Self-Efficacy, Academic Motivation, and Self-Regulation: How Do They Predict Academic Achievement for Medical Students?

Abstract

Purpose

Self-efficacy, academic motivation, and self-regulation have been identified as important factors contributing to students’ learning success in general education. In the field of medical education, however, few studies have examined these variables or their interrelationships as predictors of undergraduate medical students’ learning outcomes, especially in the context of flipped learning.

Methods

Using structural equation modeling (SEM), this study explored the impact of self-efficacy on 146 first- and second-year medical students’ academic achievement in a flipped-learning environment, and whether such impact (if any) was mediated by academic motivation and self-regulated learning (SRL) strategies.

Results

On average, students scored highest on self-efficacy (mean = 5 out of a possible 7), followed by intrinsic motivation (mean = 4.59), resource-management strategies (mean = 4.48), metacognitive strategies (mean = 4.46), extrinsic motivation (mean = 4.24), and cognitive strategies (mean = 4.17). Our SEM results suggest that, while there was a direct effect of self-efficacy on learning outcomes, academic motivation and SRL strategies did not mediate it.

Conclusions

By unpacking the structural relationships among self-efficacy, academic motivation, SRL strategies, and learning outcomes, this study provides evidence-based support for the importance of promoting students’ self-efficacy in undergraduate medical flipped-learning environments. Strategies for increasing students’ self-efficacy are also discussed.

Introduction

Medical students’ motivational beliefs and their use of learning strategies are topics of increasing interest in the study of medical education [1]. While some studies have focused on medical students’ self-efficacy beliefs [e.g., 2, 3] and self-regulated learning (SRL) [e.g., 4, 5], others have looked into the relationships between these factors, such as the effects of participation, motivation, and learning strategies on learning performance [1], and self-efficacy’s links to learning strategies [6, 7]. Yet, few studies have comprehensively examined the structural relationship among self-efficacy, academic motivation, SRL strategies, and students’ academic achievement.

Self-efficacy refers to a person’s subjective assessment of his or her ability to carry out the actions needed to attain certain goals [8]. Amid increasing attention to students’ self-efficacy in the sphere of medical education [2], it has been found that this construct is significantly and positively correlated with medical students’ academic performance [e.g., 9, 10]; and such results echo findings regarding the positive direct effects of self-efficacy on academic achievement in general [11–13]. Nevertheless, non-medical education research has also suggested that self-efficacy should not be presumed to be the direct cause of academic achievement, as the latter is usually mediated by other psychological constructs, such as academic motivation, or students’ skills, such as SRL [14]. For example, Yusuf found that undergraduates’ academic accomplishments were both directly affected by self-efficacy, and indirectly influenced by academic motivation and SRL strategies. However, this phenomenon has never been tested among medical students.

Academic motivation has been found to be predictive of medical students’ academic success [15, 16]. According to self-determination theory [17], motivation can normally be divided into two types: intrinsic—i.e., motivation that originates from genuine interest—and extrinsic motivation, which originates from external factors. From least autonomous to most autonomous, these factors include external regulation (behavior aimed at satisfying an external demand), introjected regulation (internal pressure or feelings of guilt or shame), and identified regulation (valuing a behavioral goal as personally important) [18]. Kusurkar et al. [19] have suggested that medical students’ academic achievement will be better if their motivation is either intrinsic or a more autonomous variant of extrinsic.

SRL is commonly defined as “the degree to which students are metacognitively, motivationally, and behaviorally active participants in their own learning process” [20]. From a socio-cognitive perspective, Zimmerman [21] modeled three phases of SRL: forethought/planning, performance monitoring, and self-reflection. Other scholars have suggested that learners who most frequently utilize SRL strategies tend to have better learning outcomes [e.g., 22].

Researchers have also suggested that people can be trained to use SRL, and that medical educators should therefore scaffold students’ development of such skills [23, 24]. For example, this can be achieved by introducing different types of strategies to novice learners, showing how those strategies are used in self-directed learning, and explaining the optimal times to use particular types of strategies [25, 26]. The current study explores the structural relationship among self-efficacy, academic motivation, SRL strategies, and their effects on students’ academic achievement.

