THE EFFECT of INERMEDIATE EFL LEARNERS’ LEXICAL INFERENCE from KNOWLEDGE SOURCES on LISTENING COMPREHENSION IMPROVEMENT
Author: shaghayegh Aein
Author: shaghayegh AeinAbstract This study adopted both quantitative and qualitative approaches to investigate the effect of Intermediate EFL learners lexical inference from knowledge sources on listening comprehension improvement. The participants were 63 Intermediate EFL learners homogenized by a Preliminary English Test and divided randomly into two groups of experimental and control, 33 and 30 participants in each group respectively. The researcher made use of 30 participants called the survey group as a standard source to deduce the pack of knowledge sources by means of think aloud protocols. The effects of knowledge sources were examined on experimental group, using podcast monologue audio short stories. The results indicated that practicing knowledge sources had a significant effect on the listening comprehension improvement of EFL learners at intermediate level. They also showed that majority of learners who employed knowledge sources had positive attitudes about the use of such kinds of sources on their listening comprehension improvement. Introduction According to Soria(2001), a necessity for learning a second or foreign language is that the learners should be exposed to the language, in writing and/or in speech. Such language experience or input may either be comprehensible or incomprehensible to the learners, so to figure out the incomprehensible ones, and for the reason of improbability of knowing every single one of the words in a language, learners have to make use of some compensation strategies to tackle the problem. Otherwise, the real intention of the speaker or writer may be understood wrongly or even communication may break down completely. ( Babaei & Riazi, 2005). While in production learners use communication strategies to compensate for the absence of words for which they have meanings, in comprehension inferential strategies are needed to compensate for the absence of meanings attached to unknown words.(Haastrup, 1991, p. 21, Soria, 2001, Babaee & Riazi, 2005). Communication strategies generally have been defined as devices employed by L2 learners when they encounter problems in L2 communication as their communicative ends have outrun their communicative means (Corder, 1983; Faerch Kasper, 1983; Paribakht, 1985; Soria,2001). According to Tricia Hedge(2008), when learners use gesture, mime, synonyms, paraphrase, and cognate words from their first language to make themselves understood and maintain a conversation, despite the gaps in their knowledge of the second language, they are using communication strategies. Inference is such a cognitive strategy which has been defined as the process of arriving at a hypothesis, idea or judgment on the basis of other knowledge, ideas or judgments (Richards, Platt & Platt, 1999; Soria, 2001). As evident from the definition, to generate inference, one has to rely on certain knowledge sources ( Babaei & Riazi 2005). Wesche and Paribakht (1999) argue that much if not most lexical development in both L1 and L2 appears to occur as learners attempt to comprehend new words they hear or read in context. A lot of researchers have set forth different sources of knowledge one may resort to in constructing lexical inference. Haastrup (1991 cited in Soria, 2001, Babaei & Riazi 2005) suggested three sources, namely, contextual, intralingual, and interlingual, that readers may use in making lexical inference. Contextual clues refer to one or two words from the immediate co-text of the new word, the whole sentence containing the new word, a specific part of the co-text beyond the sentence containing the new word or unspecified part of the text which may help the global understanding of the whole text. Intralingual clues pertain to the features of the new word and the inference maker's reliance on his/her information about phonology, orthography, morphology (the meaning of stem, prefix, and suffix), word class and collocations to guess the meaning. Finally, interlingual sources have to do with a language other than the second language, e.g. readers' first language. Using this clue, inference makers rely on their knowledge of mother tongue or another language to guess the meaning of a word in their second language. As far as L1 is concerned, one can rely on phonological, orthographical, morphological, lexical and collocation similarities between L1 and L2. The purpose of the present study was to figure out how practicing knowledge sources had significant effects on EFL learners' improvement according to their lexical inference on listening comprehension and how learners who employed knowledge sources thought about the use of such kinds of sources on their listening comprehension improvement. Research Questions To fulfill the purpose of this study which was to see whether exercising knowledge sources has any effect on EFL learners listening comprehension the following research questions were raised: Q1: Does practicing lexical inference from knowledge sources have any significant effect(s) on listening comprehension improvement of intermediate EFL learners? Q2: What are the learners attitudes who employed pack of knowledge sources about the impact of such kinds of sources on their listening comprehension improvement? Method Participants A group of 96 Persian-speaker adult learners of English as a foreign language at intermediate level who were selected and they were given a homogeneity test which was a Preliminary English Test (PET) as a pretest. Total number of the participants who met the necessary condition based on the level of language proficiency were 67 participants whose scores fell below and above the mean on the normal distribution