Self-reported competency, knowledge and practices of teachers teaching primary children with autism in government schools of West Malaysia: a cross-sectional study

Background: Disability brings challenges and demands in the family and society which last for a long time.  Children that are affected by disability are often kept at home, without access to opportunities other children might have, and education is one of the most effective ways to break the cycle of discrimination and poverty. Malaysia is committed to achieving the Sustainable Development Goals, and teacher quality issues are among the Ministry of Education's focuses to ensure a successful journey for pupils with disabilities. In this study, we evaluated the competency, knowledge and implementation practices of teachers serving government schools in West Malaysia regarding teaching children with autism spectrum disorder (ASD). Methods: 832 primary teachers from different states of West Malaysia participated in a web-based survey that assessed self-reported competency, knowledge and implementation practices with regards to managing children with autism spectrum disorder. Results: Respondents rated themselves as quite competent, and believed that they were knowledgeable regarding ASD and could implement ASD tasks.  Frequency of training was a consistent predictor of competency. Respondents who had never attended a training course had significantly lower self-competency, lower level of knowledge, and lower implementation ratings than those who had Open Peer Review


Introduction
Disability brings challenges and demands in the family and society which last for a long time. The financial burden associated with accessing health, education and social services is responsible for the cycle of poverty in society. Education is one of the most effective ways to break the cycle of discrimination and poverty that children with disabilities and their families often face by allowing them a greater degree of independence and better quality of life 1 . Disabled citizens are often underestimated for their sufficient capabilities to contribute to society resulting in not being prioritized. Education for All (EFA), a global movement, which aims to ensure that every child and adult receives basic education of good quality, has realized failure in its objectives because disabled children are not given proper attention. These children are often kept at home, without access to opportunities other children might have 2 . Universal primary education is one of the eight international Millennium Developmental Goals (MDGs). MDGs had been established following the Millennium Summit of the United Nations in 2000, following the adoption of the United Nations Millennium Declaration. It is a commitment made by 191 United Nation States to be achieved by the year 2015 3 . According to the MDGs report in 2015, there has been a significant improvement in terms of universal primary education. The primary school net enrolment rate in the developing regions reached 91% in 2015, up from 83% in 2000. The number of outof-school children of primary school age worldwide has fallen by almost half, to an estimated 57 million in 2015, down from 100 million in 2000. The literacy rate among youth aged 15 to 24 has increased globally from 83% to 91 % between 1990 and 2015 4 . But despite impressive strides, there are more people being left behind, especially the poorest and those disadvantaged because of their sex, age, disability, ethnicity or geographic location 2 . Targeted efforts are required to reach the most vulnerable people. 'Children with disabilities' is one of these groups identified in the report as holding back progress towards education for all. Children with disabilities demand increased attention and often require an education adapted to their needs. Ensuring that all people have an equal chance of education must be at the heart of new goals post 2015.
The Sustainable Developmental Goals (SDGs) were developed to accomplish the unmet MDGs which ended in 2015 5 . One SDG is to "ensure inclusive and equitable quality education and promote lifelong learning opportunities for all" 6 . This inclusive education is aimed at providing special education to students for them to be more successful in social interactions leading to further success in life. However, children with autism require specialized instructions and behaviour support to learn critical skills and improve long term outcomes 7 . The competency of teachers is important in the development of these special children. They should be well competent to ensure the quality of education that these children deserve.
Consequently, this research evaluates the characteristics of teachers teaching children with autism in primary government schools in West Malaysia and their self-reported competency, level of knowledge and teaching practices implemented. The implications of this study are important as the results can impact the qualifications of teachers and guide professional development that will ensure potential development of children with ASD.

Study design and setting
This cross-sectional study, conducted in West Malaysia between June 2017 and January 2018, utilized a web-based survey to determine the level of competency, knowledge and implementation practices of primary government schoolteachers registered under the Program Pendidikan Khas Integrasi (PPKI) and Sekolah Kebangsaan Pendidikan Khas (SKPK). PPKI is a special education program for primary school children with disabilities, which was introduced by the Ministry of Education in 1962 8 . It provides a separate class for disabled individuals organized under mainstream primary schools within integrated setting so they can interact and play with other children of their age. SKPK is a special education school provided by the same ministry, which provides education for primary school children with special needs. At present, there are 22 SKPK all over the country, which offer special education for different groups of disabilities; some exclusively provide education to deaf and mute children, while 12 offer programs for autism enrollees. Due to the scarcity of the available enrollees from SKPK (Table 1), regular government schools offering PPKI programs were also targeted (Table 2) in the present study.

