The integrated skills tasks used in this study are items from the Pearson Test of English Academic (PETE Academic), comprising two item types: ‘Retell Lecture’ and ‘Repeat Sentence’. The Official Guide to PETE Academic (Pearson, 2010) states that the ‘Retell Lecture’ item is designed to assess candidates’ ability to “give a presentation on information from a lecture on an academic subject’ (2010, p. 51 while the ‘Repeat Sentence’ item tests candidates’ ability “to understand and remember a sentence, then repeat the sentence exactly as you hear it sing the correct pronunciation” (2010, p. 41).
Both of the item types assess speaking and listening skills in an integrated fashion. Research design The research question for this study is: can the new integrated skills test change candidates’ use of learning strategies and learning materials? Three classes of students from the School of English Education in Asian International Studies University in China participated in this study. One class (n=31 ) served as the control group, while the other two classes (n=58) served as the experimental group. All of the students had just completed their Test for English Major Band 4 (TEM-4) and a university based speaking test.
The Test The main differences beјen TEM-4 listening tasks, the university speaking test and the PETE Academic new integrated skills tasks include: 1) the audio input of PETE Academic has a wider range of accents and the background noise has not been removed with the aim of creating a more authentic environment; 2) the content of the listening materials in PETE Academic is not an audio record of news or of daily conversation topics, but is more academic; 3) the response format of PETE Academic does not consist of ultimate choice questions and pair discussion in front Of two examiners, but asks to orally summaries or repeat the listening materials; and 4) PETE Academic is conducted through a computer Page 2 of 11 system. The candidates in the experimental group were invited to complete PETE Academic and a survey twice, at the beginning and after a six week break.
Their performance in the two constructed integrated skills tests is worth 5% of their academic overall marks in year three. Each test included two ‘Repeat Sentence’ items and two ‘Retell Lecture’ items. The Survey The first survey served as the baseline study, aiming to keep a record of antedates’ previous tests and test preparation experiences. The same survey was administered after the participants completed the new integrated skills test in order to identify any possible changes. Apart from their demographic information, the survey consists of two parts: the first part deals with test takers’ perceptions of the specific skills being tested in TEM-4 (listening tasks), the university-based speaking test and PETE Academic as used in this study.
It consists of 36 statements and each statement is judged using a 5-point skirt scale (from strongly disagree to strongly agree). These statements mom from the skills listed under the two item types ‘Repeat Sentence’ and ‘Retell Lecture’ in The Official Guide to PETE Academic (Pearson, 2010, up. 41-45, 51-56). A higher overall rating of the listed skills therefore indicates a better understanding or awareness of what skills the two item types assess. The 50 statements in the second part of the survey are taken from the Strategy Inventory for Language Learning (SILL)Version 7. 0 devised by Oxford (1990). Each statement is judged using the same 5-point Liker scale.
At the category level, there are six types of strategies: memory, cognitive, insemination, anticompetitive, affective and social strategies. Two paired t- tests were conducted through PAWS (Predictive Analytics Software) Statistics 18. 0 to check 1) whether the candidates had become more aware of the skills being assessed by PETE Academic over the two tests (by comparing their ratings to first part of survey before and after taking the integrated skills tests); and 2) whether the candidates had changed their learning strategies at any category level (sum of their ratings for each strategy under each category) or at the individual skill level (by comparing their ratings to the second part of survey).
An independent t-test was also used to check whether the candidates in the experimental group responded significantly differently from their counterparts in the control group in the second survey. The Interviews After the two tests and the two surveys, the interviews were conducted, with the aim of gaining more insight into why the participants in the experimental group had changed their learning strategies. The semi-structured interviews were designed to provide greater insight as to how and why the new integrated skills test may change participants’ learning strategies and trials, if there is any change. The interviewees were selected based on their ratings in the second part of the survey.
The prompts for the interviews come from previous studies and include questions about language learners’ beliefs about language Page 3 of 1 1 development, affective states, language learning experiences, the tasks they are required to perform and their learning outcomes (Ellis, 2008). The data was coded by researchers independently with a view to investigating the roles played by these factors in the potential link between the new integrated skills tests and adapted learning strategies and materials. The prompts of the constructed interviews have been used as the primary source for coding the data. Findings from the survey The same 58 candidates (experimental group) participated in the study by completing both the surveys and PETE Academic. The control group consisted of 31 students, who completed the survey twice.
The analysis of the first survey indicated that there was no significant difference between experimental and control group in terms of their performances in the TEM-4 and university based speaking test as well as their responses to the survey at the category level. The Cockroach’s Alpha liability coefficients suggested a high overall reliability for the survey (r=. 94) and at each category Of the learning strategies (range from . 73 to . 84) in both groups. Moreover, the paired t-test suggested that the candidates in the control group had not changed their learning strategies significantly in any of the categories. This indicates that there is no other variable which had influenced students’ use of learning strategies in this study.