Another Brick in the Wall to Understand the Complex Process of Self-Regulated Learning: General and Domain-Specific Features of SRL

Abstract

Self-regulated learning (SRL) is often defined as goal-directed behavior of learners that they display on a regular basis. Models of SRL present general stages of SRL through which learners go without making a difference between different fields of learning. However, there is also research showing that there is not one set of self-regulated actions which assist learners in regulating their behavior in every context. Instead, there are types of self-regulated actions which fit into different contexts for different learners and for different tasks as well as different domains. To shed light on this matter, this study investigated whether self-regulation is a general concept or a domain-specific characteristic. The data of the study were collected from high achieving students studying at the top-ranking science high schools of Türkiye which accept students through a competitive centralized exam. The data were collected via interviews (n = 15) whose participants were selected among the ones receiving the highest score from Self-regulatory Strategies Scale (SRSS). The results of the study showed that self-regulated learning can be depicted as a general characteristic as well as a domain-specific one, as it is a complex process that subsumes both general and domain attributes. The results of this study can be utilized to design impactful SRL interventions, as it provides a comprehensive report of the general and domain-specific phases of SRL.

1. Introduction

SRL has gained more and more significance as a factor affecting academic success, so a need to analyze this process has emerged. In an attempt to define the borders of the definition of this phenomenon, Dinsmore et al. (2008) conducted a literature review. They state that in the early studies of self-regulation, which were shaped by Bandura’s (1986) Social Foundations of Thought and Action, behavioral and emotional regulation were discussed, and the concept was dealt with from a broader perspective. While Bandura’s work keeps its prominence, self-regulation in academic contexts, that is, self-regulated learning, has emerged in research. The definition of self-regulated learning has been touched upon by both endogenous and exogenous theories of learning and has mostly been seen as a reflection of dialectical constructivism. However, the research on the SRL has still not been completed to allow for a deep understanding of self-regulation and to define it as a general or domain-specific subject. Alexander et al. (2011) call attention to this issue and state that SRL-related studies have failed to acknowledge the potential influence of domains on different aspects of SRL, such as cognition, motivation, and emotions. If educational researchers want to obtain a deep insight into the nature of SRL, they need to give immediate consideration to the issue of domain specificity. As stated by Alexander et al. (2011), there is a need to understand the complex phenomena of SRL and determine if it is a general or domain-specific concept, as it has not been sufficiently investigated to allow for definitive conclusions. In addition, comprehensive analysis of SRL data is needed to foster student engagement with SRL and to obtain a correct interpretation of student behavior (Pansri et al., 2024). This study addresses this problem and presents a report on general and domain-specific traits of SRL process. With an instrumental case study design, the study involves qualitative data collected from academically successful students studying at top ranking science high schools. It provides a deeper understanding of the nature of SRL-enabling educators to have guidelines on domain-specific and general components of the SRL process to design effective SRL instructions. Thus, the present research fills in a critical gap in the literature by offering a comprehensive and detailed insight into the nature of self-regulated learning (SRL) and elucidating how learners effectively employ SRL strategies across various contexts.

2. Review of Literature

2.1. Models of SRL

There are numerous studies handling SRL as a general concept that depict the processes of SRL accordingly without making distinctions across the subjects. Major models of self-regulation, such as the Boekarts model, Winnie and Hadwin Model, Pintrich Model, and Zimmerman model, take a general approach to SRL and report the process of SRL from a larger perspective. The Boekaerts (2011) model focuses on cognitive and affective/motivational self-regulation. In her Dual Processing model, there are two pathways that students can take, namely the growth pathway and well-being pathway. Learners utilize top-down SRL strategies to learn more and attain their goals in line with their values and needs, thus, they take the growth pathway. Learners also utilize bottom-up strategies to protect their well-being if they feel threatened or insecure. Emotions play a key role in this model, and although it emphasizes the cognitive and emotional parts of SRL, it also acknowledges the importance of the social environment, which may trigger positive or negative feelings. However, what makes this model distinctive is its emphasis on the key role of domain-specific knowledge. Boekaerts (2011) underlines that domain-specific knowledge is the starting point of SRL and lays the ground for the other processes like goal-setting and strategy selection. In this model, the learner can only go through the other phases of SRL if they have the required domain-specific knowledge.

The Winne and Hadwin (2008) model takes another approach; it centers information processing theory and emphasizes the use of metacognitive strategies and the key role of feedback loops. In this model, there are four important and iterative stages, namely task definition, goal setting and planning, use of SRL strategies and adjusting metacognitive processes to enhance learning. Even though the model revolves around metacognitive processes, the roles of social context are also recognized as a part of task conditions and as a part of standards against which goals and progress are evaluated. The only sign of domain specificity in this model is observed in the task definition phase.

The Pintrich (2005) model is also one of the most cited SRL models; it offers insights, particularly into understanding the relation between motivation and SRL. It portrays forethought, planning and activation, monitoring, control and reaction and reflection as stages of SRL. This model presents the idea that in each of these stages, cognitional, motivational/affective, behavioral and contextual regulations take place. In this model, behavioral regulation shares common insights with Bandura’s (1986) triadic model of reciprocal determinism, in which a person’s behavior impacts and is impacted by some of their personal factors, such as beliefs and ideas, as well as their environment, with all three elements—personal aspects, environment and behavior—repeatedly interrelating with each other. It also adds to this triadic model the idea of “individual’s attempts to control their own overt behavior” (Pintrich, 2005, p. 466). In addition, this model also recognizes the role of social environment in the context regulation component of the model and proves its contribution into motivational and attributional behaviors of learners. Also, they can provide a structure within the classroom, encouraging autonomy and building metacognitive skills as well as self-control skills. The model presents general guidelines on SRL rather than domain-specific instructions.

