Overview of the Study

CHAPTER FIVE DISCUSSION

The results presented in Chapter Four will be discussed in detail in this chapter.

An overview of the study will be presented first, followed by discussions on online reading strategy patterns, the effects of strategy use on reading comprehension, and feedback on English Reading Online.

Overview of the Study

The purpose of this study is to explore the reading strategies used by EFL learners in Taiwan when reading on the Internet. It focuses on identifying and explaining the reading strategies used by higher proficient and lower proficient students based on their navigating behaviors captured within English Reading Online.

It is hoped that the data and subsequent interpretation will significantly contribute to the recent discussion of research about online reading. It is also expected to provide suggestions for teachers who aim at helping their students become skillful readers on the Internet.

To collect comprehensive data about learners’ online reading behaviors, multiple tools were adopted. Two groups of students whose proficiency was differentiated via a TOEFL sample test were asked to read four articles with four different topics and with two levels of difficulty within a web-based reading program, English Reading Online.

This online reading program, whose design was based on L2 reading strategy theories, provided 15 strategy tools to facilitate students’ reading. During the course of their reading, their use of the strategy buttons and navigation history were tracked by the computer system. After reading each article, students were asked to (1) write written recalls, (2) complete post-task surveys, and (3) attend semi-structured interviews.

The Chi-square analysis was used to see if students’ online reading strategy

patterns were influenced by language proficiency, text topics, and text difficulty levels.

In addition, written recalls were scored by a combination of pausal unit and global scoring methods. The relationship between students’ strategy use and comprehension was computed using multiple regression to analyze which strategy can better predict recall scores. Furthermore, the results from the post-task survey were analyzed using descriptive statistics to show students’ feedback on English Reading Online. Finally, qualitative investigation of four case studies gathered from videotaping and semi-structured interviews was used to backup the quantitative strategy use data mentioned above. The following section will discuss online reading strategy patterns used by EFL learners

Online Reading Strategy Patterns

The online reading strategy model shown in Figure 24 has been constructed

based on the results of strategy button usage and provides the central framework for

organizing the findings of this study. In Figure 24, the overall strategy types are

depicted in the inner circle, whereas factors influencing strategy use are shown in the

outer circle. First of all, four types of reading strategies (Global, Problem-solving,

Support, and Socio-affective strategies) were the center of investigation on EFL

learners’ online reading patterns. The results of overall strategy patterns used by 30

participants showed that language proficiency influenced strategy use. In addition,

text topic and difficulty level affected strategy use only under the interaction with

language proficiency. Second, further qualitative investigation on four case studies

regarding their online reading strategy use showed that better computer skills

determined their fluent navigation of online reading and facilitated their acquisition of

background knowledge, which in turn contributed to effective online reading strategy

use.

Figure 24. A Model of Online Reading Strategies

The following discussion will first center on online strategy types, followed by

the four factors that influenced strategy use. These factors included language

proficiency and text topic and difficulty level as well as computer skills and

background knowledge.

Online Reading Strategy Types

In light of strategy types depicted in the inner circle of the model (see Figure 24), four pre-defined strategies, which included Global strategies, Problem-solving strategies, Support strategies, and Socio-affective strategies, will be discussed first.

Other strategies yielded by qualitative investigation, including Navigating strategies, Information gathering strategies, Interface changing strategies, and Usability reporting strategies, are the next focus.

Global, Problem-Solving, Support, and Socio-Affective Strategies

An analysis of the pre-defined reading strategy use revealed the following patterns: (1) Support strategies were the single category that made up the overwhelming proportion of the strategy use, (2) Individual strategies under Support strategies (e.g. Translation, Dictionary, and Highlight) were also the most frequently used strategies, (3) Problem-solving strategies were scarcely used, and (4) Support and Problem-solving strategy use differed not only in terms of frequencies, but also in time sequences.

In this study, students used Support strategies much more than any other available strategy. It seems that strategies that serve as basic support mechanisms to aid comprehension such as Dictionary and that require little conceptual processing such as Translation emerge as dominant factors in these students’ strategy choice. As strategies defined in this study are skills and techniques employed by the readers to fulfill reading tasks, it is not surprising to see that strategies that help readers efficiently achieve reading tasks are most widely used.

The finding that Support strategies dominated students’ strategy use corroborates

Shen’s (2003) study that both high and low proficient learners use predominantly

local strategies, especially translation (word-by-word). This study’s findings are also

in agreement with Huang’s study (1999) that local strategies, especially translating,

prevailed in both more or less proficient students’ choices. It seems that learners mainly depend on support mechanisms, such as translating, using dictionaries, or highlighting, to aid comprehension. The reason behind this dominant use of Support strategies leading all students’ strategy use may be due to the convenience and instant feedback that students can have by using them (Mokhtari & Reichard, 2004). For example, Translation, the most frequently used strategy, can easily turn a three-page long article into L1 equivalents within only a few seconds. It appears that support strategy mechanisms shorten the time for decoding the text and are especially helpful to students who are intimidated by the lengthy online articles. In an online reading environment of vast information resources, strategies that can help online readers quickly find, evaluate, use, and communicate information are likely to be widely accepted and adopted (Leu, et al., 2004).

Another possible explanation of the dominant usage of Support strategies is that in a typical Chinese reading class where vocabulary acquisition or reading for details is emphasized (Yeh & Wang, 2004), it is likely to find students constantly employ strategies related to vocabulary learning (Dictionary strategy button) and explanation of each sentence and word (Translation strategy button). The cultural factor of Chinese reading classes may explain the heavy use of Support strategies found in this study.

Despite that the dominant usage of Support strategies is confirmed in the

aforementioned L2 studies, findings of this study contradict a series of studies

completed by Sheorey and Mokhtari (2001) and Mokhtari and Sheorey (2002) on the

identification of metacognitive reading strategies of L2 learners. For example,

Sheorey and Mokhtari’s (2001) study has shown that both native speakers and ESL

learners attribute the same rank order of importance to Problem-solving strategies,

Global strategies, and Support strategies in reading academic texts. One possible

explanation is that reported strategy use found in Sheorey and Mokhtari’s (2001) study does not necessarily translate into actual usage of the strategies concerned in this study. Using questionnaires to collect students’ strategy use data in Sheorey and Mokhtari’s (2001) study may be different from utilizing a web-based reading program (English Reading Online) with built-in strategy functions to investigate the actual online strategy patterns in this study. Another possible explanation is that different instructional settings, such as an ESL environment in Sheorey and Mokhtari’s (2001) versus an EFL environment in this study, may contribute to the differences of strategy use.

