In schools, which are the basic production units of the education system, interpersonal relations and interaction have a significant role in both information transfer and make the students gain positive behaviors during education and training process. ?ntimidation, which affects the  relationship  particularly  between  the  teachers  and managers, other teachers and parents in schools, has frequently been on the agenda in recent years. Intimi-dation, its forms and levels experienced in public and private schools have negative effects on both the employees and the school culture. The word“mobbing” originates in Latin and it means psychological violence, oppression, siege, and harassment, abuse or afflict. The most frequently used terms by Turkish researchers for the term “mobbing” are: “intimidation” (Aras, 2012; Akkar, 2010; Asunakutlu and Safran, 2006; Cemalog?lu, 2007; Gündüz et al., 2008; Ertürk, 2011; Toker, 2006), “psychological harassment” (Palaz et al., 2007; T?naz, 2008), “emotional harassment” (Uzunc?ars??l? and Yolog?lu, 2007), and “psychological violence” (Ayd?n, 2009; Tutar, 2004). Since it is one of the most frequently adopted terms, the term “intimidation” was adopted to refer to the term “mobbing” in this study. The concept of “harassment”, which is used as a correspondent of the term “mobbing” in the dictionary of Turkish Language Association, is described as intimidating, ostracizing and humiliating one through targeting him, systematically blocking his work and causing him to be anxious in workplaces, in schools and in other similar communities (TDK, 2014). In general, mobbing is described as any kind of maltreatment, threat, violence and humiliation which is practiced systematically by superiors, subordinates and equivalents of employees (Duffy, 2012; Leyman, 1996; Mikkelsen and Einarsen, 2002; Tutar, 2004:11).

Hubert and Veldhoven (2001) made a research to determine in which sectors intimidation is most frequently experienced; and they suggested that aggressive behaviors constituting intimidation are most commonly experienced in the fields of industry and education, in municipalities and public institutions; and between managers and employees. This study also showed that 37.3% of the intimidation is experienced among the employees in education sector. In the study conducted by Fullan and Hargreaves (1991) on intimidation particularly in private schools in education sector, teachers were suggested to be intimidated by the school managers, their colleguaes and the students’ parents.

Even though intimidation is experienced in every organization, various researches made showed that intimidation is experienced more in non-profit organi-zations, schools and health sector than the profit making organizations. One reason for that was explained to be the management conducted by school managers who are inefficient in terms of management and leadership in non-profit organizations (Beswick et al., 2006; Davenport et al., 2003; Russo et al., 2008). In organizations of private sector, there is more intimidation while particularly in educational organizations intimidation is most in the form of assigning the employees duties which are below their capaticy (Ulu? and Beydo?an, 2012). The organizational climate in the organizations in which intimidation is experienced is a closed organizational climate. In other words, it comprises an atmosphere full of oppression culture, continuous tension and stress in terms of the school organization, an un-opened/closed communication system and antidemocratic attitudes and behaviors (Pšunder, 2011; Yaman et al., 2010). 

Intimidation was established to weaken the organizational trust and atmosphere and cause teachers’ work satisfaction and life satisfaction levels to decrease. Excessively competitive environment in the educational organizations, mismanagement, unsuccessful leadership types, and bad work designs; and negative effects of psycho-social working environment on communication and cooperation processes; frequent inspection of teachers and the existence of an excessive autocratic organization are considered to be the various reasons for experiencing intimidation. Those employees who are working for the institutions in which intimidation is executed deliberately or inadvertently are likely to face work dissatisfaction, lack of motivation, inefficiency and above all with a lot of physical illnesses. Particularly those intimidation types that are experienced in schools in which there is merely human factor have negative effects on teachers’ performances (Bradshaw et al., 2012; Duffy, 2012; Karakus? and C?ankaya, 2012; Korkmaz and Cemalog?lu, 2010; Mathiessen, 2006; Özen, 2009; Topa et al.,  2007).

The number and quality of the studies conducted on intimidation experienced in education sector in Turkey has been increasing day by day. Those people and institutions which have been considered in recent years in relation to intimidation as a research subject in education organizations in Turkey have not got enough knowledge about this field yet. Although intimidation is frequently experienced in public and private schools, little has been known about the meaning of this process yet. Intimidation experienced in public and private schools is frequently expressed by teachers but research results about its level and about in which school type it is experienced more need to be examined (Ayd?n, 2009; Do?an, 2009; K?rel, 2007; Özen, 2009).