Zimmerman (2000) put forward the most used and cited model of SRL, which is also utilized in this study to explain the results. The cyclical model of Zimmerman consists of three interrelated phases, namely Forethought, Performance and Self-Reflection. Each phase of his model informs and impacts the other phases and allows the learner to regulate their behavior to attain an academic goal. The forethought phase subsumes two processes named Task Analysis and Self-efficacy. In the Task Analysis phase, learners set goals, make an analysis of the task at hand to reach that goal and strategically plan their steps. In this phase, self-motivational beliefs are significant. Self-motivational beliefs are affected by self-efficacy, outcome expectations, intrinsic interest and goal orientation. Self-efficacy beliefs consist of learners’ confidence in their capacity to succeed in a specific task, while outcome expectations are the beliefs of the learner about the benefits of achieving the task at hand. Similarly, intrinsic interest is related to learner’s personal engagement with the topic, whereas goal orientation is about the underlying reason for why a learner is dealing with the task. This could be about mastering a task or simply outperforming others.

When it comes to the Performance phase, self-control and self-observation are in focus. Self-control involves attention focusing, use of task strategies, self-instruction and imagery. Self-observation subsumes two processes, namely are self-recording and self-experimentation. In the metacognitive monitoring phase, learners keep track of their own understanding and progress while they document the inhibitors and catalysts. Finally, in the self-reflection phase, learners perform self-judgment and self-reaction. In the self-judgment phase, learners evaluate their performance against certain criteria and ascribe reasons for success or failure, such as lack of enough effort, correct or incorrect strategy choice and external factors. Based on this evaluation, learners show a reaction which can be adaptive inference or affective response. In the model, the phases are iterative, and it emphasizes the central role of an active learner. Figure 1 illustrates the model and its continuous improvement cycle.

As seen in Figure 1, the Zimmerman model deals with SRL from a general perspective and presents SRL phases that are applicable across domains.

2.2. Studies on the Domain-Specific Nature of SRL

There is also research indicating that SRL is domain-specific, so the SRL process varies from one subject to another. Several researchers (e.g., Alexander, 1995; Greene et al., 2013; Poitras & Lajoie, 2013) have argued that SRL strategies carry obvious signs of being domain-specific. For example, Alexander (1995) underlines how domain-specific processes of self-regulation interact and lead to different processes of regulation in various contexts of learning. Poitras and Lajoie (2013) offer an understanding of this issue by combining empirical SRL studies and theoretical models together to reach an understanding of the process of students’ developing an understanding of historical events as well as how and why they started. While doing so, they avoided reporting declarative knowledge of how such an understanding is created. They signified that to achieve the demanding task of learning, academically flourishing students have gone through an intricate coordination of strategies that are domain-specific. They handled the issue from the perspective of history classes. Students used self-questioning techniques, corroborated subjects or identified system-related or individual factors triggering certain events in history. Such strategies were applied by students to reach an understanding of complex historical phenomena. Poitras and Lajoie (2013) offer a model called the three-phase model of Cognitive and Metacognitive activities in learning through Historical Inquiry (CMHI). An illustration of this model can be seen in Figure 2.

The model given in Figure 2 illustrates a domain-specific perspective on SRL, in which learners repeatedly adapt themselves to cognitive and metacognitive processes to develop an understanding of historical events. This model is important, as it is one of the few SRL models that put emphasis on the domain-specific traits of SRL. The model puts special emphasis on domain-specific monitoring, goal setting and the control phases of self-regulated learning. The significance of inquiry is also emphasized in the model. This is also in line with what Carretero and Lee (2014) and Wineburg (2001) stated, since these studies indicate the domain-specific nature of self-regulated learning. On the other hand, Rotgans and Schmidt (2009) signal no significant difference in the employment of SRL strategies in the fields of English, history and math.

There is also research indicating that the SRL process can have both domain-specific and general characteristics. Endedijk et al. (2014) assert that physics and history tasks share the same metacognitive activities 60% of the time. This means that 60% of the metacognitive activities applied to both physics and history tasks in their taxonomy. The differences are sourced from the use of these metacognitive skills in different domains. To illustrate, learners questioned themselves and organized their insights as a planning activity in both domains. However, while they were trying to infer meanings from complicated history texts in the history tasks, they were dealing with graph axes or coordinate systems in the physics tasks. This leads us to the implication that self-regulated learning can include both general and domain-specific components, which can be used to feed SRL process and help students learn better. Su et al. (2024) applied this notion into scaffolding programming skills of the students and designed domain-specific and general learning. The students trained with domain-specific techniques first and general second showed better performance, which is useful for program developers and practitioners. To elaborate, in this study, Su et al. (2024) designed and implemented two different education programs for two different groups of learners. In the first one, the students started to learn domain-specific SRL strategies first and then they were taught general SRL strategies. In the second group of learners, the students started to learn general SRL strategies first and then they were provided training on domain-specific strategies. The students in the first group showed a significantly better performance in SRL, which leads us to the conclusion that program developers and practitioners can start with domain-specific SRL strategies and continue with general ones to enhance SRL process.