Looking closely into each individual strategy button, it was found that the three most frequently used strategies all belonged to the category of Support strategies. For example, Translation, one of the Support strategy mechanisms, was mostly favored and most frequently used. In light of its heavy use, it is worthwhile to question if the convenience of translators can really help students arrive at the meaning. Also, special attention should be addressed to the abuses of meaning technologies, defined as tools that are based on technology and make the meaning of language items accessible on demand (Hubbard, 2004). The pedagogical implication will be discussed in detail later.

In addition to Translation, Dictionary was also a frequently used Support reading

strategy among the 15 strategy buttons. This finding is consistent with L2 reading

research that vocabulary is perceived to be the most difficult part among EFL learners

and thus looking up new words in the dictionaries is the most frequently employed

strategies when encountering difficulties (Cheng, 1998; Chi & Chern, 1989). This is

especially true for students with limited proficiency in a technical/vocational school

setting in Taiwan (Lin, 1996; Yang et al., 1994). Additionally, in a web-based reading

environment, L2 readers often seek help from annotations to compensate for their

challenged language proficiency (Chun, 2001; Ercetin, 2003). The possible explanation is that students feel frustrated about reading something incomprehensible;

therefore, the natural inclination for students is to stop reading and resort to dictionaries for help. However, one thing worthy of our attention is that it is equally frustrating to consult dictionaries for every unknown word because the act of looking up words would interrupt reading (Hubbard, 2000, 2004). The pedagogical implication is that students need to learn to strike a balance between reading for the gist and reading for improving vocabulary during the process of reading.

Highlight, a device of taking notes on unknown words, was another commonly used strategy. One interpretation for the frequent use of Highlight strategy is that users tended to lose track of what they read and they needed to conquer the fear of losing the information (Mokhtari & Sheorey, 2002; O’Malley & Chamot, 1990; Oxford, 1990). Taking down notes is a safety measure that ensures retention and stability of information. In addition, the heavy use of Highlight may be due to the traditionally teacher-centered reading classes where students have cultivated the habits of taking down notes of what is said (Yeh & Wang, 2004). Students in this study seem to transfer the act of taking notes of teachers’ lectures in class into the scenario of reading on their own.

In contrast with the dominant use of Support strategies (62.32%), the least used

strategy was Problem-solving strategies (7.18%), which included the bottom three

strategies—Semantic Mapping, Pronunciation, and Speed Reading. Participants in

this study failed to apply focused techniques, such as adopting audio aids, visualizing

information that had been read, and adjusting reading rates to solve reading

comprehension problems. The rare use Problem-solving strategies might be because

the efforts involved in directly dealing with the text such as drawing semantic maps,

reading aloud, and testing reading speed may be more demanding than those related to

support mechanisms such as consulting dictionaries and translating into L1. In an online reading environment where immediacy and speed are online users’ primary concerns (Chapelle, 2005; Hubbard, 2004; Leu et al., 2004), it may appear reasonable that Problem-solving strategies were least adopted in this study.

This finding regarding the scarce use of Problem-solving strategies is inconsistent with Anderson’s (2003) study which finds that among the 38 online reading strategies stated in the questionnaire, the top 12 strategies are mostly Problem-solving strategies. It also contradicts Sheorey and Mokhtari’s study (2001) on self-reported reading strategies where Problem-solving strategies are mostly favored by ESL and native speaking students. One possible explanation is that different data collection tools may bring about different results of strategy use.

Self-reported strategy data in previous studies such as Sheorey and Mokhtari’s and Anderson’s (2003) may differ from students’ actual online behaviors found in this study. Besides, the under-use of Problem-solving strategies observed in the current study may suggest that students would rather rely on single criteria for textual understanding, i.e., understanding of individual words by using Translation and Dictionary, than work directly with the text, i.e., re-examining the text via Semantic Mapping and re-reading for better comprehension via Pronunciation.

The differences between Support strategies and Problem-solving strategies can

be found not only in their frequencies of strategy use, but also in the sequences of

their strategy use. Support strategies, which were used most frequently, were also

employed prior to the least-used Problem-solving strategies. According to strategy

sequence analyses and the questionnaire results on strategy preference, Support

strategies were the most frequently used, were most favored, and were first used

strategies. Frequency of strategy use is positively related to sequence. This finding on

time sequence sheds additional light on online reading strategy use. Frequency and

time sequence of strategy use seem to go hand in hand in determining readers’ choice of reading strategies. Strategy use is not only a matter of “what;” “when” to use strategies is equally important (Anderson, 1991).

In summary, Support strategies, with features of instant feedback and quick reference, dominate strategy use. The most frequently used individual strategy buttons, like Translation, Dictionary, and Highlight, also belong to Support strategies.

Problem-solving strategies that require direct attention to work with textual information are scarcely used. The stunning differences of Support and Problem-solving strategy use found in both frequencies of strategy use and strategy use sequences may suggest that readers’ choices of strategy use are affected by their needs for immediate feedback in an online reading environment.

This section has discussed the use of Support strategies and Problem-solving strategies. The next section will discuss other strategies emerging from case studies.

The effects of Global and Socio-affective strategy use among different proficiency groups will be presented later.

Other Strategies: Navigating, Information Gathering, Interface Changing, and Usability Problem Reporting Strategies

Aside from the four pre-defined strategy types (Global, Problem-solving, Support, and Socio-affective), qualitative investigation into these four chosen participants yielded four additional strategy types, labeled as “Other strategies,”

including Navigating strategies, Information gathering strategies, Interface changing strategies, and Usability problem reporting strategies as shown in Figure 24. These strategies were unique assets in assisting readers in approaching online texts.

For example, Navigating strategies helped readers navigate online texts smoothly

by moving between pages patiently, connecting pictures with the text, adjusting

scrolling range, switching between windows, and moving the cursor along the words.

These strategies helped readers successfully access online texts where chunks of information were linked multi-sequentially (Konishi, 2003; McNabb, 2006; Rouet et al., 1996).

Information gathering strategies allowed readers to critically evaluate information on a Web page, monitor their pathways, and reduce the chance of becoming disoriented amid the irrelevant information (Coiro, 2005; Shetzer &

Warschauer, 2000). With the tactics of using special keywords to search websites, readers had a better chance finding information they wanted and engaging themselves in the Internet text with deeper understanding. One interesting phenomenon was that these types of strategies were mostly adopted by more experienced computer users because prior knowledge of hypertext structures brought by better computer skills helped them know the way around while reading online texts (Aleksander, 2001).

