Does University Effect Gender Perceptions of Teacher Candidates in Turkey?: Reasons and Implications

The aim of this study is to reveal the opinions of teacher candidates about gender perception according to gender, university and department variables. Survey model was used to reveal the current situation in the research. The study group consisted of 269 teacher candidates who are studying three different universities in Turkey. In the 2018-2019 academic year. The data of the study were collected using the “gender perception scale”, which consists of expressions about gender roles for women and men and developed by Altınova and Duyan (2013). According to the research findings, the gender variable has a high effect on the participants' perception of gender, and women's gender perception is more positive. The university variable has a medium effect on gender perception. The pre-service teachers' perception about gender is higher in the university, the foundation year of which is older. The variable of department has a low effect on gender perception. Based on the finding of the effect of the university variable that emerged in this research, gender equality index can be examined for all universities

ender roles may be defined as roles that society puts on women and men. This means society is a social structure that shapes the behavior of women and men and is a determining factor on the roles expected from women and men. This structure creates judgments on the social roles of women and men and puts pressure individuals to act in accordance with these judgments. This pressure causes gender-based inequalities over time. At this point, gender-based inequalities in the structure of society occupy an important place today.
The World Health Organization (WHO, 2020) defines gender as the result of socially constructed characteristics of women and mensuch as the norms, roles and relationships that exist between them. Gender expectations vary between cultures and can change over time". Accordingly, gender is a concept of being biologically female or male. Gender-based inequalities seen in social life are not related to biological differences, but rather to the fact that gender roles contain values, expectations and judgments. Connell (1987) suggests that social relations that harbor gender are not determined by biological difference, but are linked to it. In other words, Connell claims that there is a very practical connection rather than a reduction and defines gender as follows: "gender means organized practice within or in connection with the division of people in a masculine and feminine way." (Connell, 1987, 208). The meaning and expectations that societies put into the female and male genders constitute the patterns of behavior that women and men should demonstrate, and these patterns of behavior may vary according to the cultural structures of the societies, and gender roles emerge. Gender roles are defined as roles attributed to individuals based on the common beliefs of society and determined by gender definitions (Eagly, 2009). How society sees, perceives, thinks about women and men as individuals and expects certain behaviors from them reinforces the gender roles (Atli, 2017; Aydin Ozkan & Simsek Kucukkelepce 2019).
In Western societies, men are considered competent, assertive, dominant and success-oriented, while women are considered friendly, warm, unselfish, social, dependent, emotionally demonstrative and relation-focused (Eagly & Mladinic, 1989; as cited by, Eisenchlas, 2013). In Turkey, the society structure is mainly under the domination of men, like in other Islamic societies. Men are the heads of the families and women are in secondary positions within the G 261 Taşkin & Nayir-Does University Effect Gender Perceptions of Teacher Candidates in Turkey?
families (Ali et al., 2011;Adil, Shahed and Arshard, 2017). It is possible to see the impact of Islam on the construction of gender in other Islamic countries such as Indonesia, Pakistan (Adriany, 2019;Ali et al., 2011;Adil, Shahed and Arshard, 2017). The expression "being a woman in the Middle East brings negative cultural features with it" of Kandiyoti (1988) is supportive of this situation. However, while schools seem to be impartial, they are the places where male domination is maintained and certain gender definitions, relations and differences in schools are easily transferred (Arnot, 1982). It is possible that the social structure in Turkey, in which men are dominant and women are in the secondary position, can be transferred to generations in schools through education (Aydemir Uslu, 2015), because there is little practice to prevent the reconstruction of ongoing social class and gender discrimination in schools today (Sayilan, 2012). Education systems training programs continue to play a role in the reconstruction of dominant sexist judgments and values through textbooks and the cultural climate in schools (Sayilan, 2012). Such reconstruction takes place through teachers, being one of the elements that enable the continuation of the existing structure within the education system, and teachers are key to gender equality (Yasar, 2018). Students also take teachers as role models and are influenced by their behavior, attitudes and approaches (Agarwal and Shukla, 2017). As a result of this influence, when teachers repeat dominant sexism in schools, they transfer these dominant sexist behaviors to their students, thereby contributing to the said reconstruction. Similarly, Tan (2007) suggests that teachers are the most important player of sexism in Turkey, especially in the secondary education process, that they reinforce gender discrimination by their expectations and behaviors, and that they have an impact on the orientation of students towards certain professions and gender roles.
When literature is examined, it is found that studies are being conducted to determine the perceptions of high school and university students about gender roles. These studies demonstrate that students adopt traditional roles in gender roles, but female students adopt traditional roles less when compared to male students. Studies in Turkey reveal that male students have a more traditional view of gender roles than female students (Vefikulucay, Zeyneloglu, Eroglu and Taskin, 2007;Dinc and Caliskan, 2016). On the other hand, it is seen in the literature that there are studies that examine the opinions of teacher candidates on gender roles. However, these studies often focus on a single teaching field (Yazici, 2016;Bayraktar and Yagan Guder, 2019;Secgin and Tural, 2011). Whereas in a study, the opinions of teacher candidates are examined based on socio-demographic factors (Esen, Soylu and Sagkal, 2019). However, the opinions of teacher candidates are not examined based on university and department variables. In fact, institutions that educate teachers are the first socialization institutions in which teacher candidates start to understand their social identities (Arikan, 2005). Students are expected to have an egalitarian perspective about gender roles and gender equality when they graduate (Egelioglu Cetisli, Top and Isik, 2017). For teacher candidates, the institution that will bring in this perspective is the universities. The studies demonstrated that institutions that educate teachers reinforce the existing traditional roles through their teaching methods, programs, books used during the courses and the attitudes of the academic members they employ and do not create awareness of gender roles (Lumadi and Shongwe, 2009;Zittleman and Sadker, 2002;Esen, 2013). It is stated in the report published by the United Nations International Children's Emergency Fund (UNICEF, 2003) that the environment in which the teachers in Turkey are raised and the education they receive prevents them from questioning gender roles formed in society. Therefore, it may be suggested that the institutions that educate teachers have a significant impact on the perception of gender roles. At this point, it is important to examine the opinions of teacher candidates who will start their teaching professions regarding gender roles especially based on university variables. The purpose of this study, in this context, is to demonstrate the opinions of teacher candidates on gender perception based on gender, university and department variables.

