explored. Teachers’ self-efficacy as proposed by Rotter (1954) and Bandura (1977; restated in Bandura, 1997) was examined. It was mentioned how most of the scales developed to measure this construct were either plagued with factorial problems or lacked commonality. Then a scale developed by Tschannen-Moran and Hoy (2001) was introduced that did not have the aforementioned problems. Teacher’s reflection was considered historically and conceptually, and it was stated that many researchers found Schon’s reflection-in-action inapplicable in the context of teaching and promoted reflection-on-action in their practices, which is very similar to Dewey’s notion of reflection.
Next, literature regarding the theories informing the research was reviewed. Based on the literature, it can be concluded that despite the numerous and extensive researches published on teacher’s reflection and its assumed relationship with reflection has never been paid due attention. In addition, the relationship between teachers’ self-efficacy and their reflection levels was rarely examined in the context of TEFL. As it was stated earlier teachers’ self-efficacy is a construct that signifies teachers’ effectiveness.
This study, hence, is designed to explore the relationship among experience, reflectivity and self-efficacy of EFL teachers in Iran.
It was mentioned in chapter one that teacher self-efficacy is associated with numerous student and teacher positive outcomes, such as teaching effectiveness, achievement, and motivation (Knoblauch & Hoy, 2008; Eun & Heining-Boynton, 2007; Barkley, 2006; Milner, 2002; Tschannen-Moran, Hoy, & Hoy, 1998). However, researches conducted on this construct have mostly focused on its outcomes and little attention has been paid to make teachers more self-efficacious. If this study establishes the proposed relationship, a practical way towards efficacy can be suggested.
On the other hand, many researchers have scrutinized different aspects of reflective teaching, theoretically and practically. Furthermore, teachers’ self-efficacy is the core subject of many research papers and dissertations. Nevertheless, there is little published research in the professional literature investigating the relationship between teacher experience and teachers success as reflective practitioners (Akbari, Behzadpoor, & Dadvand, 2010; Akbari, 2007).
Accordingly, this research was instigated to explore the existence of a significant relationship between teachers’ efficacy and reflection considering the level of teaching experience.
This chapter, building on the conceptualizations presented in chapter two, aims at expounding the research strategy used to investigate the relationship between teacher self-reflection and teacher-efficacy of novice and experienced teachers. Consequently, the participants, instruments, procedures, design, and statistical analyses are presented in this chapter.
3.2 The Participants
As mentioned earlier there were 900 teachers employed by Safir Language School, where the study was conducted. Out of this population, 65 teachers were teaching junior classes which are the schools children classes and 835 teachers were teaching senior classes (Adult classes). In order to have a comprehensive sample senior classes’ teachers were chosen as the research population.
Their teaching experience ranged from 1 to 15 years (mean=3). As a result 383 of the participants were chosen as experienced and 324 teachers as novice teachers.
Two instruments were utilized in this study to measure the teachers’ level of reflection, and to quantify their level of self-efficacy respectively. In this section, these instruments are introduced.
3.3.1 Teachers’ Sense of Efficacy Scale (TSES)
Tschannen-Moran and Hoy (2001) developed their own self-efficacy scale and named it Ohio State Teacher Efficacy Scale (OSTES). However, later they prefered the term teachers’ sense of efficacy to refer to Bandura’s self-efficacy (Shaughnessy, 2004), and accordingly, changed the name of their scale to Teachers’ Sense of Efficacy Scale (TSES). The inventory has two versions: long form with 24 and short form with 12 items.. The author recommend that the full 24-item scale (or 12-item short form) can be used with pre-service teachers, because the factor structure is often less distinct for these respondents.” The survey has the reader answer from 1 (nothing) – 9 (a great deal) in a likert scale.
Tested on a sample of 410 male and female participants, including both pre-service and in-service teachers, factor analysis yielded in three moderately correlated factors: Efficacy in Student Engagement, Efficacy in Instructional Strategies, and Efficacy in Classroom Management. Regarding the construct validity of the inventory, Tschannen-Moran and Hoy (2001) report that it was positively related to the previously constructed self-efficacy inventories. Table 3.1 shows the details on the inventory’s reliability.
