355	3.161	2.445 *	1.051	5.690	2.301 *	.929	5.703	3.298 *	1.029	10.573
Rarely experienced	-	-	-	-	-	-	-	-	-	-	-	-	-	-	-
Note: The reference groups are Profile 1, Male, Least supportive group in L2 literacy environment, Rarely experienced in English experience before elementary school, Rarely experienced in English experienced during elementary school. Bold indicates statistically significant values. p < 0.5; p < .01; **p < .001. - indicates a reference.

In this study, decoding skill was measured separately as nonword reading and word reading, using five predictor variables. The reason for this was to determine whether students' decoding abilities rely on phonological and orthographic processing or if it is primarily driven by whole-word recognition.

This study analyzed the English reading ability of Korean students in grades 6, classified them into latent profiles, examined their characteristics, and named the six profiles based on this. After that, the learning background of each classified profile was analyzed to find out which predictive variable had the greatest influence. The results of the analysis showed that decoding ability had a great discriminatory power, especially among the lower-ranking profiles, and reading fluency showed the greatest difference among the upper-ranking profiles, indicating that there was a serious individual difference between each profile. Most of profiles 1, 2, and 3 lacked decoding ability and oral reading ability, such as nonword and word reading and reading fluency, proving that the Simple View of Reading (SVR), which claims that decoding and language comprehension are essential for reading comprehension, is also a significant model for EFL students.

One of the most surprising results of this study is that about 10% of students who received English education in public education for nearly four years lacked decoding ability, which is an essential skill. This suggests severe flaws in the current Korean curriculum regarding decoding ability, which is critical for reading comprehension and in addressing underachieving students. Failure to acquire these skills will result in students’ accumulated deficiencies in upper-level reading skills, creating a significant obstacle to their English studies in higher education. As such, the lack of decoding skills seen in sixth graders can start in elementary school and become more severe as students progress to higher grades.

According to the Korean Education Evaluation (Kim, Tae-eun et al., 2019), students who understand less than 20% of what they learn in textbooks are classified as “under-skilled.” Recent studies have found that approximately 5–7% of middle school students fall into this category, with the percentage increasing to 8% in high school (Ministry of Education, 2019, 2021). Furthermore, research results from the Korean Educational Development Institute (2020) indicate that approximately 30–50% of middle and high school students have difficulty with English reading comprehension. In particular, many students struggle with poor English reading skills, especially decoding skills.

In one study, 35% of Korean high school students rated their English reading skills as low, highlighting the need for improvement in reading skills (Nam, 2020). The English comprehension level of middle school students is deteriorating over time, with statistics showing that 14.3% of students in 2018, 24.1% in 2019, and 29% in 2020 did not understand English (Ministry of Education, 2019, 2021). These numbers indicate that many students are experiencing severe difficulties in learning English.

Therefore, efforts to improve English reading skills, especially decoding skills, are urgent. This should help students understand English better and ultimately achieve better learning outcomes. Studies have warned that the “Matthew Effect” may widen the gap between proficient and poor readers in the school system. Therefore, it is predicted that learners who do not receive appropriate treatment for the under-reading group will fall further behind in language development and academic achievement.

In addition to this gap in reading ability levels, the reading ability of low-level students, in particular, showed various deceits. This shows the inadequacy of current education, as current standardized education often has difficulty meeting the diverse needs of learners, especially in subjects with large differences in level such as English. According to this study, effective reading education should be tailored to diagnose and address specific deficiencies such as decoding, fluency, or language comprehension. In the case of the reading subgroups 1, 2, and 3 profiles, various problems were found in nonword reading, word reading, and fluency. Therefore, the curriculum should diagnose more detailed reading defects and support the continuous management of individual reading growth processes for underachieving students.

In aspect to learning background, the results of this study strongly refute the claim that early English education in an EFL environment helps overall language skills, especially reading comprehension, in relation to the learning backgrounds of learners. Many Korean parents believe that early exposure with private education to English will accelerate their children’s language acquisition, but the results of this study did not. According to studies by Lee (2002) and Cho et al. (2019), early English education, especially in preschool settings, mainly improved students’ pronunciation and speaking confidence. However, no significant results were shown for reading ability, a language skill that is much more essential for academic success in higher education. This is consistent with the study by Jang et al. (2010), which suggests that language learning is more effective when learners are cognitively mature. These findings raise the question of whether resources and policies for English education should focus on strengthening English education at the elementary school level, where public education begins, and whether it is necessary to invest opportunity costs in early private education. Ultimately, this leads to the conclusion that more strategic English education should be implemented in elementary schools, considering the critical period suitable for language education implementation.

