Does Japanese EFL Learners’ Reading Comprehension
Improve Under Time Constraints?
Sawako HAMATANI (Kansai University, Graduate School)
Keywords L2 readers; reading comprehension; reading fluency; syntactic processing; time pressure
30
1. Introduction
According to previous research on first language (L1) reading rates, the average reading speed of an adult L1 speaker of English is about 180 words per minute (Higgins & Wallace, 1989; Nuttall, 2006). Carver (1992) argued that fluent readers read for general understanding at about 250–300 words per minute. However, it is often the case that second language (L2) students read in English at 80–120 words per minute, which is one half to one-third the rate of an L1 speaker (Beglar et al., 2011; Iwahori, 2008; Taguchi et al., 2004). One of the main reasons for this difference is that reading accuracy receives more attention than reading speed in L2 language learning contexts (Grabe, 2009; Nakamori, 2002). However, reading with fluency, or with speed and comprehension, is necessary for both L1 and L2 readers for several reasons.
First, if they have fluent reading skills, L2 readers can process texts quickly and accurately. This ability is necessary to process abundant information in the digital age. Second, fluent reading allows readers to experience a much larger amount of L2 input, which increases the chance of incidental vocabulary learning
(Nation, 2001). Third, fluency enables readers to enjoy the activity of reading. Slow readers read less because they seldom find much interest or pleasure in what they read. They can escape this vicious spiral if they attain fluency in reading. Thus, developing reading fluency should significantly improve general reading ability. This holds true for Japanese English as a foreign language (EFL) learners because their reading rate has been shown to be very low (Beglar et al., 2011; Iwahori, 2008; Taguchi et al., 2004). Reading fluency is the overarching theme of this study.
This study also focuses on time constraints, because it has been demonstrated that imposing time constrains can help reveal how readers process texts. Several studies indicate that placing a time constraint on reading time influences readers’ processing and improves comprehension (Breznitz, 1987; Breznitz & Share, 1992; Breznitz et al., 1993). Other scholars (Walczyk, 1993, 1995, 2000; Walczyk et al., 2001; Walczyk & Taylor, 1996) have found that time constraints alter less-efficient readers’ processing and deteriorate comprehension. These studies suggest that time constraints affect L1 and L2 readers’ processing and comprehension of texts; however, no such study has targeted Japanese EFL learners. Therefore, this study aimed to determine what would happen to the reading comprehension of Japanese intermediate EFL learners when time pressure was added, and to establish its relationship to reading fluency.
Finally, this study addresses syntactic processing, because the automatization of it is necessary for fluent reading. Syntactic processing is essential to reading comprehension, as a means through which a reader accesses meaning as well as information from words and sentence structures. Performing syntactic processing automatically (or at least fluently), as well as other lower-level processing, is necessary because it saves storage for working memory, which can be used for higher-level reading comprehension (Grabe, 2009). However, it is assumed that Japanese EFL learners have difficulty developing fluent syntactic processing skills because of the differences between the Japanese and English syntactic systems. Does syntactic processing fluency affect passage-reading fluency?
1.1 Defining Reading Comprehension
There are several types of reading: scanning, skimming, reading to learn, reading to integrate information, reading to evaluate, and reading for general comprehension (Grabe, 2009). Among these, reading for general comprehension is the default meaning of the term “reading comprehension” (Carver, 1992). It takes place when we read a novel, newspaper article, or magazine. According to Grabe (2009), “reading for general comprehension is also a type of reading that is carried out automatically for extended periods of time and with apparently few processing difficulties” (p. 10). This type of reading is the focus of this study.
1.2 Fluency, Automaticity, and Working Memory
There is no unanimous agreement on the definition of reading fluency. In the L1 field, it is commonly accepted that “the primary components of fluency are (a) accuracy in decoding, (b) automaticity in word recognition, and (c) the appropriate use of prosodic features” (Kuhn, & Stahl, 2003, p. 5). In the L2 context, Grabe (2009) explained that “fluency in reading is the ability to read rapidly with ease and accuracy, and to read with appropriate expression and phrasing. It involves a long incremental learning process, and text comprehension is an expected outcome of fluent reading” (p. 291). In this study, fluent reading is interpreted as the ability to comprehend a text’s meaning rapidly and accurately. This study excluded prosody because it targeted EFL learners’ silent reading to measure their general comprehension.
Reading comprehension is a complex process. Therefore, researchers have examined reading in terms of its component skills, which are divided into two groups: lower-level and higher-level processes, along with their relationships with working memory (Grabe, 2009; LaBerge & Samuels, 1974; Stanovich, 2000). Lower-level processes include word recognition, syntactic parsing, and meaning encoding as propositions. Higher-level processes are text-model formation, situation-model building, inferencing, executive-control processing, and strategic processing. Both process groups operate in working memory simultaneously, interacting at certain points (Grabe, 2009). Working memory includes information that is active during processing operations and the processing directions themselves (Baddeley, 2007). Working memory has limited storage and thus limited ability to execute multiple processes simultaneously, or even nearly so. This characteristic of working memory makes automaticity crucial for fluent reading (Grabe, 2009; LaBerge & Samuels, 1974; Samuels, 1994).