There are numerous reseraches about intimidation in foreign literature. The first and the most significant one among these studies is Leymann’s research (1996) which describes the psychological intimidation and is accepted as the authority in its field. In accordance with the results of this study, 45% of the male employees experienced psychological violence whereas 55% of the female employees experienced it. Along with demographic and behavioral characteristics, the qualities and culture of the organization for which the employee work may also be factors triggering the beginning of psychological intimidation (Ertürk, 2013; Gökçe, 2006, 2008; Namie, 2003; Shallcross et al., 2008; Wet, 2010). When the school type is considered, teachers’ levels of exposure to intimidation were observed to vary in studies conducted in different countries during different times. School type, leadership approach and working conditions were observed to be the determining factors in terms of intimidation experienced by teachers in public and private schools (Bradshaw et al., 2012; Duffy, 2012; Einarsen and Skogstad, 1996; Fullan and Hargreaves, 1991; Hubert and Veldhoven, 2001; Zapf and Einarsen, 2001). Researches made by Fullan and  Hargreaves (1991)  and Zapf (2004) showed that teachers who work for private schools experience intimidation more than those teachers who work in public schools.

Teachers who work for private schools and those who work for public schools have different opinions about intimidation. Abay (2009) in his research could not find a significant relationship between teachers’ perception about intimidation and the school type. Researches conducted by Bulut (2007), E?erci (2009) and Özday? (1990) showed that teachers of public schools experience intimidation more frequently while researches conducted by Apak (2010) , Do?an (2009) and Dilmaç (2009) indicated that teachers of private schools experience intimidation more frequently. The literature survey of Halac and Bulut (2010) showed that there are numerous field researches about intimidation experienced in public schools whereas there is not a sufficient number of researches on intimidation experienced in private schools. The reserach conducted by Gökçe (2008) indicated that those teachers who work for private schools avoid from expressing their opinions because they are afraid of losing their job. Although numerous studies have been conducted on intimidation in foreign literature, there has been no meta-analysis on intimidation particularly experienced by teachers within the context of either their personal experience or the school type (Topa et al., 2007). In Turkey, various scales and different independent variables about teachers (such as sex, branch, marital status, education level, faculty from which they graduated and seniority) have been used in an increasing number of studies on intimidation within the context of teacher-school manager relationship in schools that have been conducted through qualitativr and quantitative methods since 2000. In the consequence of these studies, both significant and insignificant results, which indicate different findings in terms of the sub-groups of independent variables, were obtained. There is a strong need for meta-analyses to synthesize the results of all these studies and to develop a new direction for the following new studies on intimidation faced by teachers in schools. Considering the fact that intimidation experienced by teachers in public and private schools has negative effects on the quality of the education and the working conditions of teachers, determining the intimidation levels and raising consciousness among the related authorities and organizations increasingly become crucial. Bulut (2007), Ertürk (2013) and Gökçe (2006) stated that there is a need to make new researches and to synthesize the results of these studies since different researches made on intimidation based on the school type variable have different results. In addition, they suggested that meta-analyses of intimidation should also be made. The fact that intimidation faced by teachers in schools has been frequently studied on in recent years has led to a need for obtaining and synthesizing a common result by means of collecting these results, considering the number of samples, and analyzing them based on the variables. This study will be an original one at both national and international levels and it will also pave the way for new researches in this field based on different variables because no study in the literature covers intimidation faced by teachers in schools analyzed through meta-analysis method. Within this context, this study aims to determine the effect sizes and to determine whether there is a difference between the effect sizes when some variables that are ignored in primary researches are taken into account.

Purpose

The purpose of this study is to determine the effect sizes of teachers’ perception and their opinions about intimidation they face in schools based on the school type variable (public and private schools). 

This section covers research model, data collection and data analysis parts.

Research model

The model of this study is meta-analysis method, which is one of the methods used for synthesizing the research results. The process including analysis, synthesis and interpretation of the quantitative findings obtained from independent studies through statistical techniques is called meta-analysis. The purpose of a meta-analysis is to combine the findings of various studies conducted in different places and at different time on the same subject; thus achieving the most reliable fact in quantitative terms through increasing the number of samples (Cumming, 2012: 205; Ellis, 2012; Glass, 1976; Hartung et al., 2008; K??, 2013; Y?ld?r?m, 2014).