Based on the above results, this study contains several educational implications, especially in relation to English education in EFL environments such as Korea. First, the results of this study strongly suggest that elementary schools are an important period for students to develop their English reading skills rather than receiving English education before entering school. Considering that this study included oral reading, this is not limited to the effect of literacy education. Considering that early English education does not significantly affect basic English skills, the educational system should invest more resources in strengthening English education at the elementary level. This change is consistent with the research results that cognitive maturity plays a vital role in language acquisition (Goriot, 2019; Raising Bilingual Children, 2021; Lee & Bae, 2022; Kang, 2012; Changes in perspectives on early English education over 25 years, 2024). Second, this study revealed a significant gap in reading ability among students, suggesting that it is urgent to identify and address the condition of students who lack the essential decoding ability of 10%. Continuous customized programs should be implemented so that these students’ learning deficits persist and worsen. Schools should develop a regular evaluation system so that these students can receive the help they need. Third, considering the studies that show that early English education mainly affects speaking confidence and pronunciation, educators and parents should reconsider their massive investment in early English education and promote it more actively. Fourth, in this study, the reading ability with the greatest individual difference among learners was reading fluency ability. In order to reduce this individual difference, schools should include continuous reading fluency programs in the curriculum and manage individual growth together. They should support the development of oral reading as well as written reading through activities such as continuous English reading and repeated reading.

This study contributed to identifying factors predicting EFL learners’ English reading ability and learning background, but some limitations need to be supplemented for follow-up research. First, due to the sample's representativeness, the study recruited schools that wanted to participate among similar schools, so random sampling was impossible. More advanced results can be obtained if this situation is supplemented in follow-up research. Second, the time of the student’s oral test was somewhat different due to the impact of the pandemic during the testing process, which may have had an impact. 후속연구는 각 학교의 테스트 시기와 방식을 더 철저히 동일하게 통일할 수 있어야겠다. The third questionnaire collection was conducted on students, so it was difficult to obtain more in-depth and objective results. Therefore, it seems that more objective data can be obtained by collecting and analyzing questionnaires from parents in follow-up studies.

Overall, this study emphasizes the important role of language comprehension in L2 reading development and emphasizes the need to address decoding and other reading skills in a variety of ways to improve EFL students’ English reading ability. Considering the significant influence of elementary school English learning experiences and home literacy environments on students’ reading ability, this study suggests that continuous support and customized interventions are needed to address the diverse challenges of lean EFL learners. This study contributes to understanding the gaps in English proficiency and developing strategies to bridge these gaps to provide more equitable learning opportunities for all students.

"The study titled 'Bridging the Gap: Analyzing English Reading Proficiency Disparities Among Korean EFL Learners' was conducted with the approval of the Seoul National University Institutional Review Board (IRB). The research was carried out in accordance with the guidelines and regulations set forth by the SNU IRB (Approval No: 2108/004-013), ensuring that all participants provided informed consent, and that their privacy and confidentiality were strictly maintained throughout the study."

