Glossing as a scaffold for vocabulary learning and reading comprehension:
An action research and its implications for EFL teaching in Thai universities

Snea Thinsan, Language Education, Indiana University

            Why should “glossing” be studied and considered for use as a scaffold in EFL reading classes at undergraduate level in Thailand?  My experience as a lecturer at this level informs me that the following scenarios make a strong rationale:

            My colleague teachers have to teach reading materials that require the students to comprehend complex sentence structures, infer from sophisticated discourses, and learn many more unfamiliar words. They see their students only three times a week, 50 minutes each session. The students are of mixed abilities. The weaker ones, including students below the 50th percentile on the grading system of any previous English course, normally need a lot of teacher assistance in making sense of difficult reading texts. Most of these students also possess rather poor background knowledge about English grammar and have a limited pool of vocabulary. Most students perceive their limited vocabulary as a major barrier to their reading comprehension (Wongsri, Boon-long, & Thinsan (2000). Ironically, however, (Wiroonrat & Thinsan, 1998) found that the students obtaining poorer grades in English courses spent less time on homework or on activities related to the English language.  Essentially, there seems to be a mismatch between the students’ English background and the expectations of the undergraduate English reading courses.

            The gap between the students’ English background and the demanding reading tasks inevitably cause teachers difficulty in facilitating and finishing the lessons as required by the prescribed curricular. Many teachers use Thai to help students understand the texts and to do the tasks that accompany, which limits the students' exposure to English even further, given the only three meetings each week.  Usually, these teachers find it impossible to finish each lesson because there is too much to "explain". Likewise, other teachers who insist on using only English to aid the students’ learning seldom complete all the tasks suggested in the curriculum. Worse, they usually receive low responses from the students and often find themselves having to supply answers to their own questions. These teachers are, in turn, perceived by the students as being unhelpful, while the students, not surprisingly, are regarded by their teachers as lacking learning autonomy and motivation (Wiroonrat, et al., 1998).

            A most sensible way of assisting students in their reading and vocabulary learning in the above context is by using electronic glosses.  Gettys, Imhof, & Kautz (2001) claim that online, or electronic, glosses help students to “enhance general comprehension, improve vocabulary retention, and save student’s time and effort in reading L2 texts” (p. 91).  Nonetheless, glossary has been discouraged at Chiang Mai University because many teachers believe that students will learn better if they try to guess word meanings from context. Giving a vocabulary list is also seen as a potential cause of students’ absences because they see no need to come to class when a list of words to be tested are already in their hands.

            The above scenarios and the contradicting view against glossing cry for an investigation on electronic glossing. The literature concerning electronic glossing also shows that claims about whether it works and how it works in various situations will still need to be confirmed by more empirical evidence. This small-scale research inquiry, hence, further explored the effects of materials containing electronic glosses, in comparison with those containing paper-based glosses and no glosses, on vocabulary retention and reading comprehension of Thai students. In this paper, the aim is to weave the results of such study into EFL reading and vocabulary teaching practices. The first section clarifies definitions of a few terms used in this paper and offers a brief overview of glossing with regard to vocabulary and reading comprehension. The second section reviews some relevant studies. The third section offers details about the action research. The forth section reports and discusses the results, offering implications for classroom application where appropriate. The last section then draws the threads together.

Overview

Definitions

            A few terms used in this paper have similar meanings, but they are in fact slightly different.  Thus, it is important to distinguish them.

            Gloss:    “a brief definition or synonym, either in L1 or L2, which is provided with the text” (Nation, 2001, p. 174).

            Electronic/ Computerized glosses: Glosses in electronic, or computerized, format that can be delivered on computers through intranets, the Internet, and on standalone systems

            Online glosses:  Glosses in electronic format that are delivered either through the Internet or through the intranet

Theoretical overview
 

            Glossing, despite its long history, is not by itself an issue most EFL teachers regard as a priority topic in teaching reading at undergraduate level at Chiang Mai University. The priority is usually given to practicing sophisticated reading skills. Roby (1999) describes the historical development of glossing in language teaching, dating back to the Middle age, when glosses were used by learners of Latin as a foreign language. They generated glosses to aid their reading. Teachers adopted glossing later by producing glossary, or word list, to help their students read. The later development of glossaries resulted in dictionaries. The latest development of glossing can be seen as electronic dictionaries distributed either in CD-ROMs or through the Internet. The shift from paper-based to computer screen reading and the recent inclusion of multimedia have made glossing an issue for research because, despite the perceived potential, many questions still need to be asked and more empirical evidence is still required before its effective uses can be confirmed.

