Türkiye Ölçme Araçları Dizini’nde Yer Alan Açımlayıcı Faktör Analizi Çalışmalarının Paralel Analiz Sonuçları İle Karşılaştırılması

Abstract

The purpose of this study is to investigate the number of extracted factors in exploratory factor analysis (EFA) studies which takes part in Turkish Measurement Tools Index (TMTI / TOAD) according to parallel analysis results. By the date the study had begun, 4440 studies were indexed in TMTI, 500 studies were selected by systematic sampling procedure among them and the number of full texts investigated is 451. For these studies, values were coded by using a form which was developed by researchers such as sample size, number of items, rotation method. For 300 studies in which EFA was implied, some of them do not include information about sample size, number of items, minimum factor loading, number of factors, extracted variance. The most frequent extraction method is principal component analysis, and criterion for number of factors is eigenvalue greater than unity. For multifactor scales typical rotation, method is varimax. Amongst 67 EFA studies that could be compared to parallel analysis results, 52 overextraction and seven underextraction situation was detected. For seven studies, number of factors were accurate. Findings are discussed within existing literature and a checklist is recommended for the purpose of minimizing deficiencies in reporting EFA results.

Üst bilişsel Dinleme Farkındalığı Ölçeğinin Türkçe Uyarlama Çalışması

“…Another indicator for the one factor structure of the scale is the fact that the eigenvalue of the first factor is three times as more than the second factor and that the first factor explains more than 30% of the variance (Çokluk et al, 2010). Therefore, one-factor structure of the scale was confirmed with a parallel analysis (Akbaş, Karabay, Yıldırım-Seheryeli, Ayaz & Demir, 2019). The factor loadings of the items in the scale are presented in Table 7.…”

Evlilik Çatışma Alanları İndeksi ve Çatışma Çözme Stilleri Envanteri’nin Geliştirilmesi: Geçerlik ve Güvenirlik Çalışmaları

“…The construct validity of CAI was examined using principal component analysis (PCA), and the construct validity of CSI was examined using exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). In deciding the number of factors, while the extraction method was the main axis, in addition to eigenvalues that were greater than one, the factors' contributions to the results of explained variance, scree plot, parallel analysis (Watkins 2000), and the conceptual framework as a whole were considered (Akbaş et al 2019). For an item to be included in the scale, care was taken to ensure that the factor loading was at least .32 (Tabachnick and Fidell 2013) and that the differences between loadings under different factors were above .10 (Büyüköztürk 2021).…”

Validity and Reliability Evidence of Professional Obsolescence Scale According to Different Test Theories

“…In PA, the minimum factor loading was .50 (Todman & Dugard, 2007), and for the detection of overlapping items, the difference between the factor loading of the same item on two factors was taken as .10 (Büyüköztürk, 2019). In addition, the "Checklist for Reporting EFA" proposed by Akbaş et al (2019) was used for reporting. Although the responses to the items are graded between 1 and 5, the qualifiers for the grades differ.…”