Participants
This survey was completed by teachers specifically teaching children with ASD employed in schools involved in SKPK and PPKI. Inclusion criteria was primary schoolteachers presently working with children with autism during the study period.

Questionnaire
This research adapted the same questionnaire used by Hendricks in the study 'Skill competencies for professionals and paraprofessionals in Virginia supporting individuals with autism across the lifespan' 9 . This questionnaire was used to survey special education teachers employed in public schools in Virginia, USA The Virginia Skill Competencies is a list of guidelines for educators who serve students with autism and it was generated to develop educational standards in autism 10 .
The questionnaire consisted of four sections: Section A, background information, which comprised questions covering the following: age, gender, race, position, state, highest level of education, number of years spent/experience in teaching children with autism, number of ASD students taught in the last five years, number of students in a class, and frequency of special trainings attended; Section B, self-reported competency, which included nine statements on the self-rating of competence in handling a child with autism using a four-point Likert scale; Section C, self-rating of level of knowledge; and Section D, self-rating of level of implementation. These last two sections included statements organized under the following focus areas: Section C consisted of 16 statements on self-rating of level of knowledge and Section D included 25 statements on self-rating of the implementation practices. They used a five-point Likert scale when rating current level of knowledge, where 1 represented 'little knowledge' and 5 indicated 'very knowledgeable'. When rating current level of implementation, 1 represented 'rarely implemented' and 5 indicated 'frequently implemented'.
The questionnaire was translated into the Malay language by a professor teaching language studies. A pilot test was carried out to ensure validity and reliability of the questionnaire. Special Education Teachers at a private therapy center were asked to participate in the pilot survey and comment if the instructions were clear, comprehensive and easy to understand. They were also asked if the confidentiality was appropriately maintained in the questionnaire. Suggestions were collected and integrated into the questionnaires.

Data collection
The link to the online survey was sent to the email address of the State Education Department directors of the 13 states. The researcher followed up using telephone calls to confirm receipt of the email. Jabatan Pendidikan Negeri (State Education Department) officers were assigned to coordinate with the researcher. The online questionnaire was administered, and data collected, using Google Forms.

Data analysis
Data was exported and analyzed using SPSS version 19. Frequencies, means and standard deviations were used to summarize the data. Linear regression procedures were conducted to determine whether certain socio-demographic variables would predict self-competence in ASD management, knowledge of ASD competencies, and knowledge of implementation of ASD competencies. For all tests, the level of significance was set at 0.05.

Ethical considerations
Approval for this study was obtained from the Institutional Review Board of Taylor's University Centre for Research management and the Ministry of Education, Malaysia [KPM.600-3/2/3 Jld.34 11 ; KPM.600-3/3/3 Jld.35 37 ]. Written informed consent was obtained from the participants for participation in the study.

