This part tries to answer the basic research questions “a” and “b” aforementioned. Prior to the reflective teaching method, pretests were used to obtain reference data (baseline scores) for students’ paragraph writing performance. Two purposes were served by these reference data (baseline scores). The first purpose was to see if the two groups (control and experimental) were equivalent and that there were no differences in paragraph writing performance between them. Furthermore, the pretest findings were compared to the posttest results after the participants had completed the experiment to determine the treatment's effects. To analyze the pretests, descriptive and inferential statistics, independent samples t-test, and paired samples test were used.
3.1 Analysis of Pretest Results of Control and Experimental Groups
Before the intervention of the study, all participants (60) of the study were asked to write paragraph, which was taken as a pretest. The participants had one and half hours to write the following title: “Write a Descriptive Paragraph about You”. The paragraphs were assessed by English teacher, a paragraph writing trainer through reflective teaching, and the researchers. The secondary purpose, as mentioned in the post intervention data collection section, was to compare the pretest with the posttest, which consisted of paragraphs produced after the participants had completed the experiment to evaluate the treatment's effects. It is critical to verify that the experimental and control groups have similar English writing abilities before the experiment. The results of this analysis are summarized in this part, beginning with overall writing performance in terms of the five measuring components (See Appendix A).
3.1.1 Overall Paragraph Writing Performance
The overall writing performance of the pretest is assessed by adding all of the average (mean) scores of the five dependent variables scored by the two raters in terms of their relevance ( paragraph unity, topic sentence, adequate development, paragraph coherence, conclusion, and language unity). The experimental and control groups' paragraph writing performance was compared using descriptive and inferential statistics and an independent samples t-test in terms of overall paragraph writing performance.
As the average mean score in the above table shows, in terms of the overall paragraph writing performance, the participants of the control group (N = 30) had an average mean score of 4.60, while the participants of the experimental group (N = 30) had an average mean score of 4.79. The analysis of the descriptive and inferential statistics indicated that both the control and experimental groups had almost equal performance.
The results of the independent samples t-test indicated no statistically significant differences at the point p˂ 0.05 in the overall paragraph writing performance between the control and the experimental groups (t=-.338, p = .737) even though the average mean score of the experimental group is slightly higher than that of the control group.
3.1.2. Paragraph Writing Performance in Terms of Unity
In order to achieve the unity, a writer must stick to the point. A paragraph is unified when it states only one central idea that is developed by all other statements in the paragraph. (Stephen Bailey, 2011). In this variable, a participant could receive a maximum of 4 points and minimum of 1 point. In terms of unity, the experimental and control groups' paragraph writing performance was compared using descriptive and inferential statistics and an independent-samples t-test.
The above table indicated that in terms of unity, the participants of the control group (N = 30) had an average mean score of .76, while the participants of the experimental group (N = 30) had an average mean score of .85. The analysis of the descriptive statistics indicates the mean score of the experimental group, in regards to unity, is better than that of the control
group. However, the results of the independent samples t-test, obtained from the pretest indicated no statistically significant difference in the writing performance between the control and the experimental groups (t=-.876, p = .386).
3.1.3. Paragraph Writing Performance in Terms of Topic Sentence
Analysis of the topic sentence of the paragraphs focused on “taking the controlling role of sentences in a paragraph as it is the central idea of the paragraph” (Johnstone, 2002). In this component, a participant could get a maximum of 4 points and a minimum of 1 point. The experimental and control groups' paragraph writing performance in terms of topic sentence was compared using descriptive and inferential statistics and an independent-samples t-test.
According to the above table, the participants of the control group (N = 30) had an average mean score of 1.25 while the participants of the experimental group (N = 30) had an average mean score of 1.14 in the topic sentence of the paragraph. The analysis of the descriptive statistics indicates the mean score of the control group, in regards to topic sentence, is better than that of the experimental group. However, the results of the independent samples t-test indicated no significant difference in terms of topic sentence between the control and experimental group (t = .810, p = .422) although the average mean score of the control group was slightly higher than that of the experimental group.
3.1.4. Paragraph Writing Performance in Terms of Adequate Development
Analysis of adequate development in the paragraphs focused on “tracing the full development of the ideas for readers so they will understand the assumptions, evidence and reasoning the writer used or not leaving any significant questions in readers' minds unanswered ” (Brundage and Lahey, 2007). In this variable, a participant could obtain a maximum of 4 points and a minimum of 1 point. The experimental and control groups’ paragraph writing performance in terms of adequate development were compared using descriptive and inferential statistics and an independent-samples t-test.
