Elsevier

Computers & Education

Volume 50, Issue 3, April 2008, Pages 1069-1083
Computers & Education

Exploring differences between gifted and grade-level students’ use of self-regulatory learning processes with hypermedia

https://doi.org/10.1016/j.compedu.2006.10.004Get rights and content

Abstract

Research involving gifted and grade-level students has shown that they display differences in their knowledge of self-regulatory strategies. However, little research exists regarding whether these students differ in their actual use of these strategies. This study aimed to address this question by examining think-aloud data collected from 98 gifted and grade-level students engaging in a complex learning task: utilizing a hypermedia environment to learn about the circulatory system. We also examined both declarative knowledge and mental model measures of learning to determine whether these groups differed in their actual performance. Our results show that gifted students did outperform grade-level students in all outcome measures. In addition, gifted students more often utilized more sophisticated self-regulatory strategies (e.g. coordinating informational sources) than grade-level students. Grade-level students were more likely to use less effective strategies that are less likely to promote the acquisition of knowledge (e.g. mnemonics). Recommendations for future intervention studies are based upon these findings.

Introduction

Special educational programs for gifted students exist because these students seem to learn differently than their grade-level peers. However, the question as to why gifted students tend to be more successful remains somewhat of a mystery. Numerous authors have suggested that students in general are more successful when they engage in self-regulated learning (SRL; Boekaerts and Corno, 2005, Greene and Azevedo, in press, Pintrich, 2000, Winne, 1995, Zimmerman, 2000, Zimmerman, 2001). Recent research examining how students learn complex and challenging tasks has suggested that successful students deploy key self-regulatory strategies and processes (Azevedo and Cromley, 2004, Azevedo et al., 2004a, Azevedo et al., 2005, Azevedo et al., in press). These studies have focused on actual learning tasks and utilized more objective SRL behavior data-collection techniques such as think-aloud protocols (Ericsson & Simon, 1993), rather than self-report measures. In this study, by using think-aloud protocols and a pretest–posttest design to analyze students’ learning of a complex and challenging task with a hypermedia environment, we aim to determine if gifted students do outperform grade-level students when learning with a hypermedia environment, and whether gifted students utilize more effective SRL processes than their grade-level peers. If gifted students do perform more successfully, we assert that their use of more effective SRL processes may be one reason why these students outperform their grade-level peers. If true, these findings would suggest that specific SRL-based interventions may be effective for helping grade-level students learn complex and challenging tasks. Specifically, these interventions may be embedded in hypermedia learning environments to facilitate the SRL processes associated with learning outcomes.

SRL is a conceptual framework for describing how students engage in the learning process. This framework is used to understand students’ cognition, motivation, behavior and context as they plan, monitor, control, and reflect on those aspects of their own learning (Pintrich, 2000, Winne, 2001, Zimmerman and Schunk, 2001a, Zimmerman and Schunk, 2001). When students are effective regulators of their learning, they are able to achieve academic goals (Pintrich & Zusho, 2002).

Pintrich (2000) presents a model of SRL that is defined by four assumptions: first, all learners are active, in that they make decisions and initiate behavior to further their knowledge or understanding; second, all students have the potential to regulate their learning; third, students are aware of some goal or criterion to which they should compare their progress as they learn; and fourth, the SRL activities of a student mediate the relations between the context and the individual, and the eventual achievement for that individual. Within the context of these four assumptions, the model designates four areas that students can regulate when they are learning: their cognition (e.g., goal-setting, employing and monitoring of cognitive strategies); their motivation (e.g., self-efficacy beliefs, values for the task, interest); their behavior (e.g., help-seeking, maintenance and monitoring of effort, time use); and the learning context (e.g., evaluation and monitoring of changing task conditions). It is assumed that students will cycle through phases of planning, monitoring, controlling, and reflecting in these four areas while they learn, though the degree to which this occurs depends on the learning context (Pintrich, 2000). For example, in learning about the circulatory system with a hypermedia learning environment, successful students most likely need to coordinate the multiple representations of information; including text, diagrams, and video (Azevedo et al., 2005).