Results
Out of the 7,575 teachers teaching in West Malaysia (Table 3) in both SKPK and PPKI schools, 1073 responded and completed the survey (Table 4). A total of 242 responses were excluded due to missing information. As per Nunnally and Bernstein (1994), a measure is moderately reliable if its Cronbach's alpha is 0.70 or higher 12 . Given this criterion, the three measures (Sections B-D) were all reliable. Mean composites were created for each of the measures; as such, four was the highest possible score. As shown in Table 5, the sample of respondents rated themselves as considerably competent (M = 3.06, SD = 0.38). They also believed that they were knowledgeable regarding ASD competencies (M = 3.28, SD = 0.60) and could implement ASD tasks (M = 3.33, SD = 0.63).
Overall procedure Linear regression procedures were conducted to determine whether certain socio-demographic variables would predict self-competence in ASD management, knowledge of ASD competencies, and knowledge of implementation of ASD competencies. Prior to conducting the regressions, several variables were recoded. First, level of education was recoded into a threecategory variable: diploma, Bachelor's, and graduate degree. Second, class size was recoded into a three-category variable: 1 to 5, 6 to 10, and 11 or more. Lastly, frequency of training was recoded into a four-category variable: never, once, once a year, and more than once a year. Indicator (or dummy) variables were then created, where the first category (i.e., diploma, 1 to 5, and never) served as the reference group. In addition, because the variable, number of ASD students taught, was severely skewed, this variable was transformed using natural log function. Skewness dropped to acceptable limits (i.e., skewness index fell below three, per Kline, 2011) 13 ; thus, this transformed variable was used in subsequent regression procedures.  Assumptions of multivariate normality (assessed through a normal probability plot), linearity, and homoscedasticity (both assessed through a plot of the studentized deleted residuals by the standardized predicted values) were checked; all assumptions were met. The problem of multi-collinearity was also verified via the variables' tolerance values; multi-collinearity was not a problem as all tolerance values were above .20 14 Predictors of self-competence The findings in Table 6 reveal that the number of ASD students taught positively predicted self-competence ratings (β = .11, p = .003); the more ASD students the respondents had taught, the higher their self-competence ratings. Frequency of training also positively predicted self-competence ratings; respondents who had never attended a training course (M = 2.88, SD = .41) had significantly lower self-competency ratings than those who had attended even one course (M = 3.06, SD = .36; p < .001), those who attended a course once per year (M = 3.11, SD = .41; p < .001), and those who took more than one course a year (M = 3.11, SD = .34; p < .001).

Predictors of knowledge
The findings in Table 7 show that the number of ASD students taught positively predicted knowledge of ASD competencies (β = .18, p < .001); the more ASD students the respondents had taught, the higher their knowledge of ASD competencies scores. Level of education also significantly predicted knowledge of ASD competencies; respondents with a teaching diploma had significantly lower knowledge ratings (M = 3.25, SD = .57) than respondents with a graduate degree (M = 3.43, SD = .68; p = .022). Frequency of training also positively predicted knowledge ratings; respondents who had never attended a training course (M = 2.95, SD = .71) had significantly lower knowledge of competency ratings than those who had attended even one course (M = 3.29, SD = .58; p < .001), those who attended a course once per year (M = 3.37, SD = .58; p < .001), and those who took more than one course a year (M = 3.34, SD = .55; p < .001).

Variables n (%)
Class size (number of students)

Predictors of implementation
The findings in Table 8 indicate that number of ASD students taught positively predicted knowledge of implementation (β = .15, p < .001); the more ASD students the respondents had taught, the higher their knowledge of implementation scores. Level of education also significantly predicted knowledge implementation; respondents with a teaching diploma had significantly lower ratings (M = 3.21, SD = .59) than respondents with a graduate degree (M = 3.52, SD = .62; p = .001). Frequency of training also positively predicted implementation ratings; respondents who had never attended a training course (M = 2.99, SD = .77) had significantly lower implementation ratings than those who had attended even one course (M = 3.29, SD = .59; p < .001), those who attended a course once per year (M = 3.45, SD = .59; p < .001), and those who took more than one course a year (M = 3.40, SD = .58; p < .001).

Discussion
Increasing numbers of autistic students in primary schools demand competent and well-trained teachers who can confidently implement efficacious strategies to bring the best possible outcomes. It is crucial to understand the aptitudes of special education teachers as their understanding of autism and its spectrum, strategies to design individual teaching plans and assessment will help the students achieve their potential. This study highlights the self-reported competency, knowledge and implementation practices and its predictors among special education teachers in primary government schools in West Malaysia. Additionally, the findings can serve as a guide to realigning the content of teachers' preparation programs.
In our study, the majority of the participants had a Bachelor's degree (83.3%) and another 8.2% had a Master's degree. The average number of years of experience of teaching was 8.25 years, with an average of around 13 students with ASD, and majority of the teachers were experienced in handling a class size of 1 to 5. A similar profile was reported in studies conducted by Toran et al. and Wei et al. 15,16  Our study revealed that the frequency of training strongly predicted self-reported competency and knowledge, as well as the implementation practices. Toran et al. emphasized the need to improve special education teacher training in order to increase the level of knowledge in autism and effective use of evidence-based teaching strategies by offering hands-on activities. Hendricks also highlighted the need for increased content related to autism and evidence-based practices during pre-service training 9, 15 Scheuermann et al. recommended the use of specialized skills training and provision of technical assistance and support to teachers in the training programs in the United States. Similarly, Litton et al. recommended a oneyear induction program in conjunction with the local school districts' programs with the express purpose of retaining highly qualified teaching personnel 19,20 .
In this study, we also found that the level of education also significantly predicted knowledge of ASD competencies and implementation practices. Respondents with a graduate degree had significantly higher ratings as compared to those with a teaching diploma. This finding highlights the importance of recruiting teachers with a higher level of education in this setting.
The number of students with ASD taught by the respondents was found to be a significant predictor of self-competence and knowledge, as well as the implementation practices. This is of particular interest given the challenges educators face due to the multidimensional nature of the disability. As documented by Hendricks 9 , the wide range of cognitive abilities and verbal skills of learners affect the knowledge and competencies required. Dealing with students with varying capabilities warrant special practices that address communication and social needs and help the learners achieve academic success.
However, in our study the number of years of teaching experience did not predict self-competence or knowledge, which concurs with the finding reported by Alharbi et al. 21 . In contrast, AL Jabery et al. 11 reported statistically significant differences in the level of knowledge among teachers with a teaching experience of five years and more as compared to those with a teaching experience of less than three years