As shown in the above table, in terms of adequate development, the participants of the control group (N = 30) had an average score of .76, while the participants of the experimental group (N = 30) had .81. The average mean score indicated that the experimental group outperformed the control group in adequate development. Based on the information given in the above table, the results of the independent samples t-test, obtained from the pretest, indicated there were no significant differences in the writing performance in terms of vocabulary usage between the control and the experimental groups (t=-.369, p = .713).
3.1.5. Paragraph Writing Performance in Terms of Coherence
Analysis of the coherence of the paragraphs focused on the “maintaining the balance of ideas in a varied and rhythmic pattern (through coordination, subordination and repetition…) to have a smooth flow from sentence to sentence within a paragraph” (Schifrrin, 1987). In this variable, a participant might receive a maximum of 4 points and a minimum of 1 point. The experimental and control groups' writing performance in terms of coherence were compared using descriptive and inferential statistics and an independent samples t-test.
As shown in the above table, in terms of coherence, the participants of the control group (N = 30) had an average mean score of .76, while the participants of the experimental group (N = 30) had .81. Even though experimental group a bit outperformed than that of the control group, the results of the independent samples t-test obtained from the pretest indicated no significant difference in the writing performance in terms of coherence between the control and the experimental groups (t=-.731, p = .468).
3.1.6. Paragraph Writing Performance in Terms of Conclusion
Analysis of the conclusion of the paragraphs focused on “The last sentence of the paragraph whether or not it restates the main idea of the paragraph and indicates why the topic is important” (Schifrrin, 1987). In this variable, a participant could get a maximum score of 4 and a minimum score of 1. The experimental and control groups' paragraph writing performance was compared in terms of conclusion using descriptive statistics and an independent samples t-test.
The above table shows that, with regards to conclusion, the participants of the control group (N = 30) had an average mean score of .65, while the participants of the experimental group (N = 30) had an average score of .64. The average mean score of the two groups indicated that both the control and the experimental groups almost equally performed in conclusion. Moreover, as presented in the above table, the results of the independent samples t-test, obtained from the pretest, indicated there is insignificant difference in the paragraph writing performance in terms of conclusion between the control and the experimental groups (t = .83, p=. 934).
3.1.7. Paragraph Writing Performance in Terms of Language Use
Analysis of the language use of the paragraphs focused on “right use of vocabulary, syntax, grammar, punctuation, and mechanics” (Brundage and Lady, 2007). In this variable, a participant could get a maximum score of 4 and a minimum score of 1. The experimental and control groups’ paragraph writing performance was compared in terms of language use using descriptive and inferential statistics and an independent samples t-test.
The above table shows that, with regards to language use, the participants of the control group (N = 30) had an average mean score of .55, while the participants of the experimental group (N = 30) had an average score of .64. The average mean scores of the two groups indicated that the experimental group outperformed than that of the control group in language use. Moreover, as presented in the above table, the results of the independent samples t-test, obtained from the pretest, indicated there is insignificant difference in the paragraph writing performance in terms of language use between the control and the experimental groups (t=-.047, p=. 299).
3.2.2 Analysis of the Post-intervention Data
The posttests were used to compare the paragraph writing performance of the two groups (control and experimental) after 16 weeks of training. Descriptive and inferential statistics, paired-samples t-test, and independent samples t-test were used to analyze the results.
3.2.2.1. Analysis of the Posttests
The two raters rated the posttests of the two groups (See Appendix B). Once the raters have completed the grading, the average mean scores of each dependent variable were computed. All the mean scores of the variables were entered and organized in an excel spreadsheet. The data were then imported into SPSS for analysis. The researcher used the overall mean scores of all the 60 students’ posttest scores rated by the two raters. In the following sections, each research question is tested against the collected data.
3.2.2.2 Testing Research Question One
The first research question is ‘Will there be a statistically significant improvement in the students’ paragraph writing performance in the posttest in terms of unity, topic sentence, adequate development, coherence, conclusion, and language unity in the experimental group due to the effects of reflective teaching?’ The purpose of this research question was to investigate if the experimental group’s paragraph writing performance (for their overall paragraph writing performance in terms of unity, topic sentence, and adequate development, coherence, conclusion, and language unity) has improved in the posttest compared with their pretests.
To answer the above research question, descriptive analysis and paired samples t-tests were used to investigate any statistically significant difference in the posttest findings compared with the pretest results of the experimental group. The following is a presentation of the descriptive and inferential statistics computed for the final scores on the posttest compared with the pretest for the control and experimental groups regarding their overall writing performance (unity, topic sentence, adequate development, coherence, conclusion, and language unity).
