A Serial Ordering of Listening
Comprehension Strategies Used by
Advanced ESL Learners in Hong Kong

Recent theories of second language listening suggest that listeners are active processors of information rather than passive receivers of oral stimuli. Listeners construct meaning from the oral input by drawing upon their prior knowledge of the world and of the target language (Byrnes, 1984; Nagle & Sanders, 1986). They also generate information in their long term memory and make their own interpretation of the spoken texts (Murphy, 1985; Mendelsohn, 1994). Since listeners have limited memory capacity for the target language (Richards, 1983), they use different listening comprehension strategies, that is, steps taken by learners to help them acquire, store, retrieve, and/or use information (see O'Malley, Chamot, & Küpper, 1989; Vandergrift, 1992). Some teaching professionals believe that a better understanding of the use of these strategies enables them to develop materials which suit the needs of their learners. Carefully designed listening strategy instruction programs can enhance the performance of the learners and help promote learner autonomy (Mendelsohn, 1994; Chamot, 1995).

A few empirical studies have been conducted to uncover the listening strategies used by second/foreign language learners. Murphy (1985) worked with twelve intermediate ESL university students and concluded that the high achievers used their prior knowledge ("personalizing"), made guesses ("inferring"), and monitored their comprehension ("self-describing") more often than did the low achievers. Vandergrift (1992) found that the successful French-as-a-second-language learners monitored their own listening comprehension and identified aspects which hinder comprehension twice as frequently than their unsuccessful counterparts, but the differences were not statistically tested.

O'Malley, Chamot, and Küpper (1989) described a hierarchical order of listening comprehension strategies based on a three-stage language processing model (Anderson, 1985) and argued that the subjects selected their attention in the "perceptual processing" stage, inferred the meaning of the text in the "parsing" stage, and utilized their background knowledge in the "utilization" stage. Their conclusion, however, was not drawn from any thorough examination of the think-aloud protocols, and the systematicity of the strategy patterns was not tested.

Although previous studies have provided exhaustive lists of strategies, they do not include a hierarchical order of strategy use, which could have been a valuable source for the pedagogy (McDonough, 1996). Accordingly, this paper reports on a study which performed implicational scaling and protocol analysis on the strategy data. The combination of these two methods enhanced the precision in the description of the sequence of learners' strategy use in second language listening. The study focused on the following question: Is there a pattern underlying the strategy use of these Chinese undergraduates when they are listening to audio-texts selected from advanced ESL textbooks? This question merits investigation for two reasons: (1) the systematicity of the strategy checklists can be enhanced if there is a sequence of strategy use; (2) if such a sequence exists in second language listening, language professionals can predict the strategy use of their learners and tailor strategy instruction programs which pinpoint certain weaknesses of their students.

Subjects

The participants of this study were eighteen Chinese students enrolled in various courses at six universities in Hong Kong in the 1994/95 academic year. The criteria for sampling were that these subjects should (1) be willing to participate, (2) be enrolled in different universities, (3) be Chinese and brought up in Hong Kong, and (3) have an equal proportion of male and female students. These anonymous participants, eight males and ten females, were on average twenty-one years old and had been studying English as a second language for over fifteen years.

Instrumentation

background questionnaire. A background questionnaire was designed by the author to obtain information about the participants' age, gender, academic major, English Language examination results, previous experience in traveling to English-speaking countries and the situations in which they could listen to English.

Listening materials. Three recordings of commercial tapes for listening skills development were chosen according to five criteria: (a) the English used was natural in speed and tone; (b) the recordings were authentic or unscripted; (c) the maximum length of each recording was less than five minutes; (d) the topic of the recordings could arouse interest of the subjects; and (e) the recordings were suitable for advanced ESL learners. As reflected by the pilot subjects, these passages matched the criteria. "The Friend of the Earth" was an interview with the then director of Friends of the Earth (Harmer & Elsworth, 1989), who talked about the aims of the organization and its global campaigns for the protection of the rain forests. "Misunderstood" was a story of a baboon and a cock (Collie & Slater, 1993b), while "Teeth and Dentists" was an informal conversation of four people about their childhood memories of dental problems (Collie & Slater, 1993a).

Procedures

Data collection interviews were conducted individually in Hong Kong between October and December 1994. The time and place of each interview was decided by the participant and agreed on by the author. Noise was reduced to a minimum during the interviews. The investigator arranged the recording equipment on the subject's desk and monitored the recording processes. Each subject listened to the recordings through a pair of earphones from a portable cassette player and gave verbal report by talking into a microphone placed on a desk. The voices of the subject and the author and the oral text were synthesized by a microphone mixer and recorded by a portable cassette recorder. The data collection procedures, pilot-tested in June 1994, consisted of six sessions: (1) Briefing, (2) Warm-up, (3) Trial Listening, (4) Think-aloud, (5) Feedback, and (6) Background Questionnaire. Cantonese, the mother-tongue of the two parties, was used by the author in the interview.

