The Reading Matrix
Vol3, No.3, November 2003

Interactive Hypertext and the Development of ESL Students' Reading Skills
Loretta F. Kasper

Abstract

Recent research suggests that hypertext can provide an effective tool for developing reading skills. Because hypertext is a relatively new textual medium, and because it is likely to become more dominant in the future, research is needed so that both reading instructors and students may be empowered to use hypertext to its full advantage. This paper reports on the results of a two-year study designed to assess the effects of different types of hypertext on the development of ESL students' reading skills.

In a society that places increasing emphasis on the importance of information and communication, strong reading skills are essential not only for students' academic success, but also for their social and economic advancement. As Grabe and Stoller (2002) assert, "As we enter a new century, productive and educated citizens will require even stronger literacy abilities (including both reading and writing) in increasingly larger numbers of societal settings" (p. 1). Unfortunately, many college students, especially those for whom English is a second language, do not read well enough to ensure comprehension nor to meet minimum competency standards set by institutions of higher education. For example, during the 1999-2000 academic year at my institution, 77% of the ESL students who, after one semester of high intermediate instruction, needed to continue in developmental English did so because they had failed the end-of-semester reading examinations. Clearly these students need much support in building the critical reading skills necessary for higher-level learning.

The results of my own, as well as other recent research, suggest that through its use of interactive hypertext, Internet technology can provide an effective tool for developing reading skills (Kasper, 2002; Lomicka, 1998; Soe, Koki, Chang, 2000; Tierney, Kieffer, Whalin, Desai, Moss, Harris, & Hopper, 1997; Warschauer, 1999). This paper reports the results of a two-year study designed to explore and assess the possible effects of hypertext on the development of ESL students' reading skills.

In contrast to traditional print, in which information is presented in a linear fashion (i.e., in predefined sequences), the presentation of information in hypertext is nonlinear and represented in a semantic network in which multiple related sections of the text are connected to each other. By following links through the sections of the text, the reader actively engages with that text, choosing a path that is most relevant to his or her needs or interests. Well-designed hypertext systems can facilitate interaction between readers and texts (Rouet, Levonen, Dillon, & Spiro, 1996), thereby enhancing comprehension and building critical reading skills. Because hypertext provides easy access to multiple cross-references on related topics across several documents, or screens, it fosters a nonlinear and flexible pattern of exploration and discovery that encourages a natural juxtaposition of ideas presented (Tierney et al., 1997) and helps to promote cognitive flexibility necessary for the integration and consolidation of knowledge gleaned from a variety of sources (Mishra, Spiro, & Feltovich, 1996).

Following new and different links during subsequent online sessions encourages student readers to develop strategies for how to approach text content, as they call upon appropriate relevant schemata, or background knowledge for how to interpret that content. Reading hypertext is a naturally dynamic, recursive, and integrated process, one that provides multiple opportunities for students to acquire, test and reframe knowledge through cognitive reconstruction of text, intertextual analysis and exposure to varied perspectives on issues. Thus, hypertext may promote increased comprehension through the elaboration and integration of new information into the existing knowledge network as readers create and expand the cognitive map that guides their construction of meaning.

Although nonlinear hypertext can offer students many benefits, Rouet and Levonen (1996) advise that without overt instruction in how to navigate hypertext effectively, students may become lost in a sea of information, potentially experiencing cognitive overload. Gillingham (1996) suggests that attention to text, task, and context is necessary for effective comprehension of hypertext. Foltz (1996) cautions that hypertext may present a problem for students with poor reading skills because it causes an additional processing load by making the reader responsible for navigating the text. Having to choose where to go next can take students' attention away from processing the text, with the possible result that they generate fewer hypotheses as they read, making it harder for them to integrate the information presented. Foltz' work pointed to two key factors in hypertext comprehension: (1) the coherence of the text and (2) how the reader's goals affected strategies used. Each of these researchers points to the need for extended research to elucidate the role of hypertext in building reading skills.

Currently research on hypertext suggests that while it has the potential to be a powerful tool for building reading skills, without proper instruction, hypertext can lead to information overload and confusion, especially in developmental readers. Because hypertext is a relatively new textual medium, and because it is likely to become more dominant in the future, research is needed that explores its features and their effects on reading comprehension and performance. In this way, both reading instructors and students may be empowered to use hypertext to its full advantage.

