HAHakan Yavuz Atar

Hakan Yavuz Atar

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Hakan Yavuz Atar has a total of 6 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Comparison of Models for Simultaneous Estimation of Overall Score and Subscores: Estimation Accuracy, Reliability, and Classification Accuracy

    by Ayşenur Erdemir, Hakan Atar

    This study compares the performance of the Multidimensional Item Response Theory (MIRT), Higher-Order IRT (HO-IRT), and Bifactor models for the simultaneous estimation of total and subscale scores in multidimensional tests. Using both simulated data and real data from an English proficiency exam, model performance was evaluated in terms of accuracy (RMSE), reliability, and classification accuracy. The simulation included 5,000 respondents, 120 items, and a four-dimensional structure, manipulating item format, test difficulty, and inter-dimensional correlation. Results indicated that MIRT consistently outperformed the other models, yielding the lowest RMSE and highest reliability and classification accuracy across conditions. HO-IRT also showed strong performance, while the Bifactor model underperformed, particularly in subscore estimation. Model performance was sensitive to test characteristics and dimensional relationships. Findings from the real data analysis supported the simulation results, underscoring the value of multidimensional modeling for diagnostic feedback and informed decision-making.

  • A Comparison of the Classification Performances of the DINO Model, Artificial Neural Networks and Non-Parametric Cognitive Diagnosis

    by Emine Yavuz, Hakan Yavuz Atar

    The purpose of this study was to compare the attribute (ACR) and pattern-level (PCR) classification rates of the Deterministic-Input, Noisy-Or Gate (DINO) model, Artificial Neural Networks (ANNs), and Non-parametric Cognitive Diagnosis (NPCD) from datasets of simulation in various conditions such as the number of attributes, sample size, the number of items and missing data rate. A further purpose was to examine the similarities between the classification rates of the DINO model ANNs, and NPCD on the PISA 2015 collaborative problem-solving (CPS) datasets in various numbers of attributes and sample sizes. For the purpose of the study, simulation datasets were generated on the basis of the complex Q matrix structures and the DINO model. The conditions for the sample size factor for the real datasets were determined by simple random selection among the participants in the PISA 2015 administration. As a result, it was found that there was a similarity between the DINO model and NPCD classification rates in both simulation and real datasets. In addition, regarding the increase in sample size in both simulation and real datasets, no consistency was found in the increase or decrease of the classification rates of ANNs and NPCD and the similarities of these rates.

  • Open Access

    Investigation of the effect of parameter estimation and classification accuracy in mixture IRT models under different conditions

    by Fatıma Münevver SAATÇİOĞLU, Hakan Yavuz Atar

    This study aims to examine the effects of mixture item response theory (IRT) models on item parameter estimation and classification accuracy under different conditions. The manipulated variables of the simulation study are set as mixture IRT models (Rasch, 2PL, 3PL); sample size (600, 1000); the number of items (10, 30); the number of latent classes (2, 3); missing data type (complete, missing at random (MAR) and missing not at random (MNAR)), and the percentage of missing data (10%, 20%). Data were generated for each of the three mixture IRT models using the code written in R program. MplusAutomation package, which provides the automation of R and Mplus program, was used to analyze the data. The mean RMSE values for item difficulty, item discrimination, and guessing parameter estimation were determined. The mean RMSE values as to the Mixture Rasch model were found to be lower than those of the Mixture 2PL and Mixture 3PL models. Percentages of classification accuracy were also computed. It was noted that the Mixture Rasch model with 30 items, 2 classes, 1000 sample size, and complete data conditions had the highest classification accuracy percentage. Additionally, a factorial ANOVA was used to evaluate each factor's main effects and interaction effects.


  • Examination of Wording Effect of the TIMSS 2015 Mathematical Self-Esteem Scale Through the Bifactor Models

    by Esra Oyar, Hakan Yavuz Atar

    The aim of this study is to examine whether or not the positive and negative items in the Mathematical Self-Confidence Scale employed in TIMSS 2015 lead to wording effect. While examining whether the expression effect is present or not, analyzes were conducted both on the general sample and on a separate sample for female and male students. To this end, data of 5724 students from Turkey who participated in TIMSS 2015 were used. Six different measurement models were created in the analysis of data and tested with Confirmatory Factor Analysis. The study revealed that positive items have a higher mean than the negative ones. In addition, it was concluded that the bifactor models fit the data better compared to the traditional DFA model, in which the model where negative items were taken as a separate factor are those that best fit the data. This situation is verified both in the general sample and the subgroups of females and males. In conclusion, it is recommended that the scale items should be created carefully and whether the positive and negative items result in separate factors should be examined.


