How undergraduate students learn: uncovering interrelationships between factors that support self-regulated learning and strategy use

Abstract

Being an effective learner is an important pillar supporting success in higher education and beyond. This research aimed to uncover the extent to which undergraduate students use a set of commonly researched learning strategies, as well as to understand how learning strategy usage relates to key self-regulation factors proposed in influential models of self-regulated learning. Undergraduate students from New Zealand (N = 140) were recruited through course management systems, social media, and campus posters. Data were analysed using correlation and multiple regression. Results show that students reported using more learning strategies that are relatively effective than learning strategies that are somewhat less effective. In addition, effort regulation and student engagement were the most consistent predictors of both learning strategy use and self-reported academic achievement. Building on these findings, this research provides important new insights into the associations between learning strategy usage and pivotal factors that support effective self-regulated learning and academic achievement. As we highlight in the article, these insights have key implications for advancing theory and research on self-regulated learning.

Learning involves the acquisition of knowledge or skills as a result of experiences (Ormrod, 2016). To succeed in any educational context, students must be able to self-regulate their learning (Boekaerts, 1996, Efklides, 2011, Pintrich, 2000, Winne, 1996, Zimmerman, 2000). One important process involves the selection and implementation of learning strategies, and a key aspect of successful self-regulation includes the appropriate use of effective learning strategies (Biwer et al., 2020). Self-regulation is an umbrella term used to describe the various processes involved in appropriate goal setting and effective pursuit of the set goals (Mann et al., 2013). Although self-regulation research has uncovered influential new knowledge (e.g., Duckworth et al., 2016, Stadler et al., 2016), it has yet to provide evidence on how key factors underpinning students’ self-regulation relate to the kinds of learning strategies students use.

To start bridging this important knowledge gap, our study drew from theory and research on self-regulated learning (SRL) to identify self-regulation factors that might predict university students’ use of learning strategies. SRL is an “active, iterative process through which learners purposefully control their behavioral, cognitive, emotional, and motivational aspects of learning to fulfill learning goals” (Li & Lajoie, 2022, p. 838). As the conceptual framework of SRL helps “understand the cognitive, motivational, and emotional aspects of learning” (Panadero, 2017, p. 1), it also provides a good foundation for exploring self-regulation factors that are likely to be associated with individual differences in students’ use of learning strategies. Regarding learning strategy use, at the outset we emphasize that consistent with major models of self-regulation (e.g., Winne & Hadwin’s, 1998 COPES model), our present focus is on learning strategies that are considered applicable across a range of academic domains (Dinsmore et al., 2023). The present focus is not to imply, however, that domain-specific learning strategies do not exist and are not useful for students (Hattie et al., 1996).

Self-regulation factors influencing learning

A comprehensive review of influential SRL models (i.e., Panadero, 2017) highlighted that goal setting and the adoption of a goal orientation toward learning tasks are critical components of all major SRL models (i.e., Boekaerts, 1996, Efklides, 2011, Pintrich, 2000, Winne, 1996, Zimmerman, 2000). Specifically, the goals students set in learning environments play important roles in their choices of learning strategies and influence the effort they devote to employing chosen strategies (Li & Lajoie, 2022). Psychological research has provided convincing evidence that two key self-regulatory orientations (i.e., promotion and prevention) – that are investigated in the present research by means of strategic eagerness and vigilance, respectively (Vriend et al., 2022) – are critical determinants of the goals people adopt within and beyond learning settings (Higgins, 2012, Molden & Rosenzweig, 2016). Briefly, a promotion orientation focuses people’s attention on goals that are related to gains and advancement and is associated with an eager strategic outlook (i.e., preference for using strategies that help maximize gains/advancement even at the cost of increasing errors or losses; Higgins, 2012, Molden & Rosenzweig, 2016). A prevention orientation motivates people to set and pursue goals that focus on avoiding losses and maintaining satisfactory or neutral states. Prevention is associated with a vigilant strategic outlook (i.e., preference for employing strategies that prioritize loss avoidance even if this involves missing out on potential gains; Higgins, 2012, Molden & Rosenzweig, 2016).

In SRL, prevention/vigilance and promotion/eagerness are likely to predict goal setting and task-related goal orientation. As we noted above, goal-related processes are represented in all major SRL models (Panadero, 2017). For example, Boekaerts’ (2011) theorizing on SRL articulated that two key purposes of self-regulation during learning are to avoid/prevent (a) threats to oneself (e.g., avoid low self-esteem) and (b) loss of resources (e.g., time). Both of these purposes map well onto the prevention/vigilance orientation. At the same time, Boekaerts’ theorizing also highlighted that strivings to achieve gains (in resources) is another important driver of self-regulation in learning (e.g., the model of adaptable learning; see Fig. 6 in Panadero, 2017). This striving is well aligned with the promotion/eagerness orientation.

Metacognitive self-regulation (MCSR) refers to the ability of learners to monitor and make judgments about their learning/memory, and to use these judgments to guide their learning-related behavior (Kornell & Bjork, 2008). All major models examined by Panadero (2017) include metacognitive monitoring and/or metacognitive control among the factors/processes they propose. For example, metacognitive monitoring is an important part of the (task) performance phase of Zimmerman’s cyclical phases model of SRL (Zimmerman & Moylan, 2009). In addition, MCSR plays significant roles in Efklides’ (2011; at the task x person level), Pintrich’s (2000; in both the monitoring and control phases of cognition), and Winne and Hadwin’s (1998) SRL models. In the latter model, monitoring and control are relevant to all the critical aspects of the model, namely conditions, operations, products, evaluations, and standards (COPES).

Self-control and attention control are some of the key self-regulation factors proposed by influential SRL models (Panadero, 2017). Self-control reflects people’s ability to initiate behavior consistent with their long-term goal(s), despite the presence of temptations, difficulties, and challenges (Partsch & Danner, 2021, Zhu et al., 2016). Li and Lajoie’s (2022) integrative model of SRL engagement proposes that self-control processes operate in the performance phase of learning (i.e., during students’ engagement with learning tasks) and influence both the selection of cognitive strategies they employ and the level of mental effort they devote to using the chosen strategy/strategies (e.g., how often and for how long they use a given learning strategy). Similarly, Zimmerman’s (2000) cyclical phase model includes self-control as a key factor that contributes to performance/volitional control during learning. More generally, self-control’s importance for learning is highlighted by its positive association with educational achievement (Duckworth et al., 2016, Partsch & Danner, 2021, Stadler et al., 2016).

Attention control is the ability to direct one’s attention and focus on the learning material. During learning, students who display strong attention control can focus on pertinent tasks (Burgoyne & Engle, 2020). Attention control is a common underlying factor that influences performance in various complex cognitive tasks. For example, attention is important for working memory; that is, to move information into their working memory, people need to first pay attention to it (Ormrod, 2016). As far as SRL is concerned, attention control is an important contributor to performance/volitional control during learning; see, for example, Zimmerman’s (2000) cyclical phase model. In a similar vein, Boekaerts (2011) noted that directing (controlling) attention is an important factor underpinning effective SRL.

How learners construe the effort they expend during learning and the perceptions they have about what they do to manage (regulate) this effort are important metacognitive aspects (Efklides et al., 2006). Hence, effort regulation is another important self-regulation factor at play in learning settings (Efklides, 2011, 2019). Effort regulation is “the conscious decision to invest or stop investing effort, the actual investment of, and the fluctuations in mental effort when engaging in a learning task” (de Bruin et al., 2023, p. 4). Thus, effort regulation captures a learner’s ability to invest effort in a learning task (de Bruin et al., 2023) and persist with doing it even when that task is difficult or uninteresting (Virtanen et al., 2015). The allocation of effort to learning tasks is a pivotal facet of metacognitive control and effort regulation is considered a metacognitive skill (Efklides et al., 2006).

