The Nexus between Iranian EFL Teachers ’ self-efficacy , Teaching Experience and Gender

This study investigates the main effects of years of teaching experience and gender and their interaction effects on Iranian EFL teachers' sense of self-efficacy. To this end, the questionnaire of TEBS-Self (Teachers' Efficacy Beliefs SystemSelf) (Dellinger, Bobbett, Oliver, & Ellett, 2008) was used. It was distributed to 180 EFL teachers. 94 of them had taught for less than or equal to three years (G1), and 86 had taught for more than or equal to three and a half years (G2). Through regression analysis it was found that while experience and gender had no significant interaction effect on the participants’ efficacy beliefs, G2 had significantly more positive efficacy perceptions than G1. Also, except for two components of TEBS-Self, the female teachers were found to have significantly higher self-efficacy than the male participants on the whole questionnaire and two components of it. Conclusions and suggestions for further research are also discussed.


Introduction
1.1 Teacher self-efficacy Bandura (1994) defines self-efficacy as peoples' beliefs about their capabilities to produce designated levels of performance that exercise influence over events that affect their lives.Similarly, he defines teacher self-efficacy as the degree to which teachers believe they have the ability to affect students' performance.Thus, what are clearly emphasized in Bandura's social cognitive theory are peoples' perceptions of their capabilities rather than real/actualized capabilities since beliefs and perceptions greatly influence how one's potential is realized and utilized.Bandura (1986Bandura ( , 1995) ) believes people's beliefs about their efficacy can be developed by four main sources of influence.The most effective way of creating a strong sense of efficacy is through mastery experiences.Successful performance in the past instills more confidence and positive perceptions of the self into people, and, thus, they expect successful performance in the future.Vicarious experiences provided by social models and significant others also help create and strengthen self-efficacy beliefs.Seeing people similar to oneself succeed by sustained effort raises observers' beliefs that they similarly possess the capabilities to master comparable activities required to succeed.Social persuasion is another way through which people's efficacy beliefs are formed and strengthened.People who are persuaded verbally that they possess the capabilities to master given skills and fulfill certain activities are likely to mobilize greater effort and sustain it while negative feedback from others makes them harbor self-doubts and dwell on personal deficiencies when problems arise.Finally, Bandura has argued that physiological and emotional states affect the ways in which efficacy beliefs are formed.These states can be positive, such as happiness and excitement, or negative, like stress, tension, and anxiety.
Since the introduction of the concept of self-efficacy in social sciences, research in different areas of education has investigated contributions of teacher efficacy perceptions to various instructional variables and has shown, on most of the occasions, that there is a significant positive correlation between teacher self-efficacy and these variables.A brief discussion of some of these studies follows.
Efficacy beliefs have been shown to affect teacher activity and productivity (Ashton & Webb, 1986), the goals they identify for learning tasks in the context where they word and the amount of effort and persistence they demonstrate in fulfilling the task (Bandura, 1995;Tschannen-Moran, Woolfolk-Hoy & Hoy, 1998).Also, it has been found that, compared to low self-efficacy teachers, teachers with high efficacy beliefs set higher goals for themselves and their students, try harder to achieve those goals and persist through obstacles (Henson, 2001).In addition, they generate stronger student achievement (Goddard, Hoy & Woolfolk Hoy, 2004;Ross, 1998;Tschannen-Moran, Woolfolk Hoy & Hoy, 1998), are more emotionally intelligent (Penrose, Perry & Ball, 2007), use more effective instructional strategies in teaching mathematics (Swars, 2005), display extra-role behavior towards the team and the organization (Somech & Zahavy, 2000), tend to embrace innovations more readily, and use a greater variety of instructional strategies (Riggs & Enochs, 1990;Wenta, 2000).In an EFL context, highly self-efficacious teachers have been found to be more willing to use communicative-based strategies in EFL classes (Eslami & Fatahi, 2008), With regard to teachers' effective classroom management and interaction with students, many studies are in favor of positive influences of efficacy beliefs on these abilities.For example, teachers with high self-efficacy have been shown to make less negative predictions about students and adjust their predictions when student characteristics change (Tournaki & Podell, 2005).Also, highly efficacious teachers have been found to use more effective and positive classroom management strategies, such as integrating and compromising, (Britta, Morris & Brassard, 2006;Brouwers & Tomic, 2005;Ross & Bruce, 2007), as well as more inquiry-oriented and learner-centered teaching strategies (Czemiak, 1990).
Another group of studies have focused on factors which affect teachers' perceived efficacy.Henson (2001), for example, shows that participatory teacher research, as an active and collaborative means of teacher professional development, has positive effects on general and personal teacher efficacy.Also, teachers' cultural and social backgrounds seem to impact on teachers' sense of self efficacy (Lina & Gorrellb, 2001).The effects of other factors such as gender and years of experience on teachers' efficacy beliefs have been explored too.Since the present study investigates the possible differences in Iranian EFL teachers' efficacy beliefs based on these two variables, a brief account of some of these studies follows.

