Analysis of students ’ school engagement in terms of friendship and teachers ’ behavior

The main goal of this study is to analyze “Students’ school engagement in terms of friendship and teacher behaviors”. The population of the study is all students at 4 th grade within the boundaries of İstanbul Metropolitan Municipality. The data of the study were gathered from 2066 students from 62 schools. “School Engagement Scale”, “Peer Attachment Scale” and “Perceived Teacher Behaviors Scale” were used apart from an information form developed by the researcher. Multiple Regression Analysis, T-test and one-way analysis of variance were used to analyze the data. Results show that; students’ relationship with their friends predicts school engagement most. Female students have a higher level of school engagement. Students who do cultural and sports activities, those who get support from counseling service and those who do not have a problem with transportation have a higher level of school engagement. Students in crowded classrooms have a lower level of school engagement. It is reckoned that this study is significant in terms of presenting negative variables affecting students’ engagement in school which has an important role in their personal and social lives and in terms of taking necessary precautions.


INTRODUCTION
School is a corporal and psychological environment in which education is carried out. Schools are institutions where students are shaped and made to prosper in cognitive, physical, psychomotor, emotional and moral terms and where students can socialize apart from their families. In this sense, school takes an important place in an individual's life in terms of healthy progress of children and acquisition of the qualities mentioned above.
School engagement is defined differently by many scientists. However, the term "school engagement" is first used by Hirschi. According to Hirschi (1969), the basic determiner of crime factor is low engagement with social values and norms. In this regard, crime and violence will decrease if school engagement increases (cited in Kızmaz, 2006). Finn (1993) defines school engagement as students' feeling of belonging to school and adopting its aims. According to Maddox and Prinz (2003), school engagement is students' relation with their schools and various aspects of their academic experience. Broadly, school engagement means students' relation with their schools, school personnel and the ideals which are aimed to be achieved. Cernkovich and Giordano (1992) approached school engagement from different perspectives based on Hirschi's theory of social control and stated that school engagement is shaped under four basic subcategories (Hirschi,1969). They are school engagement, engagement to school personnel, taking responsibilities at school and joining activities at school. Newmann (1992) defines school engagement as students' psychological investment on learning. Finn et al. (1993) define school engagement as students' feeling of belonging to school and adopting its aims. According to Willms (2003, 18), school engagement has two aspects: affective aspect and behavioral aspect. Affective aspect emphasizes students' feeling of belonging to school, how much they value educational outcomes and how they identify themselves with these outcomes. Behavioral aspect emphasizes students' attendance to academic and non-academic activities. According to Jimerson et al. (2003), school engagement has three aspects. They are cognitive, affective and behavioral aspects. According to this, affective aspect involves students' feelings about school, teachers and their peers. Cognitive aspect involves students' believes and perceptions about school, students and their peers. Behavioral aspect involves students' attendance to extracurricular activities like sport, dance and theatre and so their observable behaviors and performance.
According to the definitions above, school engagement is students' feeling of belonging to school, adopting the aims and values of their schools, and consequently, adopting their duties and responsibilities about school and intense wish for academic success. There are a lot of factors which affect school engagement such as family, school environment, friendship and teacher behaviors. Friendship and teacher behaviors are mostly emphasized in this research.
According to Connell and Klem (2004), students feel more satisfied with school and develop more positive academic attitude and values in schools where there are supportive and caring group of friends.
