ADMINISTRATOR SELF-ESTEEM AND ITS RELATIONSHIP TO TRUST IN TEACHERS

This study aims to determine the level of school administrator self-esteem and its relationship to school administrator trust in teachers. The research design is a relational survey. The study was conducted at K-12 schools (a total of 131 schools) in İzmit county in the Kocaeli province of Turkey. The data were gathered through a three-part questionnaire: the first part consisted of the Rosenberg (1989) self-esteem scale, a ten-item four-point Likert-type scale; the second part included a three-factor trustworthiness instrument with 17 items by Mayer and Davis (1999); and the third part was related to six demographic variables. SPSS 13.0 was used for data analysis. Descriptive and inferential statistical techniques were applied to the data. Findings revealed that administrators have very low levels of self-esteem. Their levels of trust in teachers were also found to be low in all three dimensions of trustworthiness that were affected by administrator self-esteem. Results indicated that administrator self-esteem, education level, and seniority as an administrator in the current school were positively correlated with trust; title, school level, and seniority as an administrator were negatively correlated.


Introduction
Self-esteem and trust are two concepts that have been largely discussed in the literature.As a psychological construct, self-esteem is a central component of an individual's daily experiences, and it reflects and affects their ongoing transactions with their environment and the people around them (Kernis, 2003).Self-esteem is commonly defined as belief in oneself (Neufeldt & Sparks, 1990), as pride in oneself or self-respect in the American Heritage Dictionary (1992), as confidence in one's own worth or abilities in the Oxford Dictionary (2015), and as a feeling of having respect for yourself and your abilities, a confidence and satisfaction in oneself, in the dictionary of Merriam-Webster (2015).As a part of the general term "self-concept", self-esteem is defined as a positive or negative orientation toward oneself and an overall evaluation of one's worth or value.From Rosenberg's standpoint, the term "self-concept" characterizes the totality of a person's thoughts and feelings with reference to themselves as an object (University of Maryland, 2004).
According to Branden (1995), self-esteem is not an idea; it is a practice, the practice of (1) living consciously, (2) self-acceptance, (3) self-responsibility, (4) self-assertiveness, (5) living purposefully, and (6) personal integrity, which he suggests are the six pillars of self-esteem.In one of his earlier writings, Branden examined self-esteem as the sum of self-confidence, a feeling of personal capacity and self-respect, and a feeling of personal worth (cited in Wikipedia, 2015).Baumeister, Campbell, Krueger, and Vohs (2003) viewed self-esteem as a matter of perception rather than reality.They argued that what people believe shapes their actions in various ways; in turn, their actions shape their social reality and the social realities of the people around them.
The second focal point of this study, trust, has been studied extensively.Mayer, Davis, and Schoorman (1995) define trust as "the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party" (p.712).In a later publication, they simplified their definition of trust as being "the willingness to take risk" (Schoorman, Mayer, & Davis, 2007, p. 346), the degree of which depends on the level of trust.According to Hoy and Tschannen-Moran (1999), trust has a natural attraction because it is good to trust and to be trusted.They proposed five faces of trust: benevolence, reliability, competence, honesty, and openness.After extensive literature review, Vodicka (2006) defined trust in terms of these four components: consistency, compassion, communication, and competency.Although various factors have been proposed, three characteristics of a trustee, namely ability, integrity, and benevolence, appear in the literature, explaining a major SCHOOL ADMINISTRATOR SELF-ESTEEM AND ITS RELATIONSHIP TO TRUST... portion of trustworthiness.These three characteristics are also associated with Aristotle's Rhetoric, which suggests that a speaker's ethos is based on the listener's perception of intelligence, character, and goodwill (Mayer, Davis, & Schoorman, 1995).Krot and Lewicka (2012, p. 224) approach trust as a complex and multidimensional phenomenon that is either interpersonal or institutional in nature.They considered organizational trust to be an important part of professional relationships among all the parties of an organization.Trust is also considered to be a dynamic phenomenon, assuming different characters at different stages of a relationship; the basis of trust may change over time (Tschannen-Moran & Hoy, 2000).Henkin, Singleton, Holman, and Dee (2003) share the view that trust varies over time.
