Sakine Gocer Sahin, PhD

Dr. Sahin is a quantitative researcher with expertise in biostatistics, psychometrics, and large-scale data analysis. She has over a decade of experience managing and analyzing health-related datasets, including hospital clinical data and population health surveys. At RAC, she has worked on projects involving complex statistical modeling and validation of metrics.

She has:

• Expertise in item calibration; calculating and evaluating item-level statistics to minimize measurement error; preparing and analyzing data for formal technical research reports, using Item Response Theory (IRT) and Classical Test Theory (CTT) modeling procedures; maintaining quality control in data collection and scoring processes and procedures; and conducting data visualizations for technical research reports.
• Worked with a variety of data sets obtained from a variety of different data collection methods under a variety of research and test designs. Dr. Gocer Sahin’s experience includes practical applications of equating, instrument development, and instrument validation.
• Previously worked on projects designed to determine the validity of foreign language exams, and to equate different instruments. One of her most interesting projects was to develop a tablet-based assessment tool to determine dyscalculia tendencies of students.
• Worked as a technical expert advisor in the field of psychometrics for the Vocational Qualifications Authority in Ankara, Turkey.
• Served in a variety of consulting, training, and organizational roles. For example, she organized a multidisciplinary workshop to teach technical psychometric skills to attendees that ranged from hairdressers to teachers. In this workshop, participants learned how to evaluate the quality and fairness of their evaluations’ test items and settings.
• Been involved in a project sponsored by the European Union to write a technical report on performance assessment.

She completed her Ph.D. in Educational Measurement and Evaluation at Hacettepe University in Turkey. Her dissertation, entitled “Examining Parameter Estimation When Treating Semi-Mixed Multidimensional Constructs as Unidimensional,” argues that fitting semi-mixed structured multidimensional data as unidimensional model results in error. Upon the completion of her doctoral degree she did her post-doc training at the University of Wisconsin-Madison in the Quantitative Methods Department. Her post-doc project focused on the development of a new approach to the detection of cheating, in large scale high stakes tests.