Automaticity encompasses rapid, relatively resource-free, unconscious, and unstoppable operations. Automatized processes do not require a large number of processing resources from working memory and can hence be executed while a reader concentrates on other tasks (Grabe, 2009; Logan, 1997). While reading, an individual performs at least two interdependent tasks simultaneously: determining what words comprise the text (i.e., lower-level processes) and constructing meaning (i.e., higher-level processes; Kuhn & Stahl, 2003). By automatizing lower-level processes, a reader can more easily direct attention toward higher-level processes and achieve adequate comprehension (Grabe, 2009; LaBerge & Samuels, 1974; Stanovich, 2000).
1.3 Reading With Time Constraints
In a comprehensive series of studies, Breznitz (1987; Breznitz et al., 1993; Breznitz & Share, 1992; Leikin & Breznitz 2001) demonstrated that a moderately
accelerated reading rate boosted reading comprehension. In Breznitz’s studies, first graders in L1 Hebrew (Experiment 1) and L1 English (Experiment 4) improved their reading comprehension at a faster speed than at their normal speed. Breznitz and Share suggested that fast-paced manipulation enhanced the efficiency of various cognitive processes that are involved in reading, particularly concerning working memory. Breznitz et al. investigated cerebral activity during reading to explore the acceleration phenomenon in greater depth. They measured the event-related potentials (ERP) and electroencephalographic (EEG) components of eight college-level native English speakers. The data imply that college-level readers could be prompted to read 10% faster than they would at their regular rates, with increased comprehension. Leikin and Breznitz utilized ERP techniques to investigate the effect of a fast-paced reading rate on the recognition of a word’s grammatical functions. The participants were 20 male, native Hebrew-speaking college students. The results showed that an accelerated reading rate altered the participants’ reading strategy from the predicate-oriented (which is normal in Hebrew) to word order-oriented. This shift in processing suggests that reading rates can influence aspects of sentence processing. From her findings, Breznitz posited that “reading rate may also be perceived as an independent variable influencing the quality of decoding and comprehension reading skills” (Breznitz, 2006, p. 18).
In contrast, Walczyk (1993, 1995, 2000; Walczyk et al., 2001; Walczyk & Taylor, 1996) discovered that less-efficient readers’ comprehension deteriorated under severe time pressure, but was not impaired under no time pressure. These results correspond with Walczyk’s Compensatory-Encoding Model (1993, 1995, 2000), which hypothesizes that in the case of intermediate or advanced readers, subcomponent inefficiency typically does not affect performance during reading, as compensatory mechanisms operate when necessary. If readers can use compensatory strategies (such as slowing their reading rate or reviewing previous passages), then less-efficient reading subcomponents, a small working memory capacity, limited vocabulary, inefficient syntactic parsing, or a lack of relevant background knowledge need not hinder comprehension. However, under time constraints, readers cannot use these strategies; thus, their comprehension declines (Walczyk, 1993, 1995, 2000; Walczyk et al., 2001; Walczyk & Taylor, 1996). The Compensatory-Encoding Model extended the explanations put forth by automaticity theory (LaBarge & Samuels, 1974) and interactive theories (Perfetti, 1988; Stanovich, 1980). Those theories primarily apply to novice readers or do not provide satisfactory accounts of how automatic and controlled processes interact. Many reading interventions conducted today are categorized either as enhancing lower-level automaticity (i.e., automatic processes), or as instructing higher-level, metacognitive strategies (i.e., controlled processes). Nevertheless, they do not always improve comprehension. The Compensatory-Encoding Model helps to evince
when either type of intervention is likely to succeed and how they interact with each other. Consequently, the model aims to explicate why gains in decoding skills do not concur with better passage comprehension (Frederiksen & Warren, 1987).
1.4 Syntactic Processing
Syntactic parsing is an essential element of lower-level processing, impacting reading speed and comprehension. Syntactic information provides instructions for the construction of text comprehension (Grabe, 2009). In other words, syntax plays a role in directing how a reader parses a text into meaning units (Kintsch, 1995). Learning English syntax is difficult for Japanese EFL learners. Previous studies have found that among syntactic structures, the noun phrase is especially difficult for Japanese learners to grasp (Chujyo, et al., 2012; Clahsen & Felser, 2006; Izumi, 2003; Kimura & Kanatani, 2006). Although both pre- and post-modification are used in English (Nuttall, 2006), pre-modification is mostly used in Japanese. Further, the canonical word order is violated in some English post-modification structures. Izumi suggested that EFL learners with intermediate proficiency presumably adopt a word order-based processing in using English. However, this might not work well in fluent processing of post-modification structures, and it may take extra time to use reading strategies, which depend on information other than syntax. In this study, it is assumed that Japanese EFL readers have the necessary syntactic knowledge, but it might not be sufficient for fluent processing. This study aims to explore how syntactic processing fluency affects passage-reading fluency in the case of Japanese EFL learners.
2. Previous Studies 2.1 L2 Reading Fluency
In the L1 context, reading fluency has gained greater recognition in recent decades. Several studies and reviews of reading fluency have been conducted (Kuhn & Stahl, 2003; Samuels, 2006). In L2 settings, however, relatively little research has addressed reading fluency. Recently, it has started to receive more attention, and many researchers now emphasize the need to conduct more research on reading fluency and to pay more attention to reading fluency instruction (Chang, 2010; Chang & Millet, 2013; Shimono, 2018; Taguchi et al., 2004).