Pressley’s Good Information Processor Model, developed by Pressley et al. (1989), sets forth implications on this matter as well. The model puts emphasis on the use of cognitive strategies effectively. The model advocates five main premises. These are the employment of a range of tactics, possession of metacognitive information related to when, how and why to utilize these tactics, extensive task related information, staying away from interferences and developing habits to apply the aforementioned strategies naturally. When we focus on the first premise of Pressley et al. (1989), it is possible to see that they put emphasis on the use of cognitive strategies. Within these cognitive strategies, they make a distinction of higher and lower cognitive strategies. They state that higher order strategies are general strategies of self-regulation, and they do not depend on the domain. These higher-order strategies help the learner to manage lower-order strategies. To illustrate, as a higher-order strategy, a learner ranks the tasks at hand in accordance with importance and plans accordingly. Lower-level strategies are, on the other hand, domain-specific and controlled by higher-order strategies. Solving a complex math problem step by step can be an example of a domain-specific strategy. The researchers conclude that using higher-order strategies to manage the lower ones is imperative to self-regulate.

Recent studies of SRL also indicate that the domain-specific aspects of SRL should be taken into account when SRL interventions are designed. Lee et al. (2023) carried out their study by implementing SRL interventions, including domain-specific strategies in writing, mathematics and reading at elementary-school level. Participants of the study were placed to one of the three groups that received a different type of training: regular classroom instruction (REG), domain-specific strategy instruction (STR), and strategy instruction within the framework of an eight-phase self-regulated learning instruction (STR + SRL). There was strategy instruction within the framework of the eight-phase self-regulated learning instruction group that received training on both general and domain-specific aspects of SRL, while the other groups received either regular instruction or just domain-specific strategy instruction. The results of the study showed that the group that received STR + SRL training employed more self-regulated strategies, performed better in achievement tests and was less distracted by task-irrelevant thoughts than the STR and REG groups. This study is crucial for indicating the significance of embedding both the general and domain-specific aspects of SRL.

To conclude, despite the early writings on self-regulation, which defined it as a general concept applicable across domains, later studies have highlighted domain-specific elements within the self-regulated learning process. Nevertheless, research in this area has remained incomplete and has not reached saturation, leaving significant gaps in our understanding of how domain-specific factors impact the deployment of self-regulated learning strategies (Greene et al., 2015). This incongruence in research related to self-regulation’s domain specificity reveals the need for further exploration. In addition, although there are studies addressing the nature of SRL (Benick et al., 2021; Lau, 2020), there are not comprehensive studies exploring all stages of SRL from the perspective of domain specificity. The present study seeks to address this gap by offering a novel understanding of the nature of self-regulated learning, thereby contributing to the advancement of the field. With this purpose in mind, the present study investigates if the use of self-regulation strategies of academically successful students varies depending on the field of study. Considering the focus of the present research and the review of literature, the following hypotheses were established.

Hypothesis 1.

Self-regulated learning (SRL) encompasses both general and domain-specific processes, which may be manifested in different stages of SRL.

Hypothesis 2.

The domain specificity of self-regulated learning is influenced by the nature of the task, with high-achieving students exhibiting variations in their self-regulatory strategies, depending on the contextual demands and cognitive challenges of distinct academic domains.

Hypothesis 3.

High-achieving students demonstrate a strong alignment between their general self-regulation capabilities and domain-specific adaptations, suggesting that effective self-regulation is a dynamic interplay of general processes and domain-tailored strategies.

These hypotheses were checked through the results of the study, and they are discussed in the Discussion section.

3. Method

The present study employs a qualitative instrumental case study design, as described by Stake (1995). In this type of research, the selected case is particularly informative about the phenomenon under investigation but is not the primary focus. Instead, the case serves as a tool to explore a broader concept. Consistent with Stake’s (1995) framework, this research seeks to answer the following research question.

• Do the self-regulated learning strategies employed by academically successful students vary depending on the domain of study?
  1a. If so, what are the domain-specific and general SRL strategies?

Thus, the present research investigates if the learners achieving academic success utilize general SRL strategies across disciplines or if they use different strategies for different courses of study.

3.1. Sample

This study has a qualitative design, and the sample of the qualitative data should be determined in a way that should be directly in line with the research questions of the study and the problems that the research attempted to address (Merriam, 2014). Additionally, participants of such studies should be the ones who have personally experienced the phenomena under scrutiny, so that they can share anecdotes related to the topic (Creswell, 2013). With this requirement in mind, data were collected from high achievers studying in top-ranking science high schools in Türkiye through the criterion sampling method. First, five science high schools (SHS) were randomly selected from the top-ranking science high schools in Türkiye. Science high schools accept students based on the exam scores they took from a highly competitive national exam, and the top-ranking science high schools are the ones that accommodate students with the highest exam scores. Considering that the research is the close link between academic success and SRL, students from the latter group were selected, as they have a high potential to be self-regulated (e.g., Duckworth & Carlson, 2013; Zimmerman, 2000; Lourenço & Paiva, 2024).

Once the five schools were determined, so as to determine the students with the most competent SRL skills, the Self-regulatory Strategies Scale (SRSS), developed by Kadioglu et al. (2011), was administered to all 9th- and 10th-graders in these schools (the students studying in the 11th- and 12th-grades were excluded so as not to let the university entrance exam preparation process affect the results of the study). In total, the survey was administered to 857 students. In total, 15 of them were selected based on the score they received from the scale. Three top scorers from each school were selected and invited to an interview.

