Seçil Ömür Sünbül

Dr. Seçil Ömür Sünbül


Research Focus & Metrics

Academic Career Length

10Years (2017-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
30 June 2024 in Eğitimde ve Psikolojide Ölçme ve Değerlendirme Dergisi
Verified
Journal Article
Open access
PDF file
SASemih Aşiret
Semih Aşiret
SSSeçil Ömür Sünbül
Seçil Ömür Sünbül
This study aimed to examine the performances of item selection algorithms in terms of measurement accuracy and computational time, using factors such as test length, number of attributes, and item quality in fixed-length CD-CAT and average test lengths and computational time, using factors such as number of attributes and item quality in variable-length CD-CAT. In the research, two different simulation studies were conducted for the fixed and variable-length tests. Item responses were generated according to the DINA model. Two item banks, which consisted of 480 items for 5 and 6 attributes, were generated, and the item banks were used for both the fixed and variable-length tests. Q-matrix was generated item by item and attribute by attribute. In the study, 3000 examinees were generated in such a way that each examinee had a 50% chance of achieving each attribute. The cognitive patterns of the examinees were estimated by using MAP. In the variable-length CD-CAT, the first-highest posterior probability threshold is 0.80, and the second-highest posterior probability threshold is 0.10. The CD-CAT administration and other analyses were conducted using R 3.6.1.At the end of the study in which the fixed-length CD-CAT was used, it was concluded that an increase in the number of attributes resulted in a decrease in the pattern recovery rates of item selection algorithms. Conversely, these rates improved with higher item quality and longer test lengths. The highest values in terms of pattern recovery rate were obtained from JSD and MPWKL algorithms. In the variable-length CD-CAT, it was concluded that the average test length increased with the number of attributes and decreased with higher item quality. Across all conditions, the JSD algorithm yielded the shortest average test length. Additionally, It has been determined that GDI algorithm had the shortest computation time in all scenarios, whereas the MPWKL algorithm exhibited the longest computation time.

1 January 2024 in Uluslararası İnovatif Eğitim Araştırmacısı
Verified
Journal Article
Closed access
SGSongül Güneş
Songül Güneş
SSSeçil Ömür Sünbül
Seçil Ömür Sünbül
Bu çalışmada, bireylerin problem çözme becerilerini yoklayan açık uçlu matematik sorularına vermiş oldukları cevapların bütünsel dereceli puanlama anahtarıyla puanlanması sonucu elde edilen puanların, madde ve puanlayıcı sayısının değişmesi ve farklı desenlerin kullanılması durumlarında G ve Phi katsayıları hesaplanarak incelenmiştir. Bireylere uygulanacak olan problem çözme becerilerinin yoklanmasına ilişkin açık uçlu sorular, PISA Matematik okuryazarlığı ve problem çözme alanlarında kullanılmış sorular içerisinden güncel müfredata uygun olduğu düşünülen sorular arasından seçilmiştir. Bireylerin problem çözme becerilerini yoklamak için hazırlanan açık uçlu soruların değerlendirilmesinde, dereceli puanlama anahtarı geliştirme süreci takip edilerek bütünsel dereceli puanlama anahtarı geliştirilmiştir. Çalışma grubu 57 öğrenciden, puanlayıcılar ise, 25 öğretmeninden oluşmaktadır. Çalışmada Genellenebilirlik Kuramı’nda birey (b), madde (m), puanlayıcı (p) değişkenlik kaynakları kullanılarak, bxmxp tümüyle çaprazlazlanmış desende ve bx(p:m) yuvalanmış desende, madde ve puanlayıcı sayıları değişimlenerek G ve Phi katsayıları hesaplanmıştır. Çalışmanın sonucunda, hem bxmxp hem de bx(p:m) desen için, madde sayılarının ve puanlayıcı sayılarının artmasıyla G ve Phi katsayılarında artış olduğu görülmüştür.