Studies related to trust in the workplace mostly focus on the trust of subordinates in their superiors.Lapidot, Kark, and Shamir (2007) focused on the relationship between the behavior of formal leaders and subordinate trust in them, while exploring the differences between trust building and trust erosion.Their findings showed that behaviors reflecting a leader's ability and integrity are more salient in trust-erosion incidents, while behaviors reflecting a leader's benevolence are more salient in trust-building incidents.According to Puusa and Tolvanen (2006), trust and the creation of trust are key in creating greater commitment to an organization, and strong organizational identity creates trust by affecting the level of identification of individuals within an organization.Canipe (2006) found that trust in co-workers, trust in supervisors, and trust in the organization are positively related to both organizational commitment and perceived organizational support.On the other hand, significant negative relationships were found between these same antecedents and turnover intentions.In organizations, administrators have the great responsibility to create and sustain a trusting environment.In schools, the behaviors of both administrators and instructors are very important in forming a general atmosphere of trust (Tschannen-Moran & Hoy, 2000).Schwabsky (2014) urged principals to focus on building and enhancing trust in schools, and on affecting the development of a school culture of "giving" and promoting teacher citizenship behaviors.Tschannen-Moran (2003) suggests "a greater focus on the study of trust to understand the dynamics that foster trust and how to repair trust that has been damaged.Principal preparation programs need to focus on the development of trust as a crucial component of leadership." Few studies concerned with the trust of superiors in their subordinates have been found in the literature.Polat and Doyuran (2005) conducted a study of K-12 school principal trust in others, including superiors, vice principals, teachers, parents, and students.They found that principals trust their vice principals the most and their superiors the least.They also found that female principals trust those they interact with more than male principals do.In another study with secondary school teachers, Polat and Celep (2008) found that length of service was a significant variable.The level of teacher trust in their principal was higher in younger teachers than their older counterparts.
Some researchers were interested in the concepts of organizational trust, organizational citizenship, and leadership.One study by Podsakoff, MacKenzie, Moorman, and Fetter (1990) showed that the effects of transformational leader behaviors on citizenship behaviors are indirect and mediated by follower trust in their leaders.Polat (2007) found that teacher perceptions of organizational justice, organizational trust, and organizational citizenship behaviors were all inter-related and interactive.Yücel and Samancı (2009) found that organizational trust has an effect on the organizational citizenship behavior of teachers, and that the length of service at a school is the variable with the most effect on organizational citizenship behavior.
Relations between self-esteem and trust have rarely been studied in the literature.Among these studies, Ellison and Firestone (1974) investigated the development of interpersonal trust by dividing subjects into high and low self-esteem categories and according to how they showed willingness to disclose themselves to a listener.They found a greater willingness to disclose to reflective-style potential disclosure targets.McWhirter (1997) conducted a relational study on self-esteem and loneliness and found self-esteem to be a significant predictor of experiences of global, intimate, and social loneliness in college students.The researchers of the current study took the liberty of assuming a relationship between the experience of loneliness and a lack of trust.In a different study with audiences of performing arts, conducted by Swanson, Davis, and Zhao (2007), self-esteem enhancement is found to be positively related to trust.Amogbokpa (2010) found correlations between antecedents of trust (ability, benevolence, and integrity) and both supervisor-subordinate and subordinate-supervisor relationships.Multiple regression analyses in this study also confirmed that there was a relationship between trust and ability, benevolence, and integrity in both relationships.Weining and Smith (2012) found a significant positive correlation between self-esteem and willingness to trust in others among undergraduate students.Their study showed that a lower level of self-esteem was associated with a lower level of confidence in self and in others, resulting in less willingness to trust someone.These relational studies inspired the researchers to seek a relationship between the level of school administrator self-esteem and trust in teachers.
This study differs from others in several points.For one, it was conducted with the administrators of K-12 schools.For another, studies related to trust usually focus on subordinate trust in their superiors.This study is very likely to be one of the few studies that are concerned with superior trust in their subordinates, specifically teachers.Finally, this study examines levels of administrator self-esteem as a predictor of their trust in teachers.
The general purpose of this study is to reveal the relationship between levels of school administrator self-esteem and their trust in teachers.The researchers were also interested in determining if selected demographics were significant variables influencing the levels of school administrator self-esteem and trust in teachers.As dependent variables, self-esteem and trust were considered to differ in terms of gender, title, school level, education level, seniority as an administrator, and seniority as an administrator in the current school.The following research questions were developed: 1.What are the levels of school administrator self-esteem and trust in teachers?2. To what extent does administrator self-esteem influence their trust in teachers?3. Does the level of administrator self-esteem differ in terms of gender, title, school level, education level, seniority as an administrator, or seniority as an administrator in the current school?4. Which independent variables have an impact on administrator trust in teachers?