Data collection

MA theses and PhD dissertations on intimidation experienced by teachers in schools in Turkey are the bases of this study. The keywords “mobbing”, “psychological intimidation”, “psychological violence”, “intimidation” and “emotional harassment” were used to find the related material and researches in the National Thesis Archive of the Council of Higher Education. Following the browsing process, 11 of 42 studies on the subject of this study were found convenient for inclusion criteria. In choosing the studies to be included in this study, the following criteria were used:

(i) Criterion 1: Published or unpublished references: MA and PhD theses.

(ii) Criterion 2: Convenience of the research method of the study: the requirement for being an empirical study and use of tenure of office as an independent variable to obtain the effect size during the meta-analysis.

(iii) Criterion 3: Existence of sufficient numeric data: Sample size, mean, standard deviation, F value, t value, X² value, Kruskal Wallis value, Mann Whitney U data and p value were considered for teacher groups within the context of school type variable to determine the effect sizes necessary for a meta-analysis.

31 studies were not included in the study on the grounds that they used different variables (managers, academic members); they did not include school type variable; and they lacked the data necessary for a meta-analysis. The sample of this study is limited to MA theses and PhD dissertations on this subject written in Turkey between the years 2006 and 2013.

Research reliability: A coding protocol which includes the name, content and data of this study has been created. A secondary researcher who has an in-depth knowledge on the “Study Content” section of the Rating Protocol and on what to do is rated using an inter-rater reliability form in order to ensure the inter-rater reliability. The first rater is the researcher himself. Cohen’s Kappa statistics was used to ensure the inter-rater reliability and it was found to be 0.96. This result showed almost a perfect compliance between the raters.

Research validity: During this study, it was determined that the validity of data collection means was ensured in all of 11 studies included in the meta-analysis.

Data analysis

During this study, the effect sizes, variances and comparisons of the groups included in each study were measured through CMA ver. 2.2 [Comprehensive Meta-Analysis], Statistical Package Software for Meta-Analysis. This study includes teachers from public school teachers (the experimental group) and those from private school (control group). Thus, positive status of the effect size is interpreted as being in favor of teachers who have public schools while its negative status is interpreted as being in favor of teachers who have private schools. SSPS ver. 20.0 package software was used for the rater reliability test. Since the significance level was taken as 0.05 in the studies included in this study, the significance level of statistical analyses to be used in this study was determined as 0.05.

The related data covered in the studies included in this study were analyzed so as to find an answer to the question of the study. Findings concerning the publication bias, descriptive statistics, forest plot, fixed effect model findings, homogeneity test, random effect model findings and moderator analysis findings obtained from these analyses are given in this part.

As reflected in Figure 1, majority of 11 studies that were included in this study is located at upper side of the figure and very close to the conjoined effect size.

 

 

In case there is no publication bias, they are expected to expand symmetrically on both sides of vertical line showing the effect size (Borenstein et al., 2009: 284). One of the studies (Emiro?lu, 2011) that were included in this study to determine the conjoined effect size measured based on the school type variable went beyond the pyramid but they expanded around the top and the middle of the figure. If there was a publication level in 11 studies that were included in this study, then, the majority of the studies would be located at the bottom of the figure or only at a single part of the vertical line (Borenstein et al, 2009: 284). This cone graphic is one of the indicators of the absence of a publication bias in terms of the studies included in this study.

Orwin’s Fail-Safe N Evaluation was also conducted to test the publication bias. Orwin’s Fail-Safe N calculates the number of studies that are likely to be excluded from the meta-analysis (Borenstein et al., 2009: 285). In the consequence  of this analysis, Orwin’s Fail Safe N  was found to be 39. The necessary number of study for the average effect size was -0.12 in the consequence of the meta-analysis to increase to -0.01 (trivial) level; in other words, almost to zero effect size is 86. However, 11 studies which were included in this study are those which meet the inclusion criteria and which are available among all the studies conducted on this subject in Turkey (qualitative, quantitative, theoretical etc.). Impossibility to attain the 39 studies other than these may be accepted as another indicator of the absence of publication bias in this meta-analysis.

Conjoined findings of effect size analysis based on school type

The effect sizes of teachers’ perception about intimidation they experience in schools based on school type, standard error and its upper and lower limits based on a reliability level of 95% are given in an order from positive to the negative values on Table 1. 