In the Briefing session, the author restated the purposes of the study and reassured the anonymity of the subjects. The subjects were reminded that they were asked to "think aloud," that is, they would report whatever that came to their minds when performing a task. Each subject signed a consent form before they proceeded to the next session. In the Warm-up session, the subjects were given instructions to think aloud with three tasks selected from Ericsson and Simon (1993). For example, they were instructed to speak whatever came to their minds when they had heard of the mathematical task "36 times 24." The subjects did not explain their thoughts, but verbalized what they were thinking. In the Trial Listening session, the subjects thought aloud with a trial listening tape. They were instructed to give the investigator a signal (for example, by raising their fingers) when they thought of anything. The investigator would stop the tape and let them verbalize their thoughts.

In the think-aloud session, the subjects thought aloud with the three recordings described above. The topics of the tapes were not disclosed to the subjects, but were represented by the letters A, B, and C. As a means of counterbalancing, the subjects selected the order of listening at random. To minimize carryover effects, the subjects took short breaks between the tasks.

In the Feedback session, the subjects discussed with the investigator informally about the interview. At the end of the meeting, each subject completed the Background Questionnaire and was offered a small amount of money (HK$100) for traveling expenses. The average timing of each data collection interview was fifty-two minutes.

Transcription and Coding of the Think-Aloud Protocols

The audio-recordings of the subjects in the think-aloud session were transcribed verbatim by the author. As mentioned above, the subjects were free to use any languages in which they were thinking. Most of the subjects thought aloud in Cantonese, with some English remarks. The resulting protocols were coded for the identification of strategies using the pilot-tested Second Language Listening Comprehension Strategy Inventory (see Appendix). The following four-step procedure was used in coding the protocols. The author (1) read each think-aloud report carefully several times, (2) underlined the corresponding think-aloud verbalization, (3) categorized the strategy used in the excerpt, and (4) wrote the strategy code in the protocol.

Coding Reliability Checks

The consistency of the coding was checked by calculating the intercoder-and intracoder reliability coefficients. The target of the coding reliability checks was set at .80. Intercoder reliability coefficient was the agreement of the author and each of the two external coders over the coding of five random samples of protocols. Two external coders were invited to check intercoder reliability. One of them (expert coder) was a teacher trainer who was familiar with the language learning strategy literature, and the other (novice coder) was an experienced primary school teacher. In the individual coder instruction sessions, each coder received a copy of the coding scheme and two randomly selected samples of protocols. Following the four-step procedures of coding described above, the coders coded the protocols and discussed the results with the investigator. Any disagreement was resolved by negotiation. At the end of the session, the coders received five random samples of protocols to code independently. The investigator did not discuss the coding with them; two weeks after their first meeting, she collected the protocols.

Derived from Murphy (1985) and Scholfield (1994), the formula used to calculate the intercoder reliability coefficient was

where A, E, and N represented the author, the expert coder, and the novice coder respectively. The intercoder reliability coefficient was .80.

Intracoder reliability coefficient was the code-recode agreement of the author over five random samples of protocols. The intracoder reliability was checked by having the investigator recode the above mentioned random samples of protocols six weeks after the first coding. The investigator did not study the five sample protocols during that period, and the following formula was used to establish the intracoder reliability coefficient:

The intracoder reliability coefficient was shown to be .94.

Data Analysis

The strategy data was analyzed both quantitatively and qualitatively. In the quantitative analysis phase, implicational scaling technique was used to examine whether there was a trend underlying the subjects' choices of various listening comprehension strategies. Implicational scaling is a statistical procedure designed to order a series of items on a cumulative dimension. In the strategy literature, Chesterfield and Chesterfield (1985) have used implicational scales to show the natural order of children's acquisition of strategies over time.

The twenty types of listening comprehension strategies were coded for each listening task in the think-aloud session. The presence and absence of a strategy type in a subject's repertoire in each passage was represented by "1" and "0" respectively. To test whether there was a scale in each set of strategy data, computer software specially designed for implicational scaling analysis was used.

In the qualitative analysis phase, representative think-aloud protocols were analyzed for the identification of possible sequence in strategy use. This paper reports on the analysis of one protocol.