Adlof SM, Catts HW, Little TD (2006) Should the simple view of reading include a fluency component? Read Writ 19(9):933–958. https://doi.org/10.1007/s11145-006-9024-z
Berlin KS, Williams NA, Parra GR (2014) An introduction to latent variable mixture modeling (Part 2): Longitudinal latent class growth analysis and growth mixture models. J Pediatr Psychol 39(2):188–203. https://doi.org/10.1093/jpepsy/jst085
Cain K, Oakhill J (2006) Profiles of children with specific reading comprehension difficulties. Br J Educ Psychol 76(4):683–696. https://doi.org/10.1348/000709905X67610
Carrell PL (1988) Interactive text processing: Implications for ESL/second language reading classrooms. In: Carrell PL, Devine J, Eskey DE (eds) Interactive approaches to second language reading. Cambridge University Press, pp 239–259
Chall JS, Jacobs VA, Baldwin LE (1990) The reading crisis: Why poor children fall behind. Harvard University Press
Cho J, Choi D, Lee K (2019) Early English education and its impact on language development. Korean J Engl Educ 18(2):101–121. https://doi.org/10.15858/engedu.2019.18.2.101
Cho J-R, McBride C, Park S-G (2019) Investigating the roles of morphological awareness in early literacy acquisition in Korean. Learn Individual Differences 70:25–33. https://doi.org/10.1016/j.lindif.2019.01.001
Choi Y (2007) The social and cultural factors influencing student participation in class discussions in an EFL context. Engl Teach 62(1):37–58
Cunningham AE, Stanovich KE (1998) What reading does for the mind. Am Educ 22(1/2):8–15
Ferguson SL, Moore EW, Hull DM (2020) Finding latent groups in observed data: A primer on latent profile analysis in Mplus for applied researchers. Int J Behav Dev 44(5):458–468. https://doi.org/10.1177/0165025419881721
Garcia JR, Cain K (2014) Decoding and reading comprehension: A meta-analysis to identify which reader and assessment characteristics influence the strength of the relationship in English. Rev Educ Res 84(1):74–111. https://doi.org/10.3102/0034654313499616
Gottardo A, Yan B, Siegel LS, Wade-Woolley L (1996) Factors related to English reading performance in children with Chinese as a first language: More evidence of cross-language transfer of phonological processing. J Educ Psychol 98(2):291–311. https://doi.org/10.1037/0022-0663.98.2.291
Gough PB, Tunmer WE (1986) Decoding, reading, and reading disability. Remedial Special Educ 7(1):6–10. https://doi.org/10.1177/074193258600700104
Grabe W (2009) Reading in a second language: Moving from theory to practice. Cambridge University Press
Grabe W (2014) Key issues in L2 reading development. In: Chapelle CA (ed) The Encyclopedia of Applied Linguistics. Wiley-Blackwell. https://doi.org/10.1002/9781405198431.wbeal0653
Hatcher L (2013) Advanced statistics in research: Reading, understanding, and writing up data analysis results. Shadow Finch Media
Hoover WA, Gough PB (1990) The simple view of reading. Read Writing: Interdisciplinary J 2:127–160. https://doi.org/10.1007/BF00401799
Jeon EH (2011) Components of reading comprehension in L2 learners: The effect of vocabulary knowledge and syntactic knowledge on passage comprehension test performance. Lang Test 28(3):395–421. https://doi.org/10.1177/0265532211399331
Jeon EH, Yamashita J (2014) L2 reading comprehension and its correlates: A meta-analysis. Lang Learn 64(1):160–212. https://doi.org/10.1111/lang.12034
Jiménez RT, García GE, Pearson PD (1996) The reading strategies of bilingual Latina/o students who are successful English readers: Opportunities and obstacles. Reading Res Q 31(1):90–112. https://doi.org/10.1598/RRQ.31.1.5
Juel C (1988) Learning to read and write: A longitudinal study of 54 children from first through fourth grades. J Educ Psychol 80(4):437–447. https://doi.org/10.1037/0022-0663.80.4.437
Koda K (2005) Insights into second language reading: A cross-linguistic approach. Cambridge University Press
Krashen SD (1993) The power of reading: Insights from the research. Libraries Unlimited
Landers RN (2008) Cluster analysis and latent class analysis. In: Lavrakas PJ (ed) Encyclopedia of Survey Research Methods. SAGE Publications, Inc. https://doi.org/10.4135/9781412963947.n56
Lee J (2002) The effects of early English education on reading comprehension in Korean EFL students. Engl Teach 57(1):29–54. https://doi.org/10.15858/engedu.2002.57.1.29
Lonigan CJ, Burgess SR (2017) Phonological sensitivity and reading: Individual differences in the phonological analysis and synthesis task. J Learn Disabil 50(1):69–84. https://doi.org/10.1177/0022219415572619
Lonigan CJ, Schatschneider C (2018) Predicting reading skills: A longitudinal investigation of the simple view of reading in children learning to read in a second language. J Educ Psychol 110(2):160–175. https://doi.org/10.1037/edu0000201
McLachlan GJ, Peel D (2000) Finite mixture models. Wiley-Interscience
Mikulecky BS, Jeffries L (1986) More reading power. Addison-Wesley Publishing Company
