CHAPTER 3
Pilot Study: Collaborative Learning in Heterogeneous EFL Early Reading Groups
The purposes of this pilot study were to observe the collaborative behaviors of elementary EFL students in text reading activities, and identify the weaknesses of collaborative EFL reading in traditional elementary EFL settings (without the technological support).
3.1 Subjects
The participants were 26 third-grade students (14 boys and 12 girls, mean age of 8.7 years) from an elementary school in Taipei, Taiwan. At the beginning of the study, each of the participants had 2 school years of experience of EFL learning. Based on the elementary English curriculum standard dictated by the Taipei Municipal Educational Department, at the end of the second school year, students should master the names and sounds of the 26 English letters, 30 spoken words, and 20 sentences of basic daily conversation and classroom English.
The students were grouped into heterogeneous reading groups based on their level of English achievement in the second grade. Grade A students were classified as high reading ability (n=7), those at grade B or C were classified as medium reading ability (n=10), and those at grade D or Fail were classified as low reading ability (n=9).
Students with higher English achievement (referred to as group leaders) were grouped with two or three students with lower achievement (medium- and low-ability students) which resulted in seven reading groups: five groups with four members and two with three members.
3.2 Instruments
3.2.1 Teaching materials
Five teaching packages were included in this study. The lesson structure of the teaching materials was based on a balanced foundation: each teaching package consisted of both basic linguistic-skill training materials (phonemic awareness, phonics rules, and sight words) and text reading articles. The phonics rules, plus a set of chosen sight words, constituted the basic linguistic-skill training module of a teaching package.
Based on the basic linguistic skills contained in a teaching package, a carefully tailored written text was used as a text reading material to provide the students with opportunities to apply their basic linguistic skills to comprehend a written text.
In this pilot study, the target of the phonics rules focused on the general patterns of the five short vowels (a, e, i, o, and u). And the sight words were drawn from the most common sight words general suggested for the levels of kindergarten and first grade of English native speakers (Allen, 1999). A teaching package consisted of two two-lesson activities, which was taught over a period of 2 weeks (two lessons per week,
with a total of 160 minutes for each teaching package), and hence the treatment lasted for 10 weeks. In general, the first two-lesson activities involved the students using printed materials to practice basic linguistic skills, either in entire-class activities or with the students working in pairs. The students then collaboratively performed text reading activities with their groupmates in the second two-lesson activities, in which the students could ask for help from their groupmates when reading the assigned text.
Appendix A is the teaching materials used in this study.
3.2.2 In-class observation checklists
Two in-class observation checklists were developed: a video shooting record and an audio recording checklist as shown in Appendix B. In the video shooting record, group reading behavior (GRB) was defined as a vector GRB (I, P, G) where “I” means that students do text reading individually, “P” represents pairs of students reading together, and “G” stands for all the group members reading together. In addition, in Appendix B, IP refers to two members reading together and the others (maybe one or two members) reading individually; IG refers to one member reading individually, and the others reading together in the case of the group of four members; PP refers to two pairs of two students reading together in the case of a group of four members. Besides this, there are 3 types of behavior patterns belonging to behavior I, and they are active, helpless, and nonsense. Both behaviors P and G contain 5 different types of behavior
patterns, and they are blame, collaboration, domination, support, and nonsense. Tables 1 and 2 are the definitions of the attributes of each pattern type in GRB. On the other hand, in the audio recording checklist, four types of discourse (scaffolding, modeling, nonsense, and blame) were expected to be observed in the small group reading activities. The definitions of those discourse types are shown in Table 3.
Table 1
Definition of Pattern I in GRB
Pattern Definition
A Active: A student actively involves in reading activities.
H Helpless: A student is not able to read by himself and needs some one’s help
N Nonsense: What the student does has nothing to do with the reading activities.
Table 2
Definition of Patterns P and G in GRB
Pattern Definition
B Blame: A student is blamed because he or she does not know how to read.
C Collaborative: Two or more students collaboratively read together.
D Dominate: The reading activities are dominated by the group leader.
S Support: The student(s) read(s) the article with the group leader’s or other groupmates’ help.
N Nonsense: What students do has nothing to do with the reading activities.
Table 3
Definitions of Discourse Types in Small Group Reading Activities
Pattern Definition
C Coaching: A student tries to read with a hint or reading strategy offered by the others.
M Modeling: A student repeats after what the others read.
N Nonsense: What students talk about is nothing to do with the reading activities.
B Blame: A student is blamed because he or she does not know how to read.
3.3 Design
Both quantitative and qualitative data were gathered. The group reading behaviors of elementary EFL learners and the common weaknesses in each component were analyzed. The frequency of each identified weakness was also determined by analyzing the videotapes. The activities were recorded on videotapes, and their discourses were recorded on digital voice recorders. The video data was used to analyze students’ group reading behaviors (GRB), and the audio data was used to investigate the discourses related to peer-assisted learning supports.
After the treatment finished, two observers reviewed the collected data (audio and video) and focused on the discourse between groupmates and the intra-group interactive behaviors while doing reading tasks. Besides, in order to identify the critical moment to use mobile technology and the necessary scaffoldings needed in a CALL model, the observation also focused on weaknesses in the following
components: (a) support provided by groupmates, (b) feedback provided by groupmates, and (c) collaborative processes.
3.4 Procedure
Before the treatment, students were asked to discuss, conclude, and build the collaborative reading rules by themselves (Appendix C), and also to obey these rules.
In the first two-lesson activities of each teaching package, the EFL teacher instructed the students about sight words, phoneme segmentation, and phonics rules. After direct modeling (in which the EFL teacher pronounced sight words and explicitly demonstrated the phonics rules employed to decode and encode a word), the EFL teacher led entire-class activities and inter-group speed-reading contests. At the second of the two-lesson activities of each teaching package, the students performed collaborative text reading activities whilst sitting in groups around a desk. Each student read the text according to a step-by-step reading guide (Appendix D), after which they performed peer assessment. Whenever they encountered a problem in reading a word or comprehending its meaning, they could ask for help from their group leader or groupmates. After each two-lesson period was finished, students were asked to reflect on their collaborative behaviors whilst answering the group-process rating scale (Appendix E) which was based on collaborative reading rules.
Two video cameras were used to videotape children’s collaborative behaviors.
The video cameras were fixed in two corners of the room where four reading groups could be framed in each camera. Besides, seven digital voice recorders were used to record the discourse among group members. Then two observers watched and listened to the collected data repeatedly in order to analyze the students’ individual behaviors and their behaviors towards other students belonging to different learning groups.
3.5 Results
After the treatment finished, according to the checklists two observers coded the pattern categories via watching the video tapes repeatedly. The observations focused on the interactive behaviors among groupmates and the weaknesses that might hinder them from collaborating effectively. Then the Pearson correlation was computed from the pattern percentages of the grouping reading behavior obtained from the two copies of the records, and it was found to be 0.968. The analyzed results of the video data are shown in Table 4.
Table 4
The Whole Class Average of GRB Patterns of the Five Teaching Packages
Teaching packages (%) GRB
a I e o u Individual learning: I 49 42 37 35 29
Group learning IP 9 36 48 24 50
IG 9 6 5 13 0
PP 9 0 5 19 7
G 24 16 5 10 14
Sum 51 58 63 66 71
Table 4 gives the percentages of the GRB patterns found in the five teaching packages, where the definitions of the symbols G, PP, IG, IP, and I can be seen in Tables 1 and 2.
The chi-square analysis reveals that students’ group reading behaviors changed
(individual learning versus group learning) as the treatment progressed from package a to u, (χ2 = 9.61, df = 4, p < 0.05). The results show that as the treatment progressed,
from the first package to the last one, more collaborative behaviors could be observed.
The time period of individual working continually decreased and that of the other behavior patterns involving collaborative reading (patterns G, PP, IG, and IP) increased as the students got more proficient reading with others. This trend change also can be found in Figure 2.
0%
20%
40%
60%
80%
100%
a i e o u
G PP IG IP I
Teaching packages Percentage of GRB patterns
Figure 2. The whole class average of GRB patterns.
Besides caring about the GRB of the whole class, the intra-group interactive behavior that happened in each collaborative reading group was another observation focus. Following the cycle of coding and analysis of video tape data (Jacobs, Kawanaka, & Stigler, 1999), four different kinds of intra-group interactive behavior models were identified and named as Ardent, Be-forced, Supervisory, and Wait-on.
The Ardent model refers to group leaders being willing to help and support their groupmates whenever they need it, as well as directly giving guidance when necessary.
The Be-forced model refers to group leaders having an oppositional attitude to their groupmates, in which they are absolutely unwilling to help or support their groupmates except under pressure from the EFL teacher. The group leaders in the Supervisory model care about their groupmates’ learning and are willing to support their groupmates whenever required. However, the leaders seldom let their groupmates read individually, insisting instead that the entire group read or fill in the worksheets
together under the leader’s supervision. The group leaders in the Wait-on model are essentially stand-by helpers, being willing to help their groupmates but seldom actually giving help, support, or guidance intuitively except when they are asked to do so explicitly.
Furthermore, it was also found that each collaborative reading group had its specific intra-group interactive behavior model. Table 5 shows the typical intra-group behavior models of the seven groups. From Table 5, the typical intra-group interactive behavior model of Group 1 and Group 4 was Ardent (A), of Group 2 was Be-forced (B), of Group 5 and Group 6 was Wait-on (W), and of Group 2 and Group 7 was supervisory (S). In contrast, Group 3 switched the models between Be-forced and Wait-on because of the intervention of the EFL teacher (TI).
Table 5
The Intra-group Interactive Models of Each Reading Group in the Five Teaching
Packages
Packages Groups
a E i o u
1 A A A A A
2 S S S S S
3
B↔W TI
B↔W TI
B↔W TI
B↔W TI
B↔W TI
4 A A A A A
5 W W W W W
6 W W W W W
7 S S S S S
Besides, appendix F shows the discourse examples (the audio data) between groupmates of each intra-group interactive behavior model. In Appendix F, if students used Chinese (their native language) to communicate with each other, the discourses will be printed in italics and followed by the according Chinese version in parentheses.
The analysis of the video and audio data revealed that peer-assisted learning behavior was present in the collaborative reading groups. However, it was also found that just heterogeneously placing students in groups and assigning them group goals (to win in the inter-group speed-reading contest and acquire a reward) does not guarantee effective collaborative activities. Several weaknesses in the students’
collaborative processes were identified by the observers. First of all, the load of each
group leader was not balanced in the small-group collaborative reading activities, in that some of them were kept busy helping their groupmates, whereas others spent most of their time working on their own text. As a result, the participants in those groups with busy leaders may have experienced delays in receiving help from their leaders (which I called it ‘postponed support’) even though some group leaders were available to provide support (which I called it ‘invisible helper’). Furthermore, the medium-ability students were usually asked to read by themselves because their group leaders were busy helping their low-ability groupmates. The absence of real-time feedback resulted in those medium-ability students being unsure about the reading activity (which I called it ‘absent feedback’). Other problems in the small groups included group leaders teasing or omitting groupmates because of their slow learning rate (which I called it ‘ineffective collaborative process’), which resulted in those students becoming more passive while reading and needing more encouragement from the EFL teacher.
3.6 Discussion and conclusion
It is widely agreed by researchers that collaborative learning approach benefits students’ learning achievement. However, there remains much to be investigated on the effect of collaborative group reading approach on elementary EFL learners. It is a worthwhile issue to understand how EFL learners interact with their group members
and how these interactions influence their reading attitudes.
Based on the results of this pilot study, it was found that collaborative learning behavior and peer-assisted learning happened in small reading group. But this pilot study also finds that just heterogeneous grouping students in a small group does not generate an effective collaborative learning because of the weaknesses existing in collaborative EFL reading groups.
The group leaders’ attitude towards their group members and their mission load play important roles in collaborative learning activities. An Ardent or Supervisory group leader makes their groupmates actively learn to read. In contrast to the leaders of Ardent and Supervisory characteristics, a Be-forced or Wait-on group leader not only discourages their groupmates from getting involved in the meaning-based reading text, but also makes them look down on themselves. As a result, those children become more passive in reading activities and expressed a lack of confidence in reading a text in front of others.
Furthermore, a group leader needed to take care of two or three group members with low reading abilities. Therefore, there were always group members were waiting for group leaders’ help. The long-waiting phenomenon inhibits some students with low reading abilities from involving in group learning; moreover they showed a negative attitude to both themselves and their group members. Even some of the students with
medium reading abilities tried to read by themselves when their group leaders were busy with helping other group members, because of lack of real time feedback, they consequently did an unsure reading activity.
The finding of this pilot study shows that collaborative learning and peer-assisted learning are advisable strategies to help elementary EFL learners learn to read with their peers. Besides, it also reveals that the elementary EFL learners need adequate supports to overcome the above weaknesses and to promote the effect of EFL reading teaching and learning.
