Feedback Consistency Effect on Writing-to-Dictation Task in Chinese

Abstract The influence of feed-forward consistency (FFC) (i.e., the consistency of mapping from the orthographic form to the pronunciation of the whole character) on character recognition and reading aloud tasks have been well documented in a sizable literature in which subjects performed better on feed-forward consistent characters than inconsistent characters. Analogous to the FFC effect in reading aloud, the feedback consistency (FBC) effect (“consistent” when a pronunciation always maps onto one orthographic form, and “inconsistent” if the pronunciation maps onto multiple orthographic forms) was observed in spelling tasks at the phoneme-grapheme levels in English (e.g., Weekes, Castles, & Davies, 2006) and French (e.g., Alegria & Mousty, 1996), where subjects had better spelling performances on feedback consistent words than inconsistent words. Although Ziegler and Muneaux (2007) showed that the size of FBC effect was predicted by the reading level (determined by a feed-forward task) of children in a spoken word recognition task, the words that are phonologically similar, in fact, are also orthographically similar in alphabetic scripts. Therefore, the poor performance on feedback inconsistent words may be attributed to the difficulty in choosing the correct answers among similar orthographic forms. Unlike alphabetic scripts, homophones in Chinese characters do not always share similar orthographic forms and orthographically similar characters may have different pronunciations. These properties allow researchers to investigate the FBC effect without confounding with orthographic similarity. The aims of the present study are three-fold: first, to find out whether similar FFC effect exists in both reading aloud and writing-to-dictation in Chinese; second, to evaluate the influence of the number of homophones of a character on writing-to-dictation across grades; and, third, to observe the possible interactions between FFC and the number of homophones. Phonetic semantic compound characters were selected as stimuli. A reading- aloud task was administered to 1,590 students and a writing-to-dictation task to 2,194 students from grades 1 to 6. The stimulus characters were categorized in terms of homophone numbers and consistency values. Data collected were analyzed using ANOVA. As expected, significant positive FFC effect (consistent characters better than inconsistent characters) was observed in reading aloud across grades. In the writing-to-dictation task, significant homophone effect was found from grade 2 to grade 6. Post hoc analyses revealed that a positive FFC effect occurred when the characters had many homophones, whereas negative FFC effect occurred when the characters had fewer homophones. The findings clarify the influences of feed-forward and feedback consistencies on writing without the possible confounding orthographic similarity factor. The significant interaction effect between FFC and homophone number suggests that the orthographic similarity of the family members in a homophone family might affect the FBC effect in different tasks.