Methods

Context

This study took place in a U.S. midwestern University. At the start of the Fall semester of 2016, the medical school in this university introduced an innovative curriculum, which integrates basic science and clinical experience across all 4 years of the undergraduate medical-education program. It incorporates the flipped-classroom model, in which students are required to engage in self-directed learning of approximately 12 hours of materials before the start of every learning week, generally using learning materials posted online for them to read or watch in their own time. Face-to-face class time, meanwhile, is reserved for collective problem-solving and active discussions on topics around patients’ chief complaints and concerns. In January 2018, 380 first- and second-year undergraduate medical students were invited to participate in our online survey asking about their academic motivation, self-efficacy, and SRL strategies. The incentive for participation was that they would be entered into a drawing for a chance to win one of 20 $10 gift cards. The survey was open for 2 weeks, and we received 146 responses, representing a response rate of 38%. The research on which this manuscript was based was approved as an exempt study by the institutional review board of the university where it took place.

Measurement

The survey’s 56 items regarding students’ self-efficacy and SRL strategies (i.e., cognitive strategies, metacognitive strategies, and resource management) were adopted from the Motivational Strategies for Learning Questionnaire (MSLQ) [27]. We changed the wordings of several of these MSLQ items to fit our specific context. For example, “lectures” were changed to “class activities” since our curriculum did not have any lectures; and “in other class activities such as lecture and discussion” was changed to “in other class activities such as discussions, simulations, and anatomy.” The 28 survey items covering academic motivation, meanwhile, were adopted from Vallerand et al.’s [28] Academic Motivation Scale (AMS). All 84 items were responded to using a 7-point Likert scale, and the survey took approximately 25 min to complete. In addition, we collected the respondents’ scores on Comprehensive Basic Science Examination (CBSE) from the National Board of Medical Examiners (NBME) at the end of the Spring 2018 semester, as their academic-achievement outcome variable. An honest broker in our college (i.e., a neutral third party who was granted access to the data, but was not part of our research team) helped connect the survey data with achievement data, thus ensuring that all the information we received was deidentified and student information was kept confidential.

Data Analysis

To answer our research question, structural equation modeling (SEM) was performed to examine the structural relationships among learning outcomes, self-efficacy, academic motivation, and SRL strategies. SEM is a covariance/correlation-based multivariate analysis technique that incorporates multiple functions for investigating the unique influence of one variable on another, while accounting for the measurement errors of the variables and controlling for the influence of other related variables [29].

To examine the structural relationships within a latent-variable framework and assess the power of our model, aggregated means were used for subconstructs, including cognitive strategies, metacognitive strategies, resource-management strategies, intrinsic motivation, extrinsic motivation, and self-efficacy. Intrinsic motivation was aggregated from the responses to 12 items (Cronbach α = 0.82); extrinsic motivation from the responses to another 12 (Cronbach α = 0.67); and self-efficacy from the responses to seven items (Cronbach α = 0.92). Our latent construct, SRL strategies, was measured by three subconstructs: cognitive strategies (18 items, Cronbach α = 0.69), metacognitive strategies (12 items, Cronbach α = 0.71), and resource-management strategies (19 items, Cronbach α = 0.58). Model fit was assessed using absolute goodness-of-fit indices (χ²), parsimonious indices (root mean square error of approximation (RMSEA)), and incremental-fit indices (the comparative fit index (CFI) and Tucker-Lewis index (TLI)). Data cleaning and management were performed in R, and the model was fit using Mplus 8 [30].

Results

Descriptive Statistics

Table 1 presents descriptive statistics of each observed variable in the model. Overall, the sampled students had above-neutral scores (i.e., 3.5 or higher on the 7-point Likert scale) for motivation, self-efficacy, and SRL strategies. Among all variables, self-efficacy received the highest score, of 5.00 (SD = 1.11), followed by intrinsic motivation (M = 4.59, SD = 1.05), resource-management strategies (M = 4.48, SD = 0.78), metacognitive strategies (M = 4.46, SD = 0.75), and extrinsic motivation (M = 4.24, SD = 1.08). The students’ average Spring 2018 CBSE score was 56.41 (SD = 11.90) out of a possible 99, with the lowest individual score being 35, and the highest 90.

