Confirmatory Factor Analysis of Self-efficacy Instrument among Special Education Teachers

Self-efficacy is pivotal in education because of its influence on an individual’s personality and attitudes. It stems from the person’s belief in his or her capabilities to achieve a desired performance level and is especially important to special education teachers who deal with continual teaching challenges. Thus, to contribute to the increasingly challenging field of special education in the 21 century, this study aims to measure the validity and reliability of a self-efficacy instrument among teachers of the Integrated Special Education Program for Learning Difficulties (ISEPLD). Three subconstructs were measured, namely 1) student engagement, 2) instructional strategies, and 3) classroom management. AMOS software program version 18 was used for the data analysis and values from Comparative Fit Index, Tucker Lewis Index and RMSEA were used to retain and correlate items. An instrument with three subconstructs containing 15 items of nine-point scale was tested in this study. The instrument was administered to 500 participants across Malaysia using the proportional stratified random sampling and by the means of Confirmatory Factor Analysis, the study has confirmed that the data corresponded to the model. Therefore, it is proposed that the 15-item instrument developed from the three subconstructs can be used in measuring self-efficacy among teachers of ISEPLD in Malaysia.


Introduction
In the new Education Development Plan for Malaysia 2013-2025, teaching profession was discussed as the first area of focus, a move which can be translated into the government's intent to transform and revitalize teaching as a major profession of choice.It signifies a national priority, indicating how important the education field is to the country.As students form the future human capital for the nation, their personal growth has become an issue of concern in the Malaysian education system.In meeting this demand, it is imperative that the system is orchestrated by teachers who believe in themselves and their personal abilities and teaching abilities because teachers' quality has a direct positive relationship to their students' quality (McEwen, 2008).These qualities are of utmost importance, particularly to special education teachers who handle students with learning difficulties on daily basis.On this note, teachers of the Integrated Special Education Program for Learning Difficulties (ISEPLD) in Malaysia are expected to be equipped with high self-efficacy beliefs to function effectively and contribute towards the preparation of quality human capital for the country.
To survive the current changes and innovation in the 21 st century education field, educators need to be acquainted with high self-efficacy beliefs (Dibapile, 2011;Magno & Sembrano, 2008).This group of educators often display a character of high self-assurance, ending up being individuals who are confident in their skills and are able to implement the required changes and innovations in teaching (Harvey, 2009;Ng et al., 2010).Given its critical impact on a person's effort (Bandura, 1993), self-efficacy should be given top priority in the education system.In the research world, efficacy has been studied to understand its effect on teachers' job performance which in turn affects their work commitment (Billingsley, 2004;Korthagen, 2005).

Problem Statement
The self-efficacy theory by Bandura (1997) serves as the theoretical framework in this research project to guide the construction of the self-efficacy instrument for ISEPLD teachers.Self-efficacy is defined as a concept of individuals' personal beliefs to control various situations that occur in life (Bandura, 1997).It refers to the ability to perform a task successfully according to prescribed standards.Confidence in one's ability to perform tasks efficiently and effectively will influence the person's 1) actions and behavior control, 2) choices in approaching a situation and environment; and 3) perseverance in completing a particular task.The self-efficacy theory explains a person's levels of confidence in performing a particular task (Bandura, 1997).In this theory, the concept of reciprocal determinism is discussed at length.It is defined as the interconnected cycle between the personal factor (cognitive) with environmental factor and behaviors, influencing one another and enabling individuals to produce the required behavior based on prior knowledge in responding to their environment.Bandura (1997) defines teachers' self-efficacy as self-confidence in their ability to teach effectively in warranting student achievement.The concept of efficacy is closely related to the intensity of belief and confidence that influence a person's behavior in dealing with a particular situation.It is the teachers' judgment in their capability to carry out a teaching responsibility that encourages planning and coordinating the required behavior in meeting education goals.High efficacy facilitates teachers to be more determined to carry out teaching task by diversifying teaching activities and not to give up easily (Tschannen-Moran & Hoy, 2001).In the context of special education, teachers handling students with learning difficulties need to believe in their ability so that they can continue to understand, manage and teach students with disabilities who have different types and levels of learning difficulties.This research project, with its aim to measure constructs in a self-efficacy instrument, has applied the confirmatory factor analysis (CFA), a multivariate statistical procedure to identify items to be included in the instrument.CFA is claimed to be appropriate to use especially when researchers have little knowledge of the underlying structure of the latent variables (Byrne, 2010) and it is meant to examine the construct validation instrument to decide how well the constructs have managed to explain the variables (Stevens & Zvoch, 2007).