Interface changing strategies, on the other hand, assisted readers in building up a pleasant reading environment for their own preference. By loading the software they were familiar with and using the browser they were accustomed to, readers made their online reading experience more comfortable and in turn increased their motivation for reading (Guthrie & Wigfield, 2000; O’Mally & Chamot, 1990; Oxford, 1990).

Usability problem reporting strategies provided readers a gateway to let out their complaints about ill-designed websites and to express their needs for a more user-friendly web design. The occurrence of Usability problem reporting strategies also confirmed previous usability research that usability problems, such as non-adjustable font sizes, flashy pop-up advertisements, disorganized links, and overly literal search engines could affect reading performance (Nielsen, 1994, 1997;

Nielsen & Morkes, 1997). These strategies have reminded web designers to improve

ease-of-use during the design process (Elshair, 2002). Aside from the unique

contribution to web navigation, the use of the above four additional strategies were

influenced more by students’ computer skills and background knowledge than by proficiency levels.

The above section discussed the major strategy patterns. The following sections will discuss factors influencing strategy use, including language proficiency, the text topic and difficulty level, computer skill, and background knowledge.

Language Proficiency

Elements influencing online reading strategy use are depicted in the outer layers of the circle in Figure 24. The first factor, language proficiency, accounted for differences in strategy use. The differences were: (1) Strategy types varied with language proficiency, with the High proficiency group tending to use Global strategies, whereas the Low proficiency group resorting to Socio-affective strategies. (2) Sequences of strategy use differed with language proficiency, with the High group using strategies more effectively than the Low group. (3) Purposes for using Support strategies differed with language proficiency, with the High proficiency group using strategies more appropriately.

First of all, data concerning distribution of strategy use for the High and the Low

proficiency groups in this study show that language proficiency influenced strategy

use. Overall, the High proficiency group used more Global strategies than the Low

proficiency group, whereas the Low proficiency group used more Socio-affective

strategies than the High proficiency group. When further investigation was carried out

to determine if students with different proficiency levels used different strategies in

reading easy and difficult texts, similar tendencies were found—the High proficiency

group tended to use more Global strategies but less Socio-affective strategies than the

Low proficiency group. The High proficiency group disliked to use Socio-affective

strategies, especially when they were reading difficult texts. Additionally, the High

proficiency group used Keyword, Prediction, Preview, and Outline more frequently

than the Low proficiency group. It appears that these Global strategies give the High proficiency group a general understanding of the text prior to actual reading and therefore a careful preparation of what to read next. In contrast, the Low proficiency group tended to use Socio-affective mechanism like Music Box and Chatroom more often than the High proficiency group. It appears that less proficient readers like to use music to lower their anxiety and used Chatroom to engage themselves in communication with peers.

Similar to the overall findings on the total participants, in general, the four case studies also show that James and Chris, both from the High proficiency group used comparatively more Global strategies than Amber and Tim, both from the Low proficiency group. On the other hand, the use of Socio-affective strategies like Chatroom and Music Box was more important to the Low proficiency group than to the High proficiency group. For example, the use of Chatroom was found among the Low group students, such as Amber and Tim. However, the act of logging onto chatrooms was virtually absent for James, a student from the High group. James, with his own confidence in language, read text directly without discussing with peers.

Amber and Tim, on the other hand, heavily relied on chatroom in the process of

reading. To Amber, her major confidence and interest in continuing to read the text

was from her discussion with classmates. Likewise, the strategy of listening to music

was used by Amber and Tim, but was not utilized by James. It seems that music

released the tension in these two low EFL proficiency learners. In conclusion, the

chatroom’s capabilities in bringing global peers together in brainstorming ideas and in

creating a socially supportive environment (Beauvois, 1997; Chun, 1994; Warschauer,

1996) and healing power of music to relieve tension (Oxford, 1990) seem to be most

appreciated by low achievers. High achievers, on the other hand, would rather adopt

careful mental planning and monitoring of their reading than interact with others.

The findings of this study showed students with different proficiency levels used different strategies. This confirms previous research on comprehension strategies which shows that readers’ proficiency influences strategy use (Block, 1986, 1992), with the High proficiency group using more global/top-down strategies than the Low proficiency group (Huang, 1999; Shen, 2003; Mokhtari & Reichard, 2002). The employment of reading strategies can differentiate good readers from poor ones. For example, this study found more Global strategy use among higher proficient students.

This finding is similar to Huang’s (1999) study which shows that more proficient readers use more top-down strategies, whereas less proficient readers use more bottom-up strategies. This also corroborates Mokhtari and Reichard’s (2002) research which states that readers who rate themselves as good readers have a significantly higher use of global reading strategies than those who rate themselves as poor readers.

In addition, this study also found that the High proficiency group displayed more Global strategy use in approaching difficult texts. This is consistent with Shen’s (2003) study where in reading more challenging texts, the high proficiency students use more global strategies, while low proficiency students use more local strategies. In short, the use of Global strategies among more proficient readers is confirmed both in this study and the literature.

However, subtle differences exist between global-local distinctions about the High and the Low proficiency groups in previous studies and global-socio/affective variations in this study. Unlike previous studies that reveal the Low group’s predominant use of local strategies, this study has shown that the High group and the Low group appeared to use the same amount of so-called local strategies

, and the

1

The differences found in the literature about high-proficient and low-proficient readers were mostly global and local distinctions (Block, 1986, 1992; Chang, 1998; Chen, 1998; Huang, 1999; Shen, 2003).

Local strategies described in the literature may be termed as Problem-solving and Support strategies in

this study.

Low proficiency group differed from the High proficiency group in their heavy use of Socio-affective strategies. One possible explanation is that strategy categories are not fully mapped across studies; moreover, previous studies do not take socio-affective factors into consideration.

Another important finding regarding proficiency and strategy use is that the differences between the High and Low proficiency groups resided not only in strategy types, but also in sequences of strategy use, with the High proficiency group using strategies more effectively. For example, Global strategies were used prior to Support strategies by the High proficiency group but were used after Support strategies by the Low proficiency group. It seems that Global strategies, which are carefully planned techniques by which students manage their reading such as having a purpose in mind and previewing the text, should be adopted earlier to arouse awareness in reading (Barnett, 1989; Grabe, 1991) and are better practiced by the High proficiency group.

The pedagogical implications will be further discussed.