curve, and they were divided randomly to 34 participants in experimental and 33 participants in the control group, however, at the end due to subject mortality, 33 subjects remained in the experimental and 30 subjects remained in the control group. Also the researcher made use of the other group called the survey one composed of 30 participants from which she deduced the knowledge sources for examining their effect on the experimental & controlling the variable of knowledge sources on the control group for comparing the result with that of the experimental one. It has to be mentioned that the subjects in experimental and control group took part in the study during a semesters. Moreover, the subjects were all females for controlling the sex factor. Instrumentation The instruments used in this research study fall into five main categories; tests & rating scales, instructional tasks, think-aloud method, CALLA model and retrospective interviews. Preliminary English Test(PET) A Preliminary English Test (PET) with both multiple choice and open-ended item types was used to assure the homogeneity of each group of participants. PET consisted of three main sections; reading and listening in the first part, writing in the second part and speaking in the third. The first section included 35 multiple-choice reading comprehension questions for the reading passages, 45 multiple choice listening comprehension items, the second section consisted of 4 fill-in-the-blank items plus 6 letters writing for the writing section. The third section consisted of an interview which was divided to four parts. Each phase fulfilled specific functions. In part one, the subjects answered general questions about themselves. In part two, they had to exchange ideas and in part three the interviewer asked the students to talk about something/ picture for instance and talk about questions by help of the picture. In part four, the interviewer and the students engaged in a discussion of more abstract issues; their likes and dislikes, talk about their favorite sport or movie for instance. The test was administered in 2 sessions; reading and listening was administered in one session with one hour and forty five minutes allocated time, and the writing and speaking section was administered in another session with one hour and thirty minutes allocated time on writing and ten to twelve minutes on speaking. This test was used as a pretest for homogenizing the sample on part of their proficiency level, based on which only those subjects were selected for final data analysis whose scores fell one standard deviation above and below the mean. The test was piloted first in a way that it was administered to another group consisting 50 participants to make sure that it was a reliable test for the researchers original subjects. Listening Achievement Test To test the null-hypothesis of the study, a teacher-made test as an achievement test composed of fifty multiple choice test was used as a posttest on Thumbelina short story to measure their ability on listening. As it was a mock test, the test was piloted first, to the group consisted of 50 participants before running it to both experimental and control group to assure its reliability & validity. The audio text was selected from the site www.storynory.com Rating Scale for Rating the Retrospective Interviews In order to be objective in rating the interview, a rating scale was used by two raters. This rating scale was divided to three parts of be agreed, disagreed and undecided of the participants and received a rate ranging from 1 to 3. Podcast Audio Short Story Task A number of five podcast audio short stories texts were administered as a task to the experimental group for helping them learning knowledge sources. The genre of these audios was narrative just like previous studies. They were chosen among audio texts for the nature of the study and they were all selected at www.candlelightstories.com & storynory.com. Concurrent Think-aloud Concurrent think-aloud was supportive to deduce the knowledge sources in the survey group because it is said that if individuals' performance and verbalization do not take place simultaneously, it is probable that some part of the information is missing while individuals are producing verbal reports. Moreover, the information may be filtered in the sense that the participants may well report what they think the researcher is after and ignore what they attended to at the time of accomplishing the task (Green, 1998, Soria, 2001, Babaee & Riazi, 2005). Pair Think-aloud According to Soria (2001), Babaei & Riazi (2005), pair think-aloud is considered ideal because one can find out how learners infer on their own accord rather than if they were seated singly in front of a tape recorder and asked to verbalize alone. The interaction and negotiation that invariably occur when two people discuss enriches the wealth of information that think aloud method (TAM) offers. Retrospective Interviews According to Soria (2001), Babaei & Riazi (2005), retrospective interviwes were employed for increasing the reliability of using knowledge sources, in a way that, for identifying the difference between knowing the knowledge sources and not knowing them, an interview composed of twenty questions with was utilized for both groups of the study. One more rater assisted the researcher in the interview. Procedure In order to be able to test the null hypothesis of this research study, several steps were taken successively. Preliminary English Test In the first step, 96 female subjects were selected from the intermediate level and the standardized proficiency test (PET) was administered to those 96 selected subjects. In order to satisfy and ensure the assumption of homogeneity of the sample group, only 67 subjects whose scores