Teacher self-efficacy and experience
As mentioned above, Bandura (1977Bandura ( , 1986) ) believes mastery experience or performance experience is the most important source of self-efficacy, implying that successes strengthen and failures weaken it.Since teachers usually gain extensive experience of successful and unsuccessful performances throughout their years of teaching, this assumption has generated in-depth research into how teachers who have been involved in teaching for different periods of time perceive their teaching efficacy (e.g., Penrose, Perry & Ball, 2007;Hoy & Woolfolk, 1993;Moran & Hay, 2002;Imants & Brabander, 1996;Fives & Looney, 2009;Fives, 2010;Soodak & Podell, 1996;Campbell, 1996;Yan, 2006;Kotaman, 2010).Some studies have shown a positive correlation between years of experience and efficacy beliefs of teachers.To begin with, in a study conducted by Lin and Tsai (1999), the experts and teachers reported higher self-efficacy than novice teachers.Liu et al. (2007) also showed a positive correlation between teachers' years of experience in teaching science and their personal science teaching efficacy scores.Similarly, Wolters and Daugherty (2007) found that teachers in their first year of teaching reported significantly lower self-efficacy for instructional practices and classroom management than did teachers with more experience.However, one can find studies which have yielded results contradicting those reported above.For example, Woolfolk (1990) and Weinstein (1988) found that novice teachers reported high personal and professional efficacy.Or, Soodak and Podell (1997) observed that experienced teachers are more resistant to change in their beliefs of personal efficacy than teachers with less experience.Also, there are some other studies which show mixed results, like Gorrell and Dharmadasa (1994) which indicates that, although pre-service teachers reported higher efficacy for implementing new methods of instruction, experienced teachers reported higher efficacy for classroom management, organization of instruction, and impact on students.Finally, some researchers have also found no significant relationship between teachers' years of experience and their efficacy beliefs (e.g., Guskey, 1987).
In light of the above brief review of research, it is very difficult to conclude that studies show a generally similar direction regarding how these two variables are related to each other.