            Cubillos (1998, cited in Omaggio-Hadley, 2001), states that computers can be useful in several ways: facilitate vocabulary learning; increase the students’ awareness of linguistic forms; support reading and writing development, help to track learner behavior; enhance motivation among students; and enhance teaching resources with their ability to incorporate multimedia and many sophisticated features. Online glossed materials have been found to yield several benefits in many studies (Nagata, 1999; Lomicka, 1998; Hong, 1997; Lyman-Hager et al., 1997; Chun & Plass, 1996; Davis & Lyman-Hager, 1993). 

            The advantages of electronic glosses over the traditional paper-based glossary are often claimed (Leffa, 1992).  Gettys, et al. (2001) view electronic glosses as “a means of aiding second language reading by performing most bottom up functions, (which thus relieve) the working memory and attention from the burden of cognitive overload and allowing more of the reader’s attention to go toward top-down processes” (p. 93). Electronic glossing, they elaborate, “provides fast and easy access to the meanings of unknown words, makes up for insufficiently automatic bottom-up processes and deficiencies in processing capacity, and thus allows the reader to attend to top-down processes” (p. 93). 

            However, Nation (2001) observes that “most studies have found that glossing has a positive effect on vocabulary learning,” (p. 176) but that “the effects of glossing on comprehension are mixed” (p. 177).

Literature review

            Given the diverse findings about glosses, the literature about electronic glossing should therefore be explored so that appropriate research questions can be generated. The following review will cast light on most related studies during the past decade.

            Among the earlier studies, Leffa (1992) compared the efficacy of electronic dictionary glossaries with the traditional paper dictionary on text comprehension in a translation task among 20 university students enrolled in a beginner EAP course.  They found that the electronic dictionary users understood more of the passage (86%) than the paper dictionary users (62%) and needed 50% less time to translate the passage. Hong (1997) also found that students using Online glossed material while reading Chinese business texts spent less time than that they would have otherwise spent using paper-based glossed material. In these studies, however, there was no attempt to investigate the students’ look-up behavior and its effects on their performance, so they did not reveal how electronic glosses were used by the subjects.

            Other researchers responded by conducting studies that incorporated learner behavior tracking into the research design (see, for example, Chun & Plass, 1996; Davis & Lyman-Hager, 1997; Lomicka, 1998; Laufer & Hill, 2000). Davis et al. (1997) examined what 42 intermediate French learners at an American University actually did when they read in the L2 with computer assistance.  They found a negative correlation between the looked up information and the course grade. They also found that the type of information accessed on the computer had no effect on comprehension. In addition, the relationship between subjects’ consultation of the gloss and quantity and accuracy of recall varied greatly across the group.  Lastly, while students showed unanimous preference for the electronic dictionary glosses to the traditional mode, there was no clear evidence that it had actually enhanced comprehension.  Another similar study by Lomicka (1998) examined relationship between types and number of glosses consulted and the number of inferences generated and whether multimedia annotations allow the reader to build a situation model. They used the think aloud protocols to detect and record user behavior during a computerized reading task. The results were, again, fairly unclear whether glosses helped with reading comprehension.           

            When multimedia annotations became increasingly used, Chun and Plass (1996) investigated the effects of multimodal glosses on vocabulary acquisition.  They wanted to find out how well vocabulary was learned incidentally during a reading comprehension task; how effective annotations with different media types were for vocabulary acquisition; and the relationship between look-up behavior and performance on vocabulary test.  They conducted three separate studies employing the same methodology but a different number of subjects. In all three studies the subjects took an unannounced vocabulary test containing the words that had different glosses in the program.  The students had to provide an English equivalent to German words, and indicate which glosses they felt had helped them remember the word.  Despite generally positive results, the data suggested that there was no one mode that was helpful to all learners, but that different types of learners looked up different types of annotations and learned more successfully with the preferred modes. 

