Validity evidence based on response process in the Cloze test

Abstract This study aims to explore the errors differences presented in a Cloze test of kids’ extreme punctuations and find evidences of the validity for answer process. 277 students from third, fourth and fifth grades of public education of São Paulo participated on the survey. For data collection, which occurred collectively, it was held a traditional Cloze test that substitutes the fifth words for a dash of the same size of the word hidden, as well as verbatim correction. Results showed evidences of validity based on answer process, indicating answer divergences between students with high punctuation and low punctuation. Data discussion points some implications about cognitive process involved on the emission of kids’ answers.