Method
The design of this study is a relational survey for determining levels of school administrator self-esteem and its relationship to school administrator trust in teachers.The study was conducted at 131 K-12 schools in the İzmit district of the Kocaeli province in Turkey.The research population was 325, consisting of 105 principals, 16 deputy principals, and 204 assistant principals, with return rates of 89%, 81.3%, and 57.8%, respectively.Usable responses were received from 225 administrators, of whom 81.8% were males and 18.2% were females.The school principals were 41.8% of the responders, deputy principals were 5.8%, and assistant principals were 52.4%.With respect to school level, 11.6% of the respondents worked in pre-schools, 32.9% in primary schools, 28.0% in middle schools, and 27.5% in high schools.
In terms of the education level, 2.2% of the responders had associate's degrees, 68.5% had bachelor's degrees, and 29.3% had master's degree.The mean number of years of seniority as an administrator was 7.6, and the mean years of seniority as an administrator in the current school was 3.1.
SCHOOL ADMINISTRATOR SELF-ESTEEM AND ITS RELATIONSHIP TO TRUST...

Research instruments
The data was gathered using a three-part questionnaire.The first part consisted of the Rosenberg (1989) self-esteem scale, the second included the three-factor trustworthiness instrument by Mayer and Davis (1999), and the third part was related to demographic questions including gender, title, school level, education level, seniority as an administrator, and seniority as an administrator in the current school.
Research instrument I: Rosenberg Self-Esteem Scale (RSES): Using the data from this research, an Exploratory Factor Analysis (EFA) was applied with the RSES because the scale has one dimension.This instrument consisted of a ten-item four-point Likert-type scale, from strongly agree (3) to strongly disagree (0).Since half of the items were negatively structured, they were reverse-scored.The scale ranges from 0-30.A higher score means a higher level of self-esteem.Sample items include "I fell that I have a number of good qualities" and "I wish I could have more respect for myself".
The results of the EFA (KMO=.740;Bartlett Test=497.690df=45 P=0.000; total variance explained= %31.704; with factor loadings from .401 through .720)and reliability test (Cronbach's Alpha=.743)show that the data from this research were consistent with the Rosenberg one-dimensional scale.
Research instrument II: Trustworthiness Scale by Mayer and Davis: This instrument is a 17-item five-point Likert-type scale, from strongly agree (5) to strongly disagree (1).The original instrument consisted of three sub-scales, identified as ability, benevolence, and integrity.Sample items from these sub-scales: "I feel very confident about the skills of teachers in my school", "Teachers in my school are willing to go out of the way to help me", and "Teachers in my school have a strong sense of justice", respectively.
Item-total correlation coefficients were calculated for the reliability test.Items 7 and 15 were skipped from the scale since the values of corrected item-total correlation were under the acceptable value of 0.50.The remaining 15 items were grouped into three factors -ability (six items), benevolence (four items), and integrity (five items) -which were all compatible with the scale by Mayer and Davis.Cronbach's Alpha of the total scale was 0.92, and 0.92, 0.78, and 0.87 for the three factors respectively.Finally, a confirmatory factor analysis (CFA) was applied on Lisrel 8.54 developed by Jöreskog and Sörbom (2001).

Dependent and independent variables
Administrator trust in teachers was taken as a dependent variable in this study.Two types of procedures were applied using this variable: First, administrator trust in teachers was calculated through an analysis of the structural equation modelling (SEM) with the three dimensions of the trustworthiness instrument: ability, benevolence, and integrity.Second, two new categories of this metric variable were obtained using a cluster analysis to run a binary logistic regression.The findings of the cluster analysis are given in Table 1.The first cluster represents the administrators who trust in teachers and the second cluster represents those who do not.As illustrated in Table 1, the dependent variable has a discriminative feature since there is a statistically meaningful difference between Cluster 1 and Cluster 2. Administrator self-esteem, gender, title, school level, education level, seniority as an administrator, and seniority as an administrator in the current school are considered as independent variables in this study.Binary logistic regression analysis was used to determine which of the independent variables had an impact on administrator trust in teachers because the dependent variables had two categories that cannot be ranked.
Self-esteem was used as both a dependent and an independent variable.Self-esteem as dependent variable was applied in inferential statistics (Table 3) and as independent variable in a binary logistic regression (Table 4).