Literature Review
Self-efficacy significantly affects people's behavior, motivation and ultimately, their success or failure.Although the original focus of teacher efficacy was on the locus of control, Bandura's application of social cognitive theory to social learning theory expanded it to explore other factors that bear possible influence on teacher efficacy.Bandura (1997)  A growing body of research on teacher efficacy has suggested positive relationships with teacher performance.Teachers with a strong sense of efficacy have been reported to 1) display a higher degree of flexibility to change (Ross & Bruce, 2007) and more effective pedagogical approach as well as innovation (Chu, 2011;Pan, 2014;Ross, 2007;Thoonen et al., 2011), 2) portray better stress management skills (Ross & Gray, 2006;Tschannen-Moran & Gareis, 2004) and have higher job satisfaction level (Eichinger, 2000;Lazuras, 2006;Viel-Ruma, 2010), 3) have increasing expectations on students' academic abilities (Deemer, 2004;Ross & Bruce, 2007), and 4) prevail against workplace challenges (Ross & Gray, 2006;Skaalvik & Skaalvik, 2010;Tschannen-Moran & Gareis, 2004;Tschannen-Moran & McMaster, 2009).
In a more recent study by Berg and Smith (2014), primary school pre-service teachers from New Zealand, England, and Malaysia were studied.The participants were required to complete the "Teachers' Sense of Efficacy" (long form) (Tschannen-Moran & Woolfolk Hoy, 2001) and the "Concerns about Teaching Scale" (Smith, Corkery, Buckley & Calvert, 2012) at the onset of their second year of teaching programs.The findings suggested that culture and context are imperative in shaping the participants' concerns related to their teaching and efficacy beliefs.The Malaysian pre-service teachers were found to have the comparatively lowest efficacy beliefs which could be raised by their concerns about a large class size (50 students), high parental expectations of the students' success, and subject-based instead of class-based interaction with the students.

Research Objectives
This study aimed to determine the validity and reliability of an instrument related to the self-efficacy model by testing whether the items have measured the three constructs of special education teachers' self-efficacy, namely student engagement, instructional strategies, and classroom management using CFA.These constructs were measured with reference to the three-dimensional teachers' self-efficacy by Tschannen-Moran et al. (1998).

Research Design
This was a quantitative study of a survey design.Data were collected via paper-based questionnaires that were distributed to selected schools in every state across Malaysia.The sample size was determined based on the table by Krejcie and Morgan (1970) and Hair et al. (2010).

Participant
Five hundred teachers of the Integrated Special Education Program for Learning Difficulties (ISEPLD) throughout Malaysia were identified using the proportional stratified random sampling.The sampling was determined by the number of participants from 13 states and two Federal Territories (WP) of Kuala Lumpur and Putrajaya as obtained from the Ministry of Education Malaysia.The significantly different numbers of participants in each state have led to the use of proportional stratified random sampling technique.The selection of research sample from each sub-set (stratum) was drawn based on the percentage of the participants (Lim, 2007).Table 1 lists the number of participants of each state, its percentage from the population, and the number of sample to be used in the study.In the second stage, using the simple randomization technique, the questionnaires were mailed to schools with these criteria: a) national primary schools (government-owned and operated) with ISEPLD, and b) schools with a minimum of seven teachers who run the program, excluding substitute and temporary teachers.

Instrument
This study adapted a self-efficacy instrument of nine-point Likert scale ranging from "not sure" to "very confident" from a study by Tschannen-Moran & Hoy (2001) with their permission.The instrument was used to measure 3 subconstructs with 15 items as follows: (i) student engagement (5 items), (ii) instructional strategies (5 items) and (iii) classroom management (5 items).