Method
The general survey model is used in the study to demonstrate the current situation. The general survey model is the research model, which is conducted with a certain group in order to reach a judgment on the universe related to the research topic (Karasar, 2005).

Study group
The study group of the research consists of 269 teacher candidates who were studying at three different universities in Turkey in the 2018-2019 academic year. The year of establishment of universities and their social and cultural structures are taken into account in the selection of universities in the study group. The first of the preferred universities was established in 1946, the second in 1992 and the third in 2007. When selecting universities, the population of the city where the university is located and its social and cultural nature are examined. Then, the first level of Nomenclature of Territorial Units for Statistics (NUTS, 2002), which consists of 12 territories, is taken into account at the first level. The city, where the first university is located, has a population range of 3 -6 million and is located in the Western Anatolia Region (TR5). The city, where the second university is located, has a population range of 1 -3 million and is located in the Aegean Region (TR3). The city, where the third university is located, has a population range of 100 thousand -1 million and is located in the Western Black Sea Region (TR8). The Demographic information of participants of the study is provided in Table 1. As shown in Table 1, the majority of participants (79.6%) are women. The distribution of the number of participants based on universities is very close to each other. 26.8% of participants are students in pre-school teaching, 17.8% in intellectual disabled teaching, 15.6% in mathematics teaching, 7.8% in Turkish teaching and 26% in teaching in other fields. Departments of teaching in other fields are music, painting, tourism and hotel management, textile and fashion, and science.

Data Collection Instrument
The research data is collected using the "gender perception scale", which is comprised of statements about gender roles for women and men, developed by Altinova and Duyan (2013). A five-point Likert scale is used, ranging from the most negative to the most positive: "1-I strongly disagree" and "5 -I strongly agree". The reliability studies of the scale is conducted over 443 adults. The Cronbach's Alpha coefficient of the scale is calculated as .87. Accordingly, the scores that can be obtained from the scale are in the range of 25-125, with high scores indicating that gender perception is positive (Altinova and Duyan, 2013). In this study, the Cronbach's alpha coefficient is calculated as .92.