A leading L2 researcher, Segalowitz (2010), asserted that fluency gains represent changes in cognitive processing and can therefore be assessed in terms of automaticity development. Favreau and Segalowitz (1983) compared two groups of L2 readers (stronger and weaker) in a primed lexical decision task and found that automaticity in word recognition remained a substantial factor in distinguishing reading fluency. Fukkink et al., (2005) performed a two-day training study that
taught 41 eighth grade EFL Dutch students to read words faster. Akamatsu (2008) investigated word recognition fluency by training 49 first-year university students in Japan.
The studies mentioned above all focused on word reading fluency; only a few have explored L2 passage-reading fluency. Research by Taguchi and his colleagues (Taguchi, 1997; Taguchi & Gorsuch, 2002; Taguchi et al., 2004) demonstrated that low-grade readers’ repeated reading leads to significant improvements in their silent reading rates. Beglar et al. (2011) examined the effects of a one-year pleasure reading program on the reading-rate development of first-year Japanese university students. The results showed that all experimental groups made more significant gains than the control group. According to Chang (2010), EFL college students who participated in a 13-week timed reading activity improved their reading speed on average by 25% and their comprehension by 4%. Chang and Millet (2013) investigated the impact of repeated reading on EFL college students and found that it increased their reading rates and comprehension levels more than those of the non-repeat students. Over one academic semester, Shimono (2018) observed the effect of timed and repeated oral reading among Japanese university students with lower proficiency levels. He discovered that the reading rate and comprehension of the experimental groups increased significantly more than those of the control group. Both practices were found to be efficient in promoting foreign language reading fluency, and there was no statistically significant difference between the timed reading group and the repeated oral reading group.
The studies mentioned above investigated how interventions improved EFL learners’ overall reading fluency, but they did not focus on reading components. Componential analysis is helpful in unearthing the problems that affect EFL learners’ reading fluency, because the automatization of lower-level processes (i.e., components) is considered a prerequisite for fluent reading comprehension. Nassaji and Geva (1999), in utilizing this approach, studied the role of phonological and orthographic processing skills among L1-Farsi readers, and found that efficiency (i.e., speed and accuracy) in orthographic and phonological processing contributed to their reading comprehension. Noro (2000) conducted a similar study targeting Japanese EFL learners, and found that efficiency in orthographic processing skills was moderately correlated with reading comprehension
2.2 L2 Reading With Time Constraints
To the best of the author’s knowledge, no studies have investigated L2 reading rates and comprehension under time constraints. That said, Breznitz (1987; Breznitz et al., 1993; Breznitz & Share, 1992; Leikin, & Breznitz 2001) argued that accelerating the reading rate boosts the efficiency of various cognitive processes involved in reading. Walczyk (1993, 1995, 2000; Walczyk et al., 2001) contended
that introducing severe time constraints can worsen comprehension since the reader cannot use strategies to compensate for the weaker components of his or her reading comprehension. Hence, it is possible to use the procedure of imposing time constraints to reveal weaker components of reading comprehension, or obstacles to fluent reading that Japanese EFL learners face. Previous studies (Akamatsu, 2008; Fukkink et al., 2005) have shown that EFL readers’ word recognition speed improved after instruction, which implies that their word recognition was not fluent. No study, however, has looked at the role of syntactic processing fluency in EFL learners’ comprehension, except for Nassaji and Geva (1999), who discovered that syntactic processing efficiency (i.e., speed and accuracy) affects reading comprehension. The componential analysis approach, which Nassaji and Geva employed is helpful but it needs relatively large-scale experiment. It would be worth researching how time constraints affect the reading comprehension of Japanese EFL learners’, and particularly the reading comprehension of passages with complex syntactic structures which are presumed to be difficult for them. Therefore, if the comprehension of passages with syntactically complex sentences deteriorates under time constraints, and if the comprehension of passages without such sentences does not decline under time constraints, this means that Japanese EFL learners’ syntactic processing is weak or not fluent, and compensatory strategies that are normally used may not have been employed. This raises a question: Do time constraints differentially affect Japanese intermediate EFL learners’ comprehension of passages that differ in syntactic difficulty?
3. Method 3.1 Participants
In total, 32 Japanese EFL learners (11 males and 21 females aged 18 to 19) attending a private university in western Japan participated in the study. At the time of the study, they had studied English for about 6-and-a-half years, mainly through highly controlled formal education. They were enrolled in a required English course, the purpose of which was to improve their reading skills. Their major was literature, and their average score of Global Test of English Communication (GTEC) Academic was 217.7 out of 500 (reading 250, listening 250), equivalent to a level of B1 based on the Common European Framework of Reference (CEFR). They signed consent forms after receiving an explanation of the study’s purpose and assurance of the protection of their privacy.
Two of the participants were excluded from the analysis. One was absent from the second session onward, and the other marked the same answer for all questions in the second session. The participants were divided into two groups, Group 1 (n = 14) and Group 2 (n = 16), to counterbalance the effect of the passages’
characteristics; lots were drawn to randomize the individual differences. In the first week, under the no-pressure condition, Group 1 read passages D1 and E1, and Group 2 read passages D2 and E2. In the second week, Group 1 read passages D2 and E2, and Group 2 read passages D1 and E1 under the condition of time pressure. 3.2 Materials
Four passages were adapted from Book 2 of Reading for Speed and Fluency by Nation and Malarcher (2007). The book is written at the 1,000-word level of the British National Corpus (BNC)/Corpus of Contemporary American English (COCA) word family lists and has passages of about 300 words with five comprehension questions each. It is the second in a four-book series designed to help people practice fluent reading. The content involves eight familiar topics: art, money, communication, health, nature, people, space, and transportation.