Efklides’ (2011) and Pintrich’s (2000) SRL models attribute significant roles to effort regulation. Specifically, in Efklides’ (2011) MASLR model, effort regulation is relevant to both the person and the task x person levels. At the person level, which represents more distal antecedents of self-regulation of learning (e.g., personality traits, motivation orientations; Efklides, 2019), students’ beliefs about their ability to successfully regulate effort during learning may contribute to their “representation of situational and task demands” (Efklides et al., 2018, p. 66). At the task x person level, which “is the level of actual processing and its regulation” (Efklides, 2019, p. 83), beliefs about ability to regulate effort are likely to play out in the type(s) of cognitive processing deployed during learning activities (e.g., learning strategies used) and the ways in which the student monitors and controls their (real-time/online) engagement in learning activities (see, for example, Fig. 5.1 in Efklides et al., 2018).

All influential models of SRL are complex and involve multiple sets of factors. In addition, some factors are represented in most models (e.g., goal orientations; metacognitive self-regulation), whereas others are not (e.g., emotions). Thus, although our choice of potential predictors of use of learning strategies draws from the conceptual SRL literature, we do not propose an exhaustive set of predictors. Moreover, our research investigates a factor (i.e., student engagement) that has not been explicitly considered in leading SRL models. Yet, engagement overlaps with many aspects of SRL (Li & Lajoie, 2022) and considering engagement together with SRL factors helps build “a holistic understanding of students’ learning” (Li & Lajoie, 2022, p. 834).

Engagement reflects the extent to which students are actively involved in tasks associated with their learning (Lei et al., 2018). Therefore, it is likely that engagement has a meaningful association with the learning strategies students use. Specifically, it is plausible that engagement is positively related to general use of learning strategies. Moreover, strong engagement in one’s studies may be an important factor supporting the use of more effortful learning strategies (e.g., elaboration, self-testing, organisation). This is either because engagement is associated with strengthened attentional focus or because strong engagement renders effortful strategies more acceptable to the learner (e.g., via metacognitive beliefs that using effortful strategies leads to stronger learning; see de Bruin et al., 2023 for in-depth discussions of the relationships among perceptions of effort/learning and use of desirable difficulties).

Overall, our overview of pivotal SRL models enabled us to propose a set of self-regulation factors (eagerness, vigilance, MCSR, self- control, attention control, effort regulation, and engagement) that might predict individual differences in students’ use of learning strategies. The lack of existing research connecting these factors to specific learning strategies constrained our ability to make predictions regarding specific learning strategies. Nevertheless, considering the literature we surveyed, it seems plausible that students who effectively manage their thoughts, behaviors, attention, and emotions to achieve their learning goals (i.e., effectively self-regulate) are more likely to use more effortful (and, thus, more effective) learning strategies (e.g., self-testing) than their counterparts who struggle with their self-regulation. For example, in a new conceptual model, de Bruin and colleagues (2023) proposed that effective monitoring and control of both learning and effort – which are important aspects of self-regulation in learning settings – support students’ use of learning strategies that facilitate long-term retention and application of knowledge (i.e., desirable difficulties).

Learning strategies

To explore how students’ learning strategy usage relates to factors that theoretically influence self-regulated learning (i.e., the factors just detailed), this study focused on students’ strategy self-reports for a range of strategies. We targeted select strategies that the literature (laboratory and classroom studies) suggests are relatively ineffective, yet students still use (see, for example, the discussion in Miyatsu et al., 2018), strategies for which effectiveness seems somewhat mixed or modest, and relatively effective strategies. These strategies are also primary ones that have been targeted in the extant literature (e.g., Dunlosky et al., 2013); they are introduced next.

Rehearsal, a learning strategy that involves revising and repeating learned information (Oberauer, 2019), is a commonly used strategy (e.g., in laboratory experiments; Shaughnessy, 1981; in classroom, repeating/rereading underlined/highlighted information; Hartwig & Dunlosky, 2012; Kornell & Bjork, 2007) despite being relatively ineffective for long-term learning and retention (e.g., Bae et al., 2019; Hong et al., 2021). A plethora of laboratory experiments (with impoverished materials) have shown that repetitive rehearsal does little to increase learning (e.g., Craik & Watkins, 1973; Glenberg et al., 1977; Rundus, 1980) and produces lower levels of learning than elaborative learning strategies (e.g., Shaughnessy, 1981). Classroom studies have also failed to find strong evidence that rehearsal on its own aids students’ ability to remember information (e.g., see Sebesta and Speth (2023), for a negative relation between a strategy constellation that included rehearsal/memorizing and exam performance).

Another relatively ineffective strategy for retention of studied content is cramming: study of all content within one session or very few sessions, which are usually conducted right before an assignment is due. Cramming can be effective for immediate performance (when tests are administered immediately after study), but cramming does not support long-term retention (e.g., when tests are delayed for a day or more) (e.g., Dobson et al., 2017, with authentic educational content in a classroom context for an immediate test and 1- and 4-week delayed tests). Nevertheless, many students tend to engage in cramming and believe it to be effective (Dunlosky et al., 2013).

Rereading is another strategy that university students report using (Hartwig & Dunlosky, 2012, Karpicke et al., 2009). Research on its effectiveness is mixed, however. In the laboratory, some experiments with educational texts find that rereading can aid learning (e.g., Rawson & Kintsch, 2005), but other experiments find no benefit to rereading (versus reading once) (Callender & McDaniel, 2009). Mixed findings are also reported in the educational literature: some studies have found rereading to have a negative correlation with academic achievement, while others have found rereading to have a positive correlation with academic achievement (see Dunlosky et al., 2013, Hartwig & Dunlosky, 2012 for more extensive discussions).

We examined four strategies that have been identified as being relatively effective: elaborating, self-testing, organising, and spacing study. Elaboration involves making sense of new information by integrating and associating learned information with previous knowledge (Rhodes et al., 2020). When students engage in elaboration, they express the material in their own words, create examples for the concepts they learn, as well as paraphrase and connect ideas to what they already know (Duncan et al., 2015). Elaboration is generally effective for learning because it involves creating connections between new and pre-existing information, which in turn aids retrieval of that information (Dunlosky et al., 2013).

Self-testing (also known as retrieval practice) involves learners actively recalling information from memory (Rodriguez et al., 2021). Self-testing is effective because it helps strengthen memory for the encoded information. Self-testing also has metacognitive benefits. It improves metacognitive accuracy by helping learners identify what information they can remember and whether the remembered information is correct. Improved metacognitive accuracy, in turn, helps regulate effective study by targeting information the learner was not able to remember or remembered incorrectly (e.g., Thomas & McDaniel, 2013).

Organisation is a strategy that involves learners selecting important information from learning materials and creating connections among the to-be-learned pieces of information (Duncan et al., 2015). Organisation can take many forms, such as creating outlines of topics, maps, or flowcharts (Duncan et al., 2015, Ormrod, 2016). By organising information, learners are engaging in active learning, creating relational structures in memory that help guide retrieval over the long term (Masson & McDaniel, 1981). In alignment with these ideas, there is consistent evidence that organisation benefits long-term retention (e.g., Blunt & Karpicke, 2014, Einstein et al., 1990, for memory of text content).

Spaced study involves separating learning across multiple sessions with time between each session (Rodriguez et al., 2021). Spacing has been shown to be more effective than cramming for long-term learning in both controlled laboratory experiments (Cepeda et al., 2008, Walsh et al., 2023) and experiments conducted in authentic educational settings (Dobson et al., 2017, Küpper-Tetzel et al., 2014), though correlational studies using students’ self-reports have failed to find a significant association between spaced study and academic achievement (Hartwig & Dunlosky, 2012, Rodriguez et al., 2021).