As the descriptive and inferential statistics showed, the average posttest mean scores of the participants showed improvement in the six components of paragraph writing. This showed that implementing reflective teaching by reflection-in- action, reflection-on-action, and reflection-for-action helped participants to improve their paragraph writing performance in the posttest. A paired samples t-test was conducted to explore the effects of reflective teaching on students’ paragraph writing performance in unity, topic sentence, adequate development, coherence, conclusion, and language use in the posttests. Participants in the posttests showed better performance in all variables except conclusion in their posttests. Then, there were statistically significant differences at the p ˂.05 level in the scores, except conclusion, of the posttest and the pretest in all components as shown in the above table. Therefore, participants in the experimental group have shown statistically significant improvement in their performance in all components (unity, topic sentence, adequate development, coherence, and language use except conclusion) on the posttest. In conclusion, there were statistically significant differences at p˂.05 level in scores of the posttest than the pretest in all variables except conclusion as p˃.05 level. One major explanation for the experimental group’s improvement in their overall paragraph writing performance compared to their pretest results is that reflective teaching provided the participants with exposure to edit, comment, discuss and revise their own and peers’ drafts reflectively in EFL class. This finding is supported by studies conducted (Kumaravadivelu, 2002).
The second research question is “Which groups’ paragraph writing performance (the control group writing under teacher control method and the experimental group writing based on reflective teaching) will have a statistically significant difference, if any, in the posttest in terms of unity, topic sentence, adequate development, coherence, conclusion, and language use?” The objective of this question was to look at the difference between students’ paragraph writing performance in the control group (face-to-face /conventional paragraph writing instruction) and the experimental group (reflective teaching paragraph writing training) on the posttest in terms of unity, topic sentence, adequate development, coherence, conclusion, and language use. Descriptive and an independent samples t-test were employed to answer the study's second question. Given that there was no significant difference in their mean scores on the pretest at the start of the experimental, it is reasonable to assume that any significant difference in their mean scores on the posttest is due to the intervention. The descriptive and inferential statistics derived for the final scores on the posttest for both groups are presented below.
3.2.4 Overall Writing Performance (Posttests)
This section shows the average mean scores and independent samples t-test results of both groups in terms of unity, topic sentence, adequate development, coherence, conclusion, and language use.
The above table reveals that participants from both groups showed almost similar paragraph writing performance in unity (m = 1.186) of the experimental group and (m = .863) of the control group. To see whether there was a statistically significant difference between the two groups in terms of unity, an independent samples t-test was conducted. Results of the analysis showed that participants in the experimental group performed better (m = 1.186) compared to participants in the control group (m = .863), and there was statistically significant difference at the p ˂0.05 level in both groups’ scores (t=-.4.201, p = 0.000) as shown in the table above.
The above table shows that the experimental group performed better (m = 1.574) than the control group (m = 1.071) in writing paragraph topic sentence. The average mean scores of the two groups showed that there was a significant difference between the control and experimental groups in paragraph writing performance due to the effects of reflective teaching. An independent samples t-test was conducted to explore the effects of reflective teaching on the students’ paragraph writing performance with particular attention of the topic sentence writing in the posttest. Though the mean scores of the two groups showed different results (M = 1.071 and 1.574), there was statistical significant difference at the p˂0.05 level in both groups’ scores in a topic sentence of the paragraph (t=-4.347, p = 0.000).
The above table shows that the experimental group (m = 1.076)performed better than that of the control group (m = .723) in writing coherent paragraph. An independent samples t-test was conducted to explore the effects of reflective teaching on the students’ paragraph writing performance in terms of coherence in the posttest. Therefore, there was statistically significant difference between the control group and experimental group at the p˂0.05 level in terms of writing coherent paragraph (t=-4.192, p = 0.000).
The above table reveals that the mean score of the experimental group (m = .926) is better that the mean score of the control group (m = .698) with regards paragraph conclusion. However, according to an independent samples t-test conducted to explore the impact of reflective teaching on the students’ writing performance in conclusion in the posttest, participants in both groups didn’t show better performance in conclusion. As a result, there were no statistically significant differences at the p˃0.05 level in both groups’ scores in conclusions (t=-2.158, p = 0.35).
The above table shows that the experimental group (m = .987) performed better than that of the control group (m = .667) in using language appropriately in writing paragraphs. An independent samples t-test was conducted to explore the effects of reflective teaching on the students’ paragraph writing performance in terms of language use in the posttest. Therefore, there was statistically significant difference between the control group and experimental group at the p˂0.05 level in terms of using language in writing paragraph (t=-4.102, p = 0.000).