Here we report on MELD-SCH (MEgastudy of Lexical Decision in Simplified CHinese), a dataset that contains the lexical decision data of 1,020 one-character, 10,022 two-character, 949 three-character, and 587 four-character simplified Chinese words obtained from 504 native Chinese users. It also includes a number of word-level and character-level variables. Analyses showed that the reliability of the dataset is satisfactory, as indicated by split-half correlations and comparisons with other datasets. Item-based regression showed that both word-level and character-level variables contributed significantly to the reaction times and error rates of lexical decision. Moreover, we discovered a U-shape relationship between word-length and reaction times, which has not been reported in Chinese before. MELD-SCH can facilitate research in Chinese word recognition by providing high quality normative data and information of different linguistic variables. It also encourages researchers to extend their empirical findings, which are mostly based on one-character and two-character words, to words of different lengths.
Keywords: Chinese; Lexical decision; Megastudy; Word recognition.