We emphasize that the empirical effectiveness of a given learning strategy is not necessarily associated with the extent to which students use the strategy (e.g., Blasiman et al., 2017, Ekuni et al., 2022, Karpicke et al., 2009). Further complicating matters is that strategies can differ in effectiveness depending on how they are implemented (e.g., flashcards, Lin et al., 2018; rereading, Miyatsu et al., 2018). Thus, we do not necessarily expect that the dynamics of strategy self-regulation will be neatly determined by strategy effectiveness.

Learning strategies and achievement

As just discussed, the learning strategies we investigated in this research include a mix of strategies that differ in their levels of effectiveness. Of note, Hartwig and Dunlosky (2012) posited that the use of learning strategies “will mater most if they are related to student achievement” (p. 127) and cautioned that “a relationship between strategy use and achievement level is not guaranteed” (p. 127; Hartwig & Dunlosky provide more extensive discussions). Consistent with this viewpoint, the results reported by these authors painted a mixed picture regarding the associations between learning strategy use and achievement. For example, they found a statistically significant association between GPA and testing oneself but a nil association between GPA and employing flashcards, despite the fact that using flashcards is regarded in the learning literature as an instantiation of self-testing (however, the type of flashcards in conjunction with the nature of the summative test and the learner’s ability can moderate effectiveness of flashcards; Lin et al., 2018). In addition, Hartwig and Dunlosky (2012) found a significant association between rereading and GPA, whereas the associations involving spacing and cramming with GPA were both not statistically significant. To shed fresh light on these aspects, the present research investigates the extent to which individual differences in (self-reported) use of learning strategies are significant predictors of (self-reported) achievement. In addition, the research also investigates whether self-regulation predictors of learning strategy use make meaningful contributions to accounting for individual differences in achievement.

Research questions

The objectives of this study were to uncover the extent to which students use a set of commonly researched learning strategies and explore whether learning strategy usage could be predicted from information provided by a set of self-regulation factors that are posited to support learning. In addition, the research examined the extent to which individual differences in (self-reported) achievement could be predicted from information on learning strategies and self-regulation factors. To achieve these objectives, the study examined the following research questions (RQs):

• RQ1. To what extent do undergraduate students use relatively more effective learning strategies?

• RQ2. To what extent do undergraduate students use relatively less effective learning strategies?

• RQ3. Is students’ use of learning strategies characterized by adopting a constellation of strategies?

• RQ4. Is students’ use of learning strategies related to key self-regulation factors that support learning? These factors include eager and vigilant strategies, self-control, attention control, MCSR, effort regulation, and student engagement.

• RQ5. Do individual differences in learning strategy usage and self-regulation factors predict significant differences in students’ self-reported achievement at university?

Method

Participants and recruitment

Approval for conducting this research was received by the Human Ethics Committee at the institution with which the first author has been affiliated (approval number: 0000030139). Participants in this study were undergraduate students, who were at least 18 years old and studying at a New Zealand (NZ) university. They were recruited via three channels. The first was through promotion of the research to undergraduate courses at the university with which the first author has been affiliated. The second method involved posters with QR codes to the survey, which were displayed in the campus of this university. The final method of recruitment was via social media, through the first author’s Instagram and Facebook accounts.

Overall, 249 participants accessed the link to the survey (described in the next section). Of these, 12 were not eligible to participate because were either under 18 years old (1 participant) or were not an undergraduate student (11 participants). 140 students provided answers to each survey question (i.e., had complete data). Around half of the participants with missing data (i.e., 57) had exited the survey or stopped responding at the very beginning of the survey (i.e., the 8th question) and three quarters (83 participants) stopped responding by the 16th question. In total, 49% of participants were between 18–22 years old, 10% were between 23–30 years old, 3% were older than 30, and 38% did not report their age. Furthermore, 71% of participants were female, 21% were male, 6% selected the “Other” option, and 2% did not report their gender. Finally, 26% of participants were in their first year of study, 28% were in their second year, 33% were in their third year, and 13% of students were in their fourth year or higher. No information on ethnicity or socio-economic status was collected.

Data collection and survey

The data collection for this study occurred online, by means of a Qualtrics survey. The survey was distributed to students through a QR code or a link (depending on how they heard about the research). Items used to measure rehearsal, elaboration, organisation, effort regulation, MCSR, and one item for self-testing were taken from the motivated strategies for learning questionnaire (MSLQ) (Duncan et al., 2015). The MSLQ has been shown to accurately measure intended constructs and is considered a valuable tool for research (Credé & Phillips, 2011). To measure self-control, items were adapted from a validated four-item scale developed by Partsch and Danner (2021). The ability to direct attention was measured using the directed attention items from a questionnaire developed by Vandergrift et al. (2006). Engagement was measured with three items adapted from the Utrecht Work Engagement Scale (Schaufeli et al., 2019). This scale has been found to provide a reliable and valid measure of work engagement. The six items used to measure vigilant (three items) and eager strategies (three items) were adapted from the Regulatory Goals and Strategies Questionnaire (RGSQ). This questionnaire has been shown to have good validity (Vriend et al., 2022).

The single items for spacing of study, cramming study, and rereading were adapted from a survey previously conducted by Hartwig and Dunlosky (2012). Finally, the items for self-testing were adapted from surveys previously conducted by Yan et al. (2014) and Kornell and Bjork (2008). All adaptations to the various scales and questionnaires mentioned above were undertaken to make items relevant to the university setting.

For all items (except the demographic questions and the self-reported achievement item) participants rated the extent to which a statement applied to them using a Likert scale from 1–7, with 1 being not at all like them, and 7 being very much like them. The item for self-reported achievement asked participants to rate how well they think they are doing in their studies on a 0 – 100 scale, with 0 being “not at all well” to 100 being “extremely well”. Each participant was then assigned a score for each construct by averaging their answers to the items that measured the given construct. More information on sample items from the various scales and on the constructs’ reliability is provided in the Appendix.

Data analysis

Data analysis was conducted in IBM-SPSS 26. To answer RQ1 and RQ2, the mean score and standard deviation were calculated for each learning strategy, as well as for each self-regulation factor. To determine the size of the difference between the largest and smallest means, Cohen’s d was used. When interpreting Cohen’s d, values of 0.2 are considered small, values of 0.5 are considered medium, and values of 0.8 or above are considered large (Hedrih & Hedrih, 2022). To address RQ3, RQ4, and RQ5, we computed simple correlations and conducted multiple regression analyses. To avoid multicollinearity in the regression analyses, any independent variables that were strongly correlated to other predictors were not included in regression models; specific information on this aspect is included below.

Results

Table 1, which summarises the descriptive statistics for the use of learning strategies, provides the information needed to answer RQ1 and RQ2. The learning strategy with the lowest endorsement was self-testing (M = 3.27), followed by rereading (M = 3.39). Rehearsal (M = 3.80) had a mean score somewhat higher than rereading. Organisation (M = 4.59), spacing (M = 4.52), and cramming (M = 4.36) were endorsed more strongly than self-testing and rehearsal, but less strongly than elaboration. Elaboration had the highest mean score (M = 5.10).

Table 1 Descriptive statistics for study strategies

The Cohen’s d for the difference between the means for elaboration and self-testing was 1.03. Furthermore, the Cohen’s d value for the difference between the mean scores for elaboration and rereading was 0.96. These Cohen d values suggest there was a large difference between the extent to which students reported using elaboration (which is the most strongly endorsed learning strategy) versus self-testing and rereading (which are the strategies that were the least endorsed by students in this sample).

