The information of the participants will be presented in section 3.1

CHAPTER THREE RESEARCH DESIGN

This chapter reports the research design of the present study. The information of the participants will be presented in section 3.1. The instruments employed including a listening comprehension test, a questionnaire, and a semi-structure interview will be introduced in section 3.2. The procedures with the quantitative data and qualitative collection along with the coding steps will be described in section 3.3. The data analysis will be reported in section 3.4. Finally, a summary of this chapter will be given in section 3.5.

3.1 Participants

For both the quantitative and qualitative nature of the research design, the present study involved two steps of subject recruitment (i.e., a general group recruited for the quantitative data collection and a focused group for the qualitative study). The general group consisted of three different subgroups rendered for the need of research analysis and the focused group was selected from one subgroup of the general group for the qualitative study. The details of the general and focused groups are introduced in the following sections respectively.

3.1.1 General Group

Approximately, there were 1,772 students recruited to participate as the general group in the present study. The participants were the non-English-majoring freshmen of National Taiwan Normal University in the Fall semester, 2006. These participants were selected from the various departments which offered a bachelor’s degree in six colleges: the College of Education, College of Fine and Applied Arts, College of

Liberal Arts, College of Technology, College of Sports and Recreation, and College of Science. They were asked to fill in a questionnaire scale which examines their listening comprehension anxiety, personality, and linguistic intelligence levels. The number of the participants in each college is summarized in Table 3-1.

Table 3-1: Number of the Participants in Each College

College Number of Participants

Education 482 Fine and Applied Art 137

Liberal Arts 310

Technology 286 Sports and Recreation 121

Science 436 Total 1,772

However, for the sake of analysis, the participants with missing information such as incomplete basic information and empty questionnaire items were excluded from the subject pool. In addition, to avoid the potential bias problem, those who always chose the “Neither Agree Nor Disagree” option in the questionnaire were also excluded.¹ Therefore, there were only 1,400 participants finally selected to be the general group. Further, in order to answer the research questions, the general group was arranged into three subgroups according to their levels of listening comprehension anxiety, personality and linguistic intelligence assessed by the questionnaire scale in this study. The three subgroups are introduced below.

Anxiety Groups

Among all the participants, only the participants whose listening comprehension anxiety levels at the top 25% and the bottom 25% were selected as the two anxiety groups: the HI-ANX group and the LOW-ANX group for the present study. This led to 350 participants for each anxiety group. The criteria of selecting the two anxiety groups basically followed the study of Phillip (1992), Ganschow et al. (1994), and Cheng (2007). Table 3-2 presents the number of the participants of the two anxiety groups.

Table 3-2: Number of the Participants of the Two Anxiety Groups

Group HI-ANX LOW-ANX Total

Number 350 350 700

Personality Groups

To fit in the need of statistical analyses, all the participants (i.e., the general group) were equally divided into four personality (PER) groups (PER-1, PER-2, PER-3 and PER-4) based on their levels of extraversion. In other words, the participants of PER-1 were the learners who possessed the lowest degree of extraversion; the participants of PER-4 were the most extroverted learners. Hence, each personality group consisted of 350 participants as presented in Table 3-3 below.

Table 3-3: Number of the Participant of the Four Personality Groups

Group PER-1 PER-2 PER-3 PER-4 Total

Number 350 350 350 350 1,400

Linguistic Intelligence Groups

In addition to the previous two grouping methods, all the participants were also equally categorized into four linguistic intelligence (LI) groups (LI-1, LI-2, LI-3 and LI-4) based on their linguistic intelligence levels. Namely, the participants of the LI-1 group were the learners whose linguistic intelligence levels were at the bottom 25% of the scale of linguistic intelligence; the participants of the LI-4 group, at the top 25%

of the scale. Hence, each linguistic intelligence group consisted of 350 participants.

Table 3-4 shows the number of the participants in each group.