Four passages taken from the categories of art, health, nature, and transportation were modified into two types of passages: Type D and Type E. Type D passages (i.e., D1 and D2) include noun phrases led by relative pronouns and a sentence with a cleft construction, both of which are considered difficult for Japanese EFL learners (Chujyo et al., 2012; Clahsen & Felser, 2006; Izumi, 2003; Kimura & Kanatani, 2006). D1 has six noun phrases led by a relative pronoun and one sentence with a cleft construction, while D2 has seven noun phrases led by a relative pronoun and one sentence with a cleft construction. Type E passages (i.e., E1 and E2) do not contain sentences with those constructions, except for one short phrase led by a relative pronoun at the end of E2. Therefore, Type D passages are syntactically more difficult than Type E passages.
Table 1 shows the passages’ characteristics. There seems to be no significant difference between the readability scores of Type D and Type E, either in Flesch or Flesch-Kincaid. These were calculated with “words per sentence” and “syllables per word,” which do not indicate syntactic complexity. In order to show sentence complexity in numbers, clauses per sentence (C/S) were calculated by dividing the number of clauses by the number of sentences (Lu, 2011). C/S demonstrates that D1 and D2 have much more complex sentences than E1, whereas E2 seems to have a moderately complex structure. E2 has a relatively high score compared to E1, because E2 has two coordinate clauses and eight adverbial clauses. E1 has only one objective clause and five adverbial clauses. C/S signals a passage’s complexity, but might not well represent processing difficulty for Japanese EFL learners because it includes adverbial and other phrases, which are not assumed to be especially difficult for them compared to noun phrases.
Two inquiries were added to the original five multiple-choice questions because it was considered statistically advantageous to have seven questions for measuring comprehension. In D1, D2, and E2, the fifth question was replaced; it
concerned vocabulary instead of comprehension. The questions were framed in such a way that the participants could answer them correctly as long as they understood what the passage said, without depending heavily on memory.
D1 D2 E1 E2
Text Leonardo da
Vinci
The Secrets to a Long Life
The Chunnel Earthquakes
Words 295 308 298 299
Flesch 81.60 85.30 77.60 80.10
Flesch-Kincaid 4.90 4.10 5.10 5.10
C/S 1.72 1.72 1.22 1.52
Note. C/S represents “clauses per sentence.”
3.3 Procedure
The participants were seated in front of a personal computer in the university computer classroom known as the Computer Assisted Language Learning (CALL) room. The task for the first week was self-paced reading. Sheets of paper on which the comprehension questions were typed were handed out to the participants before the session started. The reading program was developed using Microsoft PowerPoint 2016. Each passage was displayed on the monitor screen of the personal computer line by line (Calibri, 24 point), and each line could contain up to 90 characters or about 20 words.
The participants were instructed to read at their normal speed and not to memorize the content of the passages because it could slow their reading speed. Before the session began, they practiced reading using the same program on the computer. The test began after they opened the file of the passage they were to read and clicked on the first slide, which had the title of the passage. The second slide held the first line of the passage. The participants clicked on the screen as soon as they finished reading the first line on the slide; then the next slide appeared containing the second line. The line the reader had just read remained above the new line so that the reader could look back at it (see Figure 1). In this way, the participants read through the slides until the end. As soon as each participant finished reading the passage, he/she saw the timer on the screen on the anterior wall, wrote down his/her reading time, and answered the comprehension questions printed on the paper. The participants read two passages, one of Type D followed by one of Type E. Group 1 read D1 and E1, while Group 2 read D2 and E2, both without time pressure in the first week. In the second week, the participants read the two other passages (Group 1: D2, E2; Group 2: D1, E1) under a 20% time-
Figure 1.The third slide, with the second line underneath the first line
Table 2. Means and Standard Deviations of Impressions of Pressured Speed
Group 1 Group 2
D2 E2 D1 E1
M (SD) M (SD) M (SD) M (SD)
Impression 2.14 (.64) 2.00 (.53) 1.44 (.50) 1.63 (.48)
Note. 3-point Likert scale, 1: too fast, 2: appropriate, 3: too slow.
pressured condition, or at 80% of their normal reading time (the time they took to read the passages in the first week). The purpose of exchanging passages between the groups was to counterbalance the effects of the materials. The passages were shown on the screen line by line as in the first week, but this time, the slides progressed automatically according to their reading speed. They read until the end, marked on a 3-point Likert scale how they felt about the experience (1: too fast, 2: appropriate, or 3: too slow), and answered the comprehension questions. Table 2 portrays the participants’ impressions when they read under pressure. Group 1 felt less pressured than Group 2, but there seemed to be no difference between Type D and Type E. The 20% time-pressure was calculated as follows: The participants were divided into three groups of fast, moderate, and slow readers according to their normal reading speed (the rate at which they read in the first week). The mean reading time of each group was then multiplied by 80%. This procedure was applied separately to the Type D and Type E passages, which is why a few participants belonged to different groups for Type D and Type E. Table 3 depicts the reading time and words per minute (WPM) of each passage and each group.