Teacher self-efficacy and gender
Gender is another factor which might influence, in one way or another, teachers' professional lives, in general, and their efficacy perceptions, in particular.It is believed that social relations and the dominance of either gender (usually male) affects teachers' lives.Female professionals are usually subordinate to male authorities in educational settings where professional interactions are usually characterized by marginalization of women (Bartlett, 2005;Lin et al., 2004).Since, as argued by social cognitive theorists (e.g., Bandura, 1994;1995), efficacy beliefs are constructed and reconstructed through people's social experiences and interactions, it is rather impossible to deny influences of unequal power dynamics existing in the field of TESOL and resultant explicit and implicit practices, such as unfair competition in male-dominated settings (Bandura, 1995), on women's efficacy beliefs (Davis & Skilton-Sylvester, 2004).
Studies which have focused on how gender might affect teachers' sense of efficacy are a few and have yielded vastly different results.Imants and De Brabander (1996), using a modified version of the Teacher Efficacy Scale (TES), concluded that gender influences teachers' self-efficacy.More precisely, they found that male elementary teachers appeared to have higher self-efficacy for pupil-oriented and school-oriented tasks than female teachers.On the contrary, Cheung (2006) found that female teachers have significantly more general efficacy than male teachers, while it is worth mentioning that female teachers in this study were generally older and had longer teaching experience than male teachers.Finally, some other studies, such as Ghaith andShaaban (1999), Tschannen-Moran andWoolfolk Hoy (2002), and Wilson et al. (2004), showed that gender has no signification effect on teachers' self-efficacy.
As mentioned above, studies focused on teaching experience and teachers' perceived efficacy fail to show a general direction in which years of teaching experience might contribute to teachers' efficacy beliefs.When it comes to studies on differences between the two genders in terms of teacher self-efficacy, the picture is even less clear since, first of all, these studies are a few and, second, as explained above, they present a wider variety of differing and opposing observations.
Regarding the interaction effect of years of teaching experience and gender on self-efficacy, which is a major focus in this study, the picture is much less clear.More precisely, the present researcher could only find one study, i.e.Imants and De Brabander (1996), in which it has been argued that male teachers with at least 10 years of experience seem to have lower perceived efficacy for school-oriented tasks than those with medium-level (5-9 years) experience.Conversely, female teachers with medium-level experience of teaching seem to have higher self-efficacy than those with at least 10 years of teaching experience for school-oriented tasks.In addition, generally speaking, teacher self-efficacy has not been much focused on in research in the realm of ELT (English Language Teaching) and in the context of Iran.Just a few MA theses (e.g., Abednia, 2006;Izadinia, 2008;Karrabi, 2005) and published papers (e.g., Eslami & Fatahi, 2008) have focused on this significant construct while again they have not specifically focused on the effects of gender and years of teaching experience on Iranian EFL teachers' efficacy perceptions.
Given the context-specific nature of the construct of self-efficacy (Bandura, 1995), the dearth of research in this significant area in Iran motivated the researcher to conduct the present study to contribute to the available knowledge about the place of teacher efficacy in ELT, in general, and EFL education in Iran, in particular.More precisely, this study was conducted in order to explore possible effects of teaching experience and gender as well as their interaction effect on Iranian EFL teachers' self-efficacy, in general, and its different dimensions as conceptualized by Dellinger, Bobbett, Dianne and Chad (2008).The research questions directing this study were as follows.
(1) Do years of teaching experience have any significant effect on Iranian EFL teachers' self-efficacy and its components, namely accommodating individual differences (AID), maintaining positive classroom climate (PCC), monitoring and feedback for learning (MFL), and managing learning routines (MLR)?
(2) Does gender have any significant effect on Iranian EFL teachers' self-efficacy and its components?
(3) Do gender and years of teaching experience have any significant interaction effect on Iranian EFL teachers' self-efficacy and its components?

Participants
The participants of the present study were 180 Iranian EFL teachers (90 male and 90 female).Their years of teaching experience ranged from three months to 30 years.94 of the sample had been teaching for at least three and a half years and 86 had been teaching for at most three years.They had been teaching EFL for at least 12 hours a week since they had started their career and they had, at least, a bachelor's degree in English Language and Literature or English Translation.Through convenience sampling, the participants were approached at some private language schools in Tehran, Mashhad and Isfahan as well as through personal contacts.

Instrumentation
Several measures of teacher self-efficacy have been developed so far (Dellinger et al., 2008;Henson, 2001;Tschannen-Moran et al., 1998;Tschannen-Moran & Woolfolk Hoy, 2001) and are widely used in many studies some of which were reported above.However, Dellinger et al. (2008) questioned the adequacy of most of these scales as measures of teacher self-efficacy beliefs and argued: First, the measure should clearly, and accurately, reflect the meaning of self-efficacy.Second, the measure must assess teachers' self-efficacy beliefs in the context in which the beliefs are formed.Third, the specific tasks selected for the measure should be meaningful (p.756).
Based on these assumptions, they tried to develop their own scale which, they believe, has certain advantages over the previously developed instruments.For example, they argue that they have developed it based on a conceptualization of self-efficacy that is firmly grounded in Social Cognitive Theory.Thus, they have tried to avoid confusing the construct of self-efficacy with other related and relatively more stable self-constructs such as self-esteem, locus of control, self-concept, and outcome expectancy.Also, they have taken the context-specific nature of efficacy beliefs into consideration in development of the items.
Their questionnaire which is called TEBS-Self (Teacher Efficacy Beliefs Scale-Self) includes 31 items that are assessed along a four-point Likert Scale (1=Weak, 2=Moderate, 3=Strong, 4=Very strong).Two boxes were added to the questionnaire asking respondents to show their years of teaching experience and gender (See Appendix).

