國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Stage 5: Personal Interview
At the end of the pleasure reading program, every participant in the EG was invited to an individual interview to get more specific understanding of their reaction to this program. The interview was held during the lunch break or after school in the counseling room. Before the interview started, the participants were told to feel free to talk about their various opinions on this reading program. To get more precise and specific understanding, the interview was conducted in Chinese (Appendix H). All the interviews were recorded and the researcher took some notes. The interview started with three minutes of warm-up and lasted for about 30 minutes respectively.
Data Analysis
Data collected in this study were analyzed quantitatively. Motivation questionnaires and results of GEPT reading and writing tests were analyzed quantitatively with the software Statistical Package for the Social Science (SPSS).
The .05 level of significance was selected for the analysis of data. Book reports and interview records were used for secondary references for discussion. The analysis of the two research questions was described as follows:
Question 1: How do the experimental group participants change in their reading motivation through the pleasure reading program?
Students usually come to school with different knowledge and learning
backgrounds, so their motivations are very different from each other (Day & Bamford, 2004). McCombs (1997) also claimed the diversity of reading motivation by stating
“what may motivate a particular student to read is not necessarily what may motivate another student to read” (p.126). Consequently, it was difficult to investigate the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
change of reading motivation by means of the difference between groups. Therefore, the significant difference was tested between the pretest and the posttest within groups.
For analysis, each item’s rating scale from “strongly agree” to “strongly disagree” was counted as 4 points, 3 points, 2 points, and 1 point respectively. Items with negative descriptions (items 4, 6, 24, 28, 33, 40, 50) were counted in reverse order as 1 point, 2 point, 3 points, and 4 points. The total points in the pretest and posttest were computed and compared by Paired t-test. If the outcome of the t-test was less than .05, there was a significant difference between the pretest and the posttest.
This may imply that the EG participants’ reading motivation level improved after the program. In addition to the overall motivation, three categories and 12 dimensions were also analyzed respectively between the pretest and the posttest within groups.
This was to probe which dimensions had more influence on reading motivation through the pleasure reading program.
In addition, the data in the EG’s personal interviews were utilized for secondary references for the discussion with regard to the impact of pleasure reading on their motivation to read and the feasibility of implementing pleasure reading. Before the researcher analyzed the data, the recorded were transcribed and translated into English.
The data gathered supported and explained the results obtained from the quantitative analysis and provided more insights into the results.
Question 2: To what extent do the experimental group participants improve their general reading comprehension and writing ability when compared to the control group participants?
Before the pleasure reading program started, the results of the two groups’
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
reading comprehension pretest were analyzed with Independent t-test to check if there was no difference in language proficiency between the two groups. If the p-value was more than .05, there was no significant difference between the two groups. Then they were considered homogeneous groups. After the posttest, the gain scores between the two groups were examined by Independent t-test. If p-value was less than .05, there was a significant difference. This may imply that pleasure reading could help the EG participants improve their reading ability.
Two parts were included in the writing test, sentence writing (50%) and paragraph writing (50%). Before the computer scoring and quantitative analysis, all the writing papers were graded by three raters. Besides the researcher, two raters were invited to grade the writing papers. One is a junior high English teacher, and the other is a senior high English teacher. The senior high English teacher had experience in rating GEPT writing papers in the Language Training and Testing Center (LTTC) before. All the raters had no idea which paper the pretest or posttest was and there were no names on the papers either. The writing marking protocols set by the LTTC were adapted to grade writing papers. The marking protocols for sentence writing were two points for each question. If the mistake in the sentence was the key point, then there were no grades for the question. If the mistake was the minor one, there was one point for the question. The full marks for the sentence writing part were thirty points. Paragraph writing was graded on five analytic standards. The full marks for the paragraph writing were five points. Table 3.5 displays the marking protocols for paragraph writing. Before grading the papers, both sentence and paragraph writing marking protocols were explained clearly to raters. Then raters scanned all the papers and gave some examples quoted from papers. Raters graded these examples with the marking protocols and discussed and exchanged ideas.
‧
Marking Protocols for Paragraph Writing Grades Description
5 Writing with all the main ideas of the topic and correct sentences
4 Writing with most of the main ideas of the topic and complete sentences Writing with some of the main ideas of the topic and comprehensible 3
sentences.
Writing with few of the main ideas of the topic and incomprehensible 2
sentences.