fell within the range of one standard deviation above and below the mean were selected and randomly divided into the two control and experimental groups. However, due to subject mortality later on, 33 subjects remained in experimental and 30 subjects remained in the control group for the post test. Rating Scales In the second step, speaking and writing sections of pre-test was administered by two raters. All the oral productions were recorded on an MP3 player to be subsequently marked by two raters. Then, the next step was training the other rater who was the researcher's friend with an MA degree in language teaching and sufficient teaching experience. The researcher and the other rater did couple of sample rating together to ensure the same understanding of the two scales. After that, they rated the participants' oral and written performances. Transcription of Think-aloud Protocols In the third step, the participants verbalized their thoughts in their first language or the target one up to them and transcription and translation were done on the part of the researcher. Table1.Transcription Conventions (Adapted from He, 1998, p. 113, Babaee & Riazi 2008) CAPS [ = (( )) > 73.62 ). Therefore, three subjects who had high scores in the control group and one subject who had low score in the experimental group were excluded from the study. Hence, the researcher had to check the normality of the distribution again to be able to run the t-test to make sure that excluding the subjects from the control and experimental groups made the two groups become homogenous. For this purpose, the researcher computed the Skewness of the scores and divided it by the standard error of skewness. Table.4. Checking Skewness of the scores after excluding the subjects. Statistics Experimental Statistics Control Mean 52.545 Std.Error of Mean 1.2491 Std. Deviation 7.1757 Skewnwss -.350 Std. Error of Skewness .409 Mean 46.550 Std.Error of Mean 1.2409 Std. Deviation 6.7967 Skewnwss .430 Std. Error of Skewness .427 The gained values for experimental and control groups were 0.51 and 0.2, respectively. Because the values were smaller than 1.96, it was concluded that the samples were normally distributed at each groups. Therefore, the researcher was able to run a t- test. Table.5. Independent t-test after excluding the subjects for PET. Group Statistics CODE N Mean Std. Deviation Std. Error Mean Group 1 2 33 30 77.94 76.00 3.445 2.421 .600 .442 Independent Sample Test Levens Test for Equality of variences t-test for Equality of Means F Sig t df Sig. (2-tailed) Mean Difference Std. Error Difference 95% Confidence Interval of the Difference Group Equal variances assumed Equal variances not assumed 1.359 .248 2.561 2.603 61 57.4 .013 .012 1.94 1.94 .757 .745 Lower Upper .425 .448 3.454 3.431 The t-test results indicated that there was no significant difference between the mean score of control and experimental groups with respect to their overall proficiency after excluding the subjects (t= 2.5, df = 61, p= 0.01< 0.05) ( i.e., mean of control group= 76, mean of experimental group= 77.94). Therefore, the experimental and control group were proved to be homogeneous with respect to their overall proficiency and, thus, the researcher could start the treatment phase. The Skewness of the Scores of the Posttest Table 6. Checking the Skewness of the Scores of the Posttest Statistics Experimental Statistics Control Mean 30.67 Std.Error of Mean 1.707 Std. Deviation 9.806 Skewnwss .493 Std. Error of Skewness .409 Mean 12.60 Std.Error of Mean .756 Std. Deviation 4.141 Skewnwss .059 Std. Error of Skewness .427 The gained values for the experimental group came out to be 0.82 which was lower than 1.96,showing that the sample in the experimental group was normally distributed. For the control group, however, the value came out to be 7.2 which was higher than 1.96, showing that the sample in the control group was not normally distributed. Therefore, since the scores of one of the groups were not distributed normally running t-test was not legitimated and the Mann- Whitney was used. Table 7. Mann-Whitney for Posttest of the Experimental and Control Group. Ranks GROUP N Mean Rank Sum of Ranks TOTAL experimental control Total 33 30 63 47.00 15.50 1551.00 465.00 Test Statistics TOTAL Mann- Whitney Wilcox on W Z Sig.(2-tailed) .000 .65.000 -6.824 .000 According to table, the results of the Mann-whitney test indicated that at the 0.05 level of significance there was a significant difference between the mean rank of the control group and that of the experimental group on the listening post-test (U= 0, N1= 33, N2= 30, P=0 < 0.05). In other words, the listening proficiency of the control and experimental groups differed significantly at the end of the treatment period, (i.e., mean rank of control group= 15.5 and mean rank of experimental group= 47) which showed that experimental group out-performed the control group. After the study was done, in order to make the nature of the research more qualitative, the researcher used an interview using for the experimental group to know their attitudes towards using knowledge sources, mentioning the fact that inter-rater consistency was calculated as well. Table8.The Inter-rater consistency of the experimental oral interview Correlations Rater 1 Rater 2 Rater 1 Pearson correlation sig(2-tailed) 1 . 33 .855 .000 33 Rater 2 Pearson correlation sig (2-tailed) .855 .000 33 1 . 