Procedures
To collect data, TEBS-Self (Dellinger et al., 2008) was distributed among 220 respondents via e-mail, face to face, and through asking managers of three language schools in Tehran, Mashhad and Isfahan for help.On average, it took each respondent around 20 minutes to read and answer the items.From almost 220 questionnaires that were administered, 190 were returned out of which 10 proved to be unusable due to too many unanswered items or explicitly careless or perfunctory styles of answering, manifested, for example, in just checking one point on the Likert Scale for all items.

Statistical procedure
The participants' answers to the items included in the questionnaire were fed into the SPSS for analysis.Cronbach-α was used for measuring the internal consistency reliability of the self-efficacy instrument and its components in the context of the present study.Following descriptive analysis of the data, regression analysis was run to answer the above-mentioned three research questions.

Results
To make sure the data collected through TEBS-Self can be relied upon, internal consistency reliability of the scale and its components were calculated through Cronbach's Alpha.It was found that the scale enjoyed a reliability of .97.Reliabilities of the components of the scale were also found to be high.More precisely, accommodating individual differences (AID), positive classroom climate (PCC), monitoring and feedback for learning (MFL), and managing learning routines (MLR) were estimated to have reliabilities of 0.91, 0.80, 0.92, and 0.82 respectively.
Based on their years of teaching experience, the participants were divided into two groups, those who had taught up to three years and those who had taught at least three and a half years.Since being experienced or inexperienced is a highly subjective matter, the researcher chose to refer to them as Group 1 (G1) and Group 2 (G2), respectively, instead of novice and experienced.In this section of the study, the results of the descriptive statistics and regression for exploring the effects of gender and experience as well as the interaction effect of gender and experience on self efficacy, in general, and its components are reported and discussed.
A point regarding how the analysis was done on teacher self-efficacy as a general factor and its components is in order.Obviously, all 31 items were factored into all analyses based on teacher self-efficacy as a general factor.However, to do the analyses based on its components, the researcher had to choose one of the factor structures of TEBS-Self reported in Dellinger et al. (2008).More precisely, Dellinger et al. (2008) is an outcome of three different studies done on TEBS-Self, namely Bobbett (2001), Dellinger (2001), andOlivier (2000), which have yielded three different factor structures, each composed of some of the items.The researcher chose to follow that of Dellinger (Dellinger, 2001) for three major reasons.Compared to Bobbett (2001) and Olivier (2000), the construct validation done in this study seemed to be more explicitly predicated on the distinctions between teacher efficacy, i.e. realized efficacy in the practice of teaching, and teacher self-efficacy beliefs.Also, for item selection, she was stricter with the magnitude of acceptable structure coefficients as she factored in items with structure coefficients greater than or equal to .40 while Bobbett (2001) and Olivier (2000) had kept those with coefficients greater than or equal to .33.Finally, in addition to scree plots, overall percentage of variation accounted for by the factor solution, and magnitude of eigenvalues, theoretical judgment was used for factor selection.The main four components of the resultant factor structure are Accommodating individual differences (AID) (items 1, 2, 12, 13 and 27), Positive classroom climate (PCC) (items 8, 9 and 31), Monitoring and feedback for learning (MFL) (items 16, 17, 18 and 22), and Managing learning routines (MLR) (items 3, 4 and 5).

Descriptive statistics
Table 1 shows that the self-efficacy mean score of teachers with teaching experience of less than or equal to three years (G1) was 28.93 and that of teachers with teaching experience of more than or equal to three and a half years (G2) was 71.81.Also, the standard deviations were estimated to be 10.58 for G1 and 5.62 for G2.Male teachers and female teachers' mean scores on self-efficacy beliefs were 47.38 (s.d.=23.12) and 55.27 (s.d.=22.39)respectively.A close observation of the results across the two genders and the two experience-based categories shows that male teachers in G1 and G2 were found to have the mean scores of 26.13 (s.d.=8.69) and 69.54 (s.d.=6.3) on their self-efficacy beliefs, respectively, and the mean scores of G1 and G2 female teachers on their self-efficacy beliefs were estimated to be 32.15 (s.d.=11.69) and 73.76 (s.d.=4.1), respectively.