In practice, however, students are not always effective at regulating their learning (Paris & Paris, 2001). Students can fail to use self-regulatory skills for many reasons. For example, they may not have prior content knowledge or know when to enact certain strategies to help them reach their goal (Azevedo, Winters, & Moos, 2004); they may not have the motivation, nor the control of their motivation to persist at a difficult task when they lose interest (Moos and Azevedo, 2006, Wolters, 2003); they may not plan appropriately to reach their goals (Vye et al., 1998); they may not monitor their progress towards those goals within changing task contexts (White, Shimoda, & Frederiksen, 2000); or they may not know when to seek help (Newman, 2002). Therefore, students’ regulatory behavior, or lack thereof, will have an impact upon their learning, and students who regulate their learning effectively are more likely to be successful (Boekaerts and Corno, 2005, Butler and Winne, 1995, Pintrich, 2000, Winne, 1995). It would seem intuitive that gifted students would be better regulators of their learning than grade-level students, but the research is not conclusive (see below). The question remains as to whether the differences in performance between gifted and grade-level students are at all attributable in part to their use of SRL behaviors.

Definitions of gifted students have focused upon the interaction of creativity, task commitment, and above-average ability (Renzulli, 2002). More recently, however, researchers have been examining differences in the use of metacognition between gifted and grade-level students. For example, research has found gifted students possess more declarative knowledge regarding metacognitive strategies, and more complex strategies in general (Alexander et al., 1995, Carr et al., 1996). Carr and colleagues (1996) speculated that this elaborate understanding and use of more complex strategies may be what makes gifted students more likely to transfer these strategies to other domains. Based upon these studies, gifted students clearly possess more metacognitive skills, but the question remains as to how this influences their learning.

Surprisingly, while these studies have found a difference in students’ understanding of metacognitive monitoring strategies, they have not found significant differences between gifted and grade-level students in terms of their reported use of metacognitive skills and monitoring processes (Carr et al., 1996, Zimmerman and Martinez-Ponz, 1990). It is counter-intuitive that gifted students would possess more knowledge of SRL skills than grade-level students but not actually deploy them differentially. The lack of conclusive evidence may be due to the fact that much of the research in this area has been correlational, focusing on surveys rather than learning tasks or objective measures of learning. In addition, many studies of SRL have relied upon questionable student self-report data regarding SRL behaviors (see Winne & Jamieson-Noel, 2002), as opposed to more reliable and objective methods such as think-alouds (see Azevedo, 2005, Azevedo and Cromley, 2004). Therefore, while theory would suggest that one possible explanation for the greater success of gifted students is their understanding of metacognitive skills and processes, the current research has not effectively investigated this claim. We hypothesize that studying students engaged in challenging learning tasks, coupled with the collection of think-aloud data, will reveal differences in both the performance and use of SRL strategies and processes between gifted and grade-level students. One way to capture students’ SRL use during challenging learning tasks is through studying how they learn with hypermedia.

Hypermedia is a computer-based learning environment in which the basic unit of information is represented as nodes (Jonassen & Reeves, 1996). Consisting of a video clip, sound bite, graphic, page of text, or even an entire document, the node is the unit of information storage in hypermedia. The structure of nodes in a hypermedia environment typically creates a dynamic knowledge base where students can access any node in varying sequences, depending on students’ interests and goals. Because students can access information of their choosing through the non-sequential format of the information, they can pursue personal goals when learning with hypermedia. However, though these attributes should foster students’ active participation in the construction of knowledge (Williams, 1996), empirical research has produced mixed results on the effectiveness of these learning environments. With greater freedom to explore the environment comes greater responsibility for doing so in a productive manner. Research suggests that while hypermedia fosters conceptual change in some students (Jacobson & Archodidou, 2000), other students have difficulty using these learning environments to even develop conceptual knowledge (Azevedo et al., 2005, Azevedo et al., in press, Azevedo et al., 2004, Greene and Azevedo, 2006).