Table 2, which summarises the correlation coefficients between different learning strategies and the correlations between learning strategies and academic achievement, includes information relevant to answering RQ3 and initial information regarding RQ5. Unless specifically stated, all correlations reported were statistically significant at the 0.05 level. The largest correlation reported in Table 2 was between two relatively effective learning strategies, namely organisation and elaboration (r = 0.64). This strong¹ positive correlation suggests that higher levels of using organisation are associated with higher levels of using elaboration. Notably, organisation also had a strong positive correlation with rehearsal, which is a relatively less effective learning strategy (r = 0.55). In addition, the endorsement of rehearsal as a study strategy was also moderately correlated with the endorsement of both more effective (i.e., elaboration: r = 0.48; self-testing: r = 0.43) and less effective (i.e., rereading: r = 0.42) learning strategies. A significant strong negative relationship was found between cramming and spacing (r = -0.54). This indicates that students who used cramming a lot employed spacing to a lesser degree (and vice versa). Notably, cramming was negatively correlated with all the other learning strategies investigated in this research. These correlations were generally small (see Table 2).

Table 2 Correlation coefficients between study strategies and academic achievement

Rereading had weak correlations with nearly all the other learning strategies including, elaboration (r = 0.30), organisation (r = 0.25), spacing (r = 0.23), cramming (r = -0.20), and self-testing (r = 0.13, not statistically significant). These results suggest that the extent to which students reread has only a small overlap with the degree to which they use other learning strategies. Overall, the findings in Table 2 indicate that many students who use organisation are likely to employ elaboration and vice-versa. Nonetheless, usage of these relatively more effective learning strategies was also positively related to employment of less effective strategies, such as rehearsal and rereading.

(Self-reported) academic achievement had moderate correlations with cramming (r = -0.34) and elaboration (r = 0.31), as well as weak (yet significant) correlations with spacing (r = 23), rehearsal (r = 0.21), and self-testing (r = 0.21). Achievement was not significantly correlated with rereading (r = 0.13) and organisation (r = 0.12). These results indicate that there is no meaningful relationship between rereading and academic achievement, and organisation and academic achievement. As a result, rereading and organisation were not included in the regression model for predicting achievement.

Table 3 summarises the descriptive statistics obtained for the key self-regulation factors that underpin learning. The factor that was endorsed the least by students was attention control (M = 4.36). In contrast, the self-regulation factor with the highest mean was self-control (M = 4.73). Most of the mean scores for the self-regulation factors were similar. For example, the Cohen’s d for the difference between the means for self-control and attention control was 0.41. This suggests that there is a small-to-medium difference between the extent to which students exert self-control (highest mean score) and direct their attention (lowest mean score).

Table 3 Descriptive statistics of self-regulation factors underpinning learning

Table 4 summarises the correlation coefficients between self-regulation factors that underpin learning, as well as their associations with academic achievement (setting the stage for regression analyses directed at RQ5). The largest reported correlation was between attention control and effort regulation (r = 0.75). In addition, attention control was also strongly (and positively) correlated with MCSR (r = 0.69). The ability to direct attention also had a pattern of strong positive relationships with engagement (r = 0.59) and the use of eager strategies (r = 0.59).

Table 4 Correlation coefficients for self-regulation factors underpinning learning and academic achievement

Academic achievement had a moderate positive relationship with effort regulation (r = 0.47), student engagement (r = 0.43), MCSR (r = 0.40), and attention control (r = 0.38). In contrast, weak positive relationships were found between academic achievement and self-control (r = 0.29), the use of eager strategies (r = 0.28), and the use of vigilant strategies (r = 0.12, not statistically significant). Given that the use of vigilant strategies was not significantly related to achievement, this variable was excluded from the subsequent regression models. Notably, attention control and the use of eager strategies had moderate or strong correlations with other self-regulation factors (e.g., student engagement; effort regulation; see Table 4). Therefore, to avoid multicollinearity, attention control and eager strategies were not included in the regression models predicting learning strategies and achievement.

To help answer RQ4 and set the stage for RQ5, Table 5 summarises the standardised regression coefficients (Beta) and the ${R}^{2}$ values for regressing learning strategy use and achievement (dependent variables) on self-control, student engagement, effort regulation and MCSR (predictor variables). These results show that 41% of the variability in elaboration scores were accounted for by the self-regulation predictor variables. In addition, 32% of the variability in rehearsal, and 31% of the variability in spacing scores were accounted for by the predictor variables. The ${R}^{2}$ values for the regression models predicting cramming, achievement, and organisation were similar, with 29% of the variability in cramming, 27% of the variability in achievement, and 26% of the variability in organisation, being accounted for by this set of predictor variables. Furthermore, 22% of the variability in self-testing scores were accounted for by the predictor variables. Lastly, only 12% of the variability in rereading scores were accounted for by these predictors.

Table 5 Standardised regression coefficients and r-squared values for predicting learning strategy usage on four factors underpinning learning

In the regression model predicting rehearsal, two self-regulation factors were statistically significant predictors– student engagement (Beta = 0.23) and effort regulation (Beta = 0.22). In the model predicting elaboration, student engagement was a very strong predictor (Beta = 0.42); MCSR (Beta = 0.23) was the only other statistically significant predictor of this dependent variable. The regression model predicting organisation only had one statistically significant predictor, which was effort regulation (Beta = 0.26). Similarly, spacing and cramming were both significantly predicted only by effort regulation. Self-testing was statistically significantly predicted only by student engagement (Beta = 0.30). The four self-regulation factors predicted a much smaller proportion of variability in rereading than in all of the other learning strategies that were examined. For this model, none of the predictors was statistically significant. Finally, achievement had two significant predictors – student engagement (Beta = 0.22), and effort regulation (Beta = 0.26); these findings address one aspect of RQ5, namely the relationship between achievement and self-regulation factors.

Self-control was not a significant predictor of any of the learning strategies, nor of self-reported academic achievement. Similarly, MCSR was a significant predictor of only one learning strategy (i.e., elaboration). Furthermore, MCSR did not significantly predict self-reported achievement. In contrast, student engagement was a significant predictor of three learning strategies (i.e., rehearsal, elaboration, and self-testing), as well as self-reported academic achievement. Similarly, effort regulation was also a significant predictor of four learning strategies (i.e., rehearsal, organisation, spacing, and cramming), and self-reported academic achievement.

To address the remaining aspect of RQ5, Table 6 summarises the standardised regression coefficients and the ${R}^{2}$ value for regressing self-reported academic achievement (dependent variable) on the following learning strategies: rehearsal, elaboration, spacing, cramming, and self-testing (predictor variables). The ${R}^{2}$ value for this regression model showed that 18% of the variability in achievement was accounted for by this set of learning strategies. In this model, two learning strategies were statistically significant predictors of academic achievement – elaboration (Beta = 0.24), and cramming (Beta = -0.30). These results suggest that (all else being the same) for every one standard deviation increase in elaboration, there is a 0.24 standard deviation increase in achievement. Furthermore, for every one standard deviation increase in cramming, there is a 0.30 standard deviation decrease in academic achievement. Taking into account the results of the two regression analyses conducted for achievement (see the last column in Table 5 and Table 6), a final regression model was estimated. This model included only the significant predictors of achievement in the previous models.

Table 6 Standardised regression coefficients and r-squared values for regressing achievement on study strategies

Table 7 summarises the standardized regression coefficients and the ${R}^{2}$ value for regressing self-reported academic achievement on two learning strategies (i.e., elaboration and cramming) and two self-regulation factors (i.e., effort regulation, and student engagement). The ${R}^{2}$ value for this regression model showed that 27% of the variability in achievement was accounted for by this set of predictors. In this model, two variables were statistically significantly predictors of academic achievement – effort regulation (Beta = 0.28) and student engagement (Beta = 0.21).