Data Analysis
Before starting the analysis, the distribution of the data is analyzed and it is found that the kurtosis and skewness coefficients of all data are between -2, +2 in all independent variables, that the arithmetic mean and median values are close to each other, and that the distribution is normal. In this regard, the t-test is used in the analyzes made based on gender variable and ANOVA in the analyzes made based on the university variable. In the case of a significant variation in ANOVA analysis, post hoc tests are performed to determine between which groups the variation occurs. Accordingly, the Scheffe test is used in case the variations are equal, but the number of groups is not, and the

Taşkin & Nayir-Does University Effect Gender Perceptions of Teacher Candidates in Turkey?
Dunnet C test in case the variations are not equal. Kruskal-Wallis test is used for the department variable, where the number of groups is less than 30. Mann Whitney U-test is conducted to identify the source of variation if there is significant variation in these groups. The significance level in gap analyzes is determined as .05. During the analysis, the first university is identified by the letter A, the second university by B and the third university by C. In Case of significant variation, the effect sizes of the significant variation are calculated. The effect size gives information about the percentage of variation described by the independent variable and ranges between 0 and 1 (Kilmen, 2015). The exact points for the calculated effect size for the T-test and ANOVA are 01 low, .006 moderate and .138 high, and for the Mann Whitney U-test are .01 low, .03 moderate and .05 high (Cohen, 1988; as cited by Kilmen, 2015).