Table 1. The Student Interviewees.

Student	School	Grade	Gender
Student 1	Science High School A	10th	M
Student 2	Science High School A	9th	F
Student 3	Science High School A	10th	M
Student 4	Science High School B	9th	F
Student 5	Science High School B	10th	M
Student 6	Science High School B	10th	F
Student 7	Science High School C	9th	M
Student 8	Science High School C	9th	M
Student 9	Science High School C	10th	M
Student 10	Science High School D	9th	F
Student 11	Science High School D	9th	F
Student 12	Science High School D	10th	F
Student 13	Science High School E	10th	M
Student 14	Science High School E	10th	M
Student 15	Science High School E	9th	M

shows the characteristics of the sample of the study:

3.2. Instruments

The interview protocol for the participants included semi-structured questions exploring the general and domain-specific characteristics of the SRL process. The questions asked in the interviews were created with the help of two sources. First, the guidelines proposed by Zimmerman and Pons (1986) were used as the theoretical framework. Second, a detailed review of the literature has been conducted, and the results of this review have been used to prepare questions. After the questions were prepared, feedback from three experts holding Ph.D. degrees in the field of education was received. The wording of the questions was revised considering this feedback, and some sub-questions or prompts were added. Then, the interview questions were piloted with two students studying in two different high-ranking science high schools of Türkiye. These interviews were transcribed and critically analyzed by the researcher to see if the questions served to answer the research question. The questions were refined and finalized after the pilot interviews. The data collection process started on 26 May 2022, after all the data collection instruments were prepared and the necessary permissions were granted.

During the interviews, first the students were asked to talk about the SRL strategies they employed. Questions such as “What do you want to achieve this year? What are your future academic goals? What do you do before you start to study? How do you study?” were asked at this stage. Then, the students were asked if these goals, habits or task strategies change depending on the subject matter. In this stage, the students were first given the opportunity to explain if their SRL strategies vary depending on the domain of study, then they were prompted to comment on each component of the Zimmerman model of SRL and if the SRL strategies they make use of vary in that component. While the students were prompted, courses which are not directly linked to one another were used. To elaborate, Turkish and English can be both be considered language courses, so they were not compared. Also, Biology, Chemistry and Physics are grouped as science courses, so they were not compared with each other either. Rather, they were compared with other courses such as History and Geography. Only core subjects that are tested via written exams were included. Some prompts used in this stage included “Are the task strategies you use in history course different from the ones you used in physics course? Is the way you evaluate yourself in a math course different from the way you evaluate yourself in a geography course? What do you do when you have an important Turkish exam, but you have limited time? What do you do when you have an important Chemistry exam, but you have limited time?” The students replied to some questions, stating that that particular strategy did not change across disciplines, or they underlined that they used a different strategy for a different discipline. The process was completed on 12 June 2022.

3.3. Data Analysis

The interviews were transcribed and coded within the framework of the Zimmerman model of SRL using content analysis. This method mainly involves exploring different expressions of people during the interviews to detect certain patterns in the data. By transcribing and organizing the data, the researcher can break a large amount of data into codes and identify patterns in these codes to link these small units with larger concepts or a variety of discussions in the related literature (Bogdan & Biklen, 2007; Patton, 1990). Initially, the interviews were transcribed with the help of a computer-based word processing program. This process started when the majority of interviews were completed, as suggested by Bogdan and Biklen (2007). Once all the interviews were completed and transcribed by the computer processing program, all the audio files were listened to by the researcher, and all the transcriptions were reviewed to make sure that the transcriptions were accurate and complete. Furthermore, the researcher wrote short notes next to the transcriptions at certain points to capture every nuance of the participants’ expressions. Also, a table of demographic information of the participants was created.

Researchers studying qualitative data analysis extensively (e.g., Creswell, 2013; Miles & Huberman, 1994) suggested three actions in qualitative data analysis that should be conducted in tandem. These are data reduction, data display and making conclusions. Data reduction refers to the process of organizing and categorizing the data. Data display refers to the process of placing bits and pieces of information into tables, charts, matrices and so on to keep a systematic record of the data. In the making conclusions stage, the researcher makes inferences based on the data. In this research, these steps were followed. In the data reduction stage, the data were reviewed several times, and a coding system was developed to obtain manageable units of the data (Bogdan & Biklen, 2007). The process was mostly completed with deductive coding. However, there were times when new concepts, perceptions, and unique experiences were revealed by the participants. At these times, inductive coding (Patton, 1990) or open coding (Maxwell & Chmiel, 2014) techniques were used, and new codes were created. With these techniques, the researcher carefully analyzed the content and labeled each SRL component as domain-specific or general. This was carried out considering the answers of the students, stating that they do the same thing while studying regardless of the course, or that what they do significantly changes depending on the domain of the study. However, at certain times, certain units of data allowed the researcher to create more than one category or multiple codes, contributing to the depth of the discussions that could be made with these data. For this reason, some components of SRL are labeled as both general and domain-specific, as these components were revealed to have both of these features. With these data at hand, Table 2 was created (see the Results section). Finally, to ensure the accuracy and comprehensiveness of the categories and codes created by the researcher, the researcher compared the categories and codes with their notes and the audio recordings. Minor changes were made in the codes during this process to ensure that the codes represented the true meaning of the expressions of the participants. Thus, the reliability of the codes was checked once again, and conclusive results were obtained. MAXQDA 2022 (VERBI Software, 2021) was utilized for qualitative data analysis.

