Objective: This study has done in order to provide a test to examine the selection of grapheme kind and location on spelling errors.

Materials & Methods: This study is a descriptive and cross- sectional and validity assessment research. The children were selected from the 8th territory of ministry of education (two classes from girlish school and two classes from boyish school) by non probability sampling from sample of convenience. At the first grade the words of Farsi book were analyzed based on the grapheme kind and location and some of them were selected. Different experts like speech and language pathologist, linguists and teachers scored those words. Finally 93 words were selected as words of test. They were dictated to students at first grade by the tester during 2 stages separately in one stage and the other stage the tester and the teacher do it for another group during 2 stages separately. The dictation errors were determined and data were analyzed by correlation coefficient and paired T test.

Results: Findings showed that there was no significant difference between spelling test errors of first and second stage of test performance by the tester neither in kind of grapheme selection (P=0/12), nor in location of grapheme (P=0/16). There was significant difference between spelling test errors of tester and teacher test performance in kind of grapheme selection (P<0/01), but no significant difference was seen in location of grapheme (P=0/12).

Conclusion: These findings show that this test has intra rater reliability either in kind of grapheme selection or location of grapheme, but it has inter rater reliability just in location of grapheme.