Writing with none or only one of the main ideas of the topic and 1
incomprehensible sentences.
0 No answer.
After the sentence writing and paragraph writing were graded by raters, the results were quantitatively analyzed. First, the results of both groups’ pretests were analyzed with Independent t-test to see if there was no difference in writing
proficiency between the two groups. Second, the gain scores between the two groups were examined by Independent t-test. If p-value was less than .05, there was
significant difference between the two gain scores. This may imply that the EG participants made significant progress in writing ability through the reading program.
Besides, the results of the program response questionnaires were used to support and explain the findings in reading and writing tests. At the same time, they were also used for secondary references for the discussion with regard to the impact of pleasure reading on the participants’ language proficiency. Before the researcher utilized the results, the total points in the program response questionnaires had been computed and compared by Independent t-test. Moreover, the data from the book
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
reports and book logs were also gathered and analyzed qualitatively to interpret or support the findings and the discussion about the results of reading and writing tests.
Scorer Reliability
With the marking protocols for sentence writing and paragraph writing set by the Language Training and Testing Center (LTTC), all the writing papers of both groups were graded by the researcher first. Then these writing papers were graded by the other rater to make sure that the interrater reliability of this study was not
threatened (Hughes, 2003). Twenty percent of the writing papers from each group were selected randomly and graded by the interrater (Cohen, 1960). The interrater reliability was 92% for sentence writing, 95% for paragraph writing. Wang (2009) noted that high interrater reliability in scoring EFL writing was 0.80. Thus, the interrater reliability for sentence writing and paragraph writing in the study reached the acceptable level.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Chapter 4 Results
This chapter displays the results of the study. Results gathered from GEPT reading and writing tests, motivation questionnaires, and program response
questionnaires were combined to answer the two research questions respectively. First, answers to the first research question involve with results of the overall reading
motivation, three aspects of reading motivation, dimensions of each aspect. Second, answers to the second research question include results of the reading comprehension test, writing test, and program response questionnaire.
Research Question 1: How Do the Experimental Group Participants Change in Their Reading Motivation Through the Pleasure Reading Program?
Results of Overall Reading Motivation
First, the results of the participants’ overall reading motivation changes are presented in Table 4.1. In the EG, the mean scores of the pretest and the posttest were 128.94 and 138.06. The gain score was 9.12 and there was significant difference between the pretest and posttest mean scores (p =.01). In the CG, the mean score of the pretest was 129.79 and the posttest was 125.86. The score decreased in the posttest (G = -3.93). Second, as can be seen in Table 4.2, the posttest mean score of the EG was greater than that of the CG (138.06 : 125.86). The difference between them reached a significant level (p = 0.05). Compared to the CG, the EG reached a significant level in both results. In addition, the results in the first part of motivation questionnaire (Table 4.3), which was about the participants’ belief in pleasure reading also showed significant difference. In comparison to the CG with a negative result (G
‧
= 1.39; p = .03), the EG appeared to show a more positive attitude toward pleasure reading after joining this program. All these results may indicate that the EG participants’ overall reading motivation increased after their joining the pleasure reading program.
Table 4.1
Mean Comparison of the Overall Reading Motivation in the Pretest and the Posttest within Groups
Mean Comparison of the Overall Reading Motivation in the Posttest between Two Groups
Mean Comparison of Participants’ Belief in Pleasure Reading between Two Groups
Pretest Posttest Gain p
EG 16.53 17.34 0.81 0.11
CG 16.46 15.07 -1.39 *0.03
*P＜.05
Results of Three Aspects of Reading Motivation
Table 4.4 summarizes the results of each aspect of the overall reading
motivation. As can be seen, there were two aspects with significant difference among
‧
the three in the EG. The mean score in Competence and Efficacy Beliefs was 27.81 in the pretest and 29.91 in the posttest, and it showed a p value of .01. The mean score in Achievement Values and Goals was 76.34 in the pretest and 82.44 in the posttest, and it showed a p value of .02. The difference in both aspects reached a significant level.
As for the Social Aspect of Reading, the mean score was 17.53 in the pretest and 18.16 in the posttest. Although it made a little progress (G = .63), the result did not reach the significant level. In brief, the statistic results revealed that pleasure reading may have positive influence on the participants’ reading efficacy belief and reading achievement values and goals.