Table 3-4: Number of the Participants of the Four Linguistic Intelligence Groups Group LI-1 LI-2 LI-3 LI-4 Total

Number 350 350 350 350 1,400

3.1.2 Focused Group

After identifying the two anxiety groups (i.e., the HI-ANX and the LOW-ANX group) from the general group, ten participants were randomly selected from each anxiety group to comprise the focused group with a total of twenty participants. The focused group was then interviewed by the researcher to contribute for the qualitative data of the study. The number of the participants of the focused group is shown in Table 3-5 below.

Table 3-5: Number of the Participants of the Focused Group

Group HI-ANX LOW-ANX Total

Number 10 10 20

3.2 Instruments

Three instruments were employed in the present study. They were (1) a listening comprehension placement test, (2) a questionnaire, and (3) a semi-structured interview.

The listening comprehension placement test and the questionnaire were designed to collect the quantitative data, and the semi-structured interview was conducted for the qualitative analysis.

3.2.1 Listening Comprehension Placement Test

An English placement test was administered under the university administration policy to all the freshmen at the beginning of the Fall semester in 2006. This test consisted of two sections a Reading Comprehension section and a Listening Comprehension section. The students would be placed into four proficiency levels (i.e., Advanced, High-Intermediate, Low-Intermediate and Basic) based on their performance on the placement test. The placement test has been constructed under the project of Construction and Interpretation of College English Language Curriculum at National Taiwan Normal University since 2001 (Shih & Chang, 2005). Therefore, the reliability and validity of the test were ensured.

To answer the first research question, the present study used the Listening Comprehension section taken from the placement test. By taking the test, the participants’ listening comprehension anxiety was aroused before they completed the listening comprehension anxiety questionnaire. In addition, the test scores as the measurement of the participants’ listening comprehension performance served as a means to compare to their listening comprehension anxiety level.

This Listening Comprehension Test was in a multiple choice format that consisted of three parts, each of which had 15 questions including three different text types (with a total of 45 questions). In Part I, the participants needed to listen to

questions/statements; in Part II, to conversations; in Part III, to short monologues.

More information of question distributions and their corresponding text types are shown in Table 3-6 below.

Table 3-6: Question Text Types and Number of the Questions in the Listening Comprehension Test

Part Question Text Type Number of Questions Question Number

I Question/Statement 15 51-65

II Conversation 15 66-80

III Short Monologue 15 81-95

In Part I, the participants heard 15 questions and each question only contained one question or statement. After listening to the question, they had to read four choices provided and chose the best response to the question. An example of this part is shown in Table 3-7:

Table 3-7: An Example of Part I: Listening to Questions/Statements You will hear: Ms. Lewis has worked here for a long time, hasn’t she?

You will read: A. Yes, at three o’clock.

B. No. I don’t have a watch.

C. Yes, more than twenty years.

D. No. She’s working overtime.

The best answer to the question “Ms. Lewis has worked here for a long time, hasn’t she?” is C: “Yes, more than twenty years.” Therefore, you should choose answer C.

For Part II, the participants were asked to hear 15 short conversations. After each conversation, a question about the conversation was asked. The participants then needed to choose the best answer to the question. An example of this part is shown in Table 3-8:

Table 3-8: An Example of Part II: Listening to Conversations You will hear: (Woman) Hey, I think we should make the frame first.

(Man) Right! I have some strips of light metal and some wood.

Which do you think would be better?

(Woman) Wood is better. Metal is dangerous to use because it might attract lightning. We can use either paper or cloth to cover the frame.

Question: What are they going to use to make the frame?

You will read: A. Wood.

B. Metal.

C. Cloth.

D. Leather.

The best answer to the question “What are they going to use to make the frame?” is A: “Wood.” Therefore, you should choose answer A.

In Part III, there were several short talks that the participants listened to. After each talk, two to three questions about the talk were asked. After listening to each question, the participants needed to choose the best answer to the question. Table 3-9 is an example of this part:

Table 3-9: An Example of Part III: Listening to Monologues

You will hear: Welcome, shoppers. Do you buy expensive fresh fruit juice because it’s too much trouble to make it yourself at home? Well, there’s a machine that will save you money. It’s called the Handy Juicer. Let me show you how it works. First, slice your fruit into pieces like this. Then press this button to start the machine and drop the pieces into this opening at the top. Already, you can see the juice flowing out of the spout into my glass. With this Handy Juicer, you can also combine several kinds of fruit to make your own juice creations. And cleanup is quick. As you can see, everything comes apart to make washing easy. Buy this Handy Juicer and start saving money on juice today!