Level of school administrator self-esteem and trust in teachers
As illustrated in Table 2, the mean value for self-esteem of the participating administrators was 6.89.RSES is a four-point Likert-type scale ranging from 0 to 3. The total points that participants can obtain from the ten-item scale range from 0 to 30.A higher score means higher self-esteem; a mean value of 6.89 represents a very low level of self-esteem.Table 2 also shows the mean values on three dimensions of the trustworthiness instrument.Trustworthiness was measured by a five-point Likert-type scale.The highest mean value of 2.53 was obtained from the dimension "ability" and the lowest value of 2.21 from the dimension "benevolence".All the mean values are, however, below the midpoint of the five-point scale.Influence of school administrator self-esteem on their trust in teachers A structural equation model is presented in Figure 2. In the model, the independent variable is self-esteem and the dependent variables are sub-scales of the trustworthiness instrument: ability, benevolence, and integrity.As seen in the figure, the model has three modification indices with the error co-variances between items 3 and 4 (0.35); 1 and 10 (0.29); and 3 and 10 (0.23).Standardized weights of items in self-esteem are between 0.26 and 0.74.Dependent variables for SEM are ability, benevolence, and integrity, and their standardized weights are 0.74 and 0.87; 0.62 and 0.79; and 0.62 and 0.86, respectively.According to the SEM results, administrator self-esteem has an impact on all the dimensions of trustworthiness, with the highest impact on benevolence (0.31) and the lowest on ability (0.16).T values for ability, benevolence, and integrity are 2.13; 3.61; and 2.86, respectively.Based on GOF criteria by Hair, Anderson, Tatham, and Black (1998), all fit measures of SEM are obtained at acceptable levels with three modifications (χ 2 =605.35;χ 2 /df=605.35/266=2.28;RMSEA=0.075;NFi=0.91;NNFi=0.94; and CFi=0.95).
SCHOOL ADMINISTRATOR SELF-ESTEEM AND ITS RELATIONSHIP TO TRUST...

