Asisst. Prof. Dr., Osmaniye Korkut Ata University

Published 2020-09-25


  • Housing Selection Multinomial Probit Regression Model Housing Sector
  • Konut Seçimi,
  • Multinominal Probit Regresyon Modeli,
  • Konut Sektörü

How to Cite

YAKUT, E. (2020). DETERMINATION OF FACTORS AFFECTING HOUSING SELECTION THROUGH MULTI-NOMINAL PROBIT MODEL: CASE OF OSMANİYE PROVINCE OF TURKEY. Business & Management Studies: An International Journal, 8(3), 3274-3301. https://doi.org/10.15295/bmij.v8i3.1598


Housing is a space having a physical asset that is used to meet the need for shelter, one of the basic physiological necessities (Anbarcı, Giran, Türkan and Manisalı, 2011). Independently of individuals' economic status and personal choices, housing need is described as the difference between the number and quality of housing required for individuals to meet their minimum housing needs and the number and quality of housing available at the time of measurement (Chamber of Civil Engineers, 2008). Housing demand, on the other hand, is the concept that emerges as a result of supporting the housing need with purchasing power and purchasing requests (Özlük, 2014). Today, housing demands are not only aimed at meeting the need for shelter, but it is also seen that housing is purchased as an investment instrument and status indicator to provide economic and legal security (Tosun ve Fırat, 2012). Housing is essential in terms of meeting individuals' need for shelter, protecting them against external threats and ensuring the safety of the households (Yıldırım and Başkaya, 2006:285; Çelik and Kıral, 2018:1012). In addition to the individuals' needs for shelter, housing contributes indirectly to people in social and cultural, legal, and technological aspects, while it has a vital position in developing social relations and gaining self-confidence in society (Tosun and Fırat, 2012:176). Factors affecting housing selection are classified under three titles: sociodemographic factors, economic factors, and the structure factor indicating the characteristics of the housing. Apart from these, many variables, including the customs and traditions of the society, can be shown as the factors affecting the choice of housing by individuals; however, the literature review showed that the sociodemographic factors, economic factors, and the features of the housing stand out in general (Çelik and Kıral, 2018:1012).
The purpose of this research is to calculate the change in the probability of housing selection by determining the factors affecting individuals' housing selection in Osmaniye province, Turkey. For this purpose, in our study, we will determine whether the relationship between the factors such as features of the housing, the individuals' sociodemographic and economic factors, and the possibility of housing selection is significant.
We think that the study will contribute to the individuals and contractors who will perform housing selection in Osmaniye province, and we expect that this study will contribute to the literature in terms of application area diversity of the Multinomial Probit model as an alternative to Multinomial Logistic regression analysis, one of the discrete choice models.
In terms of including statistical analysis, the study is a quantitative research study examining the relationship between independent variables and the probability of dependent variable selection.
The problem of the study is to determine the variables that are significant by applying the Chi-square test to the data obtained from the individuals participating in the questionnaire in order to determine the factors affecting the housing choice of the individuals in Osmaniye. In this sense, the problems of the research are to create a prediction model for housing selection by applying a multi-nominal probit regression model and to calculate the marginal effects of factors affecting housing selection to housing data obtained from individuals.
The necessary data set used in the research was obtained by applying a questionnaire with a convenience sampling method to 813 people who demanded housing in Osmaniye province. As a result of the literature review, factors affecting housing selection were found as follows: household income, the mortgage loan (economic factors); gender, age, marital status, educational background, profession, number of family members, average monthly income, and number of children attending school (sociodemographic factors); playground, parking lot, green area, environmental problems, location of the housing, usage area of the house, quality of the material used, age of the building and heating system (the structure factor showing the properties of the housing). The dependent variable of this study is the individuals' housing choice of individuals, and it consists of three categories: 1. Flat, 2. Housing Estate, and 3. Detached House. Independent variables in the study were determined with the help of literature review and factors related to sociodemographic and economic indicators of individuals and features of housing were determined as independent variables. Sociodemographic factors of individuals were gender (male/female), age (31 years and below, 32-40, 41-50 and 51 years and older), marital status (single/married/widow), educational background (secondary school and below/high school/university and above), professional status (civil servant, worker, tradesman/self-employment, other occupational groups), the number of family members (two and below, three, four and above), and the number of children attending school (one, two and above, none). The monthly average income (2500 TL and below, 2501-3500 TL, 3501-4500 TL, 4501 TL and above) was an economic factor. The features of the housing consist of the playground (none/available), parking lot (none/available), green area (none/available), environmental problems (none/available), credit availability (no/yes), location of the housing (close to the downtown/far from the downtown), usage area of the house (120 m2 and below, 121-150 m2, 151 m2 and above), the quality of the materials used (insignificant, neutral, significant), age of the building (0-5 years, 6-10 years, 11-15 years, 16 years and over), the heating system of the housing (coal-fired heating systems, wood coal burning stove, air conditioner and electric heaters, natural gas-operated systems).
In investigating the relationships between dependent and independent variables, if the dependent variable is qualitative, discrete choice models are used (Alpar, 2013:703). When the dependent variable has more than two categories, it is recommended to apply multinomial logistic and multinomial probit regression models. While there is an assumption that the odds ratios of the categories should be independent of other categories in the multinomial logistic model since there is no such assumption in the multinomial probit model, this method can be used in discrete choice models (Greene, 2003:724; Sigeze, 2017: 442). In the multinomial logistic model, while the error terms show a cumulative logistic distribution and there is an assumption that there is no correlation between the error terms; in the multinomial probit model, on the other hand, when the error terms are distributed normally, there may be a correlation between them (Sigeze, 2017:446; Alkan and Yarbaşı, 2020:141).
In the study, regarding the measurement model created with the help of the literature, factors affecting individuals' housing selection were tried to be determined using the multinomial probit regression model. Also, the marginal effects of independent variables estimated from the multinomial probit regression model were calculated.
The three research hypotheses determined by the purpose of the study are as follows:

