Search

We investigated the effects of just-in-time guidance at various stages of inquiry learning by novice learners. Thirteen participants, randomly assigned to an intervention (n = 8) or control (n = 5) group, were observed as they learned about DC electric circuits using a web-based simulation. Just-in-time instructional prompts to observe, predict, explain, systematically test, collect evidence, and generate rules were strongly associated with diagnosing and correcting misconceptions, and constructing correct scientific concepts. Students' repeated use of predictions, systematic testing, and evidence-coordinated reasoning often led to formulating new principles, generalizing from observed patterns, verifying comprehension, and experiencing "Aha!" moments. Justin-time prompts helped learners manage embedded cognitive challenges in inquiry tasks, achieve a comprehensive understanding of the model represented in the simulation, and show significantly higher knowledge gain. Just-in-time prompts also promoted rejection of incorrect models of inquiry and construction of robust scientific mental models. The results suggest ways of customizing guidance to promote scientific learning within simulation environments.

This special issue's editors invited discussion of three broad questions. Slightly rephrased, they are: How well do self-report data represent theoretical constructs? How should analyses of data be conditioned by properties of self report data? In what ways do interpretations of self-report data shape interpretations of a study's findings? To approach these issues, I first recap the kinds of self-report data gathered by researchers reporting in this special issue. With that background, I take up a fundamental question. What are self-report data? I foreshadow later critical analysis by listing facets I observe in operational definitions of self-report data: nature of the datum, topic, property, setting or context, response scale, and assumptions setting a stage for analyzing data. Discussion of these issues leads to a proposal that ameliorates some of them: Help respondents become better at self reporting.

Metacognition is the engine of self-regulated learning. At the object level, learners seek information and choose learning tactics and strategies they forecast will develop knowledge. At the meta level, learners gather and analyze data about learning events to draw conclusions, such as: Is this tactic a good fit to conditions? Was it effective? Was effort required reasonable? Is my ability publicly exposed? As data accumulate, learners shape, re-shape and refine a personal theory about optimal learning. Thus, self-regulating learners are learning scientists. However, without training and tools on which "professional" learning scientists rely, learners' N = me research programs are naïve and scruffy. Merging models of tasks, cognition, metacognition and motivation, I describe software tools, approaches to analyzing data and learning analytics designed to serve three goals: supporting self-regulating learners' metacognition in N = me research, accelerating professional learning scientists' research, and boosting synergy among learners and learning scientists to accelerate progress in learning science.

Background: Guidance during inquiry learning plays an important role in developing conceptual understanding and inquiry skills. This study analysed learner-tutor interactions in a simulation-assisted learning environment to investigate how tutor guidance enabled knowledge construction and fostered epistemic practice. Objectives: This research aimed to illuminate challenges learners encounter in the inquiry process and forms of guidance that support learning in both conceptual and epistemic aspects. Methods: This study uses a mixed methods approach. We analysed video recordings in which nine participants asked 72 questions and the microsequences of interactions immediately surrounding and including each question. We coded properties of each question and whether the tutors' utterances were intended to increase (upregulate) or decrease (downregulate) the complexity of the inquiry processes, and used a two-step cluster analysis to explore groupings emerged from tutors' regulation guidance and learners' questions. Results and Conclusions: The regulatory intent of tutors' utterances depended on various characteristics of student questions. The microsequences clustered in five categories: 1) upregulated investigation and inference, 2) upregulated evidence-based justification, 3) downregulated cognitive load, 4) downregulated procedural uncertainties, and 5) downregulated perceptual dissonance. Our findings suggest tutors offering guiding prompts should consider dual processes in the inquiry and, by strategically prompting them, strike a balance between the goals of guiding learners to discover scientific knowledge and grounding their conceptual understanding in concepts, data, and procedures. Implications: We emphasize conceptual and epistemic learning should be concurrently guided in scientific inquiry. We propose a bidirectional guidance model as a pedagogical approach to guide instructional practice.

