In recent years, factors affecting listening comprehension in second language settings have been discussed by many researchers. One of the important variables that affect comprehension is phoneme perception. A few researchers have tested phoneme identification training for foreign language learners to improve their English listening performance (e.g., Logan et al.,1991; Lively et al., 1994). Although these studies revealed the crucial role of the phoneme in listening at the input level, further investigation is needed to fully understand the mechanisms of English listening comprehension in foreign language learning.
To understand the information processing mechanisms in listening, the unique characteristics of listening comprehension cannot be ignored. In listening, auditory information flows almost continuously and listeners have to deal with serial and evanescent acoustic signals and process them in real time. The current study examined whether information serial processing skill is the key factor differentiating high performers and low performers in EFL listening.
Researchers de Bot, Paribakht and Wesche (1997) adapted Levelt’s L1 speech production model and proposed a lexical comprehension and production model in L2. In the model, spoken or written signals were processed through a shared route to access lexemes, and lemma then reach the concept. Hirai (1999) also reported that the optimal listening rate and reading rate are similar. These studies suggest that the information processing mechanism in listening and reading comprehension are quite similar. Therefore, the present study presented visual stimuli to evaluate participants’ information serial processing skills.
The study hypothesized that high performers in listening would have strong information serial processing skills and would be able to perform well in serial text presentation tasks. On the other hand, low performers in listening were presumed not to have developed information serial processing skills yet, and would not able to perform well in serial text presentation tasks, but perform well in non-serial presentation tasks such as normal reading tasks.
The subjects for this experiment were 21 native Japanese students (average age 21.1). All stimuli were presented by the stimulus presentation software "Super Lab 2.0.” Two conversational English skits were used in each task, and a total of 10 skits were shown to a participant.
The experiment consisted of five tasks. In the Reading Task, an English skit was presented and participants were asked to read the passage and summarize it in Japanese. In the Listening Task, an English skit was played once on a CD player and participants were asked to listen to the skit and summarize it in Japanese. In the Self-Paced Task, participants were asked to press the “space” button of a desktop personal computer, and an English skit was displayed word by word. Participants were asked to summarize it in Japanese after they finished reading it. In the Slow Paced task, an English skit was displayed word by word. Each word was presented for 472ms. After all words in the skit were presented, participants were asked to summarize it in Japanese. In the Fast Paced task, an English skit was displayed word by word. Each word was shown for 363ms and after all words in the skit were displayed, participants were asked to summarize the skit in Japanese. The summaries written by participants were graded by three language teachers from 0 (incorrect) to 5 (correct) on the Likert scale. Each task consisted of two passages, and the total points for each task was therefore 10.
The participants were divided by mean score of the listening task, and two groups, “High performers of listening (High)” and “Low performers of listening (Low)”, were formed. Statistical comparison was made between these two listening groups in the Reading Task, Self-paced Task, Slow-paced Task, and Fast-paced Task.
In the Reading Task, in which non-serial information processing was allowed and the participants could read the text in a back and forth manner, both the High and the Low group performed well. However, in the Self-Paced, Slow and Fast Tasks, the Low group showed lower performance than the High group. Therefore, when serial processing was not required, the two groups understood the information at same level, but in the serial processing requiring tasks, the Low group performed worse than the High group. On the other hand, the High group showed a high performance in all tasks. This results indicates that serial information processing skill is a key factor in differentiating high performance and low performance in the listening skills of Japanese EFL learners.