H1: Sociodemographic factors of individuals have a significant effect on housing selection.
H2: Economic factors of individuals have a significant effect on housing selection.
H3: The structural features of the housing have a significant effect on individuals' housing selection.

As a result of the multinomial probit regression model analysis, factors affecting individuals' housing selection were interpreted by calculating their marginal effect values. By the variable of a playground within the housing, individuals are 50.51% more likely to select the housing state, while they are 13.72% and 36.79% less likely to select a flat and a detached house, respectively. By the variable of a parking lot, individuals are 23.68% more likely to select the housing state, while they are 15.58% and 8.10% less likely to select a flat and a detached house, respectively. By the variable of green area within the housing, individuals are 13.09% and 21.06% more likely to select the flat and detached house, respectively, while they are 34.15% less likely to select a housing estate. In terms of credit availability, individuals are 12.36% more likely to select the housing state, while they are 14.77% less likely to select a detached house. By the variable of usage area of the housing, individuals are 24.89% more likely to select the flat with 120 m2 and below compared to 121-150 m2; while they are 17.50% less likely to select a flat with a usage area of 151 m2 and over, compared to 121-150 m2. In terms of the heating system of the housing, compared to those who use air-conditioner and electric heaters, those who use natural gas-operated heating systems are 14.06% more likely to select a housing estate, while they are 4.67% and 9.39% less likely to select a flat and a detached house, respectively.
It was understood that the H1 and H2 hypotheses were supported, which show that the variables of age, marital status, educational background and an occupational group from sociodemographic factors of individuals, and the monthly average income variable from economic factors have a significant effect on individuals' housing selection. H3 hypothesis, which shows that features of the housing such as a playground, parking lot, green area, the credit availability, the location of the housing, the usage area of the housing, the quality of the materials used and the heating system have a significant effect on individuals' housing selection, was supported.
When the studies on housing selection are examined; the findings of the studies on the relationship between sociodemographic and economic factors and housing selection (Yavuz and Çemrek 2013; Seo and Kwon, 2017; Çelik and Kıral, 2018; Samosir and Su, 2020), and the relationship between the features of the housing and the housing selection (Alkan et al., 2014; Oktay et al., 2014; Shekarian 2015; Olanrewaju and Woon, 2017; Memiş, 2019) coincide with the findings of this study.
Findings obtained as a result of multinomial probit regression analysis within the scope of the study have revealed that there is a significant relationship between the individuals' housing selection and the features of the housing such as the playground, parking lot, green area, the usage area of the housing, and the quality of the materials used in the housing. The classification success levels of the multinomial probit model established in the housing selection estimation of individuals are measured as 93.8% for the flat, 97.6% for the housing estate, 94.6% for a detached house, and 95.8% in total. It was determined that, of the sociodemographic variables, age and educational background variables have a significant effect on the selection of flats and detached houses, while the marital status variable on the selection of flats, the individual's professional status on the selection of detached houses, and the monthly average income levels on the selection of housing estates and detached houses.
In further studies, it is recommended to investigate the housing selection of individuals in other provinces and to compare the results of the studies to be obtained by applying different methods with the results of this study.

The fact that the study was conducted for individuals who were in the selection of subjects in Osmaniye, that individuals were very few in the selection of apartments and that those who were in the selection of villas could not be determined an important limitation of the study. Another limitation of the study is that house prices are not included in the study, and this will be taken into account in future studies. In addition, the fact that multi-nominal probit analysis was applied to the data of the study is another limitation of the study.