Conclusion
From our results, we conclude that teachers rated themselves as considerably competent. They also believed that they were knowledgeable regarding ASD and could implement ASD tasks. Frequency of training is a consistent predictor of this competency; respondents who had never attended a training course had significantly lower self-competency, lower level of knowledge and lower implementation ratings than those who had attended even one course. Investments in training of these teachers is highly recommended to ensure these students are provided with the quality of education they deserve.

Open Peer Review
Sample size is adequate however sampling technique is not that good.

○
Finally, I think this study is interesting and deserves indexing.

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound? Yes

If applicable, is the statistical analysis and its interpretation appropriate? Yes
Are all the source data underlying the results available to ensure full reproducibility? Yes Are the conclusions drawn adequately supported by the results? Yes (or numbers) in Results than to write long introduction.

Introduction
The last sentence contains only abbreviation (ASD).

Methods
Please make clearer about SKPK and Table 1 and . In text the authors states that 22 SKPK (12 autism) is present and in the Table 1 total number of SKPK is 53. Does it mean that total number of schools enrolled in this research is 53 SKPK + 1973 PKI? This part is somehow confusing due to specificity of the country educational organization. Authors might make it more simpler. It is important to emphasize the number of schools and number of teachers in each type of school that are enrolled in the research. Table 1 and 2 might be more appropriate as supplement since distribution of schools in the country is not so important information to the international readers, but rather specific for Malaysia. The authors state that they use adapted questionnaire. What adaption did the authors made? How many participants were used for the pilot test for validity and reliability? The authors should provide at least some information about that validation process of the questionnaire, for example (Cronbach's alpha of scales, or at least interval of alpha levels for scales B-D 0.888-0.900).

Results
The author should present the percentage of total response rate and rate of teachers enrolled in the study after excluding missing. It is more informative then just counts. In discussion the authors must explain this low percentage.
In Table 4, do the authors need to present each state? Does it change their discussion? The table is big and it would be easier to collapse some categories (states) to "Other". The Overall procedure is confusing. Why did the authors made all numerical variables to binary or ordered. Linear regression is assumed to have numerical and categorical variables (binary or ordered) as independent variables. Did the authors made any model with original variables and concluded that this models are better? There is no univariable analysis. How did the authors choose these independent variables for the modeling? Why did not use Gender or others variables presented in Table 1?
The authors might try (not necessarily) to merge Tables 6, 7 and 8 into one table with the predictors in rows (as it is) and three outcomes in column. It might provide better insight into relationship between independent predictor as specific outcome. Table 5 has alpha levels. Is Cronbach's alpha=0.98 somehow too large? Some authors suggests that.

Discussion
The authors repeat numbers from results (no need for that). First the response rate should be explained. It seems that is low and do the authors think that the sample is representative. In some paragraphs the authors leave out "self perceived" or "self-reported". It is necessarily to emphasize that their knowledge and competency is self perceived, not completely objective (if I understood the questionnaire completely).

Is the work clearly and accurately presented and does it cite the current literature? Yes
Is the study design appropriate and is the work technically sound?