15 June 2026 in Behavioral Sciences
Verified
Journal Article
Open access
PDF file
SSSeçil Ömür Sünbül
Seçil Ömür Sünbül
MSMüzeyyen SOYER
Müzeyyen SOYER
This study aimed to examine variables associated with manipulation tendency levels in adults and to describe current patterns using a decision tree method as a classification-based analytical approach. The research sample consisted of 543 adults (358 women, 65.93%; 185 men, 34.07%) residing in Turkey, aged 18 to 45 years (M = 25.79, SD = 4.23). Data were collected using a researcher-developed personal information form, the Manipulation Scale in Human Relations, the Rosenberg Self-Esteem Scale, and the Relationship Scales Questionnaire. The total composite score of the Manipulation Scale was used as the outcome variable and is referred to throughout as “manipulation tendency.” Manipulation tendency scores were dichotomized into low versus high groups using a median split to facilitate C&RT-based classification. Classification and Regression Tree was used to examine the hierarchical structure of variables related to manipulation tendency levels and to identify classification patterns among study variables. The decision tree approach was used not merely as an alternative statistical technique, but as an interpretable classification framework capable of identifying hierarchical and conditional pathways associated with manipulation tendency. Data were stratified-randomly split into training and test sets (70/30), and tree complexity was tuned via cross-validation using cost-complexity pruning. Model performance indicated acceptable classification accuracy within this sample, with a test-set accuracy of 0.81 (sensitivity = 0.74, specificity = 0.88, precision = 0.86, F1 = 0.79) and training accuracy of 0.86. The findings indicated several influential variables in classifying manipulation tendency levels, ranked by importance: preoccupied attachment style, self-esteem, age, dismissive attachment style, gender, secure attachment style, and fearful attachment style. Preoccupied attachment style was identified as the most salient variable in differentiating between high and low manipulation tendency groups. The decision tree structure showed that younger adults with higher preoccupied attachment scores were more frequently classified into the high manipulation tendency group. Self-esteem emerged as the second most influential variable, with lower self-esteem levels being more commonly observed among individuals classified in the high manipulation tendency group. Age also played a notable role in classification, with higher manipulation tendency classifications occurring more frequently among younger individuals. Dismissive attachment style contributed to the differentiation of manipulation tendency levels, particularly within specific attachment and age profiles. Gender-related patterns indicated that men were more frequently classified into higher manipulation tendency groups, especially among individuals with low self-esteem. Overall, the findings highlight the multifactorial and hierarchical nature of manipulation tendency classifications. They contribute to the literature by showing how attachment-related characteristics, developmental factors, and psychological variables jointly differentiate manipulation tendency profiles. These findings highlight the value of decision tree modelling for translating conventional psychological predictors into interpretable classification profiles of manipulation tendency.

31 March 2017 in Eğitimde ve Psikolojide Ölçme ve Değerlendirme Dergisi
Verified
Journal Article
Closed access
SSSeçil Ömür Sünbül
Seçil Ömür Sünbül
SASemih Aşiret
Semih Aşiret