Results
When examining the descriptive statistics of the scale scores of teacher candidates, the lowest score obtained from the scale is found to be 59, the highest 125, and the mean 105.83 (sd=14.82). The weighted mean for the scale is calculated as 4.23. The high total score obtained from the scale shows that the perception of gender roles is positive. At this point, it may be suggested that the perception of teacher candidates regarding gender roles is positive. When the opinions of teacher candidates are examined according to the statements about family life, 1.9% of women stated that they strongly agree with the statement "man should always be the head of the family", and 4.7% agree, while 21.8% of men stated that they strongly agree and 18.2% agree. Similarly, 1.4% of women stated that they strongly agree with the statement "the primary duty of a woman is to do housework," and 2.3% agree, while 7.3% of men stated that they strongly agree and 14.3% agree. 1.9% of women stated that they strongly agree with the statement "man should be the one making the important decisions in the family", and 2.8% agree, while 7.3% of men stated that they strongly agree and 20% agree. 1.9% of women stated that they strongly agree with the statement" a woman should not work after giving birth", and 4.2% agree, while 1.8% of men stated that they strongly agree and 7.3% agree.
When the statements about work-life of women are examined, 4.7% of women stated that they strongly agree with the statement "a woman should not work if her husband does not allow her to" and 16.8% agree; while 14.5% of men stated that they strongly agree and 12.7% agree. 1.9% of women stated that they strongly agree with the statement "man should be the one earning bread for the family", and 3.3% agree, while 7.3% of men stated that they strongly agree and 12.7% agree. 1.4% of women stated that they strongly agree with the statement "a woman should not earn more than her husband", and 2.3% agree, while 3.6% of men stated that they strongly agree and 10.9% agree. 3.3% of women stated that they strongly agree with the statement "a working woman should give her income to her husband", and 3.3% agree, while 5.5% of men stated that they strongly agree and 7.3% agree.
1.9% of women stated that they strongly agree with the statement "women should always be protected by men", and 2.3% agree, while 27.3% of men stated that they strongly agree and 18.2% agree. 2.3% of women stated that they strongly agree with the statement "a woman without a husband is like a house without an owner", and 3.7% agree, while 9.1% of men stated that they strongly agree and 10.9% agree. 2.8% of women stated that they agree with the statement "community leadership should generally be in the hands of men", while 5.5% of men stated that they strongly agree and 18.2% agree. 2.8% of women stated that they strongly agree with the statement "a woman should be younger than her husband", and 13.6% agree, while 9.1% of men stated that they strongly agree and 12.7% agree. The results of the t-test analysis conducted about the opinions of teacher candidates based on gender variables are provided in Table 2. .19. This means that 19% of the variation in gender perceptions of teacher candidates is explained by gender variable. In other words, the gender variable has a major impact on gender perception. The results of ANOVA conducted about the opinions of teacher candidates based on university variables are provided in Table 3. According to Table 3, there is a significant variation between the opinions of teacher candidates on gender perception based on university variable [F(2; 265)=14.02; p<.05]. The Scheffe analysis, which is conducted to determine among which groups the variation occurs, determined that the perception of the gender roles of teacher candidates studying at university A is higher than those of teacher candidates studying at universities B and C. The effect size value for the significant variation is calculated as .09. This indicates a moderate effect size. Accordingly, 9% of the variation in gender perceptions of teacher candidates is explained by university variable. The result of Kruskall Wallis analysis conducted about the opinions of teacher candidates based on department variable is provided in Table 4. perspective on gender roles is that women and men share responsibilities equally in all areas of life, especially in family and professional life (Zeyneloglu, 2008). Although the research findings show high perceptions of gender roles among teacher candidates, the opinions on the statements show striking results. It is seen that some of the male teacher candidates think that man should always be the head of the family, that the primary duty of a woman is to do housework, that man should be the one making the important decisions in the family, and that a woman should not work after giving birth. All of these opinions are a reflection of traditional thinking. In their study with university students, Esen, Siyez, Soylu and Demirgurz (2017) achieved similar results.
Another striking conclusion in the research findings is that some of the male teacher candidates think that a woman should not work if her husband does not allow her to. The interesting point is that female teacher candidates have a similar idea in this regard. Similarly, some of the male teacher candidates think that man should be the one earning bread for the family, that a woman should not earn more than her husband, and that a woman should give her income to her husband.
Another interesting outcome of the research results is related to the opinions of male teacher candidates on the position of women in social life. Some of the male teacher candidates think that women should always be protected by men, that a woman without a husband is like a house without an owner, that community leadership should generally be in the hands of men, and that a woman should be younger than her husband. Again, interestingly, some of the female teacher candidates also agree with those opinions.
When all these results are evaluated together, it may be suggested that the traditional view about gender roles is dominant among teacher candidates. The findings of the research demonstrate that men have a more traditional perspective (Ersoy, 2009;Esen, Siyez, Soylu and Demirgurz, 2017;Askari, Liss, Erchull, Staebell and Axelson, 2010;Unal, Tarhan and Curukvelioglu Koksal, 2017;Aslan, 2015); and according to Ersoy (2009), cultural aspects influence this perspective. As traditional roles, women have been given roles such as childbirth and child-raising, doing housework, making compromises to keep the family together, while men have been given roles such as working for their families, making a living as the head of the family, and controlling the money (Zeyneloglu, 2008;Mingione, 2010). This also reflects in the worklife of women. According to the 2019 data of the Turkish Statistical Institute, the participation of women in the labor force is 34.3%, whereas the ratio for men is 71.1%. The majority of women stated that they do not participate in the labor force because of housework (TUIK, 2020 as cited in Erikli, 2020). Again, according to 2019 data, 38.41% of public employees are female and 61.9% are male (Woman in Turkey, 2019; as cited by Erikli, 2020). Although women participate in work life, their traditional roles continue to exist. In fact, this perspective is also accepted among teacher candidates who will transform and educate society. However, given the assumption that teachers are models for students in gender roles (Brophy, 1985), the results of the research are very thought-provoking.