No te . n = 32. WPM = wor ds per m inu te Gro u p 1 G ro u p 2 D 1 E 1 D 2 E 2 T im e W P M T im e W P M T im e W P M T im e W P M n M ( SD ) M ( SD ) n M ( SD ) M ( SD ) n M ( SD ) M ( SD ) n M ( SD ) M ( SD ) Hi gh 7 128 .60 (13. 77 ) 139 .20 (17. 29 ) 7 138 .00 (14. 30 ) 131 .10 (14. 91 ) 5 125 .80 (7.9 6) 147 .90 (9.2 8) 6 132 .50 (7.1 6) 135 .80 (7 .5 0) Mo de ra te 3 14 7. 30 (6.9 4) 120 .40 (5 .5 2) 5 182 .80 (6 .7 9) 97 .95 (3 .8 0) 7 162 .71 (12. 14 ) 11 4. 22 (8.6 4) 5 160 .60 (7.3 9) 111 .9 4 (5.0 9) Lo w 5 19 8. 40 (15. 03 ) 89 .72 ( 6. 73 ) 3 214 .67 (19. 40 ) 83 .94 (7 .1 5) 5 208 .40 (4.4 1) 88 .72 (1 .8 7) 6 192 .67 (18.02 ) 93 .88 (8 .1 79 ) T ab le 3 . R ea di n g T im e an d W P M o f E ac h P as sa ge a n d E ac h G ro u p
4. Results
Table 4 outlines the means and standard deviations (SDs) of the participants’ reading comprehension scores under no pressure (the first-week conditions) and under pressure (the second-week conditions). The mean reading comprehension scores of Type E1 (Group 1, no-pressure, M = 4.64, SD = 1.04; Group 2, pressure, M
= 4.25, SD = 0.97) were considerably lower than the scores of Type E2 (Group 2, no-pressure, M = 5.81, SD = 1.29; Group 1, pressure, M = 6.21, SD = 0.77). This may have been caused by the difference in familiarity between the topics of E1 and E2. The topic of E1 is “Chunnel,” or the Channel Tunnel between England and France, which might not have been of great interest to Japanese EFL students compared to the topic of E2, earthquakes. The estimated reliability (i.e., Cronbach’s ) of the Type D and Type E comprehension tests (each of which consists of 14 questions) was 0.71 and 0.42, respectively. The low reliability score of Type E tests may be attributable to E1, as explained above.
Table 4. Means and Standard Deviations of Participants’ Comprehension
No Pressure Pressure n M (SD) M (SD) Group 1 14 D1 6.36 (0.81) D2 6.36 (0.61) 14 E1 4.64 (1.04) E2 6.21 (0.77) Group 2 16 D2 6.31 (0.58) D1 5.19 (1.55) 16 E2 5.81 (1.29) E1 4.25 (0.97)
Table 5 presents the descriptive statistics for the comprehension of the two types of passages under the two different conditions. The mean reading comprehension score of Type D under pressure (Group 1, D2; Group 2, D1; M = 5.73,
SD = 1.36) was lower than that of Type D under no pressure (Group 1, D1; Group 2, D2; M = 6.33, SD = 0.71). The mean reading comprehension score of Type E under pressure (Group 1, E2; Group 2, E1; M = 5.17, SD = 1.34) was lower than that of Type E under no pressure (Group 1, E1; Group 2, E2; M = 5.27, SD = 1.34).
Table 5. Descriptive statistics for the comprehension of two types of passages under two different conditions
M SD Min Max Skewness Kurtosis
No pressure Type D (D1+D2) 6.33 0.71 4.0 7.0 -1.210 2.559
Type E (E1+E2) 5.27 1.34 3.0 7.0 -0.248 -0.939
Pressure Type D (D2+D1) 5.73 1.36 2.0 7.0 -1.143 0.613
To answer the research question, “Do time constraints differentially affect Japanese intermediate EFL learners’ comprehension of passages that differ in syntactic difficulty?”, two non-parametric Wilcoxon paired signed-rank tests were conducted at an alpha level set at 0.025. As displayed in Table 6, the test indicated that the difference between Type D (no time pressure) and Type D (time pressure) was statistically significant (Z = 2.274, p < 0.025, r = 0.42), but the difference between Type E (no time pressure) and Type E (time pressure) was not statistically different (Z = 0.365, p = ns, r = 0.07). The results suggest that time constraints differentially affect Japanese EFL learners’ comprehension of passages that differ in syntactic difficulty.
Table 6. Statistics of the Wilcoxon paired signed-rank tests
Z p r
Np-D – P-D 2.274 0.023* 0.42
Np-E – P-E 0.365 0.715 0.07
Note. Np = no time pressure, P = time pressure *p < 0.025
5. Discussion
The purpose of this study was to investigate the effect of time constraints on EFL learners’ reading comprehension and to discern the difference between the outcomes of syntactically different passages. The results indicate that a 20% time- pressure has neither a positive nor a negative effect on EFL readers’ comprehension when a text’s construction is syntactically simpler; however, it has a negative effect when a text’s construction is syntactically more complex.
