The mediating role of responsibility and friendship between empathy skill and benevolence

Received Feb 28, 2021 Revised May 5, 2021 Accepted May 14, 2021 The purpose of this study is to investigate the mediating effect of friendship and responsibility between benevolence and empathy skill in the context of social studies courses. The data in the study, which involves a total of 646 8th grade school students from Turkey, was collected using the scales "Empathy Scale, Benevolence Scale, Friendship Scale, and Responsibility Scale" developed by the authors. It was seen that the research variables have significant positive correlations with each other. It was found that empathy predicts benevolence, friendship, and responsibility significantly and in a positive direction. Friendship and responsibility, on the other hand, were found to play a significant, fully mediating role in the relationship between empathy and benevolence. These results were discussed using the relevant literature and suggestions that were gathered.


The relationship between empathy skill and responsibility and friendship values
First of all, in altruism-based benevolent behavior, the individual's self-sacrificing attitude without any self-interest and a completely sincere absence of an expectation for external reward [46], [43] was effective in determining the values of friendship and responsibility, which are the mediating variables of the research.
Friendship is defined as the relationship that people who have a similarity of feelings, beliefs, thoughts, and goals and who love each other sincerely, without any interest, stand shoulder to shoulder and resist problems, as well as sharing happiness [47], [48]. Friendship is a need that cannot be met in the home environment, and friendships enable the child to adapt to society, to cooperate, and to participate in group works with a certain purpose and to take responsibility. Friendship enables people to exchange, be selfless and not selfish [49]. In order to comprehend the relationship between empathy and friendship, the expression "it ensures non-selfishness" can be contemplated. Because being selfish, that is, being self-centered, is a concept that is incompatible with empathy and prevents one from putting oneself in the other's shoes, in short, it is a situation that blocks the process of empathy [1]. Thus, it can be argued that egoism has no place in either friendship or empathy.
In addition, it can be argued that communication is an important tool in initiating the friendship process, since friendship is a type of interpersonal relationship [50]. Considering that communication is an action-reaction event, including both communication and empathy in the good or bad situation that may be encountered will provide an opportunity for individuals to better understand each other's problems and solve them in the most effective and efficient way [51]. At this point, it is possible to say that communication and empathy are two related skills [1]. This relationship has also been found in studies conducted [52], [53]. In fact, it is stated that individuals with a low level of empathy may be inadequate in communicating with others and expressing their feelings, and they cannot help people around them in solving their problems [54]. Considering that empathy is an important predictor of communication skill in this case and communication is an initiating factor in establishing friendship relations, it is possible to talk about empathy being present in friendship. It is also stated that individuals who have the ability to empathize are more successful in developing friendly relations than those who do not have this ability [55]. Barrio, Aluja, and Garcia [56] found a strong relationship between empathy and friendship in their study on adolescents. It is even claimed that empathy may have a positive moderator effect on friendships, not only is it more accepted by friends of those with high empathy [57], [41]. The other mediating variable of the model designed in this study, which is responsibility, is defined as a social norm, attitude, and personality and character trait, as well as being a value [58]. Similarly, in addition to the fact that responsibility is a value-laden concept, it is stated that responsibility behavior is carried out according to the perception of the person about the situation and conditions, and that a responsible person will take on their responsibilities [59]. On the subject of what is the source of responsibility, different opinions have been put forward such as environment, mind, freedom, and emotional situations. However, it has been stated that responsibility cannot be explained solely with reason [60]- [62]. On the other hand, it is pointed out that internal control constitutes the main source of responsibility, and if there is an external expectation, i.e. reward or similar situations, this will not be about responsibility [63]. Internal control increases the responsibility levels of students, moreover, additional internal control programs are provided with programs designed to develop students' responsibility behaviors. Thus, as the students' internal control increases, their responsibility levels also rise [64].
The responsible person was initially described with a simple phrase as "the person who can take on responsibilities." So, to define the concept of responsible person in a detailed manner, it can be stated that they are the people who are aware of their self-worth, who hold only themselves responsible for their thoughts, feelings, and behaviors, who can fulfill their duties, and when they do not, they endure the consequences, and they respect themselves and the environment [65,66]. In addition, it is stated that responsibility has social and individual aspects, but the prerequisite for achieving social responsibility is still development of individual responsibility [67], and it is seen that curricula are also prepared in accordance with this [68], [69]. It is seen that these curricula positively affect students' empathy skills and altruistic behaviors [70], while curricula based on responsibility and respect have a similar effect on empathy [71], [72], [12]. On the other hand, there are also studies showing that empathy skill positively predicts students' levels of responsibility [73], [42], [74]. In short, it is possible to talk about a reciprocal relationship between empathy skill and responsibility. Research also illustrates this [75], [76].

