Using Technology to Teach Vocabulary: The Role of Learners’ Motivation, Culture, and Teachers’ Perceptions

Language programs may have students from mostly one first language or, as is usually the case in North America, from several different Lis. With learners from different educational and linguistic backgrounds, finding the right set of words to teach for each LI or even each individual can be a challenge. Although the teachers in the Pitt IEP are well aware of the vital role of extensive reading in vocabulary learning, they felt that learners did not in fact engage in enough reading outside class. Although graded readers are valuable, they may not contain the vocabulary that will challenge learners once they begin their academic study.

The Pitt IEP faculty considered that one possible way around this problem was to adopt a personalized reading tutor that required learners to read and process words that had been selected from an academic word set. A tutor that fit this need was made available through a collaboration with the Pittsburgh Science of Learning Center, now http://learnlab.org. A team of computer scientists at Carnegie Mellon University developed a system, called REAP (www.lti.cs.cmu.edu/projects/language-technol-ogies-education/reap-reader-specific-lexical-practice-improved-reading), which was based on texts freely available on the world wide web (Collins-Thompson & Callan, 2004; Heilman, Collins-Thompson, Callan, & Eskenazi, 2006). The researchers used a search engine that found reading passages that satisfied very specific lexical constraints, namely AWL words in texts of LI English grades 6-8 level reading. (This level of reading was decided on because the intermediate textbook was shown to have about a US grade 7 level.) The web crawler netted a database of 50,000 documents (after filtering from 10 million), with each document about 1000 words. Because the selection was from an open corpus (the web), it satisfied a wide range of student interests and classroom needs. It had the potential to model an individual’s degree of acquisition and fluency

Lexical Development in an I EP 109 for each word in a constantly expanding lexicon so as to provide studentspecific practice and remediation.

The advantage of the tutor was that each student would receive a personalized list of words to learn. Before starting the reading course at the beginning of term, the students took a YES/NO test on a list of words from the AWL (Harrington & Carey, 2009). Students had to state if they knew a word or not. The words that formed the basis of the individual students’ lists for that term were derived from those words the student reported they definitely did not know. The web interface then generated a series of texts for each student that contained one to five key AWL words that the student reported not knowing. These words were highlighted in the text via hyperlink. When clicked on, the hyperlink directed the student to a pop-up dictionary entry that explained what the word meant in English and provided example sentences.

In 2007-2008, in an unpublished action research study, IEP faculty (led by Lois Wilson) carried out a quasi-experimental comparison of this automatic reading tutor with regular in-class activities.⁴ The procedure for the tutor was first a pre-test to establish an individual focus word list of 100 AWL words. Then the students read texts that were selected to provide examples in context of the words they didn’t know. While reading, they could choose to get help by clicking (or not) on words and looking at definitions. After the passage, there was a reading check question after reading the passage. Finally, there was a vocabulary check question. If a target word question was wrong, another passage with that word was presented right away; otherwise, the word was tested later. All words were seen at least five times: in the text, in the vocabulary questions, and in the reading check. At the end of term, post-test cloze and written output production tests were given.

In contrast, in the classroom activities only, students took a pre-test to establish an individual focus word list of 58 AWL words, none of which would appear in the REAP program. The words were chosen by the curriculum supervisor on the basis of her assessment of which words the students would know and would find useful. Students then completed interactive exercises in class, look-ups in class, and homework exercises, as part of their grade in the class. The in-class instruction was based on pair/group work, interaction, and written output practice. The final posttest was in cloze format, and there was also a final post-test of production. (We can assume based on the pre-tests that students did not know the words prior to the beginning of the instruction because the words were carefully screened.)

Over the two semesters in reading class, we compared the two approaches to vocabulary learning with the two sets of students, with a complete, combined data set for the two semesters of n = 50 for the cloze test and n = 49 for the production test (data for one student’s production were missing). Because the scales on the tests were somewhat different,

we converted the scores to percentage scores in order to make two comparisons: post-test scores on receptive vocabulary tests (cloze) and on the production tests (sentence writing) in both treatments. A comparison of the scores in Table 4.3 on the post-tests shows very different scores in accuracy on the words that were taught. It is important to recall, however, that with REAP, the learners never carried out writing as part of the learning process - the post-test writing was an assessment only.