            In their later study, Plass, et al. (1998) examined the role of individual differences in multimedia learning among 103 subjects using the same computer program and short story as in their 1996 study.  Their questions were on whether visualizers, those who prefer visual information, and verbalizers, those who prefer verbal information, differ in their behavior in multimedia learning environment, and how their differing needs can be supported to improve their overall learning outcomes.  The log files of subjects’ look up behavior were also used.  They found that the combination of visual and verbal learning modes and the subjects’ access to preferred modes were related to higher performance on the vocabulary test. The comprehension test results showed an interaction in which verbalizers did not profit from the addition of visual information to their preferred verbal type of information to the same extent as visualizers do.

            Among the most recent studies, Laufer and Hill (2000) investigated the relationship between the L2 learners’ dictionary lookup patterns and their vocabulary retention among 32 EFL students from Israel (IL) and 40 ESL students from Hong Kong (HK). Different types of dictionary information were incorporated into a CALL program called Words in Your Ear. Data analysis revealed that all target unfamiliar words were looked up, but the percentages of the retained words for the two groups were apparently different 33.3 for the IL learners and 62 for the HK learners. Despite such discrete results, the researchers concluded that electronic dictionary has a positive effect on incidental vocabulary acquisition. However, no particular lookup strategy was found to yield the best results. They finally suggested that multiplicity of lexical information was correlated with better retention, and greater attention to a word during the lookup matters more than the number of lookups.

             Similar finding was obtained from a study of intermediate Russian students by Comer and Keefe (2000). The study revealed that glossing of vocabulary alone did not ensure that learners at the intermediate level could successfully generate the text's overall factual meaning (main points and significant details) from word-level vocabulary glosses.

            In a most recent study, Gettys, et al. (2001) compared the effects of two glossing methods, namely sentence-level translation equivalent form of L2 words and basic dictionary form, on text comprehension and vocabulary retention.  They found that the group exposed to basic dictionary forms showed a significantly greater overall retention of vocabulary and better attitude toward it. They concluded that the easy and fast access to the meanings of unfamiliar words motivated the subjects to read more and, as a result, the natural redundancy and recurrence of lexical items took place; so, vocabulary gain and retention were enhanced. On the other hand, the effectiveness of the two techniques on the global comprehension of the text was not found to be significantly different nor consistent among the subjects.

            In sum, the literature above suggests several points. First, the effects of electronic glossing on vocabulary learning retention are generally positive, but those on reading comprehension are rather ambiguous. Second, there existed several variables in each of these investigations, which caused difficulty in pinpointing what had influenced the results  Third, the distinctions between the effects of electronic gloss, paper-based gloss, and non-gloss are not clear. Forth, several studies elicited positive learner attitudes towards electronic glosses, but others focus more on their learning behavior while reading. Firth, only one of the presented studies involved EFL learners and none involved Thai students. Lastly, most of these studies used convenience or purposive samples; therefore, generalizability is very low.  Therefore, this action research will readdress some of the above questions. These questions can be operationalized as:

1)      Do the gloss users have better reading comprehension than the Non-gloss user?

2)      Do electronic and paper-based glosses have different effects on reading comprehension?

3)      Does glossing help the learners have better vocabulary retention?

4)      Do electronic and paper-based glosses have different effects on vocabulary retention?

5)      What are the attitudes of the participants toward glossing?

The study
 

Participants

            The participants in this study were three female, Thai students working toward their master’s degree at Indiana University Bloomington. They were included in this study on voluntary basis.  Their background English proficiency is illustrated in table 1.

            It can be seen that these participants had had slightly different English background before the treatments. The results of the pretest on their prior vocabulary knowledge were also different. Participant 1 had the highest TOEFL scores, but knew fewer words than did Participant 3. Participant 2 scored lower in TOEFL and vocabulary pretest than the other two, but received the highest TWE score. Participant 3 had already known more target words than the other two. However, these slight differences were expected to have minimal influence on the participants’ performance after the treatments.

Data collection

            The data were collected through the following steps.