Case studies also revealed similar tendencies that the High proficiency group

used strategies more effectively. For example, unlike students from the High

proficiency group (James and Chris) who tried Global strategies (Preview and

Keyword) in the very beginning of reading to help them get an overview of the study,

students from the Low proficiency group (Amber and Tim) used Global strategies

almost towards the end of the task. Amber and Tim regarded looking up new words as

the first step towards understanding the text. Different from the High proficiency

group who used Global strategies effectively, it appears that the Low proficiency

group fail to read globally first to gain the major idea of the text. This discovery about

different time sequences used by the High and Low proficiency groups is interesting

because previous studies mostly focused on the frequencies and scarcely touched

upon the sequence of strategy use. As Anderson (1991) points out, good readers not

only differentiate from poor readers not only in terms of the amount of strategies and the quality of strategies, but also in their abilities to choose strategies at the right time under the right situations.

The last finding regarding strategy use and language proficiency is that purposes and ways for using Support strategies differed between the High proficiency group and the Low proficiency group, with the High proficiency group using strategies more appropriately. For example, James, a High proficiency group student, used the translation software purely for reference and to double-check his understanding, and he used it only when the text became difficult. Even when he used it, he did not entirely rely upon it. Instead, he read by himself and used Translator only to make parallel comparison of his understanding of the text and the translated text. However, for students in the Low proficiency group, such as Amber and Tim, the act of directly copying-and-pasting results from the translator without carefully thinking about the appropriateness of the translated content prevented them from understanding the text.

It was found that Amber and Tim kept using the translator even when they thought the text was not difficult. The reason behind Amber’s and Tim’s actions may be due to their need for quick understanding of the text. Translators provided such convenience.

It seems that the High proficiency group uses Translation not just for meaning but for learning, whereas the Low proficiency group uses Translation at the expense of building automaticity in going directly from form to meaning in L2 (Hubbard, 2000).

In addition, the High and Low proficiency groups also differed in the way they looked up dictionaries. For example, more proficient students like James and Chris read the text first time by themselves without looking up words in the dictionaries.

They tried to process as much text as possible the text. However, Low group students

like Amber and Tim tended to use dictionaries once they encountered reading

problems. They would rather look up every unfamiliar word than infer meaning from

the context. It seems that higher level students pay more attention to the actual reading activity and explore the text itself (Chun, 2001). Even in using multimedia aids, such as annotation, more proficient readers appear to use these aids more appropriately than less proficient ones (Dillon and Gabbard, 1998). Lower level students who have the habit of consulting dictionaries to cope with reading difficulties may find hypertext glossaries an even stronger temptation (Hubbard, 2000).

In summary, although the High and the Low proficiency groups were using similar amounts and types of Support strategies, how they used strategies was different and high proficiency level students seemed to be applying strategies more effectively and appropriately. This finding consolidates previous studies that high proficiency readers differ from low ones not on the number of strategies used, but how they use strategies (Anderson, 1991). The finding suggests that strategic reading is not only a matter of knowing which strategies to use, but in addition, the reader must know how to apply strategies successfully. Just as Anderson (1991) points out, strategy use is not just a matter of “what,” but “how” is equally important. This may be one factor contributing to the relationship between proficiency level and strategy use.

In conclusion, language proficiency contributes to differences in reading strategy

use in the aspects of strategy types, sequences of strategy use, and purposes for using

strategies. Overall, high-proficient L2 readers use Global strategies to a greater extent,

whereas low-proficient readers rely more heavily on Socio-affective strategies. High

proficient readers also adopt sequences of strategies more appropriately and their

purposes for using strategies make their use of strategies more effective than low

proficient readers.

Text Topic and Difficulty Level

In addition to language proficiency discussed in the previous section, the second factor affecting strategy use was text topic and difficulty level (see Figure 24). In light of text topic, as this study has shown, participants used the same strategy types regardless of the topic. Nonetheless, the study also revealed that Article 2 (on the Pyramids of Giza) prompted the use of a significantly higher number of total strategies as compared with the total number of strategies used for reading the other three articles. One possible explanation is that students may not have enough background knowledge about the process of building pyramids so they needed more strategies to process the text. The present investigation is consistent with earlier studies which suggest that reading passages with unfamiliar topics result in comprehension challenges and thus require more cognitive skills to aid comprehension than those with familiar topics (Brantmeier, 2003, 2005).

Another explanation of the larger number of strategy use on Article 2 may be due to its disorganized text structure and lack of coherence devices. For example, the title of Article 2 was “The Pyramids of Giza,” which aimed at describing the small pyramids around the Giza area. However, in the middle of the text, the process of preservation of mummies was introduced. At the end, the pyramids were solely referred to “the Pyramids of Kufu,” which was just one of the pyramids in Giza. The jumping themes easily carried readers away and posed challenges for comprehension.

More strategies were therefore needed to process the text. These results suggest that

when the text becomes difficult due to the use of unfamiliar words, long complicated

sentences, ambiguous causal relationship, and disorganized content, readers need to

use more strategies to judge their comprehension accurately (Kletzien, 1991). In

conclusion, the important finding is that students use more strategies when their

familiarity with the topic decreases. The less familiar students are with the topic, the

more strategies they use.

With regard to difficulty levels of the text, students in both the High and the Low proficiency groups tended to use a fixed set of reading strategies that they had long been accustomed to regardless of text difficulty. Their use of strategy did not increase when text difficulty increased, nor did their strategy patterns vary along with the text difficulty. This finding contradicts previous research that appears to show that readers use more strategies with more difficult materials (Smith, 1991). It could be interpreted that participants in this study are not aware of various strategies tools at hand, nor are they conscious of the goals of the reading tasks. They simply follow their habitual modes of reading, without making any adjustments for dealing with more challenging texts. The pedagogical implication is that students need to learn how to orchestrate strategy use to deal with different reading tasks.

Despite the fact that students, in general, exercised similar patterns of strategy in

reading difficult and easy texts, differences of strategy use were found when language

proficiency and the text difficulty level were taken into consideration together. In

other words, the text difficulty affected strategy use only when it interacted with

language proficiency. This interaction is shown by an arrow pointing from “Language

Proficiency” to “Text Topic and Difficulty Level” in Figure 24. When reading easy

texts, the Low proficiency group used comparatively more Socio-affective strategies

than the High proficiency group. When reading difficult texts, the High proficiency

group tended to use more Global strategies and less Socio-affective strategies than the

Low proficiency group, whereas the Low proficiency group still adhered to

Socio-affective strategies. To tackle more difficult texts, the High proficiency group

students were more likely than the Low proficiency group students to try to first grab

the main idea through a top-down approach, using such strategy buttons as Preview

and Keyword. It seems that learners with different proficiency levels apply different

strategies in dealing with texts of different difficulty levels. Good readers are more flexible to adjust their strategies in dealing with reading difficulties (Kletzien, 1991) and high-achievers apply more self-monitoring strategies to process the text (Huang, 1999).