During the data analysis process, the researcher addressed specific validity concerns, including reliability, referential adequacy, internal homogeneity, external heterogeneity and reflexivity. To ensure reliability, the researcher conducted a blind coding process in collaboration with three experts in the field of education for three randomly selected interviews. The codes assigned by the experts and the researcher were found to align significantly, and minor adjustments were made to the wording of certain codes. To address internal homogeneity and external heterogeneity, the researcher developed a coding table to ensure consistency and to clearly articulate the rationale for categorizing specific codes.

In addition, reflexivity was integral to the research process, ensuring transparency and rigor. The researcher maintained a reflexive journal to document critical reflections on how their academic background, professional experiences and personal values might shape the study’s design, data collection and analysis. This practice facilitated ongoing self-awareness and helped mitigate potential biases, particularly in interpreting domain-specific nuances within the self-regulated learning (SRL) process.

For referential adequacy, the researcher supported the codes with quotations and provided extensive data to generate implications. Trustworthiness, encompassing credibility, transferability, dependability and confirmability (Lincoln & Guba, 1985; Marshall & Rossman, 2011) were also prioritized. Credibility was enhanced through detailed feedback from field experts and by piloting interview questions prior to data collection. Transferability was achieved by offering a comprehensive description of each research stage, the participants and the context, as well as the rationale for selecting that context. Confirmability and dependability were ensured using the audit trail method (Lincoln & Guba, 1985). The researcher meticulously documented each stage of the process, which contributed to the consistency, transparency, and objectivity of the findings.

Furthermore, the researcher critically engaged with their position throughout the study. As ascholar specialized in SRL, the researcher’s theoretical insights informed the research framework, but it required vigilant reflection to prevent the imposition of preconceived notions on the data. Strategies such as peer debriefing and iterative coding were employed to challenge and refine interpretations, ensuring they were grounded in the participants’ perspectives. The researcher’s reflexive engagement and detailed documentation of each stage contributed to the consistency, transparency and objectivity of the findings, offering a balanced exploration of the domain specificity of SRL.

4. Results

The present study investigated if SRL strategies utilized by academically successful students vary depending on the field of study. In other words, if the nature and type of these strategies differ across the subjects or they stay constant across the subjects. The results of the interview showed that the process of self-regulation is complex, and it cannot be categorized as domain-specific or general in a simple way. To elaborate, when the SRL phases of these students were scrutinized, it was seen that the SRL process has both domain-specific and general characteristics, which mutually feed into one another. The following table shows SRL phases that were defined by the students as domain-specific, general, or both.

Table 2. General and Domain-specific Phases of SRL.

SRL Phase	Domain Specific	General
Forethought
Task Analysis
Goal Setting		+
Strategic Planning		+
Self-efficacy
Self-Motivation Beliefs	+
Outcome Expectations		+
Intrinsic Interest	+
Goal Orientation	+
Performance Phase
Self-control
Imagery	+
Self-instruction	+	+
Attention Focusing	+
Task Strategies	+
Self-Observation
Self-recording		+
Self-experimentation		+
Self-reflection Phase
Self-judgment
Self-evaluation		+
Causal Attribution	+
Self-reaction
Self-satisfaction/affect	+
Adaptive/defensive		+

Table 2. General and Domain-specific Phases of SRL.

SRL Phase	Domain Specific	General
Forethought
Task Analysis
Goal Setting		+
Strategic Planning		+
Self-efficacy
Self-Motivation Beliefs	+
Outcome Expectations		+
Intrinsic Interest	+
Goal Orientation	+
Performance Phase
Self-control
Imagery	+
Self-instruction	+	+
Attention Focusing	+
Task Strategies	+
Self-Observation
Self-recording		+
Self-experimentation		+
Self-reflection Phase
Self-judgment
Self-evaluation		+
Causal Attribution	+
Self-reaction
Self-satisfaction/affect	+
Adaptive/defensive		+

As Table 2 demonstrates, SRL process is intricate. To elaborate, it starts with goal setting and strategic planning, which were defined as general characteristics of the SRL process, since the student interviewees stated that when they had distal and proximal goals in mind, they would make strategic plans to reach these goals, which do not vary across disciplines.

I never study without setting my goals first. I always study for a goal…I set my goals one day earlier for the next day. I first choose the topics to study, and then I order the topics considering importance or ease. Then, I set my goals to reach for that day. I know my capacity well, so I set realistic goals considering that [capacity]. When I wake up for a new day, everything I am going to do is set. The only thing to do is to do my list.

(Student 9)

As seen in the quote by Student 9, they set a goal and made calculations of reaching that goal. However, they did not make a difference between subjects in these processes. All the participants (n = 15) mentioned the goals they set and their plans to reach these goals without making any differences between the fields of academy. In general, distal goals were about their future career and the proximal goals were about mastering the subjects. They were aware of the importance of creating a pathway for the goal they set, so they mentioned how they planned to achieve that goal immediately after they mentioned their goals.