Current study
In this study, the purpose is to investigate the mediating role of responsibility and friendship values between empathy and benevolence. As presented in the theoretical framework of the research, there are studies that reveal the relationships between the variables of the constructed model and predicting each other, thus the theoretical structure of the research was supported with these studies. In addition to these, structural models were found where empathy and social responsibility are the mediating variables, interpersonal benevolence behavior is the dependent variable [77], empathy is both the mediator and independent variable, benevolence behaviors is the dependent variable [78]- [81], empathy is the mediator variable, friendship is the dependent variable [82], empathy is both the independent and mediator variable, friendship quality is the dependent variable [83], [84]. In Turkey, on the other hand, studies were found in which the mediating role of prosocial friends and friendship quality in the relationship between autonomous motivation and prosocial behavior is examined [85], the mediating role of responsibility distributions in the relationship between attachment styles and forgiveness is examined [86], and empathy is the independent and mediating variable on various variables [87], [88]. It can be said that these are generally oriented towards the fields of psychology and social psychology, their subjects are far from the scope of education, and at the same time, they do not address empathy and values together. Moreover, although there are few domestic studies investigating the effect of empathy on values, there are no studies investigating the relationships between them with the mediation model.
In the study examining 54 doctoral theses on value education, it is seen that empathy is a frequently preferred subject in value education [89], and in a similar study, it is seen that empathy is suggested in studies in which it is examined as a skill that should be carefully considered in value education [90]. In addition to these, Baş and Beyhan [91], who examined 22 graduate theses on values education, suggest that general attitudes and opinions are taken on values and that empathy should be included in addition to these values and thus, help values to be understood better. It is known that the values discussed in this study are among the common core values aimed to be conveyed in social studies curriculum (except friendship), moreover, in all the curricula in Turkey.
From this point of view, it was necessary to know how and in what direction the interactions between empathy skill and benevolence, friendship and responsibility value occurred. With the support of the relevant literature, a structural model presented in Figure 1  As seen in Figure 1, five paths 1, 2, 3, 4, and 5 were created in the model. These pathways also illustrate the hypotheses to be tested in the study. Moreover, these hypotheses represent the conditions proposed by Baron and Kenny [92] in their analysis of mediation. If these are verified, the significance of the indirect effect will be tested. Then, the H6 hypothesis of the research, "The values of responsibility and friendship have a mediating effect between empathy skill and benevolence value" will be either accepted or rejected. There are six hypothesis in this research:

RESEARCH METHOD 2.1. The research model
The current study is designed by a predictive correlation type with multi-factors of correlational designs. In predictive correlation research, the researchers examine the associations amongst variables, and to what extent a variable predicts another one or ones [93]. Predictive correlation research with multiple factors aims to uncover direct and indirect (mediations) connections. So, it tries to explain changes in dependent variables by diagnosing any association between independent and dependent variables [94], [95].

Participants
The research was carried out with 646 8 th grade school students (female: 56.5%, male: 43.5%), who volunteered to participate and were selected from 20 public secondary schools from different socio-economic levels in Uşak city center, located in the inner west of Turkey.

Scales
Firstly, this study was adapted from the doctoral thesis "Investigation of the Relationship Between the Empathy Skills of Secondary School Students and Values: A Structural Equation Modeling," which is currently being carried out at Uşak University Graduate Education Institute. The data of the research were collected through scales named "Empathy Scale, Benevolence Scale, Friendship Scale, and Responsibility Scale," developed by authors. All the items in the scale were scored as "Not suitable to me=1, Somewhat suitable to me=2, Quite suitable to me=3, Completely suitable to me=4" Getting one point from any item in the scales means that the relevant value is low, and getting four points means that it is high. In addition, no reverse items were used in any of the scales. Data regarding the reliability, validity, and structure of the scales are given in Table 1.
As can be seen, all of the scales have a single factor structure and the explained variances are over 30%. It would be appropriate to state that the extent to which the single-factor scales explain variances is good, since it is considered sufficient for scales to explain 30% or more of the total variance [96]. In addition, for the general fit of the models, the χ2/sd parameter being below 2, the RMSEA value being below.05, and  [98]. It is seen that the scales are within these parameter values.