According to Silins and Mulford (2004, 449), teachers' school engagement will increase students' school engagement and attendance. Similarly, Skinner and Bellmont (1993) assert that teachers' attendance is highly correlated with students' school engagement and students' high level of engagement yields teachers' high level of school engagement. Research reveals that teachers' social support behaviors towards students have positive impact on affective and behavioral aspects of school engagement (Brewster and Bowen, 2004). According to Schlechty (2001), the activities that can bind students to each other increase school engagement. Shin et al. (2007) however found out that positive peer norms and peer support enhance students' school engagement. In a study conducted abroad by Leonard (2000), it has been presented that students who regard school as an unpleasant place have negative feelings towards it in all aspects of quality of life at school. In a study conducted by Green et al. (???), students who are accepted by their peers have a higher level of school engagement (cited in Osterman, 2000, 52). Murray and Greenberg (2000) have tried to find the correlation between students' relationship with their teachers and school engagement. The results of the study indicate that there is a positive correlation between student-teacher interaction and school engagement. In a study by Marks (2000) it has been found that students who are supported by their friends and teachers have a high level of school engagement. As it can be seen in the explanations above, friendship and teachers' attitude contribute to developing school engagement among students.
Why is it important to have school engagement? There are a lot of answers to this question. There is no doubt it is expected that school engagement has positive effects on students' acquisition of good attitude and healthy progress. Manlove (1998) states that high degrees of school engagement lead to decrease in school drop-out and teenage pregnancy. On the other hand, the state of lack of school engagement leads to isolation, alienation and separation. Studies conducted demonstrate that school engagement has a lot of benefits. For instance, students who have positive attitude towards school and actively participate in activities at school are more likely to finish school and become independent learners. There is positive correlation between school engagement and other educational results like increase in academic success, school attendance, and high academic expectations. However, it has been confirmed that lack of school engagement among teenagers brings in serious results such as substance abuse, teenage pregnancy, getting involved in crime and school drop-out (Finn and Rock, 1997;Caraway et al., 2003;Thompson, 2005).
Absence at school, not participating in academic activities and as a result of this; failure, damaging school property, littering the school inside and outside etc. cause loss of money and energy spent on these students. There may be many reasons for this situation, yet it is considered to be related to school disengagement. This study is believed to be a guiding light for taking necessary precautions about school engagement / disengagement. Besides, this study is very important in that there is limited research on school engagement in Turkey and this study has a huge sample which is 2066. The fact that İstanbul, where the study has been carried out, is a cosmopolitan city shows that it can represent the whole Turkey more or less. Therefore, these facts make this study important.
Based on the explanations above, the main issue of this study is "Analyzing students' school engagement in terms friendship and teacher behaviors".
In accordance with this purpose, answers to the questions below have been sought: 1) Do the points on Peer Attachment Scale and Teacher Behaviors Scale meaningfully predict the level of students' school engagement? 2) Do the points on Students' School Engagement Scale yield meaningful results with respect to students' gender, their mothers' being alive or dead, the distance between their houses and schools and the number of students in their classrooms? 3) Do the points on Students' School Engagement Scale yield meaningful results with respect to students' participation in cultural and sports activities, getting support from the counseling service, efforts to keep the class clean and grades on their transcripts?