Download data is not yet available.


  1. Alkan, Ö., & Abar, H. (2020). Determination of factors influencing tobacco consumption in Turkey using categorical data analyses. Archives of Environmental & Occupational Health, 75(1), 27-35.
  2. Alkan, Ö., & Yarbaşı, İ. Y. (2020). Investigation of Alcohol Consumption Determinants in Turkey With Multinomial Probit Model. In Global Developments in Healthcare and Medical Tourism (pp. 134-161). IGI Global.
  3. Alkan, Ö., Karaaslan, A., Abar, H., Çelik, A. K., & Oktay, E. (2014). Factors affecting motives for housing demand: the case of a Turkish province. Theoretical and Empirical Researches in Urban Management, 9(3), 70-86.
  4. Alpar, R. (2013). Uygulamalı çok değişkenli istatistiksel yöntemler. Ankara: Detay yayıncılık.
  5. Anbarcı, M., Giran, Ö., Türkan, Y. S., & Manisalı, E. (2011). Ürün Olarak Konut Kavramı ve Türkiye’deki Konut Satışlarının Ürün Hayat Eğrisi Yaklaşımıyla Değerlendirilmesi. 6. İnşaat Yönetimi Kongresi, 25-26-27 Kasım, Bursa.
  6. Avcılar, M. Y. & Kara, E. (2015). Şehir Markası Kavramı Ve Marka Şehir Yaratma Stratejilerine Yönelik Literatür İncelemesi. Sosyal ve Beşeri Bilimler Araştırmaları Dergisi, 16(34), 76-94.
  7. Bagheri, A., Habshah, M., & Imon, R. (2012). A novel collinearity-influential observation diagnostic measure based on a group deletion approach. Communications in Statistics-Simulation and Computation, 41(8), 1379-1396.
  8. Bilik, M., & Aydın, Ü. (2019). Konut sahibi olma kararlarını etkileyen faktörler: Lojistik regresyon ve destek vektör makinelerinin karşılaştırılması. Dumlupınar Üniversitesi Sosyal Bilimler Dergisi, (62), 184-199.
  9. Cihangir, M., Sak, N., & Bilgin, S. (2016). Bireysel Yatirimci Demografileri: Osmaniye Ilinde Risk Getiri Tercihlerinin Multinominal Probit Modeliyle Incelenmesi/Individual Investor Demographics: Investigation of Risk Return Preferences in Osmaniye Region with Multinomial Probit Model. Muhasebe ve Finansman Dergisi, (70), 129-142.
  10. Çamoğlu, S. M., & Çakır, E. (2020). Piyasa Değeri ve Mülkiyete Göre Hane halkı Konut Talebi: Ordu Örneği. Suleyman Demirel University Journal of Faculty of Economics & Administrative Sciences, 25(1).111-123.
  11. Çan, M. F. (2010). Kentleşme, Sanayileşme ve Kalkınma Etkileşimi. https://fka.gov.tr/sharepoint/userfiles/Icerik_Dosya_Ekleri/FIRAT_AKADEMI/KENTLE%C5%9EME,%20SANAY%C4%B0LE%C5%9EME%20VE%20KALKINMA%20ETK%C4%B0LE%C5%9E%C4%B0M%C4%B0.pdf
  12. Çelik, C. & Kıral, G. (2018). Panel Veri Analizi ve Kümeleme Yöntemi ile Türkiye’de Konut Talebinin İncelenmesi. Atatürk Üniversitesi İktisadi ve İdari Bilimler Dergisi, 32(4),1009-1026.
  13. Es, M. & Ateş, H. (2004). Kent yönetimi, kentlileşme ve göç: sorunlar ve çözüm önerileri. Sosyal Siyaset Konferansları Dergisi, (48), 205-248.
  14. Ezennia, I. S., & Hoskara, S. O. (2019). Exploring the severity of factors influencing sustainable affordable housing choice: evidence from Abuja, Nigeria. Sustainability, 11(20), 5792-5817.
  15. Gölbaş, A. (2016). Yerleşim Yerlerinin Seçiliş Biçimleri: Neolitik Kıyı Yerleşmelerine Bir Bakış. Journal of International Social Research, 9(46), 317-327.
  16. Greene, W. H. (2003). Econometric analysis. UK: Pearson Education India.
  17. Güneş, C., Birecikli, Ş. Ü., & Yıldırım, Z. (2017). Çalışan Kadının Çocuk Bakım Tercihi: Havuzlanmış Verilerle Kesikli Tercih Modellerinin Karşılaştırılması. SGD-Sosyal Güvenlik Dergisi, 7(1), 123-144.
  18. Güven, A. (2016). Kent, Kentlileşme ve Kentsel Yönetim İhtiyacı. Journal of International Management, Educational and Economics Perspectives, 4(1), 21-30.