According to research findings, the gender variable has a major impact on the gender perception of the participants, and the gender perception of women is more positive. The research found that gender perception of women is higher (Yilmaz Vefikulucay et al., 2009;Arici, 2011;Kizilaslan andDiktas, 2011, Secgin andTural, 2011;Esen, 2013;Yazici, 2016;Esen et al., 2019;Dasli, 2019;Ozden and Golbasi, 2018;Adachi, 2018). Bayraktar and Yagan Guder (2019) emphasized that this is because women are most affected by gender inequality. Women are discriminated and may suffer from inequality because women are often given roles such as being responsible for housework and not being active in work-life (Zeyneloglu and Terzioglu, 2011). Gender equality is the equality of women and men in terms of practices and the ability of women and men to access resources and services equally within the community (Yilmaz Vefikulucay et al., 2009). In the report on Gender Discrimination of the World Economic Forum (2018), Turkey ranks 130th in gender inequality among 149 countries. Given that the traditional perspective of men about gender roles is important in the emergence of inequality (Yilmaz et al., 2009), it is better understood that women have a more positive perception of gender roles. However, there are also studies that show otherwise. For example, a study conducted in Delhi found that teachers generally embrace traditional gender roles and that there is no significant variation between the opinions based on gender variables (Agarwal and Shukla, 2017). 108th ranking of India among 149 countries in the Report on Gender Discrimination of World Economic Forum (2018) Esen, 2013) states that men are members of the advantageous group and that even if they encounter courses that question male discrimination, they cannot digest the content of the course and see the critical analyzes made in the course as an attack towards them. In short, gender plays a major role in the adoption and formation of gender roles. Recent research found that the gender variable has a major impact on gender perception. According to research findings, the university variable has a moderate impact on gender perception. Teacher candidates studying at the university that was established at an earlier date have a higher gender perception. This can be explained by the culture of the university. In the organizational culture studies, the distribution of the positions of women and men in the organization is important (Coban, 2005). 52% of the total number of students at university A in the research sample is female, while this ratio is 47% at university B and C. Similarly, according to January 2020 data where the number of academic staff in universities is analyzed, 52% of the academic staff at university A is female, while 46% are female at university B and 44% are female at university C. In terms of titles, 43% of the number of professors at university A, 48% of the number of associate professors, and 56% of the number of academic staff with the doctor title is female. 30% of the number of professors at university B, 45% of the number of associate professors, and 45% of the number of academic staff with the doctor title is female. 18% of the number of professors at university C, 38% of the number of associate professors, and 44% of the number of academic staff with the doctor title is female (YOK, 2020). In the 2018-2019 academic year, the rate of female rectors in public universities was 4.7% and in private universities, 20% (TUIK, 2018 as cited in Erikli, 2020).
According to Hatunoglu, Hatunoglu, and Avci (2014), "the way women and men participate in the social life in a society, the ratio of their participation, their visibility in social life, and their representation in the society are influenced significantly by the gender perception in that society." Accordingly, it may be suggested that the higher number of women at university A causes women to be more visible and they are being represented in higher titles cause them to be role models, thus having a positive effect on the perception of gender roles. However, the size of the city where the university is located may have also been effective in this situation. The research found that people in larger cities have a more positive perception of gender roles than those in smaller cities (Esmer et al., 2012). Similarly, Esen, Soylu, and Sagkal (2019) found the perception of gender roles of teacher candidates living in big cities more positive. Ince Yenilmez (2015) describes this situation with greater awareness about freedoms and easier pursuit of innovative trends of the urban inhabitants.
According to the latest findings from the study, the department variable has a low impact on gender perception. Although the gender perception of teacher candidates studying in the fields of preschool teaching and intellectual disabled teaching is higher, the department variable has a low impact on this perception. A study by Catalcali Soyer (2009) on books for the preschool age group demonstrated that stories contain the types and expressions to set modern society understanding that emphasize emotional and social solidarity rather than traditional life and relations. The use of such books by students in the preschool teaching department as teaching materials during their teaching practices may have lead to an increase in their gender perception. In addition, the fact that the majority of the students studying in the preschool teaching department are women may also have affected the research findings. Gender perceptions of the students studying in the intellectual disabled teaching department may have improved because they work with the vulnerable portion of the society that needs special education due to their mental barriers.

Conclusion
This study examined the opinions of teacher candidates about gender roles and demonstrated that the perception of gender roles of the participants is positive and that male teacher candidates adopt traditional gender roles more. However, when examined based on gender variable, it is observed that female teacher candidates have a more egalitarian gender perception and that gender variable has a major impact on this perception. When examined based on the university variable, the perception of gender roles of teacher candidates studying at the university of the research group located in the largest city is 273 Taşkin & Nayir-Does University Effect Gender Perceptions of Teacher Candidates in Turkey? more positive, and the university variable has a moderate impact on this perception. The department variable has a low impact on the perception of gender roles of teacher candidates.

Recommendations
The opinions of teacher candidates on gender roles based on gender, university, and department variables are examined in the present study. According to research findings it can be recommended that creating an organizational culture based on gender equality in the universities that educate teacher candidates will have a positive impact on the perception of gender roles. Besides, rather than opening a university in every city, enriching the long-established universities in big cities in terms of budget, academic staff, laboratories, libraries, venues where social activities may be held, etc., may provide a versatile contribution to the society in the long run. Further in-depth research may be carried out by handling each variable one at a time. Based on the finding of the effect of the university variable, the gender equality index of all universities in Turkey may be investigated.