The study reported here had several objectives: (1) to design different types of hypertext and then assess and evaluate the effects of each on reading comprehension and performance, (2) to collect feedback from students about which types of hypertext they believe are most helpful, (3) to apply results to revise and refine hypertexts to help students improve reading skills more efficiently, (4) to apply results to teach students more effective strategies for reading and comprehending all texts, whether hypertext or traditional print, and (5) to construct and make available on the Internet a body of interactive online texts that can be used to develop students' reading skills.

THE ESL COURSE AND CURRICULUM

The study was conducted over a period of four semesters and involved 100 ESL students. The students who participated in the study were enrolled in a high intermediate course called “Developing Fluency in Reading and Writing for ESL Students” (ESL 91), which is the first of a two course developmental sequence in reading and writing. High intermediate here is defined as an entry level TOEFL score of approximately 425. The ESL 91 class meets six hours per week in three two-hour blocks for twelve weeks. Students meet in the computer lab each week for one of those two-hour blocks. My course follows a sustained content curriculum (see Pally, 2000) that develops ESL students' literacy skills through an activity called focus discipline research (Kasper, 2002a, 2002b).

To help my ESL students build key literacy skills, I have designed a curriculum of technology-enhanced focus discipline research that is based on the principles of sustained content study. An extensive description of the focus discipline curriculum is beyond the scope of this paper (interested readers are referred to Kasper, 2002b); however, briefly defined, a focus discipline is a subject area (e.g., psychology, biology) that individual students choose to research extensively over the course of the semester. I ask my students to choose a focus discipline from the ten subject areas contained in their textbook. 2 Students base their choice on personal interest and/or college major, and because students have chosen to do extensive research in that discipline, they are actively invested in a learning experience that is personally meaningful and important. 

TECHNOLOGY AS A TOOL FOR INSTRUCTION

My ESL students hone English language skills, build their overall knowledge base, and develop literacy skills through their use of text-based computer-mediated communication, intensive reading and research using Internet hypertext documents, and their production of written essays and individual and group research projects based on their research efforts. As noted earlier, my ESL class meets six hours per week in three two-hour blocks for twelve weeks. Students meet in the computer lab each week for one of those two-hour blocks. Many of my ESL students have computers at home; those that do not have additional access in the college during open computer lab times.

My focus discipline research curriculum makes extensive use of group collaboration within the physical classroom, where students work together to share and build knowledge with others studying the same focus discipline. Internet technologies have allowed me to enhance and expand this curriculum within and beyond the physical classroom. I have incorporated a course web site at http://kccesl.tripod.com/ and an online course component, called “ Interdisciplinary English and the Internet ,” on Blackboard.com. 3 I post weekly computer lab assignments to the Announcements area of the Blackboard course. Computer lab work consists of a variety of different activities, from Internet research to online reader response exercises to online practice reading and writing tests.

I design general class activities to teach students vocabulary and language structures and to provide them with day-to-day practice in complex interdisciplinary texts. These general class lessons provide guided instruction on how to dissect a text, search for clues to meaning, and compose cogent responses to inferential questions and essay prompts. Because students do a good deal of Internet research as part of the course, they must search for, evaluate and read an extensive number of hypertexts.

As Rouet and Levonen (1996) note, although hypertext can offer students many benefits in terms of developing reading skills, without instruction in how to navigate hypertext effectively, students may become lost in a sea of information, potentially experiencing cognitive overload. Moreover, students need to be able to discern which texts are reliable and valid for the purposes of their research. Guided practice that familiarizes students with criteria for evaluating web sites and that takes them through a hypertext document facilitates their acquisition of the cognitive strategies necessary for finding, navigating and comprehending nonlinear texts. To help my students build skills in searching for, sorting through and evaluating Internet texts, I give them a multi-part guided research activity . This activity, which is posted on the course web site, as well as on the online Blackboard course component, teaches students not only how to search for information on the Internet, but also how to evaluate the resources they find there. We do the guided research activity in the computer lab.