 

 

In accordance with Table 1, the standardized mean difference (SMD=SOF) based on school type in these 11 studies, varies from -0.936 in favor of private school teachers to 0.610 in favor of public school teachers. A statistically significant difference (p <0.05) was found in 2 studies while no significant difference was determined in 9 studies. The confidence interval of 11 studies was also found to vary from -1.956 to 1.318.

Forest plot of the studies including data on the school type

The forest plot of 11 studies included in this study and consisting of the data concerning the school type is given in Figure 2.

 

 

When Figure 2 is examined, a difference higher than zero in favor of pivate school teachers is observed. The fact that there is a difference in favor of teachers who work for private schools may be interpreted as a sign of the fact that they experience intimidation more in proportion to those teachers who work in public schools.

Findings of effect of size meta-analysis based on school type conjoined in accordance with the fixed effect model and random effect model

The average effect size of the perception of teachers in private and public schools about intimidation they experience in schools conjoined in accordance with fixed and random effect models (without subtracting the outliers), standard error and its upper and lower limits based on a confidence interval of 95% are given on Table 2.

 

 

In Table 2, the average effect size value obtained from the effect size values of the studies included in this study based on the school type variable in accordance with fixed effect model was calculated as d=-0,129 whereas the standard error of the average effect size, the upper limit and lower limit of confidence interval of the average effect size was calculated as SE=0,045; -0,041; and -0,218, respectively. Data obtained from 11 studies included in this study based on the calculations showed that teachers who work for private schools experience intimidation more than those teachers who work for public schools in accordance with fixed effect model.

However, there is an effect even less than the lower level in accordance with Cohen’s classification in that the effect size value is lower than 0.20 (Cohen, 1988). According to Lipsey’s classification, there is an effect even less than the lower level when the effect size is lower than 0.15. The classification of Thalheimer and Cook (2002) shows that there is an insignificant difference (-0.15-0.15).

When statistical significance is calculated according to Z test, Z=-0.280 was found. The obtained result was found to have statistical significance with p=0,005. Only 4 of the 11 studies included in this study based on the school type variable have remained within the upper and lower limits of effect size and reached a result close to the existent effect size, whereas the remaining 7 studies have remained over or below these limits.

As for the homogeneity test, in other words, Q-statistics, Q was calculated as 62,787. 10 degrees of freedom at a significance level of 95% from x² table was found to be 3.940. The hypothesis on the absence of homogeneity in terms of the distribution of effect sizes was rejected in fixed effect model because Q-statistics value exceeded the critical chi square distributionvalue (x² 0,95 =3,940) with a degree of freedom of 10 (Q= 62,787). Thus, effect sizes distribution was determined to be heterogeneous in accordance with fixed effect model.

I², which was developed as a supplement to Q statistics, put forth a clearer result concerning hetero-geneity (Petticrew and Roberts, 2006; Y?ld?r?m, 2014).I² shows the rate of total variance about the effect  size.  As opposed to Q-statistics, I²statistics are not affected by the number of study. During the interpretation of I² 25% indicates a low-level heterogeneity, 50% indicates a mid-level heterogeneity and 75% shows a high-level heterogeneity (Cooper et al., 2009: 263). Since a level of heterogeneity close to a high-level heterogeneity was found in the consequence of the homogeneity for the purpose of school type variable (Q and I²) the model to be used for conjoining process was transformed into a random model. The results of the moderator analysis made to put forth the reasons for this heterogeneity (Fixed Effect Analysis) are given in Table 3.

 

 

Based on the moderator analysis conducted, publication type (p=0,45) and education level (p=0,46) were found not to be influential on the effect sizes of the studies. The moderator effect of the geographical regions where the study was conducted was determined (p=0.00). In the studies, the samples of which were Southeastern Anatolia Region and seven geographical regions, the standardized mean difference (SMD=SOF) was observed to be in favor of public schools (Southeastern Anatolia Region; d=0,61 and Seven Geographical Regions; d=021) while  this standardized mean difference varied in favor of private schools only in the samples of Agean, Marmara and Central Anatolia regions. The moderator effect of the researcher’s sex was also determined (p=0.00). Direction of the difference was observed not to change when the researcher was male; and an increase in the difference in favor of private schools was found out.

As reflected in Figure 3, an increase tendency in school type difference by years in terms of the effect sizes of the studies is observed.

 

 

This change was also validated statistically (Intercept = 204,032, Year d=-0,101, SE=0,051 [-0,202; 0,001], Z=-1,984, p=0,047).  