Figure 2 Structural Equation Model
Level of school administrator self-esteem according to demographic variables Table 3 shows the findings of comparisons of administrator self-esteem in terms of independent variables.In this context, three different tests (t, F, and χ²) were applied based on meeting the criteria of the related tests.As seen in the table, the t test was applied for the variable of gender; ANOVA was applied for the variables of school type, seniority as an administrator, and seniority as an administrator in the current school; Kruskal-Wallis H was applied for the variables of title and education.Only one out of six variables (F= 4.539; p<.01) was found significant in these comparisons.Participants working less than five years as an administrator have significantly the highest self-esteem (mean value is 8.01) within the other two groups.However, this mean value is still low compared to the maximum score of 30.The impact of independent variables on administrator trust in teachers The fourth research question in the study was addressed with a binary logistic regression analysis, illustrated in Table 4.The independent variables used in the analysis are gender, title, school level, education level, seniority as an administrator, seniority as an administrator in the current school, and administrator self-esteem.The dependent variable is administrator trust in teachers.To run the binary logistic regression, data on ability, benevolence, and integrity were classified into two groups using a two-step cluster technique.Thus, two groups of administrators were obtained after the cluster analysis.The number of administrators with low trust levels was 126 (56%) and high trust levels was 99 (44%).
As illustrated in Table 4, the regression model fits the data appropriately (Çokluk, Şekercioğlu, & Büyüköztürk, 2010).The results of the Hosmer-Lemeshow test in a binary logistic regression must be statically insignificant SCHOOL ADMINISTRATOR SELF-ESTEEM AND ITS RELATIONSHIP TO TRUST... as a sign of the acceptable fit of a model.The Hosmer-Lemeshow value in this study was calculated as 12.395 (p=.134), and the model was found acceptable.The overall accuracy rate was 73.8%, which shows the total correct classification percentage of the administrators with either low or high trust levels.This means that the two groups obtained via a two-step cluster technique were confirmed in a binary logistic regression.
Table 4 shows that the log of the odds of administrator trust in teachers are negatively correlated with title (p<.01), school level (p<.01), and seniority as an administrator (p<.01); it is positively correlated with education level (p<.05) and self-esteem (p<.01).Two other independent variables, gender and seniority as an administrator in the current school, are not correlated with the log of the odds of administrator trust in teachers at the alpha level .05.
Findings show that administrator self-esteem is positively correlated with the administrator trust in teachers.The odds obtained from this variable are e .164= 1.178, and it could be stated that administrators having higher levels of self-esteem are more likely to trust in teachers.Gender does not affect administrator trust in teachers.Administrator title has an impact on the dependent variable.The odds of a school principal's trust in teachers are e -.1.473= .229times lower than that of an assistant principal.This means that the probability of trusting in teachers is 1/.229 = 4.367 times higher for an assistant principal than for a principal.School level also has an impact on the dependent variable.The odds of primary school administrators are e -.1.512= .221times lower than that of high school administrators.This means that the probability of trusting in teachers is 1/.221= 4.525 times higher for a high school administrator than for a primary school administrator.Education level is another significant independent variable in this study.The odds of a school administrator with an associate's degree are e 2.681 = 14.595 times higher than that of an administrator with a post-graduate degree, and the odds of school administrators with bachelor's degrees are e 1.228 = 3.414 times higher than that of administrators with master's degrees.This means that the probabilities of trusting in teachers are 14.595 times higher for a school administrator with an associate degree and 3.414 times higher for an administrator with a bachelor's degree than for administrators with master's degrees (Table 4).
As illustrated in Table 4, both seniority as an administrator and seniority as an administrator in the current school are significant independent variables.The odds of a participant working as an administrator for less than five years are e -.1.140= .320times lower than that of administrators working more than 10 years.This means that the probability of trusting in teachers is 1/.320= 3.125 times lower for a participant working less than five years than that of one working more than 10 years in a school as an administrator.In contrast, seniority as an administrator in the current school is positively correlated with the administrator trust in teachers.The odds of administrators working for two years in the current school are e 1.190 = 3.288 times higher than that of administrators working for five years and more in the current school.This means that the probability of trusting in teachers is 3.288 times higher for an administrator working for two years in the current school than that of one working for five years and more.