METHOD
The main goal of this study is to analyze students' school engagement in terms of teacher behaviour and friendship. In this regard, it is possible to assert that this study is an example of correlational model.

Population and sample
The population of this study is 4 th grade students of all primary schools within the boundaries of Istanbul Metropolitan Municipality. The data were gathered from 2066 students, 30 schools from Asian side and 32 schools from European side. The study was carried out at schools where İstanbul Sabahattin Zaim University MA students work. Due to the fact that only 4 th grade students were included in the study, homogeneous sampling, a type of purposive sampling, was used. In this study 1032 (50%) of the students are females and 1034 (50%) are males. 1868 (92%) of the students in the sample are in the age range of 9-10, while 163 (8%) of them are in the age range of 11-12.

Instrument
"School Engagement Scale" was used in this study to present students' school engagement. This scale was developed by Cernkovich and Giardona and was adapted to Turkish by Can in 2008. SES (School Engagement Scale) is a five point Likert scale. Each positive item is worth 5 points if is replied as "totally agree". They are worth 1 point if they are replied as "totally disagree". Other replies in between the two above are worth 4, 3 and 2 points. Negative items (1 st , 2 nd , 5 th , 8 th , 9 th , 10 th items) are graded negatively. The results of the first confirmatory factor analysis of School Engagement Scale are as follows: Normed Fit Index (NFI) = .80, Comparative Fit Index (CFI) = .86 and Root Mean Square Error of Approximation (RMSEA) = 0.076. Hu and Bentler (1999) indicated the critical values of the standards used for cohesiveness of items as CFI>.90; RMSEA<.05 and NFI>.90. In accordance with these findings, correction indexes yielded by LISREL were examined. The most important four corrections yielded on chisquare value were done and analyzed, and so the results are as follows: Normed Fit Index (NFI)= .93 , Comparative Fit Index (CFI)= .94 and Root Mean Square Error of Approximation (RMSEA)= .049. According to the compatibility statistics, the model used was confirmed. As a result, the findings attained from factor analysis are proof for the construct validity of School Engagement Scale.
The reliability co-efficient of the scale is .55 for "School Engagement" sub dimension, .71 for "Teacher Attachment" sub Erdoğdu 2059 dimension, .76 for "School Responsibility" sub dimension, .64 for "Attending Activities at School" sub dimension, .72 for "Family Communication" sub dimension, .60 for "Perceived Opportunities" sub dimension. As a result of the calculations, there is no factor below .50 (Can, 2008). The alpha values of School Engagement Scale were recalculated and it is concluded that the scale is proved to be reliable in this study, as well [Alpha value for school engagement subscale= .63, alpha value for teacher attachment subscale= .70, alpha value for school responsibility subscale= .58, alpha value for attending activities at school subscale= .55, alpha value for family communication subscale= .60, alpha value for perceived opportunities= .58]. "Peer Attachment Scale" was developed by Armsden and Greenberg and adapted to Turkish by Hortaçsu and Oral (1991). Peer Attachment Scale is a five point Likert scale made up of 25 items. The range of total point is 25-125. High points attained on the scale designate strong peer attachment. In Löker's study of validity-reliability (1999), Croanbach-Alpha reliability co-efficient of this scale was found to be .92. The correlation of the items on the scale varies between .51 and .75. Akkapulu (2005) found internal consistency reliability co-efficient as .91, two halves reliability coefficient as .95, test-retest reliability as .71. Total item point coefficient varies between .9 and .67 (cited in Saçar, 2007:40). The alpha value of Peer Attachment Scale was recalculated in this study and the reliability co-efficient is .90. The correlations of the items vary between .63 and .88.
"Perceived Teacher Behaviours Scale" was developed by Erdoğdu (2006). The scale is a three point Likert scale made up of options: Yes, No, Sometimes. In total, there are 29 questions on the scale. The high points on the scale show that the teacher is regarded as democratic. Test-retest reliability co-efficient has been found to be .70; two halves reliability co-efficient has been found to be .79. The item analysis of the scale and average distinctiveness value were calculated as distinctiveness value (rjx=38). At the end of the reliability study for this scale, it was observed that the scale is made up of two factors. The alpha value of democratic attitude, one of the factors, is .79 and explains 44.59% of the variance. The alpha value of "Authoritative Attitude", the other factor, is .75 and explains 17.77% of the variance (Erdogdu, 2006).

Statistical analysis
SPSS Windows 16.0 was used to analyze the data gathered for the study. Multiple Linear Regression Analysis Technique was used for predicting the dependent variable with the help of two or more independent variables. Furthermore, One Way ANOVA was used for the meaningful variance between the averages of multiple independent variables. Independent Samples T-Test was used for the meaningful variance of the averages between two variables.

Procedure
The data of the study were gathered with the help of teachers studying Educational Administration and Supervision at İstanbul Sabahattin Zaim University as MA students. As the aim of the study is to measure how students perceive their teachers' behaviors (authoritative, democratic), only volunteers from 4 th grade primary school participated in this study. Applying the questionnaire and assessment instruments took almost two course hours.