These results indicate that the focused instruction that was teacher led provided learners with better learning for the words that were taught for both receptive and production skills. The complete data set of reading activities is available on request from the Learnlab data repository: https://pslcdatashop.web.cmu.edu/Project?id=26.

It would be premature, of course, to assume that reading only should be abandoned. Recall that beyond the basic 2000 words, learners need to learn a substantial proportion of words in the 3000-9000 frequency band range. At a conservative estimate, this means that they need to learn 7000 words in three semesters. It is impossible to teach in class such a large number of words - there is simply not enough time. This word learning task means that students absolutely have to devote time to extensive reading on their own in addition to in-class work. No one denies this important component of vocabulary reading. Teachers and programs can help students in many ways, some of which are discussed in this chapter.

First, students must be encouraged to read materials that interest them, an issue considered by Cobb (2016). One of the advantages of teacher-led practice is that the teacher keeps the focus of the students on task - this is true in a wide range of teaching contexts, for example, in Spanish as a second language, it was shown that teacher led practice was more

Table 4.3 Results of Paired i-Tests on % Scores of Receptive and Productive Skills

Lexical Development in an IEP 111 effective than student group work on difficult morpho-syntactic properties of Spanish verbs (Toth, 2008). Specifically, with the REAP tutor, however, Heilman et al. (2010) showed that when using the REAP system, allowing students to select texts that were in an area of interest to them lead to greater gains in vocabulary knowledge. The role of interest has been highlighted by Krashen (1989), though the Pitt IEP does not agree that simply reading is an adequate preparation for learning vocabulary for academic study. However, given that attention is important in learning and that interest will enhance attention, it is clear that Heilman et al.’s findings support the idea that students must read articles of interest to them because this allows them to maintain greater focused attention (Schmidt, 1990, 2001). In the IEP context, this means academically oriented texts that contain words that students will encounter later in their degree programs will be of special value as long as they are not too difficult. In addition, using a system that allows translations (in moderation) can also be helpful (Saz, Lin, & Eskenazi, 2015).

However, although such intelligent tutors and online reading tutors offer extraordinarily rich opportunities for focused study at the individual level, the IEP ultimately decided that this was not the best use of class time. This decision does not mean such tutors are not valuable. Indeed, based on our research, with specific guidance, the tools can be important in helping students develop their knowledge through extensive reading. The guidance that students can benefit from emerged from several mixed-methods classroom-based studies.

Pelletreau (2006) began our review of the use of the tool. He found that some students took off-line notes while using the tutor, whereas others did not, preferring to interact with the teacher. Moreover, the topic choice was important to some students. (This was an option introduced in the system later.) Most importantly, students who were invested in the tool did best in learning, doing better on the words they were asked to focus on and on non-target words that they themselves chose to focus on. Pelletreau (2006, pp. 50-51) also notes that ‘the number of target word lookups proved to be moderately correlated with non-target production task scores’ and ‘Provided students did not overdo the attention and the resources they devoted to non-target words, they could increase their target-word acquisition by taking a difficult vocabulary task and simplifying it with intermediate steps and a combination of bottom-up and top-down processing’.

Juffs and Friedline (2014) followed up on Pelletreau (2006). Using an activity theory framework, they confirmed that learners took an individual approach to the tool - and these approaches arose to some extent from the cultural background the students came from and their attitudes to reading as part of their vocabulary learning. Juffs and Friedline (2014) compared 41 Arabic-speaking learners with 21 Korean-speaking learners, whose cultures place different emphasis on oral interaction and reading.

Recall that the reading tutor permitted the students to click on words that were highlighted (=target words from each students’ list of words they had reported they did not know) and non-target words, which were any other word the student wanted to check the meaning of. The two groups transformed the reading tutor based on their approach to learning. The Arabic-speaking learners clicked at the same rate as Korean speakers on target words - 49 on average for the Arabic speakers and 53 on average for the Korean speakers - but clicked on far fewer non-target words (74 vs. 114). In addition, in the reading check, which was a simple way to see if the learners actually read the text and not a set of comprehension questions, the Arabic speakers performed at chance (54%), whereas the Korean speakers did much better (71%).