Data used in learning analytics rarely provide strong and clear signals about how learners process content. As a result, learning as a process is not clearly described for learners or for learning scientists. Gaševic, Dawson, and Siemens (2015) urged data be sought that more straightforwardly describe processes in terms of events within learning episodes. They recommended building on Winne's (1982) characterization of traces -- ambient data gathered as learners study that more clearly represent which operations learners apply to which information -- and his COPES model of a learning event -- conditions, operations, products, evaluations, standards (Winne, 1997). We designed and describe an open source, open access, scalable software system called nStudy that responds to their challenge. nStudy gathers data that trace cognition, metacognition, and motivation as processes that are operationally captured as learners operate on information using nStudy's tools. nStudy can be configured to support learners' evolving self-regulated learning, a process akin to personally focused, self-directed learning science.

In scientific inquiry learning, within-task guidance tailored to the learner's domain knowledge and inquiry skill may be essential to promote intended learning outcomes. However, due to dynamic complexity across the timeline of inquiry learning, principles for designing tailored guidance are elusive. In this study, experienced tutors provided just-in-time guidance to 11 learners. We analysed tutor-learner interactions to investigate how tutors adapted guidance. We found tutors provided five types of guidance: prompts, support for domain knowledge, assessments, hints, and feedback. Guidance was provided when learners made errors, expressed difficulties, or asked questions; or when the tutor judged a learner successfully demonstrated a skill and was ready to progress to a follow-on skill. Based on these results, we propose a model for tailored, just-in-time guidance in simulation-assisted inquiry learning environments.

Learner modeling systems so far formulated model learning in three main ways: a learner's "position" within a lattice of declarative and procedural knowledge about highly structured disciplines such as geometry or physics, a learner's path through curricular tasks compared to milestones, or profiles of a learner's achievements on a set of tasks relative to mastery criteria or a peer group. Opening these models to learners identifies for them factors and relations among factors. Open learner models tacitly invite learners to regulate learning. However, contemporary learner models omit data about how learners have and should process information to learn, understand, consolidate and transfer new knowledge and skills. What to do with information opened to learners is opaque. I propose incorporating trace data about learning processes in learner models. Trace data allow generating learning analytics that inform self-regulating learners about potentially productive adaptations to processes they have used to learn. In a context of big data, such elaborated learner models are positioned to collaborate with self-regulating learners. Together, they can coordinate symbiotically, creating a platform for the system to improve its models of learners and for learners to more productively self-regulate learning.

Developing knowledge-transforming skills in writing may help students increase learning by actively building knowledge, regardless of the domain. However, many undergraduate students struggle to transform knowledge when drafting essays based on multiple sources. Writing analytics can be used to scaffold knowledge transforming as writers bring evidence to bear in supporting claims. We investigated how to automatically identify sentences representing knowledge transformation in argumentative essays. A synthesis of cognitive theories of writing and Bloom's typology identified 22 linguistic features to model processes of knowledge transforming in a corpus of 38 undergraduates' essays. Findings indicate undergraduates mostly paraphrase or copy information from multiple sources rather than engage deeply with sources' content. Eight linguistic features were important for discriminating evidential sentences as telling versus transforming source knowledge. We trained a machine learning algorithm that accurately classified nearly three of four evidential sentences as knowledge-telling or knowledge-transforming, offering potential for use in future research.

We investigated knowledge and attitudes before and after reading refutation texts augmented by different kinds of persuasive information and how emotions mediated the process of knowledge and attitude change. Undergraduates (N = 424) enrolled in 4 universities from 3 countries read a refutation text on genetically modified foods (GMFs) and were then randomly assigned to receive additional information about advantages of GMFs, disadvantages of GMFs, or both. After studying, students reading about advantages of GMFs had significantly more positive attitudes than students who read about disadvantages. There was also a significant reduction in misconceptions; participants in the positive-oriented text condition showed the largest learning gains, particularly those who held more positive initial attitudes. Epistemic emotions of curiosity, frustration, hope, and enjoyment mediated attitude change while confusion mediated relations between prereading attitudes and postreading knowledge. In addition, the direct relationship between prior attitudes and surprise was moderated by type of text. When reading about both advantages and disadvantages of GMFs, participants reported significantly less surprise when compared with those who read about either advantages or disadvantages of GMFs. To foster conceptual change when learning about complex topics, refutation texts may be paired with persuasive information that is aligned with accurate conceptions.