Students begin by reading the text Four Nets for Better Searching ; this text provides suggestions for how to conduct an effective Internet search. Students then study an Evaluation Criteria Chart (Kapoun, 2000), which recommends five criteria for evaluating web sites. We go over each of the criteria, using sample web sites on global warming, one of the subjects studied in the course, to illustrate the principles listed. . Finally students complete an Evaluation Exercise that directs them to four different web sites on global warming and asks them to evaluate each site based on the five criteria provided in the Evaluation Criteria Chart. They are then asked to expand their knowledge base through an independent Internet search. To do this, students need to become familiar with Internet search engines, such as Yahoo !, AltaVista , Google , or Vivisimo .  They must learn how to enter keywords to identify the information they want.  Then once the Internet search engine has returned a list of "hits" for the keyword, students must go through the list to identify the most appropriate and/or useful information.

The next part of the activity provides students with additional practice in searching for and evaluating Internet resources (see Appendix A). Students are directed to the EPA global warming web site. Students are guided through the information and links found on this web site. Together we explore and discuss all of the information presented. Then students are asked to practice accessing related web sites on their own. Students must write up a short response to the information on these sites. After completing their search, students share the resources they have found and their responses to them through focus discipline group work in the classroom.

The study reported here was conducted in class, and the texts used were an integral part of the course. I constructed several types of hypertexts 4 , each focusing on subject areas (e.g., environmental science, psychology) studied in the course. Initially, I had planned three types of hypertext conditions: (1) glosses , in which links provide popup vocabulary definitions, (2) controlled hypertexts , in which links lead to a predetermined and limited number of texts on the topic, (3) free hypertexts , in which students read a main text and then are directed to freely explore the Internet for other texts related to the topic. However, during the initial semester of the project, it became clear that two additional conditions (for a total of five) were needed: (4) controlled hypertexts with glosses, and (5) free hypertexts with glosses . These were designed and incorporated into the study during the second semester of the study. Thus, there were a total of five hypertext conditions in the overall study.

As integral components of the ESL course, the texts used in the study were online versions of print texts contained in the book Interdisciplinary English (Kasper, 1998). The selection of glossed vocabulary items in the online versions of the texts corresponded to those vocabulary items highlighted, but not glossed, in the print versions of the texts. Students reading print versions were required to complete vocabulary exercises in which they used the context of the reading passage to write a definition for each of the highlighted items. The highlighted vocabulary items and the corresponding online glossed items were selected because these words were likely to be unfamiliar to students and were words that were deemed necessary for overall comprehension of the text. Students reading online versions of the texts were provided with pop-up vocabulary definitions via glosses to use technology to help relieve some of the cognitive load associated with reading online texts. It should be emphasized that students read only one version, either print or online, of any given text.

I conducted the study as part of the ESL 91 course work during the two-hour computer lab time. I used the hypertexts as practice reading exams to prepare students for their end-of-semester reading assessment. Students read each of the hypertexts and answered comprehension questions online. They then submitted their answers to me via an online form . Students were instructed to print out a copy of their answers before submitting them. After answers were submitted, the form automatically took students to a page with the answer key 5 , where they were able to compare the answers they had submitted with those on the answer key. Thus the activities in this study were designed to be integral parts of the ESL course with the goal of helping students build both reading and writing skills. Printed texts in the course textbook served as control texts to assess the general effects of hypertext itself on reading comprehension. For the printed texts, students wrote their answers to comprehension questions on paper and handed them into me in class.

To receive full credit, responses needed to demonstrate a clear understanding of the text and the relevance of certain information to the issues presented in the text. Initially I scored the students' answers to each of the texts; as a control, another faculty member, who did not have knowledge of the specific details of the study, also scored the answers. Students' responses were scored based on the same criteria used in scoring responses to the departmental reading examination; the answers were normed to a standard prescribing information necessary for a correct answer. After all texts had been scored, I tabulated the results for each of the five types of hypertexts, as well as the print texts, which served as controls. There was an interrater reliability of .98 on the scoring.

