Design of a mathematics achievement test and examination of its applicability for high school students in subject promotion and credit system

Ill SUMMARY This work has aimed at introducing a developed and greatly standardized Mathemathics Achievement Test which should help students to define the area to which they will direct themselves. For this purpose two tests have been prepared each containing 75 multiple-choice with 5 choices. The sample of this research was derived from 20770 students studying in Izmir city centre and its attached towns. The development of the test has been achieved in a two stage application. The first application has aimed at item analysis. For this purpose the first test was applied to 449 students, the second test to 453 students. Nevertheles, the item analysis was implemented by the evaluation of 370 test results selected at random. As a consequence of the item analysis, a single test composed of 75 questions with appropriate discrimination power and difficulty has been evolved. There are 25 items with discrimination indices greater than 0.40, 33 items with discrimination indices between 0.30-0.39 and 17 items with discrimination indices between 0.20-0.29, in the whole test. Thus, the 77% of the test is composed of very good items, 23% of the items of medium discrimination power. In the test, there are 14 questions at knowledge level, 16 questions at comprehencion level, 23 questions at application level, 7 questions ât synthesis level and 4 questions at the evaluation level. The content validity of the test has been derived by evaluating judges' (specialist teacher) opinions. Criterion based validity has also been examined to determine validity. School achievement reports have been taken as a second criterion. TheIV correlation between the two has also been considered and the correlation has been found as to be r=0.6387. For reliability, test-retest and Kuder-Richardson methods have been applied. A sample of 47 students has been chosen for the test-retest method. The test has been applied twice with certain time intervals. The correlation between the results obtained has been found as r=0.84. This result shows that the reliability is high. KR-20 formula also has been used for reliability. The reliability index has been found as r=0.95. This result also is a pointer that the items are in a high degree of internal consistency. Statistical analyses have also been carried out on the test. The graphical illustrations were examined first and it was observed that the distribution did not fit the normal distribution and thus was not in a symmetrical appearence. The fact that the avarage, mode and median values are different, has been interpreted as being another pointer to an a symmetric distribution. A significant difference can be found in favour of the city centre when the achievements are compared between city and its adjoint towns in the whole test score and in all scores at levels related to targets, and in favour of private schools (p<0.05) in an achievement comparison between government schools and private schools. In the comparison based on sex, no meaningful difference has been observed (p>0.05). Most of the standardization stages have been covered. In conclusion, a Mathemathics Achievement Test which has greatly been standardized with high reliability, having discrimination power will help students to decide on the area they will choose. Guidance and counseling units may help the student to make better decisions on their selection of area and profession when this test is used along with other criteria.