Quantitative Corpus Analyses of Character Errors in Primary School Students' Chinese Writings in Taiwan

This paper quantitatively describes differences of character error patterns in terms of the psychological effects of form (F), sound (S), and meaning (M) on primary school students' substituting wrong characters for right ones. The corpora from which the character errors were collected consisted of 2,104 and 1,998 compositions written respectively by 357 second, third, and fourth grade and 301 fifth and sixth grade students in the first semester of the 1993 academic year. The error patterns were partitioned into 7 categories: F, S, M, FS, FM, SM, and FSM, the first three of which formed one significance test group and the second three of which composed another. The one-factor repeated measures ANOVA model was used in the significance tests. The results of this study showed significant differences at the level of .01 in the effects between sound and form and also between sound and meaning. There were also significant discrepancies between form-sound and form-meaning, between form-sound and sound-meaning, and between sound-meaning and form-meaning as well. When we separated the character errors into the second, third, and fourth grade corpus and the fifth and sixth grade corpus and conducted a significance test for each, both of the tests rendered the same results as that for the entire corpus composed of the two groups. This indicated that the grade factor did not significantly contribute to the differences in the effects of character features.