Another interesting finding regarding the text difficulty levels was that students generally used Support strategies followed by Socio-affective strategies in reading easy texts. When texts became difficult, students tended to use Socio-affective strategies prior to Support strategies. One possible explanation is that Socio-affective strategies like the Music strategy button play an important role in relieving tensions, especially when readers are reading difficult texts (O’Malley and Chamot, 1985;

Oxford, 1990). Also, by asking questions or cooperating with others in the Chatroom, readers increase social interactions with others, get closer to the meaning, and thus improve their understanding of the text. This may explain the priority use of Socio-affective strategies among most of the students in dealing with difficult texts.

In conclusion, text topic did not result in different uses of strategy types.

However, the total number of strategies that students used in reading Article 2 outnumbered the other three articles due to topic unfamiliarity. Besides, text difficulty influenced strategy use only when language proficiency came into interaction. For example, High proficiency group used more Global strategies in reading difficult texts, whereas Low proficiency group still resorted to Socio-affective strategies. Finally, time sequence analyses show that Socio-affective strategies were adopted prior to the most frequently used Support strategies in reading difficult texts. This result suggests that the priority use of Music and Chatroom to relax and to create social interactions is important to help readers approach difficult online texts successfully.

Computer Skill

From qualitative investigation on four case studies, it was found that in addition

to language proficiency and text topic and difficulty level, a third element—computer skill— played a role in readers’ online strategy use (see Figure 24). Among the four types of additional strategies (Navigating strategies, Information gathering strategies, Interface changing strategies, and Usability problem reporting strategies), it was found that students with better computer skills used comparatively more strategies than those who had exhibited average computers skills. In addition, students with better computer skills also used strategies more effectively to assist in their comprehension, which could be reflected in their use of Interface Changing strategies, Information gathering strategies, multimedia features of online texts, and prior knowledge on Internet reading.

With regard to Interface changing strategies, the common feature shared between students with good computer skills like Chris and Tim was that they were fastidious about the user interface. Both of Chris and Tim liked to change their browser to KKMAN, play their favorite music, and change the colors of the font. Although these strategies did not directly contribute to reading comprehension, students who were advanced and frequent computer users were strongly affected by these features when they read online (Elshair, 2002). These personalized strategies, which tailor-made their reading environment according to their preference, have made online reading more comfortable. When the online reading environment becomes pleasant, learning effects will be increased (Bulter-Wiburg, 2003). The pedagogical implication of setting up a personalized online reading environment will be discussed in detail later.

In light of Information gathering strategies, students with better computer skills

had good chances to find information to build up their background knowledge. For

example, Chris and Tim used precise keywords to search for information, decided on

the spot whether the information suited their need, and quickly switched back and

forth between web pages to expand their knowledge about the given topics (a movie

about Finding Nemo and a national holiday of St. Patrick’s Day). This shows that online reading requires a different set of strategies from printed reading (Coiro, 2005;

Shetzer & Warschauer, 2000) and computer skills are a factor determining successful online navigation. In reading online texts where information is presented non-linearly, computer skills and web navigation tactics enable readers like Tim and Chris to scaffold their learning and increase their language learning repertoires by using search engines effectively, finding the desired information by scanning and skimming, and, finally, making critical decisions as to whether to stay on a given Webpage or to go back to the search engine to start another new search. Moreover, the act of searching related information was especially beneficial for Tim, a student with limited language proficiency. He expanded his knowledge about a given topic, St. Patrick’s Day, by searching additional information and thus built up his reading confidence. It appears that Tim’s greater background knowledge gained from searching additional information has compensated for his lack of syntactic control over the language.

Information gathering strategies facilitated online reading not only because these strategies helped readers search related information by using precise keywords and critically evaluating information, but also because these strategies increased readers’

social interaction with global peers to have their problems solved. For example, while Chris had trouble interpreting the meaning of a sentence, he posted his questions and received a satisfying response from online users that could help him arrive at the meaning. In an era when the Internet and other ICTs are quickly becoming the central technology for a global community, those who are well-prepared for synchronous and asynchronous communication have a better chance to empower themselves to find solutions to their problems (Leu et al., 2004).

Aside from the benefits of using Information searching strategies to search

related information and to interact with peers, Chris and Tim, with adept computer

skills and years of experiences in online reading, also took advantage of multimedia features such as animation and pictures to boost their comprehension and increase their motivation for reading. For example, Chris’s knowledge about the religious root of St. Patrick’s Day stemmed from the picture of a snake symbolizing the pagan tradition. Tim’s knowledge about Finding Nemo was refreshed by a short movie clip in the Pixar Website. It seems that sounds and graphics that further explain or elaborate on the meaning of the text support their reading development (Mayer, 1997).

The visual and verbal presentation of the materials brought about by multimedia goes hand in hand to produce meaningful learning.

In addition to the use of precise keywords in finding relevant information needed, prior knowledge on Internet reading determined the success of Chris’ and Tim’s internet search. Their knowledge about how to define the keyword as well as their knowledge about navigating the Internet contributed to their accurate path for reading.

Chris knew very well from his previous experiences on BBS talk where to post his questions to get satisfied answers. Tim knew how to select among various links generated by the Google search based on his prior knowledge of a good tourist website that he had visited before. Their prior knowledge on how to use search engines wisely, how to use a precise keyword, and how to select the links all contributed to their successful search. These results suggest that computer skills are a factor in determining the successful online reading performance and the effects of good computer skills on building up background knowledge will be further discussed in the next section.

In conclusion, computer skills influence strategy use because they can make the

reading environment more personalized and therefore increased motivation for

reading. Computer skills also facilitate the building of background knowledge by

searching related information, the resolution of reading problems by interacting with

global peers, and the improvement of reading comprehension by utilizing multimedia features. An analysis of Tim’s adeptness in manipulating information in searching outwards to find relevant information and Chris’ endeavor to seek help from his global peers suggests that the interactivity of the digital text—typical in such activities as hyper linking, quick referencing, and synchronous and asynchronous communication— all empowers readers to find solutions to their problems (Coiro, 2005; Sutherland-Smith, 2002; McNabb, 2006). These two case studies demonstrate that in Internet reading, students with good computer skills have better chances to make good use of communication tools to enrich their understanding of the texts.