Self-regulated learning is an ongoing process rather than a single snapshot in time. Naturally, the field of learning analytics, focusing on interactions and learning trajectories, offers exciting opportunities for analyzing and supporting self-regulated learning. This special section highlights the current state of research at the intersection of self-regulated learning and learning analytics, bridging communities, disciplines, and schools of thought. In this opening article, we introduce the papers and identify themes and challenges in understanding and supporting self- regulated learning in interactive learning environments.

This special issue presents a sample of modern work on self-regulated learning (SRL) among high ability and gifted students. It includes diverse views about the construct per se, and gifted students' and their teachers' accounts about SRL and factors they believe moderate it. Zeidner and Stroeger (this issue) set the stage with a sketch of an extensive literature about SRL that has deep roots in North American educational philosophy and practice. The menu of work here is fundamentally well done and, in varying ways and degrees, slightly provocative. A trite observation would be these articles don't fully represent the multiple facets and complex articulation among them comprising SRL, especially given relatively less research with participants identified as academically talented or gifted. In this situation, I would be pedantic to point out such-and-such is omitted or this-or-that is underrepresented. Rather, using admittedly using idiosyncratic standards, I select a few matters for discussion and, hopefully, constructive critique. Other commentators would likely apply different filters.

Our article introduces the "Journal of Educational Data Mining's" Special Issue on Educational Data Mining on Motivation, Metacognition, and Self-Regulated Learning. We outline general research challenges for data mining researchers who conduct investigations in these areas, the potential of EDM to advance research in this area, and issues in validating findings generated by EDM.

A program of research is necessary to examine the psychometric properties of instruments designed to measure individuals' achievement goal orientations. Recently, research on achievement goal orientation has examined the stability of achievement goals to assess how context might influence individuals' achievement goals. Accordingly, studies are necessary to establish factorial invariance across contexts. We examined the psychometric properties of the Achievement Goal Questionnaire (AGQ; Elliot & McGregor, 2001) across task contexts within a single classroom environment. We tested the factor structure by comparing five competing models and evaluated the invariance of the factor structure across four task contexts. Results revealed that the hypothesized four-factor structure was replicated, construct- and discriminant-related evidence of validity were supported, and both internal consistency and test-retest reliability estimates were satisfactory. Moreover, invariance held at all levels across the various contexts. (Contains 6 tables.)

Self-explanation is a process by which learners generate inferences about causal connections or conceptual relationships. A meta-analysis was conducted on research that investigated learning outcomes for participants who received self-explanation prompts while studying or solving problems. Our systematic search of relevant bibliographic databases identified 69 effect sizes (from 64 research reports) which met certain inclusion criteria. The overall weighted mean effect size using a random effects model was g = 0.55. We coded and analyzed 20 moderator variables including type of learning task (e.g., solving problems, studying worked problems, and studying text), subject area, level of education, type of inducement, and treatment duration. We found that self-explanation prompts are a potentially powerful intervention across a range of instructional conditions. Due to the limitations of relying on instructor-scripted prompts, we recommend that future research explore computer-generation of self-explanation prompts.

