Abstract:
Purpose The purpose of this study was to construct effective English proficiency tests in writing for Mathayom Suksa three, four and five students, Procedure Three forms of English proficiency tests in writing were constructed: Form A, B and C, There were 150 items in each form which were divided into two sections : Section One and Section Two. After being pretested with MS. 4 and MS. 5 students, item analyses were performed. Good items were selected from each form on the basis of item difficulty and discrimination, and combined into a single form of 100 items. The resultant English proficiency test in writing was comprised of the following categories. Items 1- 20 Punctuation and Spelling Items 21 - 40 Style and Form Items 41 - 60 Structure and Syntax Items 61 - 80 The Use of Connectives Items 81 100 Organization and Rearrangement This test was post-tested with the samples of 214 MS. 3 students, 202 MS. 4 students and 215 MS. 5 students. Item difficulty, item discrimination, means, standard deviations, standard errors of means, reliability and standard errors of measurement were calculated. Results An English proficiency test in writing was produced. The major findings from the post test were that 69 items were good for all levels with the remainder good for some levels. The means of the three levels were 35.724, 45.663 and 48.986 respectively. After the t- ratios were calculated to determine the difference between the means of the three levels it came out that MS.3 was different from MS. 4, MS. 4 was different from MS. 5 and MS. 5 was different from MS. 3. As for the reliability of the test items, they were satisfactorily high at all levels.