Muter V, Hulme C, Snowling MJ, Stevenson J (2004) Phonemes, rimes, vocabulary, and grammatical skills as foundations of early reading development: Evidence from a longitudinal study. Dev Psychol 40(5):665–681. https://doi.org/10.1037/0012-1649.40.5.665
Nylund KL, Asparouhov T, Muthén BO (2007) Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Struct Equation Modeling: Multidisciplinary J 14(4):535–569. https://doi.org/10.1080/10705510701575396
Nylund-Gibson K, Choi AY (2018) Ten frequently asked questions about latent class analysis. Translational Issues Psychol Sci 4(4):440–461. https://doi.org/10.1037/tps0000176
Park J (2012) The impact of private tutoring on students' academic achievement in Korea. Korean J Educational Policy 37(2):143–168
Peugh JL, Fan X (2013) How well does multilevel modeling support cluster-level inference? Evaluation of aggregated cluster-level variables. J Educational Behav Stat 38(2):81–101. https://doi.org/10.3102/107699861244
Perfetti CA (1985) Reading ability. Oxford University Press
Perfetti CA (2007) Reading ability: Lexical quality to comprehension. Sci Stud Read 11(4):357–383. https://doi.org/10.1080/10888430701530730
Protopapas A, Parrila R (2018) Is dyslexia a brain disorder? Brain Sci 8(4):61. https://doi.org/10.3390/brainsci8040061
Raven JC (2003) Raven's Progressive Matrices. RPM). Pearson Education
Scarborough HS (2001) Connecting early language and literacy to later reading (dis)abilities: Evidence, theory, and practice. In: Neuman SB, Dickinson DK (eds) Handbook of early literacy research, vol 1. Guilford Press, pp 97–110
Schatschneider C, Torgesen JK (2004) Using our current understanding of dyslexia to support early identification and intervention. J Learn Disabil 37(5):421–429. https://doi.org/10.1177/00222194040370050701
Seidenberg MS (2017) Language at the speed of sight: How we read, why so many can’t, and what can be done about it. Basic Books
Snow CE, Burns MS, Griffin P (eds) (1998) Preventing reading difficulties in young children. National Academy
Stanovich KE (1986) Matthew effects in reading: Some consequences of individual differences in the acquisition of literacy. Reading Res Q 21(4):360–407. https://doi.org/10.1598/RRQ.21.4.1
Stanovich KE (2000) Progress in understanding reading: Scientific foundations and new frontiers. Guilford Press
Taboada Barber A, Leu DJ (2020) An expanded view of reading comprehension: The impact of internet reading on assessment and instruction. In: Hiebert EH, Pearson PD (eds) Research-based practices for teaching reading. Teachers College, pp 182–202
Tunmer WE, Chapman JW (2012) The simple view of reading redux: Vocabulary knowledge and the independent components hypothesis. J Learn Disabil 45(5):453–466. https://doi.org/10.1177/0022219411432685
van Steensel R, Oostdam R, van Gelderen A (2017) The role of word decoding and language comprehension in second language listening and reading comprehension: A three-year longitudinal study. Read Writ 30(3):501–522. https://doi.org/10.1007/s11145-016-9687-y
Verhoeven L, Perfetti C (2017) Introduction to this special issue: Developmental patterns in the acquisition of reading comprehension. Learn Individual Differences 54:1–2. https://doi.org/10.1016/j.lindif.2017.01.002
Wagner RK, Torgesen JK, Rashotte CA (1999) Comprehensive Test of Phonological Processing (CTOPP). Pro-Ed
Wang M, Koda K, Perfetti CA (2003) Alphabetic and nonalphabetic L1 effects in English word identification: A comparison of Korean and Chinese English L2 learners. Cognition 87(2):129–149. https://doi.org/10.1016/S0010-0277(02)00232-9
Wang M, Yang C-L, Cheng C (2009) The contributions of phonology, orthography, and morphology in Chinese-English biliteracy acquisition. Appl Psycholinguist 30(2):291–314. https://doi.org/10.1017/S0142716409090120
Williams JP, Shankweiler D (1994) Children's sensitivity to word-level regularities in derivational suffixes. J Exp Child Psychol 57(3):465–484. https://doi.org/10.1006/jecp.1994.1036
Ziegler JC, Goswami U (2005) Reading acquisition, developmental dyslexia, and skilled reading across languages: A psycholinguistic grain size theory. Psychol Bull 131(1):3–29. https://doi.org/10.1037/0033-2909.131.1.3
▫ Choueiry G (2020), June 1 What is an acceptable value for VIF? Quantifying Health. https://quantifyinghealth.com/acceptable-vif-value/
▫ Hill WG, Robertson A (1968) The effect of linkage on limits to artificial selection. Genet Res 12(2):143–153. https://doi.org/10.1017/S0016672300010156
▫ Vittinghoff E, McCulloch CE (2005) Relaxing the rule of ten events per variable in logistic and Cox regression. Am J Epidemiol 165(6):710–718. https://doi.org/10.1093/aje/kwk052
▫ James G, Witten D, Hastie T, Tibshirani R (2013) An introduction to statistical learning: With applications in R (1st ed., Corr. 7th printing). Springer

The authors declare no competing interests.

Bridging the Gap: Analyzing English Reading Proficiency Disparities Among Korean EFL Learners

Status:

Version 1

Abstract

Figures

Introduction

Literature Review