Question: Where is this talk most probably being given?

You will read: A. On the radio.

B. At a restaurant.

C. During a cooking class.

D. In a department store.

The best answer to the question “Where is this talk most probably given?” is D. “In a department store.” Therefore, you should choose answer D.

The participants were given about ninety minutes for the entire placement test, including forty-five minutes for the Reading Comprehension section and thirty-five minutes for the Listening Comprehension section. After completing the Reading Comprehension section, the participants went directly to listen to the tape or CD and completed all the questions in the Listening Comprehension section at the same time.

The tape or CD was played only once. For the last ten minutes, the participants needed to fill in the questionnaire about their listening comprehension anxiety, personality, and linguistic intelligence.

3.2.2. Questionnaire

To addresses the first and second research questions, a questionnaire (see Appendix B) about learners’ listening comprehension anxiety and its possible influential factors, personality traits and linguistic intelligence was employed. The questionnaire included four parts: Part I investigating the participants’ personal background; Part II, their listening comprehension anxiety; Part III, their personality traits as either extroverted or introverted orientation, and Part IV, their linguistic intelligence. In total, 33 questions were designed with 15 questions about “Listening Comprehension Anxiety,” ten about “Personality Traits,” and eight about “Linguistic Intelligence.” The distribution of the questions is shown in Table 3-10:

Table 3-10: Question Distributions in the Questionnaire Part Researched Construct Number of

Questions

Question Number

I Basic Information 4 None

II Listening Comprehension Anxiety 15 Qs 1 ~ 15

III Personality Trait 10 Qs 16 ~ 25

IV Linguistic Intelligence 8 Qs 26 ~ 33

Questions in Part I were used to gather the basic information of the participants.

They included questions about the participants’ student identification number, academic background, and number of examination classroom.²

Part II was a “Listening Comprehension Anxiety” scale adapted from Cheng’s (2007) which explored foreign language learners’ listening comprehension anxiety intensity. Cheng (2007) adopted and translated the Foreign Language Listening Anxiety Scale (FLLAS) by Elkhafaifi (2005) into a Chinese version for her study, which intended to assess Taiwanese senior high school learners’ English listening comprehension anxiety. There were a total of 20 questions in Cheng’s (2007) Chinese version. However, in order not to overwhelm the participants by too many questions given that the time was limited, some variations were made in the present study. The questions that did not seem to be highly related to the listening comprehension anxiety construct were crossed out from the scale. These questions were questions which explored learners’ “attitudes” toward English listening in Cheng’s (2007) study but may not be directly related to learners’ listening comprehension anxiety construct.

Altogether, five of them were taken away, hence a total of 15 questions were finally

2 The participants’ student identification numbers and numbers of examination classroom will be used for the ease of the coding and retrieval for their test scores of the listening comprehension test.

chosen for the present study.³ The present scale follows Cheng’s (2007) design on a five-point Likert scale from “Strongly Disagree” to “Strongly Agree” with the values one to five assigned to each option respectively.

Part III was a “Personality Trait” scale which was adapted from the section

“How Do I Deal with People” in Oxford’s (1993) Style Analysis Survey (SAS) (See Appendix C). This scale was used to investigate learners’ personality traits to see whether they were more extroverted or introverted. Originally, Oxford’s SAS was a 20 item survey with the former ten questions exploring learners’ extroversion traits and the latter ten, the introversion traits.

