دراسة تقويمية لاستخدام طرائق المقارنات المتعددة البعدية في البحوث التربوية والنفسية
The main goal of this study is to try to cast light on the different methods of Post-Hoc multiple comparisons methods that are used in the educational and psychological research in an attempt to reduce the incorrect probabilities in the results of the educational studies, as well as consolidate the accuracy of the differences between the group averages and thereby reach higher validities and strengths in the results of the studies that use the descriptive methodologies. The study sample was made of 70 studies published in academic Arabic journals during the period 2005-2014. The instrument of the study was designed such that it includes the information that may fulfill the objectives of the study. The study came up with several findings, most important among which is the use of five types of Post-hoc multiple comparisons methods only in the study sample including Scheffé, Tukey, LSD, Bonferroni and Newman-Keuls. Of these studies,18 (68.6%) used Scheffé's method, while 22 studies were distributed among the other four studies at a percentage of 31.64%. The studies with educational character were 48 (68.57%) in number, while 22 (31.43%) were of psychological character. The descriptive methodology was the one most frequently used on the study sample due to the nature of the studies as 68 (97.14%) of them used the descriptive methodology, while the empirical methodology was used only in 2.86% of the studies. All of the studies that used the multiple comparative post hoc methods did not embody any indication to verify the hypotheses that should be present in the data included in the study sample. Most of the studies, which were 68 (97.14%) in number, were of unequal sizes, against only 2 (2.86%) studies which were of equal sizes. Scheffé's method was used by 47 (67.14%) studies whose sample sizes were equal, while 23 studies were distributed among the rest of the methods. All of the uses of the multiple comparison methods were not compatible due to the big differences in the sample sizes and the failure to verify the hypotheses. The study also presented some recommendations. (Published abstract)