Recent research has begun to examine how the effectiveness of hypermedia environments varies depending upon students’ classification as grade-level, gifted, or in need of special assistance. For example, Liu (2004) explored how a hypermedia problem-based learning environment (PBL) can support the learning of 155 sixth graders. This sample was divided into ability groups through a formal nomination. Of the 155 students in this sample, 26 came from Talented and Gifted (TAG) classes and 114 came from regular education (RegEd) classes. The other 15 students were identified as students that needed additional help, beyond that of the students in either the TAG or RegEd classes. These students, who either had a learning disability or spoke English as a second language, comprised this third ability group (ESL/LD). It should be noted that while students who either have a learning disability or speak English as a second language are different types of students, this classification system was the one used by the school district from which the sample was taken. As such, Liu (2004) relied on the school’s classification system to differentiate students’ into three groups: TAG, RegED, and ESL/LD. These students used a PBL hypermedia program, Alien Rescue, during a 3-week period in their science class. In essence, Alien Rescue presents an ill-structured problem for the students to solve in which the students, acting as scientists, must determine the most suitable location for different alien species. A set of 13 cognitive tools is provided in the hypermedia environment, and these tools are designed to act as scaffolds (e.g. concept database, expert modeling) in order to assist the students in the problem solving activity.

Liu (2004) gathered different types of data, including both quantitative data (e.g. a 25-item multiple-choice test designed to assess students’ understanding of the scientific concepts introduced in this PBL) and qualitative data (e.g. multiple interviews with both the students and teachers that concentrated on several aspects including the science concepts the students learned in Alien Rescue). Results from both the quantitative and qualitative data suggest students from all three ability groups (TAG, ESL/LD, and RegEd) demonstrated a significant gain in their understanding of science concepts from pretest to posttest. Furthermore, data from the qualitative data sources suggest that students from these groups could articulate what they had learned. However, while these results provide promising implications for the use of hypermedia as an educational tool for different groups of students, it is important to note that this specific hypermedia environment included cognitive tools that served as scaffolds during learning. As such, while students from differing ability groups may be able to effectively learn with hypermedia when they are provided scaffolds, it is presently unclear whether students from different ability levels will have similar success learning with a hypermedia environment that does not provide some type of scaffold.

Some research has suggested that students have difficulty learning with hypermedia when they are not provided with scaffolds (Azevedo and Hadwin, 2005, Jacobson and Azevedo, in press). For example, Azevedo and colleagues (2005) found that while some students are able to develop conceptual knowledge of the circulatory system when using a hypermedia environment without scaffolds, other students have difficulty learning with these environments. This line of research has suggested that students may need to self-regulate their learning (e.g., activating prior domain knowledge and engaging in metacognitive monitoring) in order to develop conceptual knowledge with hypermedia, and that some students have difficulty using key SRL processes during learning with a hypermedia environment. As such, the next step in this line of research is to explore whether students from different ability groups differentially use self-regulatory processes during learning in the absence of scaffolds. The goal of this current study is to address this issue by exploring how students of different ability levels (gifted and grade-level) self-regulate their learning with a hypermedia environment that does not provide scaffolds.

This study contributes to the literature on hypermedia learning environments as well as work on gifted and grade-level students by gathering both product and process data regarding students’ engagement in a complex and complicated task, learning about the circulatory system. The use of a pretest–posttest design with an actual learning outcome allows for the analysis of knowledge gains for each group to determine whether the gifted students truly are performing at a higher level, as expected. These knowledge gains include both improvement on declarative measures such as matching items, as well as a better mental model, or conceptual understanding, of the circulatory system. In addition, the collection of think-aloud data provides a more objective measure than students’ self-reports for collecting data regarding student use of SRL processes (Azevedo, 2005, Winne and Jamieson-Noel, 2002, Winne and Perry, 2000). We believe that our research design, which includes an actual learning task with pretest and posttest measures coupled with think-aloud data collection procedures, will demonstrate that one key difference between gifted and grade-level students is their use of SRL strategies and processes.

The hypotheses of this study are as follows:

  • (1)

    Lower mental model pretest scores and classification as a grade-level student will decrease the odds of being in a higher mental model posttest score group.

  • (2)

    Gifted students’ posttest scores on measures of declarative knowledge will be statistically significantly higher than those of grade-level students, after controlling for pretest scores.

  • (3)

    Gifted students will utilize key SRL strategies and processes more frequently than grade-level students, after controlling for variations in the total number of SRL strategies and processes used by each student.

Section snippets

Participants

Ninety-eight (N = 98) middle-school (7th grade) students from a secondary school located in the mid-Atlantic region received community service credit for participating in this study during the Spring of 2004 and 2005. Forty-nine of the students attended regular, grade-level instruction classes. The mean age of these students was 12.2 years (28 girls and 21 boys).

