AI Summary & Key Takeaways
AI InsightThe study results show that the factor numbers obtained using three different strategies during the item removal process of EFA were the same; however, the number of items on the scale, the explained variance ratio, and the total scale, and reliability values differed.
📌 3 Key Takeaways
Abstract
The aim of this study is to examine how the practice of different item removal strategies during exploratory factor analysis (EFA) phase of scale development change the number of factors, factor loadings, explained variance ratio, and reliability values (α and ω) explained. In the study, data obtained from 379 university students were used for the development of a 46-item scale. As the first item removal strategy, crossloading items on two factors and where the difference between factor loadings was less than .10 were identified. Then, items were removed one by one, starting with the item with the least difference between the loadings on the factors. As the second strategy, the items that loaded on two factors and where the difference between factor loadings was less than .10 were found, and these items were removed from the scale as a whole. As the third strategy, the items that gave high loading on more than two factors and where the difference between these factors was less than .10 were identified. The item removal process was started with these items. The study results show that the factor numbers obtained using three different strategies during the item removal process of EFA were the same; however, the number of items on the scale, the explained variance ratio, and the total scale, and reliability values differed. Furthermore, the items in the factors were not all the same. The study results underscore the importance of theoretical competence in the scale development process.
Özet
The aim of this study is to examine how the practice of different item removal strategies during exploratory factor analysis (EFA) phase of scale development change the number of factors, factor loadings, explained variance ratio, and reliability values (α and ω) explained. In the study, data obtained from 379 university students were used for the development of a 46-item scale. As the first item removal strategy, crossloading items on two factors and where the difference between factor loadings was less than .10 were identified. Then, items were removed one by one, starting with the item with the least difference between the loadings on the factors. As the second strategy, the items that loaded on two factors and where the difference between factor loadings was less than .10 were found, and these items were removed from the scale as a whole. As the third strategy, the items that gave high loading on more than two factors and where the difference between these factors was less than .10 were identified. The item removal process was started with these items. The study results show that the factor numbers obtained using three different strategies during the item removal process of EFA were the same; however, the number of items on the scale, the explained variance ratio, and the total scale, and reliability values differed. Furthermore, the items in the factors were not all the same. The study results underscore the importance of theoretical competence in the scale development process.