THE COMPARISON OF THE DIFFERENT STATISTICAL ANALYSIS RESULTS USED ON THE EVALUATION OF SCALES ON A REAL SAMPLING

Authors : Fatih Sezek, Uluhan Kurt

Pages : 85-99

Doi:10.33418/ataunikkefd.558316

View : 23 | Download : 11

Publication Date : 2019-06-30

Article Type : Research

Abstract :The purpose of this study is to examine the effects of using different statistical analysis methods on the results of questionnaire or scales evaluation in education studies. The scale-survey method was used for data collection. In accordance with this purpose, "Students’ Satisfaction of Physical Space” (SSPS) developed by researchers and measuring their school satisfaction level of students was used. The sample of the study is consisted of 720 students studying in the boarding school in the province of Erzurum. Firstly, the data obtained from the sample were scored and then, both parametric and non-parametric analyzes were performed on the total score and arithmetic mean scores of the test. Evaluating of tests results were statistically compared. Thus, we tried to determine which analysis method was more accurate in the evaluation of the scales.  Accordingly, we were be able to achieve more reliable results. As a result, it was determined that performing statistical analyzes by scoring questionnaires and scales brought different problems. Contrary to the general belief that parametric tests are more reliable than non-parametric tests, it is also shown that the comparison with the chi-square test is one of the most suitable methods in the analysis of categorical datum obtained from the questionnaires. On the other hand, instead of collecting the scores of the different class levels students\' answers to each question, it is more convenient to apply the chi-square test to the total frequencies of the answers. In addition, among students of different grades (K5-8), it was determined that there was a statistically significant difference among their satisfaction levels from the physical environment of the school. 
Keywords : Parametric and non-parametric tests, Chi-square, scales