Table 7 Standardised regression coefficients and r-squared values for regressing achievement on elaboration, cramming, effort regulation, and student engagement

Discussion

The aim of this study was to gain insight into the learning strategies employed by undergraduate students and to examine the extent to which the use of specific learning strategies can be predicted from knowledge of key self-regulation factors. Given that a study of this nature has not yet been conducted, the results of this research help bridge important gaps in existing knowledge. The following sections address the research questions we proposed and map the implications of the findings.

Student use of relatively more effective learning strategies

Regarding the first research question, findings indicate that undergraduate students who participated in this research reported employing elaboration, organisation, and spacing to a greater extent that other learning strategies. These three learning strategies are considered to be relatively effective because they could help optimise long-term retention of knowledge and strengthen academic achievement (Dunlosky et al., 2013; Ormrod, 2016; Rodriguez et al., 2021).

Although in this study students reported using more effective learning strategies to a greater extent than less effective ones, the differences between endorsing specific (more versus less effective) strategies were not always large. For example, the difference between endorsing spacing and cramming was small (we consider this result more extensively in the next section). Elaboration was the most commonly reported learning strategy in this research followed by organisation. These results are consistent with a number of previously reported findings. A qualitative investigation of Spanish undergraduate students found that “participants frequently reported the use of basic elaboration and organisation strategies” (Garcia-Perez et al., 2021, p. 544). Elaboration and organisation were also the most strongly endorsed learning strategies by second-year education students enrolled in a university course in Turkey (Kucuk, 2018). In addition, as in the present research, rehearsal had a lower average level than organisation and elaboration (Kucuk, 2018). Our finding that elaboration was the most endorsed learning strategy is also in line with results reported by Pizzimenti and Axelson (2015) for a sample of US students enrolled in a first-year anatomy course.

Notwithstanding the consistencies just highlighted, several previous studies have reported results that diverge from those above. For instance, students reported creating examples (a form of elaboration) to a lesser extent than engaging with other learning strategies such as rereading (Blasiman et al., 2017; Karpicke et al., 2009). In addition, in contrast to the finding that organisation is a commonly used learning strategy (present results; Garcia-Perez et al., 2021; Kucuk, 2018), others reported that students seldom use tools associated with organisation such as, outlines, diagrams, or charts (Hartwig & Dunlosky, 2012; Karpicke et al., 2009). Moreover, unlike in the present research (where organisation was endorsed to a higher degree than rehearsal), Pizzimenti and Axelson (2015) found that rehearsal and organisation had similar mean levels of endorsement.

There are likely a range of reasons, not necessarily mutually exclusive, for why participants in this and other research, but not all studies, frequently reported using elaboration and organisation. In at least one case, the instructor in the course (from which the participants were recruited) directly “encouraged the students to organise their learning through ill-structured problem-based activities” (Kucuk, 2018, p. 266). Similarly, in a course instructing students about learning strategies, by the end of the course students reported using outlining (an organisation strategy) “often”, in contrast to being “rarely” used at the outset of the course (McDaniel & Einstein, 2023). However, direct strategy support from instructors is not experienced by a majority of students (e.g., 20% in Kornell & Bjork, 2007; 36% in Hartwig & Dunlosky, 2012). With regard to the present study, one plausible explanation for the high use of organisation may be because at the university where most of the data collection efforts were concentrated, it is common for courses to clearly state their learning objectives and provide specific links between learning objectives and assessment items. This, in turn, could likely alert students that working to organize the to-be-learned information in a coherent manner might help them identify how different parts of the learning materials support doing well in assignment tasks.

Regarding elaboration, it is possible that the benefits of acquiring knowledge by linking what one studies to what she/he already knows (rather than by rereading or rote memorization) have started to filter to the public, including (some) students in this research. Consistent with this standpoint, Rea et al. (2022) reported the US undergraduate students they surveyed understood the benefits of an elaboration-based learning strategy (i.e., generating versus listening to explanations). More generally, these authors argued that currently students might be better informed than students in the past about the effectiveness of learning strategies (e.g., from instructors, a wealth of online sources, and/or social media). Another possibility relates to our finding that student engagement was a very strong predictor of elaboration-based learning. Students who are highly engaged in their courses may elaborate as a natural consequence of striving for understanding and relating the content to prior knowledge and experience. To the extent that NZ undergraduates tend to be more engaged in their university education than other student populations (e.g., a survey at several state universities in the US showed that over 2/3 of the students did not purchase the textbook for introductory courses, and many reported not reading any of the textbook for at least one course; Sikorski et al., 2002), then their greater use of elaboration is sensible.

We suggest that variations across studies in students’ endorsement of an elaboration strategy (and potentially others, such as organisation) can be traced, at least in part, to differences across studies in the items/instruments that measure strategy use. With regard to elaboration, very different techniques that are captured under the umbrella of “elaboration” are used in different studies. For instance, in the present study, one item for elaboration was, “when doing the readings for my courses, I try to relate the material to what I already know” (for another example of an elaboration item, see the Appendix). By contrast, in studies in which elaboration is not a prominently reported strategy, the elaboration item is very different: “’creating examples” (Blasiman et al., 2017; Karpicke et al., 2009) or “making diagrams, charts, or pictures” (Hartwig & Dunlosky, 2012). Of note, in instruments prompting ratings for several different elaborative techniques, it is apparent that frequency of endorsement (from the same students) varies across these elaboration techniques, much as it does across studies (e.g., in McDaniel & Einstein, 2023, students reported using the elaboration technique “drawing” never to rarely, whereas they reported using the elaboration technique “self-explanation” often to very often). Clearly, more work is needed to fully understand the factors that underlie variable patterns of reported strategy use that have been documented.

In our study, self-testing was endorsed the least by students, despite being a relatively more effective study strategy (Dunlosky et al., 2013; Rodriguez et al., 2021). These results are consistent with findings from existing research, which reported that few students use self-testing (Blasiman et al., 2017; Karpicke et al., 2009). However, in a recent study by Ekuni et al. (2022) retrieval practice (self-testing) was used somewhat more frequently by precollege students from Brazil; this strategy had the fifth mean score on frequency of use among the 10 learning strategies investigated by these authors. Ekuni and colleagues hypothesised that this somewhat higher level of self-testing may be linked to the fact that many students in their sample completed biology degrees. This explanation is consistent with findings from Rodriguez et al. (2018) who reported that students completing degrees in science, technology, engineering and mathematics (STEM) commonly use self-testing in their study routines. Unfortunately, as our research did not collect data on students’ majors, we do not know whether our sample included a relatively small proportion of students from majors where self-testing may be more commonly used.

Student use of relatively less effective learning strategies and of learning strategies with mixed/modest effectiveness

Relevant to the second research question, participants in this research reported using rereading, rehearsal, and cramming to a non-negligible yet somewhat lesser extent than many of the relatively more effective learning strategies. These results suggest that students include in their study routines learning strategies that are not strong facilitators of meaningful learning (Blasiman et al., 2017; Dunlosky et al., 2013; Ekuni et al., 2022; Karpicke et al., 2009; Oberauer, 2019).

In our study, rereading was endorsed only somewhat more strongly than self-testing (which was the least endorsed learning strategy), suggesting that students were not heavily relying on rereading and instead were using other learning strategies. This pattern contrasts with previous studies, which have found rereading to be a commonly used learning strategy (Ekuni et al., 2022; Hartwig & Dunlosky, 2012; Karpicke et al., 2009). Although the results on rereading diverge from those of previous studies conducted in different student populations, they are consistent with other results we uncovered, namely that elaboration and organisation were (by some distance) the learning strategies most frequently endorsed by NZ students. Specifically, if organization is perceived by students as being a valuable learning strategy in some courses (see our previous discussion), this may constrain the extent to which students find it useful to use rereading in those courses (i.e., because, compared to organisation, students may not think that rereading helps them identify and interrelate aspects that are critical to achieving well in the course). Similarly, the frequent use of elaboration may also reduce the appeal of (or the need for) rereading for some students. Of course, these hypotheses need to be empirically tested.