Based on qualitative research with the learners, it was clear that the Arabic-speaking learners rushed through the readings due in part to their belief that reading was not helpful in learning vocabulary. In contrast, the Korean-speaking learners took a great deal of time, clicking on many words that were perhaps not essential to the reading, because they believed that every word needed to be understood.⁵ The study concluded that learners need help in interacting with online tutors and that teachers’ assumptions about the use of computer software have to be checked and reinforced carefully during the period that the software is used. Overall, the students preferred that class time be used through interactions with the teacher.

Finally, Wojcik (2009) looked at the problem of AWL list instruction from the point of view of teachers and students in an ethnographic study in the IEP. In a very detailed analysis, Wojcik looked into the lEP’s instruction program of 50 AWL words during the three-month term. A pre-test and a post-test were used, requiring students to provide both synonyms and good example sentences. Core vocabulary words were focused on in all classes - not just reading class. Each week, a set of 5 words from the 50 chosen was selected for focused instruction and practice. Core words were written on the board and practiced at the beginning of every class. Other activities involved crossword puzzles, synonyms, dictation, fill-in-the-blank, and sentence writing, both in class and for homework. Teachers were encouraged to ‘spontaneously’ use the target words in classroom discourse, and students were noticed paying attention to the use of these words. The total amount of time across a level - reading, writing, speaking, listening, and grammar - amounted to about 2 hours per week. In this way, the IEP hoped to raise students’ awareness of word families, the relationship among different lemmas (a persistent problem [Friedline, 2011]), and key usage.

However, the intent of the IEP during this period did not seem to make it into the consciousness of the students. Based on student interviews, Wojcik discovered that that none of the students knew about the AWL or why those particular words being focused on had been chosen. In general, students seemed not to realize that the words had been chosen for future

Lexical Development in an I EP 113 use in their academic studies but seemed focused on ‘every day’ use and were much less interested in those technical words. Learners also seemed to think that receptive knowledge was all that they needed to know.

In interviews with the teachers, Wojcik reported that they were modest concerning their knowledge of how to teach vocabulary, although many had had training through webinars and brown bag talks. The teachers did feel that they needed more guidance, however, in coordinating the teaching in the various skills with skill-appropriate activities.

In the final measurement of vocabulary learning, Wojcik notes with some concern that more than 50% of the students were not able to accurately provide synonyms and sentences for 23 out 50 of the AWL words taught during the term. In spite of this, students improved their score out of 100 possible points by an average of 32 points.

Wojcik points out some issues surrounding the teaching of words. For example, she notes that teachers who try to discuss the differences in meaning between such words as aspect and feature can eat up a large amount of class time, and such usage might better be acquired through extensive reading. Indeed, meanings of abstract words in general are a challenge for bilinguals, as reported in the experimental literature (e.g., De Goot & Poor, 1997), a fact that is finally coming into estimates of lexical sophistication (Crossley & Kyle, 2018). Wojcik also noted that students seemed to have some knowledge already of some AWL words, for example, method, alternative, secure, regions, obvious, target. She noted that 42% of students showed knowledge of these words on the pretest. This is not surprising given that obvious and target are in the 1000 most frequent words on the BNC list. Wojcik (2009, p. 95) concludes that rich instruction and exposure to five words per week does not necessarily result in deep knowledge of those words. While it is true that not every word was thoroughly learned, in fact students did improve by 32 out of 100 points on average. In addition, she notes that students are gaining important awareness of vocabulary learning, which would help future learning on an individual level when reading.

Wojcik’s work highlights some important issues for lEPs. First, teachers and program administrators need to communicate better with students regarding the reasons for instructional choices. This communication has the potential to raise student motivation in certain tasks and create better attention to these tasks. Second, instruction of vocabulary needs to be focused; teachers especially need to beware of eating up valuable class time with explanations of subtle differences among word meanings.