نبذة مختصرة : This article is intended as a sumplement to Ketidaktepatan Penggunaan Validitas Butir dan Koefisien Reliabilitas dalam Penelitian Pendidikan dan Psikologi article. The author suggests the interpretation given by earlier measurement specialists to item-total correlation indices as part of item reliability index, instead of item validity index. The index for item validity should be related to some other criteria out of the instrument being developed. Meanwhile, related to the issue of testing some hypothezised reliability coefficient (KR-20 or Cronbach Alpha), the author suggests that F distribution or its related confident interval be used, as developed and suggested by Feldt (1965).
No Comments.