Quality of children's knowledge representations in digital text comprehension: Evidence from pathfinder networks

Children in primary school read digital texts for school purposes while current research has shown that forming a coherent knowledge structure of such texts is challenging. We compared the quality of ninety 6th grade children's knowledge structures after the reading of four different hierarchically structured digital text types: linear digital text, digital text with overview, hypertext, and hypertext with overview. Psychometric pathfinder network scaling of relatedness ratings were used to assess children's knowledge structures. For each text type, we compared the similarity of the children's knowledge structures to both a sequential (linear) model and a qualitatively richer expert model. Moreover, we examined to what extent similarity of children's knowledge structures with the two models predicts their reading comprehension. Children's knowledge structures were overall more similar to the sequential model. Although similarity with the sequential model predicted reading comprehension in all four text types, similarity with the expert model accounted for additional reading comprehension variance in hypertext and hypertext with overview. Prior knowledge accounted for the variance in comprehension in linear digital text, even after controlling for similarity with the models. Evidence suggests that children can cope with the mental demands of a hierarchically structured digital text