Table 1.

Descriptive statistics of the variables of interest

Variable	Mean	SD	Median	Min	Max
Extrinsic motivation	4.24	1.08	4.25	1.25	6.92
Intrinsic motivation	4.59	1.05	4.75	1.00	6.75
Self-efficacy	5.00	1.11	5.07	1.86	7.00
Cognitive strategies	4.17	0.83	4.20	1.95	6.30
Metacognitive strategies	4.46	0.75	4.50	2.42	6.67
Resource-management strategies	4.48	0.78	4.56	2.11	6.42
CBSE exam scores	56.41	11.90	55.00	35.00	90.00
N	146

The Relationships Among Self-Efficacy, SRL Strategies, Motivation, and Learning Outcomes

According to rule-of-thumb sample-size requirements [31, 32], the estimates in the model satisfied minimum statistical power. Table 2 shows the results of the proposed SEM model, including the parameter estimates of factor loadings and regression coefficients. All the model-fit indices consistently showed that the model fits the data very well: RMSEA = 0.063, 90% CI = [0.000, 0.125], close-fit test p = 0.325; χ² test statistics = 12.579 (df = 8, p = 0.127); CFI = 0.975, TLI = 0.945, and SRMR = 0.033. Figure 1 presents our proposed model’s factor loadings and structural relationships.

Table 2.

Structural equation modeling parameter estimates

	Indicator	Latent variable	Estimate (SE)	p value
Factor loading	Cognitive strategies	SRL strategies	0.588 (0.057)	< 0.001
	Metacognitive strategies	SRL strategies	0.509 (0.050)	< 0.001
	Resource-management strategies	SRL strategies	0.393 (0.055)	< 0.001
Regression coefficient	Extrinsic motivation	SRL strategies	− 0.082 (0.103)	0.425
	Intrinsic motivation	SRL strategies	0.548 (0.120)	< 0.001
	Self-efficacy	SRL strategies	0.173 (0.094)	0.067
	SRL strategies	Test score	− 1.663 (1.093)	0.128
	Self-efficacy	Test score	2.517 (0.940)	0.007
	Extrinsic motivation	Test score	− 1.042 (1.027)	0.310
	Intrinsic motivation	Test score	2.423 (0.942)	0.010
Residual variance	Test score		128.761 (15.226)	< 0.001
	Cognitive		0.198 (0.050)	< 0.001
	Metacognitive		0.195 (0.041)	< 0.001
	Resource-management		0.395 (0.052)	< 0.001

SE, standard error; SRL, self-regulated learning

Fig. 1.

Structural equation model of the structural relationship among motivation, self-efficacy, and self-regulated learning (SRL) strategies, and their effects on test scores

The model indicated that intrinsic motivation was positively associated with SRL strategies (β = 0.548, p < 0.001), but that extrinsic motivation (β = − 0.082, p = 0.425) and self-efficacy (β = 0.173, p = 0.067) had no significant association with such strategies. In addition, self-efficacy was positively and significantly associated with students’ test scores (β = 2.423, p = 0.010), whereas no significant associations were found between CBSE scores and SRL strategies (β = − 1.663, p = 0.142), extrinsic motivation (β = − 1.042, p = 0.310), or intrinsic motivation (β = − 1.215, p = 0.340).

Discussion

The present study found that self-efficacy was the only direct predictor of medical students’ academic achievement in a flipped-classroom environment. This suggests that the more strongly students believe in their ability to carry out actions that will help them attain academic goals, the higher their test scores will be. This tends to corroborate previous studies’ findings of a significant positive relationship between self-efficacy and medical students’ academic achievement [9, 10].

