المقارنة بين طريقة الاحتمال المقترح للإجابة الصحيحة و الطريقة التجريبية لتقدير درجات أسئلة الاختيار من متعدد وفق درجة الارتباط بإحصائيات المفردة و الأفراد المقدرة بالنموذج اللوغاريتمي الثلاثي المعلم

Abstract EN

A comparison between of method of probability assigned to the correct answer and the experimental method for estimating the scores of the multiple choice questions according to the degree of correlation with the statistics of individual and Three-Parameter Logistic Model (3PL) estimates items.

The study aimed to compare between two methods of estimating the multiple-choice questions (method of probability assigned to the correct answer, and experimental method) in light of the correlation of the statistics of individuals and items that are derived from these methods with the corresponding statistics that are estimated by Three-Parameter Logistic Model (3PL).

In order to achieve the aims of this study and to answer its questions, the researcher prepared four-multiple choice test at the units (linear equations and linear Inequalities) of Math subject regarding third intermediate level.

After getting sure of the study tool propriety to the purpose of the study, it has been applied on a sample of (903 students) distributed on 15 schools and has been chosen by cluster random sample from the population of third intermediate level in Arar city at the north of Kingdom of Saudi Arabia.

After collecting data and analyzing them through using SPSS program and Bilog-MG, the statistical analyzing of data revealed the following results: 1-The statistical analysis of the students’ ability revealed that: A.

The means of the students' scores according to the two methods (probability assigned to the correct answer and the experimental method) were convergent in general.

but, the experimental method was the highest in estimating the students' scores then the method of probability assigned to the correct answer.

B.

The experimental method was most related method to the abilities scores that are estimated by the three-parameter logarithmic model with a correlation coefficient (0.76) then the method of probability assigned to the correct answer with a correlation coefficient (0.74) C.

The method that was closest to the Three-Parameter Logistic Model in the means of absolute differences between the T-scores, was the experimental method with a difference of (5.7) then the method of probability assigned to the correct answer with a difference of (5.9) 2- The results of comparisons of item statistics (difficulty and discrimination), showed:a- The method that was most related of the Three-Parameter Logistic Model in estimating the difficulty of items is the experimental method of correlation coefficient (0.83), then the method of probability assigned to the correct answer with a correlation coefficient (0.81).

The two values reveal a high positive correlation.

b- The method that was most related with the Three-Parameter Logistic Model in estimating the discrimination of items is the experimental method then the method of probability assigned to the correct answer .and The degree of their correlation was very weak and not statistically significant at (0.50).