  • An Examination of Turkish Students' PISA 2015 Collaborative Problem-Solving Competencies

    by Emine Yavuz, Hakan Yavuz Atar

    Technological advancements initially made it possible for individuals with different cultures from different parts of the world to consider working together, and eventually collaboration turned into a necessity. Based on this necessity, OECD measured collaborative problem-solving (CPS) competencies of 15-year-old students from the countries participating in PISA assessment in 2015. The objective of this study is to examine the data on Turkey from PISA 2015 CPS survey through Deterministic-Input, Noisy-Or-Gate (DINO) cognitive diagnostic model and to determine the students’ levels of CPS competencies as well as difficulty levels of these competencies. In this context, primarily, attributes of CPS competency domain were examined and the model with the most appropriate number of attributes, classification accuracy and consistency was determined for the analysis of PISA 2015 CPS data. The sample of this descriptive study consists of 435 students. As a result of the analysis, the Q-matrix of the model data fit was observed to have the best goodness-of-fit when it had three attributes whereas model fit and classification consistency decreased as the number of attributes increased. Furthermore, the easiest competency for Turkish students was determined to be ‘taking appropriate action to solve the problem’ while the most difficult was ‘establishing and maintaining shared understanding’. Finally, latent class distribution exposed that 56% of the students did not have any of the competencies, and 35% had all. Results of this examination indicate that all CPS competencies of Turkish students need improvement.


  • Open Access

    A Study on the Identification of Latent Classes Using Mixture Item Response Theory Models: TIMSS 2015 Case

    by Fatıma Münevver SAATÇİOĞLU, Hakan Yavuz Atar

    This study examined the existence of latent classes in TIMSS 2015 data from three countries, Singapure, Turkey and South Africa, were analyzed using Mixture Item Response Theory (MixIRT) models (Rasch, 1PL, 2PL and 3PL) on 18 multiple-choice items in the science subtest. Based on the findings, it was concluded that the data obtained from TIMSS 2015 8th grade science subtest have a heterogeneous structure consisting of two latent classes. When the item difficulty parameters in two classes were examined for Singapore, it was determined that the items were considerably easy for the students in Class 1 and the items were easy for the students in Class 2. When the item difficulty parameters in two classes were examined for Turkey, it was found that the items were easy for the students in Class 1 and the items were difficult for the students in Class 2. When the item difficulty parameters in two classes were examined for South Africa, it was ascertained that the items were a bit easy for the students in Class 1 and the items were considerably difficult for the students in Class 2. The findings were discussed in the context of the assumption of parameter invariance and test validity.


  • Simultaneous Estimation of Overall Score and Subscores Using MIRT , HO-IRT and Bi-factor Model on TIMSS Data

    by Ayşenur ERDEMİR, Hakan Yavuz Atar

    In educational testing, there is an increasing interest in the simultaneous estimation of the overall scores and subscores. This study aims to compare the reliability and precision of the simultaneous estimation of overall scores and sub-scores using MIRT, HO-IRT and Bi-factor models. TIMSS 2015 mathematics scores have been used as a data set in this study. The TIMSS 2015 mathematics test consists of 35 items, four of which are polytomously scored (0-1-2), and the rest of the items are dichotomously scored (0-1). The four content domains include number (14 items), algebra (9 items), geometry (6 items), and data and change (6 items). Ability parameters were estimated using the BMIRT software. The results showed that the MIRT and HO-IRT methods performed similarly in terms of precision and reliability for subscore estimates. The MIRT maximum information method had the smallest standard error of measurement for the overall score estimates. All three methods performed similarly in terms of the overall score reliability. The findings suggest that among the three methods compared, HO-IRT appears to be a better choice in the simultaneous estimation of the overall score and subscores for the data from TIMSS 2015. Recommendations for the testing practices and future research are provided.


  • Open Access

    Effects of Students and School Variables on SBS Achievements and Growth in Mathematic

    by Emine Yavuz, Şeref Tan, Hakan Yavuz Atar

    The purpose of this study is to investigate the effects of several student and school level variables on students' mathematical achievement and mathematical achievement growth in middle school. The research population consisted of students in Ankara who started middle school in 2008 and graduated in 2011. Using a non-random typical case sampling method, 3715 students were sampled from 40 middle schools. The data used in this study were obtained from the Ministry of National Education with express written permission. Student Placement raw mathematics subtest scores from 2009 to 2011 were used as the dependent variable in the analysis. "Gender, mathematics grade point average in 6h grade, and school attendance at sixth grade" comprised the student level variables whereas "school type and school size" variables comprised the school-level variables. A three-level hierarchical linear growth model was used to investigate the effects of these variables. Students' raw mathematics subtest scores in Students Placement Test were included in the model after these scores were equated. The results of the model analysis indicated that there was no growth in students' mathematical achievement. It was also observed that school attendance, sixth-grade mathematics grade point average and school type had a statistically significant effect on students' sixth-grade mathematical achievement.