The other 49 students were in a gifted program that provides highly able middle school students an interdisciplinary educational program

Hypothesis 1: lower mental model pretest scores and classification as a grade-level student will decrease the odds of being in a higher mental model posttest score group

Given the ordinal nature of the mental model scoring rubric (low, intermediate, high), we utilized an ordinal regression with cumulative logits (DeMaris, 2004, Hosmer and Lemeshow, 2000) to determine the influence of both mental model pretest score as well as grade-level versus gifted classification upon students’ posttest mental model. This analysis produced two prediction equations, one predicting the odds of being in the low mental model posttest score group versus being in a higher group,

Discussion

The literature suggests that one key difference between more and less successful learners is their use of SRL processes (Azevedo et al., 2005, Boekaerts and Corno, 2005, Greene and Azevedo, in press, Pintrich, 2000, Winne, 1995). Seeing as gifted students tend to perform differently than their grade-level peers (Winner, 2000), it would seem reasonable to assume their use of SRL processes might be one explanation. Yet, previous research examining gifted and grade-level students found differences

Acknowledgements

This research was supported by grants from the National Science Foundation (Early Career Grant REC#0133346, REC#063391) awarded to the third author. The authors thank Evan Olson for assisting with the data collection. We thank the anonymous reviewers for their feedback.

References (54)

  • R. Azevedo et al.

    Does training on self-regulated learning facilitate students’ learning with hypermedia?

    Journal of Educational Psychology

    (2004)
  • R. Azevedo et al.

    The role of self-regulated learning in fostering students’ conceptual understanding of complex systems with hypermedia

    Journal of Educational Computing Research

    (2004)
  • R. Azevedo et al.

    Adaptive human scaffolding facilitates adolescents’ self-regulated learning with hypermedia

    Instructional Science

    (2005)
  • R. Azevedo et al.

    Using computers as metacognitive tools to foster students’ self-regulated learning

    Technology, Instruction, Cognition, and Learning Journal

    (2006)
  • R. Azevedo et al.

    Scaffolding self-regulated learning and metacognition: implications for the design of computer-based scaffolds

    Instructional Science

    (2005)
  • Azevedo, R., Moos, D. C., Greene, J. A., Winters, F. I., & Cromley, J. C. (in press). Why is externally-regulated...
  • R. Azevedo et al.

    Can students collaboratively use hypermedia to learn about science? The dynamics of self- and other-regulatory processes in an ecology classroom

    Journal of Educational Computing Research

    (2004)
  • M. Boekaerts et al.

    Self-regulation in the classroom: a perspective on assessment and intervention

    Applied Psychology: An International Review

    (2005)
  • A. Bunt et al.

    Probabilistic student modeling to improve exploratory behavior

    User Modeling and User-Adapted Interaction

    (2003)
  • D. Butler et al.

    Feedback and self-regulated learning: a theoretical synthesis

    Review of Educational Research

    (1995)
  • M. Carr et al.

    Where gifted children do and do not excel on metacognitive tasks

    Roeper Review

    (1996)
  • M.T.H. Chi

    Self-explaining: the dual processes of generating inference and repairing mental models

  • M.T.H. Chi

    Commonsense conceptions of emergent processes: why some misconceptions are robust

    Journal of Learning Sciences

    (2005)
  • C. Conati et al.

    Using Bayesian networks to manage uncertainty in student modeling

    User Modeling and User-Adapted Interaction

    (2002)
  • J.G. Cromley et al.

    Self-regulation of learning with multiple representations in hypermedia

  • A. DeMaris

    Regression with social data: modeling continuous and limited response variables

    (2004)
  • Encarta. (2003). Microsoft Electronic Encyclopedia [Computer software]. Redmond, WA: Microsoft...
  • Cited by (63)

    • The influence of a web-based learning environment on low achievers’ science argumentation

      2020, Computers and Education
      Citation Excerpt :

      Metacognition could be another reason for LAs' lower learning performances (Bernacki, Byrnes, & Cromley, 2012; Greene et al., 2008). According to Greene et al. (2008), HAs demonstrated more self-reflective behaviors than LAs did in learning about the human circulatory system in a hypermedia environment. Bernacki et al. (2012) found that the HAs were also usually able to use more sophisticated metacognitive skills such as highlighting, reviewing annotations, seeking additional information, and monitoring their own understanding in a technology-enhanced learning environment, while the LAs simply passively followed instructions.

    View all citing articles on Scopus
    View full text