Background: Deep versus surface knowledge is widely discussed by educational practitioners. A corresponding construct, levels of processing, has received extensive theoretical and empirical attention in learning science and psychology. In both arenas, lower levels of information and shallower levels of processing are predicted and generally empirically demonstrated to limit knowledge learners gain, curtail what they can do with newly acquired knowledge, and shorten the life span of recently acquired knowledge. Purpose: I recapitulate major accounts of levels or depth of information and information processing to set a stage for conceptualizing, first, self-regulated learning (SRL) from this perspective and, second, how a "levels-sensitive" approach might be implemented in research about SRL. Method: I merge the levels construct into a model of SRL (Winne, 2011, "Handbook of self-regulation of learning and performance" (pp. 15-32), New York: Routledge; Winne, 2017b, "Handbook of self-regulation of learning and performance" (2nd ed.), New York: Routledge; Winne & Hadwin, 1998, "Metacognition in educational theory and practice" (pp. 277-304). Mahwah, NJ: Lawrence Erlbaum) conceptually and with respect to operationally defining the levels construct in the context of SRL in relation to each of the model's four phases--surveying task conditions, setting goals and planning, engaging the task, and composing major adaptations for future tasks. Select illustrations are provided for each phase of SRL. Regarding phase 3, a software system called nStudy is introduced as state-of-the-art instrumentation for gathering fine-grained, time-stamped trace data about information learners select for processing and operations they use to process that information. Conclusions: Self-regulated learning can be viewed through a lens of the levels construct, and operational definitions can be designed to research SRL with respect to levels. While information can be organized arbitrarily deeply, the levels construct may not be particularly useful for distinguishing among processes except in a sense that, because processes in SRL operate on information with depth, they epiphenomenally acquire characteristics of levels. Thus, SRL per se is not a deeper kind of processing. Instead, it is processing more complex--deeper--information about a different topic, namely processes for learning.

A bottleneck in gathering big data about learning is instrumentation designed to record data about processes students use to learn and information on which those processes operate. The software system nStudy fills this gap. nStudy is an extension to the Chrome web browser plus a server side database for logged trace data plus peripheral modules that analyze trace data and assemble web pages as learning analytics. Students can use nStudy anywhere they connect to the internet. Every event related to creating, modifying, reviewing, linking and organizing information artifacts is logged in fine grain with a time stamp. These data fully trace information students operate on and how they operate on it. Ambient big data about studying gathered au naturel can be tailored by configuring several of nStudy's features. Thus the system can be used to gather data across a wide range lab studies and field trials designed to test a range of models and theories.

In the context of continued growth in the accessibility of information through the internet, recent advances in theories of self-regulated learning present an opportunity to reexamine how learners work with networked resources in constructivist approaches such as problem-based learning, project-­based learning, and collaborative problem solving. We present a Resource Inquiry model consisting of five stages: (1) Set resource inquiry goals, (2) Plan for resource study, (3) Search and select resources, (4) Study and assess new knowledge, and (5) Critique and recommend resources. Our model informs designers of online tools about how to support learners' cognitive and metacognitive strategies when learning activities involve interacting with networked resources.

This document discusses the use of Information Communication Technology (ICT) in education, particularly in the context of blended or online learning environments. It highlights the need for educational institutions to analyze the data available through learning management systems (LMSs) and use it to improve teaching and learning. The document includes contributions from various authors and institutions across different countries, providing insights into how ICT is currently being used in education. It covers topics such as artificial intelligence, applied robotics, automated approaches to create smart and inclusive learning environments, emerging technologies, learning analytics, and learning strategies for smart learning environments. The document also presents recommendations for future research in the field of ICT in education. [Extracted from the article]

Learning analytics are often formatted as visualisations developed from traced data collected as students study in online learning environments. Optimal analytics inform and motivate students' decisions about adaptations that improve their learning. We observe that designs for learning often neglect theories and empirical findings in learning science that explain how students learn. We present six learning analytics that reflect what is known in six areas (we call them cases) of theory and research findings in the learning sciences: setting goals and monitoring progress, distributed practice, retrieval practice, prior knowledge for reading, comparative evaluation of writing, and collaborative learning. Our designs demonstrate learning analytics can be grounded in research on self-regulated learning and self-determination. We propose designs for learning analytics in general should guide students toward more effective self-regulated learning and promote motivation through perceptions of autonomy, competence, and relatedness.