FINDINGS
The descriptive statistics of points that students got from

School Engagement, Friendship and Perceived Teachers
Behavior scales is shown in Table 1.
As it can be seen in Table 1, the average of students' points on School Engagement Scale was calculated as 80.43. There are 19 items on the scale. The rate of average and item number is (80.43/19=4.22). It can be inferred that students' school engagement is high in general.
Are the average points of students' school engagement, friendship and teacher behavior really predictive of friendship, teacher behavior and school engagement together?
Multiple regression analysis was used to answer this question. That means analyzing multiple correlations. In this context, the correlation between independent variables (friendship and teacher behavior) is below .70. Since variance inflation factor is below 10, there is multiple correlation between the variables. When homoscedasticity of independent variables is considered, the data are found to range close to normal. As a result, the data from the statistics show that multiple regression analysis can be used for the study variables (Table 2).
Friendship and perceived teacher behavior are highly correlated with students' school engagement (R=0.49, R²=0.24, p<0.001). The two variables mentioned above explain only 24% of the variance in school engagement level.
According to standardized regression co-efficient (ß), the relative order of importance of predictor variables on students' school engagement is as follows: friendship and teacher behavior. When the results of t-test about the meaningfulness of regression co-efficient are analyzed, friendship and teacher behavior are significant predictors on school engagement.
The results of t-test on the relation between students' school engagement and gender, parents being alive or not, the distance of the school to their home and students' perception of crowd in their classrooms are shown in Table 3.
As it can be seen in Table 3, students' school engagement shows significant variance according to gender [t(2060) =2.78, p<.01 . Female students' level of school engagement (x =81.22) is higher than that of male students (x =79.66). In other words, female students have a higher level of school engagement compared to male students.
Students' level of school engagement varies according to their mothers being alive or not [t(2053) =2.50, p<.05 . Students whose mothers passed away (x =80.55) have a higher level of school engagement than those whose mothers are alive (x =77.63). In other words, students whose mothers have passed away are more engaged in school. There is no significant relation between fathers' being alive or not and school engagement.
Students' school engagement indicates significant variance according to the distance between students' school and home [t(2053) =2.32, p<.05 . Students who live close to school (x =80.68) have higher levels of school engagement than those who live far from school (x =78.72).
Students' school engagement shows significant variance according to the number of students in the classroom [t(2050) =3.70, p<.00 . Students who perceive their class as crowded (x =79.61) have lower levels of school engagement compared to those who do not (x =81.77) ( Table 4).
According to analysis results, points on "School Engagement" scale show meaningful variance as to whether students attend sports activities or not [f (2-2050) =20.13, p<.01]. The level of school engagement for  students who always attend sport events (x =81.97) is meaningfully higher than those who sometimes or never attend them (x =77.49). LSD test has been used to determine if there is significant variance within the groups and variance has been found to be significant. According to results of the analysis, the points on "School Engagement" scale show meaningful variance as to whether students attend cultural activities or not [f(2-2035)=11.82, p<.01]. The level of school engagement for students who always attend cultural activities (x =81.74) is meaningfully higher than those who sometimes or never attend them (x =78.39). LSD test has been used to determine if there is significant variance within the groups and variance has been found to be significant.
According to analysis results, the points on "School Engagement" scale show meaningful variance as to whether students get support from counseling service or not [f(2-2060) = 38.73, p < .01]. The level of school engagement for students who always get support from counseling service is (x =81.95) meaningfully higher than those who sometimes or never benefit from it (x =79.74). LSD test has been used to determine if there is significant variance within the groups and it has been determined that only students who always get support from counseling service show variance from those who sometimes or never benefit from it.
According to analysis results, the points on "School Engagement" scale show meaningful variance as to students' school marks [f (2-2060) =38.73, p<.01 . The level of school engagement for students who were "very good" students previous year (x =81.87) is meaningfully higher than those who were "good" (x =78.86) and who were "mediocre" (x =73.89). LSD test was used to determine if there is significant variance within the groups and variance has been found to be significant.
According to analysis results, the points on "School Engagement" scale show meaningful variance as to students' effort to keep the class clean [f(2-2053)=19.51, p<.01]. The level of school engagement for students who always attend classroom cleaning (x =81.54) is meaningfully higher than those who sometimes (x =78.22) and those who never attend it (x =73.10). LSD test was used to determine if there is significant variance within the groups and all variances have been found to be significant.

