نوع المستند : المقالة الأصلية
المؤلف
أستاذ علم النفس التربوى المساعد -کلية التربية – جامعة المنيا
المستخلص
الكلمات الرئيسية
الموضوعات الرئيسية
عنوان المقالة [English]
المؤلف [English]
The purpose of the current research was to assess the response styles and their impacts on Item validity, Unidimensional, Confirmatory Factor and Reliability. Four Forms of the Rosenberg Self-Esteem Scale were prepared. The paper version included an original form of the scale and a reverse form . The research was carried out on a sample of (314) undergraduate students in their fourth year of the Faculty of Education at Minia University for the academic year (2017/2018) . The electronic version included an original form of the scale and a reverse form and used a sample of( 260) undergraduate students in the fourth year of the Faculty of Education at Minia University for the academic year ( (2018/2019 ) .The results showed that response styles represented (26% to 27%) of the sample. It was also noticed that the Item Validity improved after the exclusion of the Response Styles, the improvement of the Unidimensionality and the results of Confirmatory Factor Analysis. Indices of Goodness of Fit improved and eigenvalues increased . In the Reliability estimates after exclusion of Response Styles, Reliability improved in the test-retest method, Split-Half reliability, the Guttmann reliability, Cronbach's alpha reliability and the composite reliability. Thus, the research indicated that the analysis of questionnaires without detection and control of response styles negatively affected the psychometric characteristics of the measuring instrument. Moreover, increasing the size of the sample alone was not sufficient to improve reliability and validity. Although the exclusion of the response styles led to the reduction of sample size, it could improve reliability coefficients and validity. Therefore, it is recommended to investigate the importance of detection and control of response styles to avoid distortion of results and reduce measurement errors.