Test eşitlemede Madde Tepki Kuramına dayalı yetenek parametresine yönelik ölçek dönüştürme yöntemlerinin karşılaştırılması

---

Thesis Type: Postgraduate

Institution Of The Thesis: Gazi University, Eğitim Bilimleri Enstitüsü, EĞİTİMDE ÖLÇME VE DEĞERLENDİRME ANABİLİM DALI, Turkey

Approval Date: 2015

Thesis Language: Turkish

Student: TUBA GÜNDÜZ

Supervisor: MEHTAP ÇAKAN

Open Archive Collection: AVESIS Open Access Collection

Abstract:

In this study, it was aimed that the mean-mean, the mean-standard deviation, Stocking-Lord and Haebara scale transformation methods which based on Item Response Theory in test equating were compared in terms of ability parameters (tetas) on real data. For dichotomous data under the 3PLM model, the sample of the study composed of Turkish students who participants of TIMSS 2011 fourth grade level in math chosen two forms answer (535 and 529 students). These forms have 26 items and both of forms contain 14 common item. Choosing from equating design, according to the structure of forms and methods of data collection "common item nonequivalent groups design" is used. This design of use in which to put tetas and item parameters from different forms on different scales on the same scale with the software to be used in the ST scale transformation coefficients and constants were found for all four methods. RMSD (Root mean square difference) criteria were used to decide which method is more appropriate. According to findings obtained from research on the transformation scale for the teta to do the most appropriate parameter (containing at least error method) to the method of ranking made the least favorable; (1) Mean-mean, (2) Stocking-Lord, (3) Haebara and (4) Mean-standard deviation were found. After determining the most appropriate method by which parameters are derived from the transformation of this method using the software PIE 2.0 points made IRT true score equating and interpreted. As a result of this, it is concluded that Form Y is more difficult than Form X.