DISCUSSION
Results of this study show that students' level of school engagement is generally high. According to the analysis results, being on good terms with friends is more predictive of high school engagement than democratic behavior of teachers. According to the data, it is possible to claim that friendship has a positive effect on school engagement in terms of spending time with their friends inside and outside school more frequently, taking part in activities together, being in the same age range. Furthermore, it is thought that regarding the teacher as an authority figure makes friendship more important in the sense of students' school engagement.
The results of the study show that students who perceive their teacher as positive and emotionally supportive have a higher level of school engagement. Apart from this, having healthy friendship and teachers' positive attitude enable students to have positive feelings towards school. Another finding of the present study is female students' level of school engagement is meaningfully higher than male students. Studies conducted in Turkey and abroad have yielded similar results. In studies conducted abroad, female students have higher level of school engagement compared to male students (Jenkins, 1995;Simons et al., 1999;Osterman, 2000;McNeely et al., 2002). However, in a study conducted by Eith (2005) and Bonny et al. (2000), male students' level of school engagement was found to be higher than that of female students.
Studies conducted in Turkey indicate that female students' level of school engagement is higher than male students' (Arastaman, 2006;Can, 2008). A study conducted by Ocak (2004) presented that male students' sense of belonging to school is lower than that of female students'.
There may be many reasons for female students to be more engaged in school. But especially these reasons are emphasized: female students take more responsibilities at school, are more motivated for success and have more sense of belonging because of cultural values. Besides, female students feel more engaged in school because they perceive school as social surroundings and they perceive it as a way to get out of house as females have more problems about going out compared to males and they are worried about not being sent to school if they are unsuccessful.
One of the most important findings of the study is that students whose mothers passed away have a higher level of school engagement than those whose mothers are alive. It can be interpreted as the fact that students perceive their teachers as their mothers and as their confidant and feel lonely at home. There is no other research on the correlation between school engagement and loss of mother either in Turkey or abroad. However, there is no meaningful correlation between having both parents alive or loss of father and school engagement. No meaningful correlation between school engagement and having divorced parents in a study conducted by Can (2008).
Another finding of the present study is perceiving their house is close to school increases students' school engagement. Transportation is expensive and difficult in big municipalities like Istanbul for both students and parents. In this sense, students will have a higher level of school engagement if they can go to school easily. There is, also, no research on students' perception of the distance between their house and school.
Another finding of the present study is there is a correlation between students' school engagement and their perception of crowded classrooms. According to the research findings, the degree of students' school engagement shows meaningful difference with respect to their perception of crowd in the class. According to this, the students who perceive that their class is crowded have a low level of school engagement. There is no research on the correlation between the features of class and school engagement. However, Lindsay (1982) found that high school students at lower grades have fewer absentees and join extracurricular activities more. Also, high school students who tend to drop out join intraclass activities less (cited from Arastaman, 2006). According to Purkey and Smith (1983, 445), students do not learn in noisy, distracting and unsafe classrooms. When the rules are applied equally and consistently, not only behavioral problems decrease but also students' school engagement increase.
According to research results, students who attend sports, social and cultural activities have been found to have a high level of school engagement. They get bored especially on academic subjects. On the other hand, sports and social activities which take an important place in their physical, social and psychomotor development make students pleased; therefore, school engagement increases.
A study by Sang Min and Sondra (2005) reveals that there is a positive correlation between female students' school engagement and their attendance in processes about school life from secondary school until high school. According to Finn (1989), students' identification with school affects their school engagement; so, this increases their academic success, school attendance and their sense of valuing school. Identification with school is related to their attendance in activities relevant to school (Fredricks et al., 2004). For Finn, students who attend extracurricular social activities regularly develop school engagement and school becomes an indispensable part of their lives.
According to research findings, the level of school engagement for students who always get support from counselling service has been found to be favorable. It is, also, remarkable in the sense that 722 students (35%) never get support from counselling service and it has been detected that their psychological development is affected negatively. It is considered that counselling service should work more actively in terms of students' personal and emotional development. There is no research on school engagement and benefiting from counselling service. Lee and Smith (1999) mentioned the significance of factors affecting school engagement like supplying students with their school materials, solving their catering problems, providing security and making the counselling service active apart from teacher support (cited from Ataman, 2006).
The degree of students' school engagement shows meaningful difference according to their academic achievement. In this context, students who are successful are more affiliated to school than those who are less successful. Definitely, successful students have more sense of responsibility. Studies applied abroad indicate similar results (Finn et al., 1997;Fredrick et al., 2004). According to a study by Camp, successful students attend various activities and so they are engaged in school more (cited in Lee and Smith, 1995).

Suggestions
It is not possible to crown education activities with Erdoğdu 2063 success unless the affiliation level of students, the building blocks of education, is on top and unless it is kept on top. Thus, depending on research results, implementations to foster friendship, more social and sports activities and attendance to them are considered to be important. Physical circumstances of schools, students' easy transportation to school and familial conditions are the factors which affect school engagement. In this context, those concerned should be sensitive about the factors mentioned above. This research has been conducted in schools within the borders of Istanbul Metropolitan Municipality. Broader research should be done in this field and it should raise awareness for both students and for those concerned.