Defining and Assessing Integration in the Context of Argumentative Essays Based on Multiple Sources

• In Event: 45.024 - Integration: A Critical Competency for the Digital Age

Abstract

Purpose
There is growing recognition that the ability to integrate information from multiple sources is key to 21st century learning (Gil, Braten, Vidal-Abarca, & Strømsø, 2010). Nonetheless, no agreed-upon definition of integration exists. Thus, the aims of this study were to: define integration; specify a componential model of integration; and, to test the model’s viability.

Theoretical Framework
The Integrated Framework for Multiple Texts (IF-MT; List & Alexander, in press) guided the conceptualization and operationalization of integration in this study. Specifically, the role of integration in the framework’s three stages was reflected in our model, including the specification of task parameters (preparation), synthesis of ideas expressed within and across documents (execution), and construction of an argumentative essay (production). Drawing on the IF-MT (List & Alexander, in press), we defined integration as the melding of discrete and potentially disparate information, concepts, or perspectives into a unified or cohesive mental or physical representation. Further, while the precise character of integration was expected to vary as a consequence of the task, we expected this melding to involve the synthesis and critical analysis of source content and to result in a cohesive product. Moreover, although integration occurs at different levels of specificity (e.g., idea units or paragraphs), our interest was on integration at the document level, as evidenced in students’ argumentative essays on a controversial topic.

Method
Participants were 114 undergraduates enrolled in an educational psychology course, in which they completed a MSU project, using a library of 10 pro and con documents of varying credibility. The topic was the harmful effects of overdependency on technology. No requirements, other than to read at least 4 sources and produce an argumentative essay, were given. Prior to reading, students indicated their initial positions on the issue and specified their final position after reading. Essays were scored for task parameters (e.g., Claim, Counterarguments; Table 2) and for components signifying integration (e.g., Critical Analysis). All elements were scored on a 0-2 (i.e., none, partial, or complete) scale (Cronach’s α = .93).

Results
We first specified a model of the processes underlying the production of a quality argumentative essay (Figure 1). This model was converted into a Bayesian Network (BN), which tested the interrelations among identified components and subcomponents and indicated critical junctures in the process. The BN then generated the prototypical production model (Figure 2) indicating the probability of achieving a 0-2 score for each component. Next, we examined the change in a component’s predicted influence on other components when critical analysis was fixed at 100% probability of predicting a quality essay (Figure 3). The changes in probabilities in the expected direction confirmed the viability of the specified model.

Scholarly Significance
This study provides a specific definition of integration and the operation of that definition as part of an MSU class project terminating in an argumentative essay. Using Bayesian Networks, we successfully mapped the components and procedure resulting in a well-integrated essay. Importantly, the model generated can be modified to fit other tasks involving integration within and across sources.

Authors

• Anisha Singh, University of Maryland - College Park

• Yuting Sun, University of Maryland - College Park

• Hongyang Zhao, University of Maryland - College Park