MULTİNOMİAL LOGİT MODEL İLE HANEHALKININ OKUL DIŞI EĞİTİM TÜRÜ TERCİHİNE ETKİ EDEN SOSYOEKONOMİK FAKTÖRLERİN BELİRLENMESİ DETERMINATION OF THE SOCIOECONOMIC FACTORS AFFECTING THE HOUSEHOLD'S OUT-OF-SCHOOL ECUCATION INSTITUTION PREFERENCE THROUGH MULTINOMIAL LOGIT MODEL

Education is the most important power that improves the quality of life. For this reason, families resort to different ways, such as sending their children to private courses, taking private lessons for their children, or getting them to participate in courses and to receive one to one tutorial at schools for reasons such as providing their children with a better education, helping them learn what they haven't been able to learn in school education services. In the current that the head of the household takes for his/her child for such purposes. To this end, B micro data set from the "Life Satisfaction Survey" conducted in 2018 was used. The sample of the current study is comprised of 336 households. In the analsis of the collected data, the Multinomial Logit Model was used. As a result of the estimation made with the Multinomial Logit Model, it was found that different socio economics factors were effective on the choice of type of courses. In the selection of the private course, the age, marital status and education variables were found to be effective; in the selective of school course, the age variable was found to be effective; in the selection of the private tuition, the size of household and marital status were found to be effective; in the selection of the remedial courses, the income variable was found to be effective and in the selection of remedial courses together with school courses, the age and education variables were found to be effective.


Literature Review
Education is the main tool that adds value to the individual and society. In this regard, families strive for their children to receive a better education. This endeavour drives parents towards seeking alternative education services outside the school. Out-of-school alternative education services can be defined with names such as complementary education, supportive education or additional education. Although these formations vary depending on the education policies of countries, they are generally in the form of courses and studies offered in private teaching institutions, courses or schools or in the form of private tutoring. These alternative forms of additional education taken outside the school system are preferred for many different reasons, ranging from teacher to school, from student to parent and from the way the education system is constructed to the examination system. According to the results of the existing research, the competition-induced pressure in education systems and the inadequacy of education in the school to cope with this pressure direct parents and students to alternative ways in education (Bray and Kwo, 2014;Pallegedara and Mottaleb, 2018). Low quality of education, low number of qualified teachers (Zhang and Bray, 2018), crowded classes (Darling-Hammond, 2000), the desire to eliminate lesson or subject-based deficiencies, to reinforce learning (Çokgezen and Terzi, 2008) and to maintain the high level of success (Bray, 1999) increase the demand for educational services carried out in parallel with formal education. Out-of-school education that students can receive may vary depending on the economic status of families (Tansel, 2013). Students with a high socioeconomic level can participate in both more and better quality out of school education services (Bray, 2006;Tsiplakides, 2018;Zhang and Bray, 2018). In this respect, it is recommended to examine the socioeconomic background of those who attend out-of-school education Multinomial Logit Model İle Hanehalkının Okul Dışı Eğitim Türü Tercihine Etki Eden Sosyoekonomik Faktörlerin Belirlenmesi -Determination Of The Socioeconomic Factors Affecting The Household's Out-Of-School Ecucation Institution Preference Through Multinomıal Logit Model Kübra ÖNDER, Emine ÖNDER 1123 services (Tansel, 2013). Thus, in the current study, it is aimed to determine the socioeconomic factors that are effective in choosing the types of out-of-school education the child receives.

Methodology
Individuals' preferences are analysed by many disciplines using various models. The method widely used by these disciplines is the Multinomial Logit Model (MNLM), which is one of the unordered choice models developed based on random utility maximization models (Greene, 2008: 842;McFadden, 1974). The advantage that distinguishes MNLM from other models is the simplicity it provides in interpreting the estimation and choice possibilities and flexibilities obtained as a result of the estimation (Sarrias and Daziano, 2016: 1). MNLM is the extended form of the binomial logistic regression to allow more than two dependent variables or outcome variables. Independent variables in the model can be binomial or continuous (Starkweather and Moske, 2011:1). The probability of choice within MNLM is linked to the individual choice process and the random utility maximization process that affects this process. For more detailed information about the framework in which choice models are developed, see McFadden (1974).