In the L1 context, Breznitz (1987) showed that having first graders read faster than they would at their usual pace boosted their comprehension, whereas obliging them to read slowly decreased it. The reason for this is that “fast-paced reading increases the units available in short-term memory, and thus enlarges the context within which the reading process takes place” (Breznitz, 1987, p. 241). Breznitz performed further studies to examine the effect on regular adult readers in the United States reading English (Breznitz et al., 1993), and university-level adult regular and dyslexic readers in Israel reading Hebrew (Leikin & Breznitz, 2001). According to the findings, the adult university-level regular readers could read about 10% faster in either language than their self-paced reading rate, with significantly improved comprehension (Breznitz et al., 1993; Leikin & Breznitz, 2001). Walczyk (1993, 1995) also targeted L1 readers, but proposed the Compensatory-Encoding Model (Walczyk, 1993, 1995, 2000) to capture the results
of less-efficient readers. The model signaled that less-efficient readers’ comprehension declined when under severe time pressure. In that condition, they could not employ the behaviors or strategies they normally used to compensate for their weak reading components since doing so would take up extra time. The comprehension of readers with stronger reading automaticity skills would not decrease when under time pressure, as stronger readers would not use time-consuming compensatory mechanisms or strategies, whether under time pressure or not.
In the present study, the reading comprehension of intermediate EFL learners did not improve when under time pressure. These results contradict the hypotheses of Breznitz (1987; Breznitz et al., 1993; Breznitz & Share, 1992; Leikin, & Breznitz 2001). Differences in the experimental conditions between Breznitz’s research and the present study may have played a role. First, in Breznitz and Breznitz and Share, the participants were elementary school children, whereas the participants in this study were university students. Thus, there may have been differences in the participants’ cognitive processes due to the age difference. Second, in Breznitz et al. and Leikin and Breznitz, the pressure time was 10% compared to 20% in this study. The difference in pressure time may have caused the discrepancy, and it can be inferred that the threshold at which a time pressure effect turns into an enhancement may exist somewhere between the two pressure times. Third, Breznitz and Leikin and Breznitz conducted their experiments individually and calculated pressure time individually, whereas this study only set three time-pressure conditions for three groups, which were divided according to reading times recorded in the first week. As a result, pressure times may have been exceedingly fast for some participants and not fast enough for others. This may have obscured the time pressure effects. Lastly, all of Breznitz’s studies targeted L1 readers, while EFL readers participated in this study. L1 and L2 reading share many universal aspects that mainly concern the nature of cognitive processing mechanisms. However, there are many differences between L1 and L2 reading, linguistic, developmental, and sociocultural (Grabe, 2009). These differences may have caused the discrepancy between Breznitz’s outcomes and those of this study.
The reading comprehension of syntactically difficult passages significantly decreased when the participants were under time pressure. It was not possible to apply the Compensatory-Encoding Model (Walczyk, 1993, 1995, 2000) in its entirety because the time pressure in this study (20%) was different from the 50% time pressure used in Walczyk (1995). According to the model, however, this implies that when under time pressure, the participants could not use compensating strategies such as rereading or taking extra time, which they use when there are no time constraints; consequently, their comprehension declined. Considering their GTEC score (217.7 out of 500: reading 250, listening 250), the participants had enough
knowledge to process the syntactically difficult passages, but did not utilize that knowledge to process them quickly and accurately.
Leikin and Breznitz (2001) showed that the acceleration manipulation quickened the processing speed of syntactic parsing, which in turn boosted comprehension. They also revealed that accelerating the reading rate influences the recognition of words’ grammatical functions. Under the self-paced condition in Hebrew, the participants used a predicate-centered (morphological-based) strategy, whereas under the fast-paced condition, they used a word order strategy. Although the reading comprehension of syntactically difficult passages decreased when the participants were under time pressure in this study, the time pressure was more severe, and the decline was primarily attributable to the D1 passage, whose unfamiliarity (or other aspects) may have decreased participants’ comprehension. It would be worth employing ERP measures in the future to investigate the effect of a 10% time-pressure on the comprehension of passages, including syntactically complex sentences.
Several pedagogical implications can be drawn from the results of the present study. First, introducing a program to enhance syntactical processing fluency should be considered; for example, it could involve repeated or extensive reading. Research has shown that slow readers improve their reading rate by receiving fluency instructions from teachers (Beglar et al., 2011; Chang, 2010; Chang & Millet, 2013; Nuttall, 2006; Shimono, 2018; Taguchi et al., 2004). The fluency instructions were not focused on syntactic processing, but the improvement may have been related to syntactic processing to some degree. As Nakamori (2002) pointed out, rule-based teaching has long dominated Japanese schools. Relative clauses have been introduced through the two-sentence connection, and many mechanical exercises in connecting sentences have been taught to students. The situation is changing now, and task-based learning is gaining popularity in formal education. Task-based learning focuses on meaning, but fluency training for syntactic processing is still lacking. Japanese learners seem to have adequate training to acquire knowledge of relative clauses, yet as this study’s findings imply, they do not have enough knowledge to develop processing fluency. As mentioned above, although syntactic processing is not the focus of repeated or extensive reading, introducing it as a syntactic processing fluency training could be a possibility after investigating how it affects syntactic processing fluency. Second, researchers should consider administering time-pressured reading exercises in which the learners’ reading time is set 10% to 20% faster than their usual reading time. In this study, the reading comprehension of less complex passages did not significantly deteriorate under 20% time-pressure. This result implies that readers can read relatively easy passages 20% faster than their normal rate without sacrificing comprehension. Having learners read a text under a 20% time-pressured condition
could become part of reading fluency training. It would be possible to increase the time pressure to 25% based on Chang (2010), who showed that timed reading activities improved EFL college students’ reading speed on average by 25%.