33 As the table shows, correlation between the two ratings came out t be statistically significant at 0.01 level of significance (r= 0.85); therefore, the researcher legitimately computed the mean rating between the two raters to be considered as the final score of the subjects' opinions in the interview. The researcher made use of percentile rank to make sure that what percentages of learners in the experimental group were agreed, undecided or disagreed with practicing knowledge sources for listening comprehension improvement. The data was tabulated herein. Table 9. Percentile rank for the experimental group in using the knowledge sources statistics Rater1 Rater2 Mean Percentiles 100 1.67 2.00 1.48 2.00 Rater 1 Frequency Percent Valid Percent Cumulative Percent Disagreed Undecided Agreed Total 3 5 25 33 9.1 15.2 75.8 100.0 9.1 15.2 75.8 100.0 9.1 24.2 100.0 Rater 2 Frequency Percent Valid Percent Cumulative Percent Disagreed Undecided Agreed Total 5 7 21 33 15.2 21.2 63.6 100.0 15.2 21.2 63.6 100.0 15.2 36.4 100.0 To test the second research question and make the study more qualitative, retrospective interviews were run and percentages of those learners who were disagree, undecided and agreed for using the knowledge sources with the help of a rating scale ranging 0 to 2 were calculated. According to the table, as the percentile came out to be 2 , it showed that majority of learners in the experimental group had positive attitudes towards using knowledge sources. Discussions and Conclusions The result of this study showed that there was a statistically significant between experimental and control groups' listening comprehension improvement at intermediate level by practicing the knowledge sources. Also majority of learners in the experimental group had positive attitudes towards using knowledge sources for their listening comprehension improvement. The results of this study also adds another piece of evidence to the body of literature on the importance of teaching knowledge sources in foreign language learning, this time with the specific focus on one of the most important and vital language skills, i.e. listening. Moreover, it can be asserted that the model used to teach the knowledge sources was a practical and useful one. Pedagogical Implications The findings of the present study have implications for learners, teachers, and teacher educators in the realm of TEFL in particular and education in general. Where learners of EFL have been educated in a more teacher-centered, top-down curriculum rather than one that promotes learners autonomy and independence, knowledge sources instruction could prove most effective. The students who participated in this study reacted positively to the pair think-aloud procedure. All participants reported learning something from the experience. Learner familiarity with the theme and topic of the text was an important source of clues for inferring the meanings of unknown words. Most students enjoyed the theme of the stories. The majority of the students gave favorable feedback on the texts that were used for this study. Thus, the text type and theme evidently influenced learners in terms of both their motivation and their success in lexical inferencing. It should be noted that unlike most of studies as subjects were treated as objects by ignoring their agency, the data presented in this study have provided evidence in support of a principle of a theory of activity and learners were real subjects and real agents in this study as they engaged to complete the task at hand and even their attempts to reframe the task can be seen as the dynamic nature of the activity. This study has argued for an awareness of subjects as individual agents involved in shaping their activity by speaking out their minds. It is in the same boat like what Lantolf (2009), mentioned that the study of human beings must focus on conscious individuals as they act upon themselves and their environment in a dynamic manner to show the vibrant of the activity. References Babaee, N, & Riazi A, (2008) .Iranian EFL female students' lexical inferencing and its Relationship to their L2 proficiency. Reading Matrix, 8,186-196. Cain, K., Lemmon, K., & Oakhill, L. (2004). Individual differences in the inference of wordmeaning from context: The influence of reading comprehension, vocabulary knowledge, and memory capacity. Educational Psychology, 96, 671-681. Dubin, F., & Olshtain, E. (1993). Predicting word meanings from contextual clues: Evidence from L1 readers. In T. Huckin, M. Haynes & J. Coady (Eds.) Second Language Reading and Vocabulary Learning (pp. 181-202). Norwood, NJ: Ablex Publishing Corporation,p.3-13. Garnham, A. (1992). Minimalism versus constructionism: A false dichotomy in theories of inference during reading. Psycholoquy [Online serial], 92 (3). Retrieved on December 29, 2003, from http://www.ai.univie.ac.at/archives/Psycoloquy/1992.V3/0062.html Haastrup, K. (1990). Developing learners procedural knowledge in comprehension. In R. Phillipson, E. Kellerman, L. Selinker, M. Sharwood Smith, & M. Swain (Eds.), Foreign/second language pedagogy research (pp. 120-133). Clevedon, UK: Multilingual Matters. Hudson, T. (1998). Theoretical perspectives in reading. Annual Review of Applied Linguistics, 18, 43-60. Lantolf .P .James(2009). Socio-cultural theory and second language learning. China: Oxford University Press, 3, 79-95. Nassaji, H. (2004). The relationship between depth of vocabulary knowledge and L2 learners' lexical inferencing strategy use and success. Canadian Modern Language Review, 61(1), 107-134. Paribakht, T. S., & Wesche, M. (1999). Reading and incidental L2 vocabulary acquisition: An introspective study of lexical inferencing. Studies in Second Language Acquisition, 21, 195-218. Soria, J. (2001). A study of Ilokano learners lexical inferencing procedures through thinkaloud. Second Language Studies, 19 (2), 77-110. Thomas, Barbara., & Ashton Sharon. (2006). Preliminary English Test (PET). London: Longman Group Uk Limited, 2, 94-126. Yamini,M, & Rahimi,M. (2007). A guide to statistics and SPSS for research in TEFL, linguistics and related disciplines. Shiraz: Koshamehr, 60, 111-150.About the Author: