Attitudes and Self‐Efficacy of Romanian Primary School Teachers towards Including Children with Special Educational Needs in Regular Classrooms

1 PhD, Universitatea Al. I. Cuza, Faculty of Psychology and Education Sciences, Iasi, Romania, luciana.frumos@uaic.ro Abstract: In the last years, as a consequence of inclusive policies, the number of children with special educational needs in Romanian primary classrooms has increased. In this context primary school teachers face numerous challenges when teaching children with special educational needs in their classrooms. It is assumed that regular education teachers are the critical agents in this process of inclusion. Their attitudes toward the inclusion of children with special educational needs in regular classrooms is especially important for the success of inclusionary efforts. Teacher self-efficacy has also an essential role when addressing diverse learners. In this study we examine the relationship between teacher selfefficacy and teacher attitudes toward the inclusive education. Data were collected from 126 Romanian in-service primary school teachers that have in their classrooms students with special educational needs. There is a significant positive correlation between primary school teachers’ perceived selfefficacy and their attitudes towards inclusive education. Younger primary school teachers have more positive attitudes towards children with special educational needs and inclusive education. Those primary school teachers that have attended more special education courses appear to have less positive attitudes and are less willing to accept students with disabilities in regular classrooms. Practical implications of teachers’ attitudes and self-efficacy in inclusive classrooms are discussed.


Introduction
Policies and practices of recent years in Romania reflect the global tendency to pay increasing attention to the inclusion of children with special needs in mainstream schools. Inclusive education promotes access, participation and outcomes for all students. Addressing students with special needs can be challenging and the success of inclusion can be influenced by diverse factors. It is believed that teacher attitudes towards children with special educational needs is one of them, being significant for the success of inclusive education (Norwich, 1994).
Just placing children with special needs in mainstream schools is not enough. Different strategies should be found to facilitate them the access to the curriculum and also take into account the social aspects of inclusion. This puts a lot of pressure on schools and especially on educators that need to find an answer to the challenges arising with the increased diversity of their students. Teacher must assume new roles (Avramidis, Bayliss, & Burden, 2000) and find new strategies to respond to the needs of all students in their mixed abilities classrooms.
Inclusion is defined as the process of educating children with special needs into the general educational environment, as opposed to placing them in special education settings. According to Sharma, Forlin and Loreman (2008), in the inclusive classrooms educators need to feel comfortable interacting with persons with disabilities and embrace the philosophy of inclusion. However, recent studies show that not all teachers are ready to accept the idea of inclusion. After reviewing 26 studies, De Boer, Pijl and Minnaert (2011) found out that the majority of teachers hold neutral or negative attitudes towards the inclusion of pupils with special needs in regular primary education and that there were no studies that reported clear positive results. After analyzing a large body of research, Avramidis & Norwich (2002) concluded that although there is a great evidence of positive attitudes, there is no total acceptance of inclusion.
Teachers' attitudes towards the inclusion of children with special educational needs are connected and positively influenced by their sense of self-efficacy as indicated by some studies (Meijer & Foster, 1988;Sharma, Loreman, & Forlin, 2011;Soodak, Podell, & Lehman, 1998), those teachers that have more confidence in their ability to teach will have more positive attitudes towards inclusive education. Having positive attitudes towards supporting students with special educational needs is important, but experience and concrete actions are also important (Van Laarhoven, Munk, Lynch, Bosma, & Rouse, 2007). Teachers must have the necessary skills for effectively supporting the children during class activities, for example to identify efficient inclusive strategies. Jordan et al. (2009) report research suggesting that in elementary classrooms effective teaching skills are effective for all students, with and without special needs. Several other studies on primary school teacher efficacy found that those primary school teachers that are generally effective, have excellent classroom and time management, are also effective with student that have special educational needs (Jordan, Lindsay, & Stanovich, 1997;McGhie-Richmond, Underwood, & Jordan, 2007).
In Romania, legislation and policies concerning children with special educational needs developed significantly in the period that followed the communist regime. In recent years, the concept of inclusive education is more present and to some extent has contributed to the transformation of school values, climate and practices. This change over the last three decades had mixed implications (Ghergut, 2011): more open attitudes towards teaching children with special educational needs in mainstream schools and, on the other hand, a greater emphasis on academic competition and selection for elite schools. This generates difficulties accessing the curriculum (Ghergut, 2010;Unianu, 2013;Vrasmas, 2014) that creates a barrier for the children with special educational needs, is a challenge for teachers to overcome it, and most probably the future national educational policies will have to find an answer. The number of children enrolled in special education schools decreased constantly and there is a constant increase of students included in educational support programs. General education teachers sometimes have the support of a special education teacher, but in most of the Romanian schools that is not yet the case (Horga et al. 2016). It can be concluded that not all schools embrace the philosophy of inclusive education and there are still some steps to take in order to implement inclusive practices and develop a culture of inclusion.
There are not many studies on the issue of Romanian teacher's attitudes towards inclusion. According to Gherguţ (2010) over 70% of 768 Romanian teachers questioned are theoretically in favor of inclusive education, but when asked if they were willing to act on that belief and to implement an inclusive program in their school, for example to use inclusive practices and strategies, the number of teachers that positively answered decreased to 50.5% from which only 22% would fully support such an approach. This denotes a gap between declarative support for inclusion and the personal involvement and willingness to take actions leading to the inclusion of children with special educational needs in regular classrooms.
This study attempts to explore the attitudes of primary school teachers that have children with special educational needs in their classrooms towards inclusion by answering the following research questions: 1. What is the attitude of primary school teachers that have children with special educational needs in their classrooms towards inclusion? 2. What is the effect of the demographic variables on teacher's perceived self-efficacy and their attitudes towards inclusion? 3. Does teacher perceived self-efficacy correlate with their attitudes towards inclusive education?

Participants
Primary school teachers from the north-east region of Romania that reported having at least one student with special educational needs in their classroom were asked to answer the questionnaire. Of all 285 questionnaires that were distributed, 126 were returned, at a rate of 44.2%. Demographic information of the sample are shown in Table 1.

Measuring instruments
The Teacher Sense of Efficacy Scale TSES (Tschannen-Moran & Hoy, 2001) was used to determine the perceived self-efficacy of primary school teachers. The long form has 24 items that group in three moderately correlated factors: Efficacy in student engagement (ESE), Efficacy in Instructional Strategies (EIS) and Efficacy in Classroom Management (ECM). A Likert type scale is used, ranging from 1 to 9, 1 being "not at all" and 9 "a great deal". In the study of Tschannen-Moran and Hoy (2001) the following reports for alpha cronbach were found for TSES and the three subscales EIS, ECM and ESE: Cronbach alpha = 0.94, 0.87, 0.91 and 0.90 respectively.
The instrument used for measuring teachers' attitudes towards inclusive education was The Multidimensional Attitudes towards Inclusive Education Scale (MATIES) developed by Mahat (2008). The scale addresses each of the affective, cognitive and behavioral components of teachers' attitudes towards inclusion and demonstrated acceptable content, construct, criterion and convergent validity (Ewing, Monsen, & Kielblock, 2018). A Likert type scale is used, ranging from 1 to 6, 1 being "strongly disagree" and 6 "strongly agree". According to the author (Mahat, 2008) the MATIES has good internal consistency for the three subscales Cognitive, Affective and Behavioral: Cronbach alpha = 0.77, 0.78 and 0.91 respectively.
To obtain a measure of scale reliability, internal consistency coefficients were calculated for the TSES and MATIES scales. Our data results show a good and an excellent internal consistency for TSES 0.986 and its three subscales ESE 0.960, EIS 0.967, ECM 0.962 and also for the three subscales of MATIES: Cognitive 0.728, Affective 0.817, and Behavioral 0.940.

Teacher attitudes
One of the research questions was: What are the attitudes towards inclusion of primary school teachers' that have in their classrooms children with special educational needs?
A one-sample t-test was run to determine whether primary school teachers have attitudes that are more in the favor of inclusive education. In all three situations the Cognitive, Affective and Behavioral components of attitudes towards inclusion are significantly higher than the theoretical average of the scale (3.5). As can be seen in Table 2 all the means are significantly higher on all three subscales. We can conclude that Romanian primary school teachers are in favor of inclusion. They have positive perceptions and beliefs about children with special educational needs, positive feelings and emotions associated with inclusive education, have positive intentions and are willing to get involved in working with children in inclusive classrooms.

Teacher perceived self-efficacy
A one-sample t-test was run to determine whether primary school teachers that have in their classrooms students with special educational needs, have a higher perceived teacher efficacy in inclusive practices compared to the theoretical mean of the scale (which is 5.0) and compared to the means reported by Tschannen-Moran and Hoy (2001) on TSES and its subscales ESE, EIS and ECM which that were used as test values are: 7.1, 7.3, 7.3 and 6.7 respectively. As can be seen from Table 3, the means of all subscales are significantly lower than those reported by the authors (Tschannen-Moran & Hoy, 2001). However, when using as a test value, the theoretical mean of the scale (which is 5.0) the means are significantly higher for Teacher Sense of Efficacy Scale t (125) = 7.31, p < 0.001 and for all its subscales. At the Efficacy in Student Engagement (ESE) subscale t (125) = 6.26, p < 0.001 the mean is significantly higher, suggesting a good perceived self-efficacy in engaging all students in learning activities. At the Efficacy in Instructional Strategies (EIS) subscale t (125) = 8.97, p < 0.001 the mean is significantly higher, suggesting a good perceived self-efficacy in responding to instructional challenges for the needs of all students. At the Efficacy in Classroom Management (ECM) scale t (125) = 6.18, p < 0.001 results indicate a good perceived efficacy in classroom management and controlling disruptive behavior. Romanian primary school teachers that work with children with special educational needs, seem to have a positive image on their self-efficacy and their ability to respond to the needs of the students from their classrooms.

Teachers' demographic factors influencing perceived self-efficacy and attitudes towards inclusive education
One of the research questions was: What is the effect of the demographic factors on primary school teachers' perceived self-efficacy and their attitudes towards inclusion? In order to explore the impact of teachers' demographic factors on their self-efficacy a one-way between-groups analysis of variance was conducted. Significant results were found only in case of age. Participants were divided into three age groups: Group 1: age 20-29; Group 2: age 30-39; Group 3: age 40-50. There were statistically significant differences in all TSES subscales. There was a statistically significant difference at the p = .05 level in TSES (F (2, 123) = 6.47, p = .002) ESE (F (2, 123) = 7.89, p = .001) EIS (F (2, 123) = 5.05, p = .008) ECM (F (2, 123) = 6.09, p = .003). In all four situations, as can be seen from Table 4, post-hoc comparisons using the Tukey test indicated that the mean score for Group 1: age 20-29 was significantly higher than Group 3: age 40-50. Group 2: age 30-39 did not differ significantly from the other two. These results suggest that younger teachers have a more positive perceived selfefficacy than older ones for all the subscales, and that self-efficacy seems to decrease with age. A one-way between-groups analysis of variance was conducted to explore the impact of teachers' age on their attitudes towards inclusive education. Participants were divided into three groups: Group 1: age 20-29; Group 2: age 30-39; Group 3: age 40-50. There were statistically significant differences in all MATIES subscales for the three age groups.
There was a statistically significant difference at the p = .005 level in MATIES Cognitive scale scores for the three groups (F (2, 123) = 11.96, p < .001). Post-hoc comparisons using the Tukey test (Table 5) indicated that the mean score for Group 1 was significantly higher than the mean of Groups 2 and 3. Between Group 2 and Group 3 there is no significant difference. These results suggest that teachers from the first group of age have more positive beliefs on inclusion than the teachers from the other two groups.
There was a statistically significant difference at the p = .005 level in MATIES Affective scale scores for the three groups (F (2, 123) = 16.48, p < .001). Post-hoc comparisons using the Tukey test (Table 5) indicated that the mean score for Group 2 have significantly less positive attitudes than those from groups 1 and 3.
There was a statistically significant difference at the p = .005 level in MATIES Behavioral scale scores for the three groups (F (2, 123) = 12.04, p < .001). Post-hoc comparisons using the Tukey test (Table 5) indicated that the mean score for Group 3 age 40-50 was significantly lower than Group 1 and Group 2. Between Group 1 and Group 2 there is no significant difference. Older teachers are less behavioral engaged in activities specific to inclusive education. Results show that, for all the dimensions, younger teachers hold more positive attitudes towards inclusion, they have more positive beliefs, feelings and intention behaviors in accordance with the philosophy and practice of inclusive education.
A one-way between-groups analysis of variance was conducted to explore the impact of qualification on primary school teachers' attitudes towards inclusive education. Participants were divided into three groups according to their qualifications: Group 1: Teacher Training College; Group 2: Bachelor's Degree; Group 3: Master's Degree. There were statistically significant differences at the p = .001 level in MATIES Cognitive and Affective scales scores (F (2, 123) = 6.28, p = .003) and (F (2, 123) = 13.08, p < .001) respectively. For both subscales, post-hoc comparisons using the Tukey test (Table 6) indicated that the mean score for Group 3 Master's Degree was significantly lower than the mean of Groups 1 and 2. Between Group 1 and Group 2 there is no significant difference. The beliefs and the feelings of teachers that have a Master's degree are less in the favor of inclusion than of the other two groups. A one-way between-groups analysis of variance was conducted to explore the impact of teachers' number of special education courses on their attitudes towards inclusive education. Participants were divided into three groups according to the numbers of courses they took on inclusive / special education: Group 1: none; Group 2: one to three courses; Group 3: four or more courses. There was a statistically significant difference at the p < .05 level only in MATIES Cognitive scale (F (2, 123) = 3.778, p = .026). Posthoc comparisons using the Tukey test (Table 7) indicated that the mean score for Group 3 was significantly lower than the means of groups 1 and 2. Group 1 and Group 2 did not differ significantly from each other. Results show that primary school teachers that followed four or more courses on subjects related to students with special educational needs have beliefs less favorable to inclusive education than those with less or no such courses.
A one-way between-groups analysis of variance was conducted to explore the impact of teachers' previous experience with special educational needs students on their attitudes towards inclusive education. Participants were divided into three groups according to degree of experience reported: Group 1: No experience; Group 2: Some experience; Group 3: A lot of experience. There were statistically significant differences at the p < .05 level only in MATIES Cognitive and Behavioral scales scores for the three groups (F (2, 123) = 3.778, p = .026).
For the Cognitive scale Post-hoc comparisons using the Tukey test (Table 8) indicated that the mean score for Group 3: A lot of experience (M= 3.61, SD = 0.51) was significantly different from Group 1: No experience (M = 4.28, SD= 0.87) and from Group 2: Some experience (M = 4.23, SD= 0.80). Group 1 and Group 2 did not differ significantly from each other. Teachers having more experience working with children with special educational needs have less positive beliefs comparing with those having less experience or no experience at all.
For the Behavioral scale Post-hoc comparisons using the Tukey test (Table 8) indicated that the mean score for Group 1: No experience (M= 5.01, SD = 0.59) was significantly different from Group 2: Some experience (M = 4.00, SD= 1.12) and Group 3: A lot of experience (M = 4.42, SD= 0.49). Groups 2 and 3 did not differ significantly from each other. Teachers having less experience with children with special educational needs are more willingly to do the necessary actions in order to implement the inclusive education practices in their classrooms.

Correlations between teachers' self-efficacy and their attitudes towards inclusive education
Results of the Pearson correlation indicated that there was a significant positive association between teacher efficacy and its three subscales and the all three subscales of teacher attitudes towards inclusion as can be seen in Table 9. The two-tailed test of significance indicated strong positive correlations between TSES and all its subscales and the Behavioral subscale of the MATIES, medium positive correlations with the Cognitive subscale and weak positive correlations with the Affective subscale. Primary school teachers that have a greater perceived self-efficacy tend to have a more positive attitude towards inclusive education. Squaring the correlation coefficients indicated that teacher self-efficacy can explain 45% of the variance in the Behavioral component of attitudes towards inclusive education and 18% of the variance in the Cognitive component of attitudes towards inclusive education.

Multiple linear modeling
For the Cognitive component, the model proposed (Table 10) containing the variables EIS, Age, Qualification and Previous experience with special educational needs children has an adjusted R2 = 0.422, which means that 42.2% of the attitudes' cognitive component variance can be explained by the contribution of the four variables mentioned above. For the Affective component, the model proposed (Table 11), containing the variables Age, Qualification and EIE, has an adjusted R2 = 0.313, which means that 31.3% of the attitudes' affective component variance can be explained by the contribution of the three variables mentioned above. For the Behavioral component, the model proposed (Table 12), containing the variables EIS, Age and Previous experience with children with special educational needs, has an adjusted R2 = 0.523, which means that 52.3% of the attitudes' affective component variance can be explained by the contribution of the three variables mentioned above. Multiple regression analysis were conducted to predict teacher attitudes towards inclusive education from self-efficacy and demographic variables. In the models corresponding for each of attitudes' dimensions (cognitive, affective, behavioral), besides demographic variables such as Age, Qualification or Previous experience with children with special educational needs, are two subscales of TSES. The most important seems to be Efficacy in Instructional Strategies that can be found in two of the models (Cognitive and Behavioral), and can explain in both cases most of the variance. The other TSES scale, Efficacy in Student Engagement is less significant and has limited relevance only in the Affective dimension.

Discussion
The main aim of the study was to explore the profile of Romanian primary school teachers' attitudes towards inclusive education. Results indicated high levels of self-efficacy and positive attitudes regarding inclusion among respondents. There is also a significant positive correlation between primary school teachers' perceived self-efficacy and their attitudes towards inclusive education. These results are consistent with other studies suggesting that teacher self-efficacy is an important predictor of positive attitudes towards inclusive education (Savolainen et al., 2012;Weisel & Dror 2006). Attitudes and perceptions of control (self-efficacy) are, along with subjective norms, important factors for prediction of intention and behavior (Ajzen & Fishbein, 2005: 196). From the perspective of inclusive education and teacher training it is important to maximize those factors that support the implementation of inclusive practices in the classroom.
According to Eagly and Chaiken (2007: 592) attitudes can be formed or expressed on the basis of any one of the three types of the processes (cognitive, affective and behavioral) or some mix of them. In our study results show that all three attitudes' components seem to have a significant contribution in the structure of attitudes towards inclusive education.
The cognitive dimension of attitudes towards inclusion was the only one influenced by all the demographic factors. Teacher beliefs, as core cognitive dimension of attitudes can significantly influence educational practice (Pajares, 1992). Jordan et al. (2009) consider teacher beliefs about disability as part of a broader set of epistemological beliefs (beliefs about the nature of ability/disability and about the nature of knowledge) with important consequences on teaching and learning. Our results are consistent with this perspective, the self-efficacy dimension variable Efficacy in using instructional strategies is explaining most of the variance in the cognitive dimension of attitudes towards inclusion.
The affective dimension of attitudes towards inclusion represents feelings and emotions associated with children with special educational needs. To this component of attitudes towards inclusion the primary school teachers obtained the highest mean (4.40 from maximum 5.00), proving the crucial role of positive feelings in inclusive contexts. According to Soodak et al. (1998), the most beneficial teachers' affective responses in inclusive settings is being receptive and calm to the challenges of children with special educational needs.
The willingness to act in favor of children with disabilities was measured using the behavioral scale. According to Yan & Sin (2013: 80) intention of behavior and perceived behavior control have direct effects on actual behavior. In this respect, some skills are required and teachers need to master effective teaching practices in order to maximize their instructional efficacy in inclusive classrooms (McGhie-Richmond et al., 2007). Teachers that are generally effective will be also effective with children with special educational needs (Jordan et al., 1997).
Demographic variables seems to have an important relationship with attitudes towards inclusion, but not with self-efficacy. Therefore, younger primary school teachers have more positive attitudes towards children with special educational needs and inclusive education. These results are similar to other studies. Soodak et al. (1998) report that teachers with more years of experience are significantly more hostile than inexperienced teachers. As inclusive education is not promoted for a very long time in Romania, we can assume that the attitudes of older teachers are based on the old model of segregation rather than on inclusion.
Primary school teachers that have a higher qualification hold less positive attitudes towards inclusion compared to teachers that have a lower qualification. This is an unexpected result, probably due to the prevalence of theoretical over practical training on the higher levels of qualification. Avramidis et al. (2000) suggest that it is not necessarily a matter of negative attitudes, but rather that teachers may not see the practical solutions and enhancing teachers' self-efficacy can be a matter of teacher training in using appropriate educational strategies. Also, the primary school teachers that reported having more experience with children with special educational needs seem to have, however, less positive attitudes towards inclusion. Having more experience, increase the probability to meet children with more complex situations, and according to Avramidis and Norwich (2002, p. 142) in case of more severe learning needs and behavioral difficulties, teachers will hold rather negative attitudes towards inclusion. Teachers working with children with disabilities need to have support from special education teachers and should carefully plan and support contact with children with disabilities and this could result in positive changes in their attitudes (Sharma et al., 2008).
Another unexpected result is that primary school teachers that have attended more special education courses appear to have less positive attitudes and are less willing to accept students with disabilities in regular classrooms. This result is not in line with those of other studies (Avramidis & Kalyva, 2007;Subban & Sharma, 2006;Avramidis et al., 2000) that found a positive relation between teacher training in special education and teacher attitudes towards inclusive education. A possible explanation of this inconsistencies is offered by Jordan et al. (1997) proposition of pathognomonic versus interventionist perspective towards children with special educational needs. Those teacher that hold an interventionist perspective believe that most students can benefit from instruction in the general education classroom if appropriate instruction is implemented, while for teachers that hold a pathognomonic perspective, the student with disabilities is eligible for special education services. It is possible that special education curriculum designed in a rather implicit pathognomonic perspective will consolidate the moderate attitude towards inclusive education.
Results of this study should be interpreted cautiously and some limitations should be taken into account. First, data were collected using self-reported scales, as indicators of the teacher's actual behavior and respondents might have answered in a socially desirable way. Direct observations in the classroom would be recommended for future studies. Second, the sample was selected from the north-east region of Romania and some psychosocial and cultural variables may be involved. Further studies with samples from different regions could be able to clarify the implications of such variables on inclusive education.
Traditional model of segregation implied efforts focused on educational services in special schools. The transition from integration to inclusion involved a shift from need to rights of SEN children (Pirrie, Head, Brna, 2006), imposing rethinking of training (initial and continuous) of both regular and special education teachers. Thus, the concept of inclusive education is getting closer to human rights and social justice (Florian, 2008;Florian & Spratt, 2013). In a broader perspective, Artiles and Kozleski (2007) emphasize the cultural and historical legacies that are particular to each community. For Romania it seems particularly important to take into account these issues because for many years segregation and special schools for children with disabilities were the only solution. Positive attitudes towards inclusive education still need to be developed, and one way of doing this is by better preparing future school teachers in both pre-service and inservice training programs designed on inclusive values and principles.