Results and Conclusions
In the current study, considering the model estimation results, the marginal effect (flexibility) value for each of the independent variables found to be statistically significant was also calculated.
When the calculated marginal effect values are examined, it is seen that the preference of the head of the household to send his / her child to a private course rather than other alternative out-of-school education institutions is affected by the independent variables of age, marital status and education.
Although the degree of effect varies according to the independent variable, the greatest effect belongs to the secondary school and lower education level subcategory of the education independent variable.
Individuals with secondary school and lower education level prefer sending their children to a private course approximately 21% more than individuals with the other education levels. The marital status independent variable also has a positive contribution (14.6%) to the child's enrolment in a private course if the head of the household is married. The age of the head of the household has a negative effect on the child's attendance to a private course. With the increasing age of the head of the household, the preference for sending the child to a private course rather than other alternative out-ofschool education institutions decreases. Individuals in the age groups of 31-40 and 41-50 prefer a private course rather than other alternative out-of-school education institutions about 5% less than the others. When the preference of the head of the household to send his / her child to a course organized by a state school rather than other types of out-of-school education is analyzed, it is observed that the choice of sending the child to a course organized by the school is affected by the home ownership and age independent variables. Individuals having their own houses prefer to send their children to out-ofschool education institutions other than the courses organized by a state school. The rate of this preference is 20.38%. As in the preference for a private course, the age independent variable affects the preference for a course organized by a state school negatively. While the size of the effect is the Mehmet Akif Ersoy İktisadi ve İdari Bilimler Fakültesi Dergisi -Mehmet Akif Ersoy University Journal of Economics and Administrative Sciences Faculty Cilt: 7 Sayı: Özel Sayı s.1121-1141Volume: 7 Issue: Special Issue p.1121-1141 Aralık 2020 December 1124 highest in the age group of 31-40 (0.1250), it is smallest in the age group of 18-30 (0.0257). The preference for private tutoring is negatively affected by the size of the household and marital status (single) independent variables. In the families where the household size is larger than the Turkish average (3.2), the child's likelihood of having private tutoring decreases and other out-of-school education institutions are preferred. As the per-capita income will decrease with the increasing number of the members in the family and as private tutoring is more expensive than other alternatives of outof-school education, the heads of the households prefer private tutoring 20.2% less than other alternatives of out-of-school institutions. Another independent variable affecting the preference of the head of the household for private tutoring for his/her child is his/her being single. Single individuals prefer private tutoring 05.7% less than married individuals. Only the income independent variable was found to be effective in the preference of the head of the household for remedial courses (in the 2500-3499 TL monthly income sub-category). The heads of the households whose monthly income is in this income sub-category prefer their children to attend remedial courses 3.48% more than other out-ofschool education services. In the preference for remedial courses together with private courses, the age of the head of the household was found to be effective as in other alternative out-of-school education services. The individuals in the age groups of 31-40 and 41-50 were found to be not willing to make children take remedial courses together with private courses. The probability values belonging to the individuals in these age groups are -04% and -03.7%, respectively The result of this analysis complies with the result of the analysis conducted on the preference for a private course. The preference of the individuals for their children to take remedial courses together with the courses offered by a state school is affected by the age independent variable and the education level independent variable. The individuals whose education level is secondary or lower prefer their children to take remedial courses together with school courses by 19.97% probability. The probability value for the parents who are in the age group of 18-30 to make their children take remedial courses together with school courses is -03.7% and -06% for the parents who are in the age group of 31-40. Private courses are preferred by parents to reinforce what has been learned at school and to prepare for exams and seen to be a part of the education process today. Important reasons that come to the fore from among the reasons why the heads of the household send their children to alternative out-of-school education institutions include the following: the exam system is more comprehensive than the school curriculum and schools cannot completely meet students' needs for preparing exams and the individual needs of students can be better addressed with the help of the guidance provided in out-of-school education institutions. The factors found to be the most influential on the preference for out-of-school education institutions are the education level of the head of the household, his/her income level, age, marital status and size of the household and house ownership. As the proportion of out-of-school education expenditures within the family budget increases, the demand of the head of the household for alternative out-of-school education for their children decreases. This cost-related issue can lead to inequality of rights in education among children. The number of the members in a family and house ownership are other institutions. These variables positively affect the preference of the head of the household for sending their children to out-of-school education institutions. In families whose household size is under the average in Turkey, the amount of income that can be spent for each family member increases.
Moreover, house ownership positively affects the amount of spendable income for each family member because the family do not have to pay rent and this has positive effects on the family budget.
The age of the head of the household is another independent variable affecting the preference for outof-school preference. While the individuals in all the age groups prefer the courses offered by a state school, the individuals in the age group of 31 and over also prefer alternatives together with the school courses. This might be because of the experiences of the other family members or the experiences lived in relation to other children in the family or other children in the close surrounding. The survey questions used in the current study conducted all over Turkey should be expanded to include variables such as the employment status of the mother, the father and mother's job, the mother's education level, per-capita income in the family, the number of people who earns money in the family and the number of children studying in the family and the new survey to be developed should be administered in provincial level and to different income groups. In this way, both more detailed information could be obtained about the existing state and new insights could be provided for policy makers to develop educational policies of the future.  (Bray, 1999). Yukarda ifade edilen etmenlerle birlikte ders ya da konu bazlı eksiklikleri giderme ya da öğrenmeleri pekiştirme (Çokgezen ve Terzi, 2008) ve ebeveynlerin çocuklarının ev ödevlerine nitelikli katkı sağlaması (Jayachandran, 2014) amaçları ile de örgün eğitime paralel olarak yürütülen eğitim hizmetlerine olan talep yükselmektedir.