Nevertheless, the latter five questions about each orientation were removed from the present scale. The reduction of the original scale into a smaller one was due to the time limit as mentioned earlier. Therefore, in the extroversion section (i.e., Questions 1 to 10), Question 6, I like to stay late at parties, was deleted since the party thing might not be frequently experienced in our culture. Question 7 (i.e., Interactions with new people give me energy) and Question 8 (i.e., I remember people’s names easily) were eliminated because they were about “new people,” which was similar to Question 4 (i.e., It is easy for me to talk to strangers). Question 9 (i.e., I have many friends and acquaintances) was categorized as an item surveying the participants’

“friend making” and this has already been asked by Question 2 (i.e., I make new friends easily). Finally, Question 10 (i.e., Wherever I go, I develop personal contacts) was considered an item addressing the “personal connection making,” similar to Question 5 (i.e., I keep up with personal news about other people); thus it was

3 Those questions which were taken out are as follows:

Question 10: By the time you get past the strange sound in English, it’s hard to remember what you are listening to.

Question 14: Once you get used to it, listening to English is not so difficult.

Question 15: The hardest part of learning English is learning to understand spoken English.

Question 19: English culture and ideas seem very foreign to me.

eliminated.

The same criteria for avoiding redundancy in the introversion section were applied. Question 16 (i.e., In a large group, I need to keep silent) and Question 17 (i.e., Gathering with lots of people tend to stress me) were used to ask the learners’

preferences for studying in groups; the concept was similar to Question 13 (i.e., I prefer individual hobbies and sports). Question 18 (i.e., I get nervous when dealing with new people) and Question 20 (i.e., Remembering names is difficult for me) were similar to Question 14 (i.e., It is hard for most people to get to know me), so they were deleted. Question 19 (i.e., I avoid parties if I can) was deleted because it was similar to Question 6 in the extroversion section.

Part IV was a “Linguistic Intelligence” scale, which was selected from Hsieh’s (2000) section of “Linguistic Intelligence” in the Multiple Intelligence questionnaire.

After combing the four parts above, the wordings of the designed questionnaire were slightly revised by the researcher and then proofread by two graduate students to ensure a higher level of validity and better comprehension. In terms of reliability, the cronbach’s alpha of each scale is presented in Table 3-11 below.

Table 3-11: Reliability Tests of the Three Scales in the Designed Questionnaire

Section Scale Cronbach’s

alpha

Number of Items

Valid Cases II Listening Comprehension Anxiety .897 15 1,400

III Personality⁴ .828 10 1,400

IV Linguistic Intelligence .774 8 1,400

As can be seen in Table 3-11, the reliability revealed that the cronbach’s alpha of each scale was .897, .828, and .774 for the anxiety, personality and linguistic

4 Since the personality scale contained items assessing extroversion (the former five items) and introversion (the latter five items), the latter five items were reversely coded to ensure the higher reliability and made the scale as a scale for extroverted personality.

intelligence scale respectively. All these three values were above .600, suggesting that these scales were reliable to measure the constructs intended to be measured.

3.2.3 Interviews

To answer research questions three and four, an interview was conducted with the focused group selected from the two listening comprehension anxiety groups (HI-ANX and LOW-ANX). In addition to their reports of experiencing listening comprehension anxiety, the interviewees were asked to talk about the factors affecting their listening comprehension anxiety and their views about possible solutions to it.

A semi-structured interview was used for its flexibility and the interpretative nature (Nunan, 1992) of the addressed research question. The interview questions were mainly adapted from Vogely (1998) with some sub-questions added by the researcher (see Appendix D). The sub-questions were mainly associated with the participants’ listening comprehension anxiety with “personality” and “linguistic intelligence,” two factors examined in the present study aside from the general discussion about other factors and solutions to their listening comprehension anxiety.

The interview was designed to elicit (1) whether the participants experienced listening comprehension anxiety or not and (2) whether there were potential factors affecting the participants’ listening comprehension anxiety and whether personality and linguistic intelligence played a role in associating with listening comprehension anxiety, and (3) what were the possible solutions for alleviating listening comprehension anxiety. The questions are as follows.

(1) Do you experience anxiety when participating in listening comprehension activity?

(2) What makes you feel anxious when participating in a listening comprehension

(3) What types of exercises, settings, or activities can help you lower your anxiety level?