Reliability details for Teachers’ Self-efficacy Scale (TSES) (Tschannen-Moran & Hoy, 2001)
According to Tschannen-Moran and Hoy (2001),
The results of these analyses indicate that the OSTES [TSES] could be considered reasonably valid and reliable. With either 24 or 12 items, it is of reasonable length and should prove to be a useful tool for researchers interested in exploring the construct of teacher efficacy. (p. 801)
Ever since its development, TSES has enjoyed profound acceptance among researchers and is becoming a standard instrument in the field of teacher education (Ross & Bruce, 2007). It is also used widely in the teacher education context in Iran (Rastegar & Memarpour, 2009; Moa?an & Ghanizadeh, 2009).
To determine the Efficacy in Student Engagement, Efficacy in Instructional Practices, and
Efficacy in Classroom Management subscale scores, grouping is made. Based on the answer ticked different scores from one to four was allocated to each question .
Generally these groupings are:
Efficacy in Student Engagement: Items 1, 2, 4, 6, 9, 12, 14, 22
Efficacy in Instructional Strategies: Items 7, 10, 11, 17, 18, 20, 23, 24
Efficacy in Classroom Management: Items 3, 5, 8, 13, 15, 16, 19, 21
Efficacy in Student Engagement: Items 2, 3, 4, 11
Efficacy in Instructional Strategies: Items 5, 9, 10, 12
Efficacy in Classroom Management: Items 1, 6, 7, 8
3.3.2 English Language Teaching Reflection Inventory (ELTRI)
English Language Teaching Reflection Inventory (ELTRI) was constructed by Akbari et al (2010) to quantify English teachers’ reflections. ELTRI consists of five sub-scales: Practical, Cognitive, Affective, Meta-Cognitive, and Critical. Followed the standard procedures to construct an inventory as proposed by Brown (2001; cited in Akbari et al., 2010) and Dornyei (2003; cited in Akbari et al., 2010). Prior to developing the inventory, Akbari et al. introduced a comprehensive model for reflection, because reflective thinking, as mentioned in chapter 2 has been described by many educators and it is difficult to find a universal agreed upon model for it. To develop the inventory, Akbari et al. reviewed virtually all the published models, and proposed an all-inclusive pattern. In the model proposed by him, there are five sub-scales that constitute reflection: Practical element, Cognitive element, Learner element (Affective), Meta-Cognitive element, and Critical element.
Based on the model, Akbari et al. listed 600 categories that described these elements. The number was reduced to 300 following revisions to omit overlapping items. In the next step, Akbari et al. used expert judgment, as proposed by Dornyei (2003; cited in Akbari et al., 2010), to eliminate redundant items. Ultimately, he chose 42 items to include in the inventory and devised a 5-point Likert scale ranging from “always” to “never” for it. After assessing its face validity, the inventory was piloted on 32 ELT teachers. The Cronbach’s alpha reliability of the inventory turned out to be 0.91.
To validate the inventory, they sent 650 copies to practicing English teachers at various language schools, high schools and centers of higher education across Iran. Out of the 425 completed questionnaires, 308 were chosen and the rest were discarded as they were incomplete or carelessly completed. 48% of the respondents were male and 52% were female. The respondents’ experience ranged from 1 to 40 years (Mean = 603, SD=5.9).
Subsequently, the validation scheme proposed by Mulaik and Millsap (2000; cited in Akbari et al., 2010), consisting of Exploratory Factor Analysis (EFA), Confirmatory Factor Analysis (CFA) and model evaluation, was followed. The results of the statistical analyses yielded in a 29-item inventory, discarding the moral factor from it completely.
On the final analysis of the inventory, Akbari et al. (2010) notes,
In order to determine whether this hypothetical model adequately fits the data, both the Fitting and Validation datasets underwent model-fit analysis. Crucial to such an analysis is the Validation data since it had been reserved from the clustering stage and had not met the given model before. Nonetheless, in order to cross-validate the findings, the Fitting data, which has already been subject to EFA and CFA, was also incorporated into this phase of fitting analysis. As Table 3.2shows, the assessment indices for both Validation and Fitting data outstripped the minimum cut-off points, i.e. 3 for normed Chi-Squared; .9, .85, and .08 for GFI, AGFI and RMSEA respectively; and 0.9 for (TLI) and (CFI). Despite a