In the CG, the data did not show positive results in any aspects. The gain scores in Competence and Efficacy Beliefs, Achievement Values and Goals, and Social Aspect of Reading were -.78, -2.65, and -.21 respectively. All of the three aspects revealed a decreasing gain after the program. In contrast to the outstanding progress in the EG, the skill-based reading approach may have decreased the CG participants’
motivation to read. In Achievement Values and Goals especially, the CG lost the most gain scores (G = -2.65) while the EG attained the most gains (G = 6.1). This was the most distinctive reading aspect among the three. This may imply that traditional skill-based instruction made the participants easily lose expectation of value in reading. On the contrary, pleasure reading may have a strong impact on the participants’ reading achievement values and goals.
Table 4.4
Mean Comparison of the Three Aspects of Reading Motivation between Two Groups
3 Aspects Pretest Posttest Gain t-ratio p
EG 27.81 29.91 2.1 -2.60 .01*
Competence and Efficacy
Beliefs CG 28.64 27.86 -.78 1.12 .27
‧
Results of Dimensions in the First Aspect of Reading Motivation The statistical results of the first reading motivation aspect are presented in Table 4.5. This aspect involves reading self-efficacy, reading challenge, and work avoidance. Among the three dimensions, both groups attained the most gain score (EG =1.19; CG = .65) in the dimension of Self-efficacy. The mean score of the EG was 8.25 in the pretest and 9.44 in the posttest. The result showed a very significant difference (p = .00) in this dimension. In the dimension of Challenge, the EG showed little progress (G = 1.13) after joining the program and no significant difference was found. However, the CG had a decreasing gain (G = -.75) in this dimension. In the dimension of Work Avoidance, both groups’ scores decreased after joining the program (EG = -.22; CG = -.68). In sum, pleasure reading seemed to have great influence on the participants’ self-efficacy with regard to reading.
Table 4.5
Mean Comparison of Dimensions in the First Aspect of Reading Motivation between Two Groups
‧
Results of Dimensions in the Second Aspect of Reading Motivation In Table 4.6, results of seven dimensions in the second aspect are shown as follows: First, three dimensions showed significant improvement in the EG. They were the dimensions of Recognition (G = 1.13, p =.02), Grades (G = 1.29, p =.04), and Enjoyment (G = 2.6, p =.00). In the dimension of Recognition, both groups made gain scores. The EG not only produced more gains (G = 1.13) than the CG did (G = .22) but also reached a significant difference (p = .02). The EG
participants thought their English ability was more praised by others after getting involved in the pleasure reading program. In the dimension of Grades, the EG had a score of 10.09 in the pretest and 11.38 in the posttest. The result reached a significant level (p =.04). However, the CG had a decreasing gain (G = -.03) in this dimension.
Reading for grades may have a greater impact on the EG participants’ reading motivation. The EG participants may think that pleasure reading helped them get more grades. The dimension of Enjoyment was considered the most significant dimension (p = .00), but the CG showed a decreasing gain (G = -.43) in this
dimension. Obviously, the EG participants found pleasure reading class much more enjoyable while the CG may have lost their reading interest in the skill-based reading class.
Second, compared to the positive results in the EG, most dimensions in the CG revealed decreasing gains except for the dimension of Recognition (G = .22). The dimension of Curiosity even reached a negative significant level (G = -1.46, p =.04).
‧
This may indicate that the traditional skill-based approach easily hinders the participants’ reading achievement and their values for reading, especially their curiosity about English reading.
Third, in the dimension of Involvement, although the EG did not show a significant difference, its gain score is better than the CG’s (EG = .97; CG = -.57). In contrast to the CG, the EG participants may easily engage in different contexts by reading for pleasure. Similarly, the result in the dimension of Importance in the EG was not significantly positive, but the EG made more gains than the CG (EG = .29;
CG = -.29). This may imply that pleasure reading made the EG participants place more emphasis on English reading. As for the dimension of Competition, although the EG made a little progress in this dimension (G = .88), it did not show a significant difference. Pleasure reading may not have very strong impact on the participants’
intention to outperform others in reading. However, in comparison to the result in the CG (G = -.47), the EG still showed greater intention to outperform others.
Table 4.6
Mean Comparison of Dimensions in the Second Aspect of Reading Motivation between Two Groups
Recognition EG 9.03 10.16 1.13 -2.39 .02*
‧
Results of Dimensions in the Third Aspect of Reading Motivation Subsequently, the statistic result of third aspect is presented in Table 4.7.
This aspect is related to the impact of external factors (Social and Compliance) on reading motivation. No significant difference was found in the dimensions of this aspect. However, in terms of gain score, the EG had a slightly higher score than the CG (E = .65; C = .39) in the Social dimension. In the dimension of Compliance, both groups showed a decreasing score (G = -.04; C = -.57) throughout the program. In sum, pleasure reading seemed to influence the social aspect of reading very little.
Table 4.7
Mean Comparison of Dimensions in the Third Aspect of Reading Motivation between Two Groups
‧
Research Question 2: To What Extent Do the Experimental Group Participants Improve Their Reading and Writing Ability When Compared to the Control
Group Participants?
Results of Reading Comprehension Test in GEPT
As shown in Table 4.8, the mean scores between the EG and the CG were 59.69 and 55.31 in the pretest. The result demonstrated that there was no significant
difference (p =.54) between the two groups’ reading competence. That is, the two groups had similar linguistic competence levels before they participated in this study.
After the program, the EG scored a mean of 66.03 while the CG scored 57.41 in the posttest. Both groups made some progress in terms of the gain scores (EG = 6.34; CG
= 2.10). The gain scores between groups were compared by Independent t-test and it showed no significant difference (p =.18). The EG did not outperform the CG significantly. This may imply that pleasure reading could not make a significant improvement in the participants’ English reading ability.
Table 4.8
Mean Comparison of Reading Comprehension Test between Two Groups
EG (N=32) CG ( N=29)
Mean SD Mean SD
t-ratio p
Pretest 59.69 (28.88) 55.31 (26.52) -0.61 0.54 Posttest 66.03 (29.22) 57.41 (25.65) -1.22 0.23 Gain 6.34 (16.30) 2.10 (19.99) -0.91 0.18
Although there was no significant progress in the EG group, its participants seemed to become more confident of their reading ability than those in the CG over the program. As shown in Table 4.9, the program response questionnaire, the mean score was 2.91 in the EG and 2.52 in the CG, and the difference between them was
‧
Mean Comparison of Program Response Questionnaire between Two Groups Item 3:
Reading Class Helps me Better Understand Paragraphs or Reading Passages.
N Mean Std. t-ratio p
CG 29 2.52 0.69
EG 32 2.91 0.53
-2.49 *0.02
*P＜.05
Results of Writing Test in GEPT
In Table 4.10, the EG scored a mean of 34.74 while the CG scored 29.48 in the pretest. The result demonstrated that there was no significant difference (p =.45). Both groups had similar writing proficiency levels before the program. After the pleasure reading program, the EG got a mean score of 44.43 while the CG got 36.03. Both groups made some progress in mean scores (EG = 9.69; CG = 6.55). Then, the gain scores between groups were compared by Independent t-test, and it showed no significant difference with a p value of .20. This may indicate that pleasure reading did not really result in a significant improvement in the participants’ writing proficiency.
Table 4.10
Mean Comparison of Writing Proficiency Test between Two Groups Experimental Group Control Group
(N=32) (N=29) Mean SD Mean SD
t-ratio p
Pretest 34.74 (28.50) 29.48 (25.36) -0.76 0.45 Posttest 44.43 (34.17) 36.03 (27.75) -1.05 0.30 Gain 9.69 (15.59) 6.55 (13.02) -0.85 0.20
‧
Though the EG participants did not make significant progress, they seemed to become more confident of their writing ability after doing pleasure reading in class.
Some items related to writing ability in the program response questionnaire revealed better impact on writing ability in the EG than those in the CG. For example, Table 4.11 demonstrates the results of the participants’ own perception of sentence writing ability. The mean score of the EG was greater than that of the CG (2.97 > 2.24). And the difference between groups was very significant (p = 0.00). Table 4.12 presents the results of the participants’ own perception of spelling ability. The mean score of the EG was greater than that of the CG (2.91 > 2.48) and it also reached the significant level (p = 0.02). These results may imply that the EG participants became more confident of writing sentences than the CG participants.
Table 4.11
Mean Comparison between Two Groups in Program Response Questionnaire Item 7:
Reading Class Helps me Write Sentences More Correctly.
N Mean Std. t-ratio p
N Mean Std. t-ratio p