            First, the participants provided their background information and identified the unknown words in eight reading passages taken with the author’s permission from online TOEFL practice exercises available at http://www.eslforum.net/toeflexercises.htm. The stories are of general knowledge, containing vocabulary items appropriate for the level of the participants in this study—postgraduate students.
            Second, the participants took a vocabulary pretest. They supplied parts of speech and the Thai equivalents of the selected words from three reading passages containing most commonly unknown words among the three participants. This test was also used to double check whether the participants had correctly identified the words as unknown in earlier step. It indeed informed the researcher that there were some words the participants though they had known, but could not supply the correct translations in Thai or even the parts of speech.

            Third, the three selected reading passages were redesigned into three different sets: reading passages with electronic glosses; reading messages with paper-based glosses; and reading passage passages without glosses.  The electronic glosses and the paper-based glosses contained the same gloss contents taken with permission from the Cambridge International Dictionary of English available online at http://dictionary.cambridge.org/. The electronic glosses were design to appear on a separate frame, but the same screen as each reading passage to promote ease of use, fast access, and unobtrusive interface.

            Forth, the participants were assigned randomly to receive one of the three reading packages. Then, they were told to learn the meanings of the target words (the underlined words for non-gloss user) and try to understand the content of each passage for subsequent tests.  They were also asked to record the time spent and the manner they used the materials.
            Fifth, all the participants took a comprehension test at the same time after studying the materials for two days. Exactly one week after the comprehension test, they took a vocabulary posttest. The comprehension test questions were selectively taken from the online source. The vocabulary posttest employed the cloze format. The closed sentences were chosen from the online British National Corpus to ensure authenticity of the language. Each sentence was selected on basis that it provides the clearest context and the same meaning of the target word as that the participants encountered in the passages.

            Finally, the participants answered a questionnaire designed based on the emerging issues in the logbooks they had recorded while studying the material. The aim was to elicit their attitudes and comments about the materials to which they had been exposed. Then, an informal telephone interview of each participant was conducted to check reliability of the questionnaire responses and to elicit more information.

Data analysis

            The data are tabulated where necessary. Given a very small sample, simple calculation methods are employed. Then, the qualitative data are used for triangulation of the data.

Data analysis, discussions, and implications

            The results of this study are presented and discussed in light of the five research questions listed earlier. Implications for classroom application will also be herewith argued for.

            Research question 1: Does glossing help the learners have better reading comprehension (than the Non-gloss user)?

           
           Table 2 shows the reading comprehension test scores of the three participants in three reading passages.

Analysis and discussion

            It appears that vocabulary glossing did not have straightforward effects on reading comprehension in this study. As shown in table 2, both the Electronic gloss user (Participant 1) and the Non-gloss user (Participant 3) obtained exactly the same overall reading comprehension scores (11 pts). A striking point is the much lower score gained by the Paper-based gloss user (3 pts). This finding is in line with the results of earlier studies (Davis et al., 1997; Gettys, et al., 2001) in that the effects of glossing on reading comprehension are ambiguous due to the complex factors involved in the reading process.

            Analysis of other relevant factors, however, lends some useful insights. First, the time spent by participants on the materials and the test correlated positively with the scores. That is, the more time they had spent, the better scores they received (see table 3 below). Second, the participants’ attitudes toward the materials seem to affect their scores. The Paper-based gloss user, who spent the least time on the material and the test and scored the least, showed her negative attitude about (paper-based) glossary. She stated, “Glossary maybe make readers get a little bit confused if they are not so good in reading. So they should understand passage by reading context.”  The Electronic gloss user, on the other hand, expressed her positive attitude about the convenience she was offered; likewise, the Non-gloss user expressed that it was good for her to try to help herself first. Thus, it might have happened that, with the negative attitude, the Paper-based gloss user decided to spend less time and probably pay less intention to the tasks and the test, resulting in lower scores. Lastly, the prior vocabulary knowledge appears to help with reading comprehension. The Non-gloss user, who scored the highest in the vocabulary pretest (see table 4), did quite well in reading comprehension test in spite of the absence of gloss aid. Note that the Paper-based gloss user also scored the lowest in the vocabulary pretest.