23 August 2021 in Pamukkale University Journal of Education
Verified
Journal Article
Closed access
SASemih Aşi̇rEt
Semih Aşi̇rEt
SSSeçil Ömür Sünbül
Seçil Ömür Sünbül
Bu çalışmanın amacı, Bilişsel Tanıya Dayalı Bilgisayar Ortamında Bireye Uyarlanmış Testlerde (BT-BOBUT), DINA ve DINO model için farklı nitelik sayısında, madde kalitesinde ve test uzunluklarında madde seçim algoritmalarının performanslarını ölçme doğruluğuna göre incelemektir. Çalışma kapsamında, nitelik sayısı 5 ve 8 olarak değişimlenmiş ve her madde en az bir nitelik ve en fazla 4 nitelik ölçecek şekilde sınırlandırılmıştır. Veri üretiminde, g ve s parametreleri yüksek madde kalite düzeyi için U(0,05-0,25) ve düşük madde kalite düzeyi için U(0.10-0.30) tekbiçimli dağılımdan çekilmiştir. Her bireyin her niteliğe sahip olma şansı %50 olacak şekilde 3000 bireye ait bilişsel örüntüler üretilmiştir. Sonlandırma kuralı olarak 8, 16 ve 24 sabit test uzunlukları kullanılmıştır. Çalışmada kullanılan madde seçim algoritmaları GDI, JSD, MI, PWCDI ve PWKL’dir. Madde seçim algoritmalarının performansları, nitelik ve örüntü koruma oranlarına göre değerlendirilmiştir. Çalışmada veri üretimi ve analizleri R 3.6.3 yazılımı kullanılarak gerçekleştirilmiştir. Çalışma sonucunda, madde kalitesi ve test uzunluğu arttıkça tüm algoritmaların ölçme doğruluk değerlerinin arttığı, nitelik sayısı arttıkça ölçme doğruluğunun azaldığı tespit edilmiştir. JSD algoritmasının ölçme doğruluğu tüm koşullarda en yüksek iken, PWKL algoritmasının en düşük olduğu bulunmuştur. DINA ve DINO modellerde PWKL algoritması dışındaki algoritmaların performansı yaklaşık aynı iken, DINO modelde PWKL algoritmasının ölçme doğruluğunun DINA modelden daha düşük olduğu bulgusu elde edilmiştir. The aim of this study is to examine the performance of item selection algorithms according to the accuracy of measuring through different number of attributes, item quality and test lengths for DINA and DINO models in cognitive diagnostic computerized adaptive testing (CD-CAT). Within the scope of the study, the number of attributes was manipulated as 5 and 8, and each item was limited to measure at least one attribute and at most 4 attributes. In data generation, the g and s parameters were drawn from the uniform distribution of U(0.05-0.25) for high item quality level and U (0.10-0.30) for low item quality level. Cognitive patterns of 3000 examinees were generated so that each examinee had a 50% chance of having each attribute. Fixed test lengths of 8, 16 and 24 were used as termination rules. GDI, JSD, MI, PWCDI and PWKL were used as item selection algorithms in the study. The performances of item selection algorithms were evaluated according to their attribute and pattern recovery rates. Data generation and analysis in the study were carried out using R 3.6.3 software. As a result of the study, it was determined that the measurement accuracy values of all algorithms increased as the item quality and test length increased, and the measurement accuracy decreased as the number of attributes increased. It was found that the measurement accuracy of the JSD algorithm was the highest in all conditions, while the PWKL algorithm was the lowest. While the performance of the algorithms except for PWKL algorithm in DINA and DINO models was approximately the same, it was found that the measurement accuracy of the PWKL algorithm in the DINO model was lower than that of the DINA model.

7 April 2026 in Preprints.org
Verified
Preprint
Open access
PDF file
SSSeçil Ömür Sünbül
Seçil Ömür Sünbül
MSMüzeyyen SOYER
Müzeyyen SOYER
This study aimed to examine variables associated with emotional manipulation levels in adults and to describe current patterns using a decision tree method as a classification-based analytical approach. The research sample consisted of 543 adults (358 women, 65.93%; 185 men, 34.07%) residing in Turkey, aged 18 to 45 years (M = 25.79, SD = 6.24). Data were collected using a researcher-developed personal information form, the Manipulation Scale in Human Relations, the Rosenberg Self-Esteem Scale, and the Relationship Scales Questionnaire. Emotional manipulation scores were dichotomized into low versus high groups using a median split to facilitate CART-based classification. Classification and Regression Tree was used to examine the hierarchical structure of variables related to emotional manipulation levels and to identify classification patterns among study variables. Data were stratified-randomly split into training and test sets (70/30), and tree complexity was tuned via cross-validation using cost-complexity pruning. Model performance indicated good classification accuracy, with a test-set accuracy of 0.81 (sensitivity = 0.74, specificity = 0.88, precision = 0.86, F1 = 0.79) and training accuracy of 0.86. The findings indicated several influential variables in classifying emotional manipulation levels, ranked by importance: preoccupied attachment style, self-esteem, age, dismissive attachment style, gender, secure attachment style, and fearful attachment style. Preoccupied attachment style was identified as the most salient variable in differentiating between high and low emotional manipulation groups. The decision tree structure showed that younger adults with higher preoccupied attachment scores were more frequently classified into the high emotional manipulation group. Self-esteem emerged as the second most influential variable, with lower self-esteem levels being more commonly observed among individuals classified in the high emotional manipulation group. Age also played a notable role in classification, with higher emotional manipulation classifications occurring more frequently among younger individuals. Dismissive attachment style contributed to the differentiation of emotional manipulation levels, particularly within specific attachment and age profiles. Gender-related patterns indicated that men were more frequently classified into higher emotional manipulation groups, especially among individuals with low self-esteem. Overall, the findings highlight the multifactorial and hierarchical nature of emotional manipulation classifications. They contribute to the literature by showing how attachment-related characteristics, developmental factors, and psychological variables jointly differentiate emotional manipulation profiles.