This study has several limitations. First, it was not an exact replication of time-pressure studies. This study differs from the work of Breznitz and her colleagues (Breznitz 1987; Breznitz et al., 1993; Breznitz & Share, 1992; Leikin, & Breznitz 2001) in the length of the materials and data collection methods. This study was not a precise replication of Walczyk (1995) either, in that the pressure times were different: 50% in Walczyk and 20% in this study. Second, this study only indirectly indicates compensatory strategy use due to imposed time pressure; whether the participants used such strategies cannot be known unless qualitative data (e.g., interviews, think-aloud protocol, or stimulated recall) are collected. Third, reading is a complex process; hence, low syntactic processing fluency was not the only reason for the slow reading rate of Japanese EFL learners. According to L2 theories of reading, word recognition efficiency contributes to comprehension performance (Koda, 1996; Segalowitz, 2000). Nassaji and Giva (1999) revealed that the efficiency of phonological and orthographic information processing could cause individual differences in English as a second language (ESL) reading.
Other limitations are that it is impossible to entirely control for the effects of the difference between the two groups of participants; however, the grouping was decided on by drawing lots to avoid uneven grouping. In terms of the reading materials’ effects, the comprehension of E1 (for which the topic was the “Chunnel”) was exceptionally low.
One opportunity for clarification in the future would be the extent to which each lower-level component contributes to processing speed. A componential analysis would be necessary for undertaking such research. Qualitative research should also be performed on how readers process syntactically difficult texts and possible impairments when under time pressure.
6. Conclusion
The effect of time pressure on Japanese EFL learners’ reading comprehension was not significant when the materials included passages without syntactic difficulty, but it declined significantly when the materials included passages with more syntactic difficulty. The findings suggest that the participants in this study may have used the strategy of rereading or slowing down their reading rate while reading syntactically more difficult passages. This outcome is in line with the Compensatory-Encoding Model (Walczyk, 1993, 1995, 2000) and may be one source of Japanese EFL learners’ slow reading rate. Introducing fluency instructions for
reading syntactically complex texts should be considered. Nevertheless, the comprehension of syntactically simpler passages was not hindered when the participants were under time pressure, which implies that these readers could read 20% faster than their average speed. Mildly time-pressured reading could be one training method for reading fluency.
A componential analysis of reading components and a qualitative approach would be necessary to further clarify the reasons for the low reading rates of Japanese EFL learners.
References
Akamatsu, N. (2008). The effects of training on automatization of word
recognition in English as a foreign language. Applied Psycholinguistics, 29(2), 175–193.
Baddeley, A. (2007). Working memory, thought and action. Oxford, UK: Oxford University Press.
Beglar, D., Hunt, A., & Kite, Y. (2011). The effect of pleasure reading on Japanese university EFL learners’ reading rates. Language Learning, 62(3), 665–703. Breznitz, Z. (1987). Increasing first graders’ reading accuracy and comprehension by accelerating their reading rates. Journal of Educational Psychology 79(3),
236–242.
Breznitz, Z. (2006). Fluency in reading: Synchronization of processes. Mahwah, NJ: Lawrence Erlbaum.
Breznitz, Z., DeMarco, A., & Hakerem, G. (1993). Topographic measures of cerebral activity during reading of text at fast-and slow paced rates. Brain Topography, 6(2), 117–121.
Breznitz, Z., & Share, D. (1992). Effects of accelerated reading rate on memory for text. Journal of Educational Psychology, 84(2), 193–199.
Carver, R. (1992). Reading rate: Theory, research, and practical implications.
Journal of Reading, 36(2), 84–95.
Chang, A. (2010). The effect of a timed reading activity on EFL learners: Speed, comprehension, and perceptions. Reading in a Foreign Language 22(2), 284– 303.
Chang, A., & Millet, S. (2013). Improving reading rates and comprehension through timed repeated reading. Reading in a Foreign Language, 25(2), 126– 148.
Chujyo, K., Yokota, K., Hasegawa, S., & Nishigaki, C.,(2012). Identifying the General English Proficiency and Distinct Grammer of Remedial Learners.
JOURNAL OF THE COLLEGE OF INDUSTRIAL TECHNOLOGY NIHON UNIVERSITY B, 45, 43–54.
Applied Psycholinguistics, 27(1), 3–42.
Favreau, M., & Segalowitz. N. (1983). Automatic and controlled processes in the first-and second-language reading of fluent bilinguals. Memory and
Cognition, 11(6), 565–574.
Frederiksen, J. R., & Warren, B. M. (1987). A cognitive framework for developing expertise in reading. In R. Glaser (Ed.), Advances in instructional psychology (Vol. 3, pp. 1–39). Hillsdale, NJ: Lawrence Erlbaum.
Fukkink, R., Hulstijn, J., & Simis, A. (2005). Does training in second-language word recognition skills affect reading comprehension? An experimental study.
The Modern Language Journal, 89(1), 54–75.
Grabe, W. (2009). Reading in a second Language: Moving from Theory to Practice.
New York, NY: Cambridge University Press.
Higgins, J., & Wallace, R. (1989). Hopalong: A computer reading pacer. System, 17(3), 389–399.
Iwahori, Y. (2008). Developing reading fluency: A study of extensive reading in EFL. Reading in a Foreign Language, 20(1), 70–91.