My goal is to be within the first 500 students in Türkiye and to study engineering [at] Bosporus University. I believe that is the best career path for me as … With this goal in mind, I study regularly. I have plans for the whole month, the whole week and the whole day. I review my plans from time to time. But I stick to my schedule to achieve my dreams.

(Student 2)

However, when it comes to beliefs of self-efficacy, the students (n = 9) revealed domain-specific characteristics. They have distinctive self-motivational beliefs for different courses. The quote below is a concise summary of how these beliefs change across the subjects.

I love Math and Biology. I also love studying chemistry. I am successful in these courses. However, I do not really like Physics. I am not very good at it, so usually I do not feel like studying physics. I force myself to learn or at least memorize.

(Student 1)

Based on the results of the interviews, it can be stated that although the students go through a general process of goal setting and strategic planning, their beliefs in their capacity to accomplish a particular learning task easily differ depending on the subject matter. Similarly, although students have similar outcome expectations (believing that their efforts will eventually bring them positive outcomes) across domains, their intrinsic interest and goal orientation differ across the fields of study (n = 11).

My favorite courses are Math, Physics and Chemistry. I feel more confident about these courses. I do not have to do lots of extra things to comprehend the topics. I can easily go a step further than the others. I do not like doing revision, so Math is the ideal course for me. I solve a lot of math problems, so I excel in Math. But for other classes such as Biology, I must review the topics all the time. It is a bit more challenging for me, but I can still handle it.

(Student 10)

As seen in the quote, students with high SRL skills have an optimistic perspective of learning no matter what the course is, and this is a contributing factor that makes them keep on learning. However, they have a diverse evaluation of the courses in terms of enjoyment, enthusiasm or interest. Also, their reasons for engaging in a learning task differ from one domain to another. While they may be more oriented to achieve mastery for some courses (e.g., Math or Physics in the quote), they may be more oriented to avoid failure in some others (e.g., Biology in the quote).

The performance phase of SRL is the one that domains specificity the most evident. While some students (n = 8) reported to perform a different version of self-instruction in different courses, some others (n = 7) reported to follow a certain routine of self-instruction for every course. Thus, self-instruction can be reported as a phase of SRL carrying both general and domain-specific aspects. The following quote illustrates how the self-instruction phase involves both attributes.

I can say that my learning styles both change and do not change, depending on the course. I mean, I take notes in all classes, and I organize them in accordance with their dates. I do this for all classes. But, also, my learning style changes because for classes like History or Geography or Turkish, I read the class content repeatedly. I sometimes record my voice on my phone, and I listen to my recordings. For Math and Natural Sciences, I answer a lot of multiple-choice questions.

(Student 6)

On the other hand, in the other elements of the performance phase, the students (n = 15) reported that the imagery phase, the task strategies that they employ and their attention span vary depending on the nature of the course. The interview results clearly indicated the domain-specific essence of these processes.

My methods are different for every school subject. For Physics, for instance, I do my best to keep question styles in mind. I focus on the types of questions, and I develop a method for each question type. For Biology and Chemistry, I do constant revision. I think about the subjects deeply. I revise many examples to comprehend the topic. For Math, I just solve problems.

(Student 14)

I do things differently depending on the course. For courses [that] are more about reading and memorizing, I take notes of key words, and I study on my own. For other courses, such as Math or Natural Sciences, I like studying with my friends. We answer multiple-choice questions together. That really helps.

(Student 2)

As seen in the quotes, what they do in the performance phase differs depending on the course. They justify it by emphasizing the idea that the nature of the course requires them to employ different strategies. It has also been reported by the students that while they benefit from the imagery technique for some courses, especially social sciences, for courses involving more calculation or numerical study, they benefit from other methods, such as solving complex problems. Similarly, while they preferred methods like activating prior content for lessons like History, Geography or Biology, they reported employing strategies like starting from an easy question and going on with more difficult ones for lessons like Physics or Math.

While the self-control phase of SRL was mainly described as domain-specific, the self-observation phase was significantly described as a general method (n = 15). As stated earlier, the students employ a variety of SRL strategies across domains. However, what does not change is the way they experiment with strategies and record their performance while doing that which enables them to adjust their strategy. Rather than adopting a particular strategy, the students with developed SRL skills adapt methods by experimenting with them and recording their performance, which provides them with the necessary data to adjust the strategy for future performance. Thus, they gradually excel in the particular strategy they employ for a particular course.

Finally, the self-evaluation phase of SRL contains both general and domain-specific traits, in accordance with the connection of the phase with the other phases of SRL. To elaborate, the self-evaluation phase consists of general aspects like goal setting, strategic planning, outcome expectations and self-instruction phases. This phase has been reported as a general habit of students, just as the other phases consist of general traits. The students reported that they evaluate themselves by asking themselves questions, taking some tests on their own, reflecting on their exam results or comparing their performance with their friends or other external criteria. However, the attributions that they make to their performance vary depending on their beliefs of self-efficacy related to that particular course. To illustrate, the majority of the students (f = 13) stated that they have a particular interest in math. Therefore, they show more enthusiasm for math problems. As a result, they also have higher self-efficacy beliefs when it comes to this particular field of study. They think they have the capability to solve a range of math problems and feel more satisfied when they do so. Thus, they tend to attribute their success to their math skills and correct use of strategy. On the other hand, some students (f = 5) expressed that they have lower self-efficacy when it comes to Physics. Although they regulate their behavior and keep studying to execute tasks related to this course, they stated that they are less interested in this course and feel less self-satisfaction for this course. As a result, they tend to link their low performance in Physics to the mismatch between the design of the course and their skills or other external factors like inefficient instruction methods. However, what position they take at the end of this process can be rightfully defined as adaptive rather than defensive. The participants expressed having difficulty in different courses and experienced failure in these ones, but rather than taking a defensive position, they show efforts to keep learning while addressing the possible causes of their low performance. This particular phase of SRL can be defined as general, as this was portrayed as a general routine of the participants.