Validity and Reliability of Instruments
Validity refers to the ability of an instrument to measure what is supposed to be measured in a construct.There are three types of validity required for each measurement model, namely the convergent validity, construct validity and discriminant validity (Zainudin, 2014).
Reliability, on the other hand, is the extent of how reliable the said measurement model is in measuring the intended latent construct.The assessment of reliability in a measurement model could be performed using the following criteria (Zainudin, 2014)      www.ccsen

Second ste
The fitnes items are items redu     12, it can be concluded that the discriminant validity for all three constructs are achieved (Fornell & Larcker, 1981).(Zainudin, 2012).Descriptive analysis found that all the items are in the range of +3 to + / -3 and multivariate kurtosis value is 20.483, as shown in Table 13.Therefore, it can be concluded that the data sample is normal.

Discussion
On the basis of calculations, it can be perceived that Validity is a measure of consistency of questioned items of an instrument.So, the questioned items are strongly believed to be able to measure what is to be measured.Validity requirement was achieved through the following processes: Convergent Validity: AVE > 0.50: Refer to the following table (Table 9).Construct Validity: All fitness indexes for the models meet the required level.The discriminant validity for all constructs is achieved when the square root of is higher than the values of the correlation between the respective construct.By referring to Table 12, it can be concluded that the discriminant validity for all three constructs have been achieved.
Figure 7 shows a model of the three factors used to measure the Teacher Self-Efficacy, namely classroom management, student engagement, and instructional strategies.Items for each factor is considered to measure only the latent variables respectively.Standardized loading for the three constructs were established to measure the Teacher Self-Efficacy and it shows that the factor loading is high (more than 0.6).It suggests that all three constructs formed have successfully measured Teacher Self-Efficacy.The results of the analysis of the overall fit of the model is considered fit and may be accepted based on the indicators suggested by Hair et al. (2010), Golob (2003) and Zainudin (2014).
The Reliability requirement was achieved through the following process; Internal Reliability: Cronbach Alpha > 0.70: Refer to  11 shows that the data is fit with the model.Confirmation factor analysis (CFA) has confirmed that the significant item for the three subconstructs, consisting of 5 items of classroom management, 5 items of student engagement and 5 items of instructional strategies.When the Chi Square test Degrees of Freedom (CMIN / df) show values between 1 and 5 (Chisq/df = 2.749) and RMSEA values for the hypothesis that the model is smaller than 0:08 (RMSEA = 0.059), they indicate where significant hypothetical model reserved is commensurate with the study of data collection, rather than study participants.These results are consistent with the analysis of the compatibility index TLI and CFI in excess of the value of 0.90 (0.974 and 0.979).The results of the analysis of the overall fit of a model for Teacher Self-Efficacy is fit and may be accepted based on the indicators suggested by Hair et al. (2010), Golob (2003) and Zainudin (2014).It is proposed that the 15-item-instrument based on three factors model can be used to measure self-efficacy of teachers of ISEPLD in Malaysia.
suggests a framework with four sources of self-efficacy, namely mastery experience, vicarious experience, social (or verbal) persuasion, and physiological factors.This framework is a key component in the definition of teacher efficacy and has become the focus of study of professional development program evaluations (Tschannen-Moran & McMaster, 2009; Tschannen-Moran & Woolfolk Hoy, 2007).It also affords a critical part to Tschannen-Moran et al. (1998) teacher self-efficacy model.In 2009, the framework was used by Tschannen-Moran and McMaster to study teachers' implementation of a new instructional framework for reading class by elementary teachers. Fig

Table 1 .
Number of ISEPLD teachers in each state :

Table 13 .
Assessment of normality distribution for items Table 8 (use SPSS).Composite Reliability: CR > 0.60: Refer to Table 8 (using a formula).Average tVariance Extracted: AVE > 0.50: Refer to Table 8 (using a formula).By referring to Table 11, it can be concluded that the Internal Reliability, Composite Reliability and Average Variance Extracted for all three constructs are achieved.Table