Notably, cramming and spacing, which are two contrasting strategies for scheduling study, were endorsed to similar extents by NZ undergraduate students and shared a negative association. This result is consistent with findings reported by Hartwig and Dunlosky (2012), who found that spacing and cramming were endorsed to similar extents. These authors suggested that both cramming and spacing might be useful to students, within different contexts. Specifically, when a student has limited time, cramming right before an exam can facilitate satisfactory exam performance on that exam (even though it leads to poor long-term retention of information; Hartwig & Dunlosky, 2012). In contrast, when the student has more time and the material seems difficult, then they may implement spacing. However, the present negative correlation between the rated use of spacing and cramming suggests another possibility: Individual students may tend to generally rely on spacing or on cramming (not both). For instance, cramming is commonly used by students who are prone to procrastination (Kornell, 2009). Thus, it might be that students who avoid procrastinating and manage their time well may use spacing, while students who procrastinate and have poor time management may engage in cramming (for a recent review of procrastination interventions and discussions on how procrastination could be reduced in learning settings, see Turner & Hodis, 2023). That is, although on average spacing and cramming were endorsed about equally often, there might be individual differences at play here, with some students endorsing more spacing and less cramming and the reverse being true of other students. This conjecture is also supported by our finding that effort regulation was a significant positive predictor of spacing and a negative predictor of cramming. Taking into consideration these aspects, it is possible that there exist distinct (unobserved) clusters (groups) of students who (i) procrastinate little, exert effective effort regulation during school work, and generally space their studies, others who (ii) heavily procrastinate, struggle with effort regulation during studying, and mainly cram their learning; and others who (iii) exhibit some combinations of patterns from (i) and (ii) above (e.g., little procrastination, poor effort regulation, and non-negligible cramming).

Constellations of learning strategies

This research found that (reported) use of some learning strategies correlated at strong/medium levels (e.g., elaboration, organisation, and rehearsal). At the same time, usage of other learning strategies had little or non-significant overlap (e.g., rereading, self-testing, and spacing). These results plausibly suggest that students might use a mixed constellation of relatively effective and ineffective learning strategies; for example, it is possible that some students rely more heavily on organisation, elaboration, and rehearsal and use less frequently self-testing. If this hypothesis is tenable, it may disfavour the notion that as students adopt relatively more effective learning strategies, they let go of relatively ineffective strategies they have in their toolkit, like rehearsal (e.g., McDaniel & Einstein, 2023). At the same time, as we noted above, it is possible that these constellations reflect more (vs. less) effective patterns of learning strategy usage (e.g., more spacing and less cramming in one constellation and mainly cramming in another). These are intriguing possibilities that could be profitably examined in future research by using mixture modeling analytic strategies (e.g., latent profile analyses; Harring & Hodis, 2016).

Relationships between key self-regulation factors and learning strategies

Relevant to the fourth research question, effort regulation was a significant predictor of the use of four learning strategies: rehearsal, organisation, spacing, and cramming (negatively). Specifically, stronger effort regulation predicted higher levels of engaging in rehearsal and organisation. This suggests that being able to regulate one’s effort is a key factor underpinning the use of these learning strategies. In addition, effort regulation negatively predicted the use of cramming and positively predicted spacing. Research indicates that students often view cramming as easier and faster than spacing (Kornell, 2009). This, in conjunction with our findings, suggests that supporting and strengthening students’ ability to effectively regulate effort in learning settings could be a potentially productive way to ensure that students limit the extent to which they engage in cramming and, instead, space their learning. Hence, support for effort regulation may be a cost-effective strategy to promote effective and durable student learning.

Intriguingly, we found that effort regulation was not a significant predictor of elaboration, rereading, and self-testing; one of these strategies (i.e., rereading) is generally considered less effective, whereas the other two are thought to be more effective (Dunlosky et al., 2013). Thus, our findings suggest that the effects of effort regulation on use of learning strategies are not moderated by the extent to which a strategy is effective (vs. ineffective). Hence, the role of effort regulation in learning strategy usage might be most appropriately investigated/modelled at the strategy level. This hypothesis receives some support from recent research findings. Specifically, results reported by Karabenick et al. (2021) indicate that the associations between key motivation constructs influencing students’ choice of learning strategies (i.e., how useful students perceive a learning strategy and how ‘costly’ they think is to use it) and the extent to which they actually employ the given strategy were not consistent across the learning strategies investigated. Taken together, our findings and these extant results highlight an important unanswered question: Are the effects of motivation and self-regulation factors on choice/employment of learning strategies best described by strategy-specific models?

Effort regulation had moderate positive correlations with several self-regulation factors including metacognitive self-regulation (MCSR), student engagement, the use of eager strategies, and self-control. The positive association between effort regulation and MCSR is consistent with current theorizing suggesting that perceptions of effort are metacognitive judgments (de Bruin et al., 2023) and with research indicating that students need to use their metacognition to monitor how well their learning is going, and thus make decisions about how much effort to exert (Virtanen et al., 2015). In all, findings from our study provide a strong indication that effort regulation is a key self-regulation factor that supports learning. This conclusion is consistent with results reported by Rea et al., (2022; Study 2), who found that (for US undergraduate students) the second most commonly reported barrier to use effective learning strategies was that “these strategies required too much effort” (p. 16).

In the same vein, recent conceptual work indicates that strengthening effort regulation is critical in ensuring self-regulated use of effective learning strategies (de Bruin et al., 2023). de Bruin and colleagues (2023) proposed that enhancing effort regulation vis-à-vis the use of effective learning strategies could involve helping students understand that devoting more time to using these strategies leads to stronger long-term learning and better transfer. A second family of strategies proposed by these authors centers on increasing students’ sense of fluency during learning that involves effective learning strategies. To this end, these researchers highlighted several approaches, such as (i) supporting students to frequently use effective learning strategies and providing feedback on how the effort they devoted to learning decreased across time while their learning strengthened; and (ii) advising students to “segment their study sessions, study tasks, or learning goals into smaller pieces” (de Bruin et al., 2023, p. 16).

Efklides and colleagues (2006, 2018) proposed that students develop two types of beliefs about effort. One is that devoting effort to learning is likely to lead to both success and enhanced competence through supporting perseverant engagement with learning tasks and activities. The other is that exerting considerable effort to learn is difficult and indicates that one does not have the abilities needed to do well in the given learning domain, task, or activity. Holding these different types of beliefs leads to distinct and consequential interpretations of the roles of effort and the desirability of expending effort during learning (Efklides et al., 2018). Specifically, the former indicates that effort is a valuable resource that a student could use to enhance the likelihood of her/his success in learning settings. On the contrary, the latter set of beliefs suggests that expenditure of effort is potentially costly because “the possibility of success is low” (Efklides et al., 2018, p. 76). Results in the present study suggest that promoting/strengthening students’ belief that strategic effort expended during learning leads to success is likely to have beneficial consequences regarding students’ use of learning strategies (i.e., reduced cramming and increased spacing and organization).

An intriguing possibility about the role of effort regulation in predicting learning strategy use emerges when considering (a) the analysis of Schwartz and Efklides (2012) and (b) that key decisions impacting the quality of learning (e.g., exertion of effort; choice of learning strategies) are influenced by students’ metacognitive judgments and experiences (Efklides, 2019). That is, the accuracy of students’ judgments of learning (JOLs) may moderate the extent to which the ability to regulate effort influences critical decisions affecting learning (e.g., the decision to engage in studying and the subsequent choice of learning strategies). Specifically, when students overestimate their learning (i.e., they have inflated JOLs), they may “not see the advantages of further study” and, thus, “may choose not to study even though continued study will improve performance” (Schwartz & Efklides, 2012, p. 148). Therefore, in the context of inflated JOLs (and poor overall calibration), students are likely to withhold effort directed toward further studying, regardless of their (perceived) ability to regulate their effort and knowledge (of the usefulness/efficiency) of specific learning strategies. In contrast, when students’ JOLs are accurate (well calibrated) and students recognize the need for further studying, ability to regulate effort is likely to play a significant role in whether they devote additional time and effort to learning activities.