Regression analysis
The researcher had intended to use two-way ANOVA to measure the significance of observed differences among teachers across the two genders and the two experience groups.However, since the sample sizes in the four groups across the two variables of gender and experience were unequal, the analysis of variance would become complex.Therefore, in order to test main effects and interaction effects, the researcher decided to use the regression formulation of ANOVA.The only difference is that "a reduced regression model needs to be fitted explicitly for each test of factor interactions and main effects because of the lack of orthogonality" (Kutner, Nachtsheim, Neter, & Li, 2005, p. 953).
Gender, experience and teacher self-efficacy as a whole Running a regression analysis of the data, the researcher came up with one full model as well as three reduced models showing the significance of the effect of interaction between gender and experience, gender per se, and experience per se on teacher self-efficacy as a whole (Table 2).
To explore the significance of the interaction effect of gender and experience on teachers' self-efficacy perceptions, the following test of F statistic was used.
Using the figures form the first reduced model in Table 2, i.e. the interaction effect of gender and experience (G*E), F was calculated as follows.

176
Comparing the test statistic, i.e. .597,and the critical value of F (F (.95, 1,176) =6.68), the researcher found that, since the observed value of F (.597) is less than the critical value (6.68) at .05 level of significance, the interaction effect of gender and experience on teachers' self-efficacy is not significant.
The same procedure was followed to analyze the effects of each of the variables of gender (G) and experience (E) on teachers' efficacy beliefs.

=1262.785
Following the same reasoning, since the observed values of F, i.e. 18.114 and 1262.785,exceed the critical value (F (.95, 1,176) =6.68) at .05 level of significance, both gender and experience seem to have statistically significant effects on teachers' self-efficacy.

Gender, experience and teacher self-efficacy components
The same procedure was gone through for exploring the significance of the effect of interaction between gender and experience, gender per se, and experience per se on the four components of teacher self-efficacy, i.e.AID, PCC, MFL, and MLR.
As discussed above, one of the components of TEBS-Self in Dellinger's (2001) factor structure is accommodating individual differences (AID) which subsumes items 1, 2, 12, 13 and 27 of the questionnaire.Using the test of F statistic, the researcher came up with the following observed F values for the interaction effect of gender and experience, the effect of gender, and the effect of experience on teacher self-efficacy respectively.

=718.440
Comparing these three statistics with the critical value of F (F (.95, 1,176) =6.68) at .05 level of significance, the researcher found that the interaction effect of gender and experience on teachers' self-efficacy is again statistically insignificant while gender and experience each seem to have significant effects on teachers' efficacy perceptions.These results are very similar to those found for self-efficacy as a whole.
The other component of TEBS-Self, based on Dellinger's (2001) factor solution, is positive classroom climate (PCC) which consists of items 8, 9 and 31.
Based on the following F statistic tests, it was found that the interaction effect of experience and gender ( <6.68) and the effect of gender ( <6.68) on self-efficacy were insignificant, and the only variable which was found to have a significant effect on teacher self-efficacy is experience ( 6.68).

=391.951
The next component of Dellinger's (2001) factor structure of TEBS-Self analyzed here was monitoring and feedback for learning (MFL) which composes of items 16, 17, 18 and 22.
Similar to the analysis of AID as well as that of the whole questionnaire, it was found that the interaction effect experience and gender on self-efficacy of teachers ( <6.68) is not significant while the effects of each variable of gender ( 6.68) and experience ( 6.68) are.

=726.145
The last component of TEBS-SELF whose data was analyzed in this section is managing learning routines (MLR) which includes items 3, 4 and 5.
Similar to most of the results reported above, the analysis of data of items in MLR showed that the interaction effect of experience and gender on self-efficacy of teachers ( <6.68) the effect of gender ( 6.68) are not significant, while experience ( 6.68) seems to have a significant influence on self-efficacy of EFL teachers.