Results of the Implicational Scaling Analysis

The scaling for the twenty listening comprehension strategy types used by the eighteen subjects on each of the three passages provided fairly acceptable scales by most criteria. Among the three think-aloud listening tasks, "The Friend of the Earth" had the highest scalability (CR = .87, MMR = .22, CS = .84), followed by "Misunderstood" (CR = .83, MMR = .27, CS = .77) and "Teeth and Dentists" (CR = .81, MMR = .27, CS = .74). These results suggested that the subjects in the present study had used various types of strategies and that around eighty percent of the time their strategy choices were explicable.

Table 1 illustrates the scale of the strategy data in "The Friend of the Earth." The strategy items on the triangular pattern at the right end of the matrix were "expected" to be present in the subjects' repertoires, while those strategy items on the left of the trace line were "expected" to be absent in the subjects' repertoires. The bracketed items were those which did not fit the "ideal" pattern. In other words, these strategy types were present in the subjects' repertoires when they were "expected" to be absent, and vice versa.

Table 1 Implicational Scale of the Listening Comprehension Strategy Data

The scales reflected that those subjects who were on the upper-third of the matrices (S1, S3, S8, S11, S18) were males and females with different English language achievement. Those subjects who were on the lower-third of the matrices (S15, S16, S17, S20), however, were males and females who got a Grade D or E in their English Language examinations in the Hong Kong Certificate Education Examination (HKCEE), which is a local public examination taken by the students when they finish their secondary education. The results were consistent with previous studies (O'Malley, Chamot, & Küpper, 1989; Vandergrift, 1992), suggesting that the less successful students used a narrower range of strategies in second language listening comprehension.

Further analysis of the three matrices revealed that subjects who had greater repertoires of listening comprehension strategies constantly used the following six strategies: Self-evaluation, Summarization, Elaboration, Inferencing, Feedback, and Reprise. Given the position of a subject on the scale, one could "predict" that if a subject had used five strategies in a listening task, that subject must have used Reprise, Feedback, Repetition, Inferencing/Self-evaluation, and Summarization, but not Grouping, Resourcing, Deduction, Transfer, Note Taking, Hypothesis Testing, or Uptaking.

An ideal implicational scale for the learners' listening comprehension strategy repertoire should reflect that the presence of a strategy type on the left of the matrix implied the presence of another on the right of it. The strategy pattern of S11 in "The Friend of the Earth" can be shown to interpret the results of the implicational scaling analysis. S11 used seven types of strategies when listening to the passage. According to the position of S11 on the implicational scale, his use of Translation implied his use of Imagery, Feedback, Self-evaluation, Inferencing, Elaboration, and Summarization as well. Since the trace line was set between Translation and Self-monitoring, it was "expected" that those strategy types on the left of the strategy item Translation were absent in S11's strategy repertoire.

Unlike Chesterfield and Chesterfield (1985), who used the implicational scales to show an order of strategy acquisition, this study illustrated the popularity of strategies in listening comprehension. The technique, however, could not illustrate the sequence of strategy use. Hence, a qualitative analysis of the think-aloud protocols was conducted to examine the strategy patterns on the implicational scales.

Results and Discussion of the Protocol Analysis

The results of the qualitative analysis of a full protocol of one subject (S1) are reported in this discussion. At the time of the study, S1 was a twenty-year-old third year student in Professional Accountancy at The Chinese University of Hong Kong. She had a Grade A in her English Language examination in HKCEE. S1 used nine types of strategies, of which eight were on the scale. They were Self-monitoring, Translation, Imagery, Feedback, Self-evaluation, Inferencing, Elaboration, and Summarization. She did not use Planning on the ideal implicational scale, but used Reprise instead.

The following were the transcribed think-aloud reports of S1, together with the transcripts of the original recordings of "The Friend of the Earth," a conversation between a reporter and Mr. Porritt, the director of Friends of the Earth. Special symbols such as {S1A1} were used to show the point when S1 started to report her thoughts. Each verbal report was individually numbered ({S1A1}, {S1A2} and so on). The protocol presented below was translated by the author and checked by the back-translation technique. The italicized words were translated from original Cantonese, while the words in quotations were the English remarks of the subject.

Reporter:	So...Jonathan Porritt...you're, you're the director of Friends Of The Earth, is that right?
Porritt:	That's right. I have been for three years.
Reporter:	And what exactly is Friends Of The Earth?
Porritt:	It's a campaigning pressure group, dealing with {S1A1}
文章录入：whp    责任编辑：whp
	上一篇文章： 【访谈录】情人节的真正礼物 下一篇文章： 没有了
【发表评论】【加入收藏】【告诉好友】【打印此文】【关闭窗口】