Student engagement predicted the use of three learning strategies (i.e., rehearsal, elaboration, and self-testing). Specifically, the more students were enthusiastic about their studies and immersed in them, the more likely they were to report using rehearsal, elaboration and self-testing within their study routines. Elaboration encourages students to think about the meaning of what they learn and stimulates deep level thinking (Weinstein & Sumeracki, 2019). In addition, self-testing requires that students bring to mind information that was previously learned. Although usage of both of these strategies is likely to help improve learning, they require strong engagement with one’s studies, in general, and with the learning task, in particular (Dunlosky et al., 2013; Rodriguez et al., 2021). Findings from our study support this hypothesis. Importantly, in our model, student engagement was the only significant predictor of self-testing. Given that despite its general effectiveness self-testing is under-utilized by students (Blasiman et al., 2017; Dunlosky et al., 2013; Karpicke et al., 2009), our results suggest that efforts to strengthen students’ use of self-testing may benefit from adopting a dual focus. A first focus would be to increase students’ knowledge of self-testing and enhance their confidence that they could use this strategy to their own benefit (see, McDaniel & Einstein, 2020, 2023 and McDaniel et al., 2021 for a theoretical framework and empirical findings supporting this approach). However, that may not be sufficient because though students may have knowledge of self-testing and may intend to use self-testing, students’ perceived cost (time and effort) of using self-testing remains a barrier to actually using it (Wang et al., 2023). In conjunction with those findings, our results suggest that strong student engagement might either lessen the perceived cost of self-testing or reduce the negative impact of perceived cost on using self-testing. Accordingly, an important second focus would be to enhance students’ engagement with learning in the broader university context to facilitate students’ use of self-testing.

The finding that student engagement and effort regulation positively and significantly predicted the use of rehearsal is interesting. Rehearsal is considered to be useful for simple tasks and is not thought to help students integrate new information with prior knowledge (Duncan et al., 2015). Thus, rehearsal does not appear to require strong engagement or accomplished effort regulation. However, results from our research indicate that high levels on both of these self-regulation factors are likely to be associated with more (rather than less) use of rehearsal in one’s learning. This might be because repeating information over and over (rehearsing) may require stronger effort regulation than using other relatively ineffective learning strategies (e.g., rereading). In addition, if students are successfully using rehearsal for memorization, it seems plausible that it takes strong engagement and effort regulation to do sufficient rehearsal for good memorization. Perhaps related to this conjecture, Sebesta and Speth (2023) reported that students who improved their performances from the first exam to the second exam in a biology course reported increased use of rehearsal/memorization (among other strategies).

Associations of learning strategies and self-regulation factors with academic achievement

In terms of the fifth research question, this study found that effort regulation had a significant moderate positive association with academic achievement. Previous literature has found inconsistent results with regard to the link between effort regulation and academic achievement (Richardson et al., 2012; Virtanen et al., 2015). The results from our study are in line with those reported by Richardson et al. (2012), who also found a moderate positive correlation between effort regulation and academic achievement. Of note, in the final regression model predicting achievement, effort regulation was a significant predictor of academic achievement, over and above learning strategies and student engagement. These results help reinforce the importance (and benefits) of regulating one’s effort in academic settings.

Student engagement also had a significant moderate positive association with academic achievement. This result is consistent with meta-analytic findings of a strong positive correlation between these variables (Lei et al., 2018). Furthermore, results from the regression model predicting achievement found student engagement was a significant predictor of academic achievement even after controlling for effort regulation, elaboration, and cramming. Lei et al. (2018) reported that the method of measuring student engagement moderated the relationship between engagement and achievement. Specifically, the strongest effects were found for self-reported measures, which were also used in this research.

This study also helped shed light on the relationship between the use of spacing and academic achievement. In this sample, spacing had a weak association with academic achievement. Findings from the previous literature on spacing have been mixed, with some studies suggesting spacing is an effective learning strategy vis-à-vis achievement, while other studies have failed to find a significant association between spacing and academic achievement (Dunlosky et al., 2013; Hartwig & Dunlosky, 2012; Rodriguez et al., 2021). Relevant to this issue, Rodriguez et al. (2021) suggested that using a survey to measure spacing may not accurately capture students’ actual use of this strategy. This, in turn, might explain both inconsistent prior findings regarding spacing and the relatively weak association between the use of spacing and academic achievement we uncovered in this research. Another possibility is that if students’ grades are primarily based on unit assessments (e.g., tests drawing from only a subset of the material studied during a semester), and not on cumulative assessments, as pointed out earlier, spacing may not be required for doing well in the course (i.e., cramming could be sufficient). However, we do not know to what extent this represented the situation of students who participated in this research.

The present study uncovered statistically significant associations (in the expected direction) between students’ self-reported achievement and five of the seven study strategies we investigated (see Table 2). These findings reinforce our a priori expectation that self-regulation of several strategies that we examined would be related to achievement. For example, our finding of a positive correlation between self-testing and achievement (r = 0.21) converges with the corresponding Biwer et al. (2023) finding (r = 0.34) who also investigated university students and used exam performance as an indicator of student achievement.

Miyatsu et al. (2018) proposed a novel approach to support the effectiveness of student learning, which, in turn, could have implications for the relationships between learning strategy use and achievement. In a departure from typical approaches, which (intervene to) encourage students to use learning strategies with proven effectiveness (see, for example McDaniel & Einstein, 2020), Miyatsu and colleagues suggested that it might be more productive to encourage students to augment the less optimal study strategies they were already using. For example, the effectiveness of rereading could be enhanced by strengthening students’ ability to accurately monitor their comprehension of the to-be-learned material (Miyatsu et al., 2018). This strategy is likely to be effective because “increased metacognitive accuracy can guide more effective and focused rereading” (Miyatsu et al., 2018, p. 393). Importantly, commonly used learning strategies could be employed relatively effectively or relatively ineffectively by students. Accordingly, depending on how individuals implement these strategies (e.g., rereading, self-testing, elaboration), both positive associations (when the implementation is effective) and nil/negative associations (when the implementation is ineffective) with long-term learning/achievement are possible.

Limitations and future directions of research

One potential limitation is that academic achievement was assessed with self-reports. To the extent that self-reports are inaccurate the validity of the findings pertaining to achievement could be compromised. Mitigating this concern, self-reported achievement has been found to be a good proxy for actual achievement. For example, in one study, across grades and school subjects, self-reported grades had high positive correlations with actual grades (correlations ranging from 0.76 – 0.93; Sticca et al., 2017); in another study with more than 15,000 high-school students, the self-reported grade was a reasonably accurate measure of the actual (transcript) grade (Sanchez & Buddin, 2016). Similarly, a meta-analysis found a correlation of 0.90 between self-reported grades and college GPA obtained from school reports (Kuncel et al., 2005). Thus, student self-reports of school achievement appear to be relatively accurate, at least in terms of reporting grades.

