This study is to examine the meta-analysis results acquired from Cronbach alpha reliability coefficient being used in Hacettepe University Journal of Education. Within this context, 1222 items taking place in 43 issues which were published in Hacettepe University Journal of Education between 1986 and 2012 were examined and 354 measurement tools in total were discussed according to their inclusion criteria. In this study, r index was used in calculating influence quantity in correlational data for combining data while random influence model of Fisher z method was used in correlational data for combining influence quantities. It was benefited from SPSS 20 and MetaWin 2.0 packet programs for the analysis of data. It was found in this study that effect size mean is quite strong according to various moderator varieties of Cronbach alpha reliability coefficient. As a result of this study, it was determined that effect size mean of scale reliability prepared for measuring affective structure was greater than the scales prepared for measuring cognitive structures; effect size mean of reliability coefficient in adaptation studies was greater than the studies of applying and developing prepared scale. In addition to these results, it was observed that as the education levels of individuals within the sample and the number of items in scale increase, effect size mean of alpha reliability coefficient increases, too. On the other hand, it was determined that sample size and answer category number of option items did not have a direct influence on alpha coefficient. Afterwards, independent samples t test, one way ANOVA and Kruskal Wallis Test were conducted in order to find out whether the mean of Cronbach alpha acquired in line with the determined moderator variables varied according to groups. According to these analysis results, it was determined that structure of data collection tool, content of the sample, item and number of option had significant differences on Cronbach alpha coefficient. In addition to the variables within this study, different variables which are thought to affect reliability can be handled and the effects of these variables on reliability coefficient can be examined. Keywords: Cronbach alpha, meta-analysis, reliability ÖZ: Bu çalışmada, Hacettepe Üniversitesi Eğitim Fakültesi Dergisi'nde yayınlanan çalışmalarda yer alan Cronbach alfa güvenirlik katsayılarından elde edilen meta analiz sonuçlarının incelenmesi amaçlanmıştır. Bu amaçla 1986-2012 yılları arasında Hacettepe Üniversitesi Eğitim Fakültesi Dergisi'nde yayınlanan 43 sayıda yer alan toplam 1222 makale incelenmiş ve dahil edilme ölçütlerine göre toplam 354 ölçme aracı ele alınmıştır. Bu araştırmada verilerin birleştirilmesi için korelasyonel verilerde etki büyüklüğü hesaplamalarındaki r indeksi; korelasyonel verilerde etki büyüklüklerinin birleştirilmesi için de Fisher z yönteminin tesadüfi etki modeli kullanılmıştır. Verilerin analizi için SPSS 20 ve MetaWin 2.0 paket programından yararlanılmıştır. Çalışmada Cronbach alfa güvenirlik ...

This study aimed to examine the impact of differential item functioning in anchor items on the group invariance in test equating for different sample sizes. Within this scope, the factors chosen to investigate the group invariance in test equating were sample size, frequency of sample size of subgroups, differential form of differential item functioning (DIF), frequency of items in the anchor test with differential item functioning, directionality of differential item functioning and mean differences in subpopulation ability levels.

The aim of this study is to investigate the group invariance condition according to Tucker and Levine observed score equating among linear equating methods. In the study, the 4 th and 6 th booklets of the PISA 2012 Mathematics subtest were used. Booklets were equated according to group and gender sub-variables, and then group invariance of each condition and WMSE values were calculated. Within this scope, REMSD and RMSD (x) group invariance indexes were employed. The results of the study indicated that, when WMSE values, obtained according to equating methods, were compared, Tucker observed score equating method with regard to whole-group and gender sub-groups produced the lowest error. When RMSD and REMSD values obtained according to gender sub-groups were examined by linear equating methods, it was found that group invariance value is smaller than criterion value for Tucker equating method, while it was greater than criterion value for Levine equating method. Eventually, group invariance condition was met for Tucker observed score equating, but not for Levine observed score equating.Keywords: Equating, group invariance, Tucker linear equating, Levine linear equating Öz Bu çalışmanın amacı doğrusal eşitleme yöntemlerinden Tucker ve Levine gözlenen puan eşitleme yöntemlerine göre eşitlemenin grup değişmezliği koşulunun incelenmesidir. Bu çalışmada PISA 2012 matematik alt testine ait 4. ve 6. kitapçıklardan elde edilen test puanları kullanılmıştır. Kitapçıklardan elde edilen puanlar tüm grup ve cinsiyet alt değişkenine göre eşitlenmiştir. Her bir koşula ait grup değişmezliği ve WMSE değerleri hesaplanmıştır. Bu bağlamda grup değişmezliği indekslerinden REMSD ve RMSD (x) kullanılmıştır. Araştırma sonucunda eşitleme yöntemlerine göre elde edilen WMSE değerleri karşılaştırıldığında hem tüm grup hem cinsiyet alt grubuna göre en az hata veren yöntemin Tucker gözlenen puan eşitleme olduğu görülmüştür. Doğrusal eşitleme yöntemleriyle cinsiyet alt grubuna göre elde edilen RMSD ve REMSD değerleri incelendiğinde, Tucker eşitleme yöntemi için kriter değerden küçük iken, Levine eşitleme yöntemi için kriter değerden daha yüksek çıkmıştır. Böylece grup değişmezliği koşulunun Tucker eşitleme yönteminde sağlanırken, Levine eşitleme yönteminde sağlanmadığı görülmüştür.