However, the current study did not find that either self-regulation or motivation played any mediating role in the relationship between self-efficacy and academic achievement. Studies in nonmedical fields have suggested some other mediators of the effect of self-efficacy on learning outcomes. For example, based on analysis of 797 Spanish secondary-education students’ self-efficacy and academic achievement, one study reported that the direct effect of self-efficacy on academic achievement was mediated by the participants’ expectancy-value beliefs, including their achievement expectations, their perceptions of the value of the subject matter, their process expectations regarding the teacher, and the expected cost in effort of passing the subject [33]. Similarly, Yusuf [14] examined undergraduates’ self-efficacy, motivation, and SRL strategies, and reported that self-efficacy had a direct effect on academic achievement, and that it also mediated the influences of motivation and learning strategies on such achievement.

Our findings suggest that, if undergraduate medical students are to improve their academic achievement, it is important to help them increase their self-efficacy. Prior research has suggested that self-efficacy can be increased using certain strategies, such as ascribing students’ success or failure to effort instead of ability; providing peer-observation opportunities to boost students’ beliefs that they can learn equally well; ensuring that support from trustworthy sources (e.g., teachers) is available; and providing feedback about progress and giving rewards based on it [34].

Although in the present study, neither academic motivation nor SRL strategies were found to be a significant predictor of medical students’ test scores, this could easily have been due to our small sample size, and should not be taken as positive evidence that either of these factors is unimportant to such students’ learning. Motivation—especially when it is intrinsic, or extrinsic but relatively autonomous—has long been associated with better learning outcomes among medical students [15, 19]; and students’ motivation can be shaped in such directions by autonomy-supportive learning environments. A movement in contemporary medical education away from didactic, lecture-based curricula and toward student-centered, problem-based learning is believed to reflect the need for autonomy as well as competence [35, 36]. To meet the need for competence, medical educators can set milestones to make their expectations more explicit and provide coaching that encourages reflection on performance and identifies areas for future improvement [35]. To meet the need for autonomy, teachers can allow students to peer-teach [37], identify learning objectives and resources that meet individual needs [35], and create individualized technology-assisted assessments that both facilitate and endorse the value of individually tailored learning paths as a replacement for traditional exams [36]. Medical faculty could also assign more research projects to their students and encourage them to present their results on authentic platforms, such as by co-authoring journal articles or presenting at conferences [38].

Similarly, despite our study not detecting any significant impact of SRL strategies on students’ academic achievement, we would not seek to deny that such strategies are critical to student learning, especially in our flipped-learning context. To help students develop SRL strategies, educators should help them set clear, specific, and attainable goals [23]; encourage them to adopt appropriate cognitive strategies such as paraphrasing, organizing, integrating, and positive self-talk [39–41]; aid their development of effective time-management skills [42]; provide peer learning or coaching opportunities [37, 43]; and assist their development of self-reflection skills [23].

This study is not without limitations. Its low response rate of 38%, and small sample size of 146 participants, could have resulted in insufficient statistical power, and thus to the unexpected finding that the effects of academic motivation and self-regulation on learning outcomes were both insignificant. It is also possible that students who chose to participate in the study and completed the survey were those who tended to have high motivation and self-efficacy, and to use more SRL strategies in their learning. This could have led to selection bias in our sample. Second, its dependent variable consisted solely of students’ test scores from a single time point; as such, future research could usefully collect longitudinal data to examine whether and how self-efficacy, motivation, and SRL strategies affect students’ academic achievement in a flipped medical classroom over time.

In conclusion, this study has broken new ground by unpacking the structural relationships among self-efficacy, academic motivation, SRL strategies, and learning outcomes in an undergraduate medical school flipped-learning environment. Given that more and more medical schools are transitioning from traditional lecturing to a flipped-classroom model for the education of the next generation of healthcare professionals, this study provides important evidence that self-efficacy significantly predicts medical students’ academic achievement in flipped-learning environments.

Authors’ Contributions

BZ has contributed to study design, data collection, data analysis, drafting, and finalization of the manuscript. CC contributed to data analysis, drafting, and finalization of the manuscript. CL contributed to data analysis, drafting, and finalization of the manuscript. YZ contributed to data collection, drafting, and finalization of the manuscript.

Data Availability

Upon request

Compliance with Ethical Standards

Conflict of Interest

The authors declare that they have no conflict of interest.

Code Availability

Not applicable