A Review of Self-regulated Learning: Six Models and Four Directions for Research

Hadwin et al. (2011, in press) and Järvelä and Hadwin (2013), together with other colleagues (for a review, see Panadero and Järvelä, 2015), have explored the potential of SRL theory in explaining regulation in social and interactive features of learning, e.g., use of information and communication technology (ICT) and computer-supported collaborative learning (CSCL) settings. The exploration of SRL and metacognition with this particular purpose is relatively recent, with 2003 identified as the year for the first empirical evidence published (Panadero and Järvelä, 2015). Additionally, the model is strongly influenced by Winne and Hadwin’s (1998) model, as noted in Section “History and Development of the Model.”

History and Development of the Model

One of their premises is that, despite the advantages of collaboration and computer-supported collaboration for learning (Dillenbourg et al., 1996), collaboration poses cognitive, motivational, social, and environmental challenges (Järvelä et al., 2013; Koivuniemi et al., 2017). To collaborate effectively, group members need to commit themselves to group work, establish a shared common ground, and negotiate and share their task perceptions, strategies, and goals (Hadwin et al., 2010); in other words, they need to share the regulation of their learning (SSRL). The key issue in SSRL is that it builds on and merges individual and social processes, and it is not reducible to an individual level. It is explained by the activity of the social entity in a learning situation (Greeno and van de Sande, 2007), including situational affordances that provide opportunities for SSRL to happen (Volet et al., 2009).

As mentioned above, SSRL is a field recently developed within SRL. Because of this, the model proposed by Hadwin, Järvelä, and Miller (hereinafter referred to as the SSRL model) has changed significantly from their first proposition in the 2011 handbook to the chapter in the forthcoming SRL handbook. The two biggest changes incorporated in the latest are: the authors have clarified their perspective on what is Co-regulated Learning (definition below) and they have incorporated and clarified the influence of COPES (Winne, 1997) in their model (Hadwin et al., in press).

The Model

The SSRL model (Hadwin et al., 2011, in press) proposed the existence of three modes of regulation in collaborative settings: self-regulation (SRL), co-regulation (CoRL), and shared regulation (SSRL) (Figure 12). First, SRL in collaboration refers to the individual learner’s regulatory actions (cognitive, metacognitive, motivational, emotional, and behavioral) that involve adapting to the interaction with the other group members.

Secondly, CoRL in collaboration “refers broadly to affordances and constraints stimulating the (student’s) appropriation of strategic planning, enactment, reflection, and adaptation (occurring when in interaction with other students or group members)” (Hadwin et al., in press, p. 5). This regulatory level is the one that has been in the most dispute in the field, as its use has not been consistent (Panadero and Järvelä, 2015).

Finally, the third type, SSRL in collaboration, occurs when “deliberate, strategic and transactive planning, task enactment, reflection and adaptation” are taken within a group (Hadwin et al., in press, p. 5). The key difference between SSRL and CoRL is that, in the former, the regulatory actions “emerge through a series of transactive exchanges amongst group members” whilst in CoRL they are guided or directed by (a) particular group member/s.

What are the significant changes between the 2011 model version and the forthcoming version? First, the CoRL mode has been reconceptualized based on the empirical evidence (Panadero and Järvelä, 2015). Hadwin et al. (2011) proposed three types of CoRL: (a) temporary mediation (by other than the learner) of regulated learning to promote SRL, (b) distributed regulation of each other’s learning in a collaborative task, and (c) a microanalytic approach focusing on interactions through which social environments co-regulate learning. In their forthcoming proposal, they have shifted the focus to the effects of collaborating alone and have not discussed the microanalytic approach in such detail. Another crucial change is that they have considered the reviewed empirical evidence that CoRL and SSRL could both occur “as groups progress through different phases on their collaboration and not always SSRL nor will co-regulation happen in isolation” (Panadero and Järvelä, 2015, p. 199).

Hadwin et al. (2011) conceptualized SSRL as unfolding in four loosely sequenced and recursively linked feedback loops (Figure 13) taken from Winne and Hadwin (1998). During the first loop, groups negotiate and construct shared task perceptions based on internal and external task conditions. Through the second loop, groups set shared goals for the task and make plans about how to approach the task together. In the third loop, groups strategically coordinate their collaboration and monitor their progress. Based on this monitoring activity, the groups can change their task perceptions, goals, plans, or strategies in order to optimize their collective activity. Finally, in the fourth loop, groups evaluate and regulate for future performance. In essence, when groups engage in SSRL, they extend regulatory activity from the “I” or “you” level to regulate their collective activity in agreement (Hadwin et al., 2011).

In the forthcoming model proposal, the four-phase cycle remains, but under different labels, now using the ones proposed in Winne and Hadwin’s work. Additionally, there is a crucial change: Winne’s (1997) COPES architecture is introduced for the first time in the SSRL model. This addition clarifies, especially, the (meta)cognitive processing at the three regulatory modes along with the effects on motivation and emotion (Hadwin et al., in press, Figure 1).

Empirical Evidence Supporting the Model

The SSRL authors have been working toward empirical verification of their model (e.g., Järvelä et al., 2013). Meanwhile, other researchers have also conducted a growing number of studies on SSRL. A review on CoRL and SSRL by Panadero and Järvelä (2015) extracted three main conclusions. First, different levels of social regulation were identified: a less balanced type called co-regulation, in which one member of the group takes the lead; and a jointly regulated type, in which goals are negotiated and strategies are shared, known as SSRL. Because of this, those authors proposed to reconceptualize how the CoRL and SSRL modes intertwine, which constitutes one of the main changes in the forthcoming version of the model. Secondly, empirical evidence of the occurrence of SSRL in cognitive, metacognitive, motivational, and emotional shared areas were found. This finding is important, as it shows that shared regulation happens within all SRL areas. And third, there was evidence that SSRL might promote learning and performance. Additionally, new research published after the review has continued strengthening the empirical evidence around the model (e.g., Järvelä et al., 2016a,b).

Instruments

At this time, no classical measurement instruments (e.g., questionnaires) have been developed under the SSRL model, even though there is research in the field using self-reported data (e.g., Panadero et al., 2015a). Because of the contextual nature of interpersonal regulation of learning (Vauras and Volet, 2013), new methodologies have been developed to investigate SSRL. Current instruments combine scaffolding and measures in context as, for example, a computer-supported environment to promote group awareness, planning, and evaluation (e.g., Järvelä et al., 2015). Additionally, the joint effort of the model authors has been in developing multi-modal data collections including objective data (e.g., eye tracking, physiological responses) triangulated with subjective data, such as students’ conceptions and intent (Hadwin et al., in press).