Discussion
In this section, a more conceptual analysis of the results reported above is presented.First, the observed differences between the participants in G1 (≤3 years of teaching) and G2 (≥3.5 years of teaching) and, then, the differences between the two genders in terms of their efficacy beliefs will be discussed.
Based on descriptive statistics reported in Table 1, G1 were less self-efficacious (28.93) than G2 (71.81).Regression analysis also showed that the observed difference between efficacy scores of G1 and G2 is significant with G2 identified as more self-efficacious.Since this study was not intended to explore why such a difference exists, the researcher cannot present any empirically-based reasons.However, the literature shows that this observation could be because teachers in their early years of teaching are faced with new or unexpected teaching situations and, thus, tend to experience more stress than experienced teachers (Tschannen-Moran et al., 1998).Besides, they often enter the profession with high hopes about the kind of impact they will be able to have on students' lives, while they often encounter a reality shock when they realize that bringing about change is more difficult than they had thought (Weinstein, 1988).These undesirable mastery-related experiences influence their efficacy perceptions to varying degrees.
Based on descriptive statistics in Table 1, the female teachers (55.27) seem to be more self-efficacious than the male teachers (47.38).Regression analysis also showed that the difference in perceived efficacy across the two genders is significant except for two components of self-efficacy, i.e. positive classroom climate and managing learning routines.However, the results in this study should be treated with caution since, as shown in Table 1, i.e. descriptive statistics, 44% of the female teachers are in G1 and 56% of them are in G2, whereas 51% of the male participants are in G1 and 49% of them are in G2.Moreover, the average of female teachers' years of teaching experience was estimated to be 6.05 while that of the male teachers' was 5.47.Although this difference turned out to be statistically insignificant (sig=.381),these figures are still meaningful since they show that the female teachers who participated in this study had taught for a longer period than had the male teachers.As discussed in the introduction, Cheung (2006) had come up with similar results.That is, in that study also female teachers were shown to be more self-efficacious while they had also taught for more years than the male participants.Finally, the interaction effect of gender and experience on teachers' self-efficacy perceptions turned out to be insignificant in all analyses, which means that gender does not necessarily moderate the relationship between experience and self-efficacy in any meaningful way.Therefore, one possible explanation of why the female participants were shown to have more positive perceptions of their efficacy than the male participants is the female participants' more years of teaching experience which, as shown in this study, significantly contributes to difference in EFL teachers' perceived efficacy.

Concluding remarks
"Teachers' self-efficacy is a little idea with big impact" (Tschannen- Moran & Hoy, 2006. P. 337).Through collecting data from Iranian EFL teachers regarding their self-efficacy, gender and years of teaching experience, this study concluded that experience significantly affects Iranian EFL teachers' self-efficacy, and gender doesn't seem to moderate this causal relationship in any significant way.Also, it was found that female teachers have significantly more positive perceptions of their efficacy, in general, and efficacy for monitoring and feedback for learning and accommodating individual differences, in particular.However, since the female teachers in this study had generally taught for a longer period of time than the male participants, this finding needs to be treated with caution.
Further research seems to be necessary for establishing the ways in which EFL teachers with different lengths of teaching experience differ from each other in terms of their efficacy beliefs.In addition, research on reasons behind these observed differences in their self-efficacy can make great contributions to the literature on teacher self-efficacy in the context of EFL instruction in Iran.Regarding the way gender might affect EFL teachers' self-efficacy, given the difference observed in years of teaching experience between the male and the female participants, more research needs to be done to determine if the evidence from this study is an accurate representation of how gender may contribute to perceived efficacy of teachers.For example, further research needs to keep the factor of years of teaching experience constant between male and female second language teachers and explore how different their efficacy beliefs turn out to be.

Table 2 .
Fits of Full and Reduced Regression Models (Dependent variable: teacher self-efficacy)

Table 3 .
Fits of Full and Reduced Regression Models (Dependent variable: AID)

Table 4 .
Fits of Full and Reduced Regression Models (Dependent variable: PCC)

Table 5 .
Fits of Full and Reduced Regression Models (Dependent variable: MFL)

Table 6 .
Fits of Full and Reduced Regression Models (Dependent variable: MLR)