Data analysis
The data analysis of the study was carried out in two stages-"(1) preparation stage and (2) mediation analysis stage." Preparation stage: The correlations between these variables were examined in order to causally test the effects of empathy, friendship, and responsibility values on benevolence value of middle school eighth grade students. Because, in order to explain the existence of a causal effect, the relationship between the assumed structures are examined [99]. For this purpose, descriptive statistics of these variables were calculated using Pearson Product Moment Correlation Analysis to examine the relationships between the variables. Then, skewness and kurtosis values were checked in order to see whether the data has a normal distribution. Finally, the variance magnification factor (VIF) values of independent variables (empathy, friendship, responsibility), and models used in the determining of multiple linear connections. It is recommended that the coefficient of skewness and kurtosis is between -1.5 and 1.5; and the VIF value is <2.5 [100]- [103]. All these analyses were carried out with the SPSS vol.22 package program. Thus, the data were prepared for mediation analysis.
Mediation analysis stage: A mediating role was given to friendship and responsibility values between the empathy skills of middle school eighth grade students and their benevolence value. Mediation analysis is performed to explain the role of these variables that affect the causality relations between the independent variable and the dependent variables [104]. The significant effect of the detected "full mediating role" was tested with the bias-corrected bootstrap method. This method provides strong evidence for the effect of the mediator variable with bias-corrected confidence intervals [105]. In the bias-corrected bootstrap method, it is recommended to increase the confidence interval to 95% and the sample to 2000 or 5000; whereas there should not be zero in the confidence intervals of the indirect effect, that is, none of the upper or lower values should be negative and the significance value should not be below .05 [106], [107]. To test the significance of the exact mediating role of friendship and responsibility values between empathy and benevolence value, 5000 samples and a confidence interval of 95% were preferred. All these operations were carried out using AMOS vol.24 package program.

Descriptive statistics and correlation
Descriptive statistics and correlation values of the variables (empathy, responsibility, and friendship) regarding the variance inflation factor (VIF) values of the models are given in Table 2.  Table 2 is examined, it can be said that the variables have a normal distribution, since the values of skewness and kurtosis are in the range of -1.5 to 1.5. The fact that the variance inflation factor (VIF) values of the models used in the determination of multiple linear connections were found to be <2.5, indicates that there is no multi-linearity problem between the variables. In addition, variables have significant positive correlation values with each other.

Mediation analysis
SEM analyses were conducted to test whether friendship and responsibility have mediating roles in the relationship between empathy and benevolence. In line with the conditions of mediation analysis proposed by Baron and Kenny [92], firstly, a direct connection was established between empathy and benevolence (path 1); the independent variable (empathy) predicted the dependent variable (benevolence) in a significant positive direction, and the H1 hypothesis was accepted (β =.630, SE = .060, t = 9.460, p < .001). Moreover, empathy explained 40% of benevolence (R 2 =.397).
Then, the friendship variable was included in the model with a mediating role and Ha H2 and Hb H3 hypotheses have been tested. Empathy predicted friendship positively and H2 was accepted (β=.677, SE=.070, t=9.988, p<.001). Friendship predicted benevolence significantly and in a positive direction and H3 was accepted (β=.495, SE=.062, t=6.798, p<.001). Moreover, empathy explained 53% of benevolence (R 2 =.529) and 46% of friendship (R 2 =.459). Also, as a result of the addition of the mediator variable (friendship) to the model, although empathy preserved its significance on benevolence, there was a decrease in its coefficient of effect (β=.294<β=.630, p< .001). This finding indicates a partial mediation. However, since there are two mediator variables in the model, the mediator variable (friendship) was removed from the model and the other mediator variable (responsibility) was included in the model, and H4and H5 were tested.
As expected, empathy predicted responsibility significantly and in a positive direction and H4was accepted (β=.656, SE=.055, t=9.769, p<.001). Responsibility also predicted benevolence significantly and in a positive direction, and H5 was accepted (β=.555, SE=.072, t=8.241, p<.001). Moreover, empathy explained 57% of benevolence (R 2 =.573), and 43% of friendship (R 2 =.431). The significance of empathy on benevolence was not impaired, but there was a decrease in its coefficient of influence (β=.267<β=.630, p<.001). This situation is similar to the results obtained when friendship was included with a mediating role. Then both mediating variables (friendship, responsibility) are included at the same time and H6 was tested; the final model of the study is presented in Figure 2.
SRW=Standardized Regression Weight; SE=Standart Error, ***p < .001 Figure 2. Path analysis results of the effects of empathy on benevolence mediating responsibility and friendship As presented in Figure 2, the significant positive effect of the independent variable (empathy) on the mediator variables (friendship, responsibility) continues, and similarly, the mediator variables (friendship, responsibility) have a significant positive effect on the dependent variable (benevolence). This finding shows that the H2, H3, H4, and H5 hypotheses are fulfilled. In addition, it is seen that the significant positive effect of the independent variable (empathy) on the dependent variable (benevolence) (β=.630, p<. 001), approaches 0 when the mediator variables are included in the model, and the significance disappears (β=.051, p=.477>p=.05). This finding indicates that friendship and responsibility play a full mediating role between empathy and benevolence. But in order to reach a definite conclusion, the significance of the indirect effect must be known [92].
The significance of the detected indirect effect was tested with 5000 samples at 95% confidence interval, using the bias-corrected bootstrap method and the results are given in Table 3. According to Table 3, the indirect effect coefficient of the specified direction is significant (β=.602, p<.001) and there are no zero values at the 95% confidence interval among the lower and upper values (95% CI=.476, .745). These findings show that the hypothesis H6"Friendship and responsibility have a mediating effect between empathy and benevolence" is accepted. Also, empathy explains 40% of benevolence (R 2 =.397), and it is seen that when the mediating variables (friendship, responsibility) are included, it explains 63% of the total variance (R 2 =.630). Also, it can be stated that the fit indices of the final model are within compatible ranges: χ2=693.696, df=294, p<.01, χ2/df=2.360, RMSEA=.046, CFI=.936, IFI=.936, TLI=.929, GFI=.923, AGFI=.908 [108], [109].