The results revealed that the quality of students' responses varied as an effect of the type of text read (see Appendix B for sample student responses). Controlled and controlled gloss hypertexts resulted in better performance than either print or free hypertexts. The most detailed answers of all were produced by students' responses to controlled hypertexts with glosses. When paired with controlled hypertexts, it appears that glosses aided students in understanding unfamiliar vocabulary, and the extended information provided in the controlled hypertext enabled them to form a more comprehensive answer to the question. Students' answers to questions on controlled hypertexts indicated that they actively used this extended information in forming their answers.

There was little difference between the detail in students' responses to gloss hypertexts, free hypertexts, free gloss hypertexts and print hypertexts. In their responses to questions on each of these types of text, students primarily used the information presented in the “main” text. Even though free hypertexts required them to search the Internet for an additional source of information on the topic of the text, students tended not to use this additional information in forming their responses to the comprehension questions. This result was interesting and somewhat unexpected given the fact that students had received extensive instruction on how to search for and evaluate Internet texts.

Based on students' responses to comprehension questions and feedback questionnaires, it appears that having to search for additional information on their own was problematic. This was true despite the fact that they had received specific instruction in how to search for information efficiently. In their responses to feedback questions, students indicated that they were unsure of which information gleaned from the free search was necessary to answer the question. In addition, they stated that their searches did not necessarily produce results that were useful in responding to each of the questions in the comprehension exercise. Students who were more practiced in doing Internet searches tended to find sites that were more useful in helping them respond to the questions. However, even those students who were able to find useful sites tended not to use the information in those sites in responding to the comprehension questions.

Students' scores on the comprehension exercises derived from controlled hypertexts, particularly those that contained glosses, were significantly higher (mean=88% correct) than scores on exercises derived from any of the other three types of hypertext: gloss (mean = 75%), free (mean=65%), or free gloss (mean=68%). It should be noted that only the mean scores students attained on gloss and controlled gloss hypertexts would have been considered passing on the college reading examination, where the passing score is 70%. In contrast students' scores on comprehension exercises derived from free, free gloss, and print texts (mean =60%) were lower, and would not have been considered passing on the college reading assessment.

All forms of hypertext produced higher scores on comprehension tests than did print texts; this improved performance may be due to a number of factors. Technology use encourages students to spend more time on task, providing them with increased opportunities to process linguistic and content information (Kasper, 2000a). Thus, it is possible that students in this study may have devoted more time to the task when reading any type of hypertext, and this increased attention may have yielded better performance. However, it appears that the additional information provided by gloss and controlled hypertexts further facilitated comprehension and so led to better scores on comprehension exercises. In spite of the differences in scores obtained with the various textual forms, teasing out the specific role of hypertext in building reading skill is difficult from comprehension scores alone; for this reason, getting students' feedback on the texts used in the study was critical to providing a clearer picture of how students used the texts and the benefits they derived from them.

At the end of the semester, I asked students to complete an online feedback form , which asked for their reactions to the different types of hypertexts used in the study. Students were asked what they thought of each type of text, how easy or difficult it was to use. They were asked to indicate their preferences among the different types of hypertexts and to explain how/why they believed that a specific type of hypertext was or was not useful in helping to improve comprehension. I used their feedback to revise and improve the hypertexts designed for the study.

An overwhelming majority of students (95%) reported that they found gloss hypertexts to be the most useful for improving vocabulary comprehension; of these gloss hypertexts, the controlled hypertexts with glosses were rated the easiest to use and the most effective in helping them construct a complete answer to the comprehension questions. For example, one student reported, “Hypertexts were very helpful in my readings. The most helpful were glosses and controlled hypertexts. Glosses because if I didn't know the meaning of the word I just easily could check it with my mouse. Controlled because it explained more about the subject.” Because ESL 91 focuses on developing both reading and writing skills, students are required to produce written essays on the topics of each of the readings done in the course. One student noted an additional benefit from controlled hypertext; not only was it easier to understand, but it also provided additional information to incorporate into the essays he needed to write for the course. The student said, “…the additional information from controlled hypertexts was very helpful for improving and spreading in our essays.”

In contrast free hypertexts , with or without glosses, were rated the most difficult to use by 85% of the students in the study. One student responded that “ Free hypertext was difficult to use, because I use much time on search information, and don't know the search information that I find have to be useful or not;” while another stated, “ I think the Free is difficult to use because it makes me confuse to the main text and the addition text.”