Another limitation of this study is the relatively low sample size that resulted because of a low completion rate. This low completion rate is likely due to the lack of incentive for students to fill out the survey. Thus, it could be hypothesized that many unmotivated students who initially volunteered to participate did not complete the survey. If this hypothesis is tenable, then findings from the present research might provide a better representation of the aspects under investigation for motivated and engaged students than for their less motivated and less engaged counterparts. This conjecture would need to be evaluated in future research. (We thank an anonymous reviewer for suggesting this aspect.) An additional limitation of this research is the fact that the participants’ recruitment focused on one NZ university. To address these limitations, future research should aim to increase the sample size, and recruit students from multiple universities. This will help strengthen the generalisability of the results (Johnson & Christensen, 2020). A fourth limitation of our study is that it only had access to correlational data, which precludes making causal inferences about the aspects we investigated.

A fifth limitation of this research is that several constructs investigated had less than optimal reliabilities, despite being measured with instruments that performed well in prior studies (i.e., metacognitive self-regulation, ability to direct attention, and self-control). This limitation may, at least in part, explain why neither metacognitive self-regulation nor self-control significantly predicted achievement in the model that included other self-regulation factors. Importantly, extant research findings about metacognitive self-regulation are mixed, with some studies finding that this factor did not predict academic achievement (e.g., Kitsantas et al., 2008), whereas others uncovered a positive relationship between it and exam performance (e.g., Vrugt & Oort, 2008). With regard to self-control, Stadler et al. (2016) reported that this factor was a significant predictor of GPA, but not of self-reported (subjective) academic achievement. Therefore, it may be productive for future research to use different instruments to measure these constructs in a more reliable way. In addition, using objective measures of academic achievement (e.g., GPA from official academic records) may help shed a clearer light on the associations between metacognitive self-regulation and self-control, on the one hand, and academic achievement, on the other hand.

Another important direction would be to use the findings uncovered in this study (about the relationships between self-regulation factors and use of learning strategies) as a springboard for future explorations of the interplay between self-regulation and learning strategy usage. For example, de Bruin and colleagues (2023) proposed that to adopt new learning strategies (e.g., desirable difficulties), students might need to override their existing (and potentially unproductive) learning habits and create new and effective routines for their learning; for a similar point of view, see Rea et al. (2022). Findings from our study suggest that this process could be appropriately supported by high levels of student engagement, which was a strong positive predictor of employing two relatively more effective learning strategies (i.e., elaboration and self-testing) and was unrelated to using two less effective strategies (i.e., rereading and cramming). Future research investigating this hypothesis could provide valuable new information to help students habitually use generally effective learning strategies.

We indicated earlier that procrastination may be associated with students’ decision to space versus cram learning sessions. Hence, future work could examine whether students’ procrastination patterns influence the nature and quality of students’ learning strategy usage, including but not limited to spacing/cramming. In addition, considering that “the regulation of learning can be based on a combination of information from both metacognition and affect” (Efklides et al., 2018, p. 66), future research could profitably investigate whether students’ use of different learning strategies is associated with significant differences in affective reactions during learning. For example, when using strategies that are perceived to make learning difficult (e.g., desirable difficulties) is associated with low levels of judgment of learning (de Bruin et al., 2023), students may feel frustrated or helpless (Efklides et al., 2018). Subsequently, future research could investigate whether key self-regulation factors examined in this research (e.g., effort regulation; metacognitive self-regulation) interact with task-related emotions to predict (a) the effort students devote to learning tasks; and/or (b) important cognitive aspects of their task engagement (e.g., continued use of the more difficult learning strategy vs. switching to a learning strategy that is construed as requiring less effort).

Yet another potentially productive direction for future research involves the relations between constellations of learning strategies and learning effectiveness/achievement. Specifically, future research could productively identify “benefits of combining study strategies” (Miyatsu et al., 2018, p. 401). For example, future work might probe for how students implement commonly endorsed strategies, such as rereading (e.g., they incorporate self-questioning to monitor comprehension or include retrieval practice to monitor learning) and highlighting/underlining (e.g., highlight during first reading or after the first reading – the latter is more effective). Doing so could better illuminate and provide more integrated understanding of when rereading (used with or without comprehension monitoring strategies or spacing strategies), highlighting/underlining (marking when first reading or after the first reading), and other common strategies are positively associated with long-term learning and achievement/grades. A related important direction for future research would examine “how individual differences may moderate the effectiveness of these [constellations of] study strategies. Research on this issue is sorely lacking” (Miyatsu et al., 2018, p. 401).

It is critical that future conceptual work on learning strategies and self-regulated learning provides further theoretical guidance on the factors and mechanisms that explain how students choose the learning strategies they employ and how they implement them in their study routines. Our findings provide important information for these future efforts. Specifically, results in the present study indicate that effort regulation and engagement were the most consistent predictors of students’ reported use of a set of learning strategies and achievement. Of importance, having high ability to regulate one’s effort and being engaged in one’s work at university reflect strong levels of conscientiousness; see, for example, Waldeyer et al. (2022).

Conscientiousness, which is a personality trait comprising several facets (e.g., industriousness, self-discipline, organization; Constantini & Perugini, 2016), exerts significant influences on critical psychological processes that shape learning and engagement with learning (Jach et al., 2023; Spielman et al., 2022; Waldeyer et al., 2022). Students who have high levels of conscientiousness strive for high achievement, make diligent progress toward attaining their learning goals, are organized in (as well as optimally manage) their efforts to learn, space their learning, and exhibit persistence and self-discipline in their study routines (Spielman et al., 2022; Theobald et al., 2018, 2021; Waldeyer et al., 2022). For example, in two studies of first-year university students, Waldeyer et al. (2022) found that conscientiousness positively predicted effort regulation, which, in turn, mediated the relationship between conscientiousness and student achievement. In a different strand of research, Theobald et al. (2018) found that for pre-service teachers, conscientiousness was positively related to distributing learning across a 12-week semester; results consistent with these were also reported by Theobald et al. (2021).

Future conceptual work mapping how conscientiousness influences students’ choice and use of learning strategies could profitably draw from leading models of self-regulated learning (see Panadero, 2017 for an extensive review) to theorize the roles conscientiousness plays in key phases of self-regulated learning (e.g., forethought and performance – Zimmerman, 2000; Zimmerman & Moylan, 2009; control – Winne & Hadwin, 1998; control of cognition, motivation/affect, behavior and context – Pintrich, 2000; control beliefs at the person level and regulation of effort at the task x person level – Efklides, 2011). To augment this conceptual work, future empirical research could examine how conscientiousness interacts with cognitive, metacognitive, motivational, and emotional factors/processes to predict students’ use of learning strategies. This work is critical because researchers have warned that failing to consider the role of conscientiousness in learning is likely to carry significant costs. Case in point, Waldeyer et al. (2022) stated that “providing learning strategy training to struggling students could set them up for failure if they do not have the requisite level of conscientiousness to make these strategies work” (pp. 8–9). Finally, future empirical work could aim to uncover critical mediators and moderators (e.g., coping strategies when encountering difficulties during learning, such as tenacity and flexibility; Sahdra et al., 2022) that link students’ conscientiousness (and/or its facets), self-regulation factors (e.g., effort regulation; engagement), use of learning strategies, and learning effectiveness and/or modulate their interrelationships.

Conclusion

This research examined the extent to which undergraduate students’ learning strategy usage was related to self-regulation factors that support learning and academic achievement. This investigation provides insight into key factors that underpin self-regulated learning and students’ use of learning strategies. In particular, the importance of two self-regulation factors, regulating effort and being enthusiastic about and immersed in (i.e., engaged with) one’s learning was highlighted. These self-regulation factors predicted the use of both various learning strategies and academic achievement. This study also found that students tend to use relatively more effective learning strategies somewhat more frequently than less effective strategies.

Competing interests

The authors declare that they have no competing interests of any nature that could be directly or indirectly related to the writing and publication of this article.

Appendix

Sample items for the various constructs, reliability, and number of items for each construct

Table 8 Sample items for the various constructs, reliability, and number of items for each construct