Insert Table III here:

Implication and application

            The above analysis sheds light on a few useful ideas for classroom practice. To begin with, the analysis of time on the materials suggests needs to increase students’ exposure to comprehensible inputs. It implies that the teacher, as Nation (2001) suggests, should create the opportunities for the students to read both intensively and extensively. For intensive reading, the teacher should plan, for each reading passage, different activities. These activities should expose the students to different aspects of the text, both form and meaning, by using glosses containing the help the students need. At the same time, the students should be encouraged to read extensively. Different text types on the same topic can be found on the Internet. For instance, if the reading topic is about the Titanic, the teachers can survey and select news dated back to the time of the crash, movie reviews, encyclopedia, etc., or the students can work on their own using search engines. Electronic glosses and online dictionaries can help the students deal with these difficult texts independently as a group or individually. Glossing, Nation (2001) claims, “allows texts to be used that may be too difficult for learners to read without glosses;…provides accurate meanings for words that might not be guessed correctly; and provides minimal interruption of the reading process.” (p. 175). Given the problem of the students’ low exposure to inputs and lack of learner autonomy at Chiang Mai University, the above suggestion may prove quite helpful.

            Secondly, since the negative attitude toward a glossed material might have contributed to a lack of concentration on the material, the teachers should try to explain, clarify, and colorize the materials. The Teachers can tell their students why glosses are provided and how they can best benefit using them. The elements in the glosses, if new to the students, must be introduced to them before their reading. Orientation to the materials, the software, and the teachers’ expectations is thus necessary. Also importantly, the teacher should plan using glosses to aid lessons so as to enhance the students’ motivation, for instance, by adopting fun quizzes, cooperative activities, role plays, projects, etc., in which anxiety and frustration is kept low. With increased motivation, the students will hopefully desire to read further, which then leads to more exposure to extensive inputs. Incidental vocabulary and comprehension can thus be enhanced.

            Research question 2: Do electronic and paper-based glosses have different effects on the learners’ reading comprehension?

Analysis and discussion     

            According to table 3, the Electronic gloss user scored much higher than the Paper-based gloss user (11 points vs. 3 points).  Note that the Electronic gloss user had received higher TOEFL. Some researchers suggest that better students can manipulate the information contained within the glosses better (e.g., Davis, et al., 1997; Groot, 2000). This might have been the case in this study because Participant 2 did mention her difficulty understanding the English definitions more often. Her poorer English background could also have influenced the result.

            The Electronic gloss user did mention the ease of use and the unobtrusive nature of the electronic glosses. These useful elements deserve teachers’ attention. The advantages of electronic glosses over the paper-based glosses in terms of the ease of use, faster access to word meanings, and the unobtrusive interface (e.g., Leffa, 1992; Hong, 1997) might have helped in generating better reading comprehension of the Electronic gloss user in this study.

Implication and application

            Two interesting implications are that the more knowledge of vocabulary the students possess, the better their reading comprehension and that electronic gloss is more effective for vocabulary learning than the other two treatments. Chiang Mai University is well equipped with computer infrastructure; therefore, the teachers can easily seek to benefit from it. Weaker students seem to need glossing help more and are likely to benefit from electronic glosses. Essentially, the teachers should use definitions according to the i+1 input theory to ensure that their (weaker) students can understand the content of the glosses and learn to become a more competent learner at the same time.  In another note, if glosses are to be provided on paper, the teachers should adopt the convenience, unobtrusiveness, and ease of use that electronic glosses possess. Glosses can be offered, for instance, “on the same page as the reading passage” (Participants 2 and 3).  Again, the teachers should make sure the gloss contents fit the level of most students so that they can truly benefit from it. In any case, a pilot study seeking students’ feedback on the glossed materials before use seems necessary.

Reading comprehension deals with a lot of learner factors. It is thus safe to say that the results and explanations about it need to be considered cautiously. This study did not look into each participant’s behavior and thinking at the time they dealt with the materials. The record each participant wrote while consulting the materials did not provide much information. Perhaps, a future study may either separate reading comprehension and vocabulary learning or study both but examine learner behavior more closely, i.e. by using log files or computer tracking programs.

            Research question 3: Does glossing help the learners have better vocabulary retention?

            Research question 4: Do electronic and paper-based glosses have different effects on vocabulary retention?

           
            The vocabulary pretest and posttest scores are provided in Table 4 to supply support for answers to the above two research questions.