摘要:Affable Reading Tutor (ART) is an online reading lesson designed for children who are starting to comprehend reading. A digital, human-like character (virtual peer) in ART serves as a peer model that demonstrates the use of the reading comprehension strategy called questioning to help improve the learners comprehension of expository texts. This study, with 141 boys and girls in the fourth and fifth grades in the United States, examined the effects of virtual-peer presence (presence, absence, and control) on learners text comprehension and also the effects of learner gender and virtual-peer attributes (human-like male, human-like female, robot still image) on learners perceptions of their peer and on their text comprehension. The results revealed that the virtual-peer presence group outperformed both the absence group and the control group in the immediate and delayed posttests text comprehension. There were mixed results in the impacts of learner gender and virtual-peer attributes on text comprehension. The learners perceptions of their agent were not differentiated by neither learner gender nor virtual-peer attributes. The findings are discussed with virtual-peer design implications.