DISCUSSION
Research results revealed the positive and significant relationships between empathy skills and values of benevolence, friendship, and responsibility. Accordingly, it can be stated that these values will increase as the empathy skill increases. In addition, empathy skill was found to be a predictor of benevolence value. There are also other studies with similar results [35], [2], [36], [45]. In addition, parallel results have been obtained in studies in which empathy is both a mediator and an independent variable, and helping behaviors have a dependent variable role [78]- [81]. Moreover, empathy has a significant positive effect on responsibility and friendship values.
Another result of the research is that the significant positive mediating effect of friendship and responsibility values between empathy skill and benevolence value was revealed. The striking point here is that without the mediating variables (friendship, responsibility) in the model, empathy explains 40% of benevolence (R 2 =.397), which shows that empathy is an important component of benevolence relationship [110]. When the mediating variables are included in the model, the explaining percentage of benevolence increased to 63% (R 2 =.630). While values are principles and basic beliefs that guide behaviors and are standards by which actions are judged as good or desirable [111], there is no single type of value underlying behaviors [112]. Thus, it can be inferred that the values of friendship and responsibility contribute to the explanation of benevolence behavior, which has a multidimensional nature [113]. It was also discussed that there is no external reward expectation in responsibility; it is an internal process. Hence, in the research of Bar-Tal, Raviv and Goldberg [114], it was observed that older children helped young people even if they did not expect rewards. In fact, in another study conducted with preschool children [115], it was observed that children do not display benevolent behaviors when there is an authority figure or an adult in the environment. The reason for this was explained as that children did not feel responsible and competent to help. Then, this explains the reason that Glasser [64] added an internal control program to the responsibility programs designed in order to improve the responsibilities of students. When all these are brought together, it is possible to see that responsibility is a value that has an effect on benevolence.
In addition, the significant full mediation effect found also revealed a strong structure in explaining benevolence. In a meta-analysis study examining whether values education practices in Turkey have a positive effect on desired characteristics of students, Ateş [116] found a significant effect. In addition, it is stated that there are empirical studies on social studies and science, and yet this number is not at a sufficient level in Turkey. When the contents of the studies are examined, it is observed that the values education program has a positive effect on the development of students' empathy skills, responsibility, and friendship values [10], [71], [72], [12]. On the other hand, it has been found that the empathy program applied in preschool education increases the benevolent behavior levels of students [39]. In fact, a study, where empathy is an important foothold, shows that; the peer mediation program that enables third party students to become involved in the conflict and help resolve the problem in a democratic way without adult intervention J Edu & Learn ISSN: 2089-9823  The mediating role of responsibility and friendship between empathy skill and benevolence (Ali Gokalp) 209 [117] increased the tolerance tendency, which includes values such as love and respect, of social studies 5th grade students [118]. But we are afraid that the number of such programs in Turkey are very few.

SUGGESTION
It is known that peer mediation programs are capable of raising responsible citizens and empathy is used effectively in this process similarly, social knowledge aims this and empathy skill is included in the program as a skill that needs to be developed It was also observed that the tolerance tendencies of social studies 5th grade students can be increased with 24 original scripts based on the role-playing technique by utilizing the peer mediation program Considering the positive effect of empathy on friendship and responsibility, with activities based on empathy-friendship-responsibility, studies that aim to identify their effects on students' benevolence values can be designed. Thus, in the context of social studies, contributions can be made to both the acquisition of the abovementioned values and the development of empathy skills.

CONCLUSION
A model is established toward the associations of empathy, benevolence, friendship, and responsibility. Empathy predicts positively benevolence, friendship, and responsibility. Likewise, friendship and responsibility also predict benevolence. The full mediation effect of friendship and responsibility between empathy and benevolence is significant. As a result, it can be said that empathy, friendship, and responsibility have a worthwhile contribution in becoming benevolent individuals for middle school students in social studies courses.