How Children Deal With Conflicting Advice in Web Forums: Impact of Author and Message Cues

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

Abstract

The Internet is quickly developing in the direction of user-generated-content. Web forums are a clear
example: users can post a question and wait for other users’ responses. But because in forums any user
can say anything about any topic, there is usually a great variability in terms of authors’ expertise and
message arguments. To make it more complicated, quite frequently responses in forums provide
conflicting solutions. Thus, although web forums offer great opportunities for students, they also pose
several challenges in terms of information processing. Are middle school children prepared to cope with
the demands of such scenario, e.g. to analyze the credibility of forum messages? Eastin (2008) has
proposed that credibility assessments during childhood might depend on students’ cognitive development.
Younger students (7-11 years) might be able to process credibility cues based on tangible information
(e.g. real life experience), but not jet cues demanding more abstract reasoning (e.g. scientific sources
supporting an advice). Our study aimed to empirically test this hypothesis.
Methods
One hundred seventy-two students (grades 6-8) navigated on four web forums about daily life topics
(anxiety, pets, gardening and skiing). In each forum a user requested advice on a particular issue. Then,
two messages followed that proposed conflicting advices: one from a self-reported expert on the field,
and another from a user with a random username (e.g. ‘Qwerty’). One message explicitly mentioned a
reliable source, and the other included a personal successful experience. In half of the scenarios the selfreported
expert advised using a reliable source, and in the other half she used a personal experience. To
avoid a potential confounding between the advice and the experimental manipulations, across participants
a particular advice was provided in each of the four groups. Students answered several comprehension
questions about the forum discussion and finally recommended one of the advices to a fictitious friend.
Results
Results revealed that 6th graders recommended advices based on personal experience, whereas 7-8th
graders mostly recommended messages from self-reported experts. Younger students justified this
selection by mentioning more often the author’s direct experience, whereas older students mentioned
more often her authority. Finally, the use of a reliable source to support the message arguments was
seldom considered by the participants.
Conclusions
Our results support the view that students’ recommendations and justifications in web forums can be
linked to their cognitive development (Eastin, 2008). While younger students (11 years old) might prefer
messages with real experiences given their propensity to reason based on physically tangible information,
older students (12-13 years old) might favor messages which credibility is based on more abstract
information, such as author’s expertise. Considering the use of a reliable source as a credibility cue might
be too abstract for these grades. This issue would be tested in future studies with older students.
Comprehension data, which is still being analyzed, would be used to discard an alternative hypothesis:
younger students may just understand better messages based on personal experiences.

Authors

• Ladislao Salmerón, University of Valencia

• J Francois Rouet, Université de Poiters

• Victoria García, University of Valencia

• Mônica Macedo-Rouet, National Centre for Pedagogical Documentation