Background Knowledge

Related to the factor of computer skills is the role of background knowledge in online reading. In digital reading where information is embedded in interactive multimedia and hypertext, background knowledge is not only built by relating one’s own existing knowledge, but is also constructed by searching outwards to find relevant information. Computer skills in this study then serve as a bridge to assist learners in finding relevant information to scaffold learning in the maze of the hypertext. The effect of computer skills on building up background knowledge in reading online texts is shown by an arrow pointing from “Computer Skill” outwards to “Background Knowledge” in Figure 24.

There are two determinant functions of background knowledge in this study: (1)

to relate new information with existing knowledge via readers’ old experiences and (2)

to scaffold learning by searching additional information. First, case study

investigations showed that because Chris, Amber, and Tim had heard about the movie

Finding Nemo before, their prior knowledge about the movie familiarized them with

the text structure and the content. Reading comprehension has been defined by the

construction of meaning from the text (Leu et al., 2004). Unfortunately, texts, like

Finding Nemo, did not provide necessary information for meaning construction.

Therefore, conceptual and relational gaps were likely to occur (Koda, 2005). To close the gaps, prior knowledge has become necessary when Chris, Amber, and Tim made inferences to connect the known with the unknown (Anderson, 1994). Their activation of background knowledge and the integration of new text information with existing knowledge have reaped obvious benefits in comprehension (Barnett, 1989; Carrell, 1987; Johnson, 1982; Koda, 2005).

The second function of the background knowledge was evidenced by the fact

that Tim, a student with good computer skills, tried to locate necessary information to

expand his background knowledge when his existing knowledge was not enough for

meaning construction. Computer skills in this case then served as a bridge to connect

readers to a digital reading environment with the purpose of broadening their

background knowledge. Reading comprehension on the Internet takes on a broader

definition than reading printed texts (McNabb, 2006). New skills and strategies are

required to successfully comprehend information (Leu et al., 2004). For example,

students like Tim needed to have knowledge about how to search for information

properly, how to comprehend search results, and how to make correct inferences and

synthesize information presented in multiple media formats. With the aforementioned

newly arising strategies to deal with digital texts, Tim could find necessary

information to build up his background information and further improve

comprehension. In addition, students like Chris needed to have knowledge about how

to utilize information technology such as chatroom and BBS to increase social

interaction with peers. With these skills, students could discuss and communicate their

ideas with peers and thus have their questions answered. It is true that background

information gained by searching additional information or seeking help from peers

has become useful in understanding online texts.

In conclusion, this section has discussed general online reading strategy patterns and factors that influence strategy use. The next section will discuss the second focus of this study—the connection between strategy use and reading comprehension.

The Effects of Strategy Use on Reading Comprehension

A cluster of exploratory studies has investigated the effect of strategy use on comprehension performance. Some studies suggest that the bulk of reported strategies generally contribute to reading achievement (Block, 1986, 1992; Brantmeier, 2000;

Chang, 2005; Sarig, 1987; Singhal, 2003; Tseng, 1998), whereas other studies reveal that no systematic connections exist between a particular set of strategies and reading comprehension (Anderson, 1991; Koda, 2005). Even though some empirical studies have found a particular set of strategies contributes to effective reading, relatively fewer studies have investigated how strategy use results in different levels of comprehension. In an effort to disentangle the relationship between strategy use and comprehension, this study has successfully predicted which strategies are most effective for reading comprehension at different comprehension levels (main ideas and details) among two groups of students with different proficiency levels (High and Low proficiency groups).

The results regarding the effects of strategy use on different levels of

comprehension in terms of total scores, scores on main ideas, and scores on details are

summarized below. First, the use of Support strategies contributed to the increase of

the total score among all students, among the High proficiency group, and among the

Low proficiency group. Besides, the use of Socio-affective strategies predicted the

gains of the total score among the Low proficiency group and resulted in the decrease

of total score among the High proficiency group. In reading the most difficult text

(Article 4), the use of Global strategies by the Low proficiency group predicted a

significant increase on total comprehension scores. This means that the use of Global

strategies contributed to better recall in dealing with difficult texts. When texts were too difficult, students needed to read globally to have a general understanding of the text instead of solely resorting to supporting mechanisms, such as word-by-word translation.

Second, the use of Support strategies worked well in gaining scores on main ideas for students in general, both the High proficiency group and the Low proficiency group. For all students, the use of Global strategies predicted main idea gains in reading challenging texts (Articles 3 and 4). This situation was especially obvious when the Low proficiency group read the most difficult text (Article 4).

Third, detail scores were all influenced by the use of Support strategies. In addition to Support strategies, the use of Global strategies became an obvious predictor in difficult texts (Articles 3 and 4). When the Low proficiency group students read the most difficult text (Article 4), the use of Global strategies showed the tendency of improving detail scores, whereas the use of Support strategies showed no significant gains. Global strategies not only influence the improvement of main idea scores, but also contributed to the increase of scores on details.

The models depicting the relationship between the use of strategies (Global,

Support, and Socio-affective strategies) and three levels of reading comprehension

(total score, scores on main ideas, and scores on details) among different student

groups (all students, the High proficiency group, and the Low proficiency group) are

shown in Figure 25 and Figure 26. Fgiure 25 presents the effect of strategies on all

students’ three levels of comprehension, whereas Figure 26 shows the influence of

strategy use on comprehension among the High and the Low proficiency groups. The

following section will first center on the effect of Global strategies, followed by

Support strategies and Socio-affective strategies on reading comprehension.

Figure 25. All Students’ Online Reading Strategy Use and Levels of Comprehension

Note. The lines indicated the effect of strategies on different levels of comprehension among a particular student group.

Total=total score; Main=score on main idea; Detail=score on detail

(All)=Articles 1, 2, 3, and 4; (A1,3,4) = Articles 1, 3, and 4; (A3,4) = Articles 3 and 4

Figure 26. High and Low Groups’ Online Reading Strategy Use and Levels of Comprehension

Note. The lines indicated the effect of strategies on different levels of comprehension among a particular student group.

Total=total score; Main=score on main idea; Detail=score on detail

(All)=Articles 1, 2, 3, and 4; (A1,2,3) = Articles 1, 2, and 3; (A1,3,4) = Articles 1, 3, and 4;

(A2)=Article 2; (-A2)=negative effect on Article 2

The Effects of Global Strategies on Comprehension

This study showed that Global strategies made a unique contribution to comprehension even though the use of this strategy was not particularly frequent. In agreement with Block’s (1986) study that Global strategies are associated with better comprehension, this study has three further findings: (1) Global strategies contributed to all three levels of reading comprehension and predicted better comprehension for reading difficult texts, (2) Global strategies were a stronger predictor of better comprehension than Support strategies, (3) the effect of Global strategies on comprehension was particularly obvious among the Low group, and (4) Global strategies were mostly efficiently used.