Izumi, S. (2003). Processing difficulty in comprehension and production of relative clauses by learners of English as a second language. Language Learning, 53(2), 285–323.
Kimura, M., & Kanatani, K. (2006). Eigo no ku kozo ni taisuru nihonjin chugakusei no rikaido chosa [A survey on Japanese junior high school students’ knowledge of English phrase structures: Identifying time-gaps between instruction and acquisition]. Kanto-Koshinetsu Association of Teachers of English Bulletin, 20, 101–112.
Kintsch, W. (1995). How readers construct situation models for stories: The role of syntactic cues and causal inferences. In M. A. Gernsbacher & T. Givon (Eds.),
Coherence in spontaneous text (pp. 139–60). Philadelphia: J. Benjamins. Koda, K. (1996). L2 word recognition research: A critical review. The Modern
Language Journal,80(4), 450–460.
Kuhn, M., & Stahl, S. (2003). Fluency: A review of developmental and remedial practices. Journal of Educational Psychology, 95(1), 3–21.
LaBerge, D., & Samuels, S. J. (1974). Toward a theory of automatic information processing in reading. Cognitive Psychology, 6(2), 293–323.
Leikin, M., Breznitz, Z. (2001). Effects of accelerated reading rate on processing of Hebrew sentences: Electrophsiological evidence. Genetic, Social, and General Psychology Monographs, 127(2), 193–209.
Logan, G. D. (1997). Automaticity and reading: Perspectives from the instance theory of automatization. Reading and Writing Quarterly, 13(2), 123–146. Lu, X. (2011). A corpus-based evaluation of syntactic complexity measures as
QUARTERLY, 45(1), 36–62.
Nakamori, T. (2002). Teaching relative clauses: how to handle a bitter lemon for Japanese learners and English teachers. ELT Journal, 56(1), 29–40.
Nassaji, H., & Geva, E. (1999). The contribution of phonological and orthographic processing skills to adult ESL reading: Evidence from native speakers of Farsi. Applied Psycholinguistics, 20(2), 241–267.
Nation, P. (2001). Learning vocabulary in another Language. Cambridge, UK: Cambridge University Press.
Nation, P., & Malarcher, C. (2007). Reading for Speed and Fluency 2. Seoul, Korea: Compass Publishing.
Noro, T. (2000). Kai reberu no shori gino to gaikoku-go to shite no eigo no dokkai-ryoku [Lower-level processing skills and English comprehension as EFL].
Journal of the Chubu English Language Education Society, 29, 23–30. Nuttall, C. (2006). Teaching reading skill in a foreign language. Oxford, UK:
Macmillan Education
Perfetti, C. A. (1988). Verbal efficiency in reading ability. In M. Daneman, G. E. MacKinnon, & T.G. Waller (Eds.), Reading research: Advances in theory and practice (pp. 109–143). New York, NY: Academic Press.
Samuels, S. J. (1994). Toward a theory of automatic information processing in reading, revisited. In R. Ruddell, M. Ruddell, & H. Singer (Eds.), Theoretical models and processes of reading (4th ed., pp. 816–837). Newark, DE:
International Reading Association.
Samuels, S. J. (2006). Reading fluency: Its past, present and future. In T. Rasinski, C. Blachowicz, & K. Lems (Eds.), Fluency instruction: Research-based best practices (pp.7–20). New York: Guilford Press.
Segalowitz, N. (2000). Automaticity and attentional skill in fluent performance. In H. Riggenbach (Ed.), Perspectives on fluency (pp. 200–219) Ann Arber, MI: University of Michigan Press.
Segalowitz, N. (2010). Cognitive Bases of Second Language Fluency. Oxford, UK: Routledge.
Shimono, T., R. (2018). L2 reading fluency progression using timed reading and repeated oral reading. Reading in a Foreign Language, 30(1), 152–179. Stanovich, K. (1980). Toward an interactive-compensatory model of individual
differences in the development of reading fluency. Reading Research Quarterly, 16(1), 32–71.
Stanovich, K. (2000). Progress in understanding reading: Scientific foundations and new frontiers. New York, NY: Guilford Press.
Taguchi, E. (1997). The effects of repeated readings on the development of lower identification skills of FL readers. Reading in a Foreign Language, 11(1), 97– 119.
Taguchi, E., & Gorsuch, G. (2002). Transfer effects of repeated EFL reading on reading new passages: A preliminary investigation. Reading in a Foreign Language, 14(1), 43–65.
Taguchi, E., Takayasu-Maass, M., & Gorsuch, G. (2004). Developing reading fluency in EFL: How assisted repeated reading and extensive reading affect fluency development. Reading in a Foreign Language, 16(2), 70–96.
Walczyk, J. (1993). Are general resource notions still viable in reading research?
Journal of Educational Psychology, 85(1), 127–135.
Walczyk, J. (1995). Testing a compensatory-encoding model. Reading Research Quarterly, 30(3), 396–408.
Walczyk, J. (2000). The interplay between automatic and control processes in reading. Reading Research Quarterly, 35(4), 554–566.
Walczyk, J., Marsiglia, C., Bryan, K., & Naquin, P. (2001). Overcoming inefficient reading skills. Journal of Educational psychology, 93(4), 750–757.
Walczyk, J., & Taylor, R. (1996). How do the efficiencies of reading subcomponents relate to looking back in text? Journal of Educational psychology, 88(3), 537– 545