Recent research has shown that for some topics, messages to refute and revise misconceptions may backfire. The current research offers one possible account for this backfire effect (i.e., the ironic strengthening of belief in erroneous information after an attempted refutation) from an educational psychology perspective and examines whether emotions mediate the relationship between self-concept and learning from refutation texts. In an experimental design, 120 undergraduate students responded to a questionnaire focused on their dietary self-concept and were randomly assigned to read an expository or refutation text on the topic of genetically modified foods. Immediately after reading, participants self-reported their emotions followed by completing post-test measures of their knowledge and attitudes of the topic. Results showed an interaction between self-concept and text condition on emotions wherein self-concept predicted negative emotions (i.e., confusion, anxiety, frustration) while reading a refutation text specifically. Further, negative emotions significantly mediated relations between self-concept and post-test knowledge and attitudes. Implications for educational design and future research are discussed.

New methods for gathering and analyzing data about events that comprise self-regulated learning (SRL) support discoveries about patterns among events and tests of hypotheses about roles patterns play in learning. Five such methodologies are discussed in the context of four key questions that shape investigations into patterns in SRL. A framework for this review is provided by a model that structures SRL in terms of: conditions of a task, operations, products generated by operations, evaluations of work and standards used in evaluations (COPES; Winne in "Journal of Educational Psychology, 89", 397-410, 1997). Four recommendations are made for future work on SRL as patterned activity: prune models of SRL with experimental tests, explicitly include goals in data, ensure learners have options for SRL by training them in tactics and strategies, and provide learners access to accurate displays about the events and patterns that comprise SRL.

We examined achievement goals measured by self-reports and by traces (behavioral indicators) gathered as undergraduates used software tools to study a multimedia-formatted article. Traces were operationalized by tags participants applied to selections of text and hyperlinks they clicked in the article. Tags and hyperlinks were titled to represent achievement goal orientations. Self-reported goal orientations did not correlate with goals traced as actions. In separate regression models, traces of goal orientations were stronger predictors of achievement than self reports. We suggest future research include traces in studies of achievement goals because traces reflect proximal events that comprise learning activities that can supplement static orientations that are operationally defined to be indifferent to the dynamics of learning activities. (Contains 3 tables and 1 figure.)

In theoretical models of self-regulated learning, calibration is one important component in successful learning. Two issues of calibration are explored. First, Nelson (1987) suggested the "G" (gamma) coefficient is the most appropriate measure of calibration (judgment accuracy) and rejected signal detection theory's "d'" statistic because data commonly challenge distributional assumptions. We empirically examined this issue, comparing "G" and "d'." Second, we examined whether a learner's calibration varies across three domains of knowledge: general, word, and mathematics. A sample of 266 university students volunteered to participate in the study. Participants were selected from various undergraduate and graduate courses. Participants first answered demographic items. Then they completed three knowledge tests (general, word, and mathematics) and judged correctness for each answer provided. Order of domains was randomly counterbalanced among participants. Results show that "d'" is a valid measure of calibration, that assumptions about underlying distributions can be tested, and that preliminary evidence suggests that "d'" may be a superior measure of accuracy compared to "G." Finally, calibration varied by domain.

Articles in this special issue present recent advances in using state-of-the-art software systems that gather data with which to examine and measure features of learning and particularly self-regulated learning (SRL). Despite important advances, there remain challenges. I examine key features of SRL and how they are measured using common tools. I advance the case that traces of cognition and metacognition offer critical information about SRL that other state-of-the-art measurements cannot. (Contains 1 table.)