I do self-evaluation with the help of the multiple-choice questions that I [answered]. For example, Geometry is my favorite school subject. I can see the connections immediately. However, “ratio and proportion” problems take some time to solve. I realized that I was not really good at this subject. Therefore, I used different sources to improve myself on this topic. After I feel that I can solve problems well enough, I tell myself that “OK, I have learned this subject.”

(Student 11)

As put forward by Student 11, self-evaluation is a part of their routine, and the student makes a clear connection between their beliefs of self-efficacy and the satisfaction they get as a result of this evaluation. However, at the end of the evaluation, they take an adaptive position and choose another path to learn, which is the use of different sources in this case.

As the analysis of the interviews with the students revealed, the SRL process has both general and domain-specific aspects. The analysis indicated that all the students acknowledged the cycle of self-regulation and the stages of forethought, performance and self-reflection, which feed and reshape each other. However, when we go into the details of the self-regulation process, we can observe that certain elements of SRL (e.g., self-efficacy beliefs, intrinsic interest, the task strategies employed by the students to accomplish a goal and self-satisfaction) vary depending on the course, whereas some other elements (e.g., goal setting, strategic planning, outcome expectations, self-evaluation) are general routines of the SRL process. Within this distinction, self-instruction stands out as a process embodying both domain-specific and general characteristics.

All in all, in line with the results presented above, it is appropriate to state that self-regulated learning is a convoluted process bringing both general and domain-specific aspects together, and both of these aspects are inseparable parts of the SRL process.

5. Discussion

The focus of this study was investigating the general and domain-specific aspects of SRL. The study yielded several results, which could be enlightening to develop an in-depth understanding of the SRL process. Previous research on this topic is diverse. While there are studies dealing with SRL as a general trait, there are also studies touching upon its domain-specific characteristics. However, the findings of this study suggest that domain-specific characteristics and general traits of SRL complementarily create the SRL process, which confirms the hypotheses of the study.

The first hypothesis of the current research indicates that self-regulated learning (SRL) encompasses both general and domain-specific processes, which may be manifested in different stages of SRL. The findings of this study that support this hypothesis are in line with the evidence in the literature supporting this finding. To illustrate, Zimmerman (2000) presents a cyclical model that describes the self-regulation process in general without distinguishing between academic subjects. Forethought, performance or volitional control and self-reflection processes are cyclical and valid through academic subjects and this study benefitted from Zimmerman model of SRL to report the results, as it provided a theoretical framework to put SRL process. Although the Zimmerman model did not make any specific references to domain specificity, it still provides room to accommodate domain-specific characteristics, as self-efficacy beliefs have been cited as domain-specific (e.g., Capa-Aydın et al., 2018), and task strategies are defined as “task-specific in their scope” and “context-specific in their effectiveness” by Zimmerman (2006). The results and the first hypothesis of this study are in line with these findings. To elaborate, the interview results yielded empirical findings on how these learners go through phases of SRL as a general principle and how stages inform one another, allowing the learners to adjust their goals, plans or strategies. However, the results also showed how these phases also have domain-specific characteristics that vary across fields of study. Based on these results, it is possible to deduce that one of the underlying reasons for domain specificity is the specificity of self-efficacy beliefs. The relation between self-regulation and self-efficacy has been extensively researched (e.g., Bandura, 2002; Caprara et al., 2008; Joo et al., 2000; Usher & Pajares, 2008). The research indicates that self-efficacy beliefs are the major driving force for regulating behavior to utilize strategies, and they show perseverance to accomplish a certain task. The participants of this study expressed that they feel less capable of succeeding in certain classes, while they believe in their capabilities in succeeding in some others. As a result, they reported achieving less pleasure (or no pleasure at all) in learning some courses, and they need external reasons, such as obtaining good grades or passing important exams, to regulate their behavior in such lessons. As a result, they feel less intrinsically motivated and take a more defensive approach in their self-evaluation process for such courses. However, for the courses they feel more capable of, their goal orientation comes from the joy of being a high achiever in a course and mastering the content, and their motivation is closely linked with their intrinsic interest. Thus, it is appropriate to conclude that the roots of domain specificity of self-regulated learning lie in the domain-specific nature of self-efficacy.

