The Effects of Text Structure on the Revision of Learners' Misunderstandings

• In Event: Learning From Text in Today’s Society: Reconciling Multiple Incoherent Sources of Information

Abstract

Evidence indicates that refutation texts are effective in promoting learning of scientific concepts, as they make learners detect
problems in their intuitive conceptions and contrast their views with the scientific conceptions (e.g., Ariasi & Mason, 2010; Kendeou & van den Broek, 2007). However, learners not only develop wrong conceptions before studying from texts but can also misunderstand what they are studying for the first time, especially when texts cover complex conceptual systems and learners lack sufficient prior knowledge relevant to the topic at hand (Graesser et al., 2002). We developed a technique to enable learners to revise and repair the misunderstandings that are developed while learning a new topic: revision texts.

In Pilot Study 1, 27 undergraduate students with low prior domain knowledge learned about plate tectonics from a presentation including text and graphics. Participants thought aloud while studying from the presentation. We coded their verbalizations and the results showed that 14 participants (52%) generated misunderstandings. These misunderstandings were related to the notion of ridges, the process of crust recycling, and the
differences between plate collisions. In Pilot Study 2, 50 participants with low prior domain knowledge learned from the same presentation on plate tectonics. We inserted multiple-choice questions into the presentation asking about the aspects that were typically misunderstood in Pilot Study 1. The results revealed that 43 participants (86%) made wrong choices, selecting the misunderstandings as correct responses.

We developed a new version of the presentation constructing explanations addressing the aspects that were typically misunderstood. The explanations reinstated
critical ideas of the presentation necessary to understand these aspects correctly. We further developed
two new versions of the presentation. The conventional presentation included the original presentation
and the explanations. The revision presentation included the original presentation and parts pointing out
the misunderstandings before the explanations. Thus, the revision presentation was constructed in the
same way as refutation texts, as it identified learners' misinterpretations before providing correct
information.
74 undergraduate students with low prior domain knowledge on plate tectonics participated in
an experiment in which they were randomly assigned to either the conventional or the revision text
condition. They studied from the presentation and then solved two sets of questions: critical questions (N
= 10, Max. Score = 30) asked about the aspects typically misunderstood while control questions (N = 5,
Max. Score = 15) asked about other aspects.
An independent samples t-test indicated that participants in the revision condition (M = 16.51, SD = 5.01)
outperformed those in the conventional condition (M = 14.19, SD = 4.83) in critical questions, t(72) =
1.99, p < .05, d = 0.47. There were no differences between the revision (M = 10.30, SD = 2.17) and the
conventional condition (M = 10.81, SD = 2.52) in the control questions, t(72) = -0.92, p = .36.

Authors

• Hector Garcia Rodicio, Universidad de Cantabria

• Emilio Sanchez, Universidad de Salamanca