Background: A program of research is needed that assesses the psychometric properties of instruments designed to quantify students' achievement goal orientations to clarify inconsistencies across previous studies and to provide a stronger basis for future research. Aim: We conducted traditional psychometric and modern Rasch-model analyses of the Achievement Goals Questionnaire (AGQ, Elliot & McGregor, 2001) and the Patterns of Adaptive Learning Scale (PALS, Midgley "et al.", 2000) to provide an in-depth analysis of the two most popular instruments in educational psychology. Samples and methods: For Study 1, 217 undergraduate students enrolled in educational psychology courses participated. Thirty-four were male and 181 were female (two did not respond). Participants completed the AGQ in the context of their educational psychology class. For Study 2, 126 undergraduate students enrolled in educational psychology courses participated. Thirty were male and 95 were female (one did not respond). Participants completed the PALS in the context of their educational psychology class. Results: Traditional psychometric assessments of the AGQ and PALS replicated previous studies. For both, reliability estimates ranged from good to very good for raw subscale scores and fit for the models of goal orientations were good. Based on traditional psychometrics, the AGQ and PALS are valid and reliable indicators of achievement goals. Rasch analyses revealed that estimates of reliability for items were very good but respondent ability estimates varied from poor to good for both the AGQ and PALS. These findings indicate that items validly and reliably reflect a group's aggregate goal orientation, but using either instrument to characterize an individual's goal orientation is hazardous.

School tasks interact with student motivation, cognition, and instruction to influence learning and achievement. Heeding calls for additional research linking motivational and cognitive factors in learning and instruction on specific tasks within authentic classroom settings we quantitatively and qualitatively track 90 tenth-grade science students' motivation, reported use of learning strategies, achievement, calibration, and task perceptions as they engage in a well-structured task (WST) and an ill-structured task (IST). Students achieved higher grades on, and reported more ease and value for, the WST whereas they utilised critical thinking and peer learning strategies more on the IST. Lower academic achievers calibrated their achievement less accurately on each task and experienced lower grades, interest, ease, and management capability on the IST. Conversely, higher academic achieving students reported more self-efficacy and effort regulation and lower anxiety and elaboration on the IST. Motivation--notably less intrinsic goal orientation in low academic achievers and higher task value and self-efficacy--predicted performance on the IST. The structure of tasks may provide prompts that illicit unique self-regulated learning responses in students. (Contains 4 tables and 1 figure.)

Background: A programme of construct validity research is necessary to clarify previous research on self-regulation and to provide a stronger basis for future research. Aim: A multitrait-multimethod (MTMM) analysis was conducted to assess convergent and discriminant validity of three self-regulation measures: the Learning and Study Strategies Inventory (LASSI; Weinstein, 1987), the Motivated Strategies for Learning Questionnaire (MSLQ; Pintrich, Smith, Garcia, & McKeachie, 1993) and the Meta-cognitive Awareness Inventory (MAI; Schraw & Dennison, 1994). Method bias across all three inventories was also examined. Sample and method: Three hundred and eighteen undergraduate university students (255 female, 61 male, 2 did not specify) were recruited from various courses to participate in research on perceptions about studying and study methods. Participants spent 30-60 minutes completing all three inventories. Results: Evidence for convergent validity was found at the matrix level, but was attenuated when examined at the individual parameter level. Evidence for discriminant validity among traits was modest, and common method bias was evident across all three measures. Conclusions: Results revealed the three inventories yielded different results, which suggests that researchers should be selective in the inventory they use to assess self-regulated learning (SRL).

Researching self-regulated learning (SRL) as a process that evolves across multiple episodes of studying poses large methodological challenges. While self-report data provide useful information about learners' perceptions of learning, these data are not reliable indicators of studying tactics learners actually use while studying, especially when learners are young children. We argue that self-reports about SRL need to be augmented by fine-grained traces that are records of learners' actual activities as they study. We describe how gStudy software unobtrusively collects detailed trace data about learners' use of study tactics as they engage with content presented in learning kits--collections of documents (e.g., texts, graphics, video clips) and tasks (e.g., notes, concept maps) on which learners operate to study. We suggest that trace data can advance research about how learners select, monitor, assemble, rehearse, and translate information to learn it, and provide raw materials for mapping SRL and its effects. Examples from the "Life Cycles Learning Kit" that supports grade 1 students learning about the life cycles of humans and frogs are given.