  19. Hayri, A. & Karaaslan, A. (2013). Konut Talep Edenlerin Özellikleriyle Talep Edilen Konutun Özellikleri Arasındaki İlişkinin Çoklu Uyum Analizi Yöntemi İle İncelenmesi: Atatürk Üniversitesi Personeli Örneği. Atatürk üniversitesi iktisadi ve idari Bilimler Dergisi, 27(3), 323-339.
  20. İnşaat Mühendisleri Odası (2008). “Türkiye’de Konut Sorunu ve Konut İhtiyacı” Raporu. http://www.imo.org.tr/resimler/dosya_ekler/9ca6617c167713d_ek.pdf
  21. Kährik, A., Temelová, J., Kadarik, K., & Kubeš, J. (2016). What attracts people to inner city areas? The cases of two post-socialist cities in Estonia and the Czech Republic. Urban studies, 53(2), 355-372.
  22. Kaya, E. (2003). Kentleşme ve kentlileşme. İstanbul: İlke Yayınevi.
  23. Keleş, R. (1995). Kentleşme ve Türkçe. Dilbilim Araştırmaları Dergisi, 6, 1-5.
  24. Memiş, S. (2018), Tüketicilerin Konut Seçimini Etkileyen Faktörlerin Belirlenmesine Yönelik Bir Araştırma. International Journal of Academic Value Studies, 4(20), 652-665.
  25. Memiş, S. (2019). Tüketicilerin Konut Tercihini Etkileyen Faktörlerinin AHP İle Ölçülmesi: Giresun İli Örneği. Avrasya Uluslararası Araştırmalar Dergisi, 7(16), 783-796.
  26. Oktay, E., Karaaslan, A., Alkan, Ö., & Çelik, A. K. (2014). Determinants of housing demand in the Erzurum province, Turkey. International Journal of Housing Markets and Analysis. 7(4), 586-602.
  27. Olanrewaju, A., & Woon, T. C. (2017). An exploration of determinants of affordable housing choice. International Journal of Housing Markets and Analysis. 10(5), 703-723.
  28. Ören, K. & Yüksel, H.. (2013). Türkiye’de konut sorunu ve temel dinamikleri. Süleyman Demirel Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, (18), 47-84.
  29. Özer, H. (2004). Nitel değişkenli ekonometrik modeller: Teori ve bir uygulama. Ankara: Nobel Yayınevi
  30. Özlük, S. (2014). Türkiye'de konut sektöründe talep ve arzı belirleyen faktörler. (Yayımlanmamış doktora tezi). Maltepe Üniversitesi Sosyal Bilimler Enstitüsü, İstanbul
  31. Samosir, A. C. E., & Su, H. T. (2020). Analysis of Factors Affecting Home Demand in Deli Serdang District. International Journal of Research and Review, 7(1), 458-464.
  32. Seo, D., & Kwon, Y. (2017). In-migration and housing choice in Ho Chi Minh City: Toward sustainable housing development in Vietnam. Sustainability, 9(10), 1738-1755.
  33. Shekarian, E. (2015). A novel application of the VIKOR method for investigating the effect of education on housing choice. International Journal of Operational Research, 24(2), 161-183.
  34. Sigeze, Ç. (2017). Multinomial Probit Modelinde Bayes Yaklaşımı: Türkiye'de Yağ Tüketim Tercihinin İncelenmesi. Trakya University Journal of Social Science, 19(2). 441-459.
  35. Tosun, E. K., & Fırat, Z. (2012). Kentsel mekândaki değişimler ve kişilerin konut tercihleri: Bursa örneği. Business and Economics Research Journal, 3(1), 173-195.
  36. Türkiye İstatistik Kurumu [TÜİK] (2020). Konut Satış İstatistikleri.
  37. Yavuz, S., & Çemrek, F. (2013). Konjoint Analizi İle Sağlık Çalışanlarının Konut Tercihlerinin Belirlenmesi. Atatürk Üniversitesi Sosyal Bilimler Enstitüsü Dergisi, 17(2), 379:396.
  38. Yıldırım, K., & Başkaya, A. (2006). Farklı Sosyo-Ekonomik Düzeye Sahip Kullanıcıların Konut Ana Yaşama Mekanını Değerlendirmesi. Gazi Üniversitesi Mühendislik Mimarlık Fakültesi Dergisi, 21(2), 285-291.
  39. Yilmaz, H., & Tosun, Ö. (2020). Aylık Konut Satışlarının Modellenmesi ve Antalya Örneği. Kafkas Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, 11(21), 141-158.