Discussion and Conclusion
The overall levels of school administrator self-esteem were found to be quite low (Mean=6.89)when considering the highest score of the scale is 30.The levels of trust in teachers were also found to be lower than the mid-point 3.00 in a five-point scale in all sub-scales.The reason behind these results may be the timing of this study.At the time the study was conducted, the Ministry of National Education (MoNE) was in the process of enforcing new regulations concerning school administrator appointments.Unlike previous policies, the new regulations required that, in order to be reappointed, administrators who had been working for more than four years as a principal had to go through a selection process by local authorities.This might have created uncertainty and unease among current administrators, since the new rules are considered to be open to personal and political interference, hence giving school administrators the feeling that their current positions are threatened.Administrators in the study might have been experiencing such negative feelings as stress, distrust, and uncertainty caused by the transformation and change process.As noted by Lewicka and Krot (2012), trust in vertical relations is a prerequisite for the effective implementation of changes.This study indicates that administrator self-esteem has an impact on all dimensions of trustworthiness, with the highest being on benevolence, followed by integrity and ability.This means that the higher the administrator self-esteem, the higher the trust in teachers.An interesting finding is that self-esteem has the highest impact on benevolence, although benevolence has the lowest mean among the dimensions of trustworthiness.The results of this study are similar to the findings of Weining and Smith (2012), indicating a significant positive correlation between self-esteem and willingness to trust.Similarly, Chattopadhyay and George (2001) reported a positive relationship between trust and self-esteem as measured by the Organization-Based Self-Esteem (OBSE) scale and suggested that being in a trustworthy environment has a positive relationship to a person's organization-based self-esteem (cited in Pierce & Gardner, 2004).On the other hand, the results of Krot and Lewicka (2012) show that competence may play a greater role than benevolence or integrity in establishing intra-organizational relationships in new organizations, while integrity may play a greater role than competence and benevolence in maintaining intra-organizational relationships in mature organizations.The current study's results contradict those of Krot and Lewicka (2012).
Selected demographics were analyzed as independent variables to determine whether they made significant differences in self-esteem.The only significant variable was seniority as an administrator.This indicates that the new generation of administrators with less than five years of experience have the highest levels of self-esteem.This result is not consistent with the findings of Johnson (1988), who observed no statistically significant relationship between the total length of time in administration and the level of self-esteem.Other demographics, specifically gender, title, school type, education, and seniority as an administrator in the current school, were found to be insignificant in administrator self-esteem as a dependent variable.The results of this study regarding gender are consistent with the findings of Aydoğan (2013), which show no significant difference between male and female school administrator self-esteem and the findings of Baldwin and Hoffmann (2002), who observe no significant impact on changes in self-esteem.In terms of title and education, the current study supports the findings of Johnson (1988), who shows that self-esteem is not related to level of education or administrative position in land-grant colleges and universities.
Based on the regression analysis results, three independent variables (administrator self-esteem, education level, and seniority as an administrator in the current school) were found to be positively correlated with trust as a dependent variable.Administrators in this study with higher self-esteem tend to trust in teachers more.This result is consistent with the findings of Weining and Smith (2012) and Swanson et al. (2007).Administrators with either associate's degrees or bachelor's degrees trust more in teachers than administrators with master's degrees.This could indicate that people with more education might have higher expectations of others and tend to approach them doubtfully.An interesting result of this study is that administrators working in the current school for two years trust in teachers more than those who have been working in the current school for five years or longer.This result is consistent with the findings of Polat and Celep (2008), who observed that teachers with shorter lengths of service trust more in their principals.On the other hand, these results contradict the findings of Tschannen-Moran and Hoy (2000), which indicate that as the parties gain experience, trust can develop based on knowledge of one another during the relationship.
Regression analysis results revealed that three independent variables (namely title, school level, and seniority as an administrator) were negatively correlated with administrator trust in teachers as the dependent variable.School principals tend to trust in teachers less than assistant principals do.This result might be attributed to various reasons.For one, assistant principals usually develop more informal relationships with teachers on a daily basis.The principals often have very busy schedules both in and outside the office.This situation might reduce their chances to get closer to their colleagues.As for the variable of school level, primary school principals tend to trust teachers less than high school principals do.The reason for this might be SCHOOL ADMINISTRATOR SELF-ESTEEM AND ITS RELATIONSHIP TO TRUST... that most high school teachers in Turkey have a five-year college education and they generally have a better image in society than primary school teachers.In regard to the duration of the administrative position, the results reveal that administrators with less than five years of experience have lower levels of trust than those with more than ten years of experience.This result could indicate that the younger generation of administrators is often more educated and idealistic, and tends to be less tolerant in workplace relations.Therefore, they might have a more critical approach towards their subordinates, as discussed previously.
In the regression analysis, gender was the only variable with no effect on administrator trust in teachers.Baş and Şentürk (2011) found no difference between male and female teachers in terms of trust in administrators.In another study by Vineburgh (2010), female faculty had a higher mean score on organizational trust.There were a limited number of studies examining administrator trust in subordinates on the basis of gender, making it difficult to compare the findings of this study with similar studies.
The results of this study are limited to the accessible population of a small city in northwestern Turkey.The results cannot be generalized to all Turkish school administrators.Further research is needed in larger populations to explore variables that could have the potential to explain relationships between self-esteem and trust.This study indicates that the levels of administrator self-esteem are extremely low and the levels of their trust in teachers are not satisfying.It could be properly suggested that the authorities in MoNE take the results of this study into consideration and create a supportive environment for school administrators in their challenging job.

Figure 1
Figure 1 CFA for trustworthiness scale

Table 1
Cluster descriptors: differences in mean values for administrator trust in teachers

Table 2
Descriptive statistics on administrator self-esteem and trust in teachers (n=225)

Table 3
Descriptive and inferential statistics on administrator self-esteem in terms of independent variables (n=225)

Table 4
Logistic regressions predicting administrator trust in teachers