Realizing the promise of software technologies in education requires thinking differently about how software simultaneously can serve research and contribute to learning. This article examines 3 axioms underlying contemporary educational psychology: Learners construct knowledge, learners are agents, and data include lots of randomness. By drawing out corollaries of these axioms, this research uncovers significant challenges researchers face in using classical forms of experimental research to build a basis for school reform and for testing school reforms using randomized field trials. This article describes a software system, gStudy, that is designed to address these challenges by gathering finer grained data that better support theorizing about the processes of learning and self-regulated learning. This research illustrates how this can be realized and suggests 10 ways that using software like gStudy can help pull up research by its bootstraps and bolster searches for what works.

Links between students' achievement goal orientations and learning tactics were investigated using software (gStudy) that supports a variety of learning tactics and strategies. An achievement goal questionnaire was administered to 307 students enrolled in an introductory educational psychology course. Data tracing study tactics were logged for 80 of these students who prepared for a test by studying a textbook chapter presented as a multimedia document. Using correlations and canonical correlations, we found relationships between goal orientations and activity traces indicating different forms of cognitive engagement. Notably, mastery goal orientation (approach or avoidance) was negatively related to amount of highlighting, a study tactic that is theorized to be less effective than summarizing and other forms of elaborative annotation for assembling and integrating knowledge. (Contains 5 tables.)

Although much has been discovered about relations between self-efficacy and academic achievement, questions remain about links between achievement, the structure of learning tasks, and changes in students' self-efficacy as students engage with a single, complex authentic task. Students' self-efficacy for learning (SEL) and for performance (SEP) was tracked as they worked on well- and ill-structured tasks during their regular class. Students reported higher SEL and SEP for a well-structured task. Moderate achievers reported significantly more difficulty with the ill-structured task. SEP was higher and more stable than SEL, especially in early phases of both tasks. After accounting for overall academic achievement, self-efficacy was a negligible predictor of achievement. Students may perceive various features of each task's structure as difficult. Implications concerning relations among self-efficacy, task structure, and achievement are discussed.

Playing the role of constructive critic of articles in this special issue, I adopt the term self-reliance to refer to work habits, work styles, volition, and self-regulated learning. Starting with the belief that developing self-reliance is an important goal of education, I examine key issues that influence the likelihood of reaching this goal. I assume first that learners are agents. This human quality ubiquitously affects theorizing, researching, and teaching. I examine the nature of this influence in relation to three issues: How learners experiment to develop self-reliance, conditions that influence how learners express and extend self-reliance, and considerations about how to advance theory and practice in the service of enhancing self-reliance.

The trajectory of scholarship about self-regulated learning (SRL) originates in mid-19th-century writings about learners' sense of responsibility in self education. Although Descartes's 17th-century writings implied mental activities consistent with metacognition, a central feature of SRL, these were inarticulate until Flavell and colleagues' studies circa 1970. Since then, research on metacognition and its role in SRL has approximately doubled every decade. Foundations for modeling SRL include Skinner's behaviorism, which acknowledged learners' choices about reinforcers for behavior, and Bandura's social learning theory, with its construct of agency. Research in the 1980s gathered data about SRL mainly using interviews, self-report questionnaires, and think-aloud protocols. These methods were quickly supplemented by observations of behavior and traces of learning activities tightly coupled to features of SRL. Today, SRL research is prominent across a broad spectrum of educational topics. Its importance will grow with trends toward lifelong learning and self-directed inquiries that survey vast information on the Internet, where students control what and how they will learn. Implications for future research include reconceptualizing "error variance" as arising partially due to SRL and capitalizing on software technologies that massively increase access to data about how and to what effects learners self-regulate learning.

Outlines contemporary theories of self-regulation and presents a four-phase model of self-regulation. Builds on this theory to describe features of CoNoteS2 ( a prototype electronic notebook) that support self-regulation through tacit and explicit scaffolding. Discusses how learning theory can be used to drive instructional innovation and technological enhancement in the classroom. (SLD)