First of all, Figure 25 shows that the use of Global strategies contributed to an

increase in total scores, scores on main ideas, and scores on details when students in

general were reading more challenging texts (Articles 3 and 4). It may be interpreted

that the general picture obtained by using Global strategies helps readers remove the

psychological barrier in reading texts beyond their own level. When students have a

basic understanding from utilizing Global strategies such as previewing the texts and

keywords, they have the confidence to read on and they could improve their total

score in reading more challenging texts (Block, 1986, 1992). Besides, Global

strategies are useful to increase comprehension scores on main ideas because

previewing the text as to its content and organization assist readers in detecting main

ideas (Koda, 2005). Moreover, it is interesting to find that Global strategies, which are

of a conceptual and top-down nature, including inferences, predictions, and

elaborations, also contribute to improved scores on local comprehension on details. It

appears that the all-powerful function that Global strategies have on different levels of

reading comprehension is unparalleled.

Second, when both Global strategies and Support strategies successfully predicted increased reading performance, Global strategies were a stronger predictor of comprehension scores than the most frequently used Support reading strategies in reading difficult texts. The possible explanation is that while the Support strategy is the most widely used strategy among all students, an exclusive dependence on Support strategies seems not to work when students are encountering the most difficult text. They need to use other strategies such as Global strategies to facilitate their reading.

Third, the effects of Global strategies were most obvious among the Low proficiency group. Figure 26 shows that Low group students’ total score, scores on main ideas, and scores on details in reading the most difficult text (Article 4) increased along with the use of Global strategies. It appears that as long as the Low group students are willing to adopt Global strategies, their total scores, scores on main ideas, and scores on details will have a significant increase. Combined with the literature finding that low proficiency students prefer to adopt more local strategies (Huang 1999, Shen, 2003), the significant contribution of Global strategies to reading comprehension among the low-achieving students found in this study is most joyful and precious. Global strategies help low proficient learners successfully obtain all three levels of comprehension in reading the most challenging text. The prediction of the positive relationship between Global strategy use and improved reading comprehension among the low proficient students seems to suggest that Global strategies, such as previewing the text, using keywords, and making predictions, should be encouraged. This pertains especially to students with lower language proficiency.

Last but not least, it appears that Global strategies not only effectively improve

reading comprehension, but are also mostly efficiently used as it is less

time-consuming. The pedagogical implication is that Global strategies need to be promoted and emphasized in reading strategy instruction. The details on how to maximize global strategies in class will be further discussed.

The Effects of Support Strategies on Comprehension

In addition to Global strategies, Support strategies also predicted comprehension gains in total recall scores, recall scores on main ideas, and recall scores on details, among all students, among the High proficiency group, and among the Low proficiency group. As shown in Figure 25, Support strategies increased all students’

three levels of comprehension. There were only three exceptions where Support strategies did not contribute to better comprehension. First, as shown in Figure 26, the High proficiency group’s use of Support strategies did not predict the gains of main ideas in reading Article 4. Second, when reading Article 2, the Low proficiency group’s use of Support strategies was not a predictor of main idea gains. Third, the Low proficiency group’s use of Support strategies did not predict an increase of recall scores on details in reading Article 4. It seems that Support strategies, the most frequently used among all strategies, assist reading comprehension in most cases.

However, in some situations, such as reading the most difficult text (Article 4 about St.

Patrick’s Day) and the text with unfamiliar topic (Article 2 about Pyramids of Giza), readers’ sole dependence on Support strategies will not result in better reading comprehension on main ideas and details. Students need to use other strategies, such as Global strategies (stated in the previous section) and Socio-affective strategies (will be elaborated in the next section), to improve their reading performance.

The Effects of Socio-Affective Strategy on Comprehension

Quite different from the positive effects that Global and Support reading

strategies had on reading comprehension, Socio-affective strategies resulted in both

positive and negative reading performances. As shown in Figure 26, the positive

influences were found on the Low proficiency group, whereas negative effects were on the High proficiency group. For example, the use of Socio-affective strategies predicted better reading performances on total scores and scores on main ideas while the Low proficiency group was reading Article 2. However, Socio-affective strategies produced a counter effect in achieving higher total scores and scores on main ideas when the High proficiency group was reading Article 2. The polarized effects of different proficiency groups may be interpreted that the Low group students in this study enjoyed social learning. Through interaction with peers, they felt more relaxed and more informed and therefore could better solve their problems. However, the High proficiency group may regard reading as an individual process and thus tended to search inwards to find the answers by themselves. The interaction with people may reduce the time necessary to think and thus lower their scores. The above phenomenon seems to demonstrate that the same strategy can have different effects on students with diverse proficiency levels. It also echoes the finding that readers with different proficiency levels apply different sets of strategies in coping with reading difficulties. Therefore, strategies are not good or bad in nature. The effectiveness of strategies depends on who is using under what circumstances (Anderson, 1991).

Strategies that work for a particular proficiency group are not necessarily good for another group with different language backgrounds. Most importantly, students’

personal learning styles and needs should be addressed to ensure better reading performance.

This section has discussed the influence of strategy use on different levels of comprehension. The next section will discuss students’ feedback on English Reading Online.

The Effectiveness of English Reading Online

Discussion will be geared towards four dimensions based on the results from

post-task survey and final reflection, including feedback on (1) strategy buttons, (2) learning effects, (3) text selection, (4) hyperlinks, and (5) interface design.

Feedback on Strategy Buttons

Overall, from the questionnaire results, students held positive attitudes towards the design of the strategy buttons, the learning effects, and the interface design of English Reading Online. Among the four types of strategy buttons, Global strategies had the highest number of strategy buttons favored by the users, Support strategies the second, Socio-affective strategies the third, and Problem-solving strategies the last.

This finding was consistent with what students reported in their final reflection. Two Support strategy buttons (Translation and Dictionary) and two Global strategy buttons (Preview and Keyword) were considered to be the most useful tools. Likewise, the least favored tools (Semantic Mapping, Speed Reading, Grammar, and Question) stated in the final reflection sheet were mostly in the category of Problem-solving strategies and Socio-affective strategies.

