INVESTIGATION OF CONVICT DATA BY TWO STAGE CLUSTER ANALYSIS
Abstract
Various studies have been recently conducted so as to explain crime phenomenon. In most of these studies crime phenomenon was tried to be explained on the axis of socio-cultural, economical and environmental factors and various statistical techniques were used with this aim. It can be said that the studies generally focus on crime committing reasons and socio-demographical properties of prisoners.
The aim of this study is to evaluate the sub-groups belonging to prisoners demonstrating similar properties, in other words to determine and compare their profiles of homogenous prisoner groups. In the study data from prisoner registry records of a prison was used. The basic information of the prisoners who were discharged from the prison or were transferred to another prison was analyzed. The variables under the scope of the study are gender, age, crime type and the territory where crime is committed. (Gendarmerie or police)
In the study first of all frequency tables (cross sectional tables) and descriptive statistics were evaluated. As the variables used in the study are in quantitative and qualitative type, two-stage cluster analysis (among the analysis techniques with multiple variables) was found to be suitable for determination of the homogenous sub groups among the prisoners. At the end of the application, number of groups was determined to be 4. Similar and different properties were determined among the formed groups. Clustering quality was obtained at medium level.
In the next stage of the study, analysis was repeated by reducing the number of categories of crime type variable. The reason of this application is to classify the crime types in different styles and that most common classification among various classifications is handling the crime types in 4 groups. (crimes against government, crimes against people, crimes against property, crimes against community) Although this process means to renounce from detailed information, analysis was repeated by taking this classification into consideration so as to make more general comments and to compare the findings. While 4 groups were obtained as a result of the application of previous two-stage cluster analysis in which crime types were handled in detail, it was observed that 3 groups were obtained when crime types were handled in general and incorporated o the analysis. It was observed that clustering quality was good. As a result handling crime types in 4 groups increased discrimination and clarity in clustering, in other words success. When the profiles of the prisoner groups that were formed were examined comparatively with the help of descriptive statistics, similar results were obtained.
Keywords
References
- Akdeniz, S. ve Üzümcü, A., 2013. Suç ve Sosyoekonomik Değişkenler Arasındaki Bağımlılık İlişkisi: Kars Cezaevi Üzerine Bir İnceleme, Kafkas Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi, 4 (6), s.117-138.
- Aldenderfer, M. S. ve Blashfield, R. K., 1984. Cluster Analysis. California: Sage.
- Alpar, C. R., 2000. Uygulamalı Çok Değişkenli İstatistiksel Yöntemler. Ankara: Detay.
- Altaş, D., Gençler, A. ve Giray, S., 2014. Yasadışı Göç Olgusunun Çok Değişkenli Analiz Teknikleri İle İncelenmesi. Econworld (International Conference in Economics), Prag, Çek Cumhuriyeti.
- Ata, A. Y., 2011. Ücretler, İşsizlik ve Suç Arasındaki İlişki: Yatay-Kesit Analizi. Çalışma ve Toplum Dergisi 4, s. 113-134.
- Ayhan, İ. ve Çubukçu, M., 2007. Suç ve Kent İlişkisine Ampirik Bakış: Literatür Taraması. Süleyman Demirel Üniversitesi Sosyal Bilimler Enstitüsü Dergisi 3 (5), s. 30-55.
- Aytaç, M., Aytaç, S. ve Bayram, N., 2007. Suç Türlerini Etkileyen Faktörlerin İstatistiksel Analizi. 8. Türkiye Ekonometri ve İstatistik Kongresi, İnönü Üniversitesi Malatya.
- Altuner, D., Engin, N., Gürer, C., Akyay, İ. ve Akgül, A., 2009. Madde Kullanımı ve Suç İlişkisi: Kesitsel Bir Araştırma. Tıp Araştırmaları Dergisi, 7 (2), s.87-94.
Details
Primary Language
English
Subjects
-
Journal Section
Research Article
Authors
Selay Giray
MARMARA UNIV
Türkiye
Publication Date
December 2, 2016
Submission Date
January 16, 2017
Acceptance Date
-
Published in Issue
Year 2016 Number: 25