by Annemarie Palincsar
University of Michigan
June 11, 2013
"Comprehending and learning from Internet sources: Processing patterns of better and poorer readers" by Susan Goldman, Jason Braasch, Jennifer Wiley, Arthur Graesser, and Kamila Brodowinska
Volume 47 Reading Research Quarterly (RRQ) (2012)
The Internet offers so many possibilities for supporting information gathering; that is both its blessing and its curse! Students have incredible amounts of information at their fingertips, but, of course, the quality and accessibility of that information varies dramatically. Unfortunately, Internet sites do not come with any warnings about their reliability; in addition, users have to do some sophisticated thinking to: integrate information they are likely to glean from various sites, recognize where there are disparities in the information; determine what the gaps are in the answer to their query, make judgments about the adequacy of the information, and monitor whether or not they have been successful in their search; that is, the extent to which they have satisfied the original intent with which they launched an Internet search.
Goldman and her colleagues have a history of conducting sterling research that has informed the field’s understanding about how readers make sense of text, including how they make sense of - and integrate - information across multiple documents. In the RRQ study, they were interested in capturing the differences in the activity of students using websites when those learners were characterized in terms of how much they learned from their search; hence the term “better and poorer readers” in their title.
One reason the research of Goldman and her colleagues is so helpful is the clarity and robustness of the conceptual framework that they use to describe reading comprehension. The RRQ research is no exception. The authors draw on the construction-integration model of Kintsch (1998), which, as its name suggests, proposes two processes essential to comprehension. When we read, we use the information presented to construct – or build – the meaning of the text ideas. In addition, we integrate these newly constructed ideas with the existing ideas that we already have regarding the topic. The product of this meaning construction is called a mental representation of the text. Kintsch actually proposed that there are two mental representations: the one that reflects the ideas captured in the words and phrases of the text itself, which he called the textbase and the mental representation that results from integrating the textbase with prior knowledge, which he called, the situation model.
Goldman et al. investigated readers developing and elaborating their mental representations of the causes of the eruption of Mt. St. Helens. The study was well designed; the researchers constrained the sites that were navigable so that they could control the content; they included a total of seven sites (all were representative of sites that a typical user might access via the Internet); three were deemed reliable sites (i.e., they were prepared and hosted by known reputable organizations), three were deemed unreliable, and one was a commercial educational site. The researchers used both a 30-item assessment, as well as essay writing, to evaluate learning before and after using the websites. As the participants (n=34), who were college students, navigated the various websites in order to advance their understanding of the driving question, they were asked to say everything that they were thinking about as they conducted their search (i.e., this was a think-aloud study). The students were also asked to rank order the seven sights in terms of how reliable they thought they were.
While the better and poorer learning groups did not differ in terms of the prior knowledge they had about the topic, they did differ in terms of the gains that they made from pre- to post-assessment. Furthermore - and what is particularly noteworthy - is that the good learners were distinguished from the poor learners by the finding that the better learners produced essays that integrated concepts from across the sites, resulting in more complete and accurate situation models. To understand these differences, it is helpful to look more closely at the time spent on the activity and how the time was spent. All learners, regardless of their status as better or poorer learners, spent comparable amounts of time engaged in the task, but the better learners spent more time on the reliable sites, while the poorer learners were not as discriminating about where they spent their time. This finding, in part, reflects the finding that the better learners tended to be more adept at evaluating the reliability of sites (although there was not a statistically significant difference between the two groups’ rankings).
To understand the processes the readers engaged in on the various sites, the researchers coded the think-aloud data, noting, for example, the use of self-explanations and paraphrasing, the building of connections within and between sites, the extent to which learners monitored for sense-making, and the extent to which learners made comments suggesting that they were evaluating the information on the sites. To summarize the findings from this analysis: better and poorer learners did not differ that much with respect to the processes that they used as they navigated within and across site; rather, they differed in terms of when they chose to engage in comprehension-building and comprehension-monitoring activity, with the better learners choosing to execute these processes more frequently when they were consulting more reliable sites. Better learners were more attentive to what they regarded to be more reliable, higher quality sites, while poorer learners’ activity was more guided by what they regarded as relevant to their search.
There are a number of important implications from this study for teachers across all grade levels. The first is that, in the course of reading comprehension instruction, teachers must attend not only to ensuring that their students have a repertoire of strategies to use while interpreting and learning from text (e.g., comprehension monitoring, self-explanation, question generating), but also that they attend to the flexibility with which students apply these strategies as a function of the reading material with which they are working and the purposes for which they are reading. Teachers of younger students need to be explicit about the fact that the good practices that students bring to their reading of texts in isolation are appropriate – and indeed perhaps even more necessary – when reading on the Internet. Finally, this study suggests the appropriateness of teaching students how to make judgments about the reliability of websites, and encouragement that, while unreliable websites may have fascinating facts and/or opinions, their attention should be reserved for those sites that will advance their knowledge building.
Reader response is welcomed. Email your comments to LRP@reading.org