In this study, 11 effect sizes related to 11 studies constituting a sample of 3214 people were calculated. A statistically significant difference was detected in 2 studies while no significant difference was found in 9 studies. In fixed effect model, as a result of the conjoining process, a statistically significant effect size of -0.129 in favor of those teachers who work for private schools was found. This result may be regarded as low and insignificant   in   accordance   with   the  classification  of

Cohen (1988) and Thalheimer and Cook (2002). In random effect model, as a result of the conjoining process, a statistically significant effect size of -0,109 in favor of those teachers who work for private schools was found. This result may also be regarded as low and insignificant in accordance with the classification of Cohen (1988) and Thalheimer and Cook (2002). When these results are evaluated together, they show that there is a difference which may be regarded as insignificant in within the context of social sciences among the private school teachers and public school teachers in terms of their perception about intimidation they experience in schools. Thus, not using the school type as a variable in future studies may be brought to the agenda. Since there is not any other meta-analysis regarding teachers’ perception about intimidation they experience in schools based on the school type, it is not possible to compare and contrast these results.

Teachers’ opinions about intimidation varied at a low level based on their school type. It may be said that teachers have similar opinions about intimidation experienced in schools in accordance with the school type variable (Pekdemir, 2010). No significant relation-ship between teachers’ perception about intimidation and the school type was found in researches conducted by Abay (2009), Do?an (2009) and Gökce (2006). In this study, teachers’ perception that they are victims of intimidation was observed not to change significantly based on the school type which they work. Although it is stated that those teachers who are working for private schools face intimidation more frequently than those who work for public schools in Turkish education system (Alkan, 2011; Do?an, 2009; Bucuklar, 2009), this meta-analysis indicates a low-level difference in favor of private schools. In  this respect, this low-level difference between perceptions existent in private and public schools based on school type may be evaluated as a finding showing the fact that different variables may have effects on intimidation perception (gender, marital status and seniority etc.). Findings of this meta-analysis may be regarded as suggesting that intimidation phenomenon is a problem in both private and public schools and it is perceived at equivalent rates.

The fact that teachers work on contracts, high expec-tations of students’ parents and creation of a suppressive and competitive atmosphere by school management may have influences on the finding that teachers who work for private schools have a high-level intimidation perception than those who work for public schools. Higher level of anxiety and competition among teachers who work for private schools in proportion to public schools and the fact that they are expected to exhibit organizational behaviors more frequently (Bulut, 2007; Koç and Bulut, 2009) may lead to frequent exposition of teachers to intimidation. Teachers who work for private schools work on contracts and they are more concerned about losing their job. Waggoner (2003), in his study, states that teachers who work for private schools face overt or covert intimidation particularly during contract renewal or when they feel the possibility of their discharge due to their effort to maintain their job.

In private schools, intimidation reveals itself particularly as the intimidation imposed by the school managers resulting from their control and oppression at a higher level with their expectation of a higher performance (Bradshaw et al., 2012; Einarsen, 2000; Einarsen and Skogstad, 1996; Nielsen and Einarsen, 2012). Findings of the studies conducted by Apak (2009), Ayd?n et al. (2012), Gökçe (2006), Koç and Bulut (2009) indicating that those teachers who work for private   schools subordinate to the Ministry of National Education experience intimidation more than those teachers who work for public schools support the findings of this study. In Çivilida?’s study (2003), the finding showing that Anatolian High School teachers are granted more social support and experience intimidation less frequently than private school teachers also supports the results of this study. Canbay (2007) in his research suggested that private school teachers perceive their lack of job insurance as a kind of intimidation whereas public school teachers are not affected much by factors such as discharge since they work as a government official and this provides them with job insurance. Results of the study conducted by Gül?en and K?l?ç (2014) indicating that those private pre-school education teachers expe-rince intimidation more than their counterparts in public pre-school educational organizations are in parrallel with the results of this study.

The fact that there is a difference of a low-level significance among teachers’ perception about intimi-dation depending on the school type variable within the context of the results of this study may suggest that it will not be possible to use this variable as a significant independent one in future studies. Findings of the studies conducted in recent years also support this finding. Finding, discussing and raising consciousness about the reasons for private school teachers’ high-level perception about intimidation may be recommended. Within the context of this meta-analysis, other meta-analyses may be conducted using various variables predicting the intimidation other than the school type variable such as sex, marital status and seniority.  

The author has not declared any conflict of interests.