As noted earlier, students' performance on reading comprehension exercises mirrored their preferences. Scores on gloss hypertexts, particularly controlled gloss hypertexts, were significantly higher than those on free hypertexts. Students' said that gloss hypertexts enabled them to read with greater comprehension because these texts provided easy access to the definitions of new vocabulary words. This result supports the findings of Lomicka (1998) who found that reading computer texts with glosses may promote a deeper level of text comprehension. Students in this study also said that controlled hypertexts made the text clearer by providing links to specific relevant information. Students' preference for controlled hypertext supports Foltz' (1996) claims that text coherence plays a powerful role in students' comprehension of hypertext.

Overall students disliked free hypertext because they found it confusing. They complained that free hypertext led to finding too much information, making it easy to get lost in exploring the links and forget about the main topic. They also said sorting through and evaluating the usefulness of all the different definitions and opinions on the topic was time-consuming. In addition, free hypertext was intimidating to students who were less experienced with the computer. These results support the claims of both Foltz (1996) and Rouet & Levonen (1996) and are particularly interesting since students here received instruction both in how to navigate hypertext and in how to evaluate information found on the Web.

Overall this study has provided student feedback and performance data that support the claims of previous researchers and has also helped to elucidate issues that need further attention in future studies. Student feedback and performance data indicated that controlled hypertexts with glosses led to the highest levels of text comprehension. In their responses to the feedback questionnaire, students indicated that glosses were extremely useful in helping them to understand new words in the text and in building their overall vocabulary. Therefore hypertexts designed for use in developmental reading courses should incorporate glosses.

Students in this study indicated that they did not like using the free hypertexts because they found them somewhat overwhelming. The results of the study reported here suggest that students need more extensive training in how to navigate the Internet and evaluate the information they find there. Given these results future studies that carefully examine students' exploration of free hypertexts are needed. These studies should collect specific data on the number and content of the sites visited. In addition, it is possible that students would be more comfortable working with free hypertext if the subject addressed were one with which they were highly familiar. It is also possible that with increased instruction and practice, students will experience less information overload with online resources, and they may become more proficient at finding and pursuing links that lead them to more coherent, and more useful, free hypertexts. Moreover, learning how to evaluate the quality of information is a key skill, not just in academic settings, but also in life itself. To help students build this skill, ESL courses should incorporate different textual media (Chun & Plass, 1997) and that ask students to critically analyze the information presented in print, visual and hypertexts.

The primary goal of this study was to assess the role of hypertext in developing ESL students' reading skills through a comparison of students' performance with different types of hypertexts. Unfortunately because this study was conducted as part of the actual coursework, each type of hypertext focused on a different subject area. It is possible that students may have found some subjects easier or more difficult than others, and this may have influenced their performance and their reactions to the type of hypertext used. Future studies might use the insights gained from this study and design different types of hypertexts, each of which focuses on the same subject. This would eliminate the potential confound of variations in difficulty due to subject area, and so provide a clearer picture of the effects of different types of hypertext on reading comprehension and performance.

Despite these caveats, the results of the study reported here can begin to provide insights into factors that should be considered when designing and using hypertext as a tool for reading instruction as well as directions for future research. As more and more materials become available online, it is important that ESL instructors come to recognize the potential of and learn how to use hypertext to its full advantage as a valid tool for instruction. When carefully designed and incorporated as an integral part of the ESL curriculum, Internet hypertexts can provide students with a highly motivating textual format that not only enriches their overall learning experience, but also helps them develop the literacy skills they will need to succeed in the age of information.

Endnotes:

1 This research was supported by a grant from the City of New York PSC-CUNY Research Award Program.

2 The textbook Interdisciplinary English (Kasper, 1998) contains readings in linguistics, environmental science, computer science, psychology, sociology, business, anthropology, diet and nutrition, biology, and mathematics.

3 The Blackboard.com course may be found at http://online.cuny.edu:8001/

4 A representative list of hypertexts developed for this study may be found at http://kccesl.tripod.com/hypertextstudy/

5 Both linked examples are representative of the online question forms and answer keys used in the overall study.