Another interesting finding of this study is the disclosure of how the general and domain-specific characteristics of SRL inform and feed into one another as a result of the interaction of cognitive and metacognitive processes of SRL. To elaborate, there is evidence that metacognitive activities are shaped depending on the domain, such as Math, History, or Science (Schraw, 2006; Shanahan, 2009; Van der Stel & Veenman, 2010; Veenman, 2005). Meijer et al. (2006) pointed out that 60% of the categories in their taxonomy of metacognitive activities are pertinent to the tasks of Physics and History at the same time, but, still, they were inherently different in terms of the domain and the nature of the task that required the employment of metacognitive activities. In addition, learners might do the same metacognitive activities before learning physics or history; but while they are studying history, they struggle to comprehend difficult texts or create links between the events, and they make calculations in Physics. This takes us to a premise put forward by several researchers (e.g., Winne, 2001, 2005, Winne & Hadwin, 1998, 2008, 2010) that argues that learners organize their learning methods via an iterative and adaptive process through which cognitive and metacognitive activities interact. Similarly, as Van Drie and Van Boxtel (2008) point out, learning takes place in the context of completing authentic tasks in the pertinent subject matter. These findings support the second hypothesis of this study, stating that the domain specificity of self-regulated learning is impacted by the nature of the task, with high-achieving students exhibiting variations in their self-regulatory strategies depending on the contextual demands and cognitive challenges of distinct academic domains. Thus, in line with the results of the present study, the research shows that there is a domain-specific element in the SRL process, which interacts with and supplements the other general processes. Boekaerts (1992, 1995, 1996a, 1996b) agrees with this notion and recognizes the domain-specific aspect of SRL. Boekaerts divides self-regulation into six elements, namely domain-specific knowledge and skills, cognitive strategies, cognitive self-regulatory strategies, motivational beliefs and theory of mind, motivational strategies and motivational self-regulatory strategies. Although Boekaerts and the findings of the present study are in line with the acknowledgement of the domain-specific facet of SRL, they diverge in the role of domain-specific knowledge. To elaborate, Boekaerts states that domain-specific knowledge is the base of SRL structure, and it gives direction to the other phases, such as goal setting, strategic planning, self-recording or monitoring. However, the findings of the present study indicate that learners go through general stages of goal setting and strategic planning, after which the impact of domain specificity can be seen. Although SRL has a cyclical nature, the phases of goal setting and strategic planning tend to occur in a more linear order and form the base of the process. Thus, the domain-specific traits take on roles in the later stages of SRL. However, the present study also acknowledges the co-occurrence of general and domain-specific SRL processes, which is in line with the findings of Poitras and Lajoie (2013). These findings also match with the third hypothesis of this study, indicating that high-achieving students demonstrate a strong alignment between their general self-regulation capabilities and domain-specific adaptations; this suggests that effective self-regulation is a dynamic interplay between general processes and domain-tailored strategies. These findings also match with those by Zimmerman and Schunk (2011), underlying the capabilities of self-regulated learners to effectively use SRL strategies across domains. Furthermore, they also align with the findings Alexander et al. (2011), highlighting that the reciprocal interaction between general and domain-specific SRL strategies reach its optimum level in the learning process.

The findings of this study have potential implications for practice in assisting teachers in adapting teaching strategies and designing instruction as well as SRL interventions, since, considering the domain-specific and general traits of SRL, they can lead to a much more efficient SRL training, as also stated by Lee et al. (2023). To elaborate, tailoring instructional approaches to foster domain-specific self-efficacy in subjects where students struggle, emphasizing achievable short-term goals, scaffolding and providing frequent positive feedback can enhance self-efficacy, which, in turn, enhance SRL process (Bai & Wang, 2023). To illustrate, increasing intrinsic motivation in domains where students feel capable and motivated through challenging yet engaging tasks that align with their interests can enhance SRL development. Additionally, based on the findings of this study, it is possible to state that because metacognitive processes vary across domains, educators can provide explicit instruction on how to employ strategies tailored to each subject. For instance, in science classes, focusing on hypothesis formation and iterative testing can be prioritized, while in history classes, teaching students to identify causal links between events and construct timelines can enrich SRL development (Greene et al., 2015). Also, in mathematics, emphasizing step-by-step problem-solving, attributing failure to inefficient selection or use of task strategies and reflecting on the evaluation of solutions can be helpful for SRL development (Throndsen, 2011).

6. Conclusions

To conclude, the present study acknowledges the domain-specific elements of the SRL process as well as the general processes of SRL, which are inherent in various academic subjects. This takes us to the conclusion that self-regulated learning is a multifaceted process that encompasses both domain-specific and general characteristics, occurring in different phases of SRL as well as concurrently. While SRL models provide a theoretical framework for understanding how SRL takes place, this research emphasizes the need to acknowledge that specific stages of SRL can be domain-specific, some can be general, and others may exhibit both characteristics. By using the detailed guidelines on identifying which stages fall into these categories, this study can be used to support educators and researchers in advancing both future studies and practical applications.

On a final note, it is appropriate to state that the present research has some limitations, which can be addressed in future research. Although this study enlightens the specific phases of SRL and their characteristics, its limitation is that it does not comprehensively focus on specific fields of study and the domain-specific task strategies they require. Also, the current study does not deal with the impact of assessment methods on strategy use, so future studies addressing these issues can develop our understanding of the domain-specific and collective aspects of SRL. Additionally, the interaction between general and domain-specific characteristics can be investigated in the future, which could shed light on how these different dimensions of SRL are combined in varied contexts. In addition, the influence of additional external factors, such as the impact of assessment methodologies and learning environments on self-regulation strategies, can be dealt with by considering the domain-specific and general characteristics of the SRL process and using longitudinal analysis of SRL development to observe how self-regulation strategies evolve over time and in response to changes in the learning context, which can also be explored via the same lenses. Finally, for future investigations of SRL development, this study offers hints on the essential qualities of specific SRL phases. Designing SRL investigations considering the domain-specific and general traits of SRL can be much more efficient for the students.