As mentioned previously, the sub-questions were designed to elicit the participants’ possible factors and suggested solutions associated with their personality traits and linguistic intelligence. For example, the questions were asked like “Being an extroverted or introverted person, do you think this personality trait affect your listening comprehension anxiety when participating in English listening activities?’

and “Do you think your linguistic intelligence ability has anything to do with your listening comprehension anxiety?” Further interview questions were asked to gain more insights and understanding about the participants’ perspectives if necessary.

3.3 Procedures

The data collection procedures proceeded in the following manner. First, the quantitative data was collected via the questionnaire survey after the listening comprehension placement test; second, the semi-structured interview was conducted to examine the qualitative data.

3.3.1 Quantitative Data Collection

Due to the administration policy, all the freshmen were required to take a Freshmen English Placement Test in the beginning of the Fall semester in 2006. Right after they finished the listening comprehension test, they were asked to fill out the questionnaires given that they might experience the highest listening comprehension anxiety. The data was then coded and entered into the SPSS statistical analysis package and constituted the quantitative data bank of the study.

3.3.2 Qualitative Data Collection

Descriptive statistics was employed to identify the participants who were the higher and lower anxious listeners regarding their responses in the listening comprehension anxiety scale. The screening criteria relied on the statistical analysis of the participants’ anxiety level from the top 25% and the bottom 25%. Two anxiety groups (i.e., the HI-ANX group and LOW-ANX group) were then identified. Later, ten students of each anxiety group were chosen to constitute the focused group. These participants were interviewed with the list of interviewed questions. Approximately, the interview for each participant lasted for about ten to fifteen minutes.

Tape-recorded devices were employed to ease the following transcription of the interview content. The interviewed data was then coded and categorized with the qualitative analysis method to generalize a pattern.

Simply put, the data collection procedures of the present study are as follows:

Figure 3-1: Data Collection Procedures of the Present Study Administering the Listening Comprehension Placement Test

Distributing and Collecting the Questionnaire

Identifying the Focused Group

Conducting the Semi-structured Interview

Coding and Analyzing the Data

3.4 Data Analysis

The data of the present study was analyzed as follows.

First, descriptive statistics and a correlation matrix of the constructs of interest were performed to provide a general picture of the present findings.

Second, in order to probe into the relationship between listening comprehension anxiety and performance, descriptive statistics of the two constructs were done to see the general performance of all the participants including the mean scores, standard deviations and other related information. A correlation analysis was then carried out to investigate the relationship intensity between listening comprehension anxiety and performance. In addition, the same analysis procedure was performed on the data of the two anxiety groups in order to examine the group effect. A t-test comparison was employed to see whether the performances of the two groups were significantly different.

Third, exploring the relationship between the participants’ listening comprehension anxiety with their personality traits and linguistic intelligence was done through the descriptive statistics of the three constructs and the correlation analyses. The group effect was also examined by comparing the two anxiety groups, four personality groups, and four linguistic intelligence groups. A t-test comparison and ANOVA were carried out to ensure the significant differences. As proposed by Brown (1992), when there are more than one independent variable with a single dependent variable and when the prediction power of the dependent variable from the independent variables is of interest, multiple regression analysis should be conducted.

Thus, a multiple regression analysis was executed to get a better understanding of the participants’ listening comprehension anxiety, personality and linguistic intelligence.

Fourth, for the qualitative data analysis of other factors and solutions to listening comprehension anxiety, the interviewed data was coded according to the idea

unit obtained. In this way, similar ideas contributed by the interviewees were categorized into a single type.

Finally, a path analysis was performed integrating all the constructs examined in the present study in order to offer a thorough picture of the interplay of these complicated psychological and affective constructs.

3.5 Summary of Chapter Three

This chapter introduced the research design of the present study including the participants involved (a general group and a focused group), the instruments employed (a listening comprehension test, a questionnaire, and an interview), the procedures taken (a quantitative data collection and a qualitative one), and the data analysis employed (descriptive statistics, comparison testing, correlation analysis, multiple regression, path analysis and qualitative data arrangement).