The aforementioned results of students’ dominant preference on certain strategy

buttons revealed one interesting phenomenon; that is, students were concerned with

the immediacy of the function provided in the CALL program. The strategy tools that

provided quickest delivery such as Translation, Dictionary, Preview, and Keyword

won students’ hearts. Only by clicking the function buttons could L1 translation,

definition of words, short introduction of the text, and important words of the articles

automatically come out. These strategy buttons, which demanded relatively short time

to use found in the analysis of time on strategy use, were perceived to be effective

strategy buttons by most students. On the other hand, strategy buttons that required

more efforts and demanded more time to use, such as Semantic Mapping, Speed

Reading, and Grammar, were not favored by students. In the online reading

environment where fast and circulated information crowded before students’ eyes,

drawing maps based on the text on their own, testing their speed, and consulting grammar rules seemed to be unappreciated tools. They were perceived as tools that required a great deal of efforts but showed little immediate or widely-noticeable return.

The above findings on students’ concern of immediacy prove that “speed counts in important ways within new literacies” (Leu et al., 2004: 1597). In the Internet era of vast information resources, the new literacies will be directly related to the rate that students can read, write, and communicate. The ability to quickly find, evaluate, communicate, and solve problems determines successful information acquiring. It is true that highly literate students will quickly skim pages and link to other web pages in a short time, while low literate students are still far behind, staying on one screen.

Instructional issues arise in turn. How to enable our students to succeed in a competitive environment defined by speed and information has become very important.

In addition to surveying preferences on strategy use, results of correlation analysis on the linkage between the actual usage of each strategy button and their attitudes towards its design showed that students’ perceived usefulness of strategy buttons positively correlated with their actual usage of these strategies. In other words, students used strategies they considered useful more often than those they did not think useful. This finding is consistent with reasoned action theory that learners’

intention to perform a behavior determines the action, and their intention depends on attitudes towards the behavior (Ajzen & Fishbein, 1980). The pedagogical implication is that because attitudes lead to actual behaviors, teachers may survey students’

attitudes towards strategy use prior to strategy instruction. In this way, individual

differences in learning attitudes can be addressed to ensure the best results of

comprehension instruction.

Feedback on Learning Effects

From the post-task survey and final reflection, it is interesting to find that students’ ratings of each individual strategy button were lower than the overall impression of the program in terms of the learning effects. Despite that their feedback on the usefulness of the 15 strategy buttons was only above average, students still held positive attitudes towards the effectiveness of this program. It can be interpreted that students did not master the strategies during the course of reading so they only mildly agreed with the usefulness of the strategy buttons. However, they were optimistic about the functions of this program and would like to try these strategy function tools more when they became more familiar with them. Students reported that they would like to read more online materials because they knew that the tools were useful and user-friendly. They would even recommend the program to their friends.

Another phenomenon worthy of our attention is despite that Translation function was favored by most of the students and translated L1 text could be obtained easily by a magic click, some High proficiency group students doubted the usefulness and learning effects of relying on Translation entirely. They did not think overusing the translator without digesting information by themselves would be good for learning.

This finding echoes the recent debates on the use of L1 translation in learning. For many years, the common wisdom is that learners need to develop their strategies to arrive at the meaning rather than depending on first language (Chapelle, 2003).

Therefore, L1 translation that provides access to the meaning of the input in many language materials or CALL programs should be abandoned. Recently, the use of L1 as a modified input has regained its legitimate place (Cook, 1999). From students’

feedback on Translator in the study, it is found that Translator is beneficial only if it is

used intelligently. Direct copying and pasting translated text is merely a mirage image

of understanding. Students need to read by themselves to get meaning across.

In addition, the comments that students made on the application of strategy tools in reading deliver important messages on the future CALL program design. In the case of using strategy tools in English Reading Online, students admitted that all of the tools were only facilitating comprehension. They needed to synthesize information they gathered from Keyword, Preview, Translation, etc., digest them, and read by themselves. The truly internalized knowledge that underwent the process of learning would be retained longer and would be useful for their future learning (Tseng, 1998). It seems to suggest that, when CALL researchers devote extensive efforts in developing help options that would pay-off in students’ learning, equally important is how to teach learners to use online help effectively and wisely (Chapelle, 2005).

Feedback on Text Selection

The results on text selection from final reflection have also shown interesting findings. The most favored article was the easiest one—Finding Nemo. The least favored was the most difficult text— St. Patrick’s Day. Student preferred Finding Nemo because the vocabulary was easy for them and the topic was motivating and familiar. They had enough background knowledge to deal with the text. In reading the most difficult article, St. Patrick’s Day, students’ lack of necessary background knowledge and vocabulary discouraged them from exploring the text.

The above finding shows that text difficulty determines students’ preference in

reading. Text difficulty, to most of the students in this study, not only depends on the

readability level judged from ready-made software (Readability), but also consists of

vocabulary and background knowledge. These two are the primary factors that

determine the ease of reading and thus influence their preference in reading. The need

for more known words in dealing with more challenging texts is consistent with Chi

and Chern’s study (1989) that vocabulary is the gravest concern among EFL learners

and looking up new words in the dictionaries is the most frequently employed strategies when encountering difficulties. Findings in this study also consolidate previous research that content schemata or cultural orientation is a crucial factor in L2 reading comprehension (Barnett, 1989; Carrell et al.; Johnson, 1982). Readers understand what they read only as it is related to their existing knowledge (Anderson, 1994). In addition to the knowledge of the vocabulary and syntax, text selection needs also take students’ background knowledge into consideration. Finally, students express their needs to choose text on their own instead of being assigned texts to read.

In curriculum planning, teachers need to strike a balance between teacher preparedness and student autonomy. Pedagogical implications will be further elaborated.

Feedback on Hyperlinks

Overall, students showed positive feedback on consulting hyperlinks to gain more information and the usefulness of pictures in associating with the text. However, still some students reported that they were overloaded with links and they felt lost in consulting links. The above findings direct our attention to the nature of hyperlinks, their advantages, and disadvantages. The rise of hypertext revolutionizes our way of looking at the text. Instead of arranging the text in a traditional linear way and assuming the reader to start reading from the beginning to the end, a reading text can be created through the combinations of links, providing tables of content and indexes as organizing maps (Alessi & Trollip, 2001). Just like what we have found in this study in which hyperlinks provided immediate and relevant information to aid comprehension, advantages of hyperlinks lie in that hyperlinking can take readers to other relevant sections, which may facilitate knowledge acquisition (Dillion &

Gabbard, 1998; Valmont, 2003).

However, the disadvantage is that the structure of hypermedia itself will increase