Abstract

The study aimed to construct a measure of students' attitudes towards English language by using the Thurston method with equal appearing intervals. Basically, the study targeted two samples: the first one was the exploratory and the second was the main sample; both samples targeted male and female students. The constructed measure was built with equal appearing intervals principles: first, there were (52) items which were presented to (14) judges to check the propriety of the items and its linguistic construction, second, after reviewing judges' notes, the measure was written in the primary shape which included (30) items. Next, the items were presented to (40) judges with the aim of grading the items among (11) positions. After collecting judges' data, the researcher used the median to find out the scale value; beside that, the Interquartile Range was used to explore the range of variance among the judges' judgments. Based on that, the final measure was revealed with (20) items which covered all grades on the measure. To verify the validity of the constructed measure, the face validity and Exploratory Factor Analysis where used. The results of Exploratory Factor Analysis showed that there is one dominant dimension, and it interprets the variation in students' attitudes toward English language. To estimate the measure reliability, three methods where used: the Spearman Brown split -half correlation coefficient where the estimated value was 0.81, the Test- retest method where the estimated value was 0.80, and Spearman correlation for reliability of estimates where the estimated value was 0.83 which revealed that the measure has good reliability. The discriminating ability of the items by using the Point Biserial Correlation Coefficient between the grades on the items and the overall grade on the measure referred to the measure has a good ability to discriminate between students who have different levels of attitudes toward English language the values were ranged between 0.59 to 0.31.